© 2011 by Taylor and Francis Group, LLC

GREEN
NANOTECHNOLOGY
Solutions for Sustainability and
Energy in the Built Environment

Geoffrey B. Smith
Claes G. Granqvist

Boca Raton London New York

CRC Press is an imprint of the

Taylor & Francis Group, an informa business

© 2011 by Taylor and Francis Group, LLC

 CRC Press
Taylor & Francis Group
6000 Broken Sound Parkway NW, Suite 300
Boca Raton, FL 33487-2742

© 2011 by Taylor and Francis Group, LLC

CRC Press is an imprint of Taylor & Francis Group, an Informa business

No claim to original U.S. Government works

Printed in the United States of America on acid-free paper

10 9 8 7 6 5 4 3 2 1

International Standard Book Number: 978-1-4200-8532-7 (Hardback)

This book contains information obtained from authentic and highly regarded sources. Reasonable efforts
have been made to publish reliable data and information, but the author and publisher cannot assume
responsibility for the validity of all materials or the consequences of their use. The authors and publishers
have attempted to trace the copyright holders of all material reproduced in this publication and apologize to
copyright holders if permission to publish in this form has not been obtained. If any copyright material has
not been acknowledged please write and let us know so we may rectify in any future reprint.

Except as permitted under U.S. Copyright Law, no part of this book may be reprinted, reproduced, transmit-
ted, or utilized in any form by any electronic, mechanical, or other means, now known or hereafter invented,
including photocopying, microfilming, and recording, or in any information storage or retrieval system,
without written permission from the publishers.

For permission to photocopy or use material electronically from this work, please access www.copyright.
com (http://www.copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC), 222 Rosewood
Drive, Danvers, MA 01923, 978-750-8400. CCC is a not-for-profit organization that provides licenses and
registration for a variety of users. For organizations that have been granted a photocopy license by the CCC,
a separate system of payment has been arranged.

Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used
only for identification and explanation without intent to infringe.

Library of Congress Cataloging‑in‑Publication Data

Smith, Geoffrey B. (Geoffrey Burton)

Green nanotechnology : solutions for sustainability and energy in the built
environment / authors, Geoffrey B. Smith and Claes-Goran S. Granqvist.
p. cm.
“A CRC title.”
Includes bibliographical references and index.
ISBN 978-1-4200-8532-7 (alk. paper)
1. Sustainable buildings--Design and construction. 2. Nanotechnology. 3.
Buildings--Electric equipment. 4. Buildings--Energy conservation. 5. Sustainable
construction. 6. Electronic apparatus and appliances. I. Granqvist, Claes G. II. Title.

TH880.S65 2010
690.028’6--dc22 2010007968

Visit the Taylor & Francis Web site at

http://www.taylorandfrancis.com

and the CRC Press Web site at

http://www.crcpress.com

© 2011 by Taylor and Francis Group, LLC

v
vi Contents
 Contents vii
viii Contents
 Contents ix
x Contents
 Contents xi
xii Contents

8.3 Thermal Insulation with Nanomaterials 378

8.3.1 Thermal Conductance of Porous and
Nanoporous Materials 379
8.3.2 Vacuum Insulation Panels 381
8.3.3 Silica Aerogel 382
8.4 Green Energy Storage 383
8.4.1 Energy Storage: Survey of a “Missing
Link” 383
8.4.2 Electrical Storage Using
Electrochemistry 385
8.4.3 Electrochemical Super- and
Ultracapacitors 387
8.4.4 Nanomaterials for Advanced Batteries 389
References 393

Chapter 9 Conclusions: Nanotechnologies for a

Sustainable Future 399
9.1 Energy and the Future 399
9.2 New Technologies and Growing Uptake of
Proven Technologies 402
9.3 Toward a “Nanoworld” 403
References 410
Appendix 1: Thin-Film Deposition 413
Appendix 2: Abbreviations, Acronyms, and Symbols 429
Index 437
xiii
xiv Preface
 Preface xv
xvi Preface
 Preface xvii
xviii Preface
 Preface xix
xx Preface
 Preface xxi
xxii Preface
 Preface xxiii
1
2 Green Nanotechnology
 Green Nanotechnology 3
4 Green Nanotechnology
 Green Nanotechnology 5
6 Green Nanotechnology
 Green Nanotechnology 7
8 Green Nanotechnology
 Green Nanotechnology 9
10 Green Nanotechnology
 Green Nanotechnology 11
12 Green Nanotechnology
 Green Nanotechnology 13
14 Green Nanotechnology
 Green Nanotechnology 15
16 Green Nanotechnology
 Green Nanotechnology 17
18 Green Nanotechnology
 Green Nanotechnology 19
20 Green Nanotechnology
 Green Nanotechnology 21
22 Green Nanotechnology
 Green Nanotechnology 23
24 Green Nanotechnology

36. IPCC, Climate Change 2007: Mitigation. Contribution of Working

Group III to the Fourth Assessment Report of the Intergovernmental
Panel on Climate Change, edited by B. Metz, O. R. Davidson, P. Bosch,
R. Dave, L. A. Meyer, Cambridge University Press, Cambridge, U.K.
and New York, 2007.
37. M. Kolokotroni, Y. Zhang, R. Watkins, The London heat island and
building cooling design, Solar Energy 81 (2007) 102–110.
38. World Energy Outlook 2009, International Energy Agency, Paris,
France, 2009; http://www.worldenergyoutlook.org.
39. G. Pearman, personal communication. Unpublished plot based on
the approach in R. Socolow, S. Pacala, J. Greenblatt, Wedges: Early
mitigation with familiar technology, in Proceedings of GHGT-7,
the Seventh International Conference on Greenhouse Gas Control
Technology, Vancouver, Canada, September 5–9, 2004; see also
http://www.princeton.edu/~cmi/.
40. J. J. Louviere, D. J. Street, L. Burgess, A 20+ years’ retrospective on
choice experiments, in Marketing Research and Modeling: Progress
and Prospects, edited by Y. Wind, P. E. Green, Kluwer Academic,
New York, 2003, chap. 8, pp. 201–214.
41. P. Eichholtz, N. Kok, J. M. Quigley, Doing well by going good? Green
office buildings, Center for the Study of Energy Markets, Berkeley,
CA, Working Paper CSEM WP-192 (2009); http://www.ucei.berke-
ley.edu/PDF/csemwp192.pdf.
42. L. Heshong, R. L. Wright, S. Okura, Daylighting impact on retail
sales performancs, J. Illum. Engr. Soc. 31 (2002) 21–25; Daylighting
impacts on human performance in school, J. Illum. Engr. Soc. 31
(2002) 101–114.
43. EIA, International Energy Outlook 2007, Energy Information
Administration, US Department of Energy, Washington, DC, 2007;
http://www.eia.doe.gov/emeu/international/energyconsumption.
html.
25
26 Green Nanotechnology
 In Harmony with the Environment 27
28 Green Nanotechnology
 In Harmony with the Environment 29
30 Green Nanotechnology
 In Harmony with the Environment 31
32 Green Nanotechnology
 In Harmony with the Environment 33
34 Green Nanotechnology
 In Harmony with the Environment 35
36 Green Nanotechnology
 In Harmony with the Environment 37
38 Green Nanotechnology
 In Harmony with the Environment 39
40 Green Nanotechnology
 In Harmony with the Environment 41
42 Green Nanotechnology
 In Harmony with the Environment 43
44 Green Nanotechnology
 In Harmony with the Environment 45
46 Green Nanotechnology
 In Harmony with the Environment 47
48 Green Nanotechnology
 In Harmony with the Environment 49
50 Green Nanotechnology
 In Harmony with the Environment 51
52 Green Nanotechnology
 In Harmony with the Environment 53
54 Green Nanotechnology
 In Harmony with the Environment 55
56 Green Nanotechnology
 In Harmony with the Environment 57
58 Green Nanotechnology
 In Harmony with the Environment 59
60 Green Nanotechnology
 In Harmony with the Environment 61
62 Green Nanotechnology
 In Harmony with the Environment 63
64 Green Nanotechnology
 In Harmony with the Environment 65
66 Green Nanotechnology

18. C. M. Lampert, C. G. Granqvist (eds.), Large-Area Chromogenics:

Materials and Devices for Transmittance Control, SPIE–The
International Society for Optical Engineering, Bellingham, WA,
1990.
19. M. Leijon, H. Bernhoff, M. Berg, O. Ågren, Economical con-
siderations of renewable electric energy production: Especially
development of wave energy, Renewable Energy 28 (2003)
1201–1209.
20. M. Leijon, A. Skoglund, R. Waters, A. Rehn, M. Lindahl, On the
physics of power, energy and economics of renewable energy sources:
Part I, Renewable Energy 35 (2010) 1729–1734.
21. F. L. Curzon, B. Ahlborn, Efficiency of a Carnot engine at maximum
power output, Am. J. Phys. 43 (1975) 22–24.
67
68 Green Nanotechnology
 Optical Materials Science for Green Nanotechnology 69
70 Green Nanotechnology
 Optical Materials Science for Green Nanotechnology 71
72 Green Nanotechnology
 Optical Materials Science for Green Nanotechnology 73
74 Green Nanotechnology
 Optical Materials Science for Green Nanotechnology 75
76 Green Nanotechnology
 Optical Materials Science for Green Nanotechnology 77
78 Green Nanotechnology
 Optical Materials Science for Green Nanotechnology 79
80 Green Nanotechnology
 Optical Materials Science for Green Nanotechnology 81
82 Green Nanotechnology
 Optical Materials Science for Green Nanotechnology 83
84 Green Nanotechnology
 Optical Materials Science for Green Nanotechnology 85
86 Green Nanotechnology
 Optical Materials Science for Green Nanotechnology 87
88 Green Nanotechnology
 Optical Materials Science for Green Nanotechnology 89
90 Green Nanotechnology
 Optical Materials Science for Green Nanotechnology 91
92 Green Nanotechnology
 Optical Materials Science for Green Nanotechnology 93
94 Green Nanotechnology
 Optical Materials Science for Green Nanotechnology 95
96 Green Nanotechnology
 Optical Materials Science for Green Nanotechnology 97
98 Green Nanotechnology
 Optical Materials Science for Green Nanotechnology 99
100 Green Nanotechnology
 Optical Materials Science for Green Nanotechnology 101
102 Green Nanotechnology
 Optical Materials Science for Green Nanotechnology 103
104 Green Nanotechnology
 Optical Materials Science for Green Nanotechnology 105
106 Green Nanotechnology
 Optical Materials Science for Green Nanotechnology 107
108 Green Nanotechnology
 Optical Materials Science for Green Nanotechnology 109
110 Green Nanotechnology
 Optical Materials Science for Green Nanotechnology 111
112 Green Nanotechnology
 Optical Materials Science for Green Nanotechnology 113
114 Green Nanotechnology
 Optical Materials Science for Green Nanotechnology 115
116 Green Nanotechnology
 Optical Materials Science for Green Nanotechnology 117
118 Green Nanotechnology
 Optical Materials Science for Green Nanotechnology 119
120 Green Nanotechnology
 Optical Materials Science for Green Nanotechnology 121
 Chapter 4
Visual Indoors–Outdoors
Contact and Daylighting
Windows

This chapter will bring us deep into green nanotechnology by consider-

ing a range of materials—often in the form of thin films with nanofea-
tures—for controlling the radiation going in or out through windows.
“Windows” should be taken in a wide sense to include not only apertures
in domestic buildings but also transparent façades such as those com-
monly used in modern commercial buildings. And the windows do not
have to be integrated in walls; they can be skylights or any other opening
in the envelope in a building. Figure 4.1 serves as a point of departure
and shows one emerging type of windows, referred to as “smart” or
“switching” windows. These are capable of changing the throughput of
visible light and solar energy, thereby giving the user of the building a
comfortable indoor climate, as well as minimizing energy for air condi-
tioning while preserving visual indoors–outdoors contact.
This chapter discusses many aspects of the “view window” capable
of giving both visual indoors–outdoors contact and daylighting. Chapter
5 on luminaires then continues the treatment of light in buildings, and
covers artificial light and daylighting, primarily with systems that do not
provide visual contact between the two sides.
In fact, (day)lighting and visual contact are two functions that are
intertwined in a fascinating and poorly understood way. The dichot-
omy even appears in the very term for this multifaceted object. Thus,
the English window literally means “wind eye” (from vindauga in Old
Norse), and in this way is connected to vision, whereas the correspond-
ing term is Fenster in German and fenêtre in French both going back
to Latin fenestra meaning to provide light. Incidentally, there are other
ways to represent the meaning of a window, such as the Greek παράθυρο,
literally meaning “opening near a door.”

123

© 2011 by Taylor and Francis Group, LLC

 124 Green Nanotechnology

Figure 4.1  Prototype of a “smart” window with four panes, two of

which are dark and the other two transparent. The properties of the
panes can be changed, gradually and reversibly, during the course of
some tens of seconds. The smart functionality is obtained via a poten-
tially low-cost foil that can serve as a lamination material and also pro-
vide mechanical ruggedness and acoustic damping.

4.1  General Introduction

Architectural windows and glass façades are problematic from an

energy perspective, and the same is true for windows used in cars,
trucks, and other vehicles. Nevertheless, the windows are absolutely
essential for the users of a building and provide them with contact to
the surrounding world. So diminishing window areas to the extreme
may be good with regard for energy, but it is generally highly unde-
sirable for the persons who spend their time in the building. Indeed,
the primary function of fenestration—to provide unmitigated visual
contact between indoors and outdoors, as well as daylighting—always
must be kept in mind so that window areas are not made too small.
Psychological well-being and health benefits associated with visual con-
tact to the ambience and with daylight are issues whose importance is
being increasingly recognized.
In a practical situation, there are normally unwanted energy flows
with too much energy leaving or entering buildings via their windows,
with concomitant demand for energy guzzling heating and cooling.
Supplying daylight without glare can also save a lot of lighting energy
and some thermal energy. Present architectural trends are to increase

© 2011 by Taylor and Francis Group, LLC

 Visual Indoors–Outdoors Contact and Daylighting 125

the window areas, and hence the energy issue may become even more
pressing in the future.
This increase is especially marked with regard to roof glazing for
homes, large buildings, and in atria attached to homes in cold climates.
The average solar flux onto a roof is several times that onto a vertical
façade, while heat loss via a glass roof during the winter is also much
larger than for a wall. Glare can also be a major problem, though the
daylighting benefits from roof windows are marvelous if these problems
can be managed. Many building codes severely limit the area of roof
glazing apertures unless special glazing is used. Glazings with variable
optical properties—as illustrated in Figure 4.1 and discussed in the text
following—is an ideal solution for roof windows since their problems
vary with time and solar irradiation intensity, and such glazings can
control heat flows as well as glare. Roof glazings are also combined
increasingly with solar cells for power generation; these solar cells can
reduce glare and solar heat gain while still admitting plenty of daylight.
The solar cells, which are essentially opaque, may create illuminance
patterns inside the building, which may be advantageous or problem-
atic depending on the specific building; if such patterns are unwanted,
they can be removed by having diffusing glass in the roof windows.
Figure 6.1c shows a glazing with integrated solar cells.

4.1.1  Strategies for Energy Efficiency

Several strategies can be used to improve the windows’ energy perfor-

mance. All of them are based on the rational use of energy in our ambi-
ence, as discussed at length in Chapter 2. Three main strategies are that

• Spectral selectivity can be used to let in only the desired radia-

tion whereas other radiation is removed or at least diminished.
• Angular properties can be taken advantage of, typically by
maintaining good transparency along a desired line-of-sight
while preventing radiation from other directions.
• The properties of windows can be varied in response to an exter-
nal stimulus, so that visible light and solar energy are introduced
or excluded depending on the need; materials capable of such
dynamic performance are referred to as chromogenic.

The green nanotechnologies of particular importance for these strategies

are discussed in detail below. But before we do so, it is important to have
some basic knowledge about the properties of normal window glass and
about plastic foils that are used in windows.

© 2011 by Taylor and Francis Group, LLC

 126 Green Nanotechnology

A note on terminology is useful for avoiding confusion. The thin sheet

of plastic, typically a fraction of a millimeter in thickness, that can be used
for upgrading windows are referred to here as “window foil,” whereas the
term window film is also commonly used. We prefer the former terminol-
ogy and reserve the term film for the surface coatings—typically less than
a micrometer thick—that can be applied to the surface of a glass pane or
a foil, or any other substrate, in order to change its optical and electrical
performance and, as a consequence, improve its properties.

4.1.2  Uncoated Glass and Plastic Foil

Most windows use glass panes. In principle, there are many different
glass types, but float glass is employed in the great majority of all build-
ings—at least in the industrialized world—and is a highly standardized
product [1,2].
Float glass is made by a process in which the glass melt is solidified
on the surface of molten tin. As a result, this glass is characterized by
uniformity and flatness almost on the atomic scale, which is important
for its usefulness as a substrate for thin-film deposition. The middle curve
in Figure 4.2 shows the spectral transmittance of 6-mm-thick standard
float glass within the wavelength range relevant to solar radiation [3]. A
characteristic absorption feature at λ ≈ 1 μm is clearly seen and limits Tsol

100

Low Fe2O3
80
Transmittance (%)

60 Standard clear

40

20
High Fe2O3
0
0.3 0.5 1 2
Wavelength (µm)

Figure 4.2  Spectral transmittance for float glass with three different
amounts of Fe2O3. The upper two curves refer to 6-mm-thick panes and
the lower curve to a 4-mm-thick pane. (From C. G. Granqvist, Materials
Science for Solar Energy Conversion Systems, edited by C. G. Granqvist,
Pergamon, Oxford, U.K., 1991; chap. 5, pp. 106–167. With permission.)

© 2011 by Taylor and Francis Group, LLC

 Visual Indoors–Outdoors Contact and Daylighting 127

to some noticeable extent. This absorption is due to iron oxide, which is

often added to the raw material of the glass in order to make the glass
produced in one plant indistinguishable from that made in other plants.
Special float glass is available with varying amounts of Fe oxide, and
Figure 4.2 shows that low Fe content leads to glass with very high values
of Tlum and Tsol, and with substantial transmittance also in the UV. These
properties are important for several different solar-energy-related applica-
tions such as for solar collectors for hot water production (cf. Figure 6.2).
Glass with a high Fe content limits Tsol, while Tlum remains rather large;
such glass has a greenish tint, which becomes increasingly distinct at large
thickness. Other metal additions can provide different colors.
The reflectance of each interface between glass and air is about 4%
in the 0.4 < λ < 2 μm range, implying that the maximum transmittance
for a glass pane is 92%. Glass is strongly absorbing for λ > 3 μm, and
E therm is as large as about 87%. In other words, uncoated glass is almost
a blackbody radiator.
The production of float glass is enormous and currently (2009)
amounts to more than 4 billion m 2 per year, with an overall annual
growth of some 6%. This glass is used for different purposes, with 37%
for residential buildings, 23% for commercial buildings, and 27% for
vehicles. Some 1.8 billion m 2 is produced in Asia, whereas the numbers
for the Americas and Europe are 1.0 and 0.9 billion m 2 , respectively. The
total value is expected to be about US$ 8 billion in 2010. Considering
these stunning numbers, and the large role of buildings in the world’s
use of primary energy, one conclusion to draw is that the potential for
energy savings is huge for any technology that can significantly improve
the properties of windows.
Polymers are also of great importance for windows, and transparent
polymers can, in principle, replace glass. However, polymers degrade
more easily than glass and hence have fewer applications. Polyethylene
terephthalate (PET) foil deserves special mention since it can be coated
and applied to windows in different ways: glued onto individual glass
panes, suspended between glass panes, or integrated in laminates joining
glass plates. Such foil can be retrofitted to existing windows or applied
in new windows. An alternative to PET is polyethylene naphthalate
(PEN) and some fluorocarbon plastics such as ethylene tetrafluoroeth-
ylene (ETFE). The latter plastics are less temperature sensitive and more
UV resistant—but also more expensive—than PET. ETFE is employed
in contemporary “membrane architecture,” which we will touch upon
in Chapter 9.
Foils comprising many alternating layers make it possible to achieve
interesting optical effects. If the difference in refractive index among the
layers, as well as the number of layers, is large it is possible to obtain high
reflectance in a prescribed wavelength range, as can be understood from

© 2011 by Taylor and Francis Group, LLC

 128 Green Nanotechnology

thin-film optics [4]; the principles for the optical functionality are dis-
cussed briefly in Section 5.3. Figure 4.3 shows data for two commercially
available foils with different values of Tlum constructed in this way and
backed by glass [5]. Obviously, the reflectance covers a significant part
of the infrared solar radiation, specifically the 0.85 < λ < 1.2 μm range,
where the reflectance from the foil side is ~50%, and hence such a foil can
be used to limit the energy inflow through windows. In the thermal infra-
red, the multilayer foil behaves as a homogeneous one and E therm is high.
Multilayer foils can be manufactured by coextrusion—which in principle
is a very low-cost technique—and have been commercialized recently.
Laminated glass may be used for safety and other reasons. Such glass
normally has a layer of polyvinyl butyral (PVB) sandwiched between two
glass panes and bonded to the glass under heat and pressure. Laminated
glass is of growing importance since it has to be used for safety reasons
in high-rise buildings, and such buildings are the only option for hous-
ing the world’s growing population, more than half of which lives in big
cities today (2010). Car windscreens always use PVB laminates in order
to avoid glass splinters in the case of breakage. The PVB is an efficient
absorber of UV light, and the transmittance at λ < 0.38 μm is almost
zero for a layer thickness of the order of 1 mm.

4.2  Spectral Selectivity:

The Energy Efficiency That Is Possible

Consider a window with two or more vertical panes separated by air

gaps. Generally speaking, the heat transfer should be low so that a desired
indoor temperature can be maintained irrespective of the outdoor tem-
perature being higher or lower than this chosen “comfort temperature.”
Heat transfer takes place via three mechanisms: radiation between the
glass surfaces, conduction across the air gap, and convection (moving
gas) in the air gap. The heat transfer can be decreased by minimizing
each of these transfer mechanisms. The one that is easiest to influence
is the radiative part, which can be cut back by coating at least one of
the glass panes so that it does not emit as much thermal radiation as the
blackbody-like bare glass surface.
What is the energy savings one can accomplish by having glass with
low thermal emittance? To answer this question we now discuss some
computations of the heat transfer in a window with one, two, or three
vertical panes separated by air gap(s) of 12.7 mm, which is a typical
value for insulated glass units [3]. The glass surfaces are designated by
consecutive numbers, with the outside surface labeled 1. One of the sur-
faces is taken to have an emittance in the zero to 85% range. The upper
value is characteristic for uncoated glass, as pointed out above, so the

© 2011 by Taylor and Francis Group, LLC

 Visual Indoors–Outdoors Contact and Daylighting 129

Prestige-PR70
1
T
0.9 Rf
Rb
0.8
T
0.7

0.6
Intensity

Rf
0.5

0.4
0.3 Rb

0.2

0.1

0
0 0.5 1 1.5 2 2.5
Wavelength [µm]

Prestige-PR40
1
T
0.9 Rf
Rb
0.8

0.7

0.6
Intensity

Rf
0.5
T
0.4
0.3
0.2 Rb

0.1

0
0 0.5 1 1.5 2 2.5
Wavelength [µm]

Figure 4.3  Spectral transmittance of two commercial foils comprising

multilayer polymers and backed by glass. Data are given on transmit-
tance T and reflectance from the front (foil) side Rf and back (glass)
side Rb. Upper data correspond to Tlum = 69.3% and Tsol = 36.5%; lower
data correspond to Tlum = 39.5% and Tsol = 21.9%. (From International
Glazing Database (2009), Lawrence Berkeley National Laboratory, CA;
http://windows.lbl.gov/software/window/window.html.)

© 2011 by Taylor and Francis Group, LLC

 130 Green Nanotechnology

computations illustrate the role of having one surface with a low value
of E therm. The calculations assumed a low outside temperature and that
there were thermal losses due to wind. Data are shown in Figure  4.4.
The upper two curves indicate that if the emittance of one of the surfaces
in a single pane window is lowered there is a minor drop in the heat
transfer (thermal conductance) from an initial value of about 6 Wm−2K−1,
the effect being largest when the inner surface has a low emittance. For
double glazing, the heat transfer can drop from about 2.8 to 1.4 Wm−2K−1
when either of the surfaces facing the air gap has a low value of E therm.
For triple glazing, finally, the corresponding improvement is from 1.8 to
1.2 Wm−2K−1. The most significant result with regard to a typical double-
pane window is that the heat transfer can be cut to almost half by the use
of a surface treatment giving a low thermal emittance.

6
1
Single glass

5 2
Thermal Conductance (Wm–2 K–1)

3 Double glass
1
4
2 or 3
2 1

6 Triple glass
4 or 5
1

0
0 20 40 60 80
Emittance (%)

Figure 4.4  Computed thermal conductance for different window

constructions with one glass surface having low emittance. The glass
surfaces are numbered consecutively with the outermost surface desig-
nated 1. (From C. G. Granqvist, Materials Science for Solar Energy
Conversion Systems, edited by C. G. Granqvist, Pergamon, Oxford,
U.K., 1991; chap. 5, pp. 106–167. With permission.)

© 2011 by Taylor and Francis Group, LLC

 Visual Indoors–Outdoors Contact and Daylighting 131

Detailed computations of the thermal performance of windows are

technical, and rather complicated, issues which we do not wish to dis-
cuss here. Detailed information is available in the literature [6]. This
literature also describes the role of gas fillings to diminish the thermal
conductance still further, problems associated with thermal leaks at the
frames in the windows, etc. Generally speaking, the effect of a low emit-
tance treatment is accentuated when conductive and convective trans-
fers are diminished. A judiciously chosen window design, with vacuum
between the panes in a triple-layer construction and with low-emittance
treatment, has given a thermal conductance coefficient as low as ~0.2
Wm−2K−1 [7]. That number probably gives a good estimate for what is
feasible for the thermal insulation in practical fenestration. Vacuum
insulation is discussed later in Section 8.3.
Controlling thermal conductance in a window is only one of the
important tasks. Another, obviously, is related to the transmission of
solar energy and visible light. As emphasized in Chapter 2, only about
half of the solar energy falls in the spectral range for which the eye is
sensitive. Since Figure 4.2 showed that regular float glass lets through
most of the solar energy, the energy inflow through a window can be
about twice as high as the one needed to provide full daylighting or, in
other words, the radiative energy inflow through a window, in principle,
can be cut in half without affecting the windows’ primary functions of
providing visual contact as well as daylighting.
We now return to the discussion of the ambient radiative properties
in Chapter 2 and repeat, mainly for convenience, some of the essential
facts. The upper part of Figure 4.5 shows the most relevant aspects of
the ambient radiation: the solar irradiance after the radiation has gone
through one air mass, the luminous efficiency of the eye, and blackbody
spectra for some different temperatures [8]. As noted before, the black-
body spectra define the spectral ranges for thermal radiation.
The lower part of Figure 4.5 shows ideal properties of a coating that
transmits in the solar range but does not emit thermal radiation. This is
known as a “low emittance coating” (“low-E”) and is characterized by

T(λ) = 1 for 0.4 < λ < 3 μm (4.1a)

R(λ) = 1 for 3 < λ < 50 μm (4.1b)

The other idealized coating specified in Figure 4.5, which transmits only

visible light, defines a “solar control coating” characterized by

T(λ) = 1 for 0.4 < λ < 0.7 μm (4.2a)

R(λ) = 1 for 0.7 < λ < 50 μm (4.2b)

© 2011 by Taylor and Francis Group, LLC

 132 Green Nanotechnology

(GW/m3) Relative luminous

efficiency of the eye Blackbody
1.5 spectra
Solar spectrum (AM1) (MW/m3)
1.0 + 50°C 50
± 0°C
0.5 – 50°C

0 0

Transmittance
Exp.
Solar Low emittance
0
1 control coating
coating

Reflectance
Exp.
0

0.2 0.5 1 2 5 10 20 50
Wavelength (µm)

Figure 4.5  Upper part shows relative luminous efficiency for the eye,
a solar spectrum for one air mass (AM1), and blackbody spectra for
three temperatures. Lower parts show transmittance and reflectance for
idealized window coatings designed for “solar control” and for “low
emittance.” Dotted curves refer to experimental data to be discussed
in Section 4.4. (From I. Hamberg, C. G. Granqvist, J. Appl. Phys. 60
(1986) R123–R159. With permission.)

Both of the spectral profiles were given earlier in Figure 2.16. We empha-

size that the spectral selectivity inherent in the relationships is idealized,
and the dashed curves in the lower part of Figure 4.5 illustrate some typi-
cal experimental results. Practical ways to accomplish spectral optical data
that approximate Equations 4.1 and 4.2 will be discussed at length next.

4.3  Spectral Selectivity of

Noble-Metal-Based Films

Very thin films of free-electron-like (“noble”) metals can combine lumi-

nous transmittance with infrared reflectance and electrical conductivity, as
has been known for a long time. Thus, they can have properties that more

© 2011 by Taylor and Francis Group, LLC

 Visual Indoors–Outdoors Contact and Daylighting 133

or less resemble those in Equations 4.1 and 4.2 and, therefore, provide a
good starting point for development of spectrally selective window coat-
ings. Very thin layers of Cu, Ag, and Au are of particular interest; alterna-
tive materials are TiN and ZrN as well as Pt and Al. Whereas Ag-based
films have unsurpassed optical properties (cf. Section 3.7), Au-based films
may be of interest under corrosive or otherwise chemically demanding
conditions. The nitrides are notable for their hardness and durability. A
discussion of these matters is found elsewhere [9], and this earlier article
can serve as a general reference to much of the material covered below.
Ag-based films are used on a very large scale in practical fenestra-
tion. Such films are delicate and are only suited for application on sur-
faces facing insulated gas-filled cavities in double-glazed windows. The
films are sometimes referred to as “soft coats” and are normally applied
by high-rate magnetron sputtering.

4.3.1  Thin Metal Films Are Not Bulk-Like

The idealized performance of a uniform metal-based thin film is read-

ily computed using formulas for thin film optics. The metal is uniquely
characterized by two parameters—known as the optical constants n and
k or the complex dielectric function ε—pertinent to a uniform bulk-
like material. Figure 4.6 shows computed results of Tlum, Tsol, R sol, and
E therm as a function of thickness d for Ag [3,10]. It is evident that d = 5
nm yields an impressive set of data, specifically with Tlum = 85%, Tsol =
74%, and E therm = 8%. A glass pane with such a thin film would make an
excellent low-E glazing. Increasing the film thickness predictably makes
the transmittance drop and the reflectance go up.
But can a uniform film, such as the one just discussed, really be made
as a window coating? The answer may be easy to guess, since we empha-
sized the idealized nature of the films considered in Figure 4.6. Before
discussing this matter further, let us look at experimental data for thin
silver films, such as those given in Figure 4.7 [11]. Here are shown mea-
sured reflectance and transmittance for conventionally evaporated Ag
films on glass. At d = 6 nm one observes that Rtherm is low, implying that
E therm is high, which is very much in opposition to the data in Figure 4.6.
It is only for films with d ≈ 9 nm or thicker that E therm becomes small, but
then Tlum and Tsol are rather low. Specifically demanding a high reflec-
tance at long wavelengths—that is, low E therm—evidently confines us to
Tlum < 50% and Tsol < 40% for evaporated Ag films. Somewhat superior
properties, with Tlum ≈ Tsol < 60% are possible for other types of film
deposition. If a much higher transmittance is required—as is frequently
the case for architectural windows and glass façades—one must use
multilayer coatings as considered in Section 4.3.4.

© 2011 by Taylor and Francis Group, LLC

 134 Green Nanotechnology

100

80 Rsol

Integrated Optical Property (%)

60

40 Tlum

Tsol
20
Etherm

0
0 10 20 30 40 50
Thickness (nm)

Figure 4.6  Integrated optical properties as a function of thickness

computed from an ideal plane-parallel Ag film. (From C. G. Granqvist,
Materials Science for Solar Energy Conversion Systems, edited by C. G.
Granqvist, Pergamon, Oxford, U.K., 1991; chap. 5, pp. 106–167. With
permission.)
4.3.2  Very Thin Metal Films Are Nanomaterials

A thin metal film does not grow uniformly on a glass pane or a poly-
mer foil. Rather the growth resembles what happens when water vapor
condenses on a fatty surface—though with the difference that the metal
“droplets” are on the scale of nanometers to micrometers. This effect of
nonuniform “wetting” has been demonstrated many times.
Information on thin-film growth and evolution has usually been
gained from transmission electron microscopy applied to films deposited
onto electron-transparent substrates such as thin carbon films, that is,
onto substrates that are different from the ones actually used in fenestra-
tion. Figure 4.8, however, shows recent scanning electron micrographs
for Au films made by sputter deposition onto glass and the images hence
are appropriate for assessing what really happens for window coat-
ings [12]. The thicknesses of the films are “equivalent” ones, that is,
the thicknesses hypothetical uniform films would have; the term “mass
thickness” is sometimes used. The initial deposition clearly leads to tiny
metallic nuclei formed at certain sites on the substrate, and continued
deposition makes these nuclei grow, which is expected to occur via dif-
fusion of atoms or molecules over the substrate surface as well as by
direct impingement of atoms or molecules. The metal “islands” that are

© 2011 by Taylor and Francis Group, LLC

 Visual Indoors–Outdoors Contact and Daylighting 135

100

Transmittance (%)
50
d = 6 nm

9
12
16
0
100
16
12
Reflectance (%)

9
50
d = 6 nm

0
0.5 1 2 5 10 20 50
Wavelength (µm)

Figure 4.7  Spectral normal transmittance and near-normal reflec-

tance measured for Ag films on glass. The film thickness is denoted d.
(Replotted from E. Valkonen et al., Solar Energy 32 [1984] 211–222.
With permission.)

deq [nm]: 2.6 4.4 6.2 8.0 9.8

200 nm

Figure 4.8  Scanning electron micrographs for Au films made by

sputtering onto glass to the shown equivalent thicknesses deq. The gold
appears bright and the uncoated parts of the substrate look dark. (From
P. C. Lansåker et al., Thin Solid Films 518 [2009] 1225–1229. With
permission.)
then formed can have shapes somewhat resembling ellipses, thus imply-
ing that the “islands” can be represented in geometric terms by a more or
less well-defined eccentricity. Progressing deposition makes some of the
“islands” touch and reorganize into larger and more irregular objects;
this is normally referred to as “coalescence growth.” The growing
film then passes through what can be called “large-scale coalescence,”

© 2011 by Taylor and Francis Group, LLC

 136 Green Nanotechnology

signifying that a meandering metallic network of macroscopic extent is

formed. This point in the growth sequence is referred to as the “percola-
tion threshold.” It is only then that metallic conduction can be measured
across the film. Optically, the film begins to display plasmonic features
[13] with increased IR reflection and an “effective” plasma frequency
ωP* due to the still low average density of carriers and their difficulty
in getting through the tortuous metallic network (so the effective mass
m* is high). Subsequent deposition makes the voids between the metallic
paths become smaller and increasingly regular so ωP* rises. Again, an
eccentricity may be useful for representing the geometry. The filling-in
of the voids may not be a continuous process, though, since crystalliza-
tion occurs concurrently in the Au film, as can be realized by comparing
the films with thicknesses being 8.0 and 9.8 nm in Figure 4.8. A close-to-
uniform layer with atomically flat grain tops may ultimately be formed.
Figure 4.9 shows the structural evolution of a metal film in a pictorial
way and also points out the important fact that the thickness range gone
through toward the growth of a continuous film depends on the method
for making the films [14]. Thus, a gold film made by conventional evapo-
ration reaches large-scale coalescence at a “critical” thickness d c of ~10
nm, whereas a film made by evaporation under simultaneous bombard-
ment with energetic ions (known as ion-assisted evaporation) has d c ≈ 4
nm. For the sputter-deposited Au film reported on in Figure 4.8, d c was
~4 nm, too, which indicates the important effect of ion bombardment
for a film undergoing growth via sputter deposition.
It is worth noting that the percolation thresholds in sputter-depos-
ited Au and Ag films are specific and nonidentical despite the fact that
both types of films go through the same structural evolution [13]. Thus,

Growth Stage Structure Thickness Model for Optical Properties

(nm)
Metal nuclei Maxwell Garnett theory
c.e. i.p.
(metal particles in air)
Irregular metal islands 3 Increasing particle
3 eccentricity and interaction
Large-scale coalescence 10 Continuum percolation?
Bruggeman theory
Non-uniform film 6 (metal-air composite)
20 Decreasing void eccentricity
Continuous metal film 9
Bulk metal

Figure 4.9  Survey over growth stages, structures, and thickness scales
for thin Au films deposited onto glass by conventional evaporation (c.e.)
and ion-assisted evaporation (i.e.). Optical models for representing the
optical properties are shown. The Maxwell Garnett and Bruggeman
theories for optical homogenization were discussed in Section 3.9. (From
G. B. Smith et al., J. Appl. Phys. 59 [1986] 571–581. With permission.)

© 2011 by Taylor and Francis Group, LLC

 Visual Indoors–Outdoors Contact and Daylighting 137

fc ~ 0.3 for Au and fc ~0.5 for Ag, with f being metal volume fraction or
area coverage and c denoting the “critical” value. This difference ensues
because highly mobile gold atoms grow together anomalously near the
percolation transition. It implies one can get an Au film to behave as a
metal with less material than in the case of Ag.

4.3.3 Toward a Quantitative Theoretical Model

for the Optical Properties

We now endeavor to go one step beyond the purely empirical descrip-

tion of the optical data and see to what extent these data can be under-
stood from models for nanomaterials, that is, we want to compare
experimental data with predictions from the optical homogenization
theories introduced in Chapter 3. Specifically, we will discuss some
results from an investigation by the authors [14]. Figure  4.10 reports
spectral transmittance and reflectance for Au films made by ion-assisted
deposition. The transmittance has a plateau-like appearance except for

100
R
T
Transmittance (%)

Au, d
glass
T
50

d (nm)
0 6
100 7
7.5
8.5
9
Reflectance (%)

50
R

0
0.5 1 1.5 2 2.5
Wavelength (µm)

Figure 4.10  Spectral transmittance T and reflectance R for Au films

made by ion plating and having the shown thicknesses d. (From G. B.
Smith et al., J. Appl. Phys. 59 [1986] 571–581. With permission.)

© 2011 by Taylor and Francis Group, LLC

 138 Green Nanotechnology

the thickest film with d ≈ 9 nm, which has a distinctively metallic-

like performance with a transmittance that drops and a reflectance
that increases monotonically with increasing wavelength in the IR. The
value of Rtherm exceeds 75% for d > 7.5 nm. But why are there plateaus
in the optical data?
The optical homogenization theories can be modified to include the
role of nonspherical voids represented by a depolarization factor L. The
calculations used Equation 3.23 with L A = L and L B = 1/3 and led to the
data in Figure 4.11, indicating that the transmittance develops a plateau
for the case of highly irregular voids specified by a large depolarization
factor [15]. The analysis is not expected to be quantitative for the simple
reason that the nanostructure of the film is highly irregular and not pos-
sible to capture accurately by one single parameter. Nevertheless, the
theoretical model leads to an intuitive understanding of the relationship
between the nanostructure and the characteristic optical data for the
films. Further analysis leads us to expect that different homogenization
theories are appropriate to the various growth stages, as illustrated sche-
matically in the right-hand part of Figure 4.9. Recent work has made it
possible to complement and extend the theoretical analysis, as discussed
in Box. 4.1, and the last word on the optical properties of noble-metal-
based films may not have been spoken.

1
Calculated Transmittance

0.8

L = 0.8
0.6

0.4 0.7
0.5
0.2
0.2

0.2 0.6 1.0 1.4 1.8 2.2

Wavelength (µm)

Figure 4.11  Computed spectral transmittance for an Au film with

thickness d = 9 nm, filling factor f = 0.2, and the averaged optical prop-
erties represented by the Bruggeman theory. The structural entities
(voids) are represented by the values shown of one depolarization factor
L. (From G. A. Niklasson, C. G. Granqvist, Appl. Phys. Lett. 46 [1985]
713–715. With permission.)

© 2011 by Taylor and Francis Group, LLC

 Visual Indoors–Outdoors Contact and Daylighting 139

Box 4.1 A Special Local “metal-in-Metal”

Plasmonic Resonance, First Identified
in Growing Thin Metal Films
The “anomalous” spectral region at λ ~ 1 μm in Figure 4.11 is due
to absorption. An analogous anomaly has also recently been seen
in nanoporous gold made by etching [16]. The large values of L
used in Figure 4.11 showed that there was a special resonance, but
its physical origins were unraveled only recently [17]. The optical
effect is related to residual isolated islands that have not joined up
with the surrounding percolating network; examples of this are
circled in Figure B4.1.1.

100nm EHT-20.00 kv

Figure B4.1.1  Percolating nanoporous Ag network containing

a small number of nanoislands which are not electrically con-
nected to the network (encircled). (From G. B. Smith, A. A. Earp,
Nanotechnology 21 [2010] 015203 1–8. With permission.)

This new class of local surface plasmon resonance is extraor-

dinarily strong because the matrix is a metal, not an insulator. To
understand and model the optical response of this nanostructure,
we can go back to the random unit cell approach in Section 3.9,
modified as in Figure B4.1.2. The entity of interest is a metal nano-
particle with a very thin insulating shell, embedded in a conducting
matrix. The equations for the core-shell structure introduced in
Section 3.9 apply here if suitably modified. The conducting matrix
surrounding the core-shell particle is porous, but it can be a fully
dense metal, which brings forth particularly interesting results.

© 2011 by Taylor and Francis Group, LLC

 140 Green Nanotechnology

Conductor surface
r
(dense or
inhomogeneous R
/nanoporous)

Figure B4.1.2  Schematic cross-sectional and plan views of a res-

onant entity being a core-shell structure. (From G. B. Smith, A. A.
Earp, Nanotechnology 21 [2010] 015203 1–8. With permission.)

The structure shown is two-dimensional and applicable to thin

films, including those in Figure B4.1.1. However, the principle can
be extended to three-dimensional core-shell systems wherein the
resonating entity is totally buried [17]. In this latter case one can
speak of a “Babinet’s principle structure” [4], and the terminol-
ogy indicates that metal is replaced by insulator and vice versa.
Babinet’s structures have begun to attract attention in several areas
of metamaterials science, in part because they can give enhanced
responses to external conditions, in part because they are easy to
engineer in thin metal films and metal foils, including with new
patterns which lead to novel resonant and plasmonic transport
properties. To amplify and shift the resonance, and hence show it
very clearly, one can overcoat to fill the voids with oxide. The core-
shell equations introduced in Section 3.9, modified for a Babinet’s
structure, fit the anomalous resonance very well with a distribu-
tion of shell thicknesses from 1.5 to 3.5 nm [17]. The resonance
wavelength is very sensitive to small changes in shell thickness.
These metal-in-metal systems make it possible to engineer
strong and very narrow absorption features within a reflector by
rather simple means. Figure B4.1.3 shows data for a 40-nm-thick
continuous Au film on glass and for such a film on top of a discon-
tinuous Au layer with 12-nm-diameter nanoparticles coated with
2 nm of alumina. The measurements are done from the substrate
side and show a prominent reflectance minimum in the NIR for the
metal-in-metal system. These data indicate a new way for tuning
solar properties and color in metal thin films.

© 2011 by Taylor and Francis Group, LLC

 Visual Indoors–Outdoors Contact and Daylighting 141

1.0
40 nm
0.9 Dense Au
Reflectance from the glass side
0.8
40 nm gold on 12 nm gold
0.7 particles coated with 2 nm
0.6 alumina
0.5
0.4
0.3
Resonant structure
0.2
0.1
0.0
250 500 750 1000 1250 1500 1750 2000 2250 2500
Wavelength (nm)

Figure B4.1.3  Spectral reflectance of an Au-based metal-in-metal

system and a corresponding continuous Au film, as described in
the main text. (From G. B. Smith, A. A. Earp, Nanotechnology 21
[2010] 015203 1–8. With permission.)

4.3.4  Multilayer Films for Spectrally Selective Windows

The low values of Tlum and Tsol in contiguous noble metal films hav-
ing low values of E therm are largely due to reflectance at their surfaces.
Consequently, it is possible to increase the transmittance by adding lay-
ers that anti-reflect the metal. One is then led to dielectric/metal and
dielectric/metal/dielectric multilayers. Dielectrics with high refractive
indices—such as Bi2O3, In 2O3, SnO2 , TiO2 , ZnO, and ZnS—give the
largest enhancement of the transmittance. It is possible to optimize for
low-E or for solar control trough proper choices of film thicknesses.
Figure 4.12 shows measured transmittance and reflectance spectra
for TiO2 /Ag/TiO2 films on glass [3]. These data are characteristic for
what one can accomplish by using the three-layer design. It is concluded
that Tlum > 80% and E therm << 20% are obtainable for the two types
of coatings. The solid curves pertain to the coating with maximum IR
reflectance; it yields Tsol ≈ 50% and R sol ≈ 42%. The dotted curves cor-
respond to Tsol ≈ 67 % and R sol ≈ 26%.
Joint development of electrical and optical properties in three-layer
films is of interest for a number of applications, and Figure 4.13 shows
some recent data for ZnO/Ag/ZnO films with each of the ZnO films
being 20 nm thick [18]. The sheet resistance R◽ is given; it is defined by
R◽ = ρ/d where ρ is the resistivity. The stated transmittance corresponds

© 2011 by Taylor and Francis Group, LLC

 142 Green Nanotechnology

100

Transmittance (%)
Glass/TiO2/Ag/TiO2
50
1 mm/18/18/18 nm
1 mm/33/13/33 nm

0
100
Reflectance (%)

50

0
0.5 1 2 5 10 20 50
Wavelength (µm)

Figure 4.12  Spectral normal transmittance and near-normal reflec-

tance measured for TiO2 /Ag/TiO2 coatings on glass. Film thicknesses are
shown. (From C. G. Granqvist, in Materials Science for Solar Energy
Conversion Systems, edited by C. G. Granqvist, Pergamon, Oxford,
U.K., 1991; chap. 5, pp. 106–167. With permission.)

to its maximum value. Clearly d < 4 nm yields poor electrical properties

while d > 8 nm produces poor optical properties. The optimum is at d ≈
6 nm, where one finds R◽ ≈ 3 Ω. The results are in basic agreement with
those discussed above.
The optimum Ag thickness in large-scale manufactured multilayer
coated windows is typically 10 to 11 nm, as pointed out in Section 3.7,
which gives a value of Tlum approaching 90%. Two features are impor-
tant for achieving the highest quality: to keep the Drude relaxation rate
(Equation 3.8) in the Ag as low as possible and to minimize surface
nanostructure. Doing this leads to sharper spectral transitions, more dis-
tinct colors, and lower values of E therm. Silver grown on TiO2 has a lower
relaxation rate than on glass, and on ZnO it is lower still. This is so
because the silver tends to grow more epitaxially on some substrates than
on others, which allows electrons to pass more easily between neighbor-
ing Ag grains. Surface nanostructure can introduce losses by interface
plasmons, as discussed above in Section 3.10, which reduces transmit-
tance. These plasmons are exacerbated by adjacent thicker high-index

© 2011 by Taylor and Francis Group, LLC

 Visual Indoors–Outdoors Contact and Daylighting 143

100
60
90
Sheet Resistance (ohm/sq.) 50

Transmittance (%)
40 80

30 70

20
60
10
50
0
40
0 2 4 6 8 10 12 14 16
Ag Thickness (nm)

Figure 4.13  Sheet resistance and maximum transmittance as a func-

tion of Ag thickness in a three-layer structure comprising ZnO(20 nm)/
Ag/ZnO(20 nm). (From D. R. Sahu et al., Appl. Surf. Sci. 252 [2006]
7509–7514. With permission.)

films [19,20], which localize the plasmon fields closer to the Ag. This is
the likely reason the coatings in Figure 4.13 perform worse than those of
optimized thin-film stacks in commercial products, such as those illus-
trated in Figure 4.14 below. The adjacent high-index layers in the latter
coatings are either thinner or slightly separated from the Ag, and the Ag
itself is very flat.
The examples above were thus selected to illustrate scientific princi-
ples rather than today’s (2010) commercial products. In fact, Ag-based
thin films for energy efficient fenestration are now highly optimized, and a
very large number of products are available on the market. The production
is of the order of 60 million m2 per year, with specified thermal, solar, and
luminous (including color) properties. The current technologies are propri-
etary, and the reader is directed to vendors’ technical publications or com-
prehensive databases [5] for specific details to the extent such are given.
Typical coatings go far beyond the three-layer design, and a “simple” low-E
coating can have six layers comprising glass/TiO2(15–20 nm)/ZnO(5–10
nm)/Ag(9–12 nm)/ITO(2–5 nm)/SnO2(35–40 nm)/TiO2(1–5 nm) [21]. An
advanced solar control coating can have as many as nine layers, including
two Ag films. The layer sequences around the silver must ensure imped-
ance matching, just as in the simple case of a three-layer coating.
Figure  4.14 illustrates two sets of data for commercially available
glass with coatings of these general types. The curves designate trans-
mittance, reflectance from the film side of the glass, and reflectance from

© 2011 by Taylor and Francis Group, LLC

 144 Green Nanotechnology

Suncool HP Neutral 70-40

1
Rf T
0.9 Rf
T Rb
0.8

0.7
Rb
0.6
Intensity

0.5

0.4

0.3

0.2

0.1

0
0 0.5 1 1.5 2 2.5
Wavelength [µm]

Optitherm SN
1
T
0.9 T Rf
Rb
0.8
Rf
0.7 Rb

0.6
Intensity

0.5

0.4

0.3

0.2

0.1

0
0 0.5 1 1.5 2 2.5
Wavelength [µm]

Figure 4.14  Spectral transmittance and reflectance from the front

(film) side (Rf) and the back side (Rb) for two commercial window coat-
ings. The upper set of data corresponds to Tlum = 79.8% and Tsol = 40.9%;
lower set of data corresponds to Tlum = 86.7% and Tsol = 57.5%. (From
International Glazing Database [2009], Lawrence Berkeley National
Laboratory; http://windows.lbl.gov/software/window/window.html.)

© 2011 by Taylor and Francis Group, LLC

 Visual Indoors–Outdoors Contact and Daylighting 145

the back side (so that the light has passed through the glass before being
reflected). It is remarkable that these multilayer window coatings are
prepared with thickness control approaching atomic precision in high-
performance production environments handling glass sheets up to 30
m 2 in size. Figure A1.4 in Appendix 1 shows such a machine. Extremely
high cleanliness of the starting glass is essential, which means multiple
washes and rinses.
Section 4.1.2 contained a brief discussion of multilayer polymer foils
with strong spectral selectivity, and at this point one may wonder why
a similar approach is not used to create solar-control coatings based on
dielectric multilayer thin films with alternating high and low refractive
indices. The reason basically is cost, and coating large surfaces—such as
window panes—with a very large number of layers is time consuming
and, hence, costly. Furthermore, window coatings normally should be
able to provide thermal insulation and have high reflectance for 3 < λ
< 50 μm which is far too wide a range to be accomplished by dielectric
multilayers. Thus, one cannot avoid a metallic-type coating, either of
the kind described above or based on a heavily doped semiconductor as
discussed next.

4.4  Spectral Selectivity of Oxide-

Semiconductor-Based Films

There are a number of oxides with band gaps which are wide enough
so that they are transparent for luminous and solar radiation and which
also allow doping to sufficiently high levels, in practice up to several
percent, so that they can achieve good electrical conductivity and hence
IR reflectivity. They are often referred to as “transparent conducting
oxides” or TCOs. In practice, these are oxides based on In, Zn, Sn, and
combinations of these. Particularly good properties have been obtained
with In 2O3:Sn, ZnO:Al, ZnO:Ga, and SnO2:F. The first of these is often
referred to as indium tin oxide or ITO and can serve as a model example
for this class of materials. A thorough discussion of these oxides and of
their achievable properties has been given recently [22], and a review
of their performance with regard to solar energy–related applications
is available as well [9]. A recently discovered alternative to the oxides
mentioned is TiO2:Nb [23,24].
Why are these oxides conducting? Figure 4.15 illustrates the origin
of the conductivity in a highly schematic way for the case of ITO [3]. The
undoped In2O3 has a regular arrangement of oxygen (O2−) and indium
(In3+) ions. When dopants are added, such as tin (Sn4+), they can replace
some of the indium. If the tin density is sufficient, each dopant atom
can be singly ionized by giving off a free electron which, together with

© 2011 by Taylor and Francis Group, LLC

 146 Green Nanotechnology

Oxygen

Indium

e– Tin

Sn+

Figure 4.15  Oversimplified crystal structure and doping model for

ITO. The actual crystal structure of In 2O3 is complicated. (From C. G.
Granqvist, in Materials Science for Solar Energy Conversion Systems,
edited by C. G. Granqvist, Pergamon, Oxford, U.K., 1991; chap. 5, pp.
106–167. With permission.)

other free electrons with the same origin, renders the material metal-
lic. The onset of metallic properties occurs at a doping level of ~10 −19
cm−3. Electron densities up to ~1021 cm−3, or somewhat above this value,
can be obtained by maximum doping. If one tries to push the doping to
even higher levels, the result will be that the dopant atoms start to form
strongly absorbing metallic clusters so that transparency is lost. The min-
imum resistivity is ~10 −4 Ωcm, or slightly below, and similar data have
been recorded for ZnO-based films. We emphasize the simplified nature
of the doping model outlined here. In fact, In 2O3 has a complicated crys-
tal structure with 80 atoms per unit cell, and there are two nonequivalent
In3+ positions. Also not every tin atom is active as a donor center [25].
A noble metal such as Ag or Au has an electron density of ~6 ×
1022 cm−3, that is, much higher than the electron density for the doped
oxides, and their resistivity is ~2 × 10 −6 Ωcm. These differences imply
that a good low-E coating made of a doped semiconductor must be much
thicker than a metal-based low-E coating with comparable properties.
In fact, the oxide coating must have a thickness of ~200 nm or higher,
as we will find below. This means that details of the initial film growth
are not at all as important as they are for metal-based window coatings.
Another consequence of the fact that the oxide films must have thick-
nesses comparable with the wavelengths of visible light is that minor
variations of this thickness—which may be hard to avoid in practical
large-area coating—can lead to optical effects such as iridescence (some-
what related to the exhibition of color in a soap bubble).
Another important difference between noble-metal-based and
doped-semiconductor-based films is that the latter are much more dura-
ble and can be used even on the exterior surfaces of window panes. The

© 2011 by Taylor and Francis Group, LLC

 Visual Indoors–Outdoors Contact and Daylighting 147

term “hard coat” is sometimes used to distinguish them from the metal-
based “soft coats.” In particular, SnO2:F films are used for practical fen-
estration; these coatings are normally made by spray pyrolysis directly
onto the hot surface of float glass as it emerges from its solidification on
the liquid tin bath. A similar technique is used to coat the windows used
in cockpits of commercial aircraft (to allow electrical heating and main-
tain impact strength); this glass usually appears iridescent when viewed
from the outside, which is a consequence of film thickness variations.

4.4.1  Some Characteristic Properties

Thin films of transparent and electrically conducting oxides require

well-defined deposition conditions in order to have optimized proper-
ties. These features are illustrated next by showing some typical data for
ITO made by reactive e-beam evaporation [8].
The left-hand panel of Figure 4.16 shows transmittance and reflec-
tance of a 0.4-μm-thick ITO film produced on a substrate heated to
~300°C. The values of Tlum and Rtherm are high. Strong interference effects
show up as oscillations in the optical data. The onset of reflectance occurs
rather steeply in the IR; it takes place at a wavelength that is somewhat
too large for the coating to work well for solar control purposes. The
Energy Advantage Low-E
1
Transmittance and Reflectance (%)

100
0.9
0.8 T
80
T R 0.7
60 0.6
Intensity

Rf
R 0.5
0.41 µm ITO 40 0.4
CaF2 Rb
0.3
T 20
0.2
0 0.1
0
0.2 0.5 1 2 5 10 20 50 0 0.5 1 1.5 2 2.5
Wavelength (µm) Wavelength (µm)

Figure 4.16  Spectral transmittance T and reflectance R for an ITO

film made by e-beam evaporation onto a substrate at ~300°C accord-
ing to the configuration shown in the inset (left-hand panel) and for
float glass with a commercial SnO2-based coating characterized by Tlum
= 83.1% and Tsol = 71.3% (right-hand panel). The reflectance for the lat-
ter film was measured from the front (film) side (Rf) and from the back
side (Rb). (Results for ITO from I. Hamberg, C. G. Granqvist, J. Appl.
Phys. 60 (1986) R123–R159. With permission; results for SnO2 from
International Glazing Database [2009], Lawrence Berkeley National
Laboratory; http://windows.lbl.gov/software/window/window.html.)

© 2011 by Taylor and Francis Group, LLC

 148 Green Nanotechnology

100 100

Luminous Absorptance (%)

80 80
Infrared Reflectance (%)

60 60
R
40 R In2O3:Sn 0.3 µm
40
In2O3:Sn 0.3 µm Glass
Glass T
20 20

0 0
0 100 200 300 400 500 0 100 200 300 400 500
Substrate Temperature (°C) Substrate Temperature (°C)

Figure 4.17  Infrared reflectance R at λ = 10 μm (left-hand panel) and

luminous absorptance (right-hand panel) for an ITO film made by e-beam
evaporation according to the configuration shown in the inset. (From I.
Hamberg, C. G. Granqvist, J. Appl. Phys. 60 [1986] R123–R159. With
permission.)
data are comparable with those for commercial low-E-coated float glass,
with the coating based on doped SnO2 , as shown in the right-hand panel
of Figure 4.16 where the presentation is analogous with the one for com-
mercially available Ag-based products in Figure 4.14 [5].
The importance of a hot substrate for making good transparent con-
ductors and IR reflectors can be inferred from Figure  4.17, where it is
evident that a temperature of about 150°C is needed to produce ITO with
a low visible absorptance and a high IR reflectance. It is noteworthy that
Alum can be as small as ~1% if the deposition takes place at a substrate
temperature of ~300°C. The resistivity of those films was ~2 × 10 −4 Ωcm.
The fact that Alum is so low makes it of interest to explore the effect of
antireflection treatments. This has been done by coating an ITO film with
MgF2; the top layer fulfills the requirements for a quarter-wave antireflec-
tion coating and, as expected, the transmittance is significantly increased.
In fact, Figure 4.18 shows that the transmittance is higher than that of the
uncoated glass over almost the entire luminous spectrum. Hence, there is
no ground for believing, as is sometimes done, that a low-E coating nec-
essarily degrades the luminous properties of a window. The MgF2 layer
enhances the thermal emittance by a few percent, that is, only marginally.

4.4.2  Typical Nanostructures of ITO Films

We first consider the structure of ITO films such as those discussed

above. Figure 4.19 shows a transmission electron micrograph of a film

© 2011 by Taylor and Francis Group, LLC

 Visual Indoors–Outdoors Contact and Daylighting 149

100

Glass + In2O3:Sn + MgF2

90 Glass
Transmittance (%)

Glass + In2O3:Sn

80

MgF2 0.09 µm
70 In2O3:Sn 0.27 µm
Glass
T
60
0.4 0.5 0.6 0.7
Wavelength (µm)

Figure 4.18  Spectral transmittance T for an ITO film made by e-beam

evaporation as deposited onto glass and after coating with an antireflect-
ing MgF2 layer. (From I. Hamberg, C. G. Granqvist, J. Appl. Phys. 60
[1986] R123–R159. With permission.)

0.2 µm

Figure 4.19  Transmission electron micrograph of a 40-nm-thick ITO

film made by e-beam evaporation onto an electron-transparent amor-
phous carbon film at 230°C. Inset depicts part of an electron diffraction
pattern. (From I. Hamberg, C. G. Granqvist, J. Appl. Phys. 60 [1986]
R123–R159. With permission.)

© 2011 by Taylor and Francis Group, LLC

 150 Green Nanotechnology

(a) (b)

100 nm 100nm

Figure 4.20  Scanning electron micrographs for an ITO film sputter

deposited at ambient temperature and annealing posttreated at 200°C
(left-hand panel) and for a similar film sputter deposited onto a substrate
kept at 200°C (right-hand panel). (From U. Betz et al., Surf. Coating
Technol. 200 [2006] 5751–5759. With permission.)
made at a substrate temperature of 230°C. It is seen to consist of crystal-
lites with an average size of the order of 100 nm. This structure is typical
for films with the highest conductivity and transmittance.
Generally speaking the nanostructure of ITO films is critically depen-
dent on details of the deposition technology. Figure 4.20 illustrates this
fact clearly for ITO films made by sputter deposition [26]. The left-hand
panel shows a scanning electron micrograph for a film deposited onto a
substrate at room temperature and subsequently annealing post-treated
at 200°C, and the right-hand panel applies to a film that was sputtered
onto a substrate at 200°C. A first assumption would be that the films
should have similar nanostructures, but clearly this was not the case.
The granular nature apparent in Figure 4.20 can be even more pro-
nounced for ITO films made by sol gel deposition using dispersions of
nanoparticles. This kind of technique is of interest for some applications
since it may open avenues toward simple and cheap deposition of pat-
terned surfaces. As found from a recent study [27], the sol-gel-prepared
films consisted of clusters of individually connected nanoparticles. Each
of the nanoparticles had a resistivity of ~2 × 10 −4 Ωcm, that is, as low
as the resistivity of ITO films made by evaporation or sputtering under
optimum conditions. However, clusters of such particles were weakly
coupled to one another, so the overall resistivity of the nanoparticle-
based films was not better than ~10 −2 Ωcm. This clustering also led to a
poor value of E therm.
For some applications it is of interest to intentionally produce a
nanostructure in order to create controlled light scattering. Such trans-
parent conductors, sometimes referred to as “milky,” can be applied as
electrodes on thin-film solar cells and have the effect of increasing the
optical path length of incident solar radiation and thereby the absorption
of this radiation. Recent work has been made in particular on ZnO:Al
prepared by sputtering and subsequent post-treatment in diluted HCl

© 2011 by Taylor and Francis Group, LLC

 Visual Indoors–Outdoors Contact and Daylighting 151

Figure 4.21  Scanning electron micrographs for a sputter-deposited

ZnO:Al film before (left-hand panel) and after (right-hand panel) etch-
ing in 0.5% HCl. (From F. Ruske et al., Thin Solid Films 515 [2007]
8695–8698. With permission.)

[28]. Figure  4.21 illustrates the roughening that is then accomplished

and that the treated film displays unevenness on a scale comparable to
the wavelength of solar light. Films prepared under optimum conditions
displayed a haze—defined as the diffuse transmittance divided by the
total transmittance—of ~30% for λ = 550 nm. The haze then dropped
off steeply toward longer wavelengths.

4.4.3 Theoretical Models for the Optical and

Electrical Properties of ITO Films

Different deposition technologies lead to very different nanostructures,

which makes it obvious that the optical and electrical properties depend
on the deposition parameters. Hence, theoretical models of these prop-
erties are not universally true for all types of films. In order to develop
a theoretical framework we first consider optimized e-beam-evaporated
films, such as those reported on in Figures  4.16–4.19 above. The the-
ory has been worked out in considerable detail [8] and the results in
Box 4.2, which uses a somewhat simplified approach, show that theory
and experiments are in excellent agreement [29].
Figure 4.22 illustrates the essentials of the physics that governs the
optical and electrical properties of ITO films. At the shortest wave-
lengths, well beyond those of relevance for visible light and solar energy,
the material is absorbing as a result of electron transitions across the
fundamental band gap Eg. When the material is electrically conducting,
the free electrons block the lowest states in the conduction band, imply-
ing that Eg becomes shifted toward larger wavelengths. The shape of
the band gap is somewhat blurred (the band edge is logarithmic) by the

© 2011 by Taylor and Francis Group, LLC

 152 Green Nanotechnology

Box 4.2 Theoretical Models for the

Optical Properties of ITO and Other
Transparent Conducting Oxides
The dielectric function of ITO is represented by additive parts from
the undoped In 2O3 lattice and the free electrons according to

ε(ω) = εIn2O3 + i/ε0ωρ(ω) (B4.2.1)

Here ρ(ω) is the dynamical resistivity due to the free electrons and
εIn2O3 contains contributions from valence electrons and polar pho-
nons of In 2O3. This particular formulation is convenient, as will
become obvious later, since it clearly exhibits the role of electron
scattering. The free-carrier part is of the greatest interest since it
is responsible for the electrical conductivity and the IR reflectiv-
ity. We first use the Drude theory that was earlier introduced in
Section 3.4. It can be written

ρDrude = 1/ε0ω2pτe − iω/ε0ω2p (B4.2.2)

where ωp is plasma frequency as given by Equation 3.7 and τe is a

relaxation time.
One may get an approximate model for ε(ω) in this way, but the
approach nevertheless is unsatisfactory, because τe is introduced as
a constant fitting parameter without any connection to the actual
mechanism for the electron scattering. However, it is possible to cal-
culate τe(ω) for a variety of scattering mechanisms in considerable
detail [30]. The one of particular interest here concerns scattering
of free electrons against ionized impurities; this type of scattering
indeed is unavoidable for ITO and similar transparent conducting
oxides, because the free electrons originate from the ionization of
the tin atoms, so ionized impurity scattering represents a baseline
giving the lowest electrical resistivity and the highest IR reflectivity
(for frequencies below the plasma frequency). Other types of scat-
tering, if important, can only degrade the properties.
Figure B4.2.1 shows that optimized ITO films come close to the
theoretically predicted properties for a material having nothing
but the unavoidable ionized impurity scattering. The upper panel
reports experimental data of ρ(ω) = ρ1(ω) + iρ2(ω) and the lower
panel gives computed results using experimental parameters for Ne
and m*. The two sets of data are in good agreement, which shows
that, indeed, ionized impurity scattering gives a good description
of the experimental results. The data on ρ1(ω) are particularly

© 2011 by Taylor and Francis Group, LLC

 Visual Indoors–Outdoors Contact and Daylighting 153

interesting because they are directly related to the electron scatter-

ing. Specifically, ρ1(ω) is constant, and consistent with the dc resis-
tivity, at energies below ħωp and falls off according to ω–3/2 above
ħωp. Other power dependencies are valid for other scattering mech-
anisms, and hence the assignment to ionized impurity scattering is
unique. The drop of ρ1(ω) at high energies causes other scattering
mechanisms to become important and, in particular, grain bound-
ary scattering can then start to influence the optical properties.
Wavelength (µm)
50 20 10 5 2 1 0.5 0.3

10–2 Experiment
Dynamical Resistivity (Ωcm)

–ρ2

dc
10–3
resistivity

ρ1
10–4

0.02 0.05 0.1 0.2 0.5 1 2 5

Energy (eV)

Wavelength (µm)
50 20 10 5 2 1 0.5 0.3

Theory
10–2
Dynamical Resistivity (Ωcm)

–ρ2

10–3

ρ1
10–4

0.02 0.05 0.1 0.2 0.5 1 2 5

Energy (eV)

Figure B4.2.1  Experimental (upper panel) and calculated (lower

panel) data on the real and imaginary contributions to the dynami-
cal resistivity ρ = ρ1 + iρ2 for e-beam-deposited ITO films as a func-
tion of photon energy and wavelength. The dc resistivity is shown.
(From I. Hamberg, C. G. Granqvist, Appl. Phys. Lett. 44 [1984]
721–723. With permission.)

© 2011 by Taylor and Francis Group, LLC

 154 Green Nanotechnology

Interband absorption
Eg Shifted bandgap
0.3 Logarithmic bandedge
Electron scattering by grain boundaries etc.
Electron scattering by partly screened Sn ions
1
ħ p
Electron scattering by screened Sn ions
3 Free electron properties

10

ħ PH Phonon absorption
30

Wavelength (µm)

Figure 4.22  Overview of the fundamental physical properties and

mechanisms for high-quality films of ITO and similar materials. (From
I. Hamberg, C. G. Granqvist, Evaporated Sn-doped In 2O3 films: Basic
optical properties and applications to energy efficient windows, J. Appl.
Phys. 60 [1986] R123–R159. With permission.)

electron scattering. Electron scattering by ionized impurities, screened

by the free electron plasma, dominate the properties in most of the
energy range from the plasma energy ħωP* and downward, although
grain boundary scattering can be of importance at the highest energies,
including in the luminous part of the spectrum. Phonon absorption at
the energy ħωPH lies far into the thermal IR; its influence tends to be neg-
ligible owing to the dominant influence of the free electrons.
The effects of the specific nanostructure are rather weak in ITO
films of the highest quality, which is due to the fact that the individual
crystallites are much larger than the distance between the ionized impu-
rities. If the crystallites are smaller, as illustrated for specific films in
Figure 4.20, the situation can be different and grain boundary scattering
can be more dominant than ionized impurity scattering. Whether or not
ITO, or similar heavily doped oxides, should be regarded as “nanoma-
terials” is somewhat of a matter of definition and—strictly speaking—
depends on the quality of the samples.

4.4.4  Computed Optical Properties

Figure 4.22 outlined the backbone of a fully quantitative theory for ITO

films that was also discussed in Box. 4.2. It is illustrative to use this
theory to calculate T(λ) and R(λ) as a function of Ne. This is done in

© 2011 by Taylor and Francis Group, LLC

 Visual Indoors–Outdoors Contact and Daylighting 155

100
T R

Transmittance (%) 0.2 µm In2O3:Sn

substrate
T
50

10 6 3 Ne = 1×1020 cm–3

0.3 0.5 1 2 3 5 10 20 30 50
Wavelength (µm)

100
R
Reflectance (%)

50

10 6 3 Ne = 1×1020 cm–3

0
0.3 0.5 1 2 3 5 10 20 30 50
Wavelength (µm)

Figure 4.23  Spectral transmittance and reflectance computed from a

quantitative theoretical model for ITO as a function of the electron den-
sity Ne. (From I. Hamberg, C. G. Granqvist, J. Appl. Phys. 60 [1986]
R123–R159. With permission.)

Figure 4.23, where the largest value of Ne lies above the range of what
is possible to achieve in practice. The main result is that the IR reflec-
tance sets in at a wavelength that scales with Ne. Phonon effects, visible
as three peaks in the thermal IR, become entirely swamped as Ne is
increased.
The spectral data allow calculations of various integrated optical
properties. Figure 4.24 shows results for Tsol as a function of film thick-
ness for three values of E therm with Ne as parameter. As discussed in
Section 4.2, there are different heat transfer mechanisms in a window
and, unless one deals with “vacuum windows,” it does not make much
sense to try to decrease E therm to the extreme. Putting E therm = 0.2 lim-
its Tsol to 78% and this is obtained for a 200-nm-thick film. Setting
E therm = 0.15 limits Tsol to 70%, and allowing E therm = 0.25 enhances
Tsol marginally to 80%. The shaded region in Figure  4.24 pertains to

© 2011 by Taylor and Francis Group, LLC

 156 Green Nanotechnology

100

Solar Transmittance (%) 80 Etherm= 0.25

0.20

0.15
60 Ne > 1021cm–3

0 0.2 0.4 0.6

Film Thickness (µm)

Figure 4.24  Solar transmittance as a function of film thickness com-

puted from a quantitative theoretical model for the optical proper-
ties of ITO. Data are shown for three values of the thermal emittance
E therm. Shaded region corresponds to electron density Ne larger than
1021 cm –3. (From I. Hamberg, C. G. Granqvist, J. Appl. Phys. 60 [1986]
R123–R159. With permission.)

Ne > 1021 cm –3, which is not practically achievable. Antireflection treat-

ments can be used to obtain a solar transmittance in excess of that in
Figure 4.24 as already shown in Figure 4.18.

4.5  Spectral Selectivity: Novel

Developments for Films and Foils

There are many emerging techniques to achieve spectral selectivity. Next,

we discuss nanomeshes and nanotubes based on metal and carbon,
foils with conducting nanoparticles, possibilities to use metal films with
designed nanoholes, and photonic crystals. There are still other ways.
One of them uses films of poly(3,4-ethylenedioxytiophene):poly(styrene-
sulfonate), known as PEDOT:PSS. Their electrical properties are not as
good as for ITO but can be better than for carbon nanotubes. The trans-
mittance can be large in the luminous range but drops in the NIR [31].

4.5.1  Silver-Based Nanowire Meshes

A new way to make transparent conducting window films that is radi-

cally different from those discussed above is to start from nanowires and

© 2011 by Taylor and Francis Group, LLC

 Visual Indoors–Outdoors Contact and Daylighting 157

500 nm

10 µm

Figure 4.25  Scanning electron micrographs of Ag nanowire meshes.

(From J.-Y. Lee et al., Nano Lett. 8 [2008] 689–692. With permission.)

deposit those onto a substrate so that an electrically conducting mesh is

formed. This is a novel approach and a direct manifestation of the poten-
tial of nanotechnologies for “green” applications. Recent work has dem-
onstrated that silver nanowires with lengths of ~10 μm and diameters of
~100 nm can be produced in large quantities via potentially inexpensive
reduction of liquid Ag nitrate [32]. After deposition and annealing to
promote electrical conduction between individual rods, a structure such
as the one in Figure 4.25 can be observed.
Figure  4.26 shows the high spectral transmittance that can be
obtained in Ag nanomesh films with a low resistance, specifically being

1.0

0.8
R = 10.3Ω
0.95
R = 22.1Ω
Solar Transmittance

0.90
Transmittance

0.6
0.85

0.4 0.80

0.75
0.2
0.70
0.0 0.1 0.2 0.3 0.4 0.5 0.6
Nanowire Density (µm–2)
0.0
400 500 600 700 800
Wavelength (nm)

Figure 4.26  Spectral transmittance of Ag nanowire meshes with two

values of the resistance. (From J.-Y. Lee et al., Nano Lett. 8 [2008] 689–
692. With permission.)

© 2011 by Taylor and Francis Group, LLC

 158 Green Nanotechnology

10.3 and 22.1 Ω/◽. The data compare very well with those for Ag-based
multilayer films in Figure  4.13. The meshes consist of structural units
that are large enough to cause significant scattering, and ~20% of the
visible light is scattered at angles exceeding 10°. This haze is a severe lim-
itation for most windows in buildings, whereas it is insignificant, or can
even be advantageous, for applications related to solar cells and LEDs.

4.5.2  Carbon Nanotubes and Graphene

Carbon can form an infinite number of nanostructures, some of which

are illustrated in Figure 4.27. Graphite has a layered structure such as
the one to the right. If the individual layers are separated, one gets gra-
phene. Carbon can also form nanotubes, as indicated in the middle, and
nanoparticles consisting for example of 60 atoms in a C60 unit (also
known as a buckminsterfullerene molecule or a “buckyball”).
An emerging type of transparent conducting films, with possibly
even larger potential than the Ag nanowires, uses meshes based on car-
bon nanotubes. Single-walled carbon nanotubes, which can have lengths
up to several centimeters, consist of one layer of hexagonal graphite lat-
tice rolled to form a seamless cylinder with a radius up to a few nano-
meters. The nanotubes are of two types: metallic and semiconducting

Figure 4.27  Schematic illustration of carbon-based nanostructures

of different types and of their formation. (From A. K. Geim, K. S.
Novoselov, Nature Mater. 6 [2007] 183–191. With permission.)

© 2011 by Taylor and Francis Group, LLC

 Visual Indoors–Outdoors Contact and Daylighting 159

1 µm

Figure 4.28  Transmission electron micrograph of a film comprised of

carbon nanotubes. (From Y.-X. Zhou et al., Appl. Phys. Lett. 88 [2006]
123109 1–3. With permission.)
with conductivities comparable to those of Cu and Si, respectively. The
individual nanotubes are electrically weakly coupled to one another.
There are several different methods for preparing such nanotubes on
substrates, including solution-based techniques such as dipping, spray-
ing, spin coating, and printing. The ensuing layer is a mesh of nano-
tubes; a typical example is shown in Figure 4.28 [34]. Continuous and
scalable manufacturing is feasible [35].
The optical transparency of carbon nanotube meshes can be high
in the visible and IR, as apparent from Figure 4.29 pertaining to a film
with a resistance of 120 Ω/◽. It is apparent that the properties cannot
yet fully match those of ITO and the best metal-based window films,
but advantages regarding cost can outweigh this limitation depending
on application. In the future, nanotube layers with a higher fraction of
the metallic variety, and perhaps means to diminish the barriers between
the individual tubes, conceivably will lead to better performance. A par-
ticular feature is the possibility to have large transparency even in the
thermal infrared range [36].
Another possibility to make transparent conducting films is offered
by graphenes, that is, two-dimensional graphite sheaths. These sheaths
can be produced via mechanical exfoliation by repeated peeling of highly
ordered graphite. The data obtained so far point at the feasibility of
using graphenes [37], but the optical and electrical performance data are
still far from those of the other coatings discussed above.

© 2011 by Taylor and Francis Group, LLC

 160 Green Nanotechnology

100

Transmittance at 550 nm (%)

90
ITO
80

70

60

50
100 1000 10000
Sheet Resistance (Ω)

Figure 4.29  Transmittance at a mid-luminous wavelength versus sheet

resistance for carbon nanotube films (dots). The horizontal bar denotes
typical data for ITO films. (After Y.-X. Zhou et al., Appl. Phys. Lett. 88
[2006] 123109 1–3. With permission.)

4.5.3  Foils with Conducting Nanoparticles

Narrow-band surface plasmon resonances in conducting nanoparticles

were introduced in Section 3.10. These resonances allow a simple and
practical approach to solar control if they are used in polymer foils [38],
and foils of this type—based on PVB as well as PET—have been applied
commercially. Conducting nanoparticles, with little effect on Tlum while
blocking in the NIR, are needed and the sizes of the particles must be
below ~50 nm as they otherwise create haze and destroy the possibility of
giving a clear view in transmittance. Figure 4.30 shows spectral absorp-
tion coefficient for a polymer with a refractive index of 1.5 and containing
0.3% by weight of a number of readily available conducting nanopar-
ticles, and illustrates the sharp surface plasmon resonances that are then
manifest. Clearly, the resonances can occur anywhere in the wavelength
range for solar radiation, depending on the type of nanoparticles.
It is evident that metallic nanoparticles, such as TiN and ZrN, have
too high plasma frequencies to yield a large value of Tlum, but they can
produce distinct colors of glazings, as illustrated for Roman gold-doped
glass in Figure 3.1 (and as will be seen also in heat reflecting paints; cf.
Chapter 5). In stark contrast, commonly available transparent conductors
such as ITO in nanoparticle form clearly have their absorption peaks well
beyond the wavelengths for most of the IR in the solar radiation (i.e., 0.7
< λ < 1.1 μm). Foils with this type of nanoparticles look clear, and some
weak damping of Tsol takes place only by the leading edge of the plasmon
resonance. The best material so far for blocking NIR solar radiation in

© 2011 by Taylor and Francis Group, LLC

 Visual Indoors–Outdoors Contact and Daylighting 161

LaB6
600
ITO

Absorption Coefficient (mm–1)

SnO2:Sb
500
TiN
ZrN
400

300

200

100

0
500 1000 1500 2000 2500
Wavelength (nm)

Figure 4.30  Spectral absorption coefficient for a polymer containing

0.3% by weight of the shown nanoparticles. Only a fraction of the initial
part of the resonance due to SnO2:Sb is shown.
transparent foils is LaB6 [39] for which the resonance is centered at λ ≈ 1
μm as shown in Figure 4.30. This material is excellent for glare reduction;
it displays some enhancement of the coloration when the nanoparticle
level is increased from zero to 0.03%. Figure 4.31 shows that very small
amounts of LaB6 are needed to significantly decrease Tsol, while Tlum is not
affected to any major extent. ITO nanoparticles can serve as a replace-
ment for LaB6, but an order of magnitude higher nanoparticle contents
must then be used to get the same blocking of the NIR solar radiation.
Transparent conducting oxide films can be made with doping lev-
els from zero to a value exceeding 1021 cm−3, as discussed at length in
Section 4.4 above, and hence the plasma frequency can be tuned to a
desired position for blocking NIR solar radiation in nanoparticle-con-
taining foils. Calculations for 0.7-mm-thick PVB foils with 0.05 wt% of
optimized ITO and ZnO:Al point at possibilities to achieve properties
even better than those for LaB6.

4.5.4  “Holey” Metal Films

An optically thick metal film with subwavelength holes can show

“anomalous” optical transmission, that is, the transmission can be sub-
stantially larger than the area fraction that is occupied by the holes [40].
There is much interest in this novel phenomenon today, and we include
some information here partly for completeness and partly because there
may be a relationship to the optical data for thin metal films including
holes discussed in Section 4.3.3 above.

© 2011 by Taylor and Francis Group, LLC

 162 Green Nanotechnology

90

80
Luminous and Solar Transmittance (%)
70 Tlum

60

50

40 Tsol

30

20
0 0.005 0.01 0.015 0.02 0.025 0.03 0.035
LaB6 Nanoparticle Concentration (wt%)

Figure 4.31  Luminous and solar transmittance as a function of LaB6

nanoparticle concentration for 0.7-mm-thick doped PVB foils. (From
S. Schelm, G. B. Smith, Appl. Phys. Lett. 82 [2003] 4346–4348. With
permission.)
The effect is illustrated in Figure  4.32 showing transmittance
through gold films with the indicated nanostructures; the holes are 0.25
μm in diameter and have a lattice constant of ~2 μm [41]. An “extraor-
dinary” transmittance peak centered at λ ≈ 615 nm stands out clearly.
Peak transmission levels up to 39% have been reported. Multiple trans-
mission bands can be created with periodic arrays of nanohole aper-
tures, and maxima and minima of transition spectra can be suppressed
by adequate nanohole arrangements so that the transmittance can be
tailored to different needs. The transmittance may not be sufficient for
most energy-related applications, and the “holey” films may not allow
visual indoors−outdoors contact, but the “anomalous” transmission
nevertheless deserves some attention here because it can affect Tlum and
Tsol for noncontinuous metal films.

4.5.5  Photonic Crystals

Also for completeness we now consider nanostructures based on arrays

of identical oxide or polymer nanoparticles. They are known as “pho-
tonic crystals” and can be used for precise spectral control in either
windows or mirrors. Particle arrays yield two- or three-dimensional

© 2011 by Taylor and Francis Group, LLC

 Visual Indoors–Outdoors Contact and Daylighting 163

0.5

100 nm
0.4
Transmittance

180 nm
0.3

0.2

180 nm
0.1

180 nm
0.0
400 500 600 700 800
Wavelength (nm)

Figure 4.32  Spectral wide-angle transmittance through Au films with

the shown configurations. The subwavelength hole arrays are seen to lead
to extraordinary transmittance in a band centered at λ ≈ 615 nm. (From
H. Gao et al., Nano Lett. 6 [2006] 2104–2108. With permission.)

photonic crystals while multilayer films—such as the ones producing the

data in Figure 4.3—can be described as one-dimensional photonic crys-
tals. Lattice spacings or layer thicknesses must be repeatable to a high
degree of accuracy, though in some applications a deliberate slight ran-
domness in spacing may also be of use. These systems can act as near-
perfect all-dielectric mirrors at wavelengths that cannot propagate inside
the structure at the forbidden “band gaps” for photons. Such effects are
well known for electrons in crystalline solids and result from destruc-
tive interference between electron waves. Figure 4.33 depicts a photonic
band gap structure and illustrates optical interference leading to reflec-
tion depending on viewing angle [42,43]. Figure  4.34 shows viewing-
angle-dependent colors and also indicates that there is a dependence on
particle size, that is, on the lattice constant [42,43].

4.6  Optimized Angular Properties:

The Energy Efficiency That Is Possible

Spectral selectivity can give well-specified advantages with regard to

energy efficiency for a window, as discussed in Section 4.2 above. The
choice of appropriate angular performance for the fenestration is impor-
tant, too, but the advantages cannot be as sharply defined.

© 2011 by Taylor and Francis Group, LLC

 164 Green Nanotechnology

Interference of Diffused Light

Angle-dependent color

Viewing
θ angle
Matrix

Nanoparticles

Lattice constant a

Figure 4.33  Interference of light in a photonic crystal consisting of an

array of polymer nanoparticles embedded in a transparent crystal. (R. J.
Leyrer, personal communication to one of the authors [GBS].)

Figure 4.34  A color version of this figure follows page 200. Color
shifts, caused by a range of viewing angles, in reflected light from a pho-
tonic crystal such as the one depicted in Figure 4.33. Data are shown for
different particle sizes. (R. J. Leyrer, personal communication to one of
the authors [GBS]. Figure courtesy of BASF—The Chemical Company.)

Why are angular properties important for providing energy efficiency

of windows? The reason is that the angles appropriate for viewing usu-
ally are not the same as the one from which the solar radiation emerges.
Normal indoors–outdoors visual contact through a window is most often
along a more or less horizontal line-of-sight. Of course, this is not always
the case, and, for example, a person who is indoors may wish to look at
the sky before taking a walk in order to decide whether or not to bring
an umbrella. Or someone high up in a tall building may want to see what
happens on the ground just outside the building. But these cases are not
the most typical. Solar irradiation comes for most of the day from a small
element of solid angle high up on the vault of heaven, and if this radiation

© 2011 by Taylor and Francis Group, LLC

 Visual Indoors–Outdoors Contact and Daylighting 165

penetrates freely into a building it may add significantly to the cooling

load, and it would then be energetically advantageous to remove at least
the direct impingement of solar radiation. This is what one can accom-
plish by the use of traditional techniques such as overhangs, awnings, and
the like. But the problem is then that the basic function of the window—
to give unimpeded indoors–outdoors contact—is severely eroded.
Daylight management is another bonus for glazings with properly
chosen angular properties. There is usually plenty of daylight from the
very large solid angle of sky not occupied by the sun and its circumsolar
bright halo. Using this indirect radiation and blocking most of the direct
solar rays can eliminate glare, which is a common reason for people to
close blinds and turn on lights. Another type of glazing with carefully
designed angular properties can be very useful when employed on a roof
for collecting low-angle light in the morning or in the winter but reject-
ing much of the solar heat in the middle of the day and in mid-summer.
The fact that Tlum and Tsol often pertain to different angles leads
naturally to the consideration of coatings with angular dependent prop-
erties. For the most common case of a vertical window, a simple coating
such as a metal film will be most transparent at normal incidence simply
because the thickness through which the optical radiation has to pen-
etrate is then at a minimum. Hence, the desired angular properties are
easily reached to some extent.
If the window is not vertical, one may want a higher transmittance
at an off-normal angle of incidence (along a horizontal line-of-sight,
say) than at normal incidence. This can be accomplished with multilayer
structures having more than one metallic film and, essentially, depends
on destructive and constructive interference of light reflected between
the two metal films. As noted in Section 4.3.4, solar control films are
often of this general type. In order to get an idea of the orders of mag-
nitude that are involved, we can look at some computed data for a five-
layer coating of the type discussed above, specifically incorporating two
12-nm-thick Ag films embedded between three SiO2 films of equal thick-
ness. It is straightforward to compute the optical properties of this struc-
ture using formulas for thin-film optics [4]; of course these properties are
symmetrical with regard to the normal to the window pane. Figure 4.35
shows that Tsol can display very different angular variations, depending
on the thickness of the SiO2 layers [44]. If this thickness is 120 nm, Tsol
falls off monotonically with increasing incidence angle θ, whereas the
behavior is more complicated when the SiO2 films are 170 nm thick. In
the latter case, Tsol is about twice as high for θ = 45°, and three times as
high for θ = 60°, compared to the case for normal incidence. It is clear
that the use of proper angular properties will lead to energy efficiency,
but the magnitude of this efficiency depends on many external param-

© 2011 by Taylor and Francis Group, LLC

 166 Green Nanotechnology

100

80 d = 120 nm

Solar Transmittance (%)

60

40 d = 170 nm
θ

20
SiO2 (d) Ag (12 nm)

T
0
0 30 60 90
Incidence Angle (deg.)

Figure 4.35  Solar transmittance for thin films of SiO2/Ag/SiO2/Ag/

SiO2 as a function of SiO2 thickness d and incidence angle θ. (From G.
Mbise et al., Proc. Soc. Photo-Opt. Instrum. Engr. 1149 [1989] 179–199.
With permission.)
eters and cannot be generally quantified. What can be stated, though, is
that the effect is far from insignificant.
An assessment of the energy efficiency for a practical case is ham-
pered by the fact that commercially available coated window glass and
foil are rarely specified with regard to their angular properties; however,
empirical formulas can be used to estimate this dependence [45,46].

4.7  Angular Selectivity of Films with

Inclined Columnar Nanostructures

Window coatings that have different optical properties at equal angles

on the two side of the surface normal are of potential interest for energy-
efficient windows, although they have not yet been implemented in prac-
tical fenestration. This angular selectivity cannot be accomplished with
uniform films, such as the ones in Figure 4.35, whereas it is possible in
films with a pronounced nanostructure comprised of inclined columnar
features. The first issue to consider is a practical one: how to produce
such nanostructured films.
Consider a film that is built up atom by atom or molecule by molecule
using a process—such as vacuum evaporation or sputtering—wherein

© 2011 by Taylor and Francis Group, LLC

 Visual Indoors–Outdoors Contact and Daylighting 167

these atoms or molecules impinge from a more or less well-defined direc-

tion. Furthermore, assume that these impinging species have an energy
that is not very high, meaning that they cannot diffuse widely over the
surface where they hit and that they are incapable of causing major recon-
struction of the growing film. Finally, assume that the species have an
angle α with regard to the substrate normal. The film that is built up will
then consist of columns separated by void spaces, and the columns will
be oriented at an angle β toward the direction of the impinging species.
The angles α and β do not coincide but are related to each other [47]. The
relationship is often expressed in terms of a “tangent rule,” tan β = ½ tan
α, which, however, cannot be taken as generally valid.
The evolution of the inclined columnar structure is made clear from
simple model calculations illustrated in Figure 4.36 [48]. Here “atoms,”
depicted as two-dimensional hard discs, are allowed to travel with a
well-defined direction but otherwise randomly toward a “substrate.”
The “atoms” stick where they hit this substrate or a previously deposited
“atom” and relax into a stable position between other “atoms.” Columns
are then formed basically for the simple reason that a randomly formed
protrusion in the film tends to shade whatever is behind it from further
deposition, so there is a “self-shadowing” mechanism in place. The nature
of the columns and the density of the nanostructure depend on the energy
the “atoms” are taken to have in the simulation, that is, to the mobility
of the deposition species. The panels in Figure 4.36 refer to three values
of this mobility. As expected, an increase of the mobility leads to a denser
“film.” By moving the substrate during the deposition, one can accom-
plish an entire “zoo” of nanostructures [49], as shown in Figure A1.6.
Theoretical modeling of films with inclined columnar nanostructures
is possible via the theories for optical homogenization presented in Section
3.9, essentially by representing the columnar features, in three dimen-
sions, as elongated prolate spheroids (cigar-shaped ellipsoids) with their
long axis coinciding with the columns’ symmetry axis [50]. The modeling
is technically rather complicated, though, and is not pursued here.
Particularly detailed data on angular selectivity are available for Cr
films made by oblique angle deposition [51]. They are discussed in Box
4.3, where it is shown that empirical film structures can be strikingly
similar to those in Figure 4.35, and that Tlum can differ by a factor of
about two at equal angles on either side of the surface normal.

4.8  Chromogenics: The Energy

Efficiency That Is Possible

Chromogenics is a relatively new term which was introduced, in the

sense it is used here, only in 1990 [52]. Chromogenic devices are able

© 2011 by Taylor and Francis Group, LLC

 168 Green Nanotechnology

(a)

(b)

(c)

Figure 4.36  Simulated deposits grown at an angle of incidence α =

50° and at low (a), medium (b), and high (c) mobility of the impinging
“atoms.” (From M. J. Brett, J. Mater. Sci. 24 [1989] 623–626. With
permission.)

© 2011 by Taylor and Francis Group, LLC

 Visual Indoors–Outdoors Contact and Daylighting 169

Box 4.3 Angular Selectivity in Obliquely

Evaporated Chrome-Based Films: A Case Study
Figure B4.3.1 shows a scanning electron micrograph of the cross-
section of a Cr-based film prepared by evaporation with α = 75°. A
pronounced columnar structure is clearly seen with β = 40 ± 5°.

0017 15KV X20,000 1 µm WD48

Figure B4.3.1  Scanning electron micrograph of a Cr film pro-

duced with oblique angle evaporation at α = 75°. (From G. W.
Mbise et al., J. Phys. D: Appl. Phys. 30 [1997] 2103–2122. With
permission.)

Measurements of polarization-dependent optical transmittance,

Ts(θ,λ) and Tp(θ,λ), were performed on the Cr-based films with the
light beam incident at an angle θ to the film’s normal and in the
incidence plane (spanned by this normal, the direction of the evap-
orated flux, and the column direction). The geometry is explained
in Figure B4.3.2, where ϕ = 0 corresponds to the incidence plane.
Data were recorded for −64° < θ < +64°, with positive angles taken
to be on the same side of the normal as the deposition direction and
the column orientation. It is necessary to consider s and p polar-
ized light separately; the transmittance appropriate for unpolar-
ized light is then the arithmetic mean of the two values. Figure
B4.3.3 shows data for Ts(0,λ), Ts(±55°,λ), Tp(0,λ), and Tp(±55°,λ)
for a 45-nm-thick film evaporated at α = 85°. The curves display
pronounced angular selectivity with the transmittance increas-
ing monotonically with wavelength in the displayed solar range.

© 2011 by Taylor and Francis Group, LLC

 170 Green Nanotechnology

It is found that Ts(−55°,λ) = Ts(+55°,λ) < Ts(0,λ) and Tp(−55°,λ) <

Tp(0,λ) < Ts(+55°,λ), that is, selectivity is confined to the p polar-
ized component, as expected from theory.
z
z
y

t
gh
Li
x
θ y
ø
x

Figure 4.3.2  Geometry for describing angular selective opti-

cal properties of a film with inclined columnar nanostructure.
Light is incident at a polar angle θ and an azimuthal angle ϕ onto
a structure with nanofeatures represented by prolate spheroids.
The theoretical modeling of the optical properties represents these
spheroids in a coordinate system spanned by (x,y,z). The thin film
is described in the coordinate system (x′,y′,z′). The theory accounts
for the transformation between these coordinate systems.
100

Tp(+55°)
Ts(0)

Tp(–55°)
Transmittance (%)

Tp(0)
Ts(±55°)
50

Evaporated Cr film

0
0.5 1 1.5 2 2.5
Wavelength (µm)

Figure 4.3.3  Spectral transmittance T for a 45-nm-thick Cr film

made by oblique angle evaporation at α = 85°. The curves refer to
different incidence angles and states of polarization for the light.
(From G. W. Mbise et al., J. Phys. D: Appl. Phys. 30 [1997] 2103–
2122. With permission.)

© 2011 by Taylor and Francis Group, LLC

 Visual Indoors–Outdoors Contact and Daylighting 171

Figure B4.3.4 illustrates Ts,lum(θ), Tp,lum(θ), and Tu,lum(θ), where

u denotes unpolarized light, at −64° < θ < +64° for the same film
as in Figure B4.3.3. Expectedly, Ts,lum(θ) is symmetrical around θ
= 0, whereas Tp,lum(θ) is strongly angular selective. Tu,lum(θ) varies
from ~0.3 to ~0.7 when θ goes from −64° to +40°. The maximum
transmittance occurs at an angle θmax. It was found that θmax ≈ β,
implying that the column orientation coincides with the direction
for maximum transmittance.

Evaporated
80 Cr film Tp
Luminous Transmittance (%)

Tu
60

Ts
40

20
–80 –40 0 40 80
Incidence angle, θ (deg.)

Figure B4.3.4  Luminous transmittance as a function of inci-

dence angle for the same Cr film as in Figure B4.3.3. Data are given
for s polarized, p polarized, and unpolarized light. (From G.  W.
Mbise et al., J. Phys. D: Appl. Phys. 30 [1997] 2103–2122. With
permission.)

Data were recorded not only with regard to the incidence plane
but also for other directions. A polar representation of T(θ,ϕ,λ) is
then appropriate; ϕ = 0 and ϕ = 180° correspond to the incidence
plane as shown in Figure B4.3.2. Measurements were taken with
laser light at λ = 543 nm and a photodetector matched to the eye’s
sensitivity. Figure B4.3.5 shows data for the same film as above. The
transmittance is illustrated by use of a gray scale with five 5%-wide
intervals. The asymmetry between the left-hand and right-hand

© 2011 by Taylor and Francis Group, LLC

 172 Green Nanotechnology

halves of the diagram provides striking evidence of angular selectiv-

ity for monochromatic as well as for luminous light, whereas the
good agreement between the upper and lower halves of the diagrams
indicates the symmetry expected from the geometrical features of
the obliquely evaporated films. Data corresponding to ϕ = 0 and ϕ =
180° can be directly compared with results in Figure B4.3.4.
λ = 543 nm

θ
–50 50

(a)
Luminous

θ
–50 50

(b)
Transmittance: 35–40%
40–45%
45–50%
50–55%
55–60%

Figure B4.3.5  Polar plot of fixed wavelength and luminous

transmittace through the same Cr film as in Figures B4.3.3 and
B4.3.4. Data are given as a function of incidence angle θ and azi-
muthal angle ϕ. (From G. W. Mbise et al., J. Phys. D: Appl. Phys.
30 [1997] 2103–2122. With permission.)

to vary their optical properties, reversibly and persistently, in response

to an external stimulus such as ultraviolet irradiation, temperature, or
electrical charge or voltage. Variations of the optical properties are of
obvious interest for achieving energy efficiency for the simple reason that
the ambient properties vary during the day or season.
The purpose of chromogenic fenestration can be to create energy
efficiency in a building by admitting solar energy when there is a

© 2011 by Taylor and Francis Group, LLC

 Visual Indoors–Outdoors Contact and Daylighting 173

heating demand and rejecting solar energy if there is a cooling demand.

Considering such fenestration, which is also referred to as “smart” or
“switchable,” there is also a human dimension related to comfort imply-
ing that there may be a need for good lighting, glare elimination, and
prevention of thermal stress. Obviously the notion of “comfort” is a
subjective one, even if quantification is possible to some extent, so chro-
mogenics is also of interest for meeting individual demands.
How much energy can one save by using windows with variable
properties? The answer to this question is complicated and multidimen-
sional as will become apparent immediately below. In the rest of this
section we then discuss three types of materials and their associated
technologies in some detail. They are photochromic and change under
ultraviolet irradiation, thermochromic and change when the tempera-
ture is altered, and electrochromic and can be changed electrically.
These are not the only chromogenic technologies, though, and another
one that has attracted some attention recently is gasochromics [53]. It con-
siders double-pane windows with facilities to introduce reducing (hydro-
gen containing), as well as oxidizing gases between the panes. One of the
glass surfaces facing these gases has a coating of, for example, tungsten
oxide and a very thin layer of a catalyst such as Pd. Tungsten oxide colors
uniformly in the presence of a reducing gas and bleaches in an oxidiz-
ing gas. The physics underlying these phenomena are similar to that of
the electrochromic materials to be discussed in Section 4.11. Gasochromic
window prototypes have been built and have demonstrated good optical
modulation. The necessity to invoke a gas handling system is an undeniable
drawback, however, and gasochromics will not be discussed further here.
Still another magnetochromic option was mentioned in Section 3.2.3.
How large is the possible energy savings with windows that are
able to change their transparency? We first attempt to answer this ques-
tion by a simple and intuitive line of reasoning and consider a building
requiring cooling.
The solar energy falling onto a vertical surface, such as a normal
window, each year is chosen here to be 1000 kWh/m 2. This energy can
serve as a nominal value, whereas more correct numbers for south-
facing/north-facing/horizontal surfaces are 850/350/920 kWh/m 2 for
Stockholm (Sweden), 1400/450/1700 kWh/m2 for Denver (Colorado), and
1100/560/1800 kWh/m2 for Miami (Florida). Only about half of the solar
energy corresponds to visible light (i.e., 500 kWh/m 2). This number is
used in the analysis below because IR radiation can be reflected off, at
least in principle, by use of a spectrally selective surface that does not
exhibit variable transmittance (cf. Figure  2.16 or Figure  4.5). We now
assume that the transparency of visible light can be altered between 7%
and 75%, as has been demonstrated by use of electrochromic technology
to be discussed further below. The difference between having the window

© 2011 by Taylor and Francis Group, LLC

 174 Green Nanotechnology

constantly colored and constantly bleached then is 340 kWh/m2. The

next issue is to consider when the window should be colored and when it
should be bleached. Assuming for simplicity that the window should be
fully transparent when the room is in use, and that it is in its darkest state
when there is nobody in the room, the relevant question regards the frac-
tion of time that the room is “inhabited,” or, more precisely, the fraction
of the energy that comes into the room when it is not in use. Considering
that a normal (office) room is empty during weekends, holidays, and vaca-
tions, early mornings and late afternoons (when the sun is near the hori-
zon), etc., it is probably a conservative estimate that 50% of the energy
in principle can enter the room when there is no one to look through the
window. Hence, the estimate yields that 170 kWh/m2 is the amount of
energy saved annually by adopting the given control strategy.
Is this energy savings significant or not? To answer this question,
one may note that 17% is a typical value for the photoelectric conversion
efficiency of today’s best solar cells and thin-film submodules devised for
terrestrial applications (somewhat below the data for the research-type
solar cells whose efficiencies are given in Figure 6.15). Thus, these solar
cells would be able to generate 170 kWh/m 2 in the example given above.
Of course, the analogy between energy savings in “smart” windows and
energy generation in solar cells is not tied to the choice of the incident
solar energy being 1000 kWh/m 2 but applies generally, irrespective of
the orientation of the surface under consideration. The “smart” window
saves thermal energy, but if a cooling machine—operating with an effi-
ciency of 300%, say—runs on electricity generated with an efficiency of
33%, then the analogy becomes perfect.
We can summarize the reasoning above as follows: A “smart win-
dow” can save as much electricity for a cooling machine as a solar cell
module, with the same position, could have generated.
This analogy between energy savings and energy generation is very
simplistic, though, and more correct assessments of the potential of the
“smart windows” technology must consider issues such as downsizing
of cooling equipment, reductions in peak electricity demand, possibili-
ties to achieve energy-efficient daylighting, and user aspects such as pos-
sibilities to avoid glare and thermal stress. Some studies for European
climates have indicated that the energy for space cooling, on an annual
basis, could be reduced by as much as 40% to 50% when using “smart”
windows instead of conventional static solar control windows [54,55].
Other analyses, performed for California, also indicated very substan-
tial energy savings [56]. A study from which Figure 4.37 was taken gives
a very schematic illustration of the energies for cooling and for elec-
tric lighting with a number of fenestration types [57]. Not surprisingly,
clear glass gives a comparatively small need for artificial lighting but
is disadvantageous with regard to cooling energy. Going to tinted and

© 2011 by Taylor and Francis Group, LLC

 Visual Indoors–Outdoors Contact and Daylighting 175

40

Clear glass
30

Cooling Energy (kWh/m2)

Tinted glass
Photo-
20 chromic

Reflective
glass
10
Electro- Thermochromic
chromic

0
0 20 40 60 80
Electric Lighting Energy (kWh/m2)

Figure 4.37  Schematic relationship between electrical lighting energy

and cooling energy for a number of fenestration types. (After UNEP,
Buildings and Climate Change: Status, Challenges and Opportunities,
United Nations Environment Programme, Paris, 2007.)

reflecting glass diminishes the cooling energy but increases the demand
for lighting. Chromogenic technologies—especially the one based on
electrochromics—are found to have strong advantages both for cooling
energy and electric lighting energy.
The detailed numbers in the mentioned studies should be regarded as
tentative. However, there can be no doubt that chromogenic technologies
are able to provide very significant energy savings especially in buildings
that are cooled during a large part of the year. And these energy savings
can be combined with increasing comfort [58]. It is expected that more
analyses will be performed during coming years so that the energy sav-
ings potential can be set on firmer ground.

4.9  Photochromics

Photochromic glass darkens under UV irradiation and bleaches spon-

taneously in the absence of such irradiation. This kind of glass is well
known and has been used in sunglasses for decades. In principle, there
is nothing to prevent its use in windows, and test production of photo-
chromic glass panes has been made in the past. There are also numerous
thin films that show photochromism, but usually their ability to sustain
many darkening and bleaching cycles is low.

© 2011 by Taylor and Francis Group, LLC

 176 Green Nanotechnology

Ordinary photochromic glass is based on photosensitive compounds

that are randomly dispersed in the vitreous matrix. Metal halides, espe-
cially silver halide, are widely used as the light-absorbing substance. The
glass that is most common also contains some chlorine and bromine
ions as well as a small amount of copper ions. After appropriate heat
treatment, the glass contains Cu-doped halide particles with sizes of 10
to 20 nm (i.e., the glass is a nanomaterial) [59]. Photo-induced reactions
lead to the formation of clusters of metallic Ag, and therefore this mate-
rial has some principle kinship to the selectively solar-absorbing coat-
ings discussed in Section 6.1. When the UV irradiation ceases, the glass
returns to its initial state. The details of the coloration and bleaching are
still poorly understood.
The practical performance of photochromic glass is governed by the
transmittance in dark and transparent states and the coloring and bleach-
ing rates. The transmittance levels depend on the glass composition and
values of Tlum up to 90% and down to 5% are possible. There is also
some temperature dependence, and a temperature rise leads to enhanced
dynamics. Figure  4.38 shows Tlum during darkening and clearing for
two photochromic glass types measured at different temperatures  [3].

60
Darkening Clearing
Luminous Transmittance (%)

40

20

0
0 30 60 90 120
Time (minutes)

Figure 4.38  Luminous transmittance versus time during darkening

and clearing of two photochromic glasses that had been cleared over-
night prior to the measurements. Upper data were recorded at 25°C and
lower data at 3°C. (From C. G. Granqvist, in Materials Science for Solar
Energy Conversion Systems, edited by C. G. Granqvist, Pergamon,
Oxford, U.K., 1991; chap. 5, pp. 106–167. With permission.)

© 2011 by Taylor and Francis Group, LLC

 Visual Indoors–Outdoors Contact and Daylighting 177

100
Clear

80

Transmittance (%)
60
Dark

40

20

0
0.4 0.6 0.8 1.0 1.2
Wavelength (µm)

Figure 4.39  Spectral transmittance for a photochromic glass in clear

and dark states. (From C. G. Granqvist, in Materials Science for Solar
Energy Conversion Systems, edited by C. G. Granqvist, Pergamon,
Oxford, U.K., 1991; chap. 5, pp. 106–167. With permission.)

Irradiation makes Tlum drop swiftly and, generally, darkening to 80% of

the full range occurs during 1 min. However, some residual darkening
is noticeable after times as long as 1 h. The clearing in the absence of
irradiation is much slower and is incomplete even after 1 h.
Figure 4.39 shows typical spectral transmittance in the 0.35 < λ <
1.3 μm range for photochromic glass in dark and clear states [3]. It is
evident that the difference in the optical properties is confined mainly to
the 0.35 < λ < 1 μm range, implying that the modulation of Tsol is much
smaller than that of Tlum, thus pointing at a limitation in the perfor-
mance as a window coating.
Photochromic plastics have been known for many years, too,
and have optical properties similar to those for photochromic glass.
Photochromic spiropyran-silica mixtures can show both photochromic
and thermochromic properties [60].

4.10  Thermochromics

Thermochromic materials are of interest for energy-efficient windows

if they allow a substantially larger transmission below a “critical” tem-
perature τc than above this temperature, and if τc is near room tem-
perature. Under these conditions, the thermochromic material admits

© 2011 by Taylor and Francis Group, LLC

 178 Green Nanotechnology

energy when there is a need for heating and rejects energy when heating
is not required.

4.10.1  Metal-Insulator Transition and Its Nanofeatures in VO2

There are several principles that can be used to create thermochromism.

We first consider materials capable of undergoing reversible transitions
between a high-temperature metallic-like, and hence reflecting, state and
a low-temperature dielectric-like state that is transparent. Figure  4.40
displays temperature-dependent electrical conductivity for several metal

Temperature (°C)
300 100 0 –100
104 NiS
Ti5O9

102
FeSi2 Fe3O4
Ti2O3

100
Conductivity (S/cm)

10–2
VO2

NbO2
10–4
Ti4O7

10–6
V2O3

10–8
0 1 2 3 4 5 6 7 8 9
103/T (K)

Figure 4.40  Electrical conductivity versus reciprocal temperature for

several metal compounds. (From G. V. Jorgenson, J. C. Lee, in Large-
Area Chromogenics: Materials and Devices for Transmittance Control,
edited by C. M. Lampert, C. G. Granqvist, The International Society
for Optical Engineering, Bellingham, WA, 1990; pp. 142–159. With
permission.)

© 2011 by Taylor and Francis Group, LLC

 Visual Indoors–Outdoors Contact and Daylighting 179

compounds [61]. Only two of them show transitions in the vicinity of

room temperature, and only one has been investigated in detail, namely,
vanadium dioxide, which has τc ≈ 68 °C. Its conductivity is shifted
abruptly by several orders of magnitude at the transition.
Bulk crystals of VO2 transform between a semiconducting, non-
magnetic, and relatively IR-transparent state with monoclinic structure
below τc to a metallic, paramagnetic, and IR-reflecting state with tetrag-
onal structure above τc. The physics behind the structural transforma-
tion has been discussed for decades but is still not fully understood. The
metal phase is most unusual and belongs to the class of materials known
as “bad metals” because the relaxation rate ωτ is so high that the carriers
would travel less than one lattice spacing between collisions if they had
the normal Fermi velocity. Properties in both phases depend on nano-
structure and doping, but not in the expected way through increased
relaxation rate but via changes in band structure [62]. Crystals of VO2
tend to disintegrate after many structural transitions but this is not so
for VO2-based films.
Indirect evidence for percolation at the metal-insulator transi-
tion in VO2 films has existed for many years. Recently, the notion of a
gradual two-phase transition has been verified in direct measurements
using scanning near-field infrared microscopy. Figure 4.41 shows such
images for a series of temperatures between 68°C and 70°C [63] for
films made by sol-gel technology. Clearly, the metallically conducting
regions nucleate at different places and merge gradually as the tempera-
ture is increased (i.e., VO2 is a clear case of a nanomaterial at the metal-
insulator transition). There is some resemblance to the formation of a
continuous metallic film on a glass substrate (cf. Figures 4.8 and 4.9).
The main percolation transition appears to be unusual in that it occurs
at different fractions of the metal phase under heating and cooling [62].

4.10.2 Thermochromism in VO2-Based Films, and How

to Adjust the Metal-Insulator Transition

Figure 4.42 shows typical data on the thermochromism in the IR for VO2

films, in this case for 160-nm-thick films made by sputter deposition onto
glass at 400°C [64]. The VO2 film transforms between states with high
and low transmittance at τc ≈ 60°C upon heating, and cooling brings
back the original transmittance with ~10°C hysteresis (not shown).
Clearly, the metal-insulator transition takes place at an undesirably high
temperature for buildings-related applications, but τc can be lowered to
room temperature by addition of a few atomic percent of tungsten, as
also shown in Figure  4.42. The addition of W also tends to blur the

© 2011 by Taylor and Francis Group, LLC

 180 Green Nanotechnology

Figure 4.41  A color version of this figure follows page 200. Scanning
near-field micrographs of 4 x 4 μm areas of VO2 films as measured at the
shown temperatures. Dark blue represent an insulating state and light
blue, green, and red colors represent metallic conductivity. (From M. M.
Quazilbash et al., Science 318 [2007] 1750–1753. With permission.)

© 2011 by Taylor and Francis Group, LLC

 Visual Indoors–Outdoors Contact and Daylighting 181

100

x: WxV1–xO2
80
0.006
Transmittance at λ = 2.5 µm (%) 0
0.010
0.017
60
0.030

40 0.032

20

0
0 20 40 60 80 100
Temperature (°C)

Figure 4.42  Temperature dependent transmittance at λ = 2.5 μm for

sputter-deposited WxV1–xO2 films having the shown magnitudes of x.
(From M. A. Sobhan et al., Solar Energy Mater. Solar Cells 44 [1996]
451–455. With permission.)

metal-insulator transition. Figure 4.43 shows that τc drops linearly with

the amount of tungsten doping [64]. The rate of this decrease seems to
depend on the quality of the VO2 films, and τc falls off by ~28°C/at %W
in bulk-like samples [65].
A very sharp transition between the two states may not be desired
for a practical application to windows. The reason is that a sharp transi-
tion in a narrow temperature range would tend to give rise to undesired
optical effects with visible “color fronts” moving across the window
under heating and cooling through τc.
Figure 4.44 shows spectral transmittance in the solar range for films
of VO2 and W0.032V0.968O2 below and above τc [64]. It is evident that
the transmittance is undesirably low and Tlum does not exceed 40%. Of
course, a thinner film would display higher transmittance, but at the
expense of degraded thermochromism (i.e., a less pronounced switching
of the NIR transmittance and hence of Tsol). This points at a real prob-
lem of too high luminous absorptance in pure VO2 as well as in W-doped
VO2 . This feature has severely curtailed the applicability of thermochro-
mic window coatings for many years. However, there are some recently
discovered ways to boost the transmittance, as considered next.

© 2011 by Taylor and Francis Group, LLC

 182 Green Nanotechnology

80

60

Transition Temp. (°C)

40

20
WxV1–xO2

0
0 0.02 0.04 0.06
Doping Level, x

Figure 4.43  Thermochromic transition temperature τc versus tung-

sten content in sputter-deposited WxV1–xO2 films. The line was drawn as
a fit to the data points. (From M. A. Sobhan et al., Solar Energy Mater.
Solar Cells 44 [1996] 451–455. With permission.)

100
100

VO2 W0.032 V0.968O2

Transmittance (%)
Transmittance (%)

20°C
50 50
20°C

80°C 80°C

0 0
0 1 2 0 1 2
Wavelength (µm) Wavelength (µm)
(a) (b)

Figure 4.44  Spectral transmittance of films of VO2 (panel a) and

W0.032V0.968O2 (panel b) at the shown temperatures. (From M. A.
Sobhan et al., Solar Energy Mater. Solar Cells 44 [1996] 451–455. With
permission.)

© 2011 by Taylor and Francis Group, LLC

 Visual Indoors–Outdoors Contact and Daylighting 183

50

Luminous Transmittance (%) 45

40

0 2 4 6 8
Doping Level (%)

Figure 4.45  Luminous transmittance versus doping level in 50-nm

thick Mg xV1-xO2 films made by sputtering. (From N. R. Mlyuka et al.,
Appl. Phys. Lett. 95 [2009] 171909 1–3. With permission.)

4.10.3  How to Enhance the Luminous Transmittance in VO2

There are several ways to enhance Tlum in thermochromic VO2-based

films. One recently discovered possibility is to dope with Mg, and
Figure 4.45 shows that Tlum increases noticeably when the value of x in
Mg xV1-xO2 goes up [66]. Interestingly, τc drops upon Mg doping, though
not as fast as in WxV1-xO2 .
Fluorination offers another possibility to boost Tlum, and Figure 4.46
illustrates the difference between films made in the presence and absence
of a fluorine-containing gas in the sputter plasma [67]. Clearly, the fluo-
rination can enhance Tlum by up to ~10%. Antireflection treatment is,
of course, possible in order to increase the transmittance of VO2-based
films. This is illustrated in Figure  4.47, where the upper panel reports
spectral transmittance for a TiO2 /VO2 /TiO2 stack with a 50-nm-thick
VO2 film and 75-nm-thick dielectric TiO2 films as measured at 20°C
and 100°C, together with results for a single 50-nm-thick VO2 film
[68]. The maximum transmittance is strikingly higher in the three-layer
structure and reaches 81.5% at λ ≈ 740 nm, while the single-layer film
peaks at 56.1% for λ ≈ 710 nm. Integrated values of Tlum were 62.8
and 58% below and above τc , respectively, for the three-layer struc-
ture, whereas the single-layer film had 42 and 45% as corresponding
measures. Considering Tsol, the three-layer structure had 63 and 57%
below and above τc , respectively, and the single-layer film correspond-
ingly displayed 47 and 42.5%.

© 2011 by Taylor and Francis Group, LLC

 184 Green Nanotechnology

80
VOx Fy
VO2
60

Luminous Transmittance (%)

40

20

20 50 100 200 500

Thickness (nm)

Figure 4.46  Luminous transmittance versus film thickness for sputter

deposited films of VO2 and VOxFy.Curves were drawn for convenience
only. (From K. A. Khan, C. G. Granqvist, Appl. Phys. Lett. 55 [1989]
4–6. With permission.)

The lower panel of Figure  4.47 shows spectral transmittance for

a TiO2 /VO2 /TiO2 /VO2 /TiO2 stack with 50-nm-thick VO2 films and
130-nm-thick TiO2 films [68,69]. These data are compared with cor-
responding ones for a 100-nm-thick VO2 film; the total amount of VO2
hence was the same in the five-layer and single-layer configurations.
Here Tlum was 45 and 42.3% below and above τc , respectively, for the
five-layer structure, whereas the single-layer film had 41 and 40% as cor-
responding data. Considering Tsol, the five-layer structure had 52.1 and
40% below and above τc , respectively, and the single-layer film displayed
41 and 34.3%. The strong enhancement of the modulation of Tsol, as a
result of the three dielectric layers, is noteworthy. The use of multiple
coatings makes it possible to induce stress in the VO2-type thin films. In
this manner it is possible to influence τc to some extent too.
A different approach to high-transmittance thermochromic materi-
als may be to incorporate VO2-based pigments in polymer foils in the
manner described for LaB6 and other pigments in Section 4.5.3. Recent
results on plasmonically resonant VO2 pigments have given interesting
results [70].
There are several other ways to create a thermochromic effect in
windows. One of them uses cloud gels and is outlined in Box 4.4.

© 2011 by Taylor and Francis Group, LLC

 Visual Indoors–Outdoors Contact and Daylighting 185

80 20°C

20°C
60
Transmittance (%)

40
100°C
100°C
20 TiO2/VO2/TiO2
VO2

0
500 1000 1500 2000 2500
Wavelength (nm)
100

80
20°C
Transmittance (%)

60 20°C

40
100°C

20 100°C
TiO2/VO2/TiO2/VO2/TiO2
VO2
0
500 1000 1500 2000 2500
Wavelength (nm)

Figure 4.47  Spectral normal transmittance for three-layer TiO2 /VO2 /

TiO2 films (upper panel) and five-layer TiO2 /VO2 /TiO2 /VO2 /TiO2 films
(lower panel), as specified in the main text, at two temperatures, one
lying below the metal-insulator transition temperature and the other
above this temperature. The data are compared with those for corre-
sponding VO2 films. (From N. R. Mlyuka et al., Solar Energy Mater.
Solar Cells 93 [2009] 1685–1687. With permission.)

© 2011 by Taylor and Francis Group, LLC

 186 Green Nanotechnology

Box 4.4 Thermochromism
in Cloud Gels
Thermochromic windows can include a layer of a polymeric “cloud
gel” [71,72]. Clouding—that is, transition to a diffusely scatter-
ing state—can set in above a temperature τc due to reversible ther-
mochemical dissolution and a thermally induced modification
of the length of the polymer molecules. This means, essentially,
that the material changes between a “grainy” and light-scatter-
ing state and a homogeneous and nonscattering state, as indi-
cated in Figure B4.4.1. The value of τc can lie anywhere within
a wide range by proper choice of material. Figure B4.4.2 shows
the total transmission (direct and diffuse) of radiation through a
one-millimeter-thick cloud gel layer between two glass panes [3].
Both Tlum and Tsol can drop by ~50% when τc is exceeded. Apart
from windows-related applications, “cloud gels” are of interest
for overheat protection, for example in plastic solar collectors (cf.
Section 6.1) [73].

Clear state
(low temperature) (high temperature)

Homogeneous mixture Scattering material

Cover/ substrate Matrix material

Figure B4.4.1  Schematics of the low temperature clear state

and the high temperature backscattering state in a thermochromic
glazing unit. (After P. Nitz, H. Hartwig, Solar Energy 79 [2005]
573–582. With permission.)

© 2011 by Taylor and Francis Group, LLC

 Visual Indoors–Outdoors Contact and Daylighting 187

100

Clear
80
Transmittance (%)

60

40 Clouded

20

0
0.5 1 1.5 2 2.5
Wavelength (µm)

Figure B4.4.2  Spectral total transmittance through a cloud gel

layer between glass panes. The data refer to a low-temperature clear
state and a high-temperature clouded state. (From C. G. Granqvist,
in Materials Science for Solar Energy Conversion Systems, edited
by C. G. Granqvist, Pergamon, Oxford, U.K., 1991; chap. 5,
pp. 106–167. With permission.)

4.11  Electrochromics

Electrochromics is different from photochromics and thermochromics

in that it allows the transmittance to be easily changed by an opera-
tor, and it is hence more flexible regarding applications than the other
“chromics.” It is also possible to construct devices with a large optical
modulation, as illustrated in Figure  4.1. However, electrochromism is
more complicated than photochromics and thermochromics, which rely
on the change of a single material. In contrast to this, an electrochromic
device includes several layers, and the optical modulation is connected to
electrical charge being shuttled between different layers.
Electrochromism was discovered and made widely known in the late
1960s and early 1970s [74]. Early applications were sought for informa-
tion displays, but electrochromic (EC) displays did not stand up to the
then rapidly developing liquid crystal displays. The focus on EC research
was changed to windows during the mid-1980s when it was realized that
the technology had “green” attributes and could yield energy efficiency

© 2011 by Taylor and Francis Group, LLC

 188 Green Nanotechnology

and user comfort in buildings. The concept of a “smart” window was

then coined and captured interest both from researchers and the general
public [75].
Electrochromism is well known both in a number of transition
metal oxides and in several classes of organic compounds, but only the
former are likely to be durable enough for long-time uses in architectural
windows. Prior work on EC materials and devices has been summarized
in considerable detail [76,77].

4.11.1  How Do Electrochromic Devices Work?

Figure  4.48 shows a standard design of an EC device [76]. There are

five layers backed by one substrate or positioned between two substrates
by use of lamination. The substrates are normally of glass, but plastic
works, too, and flexible PET foil allows for interesting devices. The cen-
tral part of the device conducts ions but not electrons. This can be an
organic material, such as a transparent polymeric electrolyte or an ionic
liquid, preferably with adhesive properties. It can also be a thin film, for
example, a porous oxide incorporating ions. The ions should be small in
order to be easily mobile in an electric field, and protons (H+) and Li+ are
of concern in most cases.

)
r

ELECTROCHROMIC r
cto

to lm
uc c fi
t c lm

DEVICE
du

d i
en ic fi

on m om
on

t c e fil chr or
sp om

n
e ag ro uct te)
an hr

r
a r t d ly
ar

s) nsp sto elec on ctro

Tr troc

a s c
ass

l a n r n le
Tr Io (o
ec

(G Io (e
Gl
El

++ ––
+ –
+ –
+ –
+ –
+ –
+ –
+ Ions –

+ –

Figure 4.48  Basic design of an electrochromic device. The transport

of positive ions in an electrical field is indicated. (From C. G. Granqvist,
Handbook of Inorganic Electrochromic Materials, Elsevier, Amsterdam,
the Netherlands, 1995. With permission.)

© 2011 by Taylor and Francis Group, LLC

 Visual Indoors–Outdoors Contact and Daylighting 189

The ion conductor is in contact with an EC film capable of conduct-

ing both ions and electrons (i.e., a mixed conductor). Tungsten oxide
is a typical example. On the other side of the ion conductor is an ion
storage film, which also is a mixed conductor for ions and electrons.
Ideally, this has EC properties which are complementary to those of the
first EC film. This central three-layer stack is positioned between two
transparent conducting films. ITO is often preferred because it has an
unsurpassed combination of optical transparency and electrical conduc-
tivity, but SnO2-based films can be used and offer cost benefits, espe-
cially if the SnO2-based film is produced on glass by spray pyrolysis.
Metal-based transparent conductors have not been tried to any large
extent but should not be written off.
When a voltage of the order of 1 V is applied between the transpar-
ent conductors, ions can be transported between the EC film and the
ion storage film. The charge of the ions is then balanced by electrons
that are injected into or withdrawn from the EC film and ion storage
films via the transparent conductors, and these electrons are the cause
of the optical absorption as we will see shortly. Reversal of the volt-
age or, with suitable materials, short circuiting brings back the original
properties. The coloration can be halted at any intermediate level, which
means that the EC device has open circuit memory. The fact that power
is needed only to change the optical properties is important for window-
type devices designed for energy savings. The memory effect hinges on
the fact that the ion conductor in the middle of the EC device does not
conduct electrons, which is easier to accomplish with a laminate layer
having a thickness on the order of several micrometers than with a thin
ion-conducting film. The voltage level needed to move the ions is of the
order of a few volt dc, which can be easily obtained by solar cells, and a
number of EC devices with integral solar-cell-based powering have been
researched [78].
It should now be clear that the EC device resembles an electrical
battery with a charging state that corresponds to a degree of optical
absorption. The analogy is useful and the two types of devices share
many pros and cons. For example, both of them can easily degrade if
they are mistreated by overcharging or overheating, but they also exhibit
certain “self healing” features which are poorly understood. Also, both
kinds of devices are unable to change their properties abruptly, and in
the case of an EC device, the time for going from fully colored to fully
bleached may vary from a few seconds in a device of the size of a few
square centimeters to minutes, or even tens of minutes, for a window
with a size of square meters.
What does it mean that the base EC film and the ion storage film
should have complementary properties? The background is that there
are oxides with two types of EC properties: those coloring under ion

© 2011 by Taylor and Francis Group, LLC

 190 Green Nanotechnology

ELECTROCHROMIC OXIDES:

H Cathodic coloration He

Li Be Anodic coloration B C N O F Ne

Na Mg Al Si P S Cl Ar

K Ca Sc Ti V Cr Mn Fe Co Ni Cu Zn Ga Ge As Se Br Kr

Rb Sr Y Zr Nb Mo Tc Ru Rh Pd Ag Cd In Sn Sb Te I Xe

Cs Ba La Hf Ta W Re Os Ir Pt Au Hg Ti Pb Bi Po At Rn

Fr Ra Ac

Figure 4.49  Periodic system of the elements (apart from the lan-
thanides and actinides). The differently shaded boxes indicate transi-
tion metal oxides whose oxides have clear “cathodic” and “anodic”
electrochromism. (From C. G. Granqvist, Handbook of Inorganic
Electrochromic Materials, Elsevier, Amsterdam, the Netherlands, 1995.
With permission.)
insertion and known as “cathodic” and those coloring under ion extrac-
tion and known as “anodic.” The terminology clearly points at the kin-
ship between EC technology and battery technology. Figure 4.49 shows
the metallic elements whose oxides exhibit the two kinds of coloration
[76]. Among the cathodic oxides, most attention has been on oxides of
W, Mo, and Nb. Among the anodic oxides, those based on Ir and Ni
stand out as most interesting. It should be noted, though, that Ir is very
rare and precious and hence ill suited for commodity-scale applications.
The only metal with different properties is vanadium for which the V2O5
can exhibit cathodic and anodic features in different wavelength regions,
but this is a special case which we return to briefly in the text to come.
By combining, say, a cathodic EC film with an anodically coloring ion
storage film, one can accomplish devices with both films becoming dark
when charge is moved from one to the other and both films bleaching
when the original charging state is brought back. This complementary
feature also can be used to create color neutrality, provided that ade-
quate combinations of “cathodic” and “anodic” films are used.

4.11.2  Facile Ion Movement Due to Favorable Nanostructures

Most of the EC oxides can be viewed as built from octagonal build-

ing blocks in different arrangements. The spaces between these octahe-
dra are large enough to allow at least some ion transport. Furthermore,
clusters of octahedra can aggregate into disordered and loosely packed

© 2011 by Taylor and Francis Group, LLC

 Visual Indoors–Outdoors Contact and Daylighting 191

Tungsten atoms
(in the plane of the page)
Oxygen atoms

Figure 4.50  Schematic illustration of corner-sharing and edge-shar-

ing octahedra in a W oxide crystal. (From C. G. Granqvist, Handbook
of Inorganic Electrochromic Materials, Elsevier, Amsterdam, the
Netherlands, 1995. With permission.)

clusters with significant intergranular spaces. The nano features there-

fore enter on two levels as further elaborated shortly. Most of the dis-
cussion concerns W oxide, which is the most widely investigated EC
material [76,79–81].
Figure 4.50 describes nanostructural features of W oxide and shows
WO6 octahedra with six oxygen atoms and a tungsten atom in the
center [76]. Stoichiometric WO3 corresponds to a structure with each
octahedron sharing corners with adjacent octahedra. W oxide has a
tendency to form substoichiometric phases in which some of the octa-
hedra are not corner sharing but edge sharing, as is also illustrated in
Figure 4.50. It is easily realized that a three-dimensional arrangement
of octahedral “building blocks” leaves a three-dimensional network of
“tunnels.” They are wide enough to serve as conduits for small ions.
The schematic crystal structure in Figure 4.50 is, in fact, not entirely
adequate because it refers to a cubic structure of W oxide, which does
not form except under high pressure. The transition metal oxides are
notorious for their large number of possible structures, and a tetragonal
structure is appropriate for WO3 crystals at normal temperature and
pressure. It is shown in Figure 4.51 that the spaces between the octahe-
dral “building blocks” are then larger than for the cubic structure, and
an even more favorable structure with regard to the possibility of ionic
movement is found in a hexagonal atomic arrangement [76]. The latter
structure seems to be common in thin films of W oxide.
Nanostructures formed in thin films of W oxides have been investi-
gated many times. Figure 4.52 reports results for films made by evapora-
tion onto substrates at different temperatures τs and represent modeling
based on x-ray scattering [76,82]. Clearly, the films exhibit cluster-type

© 2011 by Taylor and Francis Group, LLC

 192 Green Nanotechnology

Cubic

(a)
Tetragonal

(b)
Hexagonal

(c)

Figure 4.51  Atomic arrangements for W oxide with (a) cubic, (b) tet-
ragonal, and (c) hexagonal structure. Solid dots indicate sites available
for ions in the open spaces between the WO6 octahedra. Dashed lines
show the extents of the unit cells. (From C. G. Granqvist, Handbook
of Inorganic Electrochromic Materials, Elsevier, Amsterdam, the
Netherlands, 1995. With permission.)

features with clusters growing in size as τs rises. The individual clus-

ters are believed to be linked by hydrogen bonds via water molecules.
The cluster size is ~3 nm for τs = 150°C, and further cluster growth has
taken place at τs = 300°C. The clusters start to interconnect at the latter
temperature and a long-range ordered structure then starts to prevail.
Two features in Figure 4.52 should be emphasized: the existence of large
open spaces between the clusters, especially in the absence of substrate
heating, and the hexagonal nature of the individual clusters (as apparent
from a comparison with Figure 4.51c). Both of these features are advan-
tageous for ionic mobility in the oxides.
The structural models discussed above are believed to be essentially
correct, at least for the cathodic EC oxides. For the anodic ones, the situ-
ation is more complicated and the atomic distances are so small that any

© 2011 by Taylor and Francis Group, LLC

 Visual Indoors–Outdoors Contact and Daylighting 193

Substrate temperature

RT 150°C 300°C
Structural Model

Y
Y

X X X

Figure 4.52  Structural models based on connected WO6 octahedra

for W oxide films made by evaporation onto substrates at room tem-
perature (RT) and two elevated temperatures. Arrows in the x and y
directions denote 2 nm. (From C. G. Granqvist, Handbook of Inorganic
Electrochromic Materials, Elsevier, Amsterdam, the Netherlands;
T.  Nanba, I. Yasui, J. Solid State. Chem. 83 [1989] 304–315. With
permission.)
direct insertion on ions is unlikely; therefore, it seems that film porosity
and processes on the grains’ surfaces are of overriding importance.

4.11.3  What Causes the Optical Absorption?

The origin of the optical absorption in electrochromic oxides has been

the subject of much research. It is a complicated subject for various rea-
sons, one being that oxides are poorly defined with regard to crystallinity
and can incorporate mobile ions and water molecules to varying degrees.
Another difficulty ensues from the fact that even the intrinsic electronic
structure for materials such as NiO has been debated for decades with-
out any consensus having been reached. We approach the optical absorp-
tion mechanisms in the EC materials step by step in Box 4.5 and focus
on a particularly interesting combination of “cathodically” and “anodi-
cally” coloring EC oxides: those based on W and Ni [81].

4.11.4  Some Device Properties

Many different types of electrochromic devices have been studied over

the years. Recently much interest has been focused on combinations of
cathodic W oxide and anodic Ni oxide joined via with an electrolyte
being an ion-containing inorganic thin film or an organic ion-conducting

© 2011 by Taylor and Francis Group, LLC

 194 Green Nanotechnology

Box 4.5  Electrochromism in W Oxide

and Ni Oxide: A Detailed View
We first consider insertion and extraction of protons (hydrogen
ions) and electrons in WO3 by the simple electrochemical reaction

[WO3 + H+ + e−]bleached ↔ [HWO3]colored (B4.5.1)

where it should be remembered that H+ could be replaced by some

other small ion such as Li+, and that the reaction should only be
partial in order to be reversible so that the colored compound
should be written H xWO3 with x < 0.5. For the Ni-based oxide,
the corresponding reaction, which is expected to be confined to
hydrous grain boundaries, is
[Ni(OH)2]bleached ↔ [NiOOH + H+ + e−]colored (B4.5.2)
An understanding of the optical properties under ion and electron
exchange and of the principal differences between the cathodic
and anodic can be developed by considering the electronic band
structure appropriate for oxides comprised of octahedral “building
blocks.” The oxygen 2p bands are separated from the metal d lev-
els, and octahedral symmetry leads to the splitting of the d levels
into bands with the conventional notation eg and t 2g. We refer to
the literature for a detailed discussion of these matters [83].
Figure B4.5.1 illustrates schematically the cases believed to be
relevant for typical cathodic and anodic oxides. The left-hand
panel—for H xWO3, say—shows that the O2p band is separated
from the split d band by an energy gap. Pure WO3 has a full O2p
band and an empty d band and hence is transparent as any semi-
conductor characterized by a wide-enough band gap. Insertion of
small ions and accompanying electrons leads to a partial filling of
the d band, accompanied by optical absorption as discussed in the
text following. Incidentally, the filling of the lowest states in the
band permits optical transitions across the band gap only with a
larger energy than in the case of undoped WO3, which is the same
mechanism as the one leading to band gap widening, for example,
by Sn doping an In 2O3 to make ITO (cf. Figure 4.23).
The middle panel in Figure B4.5.1 applies to the anodic oxides.
Here, the “pure” oxide has some unoccupied t 2g states, and inser-
tion of ions and electrons may fill these states to the top of the band
so that the material becomes characterized by a band gap between
the eg and t 2g states. Transparency then prevails provided that this
gap is wide enough.

© 2011 by Taylor and Francis Group, LLC

 Visual Indoors–Outdoors Contact and Daylighting 195

HxWO3 HxNiO2? V2O5

E E E
eg eg

t2g t2g V3d

O2p O2p O2p

Figure B4.5.1  Schematic band structures for different classes

of EC oxides. Shaded regions denote filled states, and E signi-
fies energy. (From C. G. Granqvist, Handbook of Inorganic
Electrochromic Materials, Elsevier, Amsterdam, the Netherlands,
1995. With permission.)

For V2O5, finally, the structure deviates sufficiently from the

octahedra-based one that the d band displays a narrow low-energy
portion lying in the band gap, as illustrated in the right-hand part
of Figure B4.5.1. Low-level doping of V2O5 leads to filling of this
split-off band and subsequent doping makes the split-off band fully
occupied so that the “effective” band gap is widened. These fea-
tures account for the fact that V2O5 is neither purely cathodic nor
anodic [84].
The detailed mechanism for the optical absorption is considered
next. It has been studied in greatest detail for W oxide. When ions
and electrons are inserted, the electrons are localized on tungsten
sites, so some of the W6+ sites are transformed to W5+ sites. By
absorbing a photon, the inserted electrons can gain enough energy
to be transferred to a neighboring site. Such transfer between sites
i and j, say, can be written schematically as

Wi5+ + Wj6+ + photon → Wi6+ + Wj5+ (B4.5.3)

This mechanism is effective only as long as transfer can take place

from a state occupied by an electron to an empty one available
to receive the electron. If the ion and electron insertion is large
enough this is no longer the case and then not only transfer of the
type W5+ ↔ W6+ is important but also W4+ ↔ W5+ and W4+ ↔ W6+.

© 2011 by Taylor and Francis Group, LLC

 196 Green Nanotechnology

In a practical situation, such “site saturation” effects may not be so

important, though, since the possibilities to have a highly revers-
ible ion exchange tends to limit the insertion levels to those where
W5+ ↔ W6+ is dominant [85].

laminate [81]. The latter configuration has been implemented on flexible

PET foil, which allows facile roll-to-roll manufacturing (cf. Figure A1.5)
of products suitable for new windows as well as for retrofitting win-
dows. The properties of such a foil are described in Box 4.6.

4.11.5  Alternative Electrochromic Devices

Several alternative EC device types have been studied over the years
[9]. Thus, there are metal hydrides with some added catalysts that can
change from a transparent, via an absorbing, to a reflecting state upon
hydrogen exchange. The fact that a reflecting state can be reached is an
asset since such a film does not heat up as much as an absorbing one in
a typical windows-related application. However, the values of Tlum have
so far been too low for general applications and long-term durability has
not been demonstrated. These metal hydrides also exhibit gasochromic
features [86].
Reversible electroplating of a metal onto a glass surface from an
adjacent electrochemically active substance is another conceivable tech-
nique for reflectance modulation [87]. It has been researched for years,
but no practical solution has yet been demonstrated for windows.
Other types of materials require constant electrical powering in
order to remain dark or transparent and revert spontaneously to the
other state in the absence of such powering. The drawbacks with regard
to energy savings are obvious. These technologies include solution- and
gel-based EC redox systems such as those used in today’s antidazzling
rear view mirrors for cars and trucks, and polymer-encapsulated liquid
crystal devices that can switch between states of high and low scatter-
ing [88] and being somewhat analogous to the “cloud-gels” discussed in
Box 4.4. Systems incorporating suspended rod-like particles can serve as
“light valves” and can change their transparency, depending on whether
or not the particles are aligned in an electrical field [89].

© 2011 by Taylor and Francis Group, LLC

 Visual Indoors–Outdoors Contact and Daylighting 197

Box 4.6  Properties of an Electrochromic

Foil Based on Films of W Oxide and Ni Oxide
Figure B4.6.1 shows optical properties for a laminated device with
W oxide and Ni oxide. The transmittance in the luminous and
solar ranges can be modulated between widely separated extrema,
which is due to a change of the absorption rather than of the reflec-
tion. The reflectance is not completely identical when the device is
viewed from the two sides.

90 Bleached Bleached
Colored Colored
80
Transmittance (%)

70
60
50
40
30
20
10

35

30 PET PET
Reflectance (%)

TCO TCO
25 WOx NiO
PE PE
NiOx WOx
20 TCO TCO
PET PET
15

10

90
80
Absorptance (%)

70
60
50
40
30
20
10

550 1100 1650 2200 550 1100 1650 2200

Wavelength (nm) Wavelength (nm)

Figure B4.6.1  Spectral transmittance, reflectance and absorp-

tance as measured for the EC device shown in the inset. It incorpo-
rates films based on W oxide, Ni oxide, and transparent conducting
oxide (TCO), as well as a polymer electrolyte (PE) and PET foils.

© 2011 by Taylor and Francis Group, LLC

 198 Green Nanotechnology

The transmittance modulation during several color/bleach

cycles for a similar EC foil device, 240 cm 2 in size, is shown in
Figure B4.6.2. It is evident that the optical changes are not abrupt
but take place during a time span of tens of seconds for devices
of this size. The modulation range can be changed by having dif-
ferent thicknesses of the EC films. Using foil devices it is also
feasible to make tandem foils in order to reach very low transmit-
tance levels in the colored state. An example of this is shown in
Figure B4.6.3.

70

60

50
Transmittance (%)

40

30

20

10
500 1000 1500 2000
Time [s]

Figure B4.6.2  Transmittance versus time at λ = 550 nm for

repeated coloring and bleaching of an electrochromic foil device of
the type shown in Figure B4.6.1.

© 2011 by Taylor and Francis Group, LLC

 Visual Indoors–Outdoors Contact and Daylighting 199

50

40

Transmittance (%)
30

20

10

0
0 50 100 150 200 250 300
Time [s]

Figure B4.6.3  Transmittance versus time at λ = 550 nm for

repeated coloring and bleaching of two superimposed electrochro-
mic foil devices of the type shown in Figure B4.6.1.

References

1. L. A. B. Pilkington, The float glass process, Proc. Roy. Soc. London

314 (1969) 1–25.
2. M. Wigginton, Glass in Architecture, Phaidon, London, 1996.
3. C. G. Granqvist, Energy efficient windows: Present and forthcom-
ing technology, in Materials Science for Solar Energy Conversion
Systems, edited by C. G. Granqvist, Pergamon, Oxford, U.K., 1991;
chap. 5, pp. 106–167.
4. M. Born, E. Wolf, Principles of Optics, 7th edition, Cambridge
University Press, Cambridge, U.K., 1999.
5. International Glazing Database (2009), Lawrence Berkeley National
Laboratory, CA; http://windows.lbl.gov/software/window/window.
html.
6. K. G. T. Hollands, J. L. Wright, C. G. Granqvist, Glazings and coat-
ings, in Solar Energy: The State of the Art, edited by J. Gordon, James
& James Sci. Publ., London, U.K., 2001; chap. 2, pp. 29–107.
7. H. Manz, S. Brunner, L. Wullschleger, Triple vacuum glazing: Heat
transfer and basic mechanical design constraints, Solar Energy 80
(2006) 1632–1642.
8. I. Hamberg, C. G. Granqvist, Evaporated Sn-doped In2O3 films:
Basic optical properties and applications to energy efficient win-
dows, J. Appl. Phys. 60 (1986) R123–R159.

© 2011 by Taylor and Francis Group, LLC

 200 Green Nanotechnology

9. C. G. Granqvist, Transparent conductors as solar energy materi-

als: A panoramic review, Solar Energy Mater. Solar Cells 91 (2007)
1529–1598.
10. E. Valkonen, B. Karlsson, Optimization of metal-based multilay-
ers for transparent heat mirrors, Int. J. Energy Res. 11 (1987)
397–403.
11. E. Valkonen, B. Karlsson, C.-G. Ribbing, Solar optical properties
of thin films of Cu, Ag, Au, Cr, Fe, Co, Ni, and Al, Solar Energy 32
(1984) 211–222.
12. P. C. Lansåker, J. Backholm, G. A. Niklasson, C. G. Granqvist, TiO2/
Au/TiO2 multilayer thin films: Novel metal-based transparent con-
ductors for electrochromic devices, Thin Solid Films 518 (2009)
1225–1229.
13. G. B. Smith, A. I. Maaroof, M. B. Cortie, Percolation in nanopo-
rous gold and the principle of universality from two dimensions to
hyperdimensions, Phys. Rev. B 78 (2008) 165418 1–11.
14. G. B. Smith, G. A. Niklasson, J. S. E. M. Svensson, C. G. Granqvist,
Noble-metal-based transparent infrared reflectors: Experiments and
theoretical analyses for very thin gold films, J. Appl. Phys. 59 (1986)
571–581.
15. G. A. Niklasson, C. G. Granqvist, Noble-metal-based transparent
infrared reflectors: Improved performance caused by nonhomoge-
neous film structure, Appl. Phys. Lett. 46 (1985) 713–715.
16. A. I. Maaroof, A Gentle, G. B. Smith, M. B. Cortie, Bulk and surface
plasmons in highly nanoporous gold films, J. Phys. D: Appl. Phys.
40 (2007) 5675–5682.
17. G. B. Smith, A. A. Earp, Metal-in-metal localized surface plasmon
resonance, Nanotechnology 21 (2010) 015203 1–8.
18. D. R. Sahu, S.-Y. Lin, J.-L. Huang, ZnO/Ag/ZnO multilayer films for
the application of a very low resistance transparent electrode, Appl.
Surf. Sci. 252 (2006) 7509–7514.
19. Y. Tachibana, K. Kusunoki, T. Watanabe, K. Hashimoto, H. Ohsaki,
Optical properties of multilayers composed of silver and dielctric
materials, Thin Solid Films 442 (2003) 212–216.
20. G. B. Smith, A. Maaroof, Optical response in nanostructured thin
metal films with dielectric overlayers, Opt. Commun. 242 (2004)
383–392.
21. J. Ridealgh, Large area coatings for architectural glass: Design, man-
ufacturing, processing faults, MRS Proc. 890 (2006) Y01–10.
22. D. S. Ginley, H. Hosono, D. Paine, editors, Handbook of Transparent
Conductors, Springer Science and Business Media, 2010.
23. Y. Furubayashi, T. Hitosugi, Y. Yamamoto, K. Inaba, G. Kinoda, Y.
Hirose, T. Shimada, T. Hasegawa, A transparent metal: Nb-doped
anatase TiO2, Appl. Phys. Lett. 86 (2005) 252101 1–3.

© 2011 by Taylor and Francis Group, LLC

 Visual Indoors–Outdoors Contact and Daylighting 201

24. C. M. Maghanga, G. A. Niklasson, C. G. Granqvist, Optical proper-

ties of sputter deposited transparent and conducting TiO2:Nb films,
Thin Solid Films 518 (2009) 1254–1258.
25. Y. Shigesato, D. C. Paine, Study of the effect of Sn doping on the
electronic transport properties of thin indium oxide, Appl. Phys.
Lett. 62 (1993) 1268–1270.
26. U. Betz, M. Kharazzi Olsson, J. Marthy, M. F. Escolà, F. Atamny,
Thin films engineering of indium tin oxide: Large area flat panel dis-
plays application, Surf. Coating Technol. 200 (2006) 5751–5759.
27. J. Ederth, P. Heszler, A. Hultåker, G. A. Niklasson, C. G. Granqvist,
Indium tin oxide films made from nanoparticles: Models for the opti-
cal and electrical properties, Thin Solid Films 445 (2003) 199–206.
28. F. Ruske, C. Jacobs, V. Sittinger, B. Szyszka, W. Werner, Large area
ZnO:Al films with tailored light scattering properties for photovol-
taic applications, Thin Solid Films 515 (2007) 8695–8698.
29. I. Hamberg, C. G. Granqvist, Optical properties of transparent and
heat-reflecting indium tin oxide films: The role of ionized impurity
scattering, Appl. Phys. Lett. 44 (1984) 721–723.
30. E. Gerlach, Carrier scattering and transport in semiconductors
treated by the energy-loss method, J. Phys. C: Solid State Phys. 19
(1986) 4585–4603.
31. C. S. S. Sangeth, M. Jaiswal, R. Menon, Charge transport in trans-
parent conductors: A comparison, J. Appl. Phys. 105 (2009) 063713
1–6.
32. J.-Y. Lee, S. T. Connor, Y. Cui, P. Peumans, Solution-processed
metal nanowire mesh transparent electrodes, Nano Lett. 8 (2008)
689–692.
33. A. K. Geim, K. S. Novoselov, The rise of graphene, Nature Mater. 6
(2007) 183–191.
34. Y.-X. Zhou, L.-B. Hu, G. Grüner, A method of printing carbon nano­
tube thin films, Appl. Phys. Lett. 88 (2006) 123109 1–3.
35. B. Dan, G. C. Irvin, M. Pasquali, Continuous and scalable fabrica-
tion of transparent conducting carbon nanotube films, ACS Nano 3
(2009) 835–843.
36. L. Hu, D. S. Hecht, G. Grüner, Infrared transparent carbon nano-
tube thin films, Appl. Phys. Lett. 94 (2009) 081103 1–3.
37. X. Wang, L. Zhi, K. Müllen, Transparent, conductive graphene elec-
trodes for dye-sensitized solar cells, Nano Lett. 8 (2008) 323–327.
38. G. B. Smith, C. A. Deller, P. D. Swift, A. Gentle, P. D. Garrett, W. K.
Fisher, Nanoparticle-doped polymer foils for use in solar control
glazing, J. Nanoparticle Res. 4 (2002) 157–165.
39. S. Schelm, G. B. Smith, Dilute LaB6 nanoparticles in polymer as
optimized clear solar control glazing, Appl. Phys. Lett. 82 (2003)
4346–4348.

© 2011 by Taylor and Francis Group, LLC

 202 Green Nanotechnology

40. C. Genet, T.W. Ebbesen, Light in tiny holes, Nature 445 (2007)
39–46.
41. H. Gao, J. Henzie, T. W. Odom, Direct evidence for surface plasmon-
mediated enhanced light transmission through metallic nanohole
arrays, Nano Lett. 6 (2006) 2104–2108.
42. R. J. Leyrer, personal communication to one of the authors (GBS).
43. X. He, Y. Thomann, R. J. Leyrer, J. Rieger, Iridescent colors from
films made of polymeric core-shell particles, Polymer Bull. 57
(2006) 785–796; W. Wohlleben, F. W. Bartels, S. Altmann, J. R.
Leyrer, Mechano-optical octave-tunable elastic colloidal crystals
made from core-shell polymer beads with self-assembly techniques,
Langmuir 23 (2007) 2961–2969.
44. G. Mbise, G. B. Smith, G. A. Niklasson, C. G. Granqvist, Angular
selective window coatings: Theory and experiment, Proc. Soc.
Photo-Opt. Instrum. Engr. 1149 (1989) 179–199.
45. A. Roos, P. Polato, P. A. van Nijnatten, M. G. Hutchins, F. Olive, C.
Anderson, Angular-dependent optical properties of low-E and solar
control windows: Simulations versus measurements, Solar Energy
69 (supplement) (2000) 15–26.
46. I. R. Maestre, J. L. Molina, A. Roos, J. F. Coronel, A single-thin-film
model for the angle dependent optical properties of coated glazings,
Solar Energy 81 (2007) 969–976.
47. A. G. Dirks, H. J. Leamy, Columnar microstructures in vapor-depos-
ited thin films, Thin Solid Films 47 (1977) 219–233.
48. M. J. Brett, Simulation of structural transitions in thin films, J. Mater.
Sci. 24 (1989) 623–626.
49. J. J. Steele, M. J. Brett, Nanostructure engineering in porous colum-
nar thin films: Recent advances, J. Mater. Sci: Mater. Electron. 18
(2007) 367–379.
50. G. B. Smith, Theory of angular selective transmittance in oblique
columnar thin films containing metal and voids, Appl. Opt. 29
(1990) 3685–3693.
51. G. W. Mbise, D. Le Bellac, G. A. Niklasson, C. G. Granqvist, Angular
selective window coatings: Theory and experiments, J. Phys. D:
Appl. Phys. 30 (1997) 2103–2122.
52. C. M. Lampert, C. G. Granqvist, editors, Large-Area Chromogenics:
Materials and Devices for Transmittance Control, The International
Society for Optical Engineering, Bellingham, WA, 1990; Vol. IS4.
53. A. Georg, A. Georg, W. Graf, V. Wittwer, Switchable windows with
tungsten oxide, Vacuum 82 (2008) 730–735.
54. A. Roos, M.-L. Persson, W. Platzer, M. Köhl, Energy efficiency
of switch­able glazing in office buildings, in Proceedings Glass
Processing Days, Tampere, Finland, 2005; pp. 566–569.

© 2011 by Taylor and Francis Group, LLC

 Visual Indoors–Outdoors Contact and Daylighting 203

55. A. Jonsson, Optical Characterization and Energy Simulation of

Glazing for High-Performance Windows, Ph.D. Thesis, Department
of Engineering Sciences, Uppsala University, Uppsala, Sweden,
2010.
56. E. S. Lee, S. E. Selkowitz, R. D. Clear, D. L. DiBartolomeo, J. H.
Klems, L. L. Fernandes, G. J. Ward, V. Inkarojrit, M. Yazdanian,
Advancement of electrochromic windows, California Energy
Commission, PIER, 2006; CEC-500-2006-052.
57. UNEP, Buildings and Climate Change: Status, Challenges and
Opportunities, United Nations Environment Programme, Paris,
France, 2007.
58. R. D. Clear, V. Inkarojrit, E. S. Lee, Subject responses to electrochro-
mic windows, Energy Buildings 38 (2006) 758–779.
59. H. J. Hoffman, in Photochromic Materials and Systems, edited by H.
Dürr, H. Bouas-Laurent, Elsevier, Amsterdam, the Netherlands, 1990.
60. L. Malfatti, S. Costacurta, T. Kidchob, P. Innocenzi, M. Casula, H.
Amenitsch, D. Dattilo, M. Maggini, Mesostructured self-assem-
bled silica films with reversible thermo-photochromic properties,
Microporous Mesoporous Mater. 120 (2009) 375–380.
61. G. V. Jorgenson, J. C. Lee, Thermochromic materials and devices:
Inorganic systems, in Large-Area Chromogenics: Materials and
Devices for Transmittance Control, edited by C. M. Lampert,
C. G. Granqvist, The International Society for Optical Engineering,
Bellingham, WA, 1990; pp. 142–159.
62. A. Gentle, G. B. Smith, Dual metal-insulator and insulator-insulator
switching in nanoscale and Al-doped VO2, J. Phys. D: Appl. Phys. 40
(2008) 1–5.
63. M. M. Quazilbash, M. Brehm, B.-G. Chae, P.-C. Ho, G. O. Andreev,
B.-J. Kim, S. J. Yun, A. V. Balatsky, M. G. Maple, F. Keilmann, H.-T.
Kim, D. N. Basov, Mott transition in VO2 revealed by infrared spec-
troscopy and nano-imaging, Science 318 (2007) 1750–1753.
64. M. A. Sobhan, R. T. Kivaisi, B. Stjerna, C. G. Granqvist,
Thermochromism of sputter deposited WxV1-xO2 films, Solar Energy
Mater. Solar Cells 44 (1996) 451–455.
65. J. B. Goodenough, The two components of the crystallographic
transition in VO2, J. Solid State Chem. 3 (1971) 490–500.
66. N. R. Mlyuka, G. A. Niklasson, C. G. Granqvist, Mg doping of
thermochromic VO2 films enhances the optical transmittance and
decreases the metal-insulator transition temperature, Appl. Phys.
Lett. 95 (2009) 171909 1–3.
67. K. A. Khan, C. G. Granqvist, Thermochromic sputter-deposited
vanadium oxyfluoride coatings with low luminous absorptance,
Appl. Phys. Lett. 55 (1989) 4–6.

© 2011 by Taylor and Francis Group, LLC

 204 Green Nanotechnology

68. N. R. Mlyuka, G. A. Niklasson, C. G. Granqvist, Thermochromic

multilayer films of VO2 and TiO2 with enhanced transmittance,
Solar Energy Mater. Solar Cells 93 (2009) 1685–1687.
69. N. R. Mlyuka, G. A. Niklasson, C. G. Granqvist, Thermochromic
VO2-based multilayer films with enhanced luminous trans-
mittance and solar modulation, Phys. Stat. Sol. A 206 (2009)
2155–2160.
70. H. Bai, M. B. Cortie, A. I. Maaroof, A. Dowd, C. Kealley, G. B.
Smith, The preparation of a plasmonically resonant VO2 thermo-
chromic pigment, Nanotechnology 20 (2009) 085607 1–9.
71. P. Nitz, H. Hartwig, Solar control with thermotropic layers, Solar
Energy 79 (2005) 573–582.
72. O. Muehling, A. Seboth, T. Haeusler, R. Ruhmann, E. Potechius, R.
Vetter, Variable solar control using thermotropic core/shell particles,
Solar Energy Mater. Solar Cells 93 (2009) 1510–1517.
73. K. Resch, G. M. Wallner, R. Hausner, Phase separated thermotropic
layers based on UV cured acrylate resins: Effect of material formula-
tion on overheating protection properties and application in a solar
collector, Solar Energy 83 (2009) 1689–1697.
74. S. K. Deb, Opportunities and challenges in science and technology
of WO3 for electrochromic and related applications, Solar Energy
Mater. Solar Cells 92 (2008) 245–258.
75. J. S. E. M. Svensson, C. G. Granqvist, Electrochromic coatings for
“smart windows,” Solar Energy Mater. 12 (1985) 391–402.
76. C. G. Granqvist, Handbook of Inorganic Electrochromic Materials,
Elsevier, Amsterdam, the Netherlands, 1995.
77. P. Monk, R. Mortimer, D. Rosseinsky, Electrochromism and
Electrochromic Devices, Cambridge University Press, Cambridge,
U.K., 2007.
78. C. M. Lampert, Large-area smart glass and integrated photovolta-
ics, Solar Energy Mater. Solar Cells 76 (2003) 489–499.
79. C. G. Granqvist, Electrochromic tungsten oxide films: Review of
progress 1993–1998, Solar Energy Mater. Solar Cells 60 (2000)
201–262.
80. C. G. Granqvist, E. Avendaño, A. Azens, Electrochromic coatings
and devices: Survey of some recent advances, Thin Solid Films 442
(2003) 201–211.
81. G. A. Niklasson, C. G. Granqvist, Electrochromics for smart win-
dows: Thin films of tungsten oxide and nickel oxide, and devices
based on these, J. Mater. Chem. 17 (2007) 127–156.
82. T. Nanba, I. Yasui, X-ray diffraction study of microstructure of
amorphous tungsten trioxide films prepared by electron beam vac-
uum evaporation, J. Solid State. Chem. 83 (1989) 304–315.

© 2011 by Taylor and Francis Group, LLC

 Visual Indoors–Outdoors Contact and Daylighting 205

83. J. B. Goodenough, Metallic oxides, Prog. Solid State Chem. 5 (1971)

145–399.
84. A. Talledo, C. G. Granqvist, Electrochromic vanadium-pentoxide-
based films: Structural, electrochemical and optical properties, J.
Appl. Phys. 77 (1995) 4655–4666.
85. L. Berggren, J. C. Jonsson, G. A. Niklasson, Optical absorption in
lithiated tungsten oxide thin films: Experiment and theory, J. Appl.
Phys. 102 (2007) 083538 1–7.
86. J. L. Slack, J. C. W. Locke, S.-W. Song, J. Ona, T. J. Richardson,
Metal hydride switchable mirrors: Factors influencing dynamic
range and stability, Solar Energy Mater. Solar Cells 90 (2006)
485–490.
87. J. P. Ziegler, Status of reversible electrodeposition electrochromic
devices, Solar Energy Mater. Solar Cells 56 (1999) 477–493.
88. D. Cupelli, F. P. Nicoletta, S. Manfredi, M. Vivacqua, P. Formoso, G.
De Filpo, G. Chidichimo, Self-adjusting smart windows based on
polymer-dispersed liquid crystals, Solar Energy Mater. Solar Cells
93 (2009) 2008–2012.
89. R. Vergaz, J.-M. Sánchez-Pena, D. Barrios, C. Vásquez, P. Contreras-
Lallana, Modelling and electro-optical testing of suspended particle
devices, Solar Energy Mater. Solar Cells 92 (2008) 1483–1487.

© 2011 by Taylor and Francis Group, LLC

207
208 Green Nanotechnology
 Electric Lighting and Daylighting 209
210 Green Nanotechnology
 Electric Lighting and Daylighting 211
212 Green Nanotechnology
 Electric Lighting and Daylighting 213
214 Green Nanotechnology
 Electric Lighting and Daylighting 215
216 Green Nanotechnology
 Electric Lighting and Daylighting 217
218 Green Nanotechnology
 Electric Lighting and Daylighting 219
220 Green Nanotechnology
 Electric Lighting and Daylighting 221
222 Green Nanotechnology
 Electric Lighting and Daylighting 223
224 Green Nanotechnology
 Electric Lighting and Daylighting 225
226 Green Nanotechnology
 Electric Lighting and Daylighting 227
228 Green Nanotechnology
 Electric Lighting and Daylighting 229
230 Green Nanotechnology
 Electric Lighting and Daylighting 231
232 Green Nanotechnology
 Electric Lighting and Daylighting 233
234 Green Nanotechnology
 Electric Lighting and Daylighting 235
236 Green Nanotechnology
 Electric Lighting and Daylighting 237
238 Green Nanotechnology
 Electric Lighting and Daylighting 239
240 Green Nanotechnology
 Electric Lighting and Daylighting 241
242 Green Nanotechnology
 Electric Lighting and Daylighting 243
244 Green Nanotechnology
 Electric Lighting and Daylighting 245
246 Green Nanotechnology
 Electric Lighting and Daylighting 247
248 Green Nanotechnology
 Electric Lighting and Daylighting 249
250 Green Nanotechnology
 Electric Lighting and Daylighting 251
252 Green Nanotechnology
 Electric Lighting and Daylighting 253
254 Green Nanotechnology
 Electric Lighting and Daylighting 255
256 Green Nanotechnology
 Electric Lighting and Daylighting 257
258 Green Nanotechnology
 Electric Lighting and Daylighting 259
260 Green Nanotechnology

54. A. Y. Vorobyev, V. S. Makin, C. Guo, Brighter light sources from

black metal: Significant increase in emission efficiency of incandes-
cent light sources, Phys. Rev. Lett. 102 (2009) 234301 1–4.
 Chapter 6
Heat and Electricity
Solar Collectors and Solar Cells

This chapter deals with devices for solar absorption that produce heat
(photothermal conversion in solar collectors) and electricity (photo-
electric conversion in solar cells). These devices can be building-inte-
grated on roofs, walls and even windows, and they can also be free
standing. Figure 6.1 shows some illustrative examples, specifically of
a solar collector installation (a), a wall essentially covered with solar
cells (b), and solar cells integrated in a wall glazing (c). The latter
example shows solar cells that are very well protected by the glazing,
but obviously these windows do not permit a good visual indoors–
outdoors contact. Both photothermal and photoelectric devices lead to
heating of the air around them, which then contributes to the urban
heat island effect in big cities (cf. Section 7.2). The focus of this chapter
is on nanostructural features of essential materials for applications to
heat and electricity.

6.1  Solar Thermal Materials and Devices

Solar collectors for hot water production have been used at least since
the 1950s. Roof mounting is commonplace, and the collectors are ori-
ented to maximize the inflow of solar energy. Figure 6.2 shows the prin-
ciples of the flat plate collector comprising an absorber plate positioned
in a box which is thermally well insulated on the back and sides and has
a transparent cover—normally of glass—toward the sun. Obviously,
the glass should not absorb solar energy excessively, so it is preferable
that it has low iron content (cf. Figure 4.2) and is antireflection coated.
The absorber plate is thermally linked to a heat transfer medium that is
normally water; it goes through “riser” tubes and emerges heated at the

261

© 2011 by Taylor and Francis Group, LLC

 262 Green Nanotechnology

(a)

(b) (c)

Figure 6.1  A color version of this figure follows page 200. Photos
illustrating building-integrated devices for photothermal and pho-
toelectric conversion of solar energy. Panel (a) shows a roof-mounted
solar collector arrangement from Sweden, used to heat a number of pri-
vate houses; panel (b) is a solar-cell-covered façade at Solar-Fabrik in
Freiburg, Germany; and panel (c) is a solar cell installation forming an
architectural element on a building belonging to Fraunhofer Institute of
Solar Energy Systems in Freiburg, Germany.

upper end. Construction aspects and thermodynamics of solar collec-

tors have been described in detail in the literature [1].
The flat plate collector is only one of many different types. Another,
which is used on a large scale, particularly in China, has the absorber
placed in a vacuum tube [2]. This design cuts down on the thermal
losses but adds complexity. There is also a large variety of concentrating
solar collectors, with low concentration and possible for building inte-
gration as well as with high concentration and suitable for large-scale
power generation. High concentration is achieved by tracking reflectors
that follow the movement of the sun, either using linear trough collec-
tors with focusing onto tubular absorbers or using reflector fields with
three-dimensional tracking and concentration onto centrally positioned
tower-mounted absorbers. Direct absorption in fluids, with or with-
out blackening by added particles, has received some attention, as has

© 2011 by Taylor and Francis Group, LLC

 Heat and Electricity 263

Fluid out
Header

Risers
Cover

Absorber plate

Insulation

Fluid in

Figure 6.2  Principle design of a flat-plate solar collector.

absorption in gases containing absorbing particles [3]. The discussion

below is mainly geared toward buildings-related applications.

6.1.1  Spectral Selectivity and Its Importance

What are the desirable optical properties of the surface of the absorber
plate (i.e., of the surface that should absorb the impinging solar energy
and convert it into useful energy in a heat transfer medium)? This issue
was discussed at length in Chapter 2 but is repeated briefly here for
convenience. The primary requirement is that solar energy, at 0.3 < λ
< 3 μm, should be absorbed (i.e., Asol = 1); second, the absorbed energy
should not be reradiated as useless heat to the ambience so that thermal
emission, at 3 < λ < 50 μm, should be avoided (i.e., E therm = 0). Figure 6.3
emphasizes again that the overlap between solar and thermal proper-
ties is almost nil so that optimization in the two spectral ranges can
be achieved separately. The absorber is nontransparent, and hence the
desired spectral reflectance is

R(λ) = 0 for 0.3 < λ < 3 μm (6.1)

R(λ) = 1 for 3 < λ < 50 μm (6.2)

These properties are shown in the lower part of Figure 6.3. A surface

with a reflectance profile approximating this is referred to as a “selective
surface” or, more specifically, a “selectively solar absorbing surface.”
The magnitude of A sol and its deviation from unity, and the value of
E therm and its deviation from zero, constitute quality measures of the
surface as further discussed below. A good selective surface should have,

© 2011 by Taylor and Francis Group, LLC

 264 Green Nanotechnology

(GW/m3)
Blackbody
1.5 spectra
Solar spectrum (AM 1) (MW/m3)
1.0 +50°C 50
±0°C
0.5

0 0
1

Reflectance

0.2 0.5 1 2 5 10 20 50
Wavelength (µm)

Figure 6.3  The upper part shows a solar spectrum for one air mass
(AM1) and blackbody spectra for two temperatures. The lower part shows
reflectance for an idealized surface designed for absorbing solar energy
with minimum thermal loss. This figure is a companion to Figure 4.5.

say, Asol > 95% for normal incidence of the solar rays and E therm < 10%
for hemispherical thermal emission. The change of the optical properties
takes place at a “critical” wavelength λc that was put, somewhat arbi-
trarily, at 3 μm in Figure 6.3. This value of λc is good for applications
in which the heat transfer medium is not too hot (in practice < 100°C),
whereas the high temperatures needed in a thermal power station may
require λc ~ 2 μm or even shorter.
It is illustrative to show the effectiveness of the selective absorber
over the blackbody absorber via a calculation of the ratio of the energy
absorbed and retained by the two surfaces. This is done in Figure 6.4
for a concentration equal to unity (appropriate for a flat plate collector)
and for concentration of the solar energy by 10 times [4]. In the absence
of concentration and at room temperature, the blackbody absorber only
gives some 60% of the energy that is produced and retained by the ideal
selective absorber, and the blackbody absorber becomes practically use-
less at higher temperatures. For a solar concentration of 10 times, the
difference between the two types of surfaces is not at all as large at low
to moderate temperatures, and the difference is even less significant at
higher concentrations. Clearly, spectral selectivity is a key concept for the
regular flat plate solar collector, which is the design of most interest for
building integration. It may be less important in a highly concentrating

© 2011 by Taylor and Francis Group, LLC

 Heat and Electricity 265

1.0
0.9

0.8

0.7 X = 10

Fraction of Energy
0.6

0.5 X=1

0.4

0.3

0.2
0.1

0 200 400
Absorber Temp. (°C)

Figure 6.4  Fraction of energy absorbed and retained for a blackbody

as a function of temperature, as compared with the case of an ideal
spectrally selective absorber with λc = 2 μm, for two values of the solar
concentration X. (From B. O. Seraphin, in Solar Energy Conversion:
Solid-State Physics Aspects, Springer, Berlin, 1979; Topics in Applied
Physics Vol. 31, chap. 2, pp. 5–55. With permission.)

set-up for power generation, but spectral selectivity nevertheless is not

insignificant because the absorber constitutes only one part in a more
complicated device, and improving it even to a minor degree may make
sense economically.
Spectrally selective transmittance of the cover glass of the solar col-
lector in principle is an alternative to the spectrally selective absorber. For
this latter case one would like to have, ideally, a glass with Tsol = 1 and
Rtherm = 1 so that thermally emitted radiation from the solar absorber is
reflected back to the same surface. The infrared reflecting coating on the
downward facing glass surface could be made of doped SnO2 , In 2O3, or
ZnO, as discussed in Section 4.4. However, in order to compete with a
good spectrally selective absorber, the IR reflector must be antireflected,
not only for perpendicularly incident radiation, as in Figure  4.18, but
over a wide angular range.

6.1.2  Principles for Spectral Selectivity

Spectrally selective surfaces can be designed according to a number of dif-

ferent principles [5,6], as indicated in Figure 6.5. The most straightforward

© 2011 by Taylor and Francis Group, LLC

 266 Green Nanotechnology

Intrinsic Metal
selective material Dielectric
Substrate Metal
(a) (d)
2 µm
Antireflection coating
Silicon
Metal
Metal
(b) (e)
Dielectric SnO2 : F ,...
Metal
Dielectric Black enamel
Substrate Substrate
(c) (f )

Figure 6.5  Schematic designs of six different approaches to spec-

trally selective absorption of solar energy. (From G. A. Niklasson, C. G.
Granqvist, in Materials Science for Solar Energy Conversion Systems,
edited by C. G. Granqvist, Pergamon, Oxford, 1991; pp. 70–105. With
permission.)
of these is to have a material whose intrinsic optical properties have the
desired spectral selectivity. This approach has not been fruitful, though,
and no material with sufficiently good properties has yet been identi-
fied. Another possible way to create spectral selectivity is to coat a low-
reflecting metal with a semiconductor having a band gap corresponding
to λc. Silicon, with a band gap of ~1.1 eV, is a candidate material; another
possibility is PbS with a band gap of ~0.4 eV [7]. A problem with silicon
is that its refractive index is so high that reflective losses become exces-
sive. Antireflection treatment is possible, but the weak absorption in Si,
which demands a large thickness, makes the approach unattractive with
regard to thermal applications. Instead, the desired properties have been
obtained by use of thin films with a suitable nanostructure in the compo-
sition and/or in the surface roughness as discussed next.
Multilayer absorbers backed by metal can be tailored to have the
desired optical properties. Basically, the design principles are the same as
those underlying the metal-based transparent heat reflectors discussed in
Section 4.3.4, although the coating should have maximum solar absorp-
tance (rather than maximum Tlum or Tsol) and maximum thermal reflec-
tance. Three-layer coatings of the type Al 2O3/Mo/Al 2O3 have been rather
successful and can be manufactured with large area technologies [8].
Nanostructures with a mixture of metal particles in a dielectric host
represent another approach to spectral selectivity. The materials are often
referred to as “ceramic-metallic mixtures” or “cermets.” Their principal
optical properties can be illustrated from an effective medium calcula-
tion according to Equation 3.19, wherein for simplicity we consider an

© 2011 by Taylor and Francis Group, LLC

 Heat and Electricity 267

100

1
80

3
Transmittance (%)

10
60

30

0
10
300
40

10 3
20

0
0.2 0.5 1 2 5 10 20 50 100
Wavelength [µm]

Figure 6.6  Computed spectral transmittance for a layer comprised to

1% of Cr particles in air. Data are given (in μm) for the film thicknesses
shown. (From C. G. Granqvist, Phys. Scripta 16 [1977] 163–164. With
permission.)

essentially nonreflecting medium with 1% of Cr particles surrounded by

air [9]. As seen from Figure 6.6, the transmittance is low at short wave-
lengths and high at long wavelengths and the change in transmittance
depends on the thickness of the layer. For practically useful surfaces—
such as those discussed below—the coating is

• Backed by a reflecting metal

• The particles are embedded in a dielectric material (normally
an oxide)
• The filling factor of the metallic component is comparable to
that of the dielectric
• The nanostructural entities are nonspherical

The latter three items make it possible to decrease the coating thickness
to only a fraction of a micrometer in a coating for spectrally selective
solar absorption.
Surface roughness leads to other possibilities to create spectral selec-
tivity. If the surface has metallic protrusions separated by distances of
the order of the wavelengths for solar irradiation, this radiation will pen-
etrate into the structure and undergo multiple reflections and thereby
become absorbed, whereas radiation with longer wavelengths will not
“see” the surface as rough but as smooth and hence reflecting. These

© 2011 by Taylor and Francis Group, LLC

 268 Green Nanotechnology

features are clearly illustrated in the recent calculations reported on in

Figure 6.7 [10]. They regard a tapered subwavelength grating comprised
of a square lattice array of tungsten pyramids with a 250-nm-period in
both transverse dimensions and a height of 500 nm. The spectral reflec-
tance was computed (by a technique we do not discuss here) as a function
of light incidence represented by polar and azimuthal angles θ and ϕ,
respectively, defined in the same way as in Sections 2.3 and 4.7. Panels (b)

500

250 750
[nm]

0 500

]
[nm
750 250
500
250
[nm] 0 0
(a)
100 100
e q
e q
e q
80 80 q
q
Reflectance (%)

Reflectance (%)

60 60

40 40

20 20

0 0
1 2 3 1 2 3
Wavelength [+m] Wavelength [+m]
(b) (c)

Figure 6.7  Computed spectral reflectance for a surface consisting of

tungsten nanopyramids according to panel (a). Data are shown in (b)
and (c) as a function of polar angle θ and azimuthal angle ϕ, respectively.
(After E. Rephaeli, S. Fan, Appl. Phys. Lett. 92 [2008] 211107 1–3.
With permission.)

© 2011 by Taylor and Francis Group, LLC

 Heat and Electricity 269

and (c) in Figure 6.7 demonstrate a very low reflectance—irrespective of

the direction of the light incidence—for solar energy and a sharp onset of
reflectance for λ > λc. The approximate independence of light incidence
is, of course, a very important asset for a stationary solar collector such
as the one depicted in Figure 6.1a. If the height of the pyramids was put
to 250 nm, the value of Asol was significantly degraded, and the same was
the case if the periodicity was significantly larger than 250 nm [10].
The final approach to spectral selectivity is different and is based
on a blackbody-type surface coated with a film with selective transmit-
tance, specifically transmitting at λ < λc and reflecting at λ > λc. Such a
surface can consist of black enamel. The heavily doped wide band gap
semiconductors discussed in Section 4.4 are natural candidates for the
top layer, and SnO2:F stands out as particularly interesting on account of
its ruggedness and possibility for large-scale manufacturing by low-cost
spray pyrolysis.
Particularly detailed studies of selectively solar absorbing surfaces
have been made on Co-Al 2O3 nanocomposite films prepared by co-evap-
oration [11]. The results of that study are summarized in Box 6.1.

6.1.3 Selectively Solar-Absorbing Coatings Based on

Nanoparticles: Some Practical Examples

The case study on well-characterized Co-Al 2O3 films discussed in

Box  6.1 has set the scene for appreciating spectrally selective surfaces
used in real life, or at least developed with that in mind. The choice of
Co and Al 2O3 is less restrictive than one perhaps would believe, and the
data for Co are representative of transition metals in general (including
Cr and Ni); Al 2O3 is a typical oxide with good thermal stability and
hence well suited for solar-related applications.
In fact, there are numerous selectively solar absorbing coatings in
practical use. They are made by a variety of coating technologies such as
electroplating, anodization and other chemical conversion techniques,
evaporation, sputtering, etc. [6]. All of these coatings acquire their opti-
cal properties by a combination of metal-dielectric nanostructures and
surface roughness integrated with multilayer features (i.e., by a concoc-
tion of at least three of the mechanisms illustrated in Figure 6.5). Rather
than trying to cover the different coatings in detail, we will next pres-
ent a series of Ni-based coatings prepared by different technologies. By
doing this we will also be able to give a view on the development of the
field during the past couple of decades and of the refinements that have
been accomplished.
Electrochemical methods were used in the earliest work to produce
selectively solar absorbing surfaces. Thus electroplating was employed to

© 2011 by Taylor and Francis Group, LLC

 270 Green Nanotechnology

Box 6.1 Case Study for Co-Al2O3

Nanocomposite Films
Here we present some data from a very detailed study of Co-Al 2O3
films in order to elucidate critical parameters for the use of metal-
dielectric-based films as spectrally selective absorbers of solar
energy [11]. The films were made by e-beam co-evaporation of Co
and Al2O3 under well-controlled conditions onto unheated sub-
strates. Electron microscopy showed that a nanostructure with hcp
Co nanoparticles embedded in Al2O3 prevailed as long as the Co
volume fraction fCo remained below ~0.3. At higher fCos, the two
components formed a two-phase mixture of a more random nature.
The average size of the Co nanograins was ~1 nm for small fCos and
rose approximately linearly with the Co content to ~2.5 nm for fCo
≈ 0.6. The effective dielectric function was evaluated from meas­
urements of T(λ) and R(λ) on samples with 0.1 < fCo < 0.6. These
data could be reconciled with theories for optical homogenization
(cf. Section 3.9) at least for fCo < 0.3 and assuming that a minor
part of the Co was atomically dispersed in the Al 2O3 matrix.
Empirical data on the complex dielectric function were then
used to compute solar absorptance for normally incident solar
radiation as well as hemispherical thermal emittance at 100°C.
Figure B6.1.1 shows these quantities as a function of thickness for
uniform Co-Al 2O3 films backed by a substrate characterized by the
dielectric function for Al and Ni. Requiring a high value of Asol and
a low value of E therm leads to Co-Al 2O3 films with high fCos, and it is
found that the Al substrate yields a particularly low value of E therm.
There is an optimum film thickness for giving maximum Asol; its
value is around 70 nm, and the beneficial effect on the absorption
is due to antireflection of the underlying metal.

© 2011 by Taylor and Francis Group, LLC

 Heat and Electricity 271

100
(a)
80
Asol
60

fCo ≈ 0.11
40

20 Etherm

0
100
(b)
80
Asol
60 fCo ≈ 0.23
Solar Absorptance and Thermal Emittance (%)

40

20 Etherm

0
100
(c)
80
Asol
fCo ≈ 0.40
60

40
Etherm
20
Ni substrate
0 Al substrate
100
(d)
80 Asol

fCo ≈ 0.60
60
Etherm

40

20 Co-Al2O3
Substrate
0
0 200 400 600 800 1000
Thickness (nm)

Figure B6.1.1  Computed normal solar absorptance and hemi-

spherical thermal emittance at 100°C for Co-Al 2O3 as a function
of film thickness. The Co contents and substrate metals shown here
were studied. (From G. A. Niklasson, C. G. Granqvist, J. Appl.
Phys. 55 [1984] 3382–3410. With permission.)

© 2011 by Taylor and Francis Group, LLC

 272 Green Nanotechnology

The computational study in Figure B6.1.1 clearly leads to a

rather well defined design of a spectrally selective surface. This
design was then produced experimentally and is further investi-
gated in Figure B6.1.2, showing data on spectral reflectance for a
Ni surface coated with a 70-nm-thick Co-Al 2O3 film having fCo ≈
0.6. Spectral selectivity with λc ≈ 2 μm is manifest. The reflectance
shows a peak at around 25% at a wavelength of ~0.5 μm which
obviously is undesired and tends to deteriorate Asol significantly.
However, applying an antireflection coating consisting of 70 nm
of Al2O3 leads to a double minimum in the reflectance and a much
improved solar absorptance. A still lower value of Asol could have
been achieved by having a graded composition of the Co-Al 2O3
film, but it should be remembered that the exercise at hand aims
at showing principles rather than optimized products—which we
deal with shortly.

100

80

1
Reflectance (%)

60 1 + 2
1 + 2 + 3

40
70 nm Al2O3 3
70 nm Co-Al2O3 2
20 Ni substrate 1

fCo ≈ 0.60
0
0.5 1 2 5 10 20 50
Wavelength (µm)

Figure B6.1.2  Measured spectral reflectance of a Ni surface (dot-

ted curve), after deposition of ~70 nm of Co-Al 2O3 with fCo ≈ 0.6
(dashed curve), and after overcoating with 70 nm of Al 2O3 (solid
curve). (After G. A. Niklasson, C. G. Granqvist, J. Appl. Phys. 55
[1984] 3382–3410. With permission.)

We finally widen the perspective by considering angular depen-

dent properties, specifically Asol(θ) and E therm(θ) where θ is the angle

© 2011 by Taylor and Francis Group, LLC

 Heat and Electricity 273

with regard to the surface normal. These data are shown in Figure
B6.1.3, where we also demonstrate the internal consistency of the
evaluations by plotting closely agreeing data as computed from the
complex dielectric function and as derived from angular depen-
dent spectral reflectance for s- and p-polarized light. It is seen that
the optimized design leads to Asol ≈ 95% at normal incidence, and
that this value does not fall off drastically until high angles are
reached. The slow drop at increasing angles obviously is favorably
for stationary solar collectors. E therm, on the other hand, is about
5% at small angles and rises gradually for increasing angles and
has a peak value of ~15% at θ ≈ 85°. This latter feature makes the
hemispherical value of E therm increase to ~7%, i.e., a by fraction
that is noticeable though not very prominent. Similar results for
Asol and E therm have been obtained for Ni-based coatings such as
those discussed next [6,12].

100

Asol

80
Computed
Thermal Emittance (%)
Solar Absorptance (%)

60
Measured

40
70 nm Al2O3
70 nm Co-Al2O3, fCo ≈ 0.6
20 Ni

Etherm
0
0 30 60 90
Angle (degrees)

Figure B6.1.3  Angular-dependent solar absorptance and ther-

mal emittance at 100°C as evaluated from reflectance data (circles)
and as computed from dielectric functions (solid curves) for the
optimized two-layer coating sketched in the inset. (After G. A.
Niklasson, C. G. Granqvist, J. Appl. Phys. 55 [1984] 3382–3410.
With permission.)

© 2011 by Taylor and Francis Group, LLC

 274 Green Nanotechnology

1
Ni-Al2O3

0.8 Exp.
Ideal
Reflectance (%)

0.6

0.4

1
0.2 2
3
4
Al
0
1 10
Wavelength (µm)

Figure 6.8  Spectral reflectance of Ni-pigmented alumina made by

anodization followed by ac electrolysis. The data combine two sets of
measurements in different wavelength ranges. Dotted lines indicate the
ideal performance. Inset shows the nanostructure with a compact barrier
layer of Al2O3 (4), a Ni-Al2O3 composite layer (3), and a porous Al 2O3
layer (2) whose pores widen at the top surface (1). (After Å. Andersson et
al., J. Appl. Phys. 51 [1980] 754–764. With permission.)
make “black nickel” and “black chrome,” both consisting of a layer with
metallic nanoparticles in an oxide-based matrix of unspecific nature. A
later generation of Ni-based coatings used a two-step procedure with
initial anodization of aluminum in dilute phosphoric acid so that the
surface layer of the metal was transformed into ~0.7 μm thick porous
Al2O3 with channels perpendicular to the surface and extending through
the pore layer [13]. Nanostructures of this kind are described briefly in
Appendix 1. In a second step, metal was precipitated inside the pores by
ac electrolysis in a bath containing nickel sulfate. Metal particles were
then formed as nanorods with diameters of 30 to 50 nm and a length
of ~300 nm [14]. With suitable anodization parameters, the metal frac-
tion was ~30%. Essentially, the coating comprises Ni-pigmented Al 2O3
under a porous Al 2O3 antireflection layer, and the bottom of the coating
has a layer of compact Al 2O3 serving as diffusion barrier. The inset of
Figure 6.8 illustrates this structure, whereas the main part of the figure
shows spectral reflectance. The reflectance resembles the desired profile
indicated by the dotted lines. Commercial production of such coatings

© 2011 by Taylor and Francis Group, LLC

 Heat and Electricity 275

100

80 Exp.
Ideal
Reflectance (%)

60

40
1
2
20 3
4

0
0.2 0.5 1 2 5 10 20 50
Wavelength (µm)

Figure 6.9  Spectral reflectance of a Ni-based coating made by sput-

ter deposition and roll-to-roll manufacturing. Dotted lines indicate the
ideal performance. Inset shows the nanostructure with Al substrate (4),
barrier layer (3), graded Ni-NiO composite layer (2), and antireflecting
Al2O3 layer (1).

typically yielded Asol ≈ 96% and E therm ≈ 15%. The manufacturing

involves treatment of large amounts of hazardous chemicals and gener-
ally has been superseded by alternatives with “greener” attributes.
The durability and degradation mechanism of the Ni-pigmented
anodic Al 2O3 coatings have been investigated in considerable detail [5,6].
Prolonged treatment at high temperatures led to a gradual decrease of
the solar absorptance, which could be reconciled with uniform oxida-
tion (rather than a progressing oxidation front) of the Ni nanoparticles.
Detailed work has led to lifetime predictions, which have been found to
agree well with real ageing data [6,15,16].
Sputter deposition was developed into a manufacturing technol-
ogy for spectrally selective coatings during the 1990s [17]. The inset
of Figure  6.9 illustrates the composition of a fully developed coating
with a graded Ni-NiO film backed by Al with a highly reflecting barrier
layer and having an 85-nm-thick antireflecting Al 2O3 layer at the top.
The graded layer was made by the special sputter technology shown in
Figure A1.8. Deposition used roll-to-roll coating onto 30-cm-wide rib-
bons comprised of two Al sheaths with a flattened Cu tube in the center.
After sputter deposition, this ribbon was cut into suitable lengths, and
the Cu tube was opened by pressurized air to turn it into “riser” tubes
(cf. Figure 6.2). The spectral reflectance of the coated Al-based ribbon
is closer to the ideal performance than the electrochemically deposited

© 2011 by Taylor and Francis Group, LLC

 276 Green Nanotechnology

100

80 Exp.
Ideal
Reflectance (%) 60

40
1
2
20 3
4
5

0
0.2 0.5 1 2 5 10 20 50
Wavelength (µm)

Figure 6.10  Spectral reflectance of a Ni-based coating made by sol-

gel deposition. Dotted lines indicate the ideal performance. Inset shows
the nanostructure with Al substrate (5), barrier layer (4), Ni-Al 2O3 com-
posite layers with two compositions (3 and 2), and antireflecting SiO2
layer (1).

coating in Figure 6.8 and corresponds to Asol = 97% and hemispherical

E therm = 5% for 100°C [18].
During the 2000s, another Ni-based coating, based on sol-gel depo-
sition, has emerged [19]. Its cross section was illustrated as an example
of a sol-gel coating in Figure A1.2, which showed a 85-nm-thick bot-
tom layer of 80% Ni + 20% Al2O3 and an upper 60-nm-thick layer of
40% Ni + 60% Al 2O3 under a 65-nm-thick antireflecting SiO2 layer. The
Ni-Al2O3 layers were heat treated at 550°C and the SiO2 layer at 400°C.
The coating was deposited onto Al having a barrier layer. The optical
performance is reported in Figure 6.10. Just as for the previous film, the
spectral reflectance corresponds to Asol = 97% and hemispherical E therm
= 5% for 100°C.
Evaporation serves as an alternative to sputter deposition, and an
interesting type of selectively solar absorbing coatings has been devel-
oped by reactive evaporation to make TiOxNy (known as “TiNOX”)
coatings [20]. Asol ≈ 95%, E therm ≈ 5%, and stability up to 400°C, have
been reported.
The coatings discussed thus far are examples of selective absorbers
suitable for buildings-related applications. As mentioned above, there are
also many large-scale applications for solar-driven power generation,
and then the absorber should work at high temperatures so that λc ≈ 2
μm may be appropriate. Figure 6.11 shows an example of such a coating
deposited onto stainless steel and designed to work on tubular collectors

© 2011 by Taylor and Francis Group, LLC

 Heat and Electricity 277

100
Mo-Al2O3

80 Exp.
Ideal
Reflectance (%)

60

40 1
2
3
4
20 5
SS

0
1 10
Wavelength (µm)

Figure 6.11  Spectral reflectance of a Mo-Al 2O3 -based coating made

by sputter deposition. The data combine two sets of measurements in
different wavelength ranges. Dotted lines indicate the ideal performance.
Inset shows a five-layer structure with a stainless steel (SS) substrate,
Al2O3 (5), Mo (4), graded Mo-Al 2O3 (3), graded Mo-SiO2 (2), and a top
layer of SiO2 (1). (After M. Lanxner, Z. Elgat, Proc. Soc. Photo-Opt.
Instrum. Engr. 1272 [1990] 240–249. With permission.)

at temperatures exceeding 300°C [21]. The nanostructure embodies two

Mo-dielectric composite layers backed by an IR reflecting, and hence low-
emittance, Mo layer. There is also an antireflecting SiO2 layer at the top
and a diffusion barrier of Al2O3 at the bottom. The coating shown has
Asol = 97 ± 1%, E therm ≈ 10% at 50°C and E therm = 17 ± 1% at 350°C.

6.1.4  Colored Absorbers and Paints: Novel Developments

It was demonstrated above that spectrally selective solar absorbers with

close to ideal properties can be produced successfully not only in the
laboratory but as commercial products. So what remains to be done?
Current development progresses in two different directions: one of
them—at first sight perhaps paradoxically—addresses the fact that the
earlier discussed spectrally selective surfaces are too good! The latter
statement can be interpreted as saying that the surfaces, being almost
ideal solar absorbers, also look very black to the eye which often is not

© 2011 by Taylor and Francis Group, LLC

 278 Green Nanotechnology

100 100
Asol = 0.86 Asol = 0.81
80 Etherm = 0.04 80 Etherm = 0.06
Reflectance (%)

Reflectance (%)
60 60
575 nm
40 683 nm 40

20 20

0 0
0.2 0.5 1 2 5 10 20 50 0.2 0.5 1 2 5 10 20 50
Wavelength (µm) Wavelength (µm)
100 100
Asol = 0.92 Asol = 0.86
80 Etherm = 0.08 80 Etherm = 0.08
Reflectance (%)

Reflectance (%)
60 60
40 515 nm 40 463 nm
20 20

0 0
0.2 0.5 1 2 5 10 20 50 0.2 0.5 1 2 5 10 20 50
Wavelength (µm) Wavelength (µm)

Figure 6.12  A color version of this figure follows page 200. Spectral
reflectance for TiAlOxNy films made by sputter deposition onto Al.
Different colors (shown) were obtained by selecting film thicknesses
leading to reflectance maxima at the wavelengths shown in the luminous
wavelength range. Corresponding values of Asol and E therm are stated.
(From D. Zhu, S. Zhao, Solar Energy Mater. Solar Cells [2010] in press.
With permission.)

desired for aesthetic reasons. Hence, there is a need to develop colored

selective absorbers for which the coloration is achieved without an exces-
sive loss of solar absorptance. The other development line is related to
cost issues and deals with spectrally selective paints.
We first consider colored coatings produced by sputtering and show
data for TiAlOxNy films backed by Al in Figure 6.12 [22]. Single homo-
geneous films lead to two interference minima when their thickness is
~200 nm. The precise location of the intervening reflectance maximum
is governed by the thickness, which can be tuned so that the perceived
color is red, yellow, or dark blue. A green color is somewhat more difficult
to obtain and requires a double layer coating with TiAlOxNy films hav-
ing different compositions and hence refractive indices. Corresponding
values of Asol lie between 81 and 92%, whereas the emittance values,
based on near-normal reflectance measurements, lie between 4 and 8%.
These results are not as good as those for optimized spectrally selective

© 2011 by Taylor and Francis Group, LLC

 Heat and Electricity 279

Figure 6.13  Cross section (left) and top view (right) of a thickness-
insensitive spectrally selective paint with Al flakes. (From B. Orel, pri-
vate communication [2009].)

solar absorbers, but the colors allow applications in situations for which
black surfaces may not be acceptable.
Selectively solar absorbing paints represent a low-cost alternative
to the coatings applied by other techniques. The paint formulations
typically contain a strongly absorbing pigment such as FeMnCuOx or
CuCr2O4 as well as highly reflecting metal flakes, typically of Al in a
polyurethane binder [23]. Figure 6.13 depicts a cross section and a top
view of such a coating. The flakes align more or less in parallel and serve
as “artificial low-emittance substrates” so that the optical properties are
thickness insensitive and independent of the substrate. In the absence
of metal flakes, the coating would have a thermal emittance typical for
the organic binder (i.e., E therm ≈ 90%). A proper paint formulation, how-
ever, leads to coatings with 80 < Asol < 90% and 30 < E therm < 40%.
Figure 6.14 shows typical reflectance spectra for three paint layers, and
demonstrates that spectral selectivity is at hand, though far from the
optimum. The samples are visibly colored and appear green, red, and
blue. Very recently, paint layers with self-cleaning properties have been
developed [24].

6.2  Photovoltaic Materials and Devices

Solar cells, also called photovoltaic (PV) cells, convert solar energy into
electricity and are widely touted today (2009) as a technology for large-
scale electricity production. The efficiency of the solar cells is less than
that for converting solar energy into heat, but electrical energy, of course,
can be used for many more purposes than thermal energy. What is not
converted to electricity in the solar cell turns into heat, and solar cells
therefore become heated during operation. This heating tends to erode

© 2011 by Taylor and Francis Group, LLC

 280 Green Nanotechnology

100

red
80

Reflectance (%)
60
black

40
green
20

0
1 10
Wavelength (µm)

Figure 6.14  Spectral reflectance for visibly green, red, and blue layers
of thickness-insensitive spectrally selective paint. Corresponding values
of Asol /E therm are 84/36, 83/40, and 81/32%, respectively. (From B. Orel,
private communication [2009].)

the photoelectric efficiency as well as cause degradation and diminish

the operating life time of the cells. Systems combining electricity produc-
tion with heat production—using a heat transfer fluid for cooling the
solar cells—have been discussed and implemented many times, and may
be technically elegant, but such systems have not gained much popular-
ity owing to their technical complexity.
The heating that would ensue from widespread deployment of build-
ing integrated photovoltaics, known as BIPV, has not received much
attention so far but clearly is an effect that contributes to urban heat
islands (cf. Section 7.2) and hence stands in opposition to the global
cooling one might accomplish via increasing urban albedos (surface
“whiteness”) [26]. On the other hand, if BIPV power displaces coal
power it provides net environmental benefits quickly despite its addi-
tional heat output [27]. Net benefits require the accumulated CO2 reduc-
tions, which arise from PV cells replacing coal power, to surpass the
impact of associated heat emissions. But also necessary to consider are
alternative mitigation strategies, such as global cooling by raising roof
albedos. Consider, for example, a 1 kWp solar cell installation (where the
subscript indicates peak output). It might occupy about 10% of a home’s
roof area; what if instead the whole roof were specially painted to raise
its solar albedo by 30 to 40%? It would take 50 to 80 years for the accu-
mulated CO2 reductions from these solar cells to match the initial albedo
offsets from the complete white roof [27]! Doubling the cell area would
cut these times into half, but the cost differential—which is very large—
would also double. In addition, the production of silicon for solar cells

© 2011 by Taylor and Francis Group, LLC

 Heat and Electricity 281

generates a lot of heat and CO2 emissions, and this issue is so important
that it is specifically addressed for some emerging routes for solar cell
production. One possibility is to use concentrated solar radiation in the
melting and refinement process.
Below we first give a brief overview of solar cell types and then dis-
cuss how nano-features can boost the efficiency of silicon-based solar
cells and how nanotechnology leads to new types of solar cells, such as
dye-sensitized and organic cells.

6.2.1  Technology Overview

In principle, all that is required for an electricity-generating material is

a capacity for having excited states into which electrons can be brought
from a ground state via photon absorption, and means to extract elec-
trons from these states. There is a plethora of options but—despite intense
research and development over many years—attention has been focused
only on a few of these. A common feature of solar cells is that the side
exposed to the sun must have some current collecting arrangement, either
a metallic pattern, a transparent conducting film (cf. Section 4.4) or a
combination of these. We do not go into details about solar cell technol-
ogy here but refer the reader to excellent texts on the subject [28–33].
Figure 6.15 gives a bird’s eye view on the development of a number of
solar cell technologies that has followed the first solar cell created in 1954
[34]. The highest efficiencies are for advanced, and expensive, designs
with superimposed two-junction and three-junction cells (see below),
which can have a conversion efficiency η exceeding 40%. Other tech-
nologies, including today’s dominating Si cells and various thin-film tech-
nologies, have efficiencies with maximum values in the 20 to 25% range.
It is the latter that are of interest for BIPV in the foreseeable future.
Silicon is an excellent material for solar cells and it is by far the most
used one today [32]. There are different types of Si-based solar cells,
as apparent from Figure 6.15, with single crystalline ones reaching η ≈
25%. Silicon is nontoxic, very abundant, and has a thorough technologi-
cal base as a result of its ubiquitous uses in microelectronics. In essence
a slab or film of Si is n-doped and p-doped in its two surface regions,
and this structure is put between a transparent front electrode and an
opaque metallic back electrode. The device is supported by glass or by a
foil of a metal or polymer. The Si can be of various types:

• Crystalline and cut out from a single crystal ingot

• Polycrystalline and prepared from a multi-crystalline ingot or
via “ribbon growth”

© 2011 by Taylor and Francis Group, LLC

 44

282
Best Research-Cell Efficiencies Boeing-
Multijunction Concentrators Spectrolab
Three-junction (2-terminal, monolithic) (metamorphic)
40
Two-junction (2-terminal, monolithic)
Crystalline Si Cells Boeing-
NREL
Spectrolab
36 Single crystal (inverted, semi-mismatched)
Spectrolab
Multicrystalline
Thick Si film Japan
32 Energy
NREL/
Thin Film Technologies
NREL Spectrolab
Cu (In, Ga)Se2 NREL
28 CdTe
Amorphous Si:H (stabilized)
Nano-, micro-, poly-Si UNSW UNSW
Efficiency (%)

Spire
UNSW
24 Multijunction polycrystalline

Green Nanotechnology
UNSW NREL
Emerging PV Stanford UNSW Cu(In, Ga)Se2
Spire 14× concentration FhG-ISE
Dye cells Georgia Tech UNSW
20 Organic cells ARCO Sharp
Georgia Tech NREL NREL NREL NREL
(various Westing- Varian NREL
NREL Univ. Stuttgart Sharp
technologies) house University (large area)
16 No. Carolina NREL (45 µm thin-film
So. Florida AstroPower NREL transfer)
State Univ. NREL (small area)
ARCO Boeing Euro-CIS
NREL
Kodak Solarex Boeing (CdTe/CIS)
12 Boeing EPFL United
AMETEK Photon Energy Solar
Matsushita Sharp
United Solar Kaneka
Monosolar Kodak Boeing Solarex (2 µm on glass)
8
RCA NREL Konarka
Boeing Univ.Linz UCSB
University EPFL Groningen
4 of Maine RCA
RCA Siemens
RCA RCA
RCA RCA University Linz University
Linz
0
1975 1980 1985 1990 1995 2000 2005 2010

Figure 6.15  Best efficiency for research-type solar cells of a variety of kinds versus year when the results were presented. The
symbols have designations mentioning the laboratories responsible for the data; the symbols are joined by straight lines to guide the
eye. Acronyms used are NREL (National Renewable Energy Laboratory, United States), UNSW (University of New South Wales,
Australia), FhG-ISE (Fraunhofer Institut für Solare Energiesysteme, Germany), EPFL (École Polytechnique Fédérale de Lausanne,
Switzerland), and UCSB (University of California–Santa Barbara). (After J. Berry, private communication [2009].)

© 2011 by Taylor and Francis Group, LLC

 Heat and Electricity 283

• Amorphous and hydrogenated, typically being a film made by

glow discharge deposition from silane gas

High values of η demand low reflection losses, which can be accom-

plished by an antireflecting film or—for single crystalline cells—by using
anisotropic etching to make pyramid-type texture. Single crystalline and
polycrystalline cells normally have metallic “fingers” as current collec-
tors, which partially shade the cell. Higher efficiency (but more expen-
sive) cells use “buried” vertical electrodes formed inside laser-etched
grooves. These shade less and collect more current. SnO2-based transpar-
ent conductors are often used as electrodes in thin-film devices. A thin
TiO2 layer can serve as protection against harmful exposure to hydrogen
radicals in some techniques for making amorphous Si solar cells.
Compound semiconductors of the III-V type—based on GaAs,
(Al,Ga)As, InP, or (In,Ga)P—can display high conversion efficiencies
and good resistance to ionizing radiation. However, they are costly and
mainly of interest for space applications; no information for this type of
cells in shown in Figure 6.15.
II-VI compounds represent another option, and solar cells based on
CdTe and Cd(S,Te) are noteworthy for their robust and cheap manu-
facturability and are of much interest for terrestrial applications [35].
Potential environmental problems regarding large-scale deployment of a
technology involving Cd-containing materials must not be overlooked,
though. Figure 6.15 shows that η ≈ 17% has been achieved.
Ternary I-III-IV2 compounds attract much interest for PV applications
[35,36]. In particular, CuInSe2, Cu(In,Ga)Se2, and Cu(In,Ga)S2 are notable
for their ability to yield high efficiency in thin-film solar cells with values
up to ~20% shown in Figure  6.15. The sulfur-containing films seem to
allow cheaper manufacturing techniques than the selenide-containing ones.
However, the fabrication of ternary oxides with strict demands on compo-
sition is always demanding with regard to large-scale manufacturing.
It is customary to speak of first, second, and third generation solar
cells, as illustrated in Figure  6.16 [31]. Here “generation I” refers to
the well-established crystalline Si wafer technology and “generation II”
refers to today’s thin-film technologies. “Generation III” draws on a
number of advanced concepts that in the future may allow solar cells
to get closer to the thermodynamic limit of ~93% while the cost is kept
low. “Generation III” solar cell technologies include multi-junction
devices, arrangements leading to the creation of multiple electron-hole
pairs, cells utilizing hot carrier effects, thermo-photovoltaics, as well
as different thermoionic and thermoelectric approaches; these concepts
are generally considered to be outside the scope of the present book,
and the reader is referred to the literature [31] for in-depth information.

© 2011 by Taylor and Francis Group, LLC

 284 Green Nanotechnology

US$0.10/W US$0.20/W US$0.50/W

100

Thermodynamic
80 limit
Efficiency, %

60
US$1.00/W

40 III Present limit

20
I US$3.50/W
II

0 100 200 300 400 500

Cost, US$/m2

Figure 6.16  Efficiency versus cost (in 2003 US$) for solar cell technolo-
gies belonging to generations I–III. (From M. A. Green, Third Generation
Photovoltaics: Advanced Solar Energy Conversion, Springer, Berlin,
Germany, 2003. With permission.)

Some emerging PV technologies, represented in the lower right-hand

corner of Figure 6.15, may also be considered “generation III”; they are
considered below.

6.2.2 Nanofeatures for Boosting the Efficiency

of Silicon-Based Solar Cells

Thinner wafers would lead to a reduction in the production cost for Si

solar cells and a lowering of their energy payback time. An even more
radical approach to saving on the amount of Si is to use devices or wires
in which the p-n junction is radial rather than planar [37]. Thickness
reduction is a nontrivial issue, though. Surface texturing at the micro
scale allows thinner c-Si, but if the thickness is too small there will nev-
ertheless be an excessive number of photons that pass through the cell
without being absorbed, so η becomes undesirably small.
Nanostructures can improve the situation and allow thinner c-Si
cells. One approach uses micrometer-thick c-Si under glass arranged so
that the light coupling into the Si is enhanced either by having the glass
surface nanostructured or by having the transparent conducting oxide
electrodes between the smooth glass and the Si film deposited so as to be
nanostructured. A problem with c-Si layers having thicknesses of a few

© 2011 by Taylor and Francis Group, LLC

 Heat and Electricity 285

micrometers is that they require a very long high-temperature anneal—

as much as about 24 h—in order to achieve the desired electrical quality,
so energy, emissions, and urban heat island impacts may be as high as for
Si wafers, even if total cost is down. Lower production energy would be
needed for this type of cell if the c-Si layer and the nanostructured TCO
layer were deposited pyrolytically on glass sheet in conjunction with
its production in a float line. This idea is under serious consideration
by some glass manufacturers. Another attraction of such a production
route is that it opens the way toward commodity-scale manufacturing
with associated low costs for solar cells. As pointed out before in Section
4.1.2, float glass is made on the scale of billions of square meters per
annum, and new plants are easily built. The electrical quality of the c-Si
made in this way will be the main challenge.
Nanostructured TCOs for solar cells are already being deposited
pyrolytically onto glass sheet during its production on a float line by
use of a modification of the processes for making low-emittance win-
dows, as described in Figure A1.1. The physics behind enhanced light
coupling to Si due to nano-TCOs depends on the scale of the nano-
structure. A combination of scattering and surface plasmon resonance
coupling can be involved. Plasmon resonances can lead to coupling into
the Si layer both by scattering and more directly if the Si layer’s light
transport modes and the nanoparticle’s SPR modes form a coupled
resonant system. These types of effects were discussed in Section 3.11.
The effects are most obvious when metal nanoparticles are deposited
onto a very thin Si-oxide layer on the Si. This idea was first discussed
for Cu, Ag, and Au particles [38], and subsequently for Ag nanopar-
ticles on the LED structure shown in Figure  6.17 [39]; such a LED
structure can be easily adapted for use as a solar cell. Enhanced pho-
tocurrents arise at wavelengths that depend on the particle shape and
size, as well as on the oxide and waveguide thicknesses in structures
such as that in Figure 6.18. Overcoating the Ag particles with a thin
insulator, as shown for ZnS, red-shifts and enhances the coupling and
protects the silver. Details of the relevant SPR mechanism are elabo-
rated in Box 6.2.
Surface texturing of the Si itself with pyramids is used to enhance
light coupling, but such texturing can degrade the electrical properties.
In the plasmonic approach, however, the nanoparticles are not in contact
with the Si, so electrical degradation does not occur.
The structures discussed above involve Si layers with thicknesses of
at least a few micrometers. It may also be possible to enhance the cou-
pling into much thinner Si layers by use of SPR nanoparticles, but the
coupling mechanism is different from that with thicker layers since nano-
thin Si cannot act as a waveguide for light. In this latter case, the extra
coupling is via the increased electric field strength close to the resonant

© 2011 by Taylor and Francis Group, LLC

 286 Green Nanotechnology

– +
Dielectric
overcoating Ag islands Metal

SiO2

n+ p-Si p+

SiO2

Si

Figure 6.17  Schematic of a simple LED or thin crystalline Si solar cell

structure. (From S. Pillai et al., Appl. Phys. Lett. 88 [2006] 161102 1–3.
With permission.)

10

8
Enhancement

With Ag
6
With Ag+ZnS
4 Reference variation

0
800 900 1000 1100 1200 1300
Wavelength (nm)

Figure 6.18  Enhancement of light coupling in the case of bare and

ZnS coated Ag nanoparticles for the structure in Figure 6.17. The actual
data refer to electroluminescence. The dotted curve shows data taken in
the absence of nanoparticles. (From S. Pillai et al., Appl. Phys. Lett. 88
[2006] 161102 1–3. With permission.)

© 2011 by Taylor and Francis Group, LLC

 Heat and Electricity 287

Box 6.2  How Can Resonant Metal

Nanoparticles, Which Normally
Convert Light into Heat, Enhance the
Conversion of Light to Electrical
Current When Placed on a Layer of Si?
The trick is to have the surface plasmon resonant energy in the
metal nanoparticles convert into waves of traveling light inside the
Si layer before it can dissipate as heat within the nanoparticle. It is
also necessary to produce electric current before any reverse cou-
pling of this light back into other nanoparticles occurs inside the
Si. There are two distinct situations.
In the first of these the Si layer is thick enough to act as a wave-
guide, which means that certain wavelengths or modes can travel
sideways through the layer long enough to be absorbed by the Si
as in Figure B6.2.1. The illuminated resonant metal nanoparticles
behave like little dipoles or antennas and radiate in all directions,
in effect scattering a lot of light into the Si at an angle so that light
travels much further on average inside the Si than if the nanopar-
ticles had not been there. The effective dipole sits a few nanometers
above the oxide layer. Models for this effect have been established
[41] and show that a point dipole just above an Si layer can couple
very strongly (with nearly 100% efficiency). Finite-size nanopar-
ticles are not as efficient but still couple well. Both shape and size
of the nanoparticles influence the coupling efficiency [42].
A nanocylinder and a nanohemisphere couple more strongly,
at about 90% efficiency, than a nanosphere if they all have the
same radius. The latter’s optical coupling efficiency varies with
particle size, and 150-nm-diameter and 100-nm-diameter spheres
are about 50 and 70% efficient, respectively. The Si layer is rather
thick, around 1 μm, but this is much less than a typical Si-wafer
thickness. It is interesting that the resonant properties are rather
independent of the choice of metal. While Cu, Ag, and Au nano-
particles on glass have their resonances at widely different wave-
lengths, the enhancement effect is peaked at almost the same
wavelength in the NIR when the nanoparticle is located immedi-
ately above an Si waveguide layer [38]. The waveguide coupling is
responsible for this effect as it couples to the tails of the SPRs.
The second situation regards a less-than-60-nm-thick semicon-
ductor layer under surface plasmon resonant nanoparticles. This
was the case for the data presented in Figure 6.19. Very strong eva-
nescent or localized electric fields occur just outside the resonant

© 2011 by Taylor and Francis Group, LLC

 288 Green Nanotechnology

particles. These fields extend only a short distance, but if the particle
is close enough to a semiconductor layer whose thickness matches
the extent of the local field, then enhanced absorption becomes
possible in the semiconductor at wavelengths where transmittance
would normally take place. As noted above, charge transport is a
concern in such layers.

Incident light

Metal islands Spacer layer

Absorbing
dielectric waveguide

Figure B6.2.1  Metal nanoparticles on a thin spacer layer on top

of a layer which acts as a waveguide. (From H. R. Stuart, D. G.
Hall, Appl. Phys. Lett. 69 [1996] 2327–2329. With permission.)

nanoparticles. It has been estimated that Si solar cells as thin as 15 to 50

nm may be made up to 40% optically efficient by using an overlayer of
plasmonic nanoparticles, as evident from the data in Figure 6.19 [40].
The challenge lies in achieving good electrical properties in such layers.

6.2.3  Dye-Sensitized Solar Cells

Dye-sensitized solar cells (DSSCs) are based on electrochemical conver-

sion of solar energy. The pivotal work was by Grätzel in 1991 [43], and
the cells are often referred to as “Grätzel cells,” although, as always,
there was a large body of earlier work to build on. The highest value
of η so far is 11.1%, which was measured under standard AM1.5 solar
irradiation onto an aperture that was ~0.2 cm 2 in size [44]. Apertures
of these magnitudes are typical for many efficiency assessments, and
much lower efficiencies are typically observed for module sizes of more
practical interest. Figure 6.15 shows that η ≈ 10% was reached rapidly,
but the progress manifested since the mid-1990s seems to have been
rather slow despite huge research efforts worldwide [45]. Low manu-
facturing cost is often put forward as a specific asset for these cells, and
was done so already in the first publication on DSSCs; however, since no

© 2011 by Taylor and Francis Group, LLC

 Heat and Electricity 289

0.4
60 nm
40 nm

Plasmon Induced Absorption

0.3 20 nm
10 nm
5.0 nm
0.2
2.5 nm

0.1

0
0.001 0.01 0.1 1
PV Layer Extinction Coefficient

Figure 6.19  A color version of this figure follows page 200. Absorption
induced in various thicknesses of very thin Si by neighboring Ag nano-
particles (peaked curves to right). The x-axis is the absorption coefficient
of the Si material. The particles themselves absorb according to the plots
with crosses. (From C. Hägglund, B. Kasemo, Opt. Express 17 [2009]
11944–11957. With permission.)
reliable technology for mass fabrication seems to be at hand (in 2010),
such statements must be regarded as tentative, at best.
Figure 6.20 is a schematic representation of the functional principles
of a DSSC [46]. A nanoporous and nanocrystalline oxide layer is at the
heart of the device. It is normally made of connected TiO2 anatase nano-
particles, but other oxides, such as ZnO and SnO2 , may be used, too.
The surfaces of the nanoparticles have an attached monolayer of a charge
transfer dye (“sensitizer”). When the latter is exposed to solar irradia-
tion, it is photo-excited to a state denoted S* in Figure  6.20, and an
electron is injected into the conduction band of the oxide. The original
state of the dye is then restored by electron donation from an electrolyte,
which normally is a liquid organic solvent containing the iodide ion/
triiodide ion (I – /I3 –) system. The regeneration of the sensitizer by iodide
intercepts the recapture of the conduction band electron by the oxidized
dye. The iodide, in its turn, is regenerated at the counter electrode (cath-
ode), which is usually Pt-coated glass, and the circuit is completed via
the external load. The voltage generated under illumination corresponds
to the difference between the Fermi level of the oxide and the redox
potential of the electrolyte, and electric power is generated without any
permanent chemical transformation.
The use of a liquid electrolyte is problematic from a device perspec-
tive, but it is possible to replace it—albeit then having a decrease in

© 2011 by Taylor and Francis Group, LLC

 290 Green Nanotechnology

Conducting
glass
TiO2 Dye Electrolyte Cathode
E (V) n
tio
jec S*
In
–0.5

Da
rk Maximum
0 cu
rre voltage
nt

hν
0.5 Red Ox

n
Electrolyte
tio
ep
Diffusion
erc
Int

1.0
S°/S+

Electron flow
Load

Figure 6.20  Functional principles for a DSSC, showing the excitation

of the dye by photons with the energy hν, the redox reaction in the elec-
trolyte, and the ensuing development of a current with a certain maxi-
mum voltage. An energy scale is shown to the left. (After A. Hagfeldt,
M. Grätzel, Acc. Chem. Res. 33 [2000] 269–277. With permission.)

efficiency—with a wide band gap inorganic p-type semiconductor or a

hole-transmitting solid. The nature of the TiO2 layer is critical for the
proper functioning of the DSSC—in particular, for its sustaining good
electrical transport through the nanoporous structure; details of this
transport still remain poorly understood. Figure 6.21 illustrates typical
nanostructures of TiO2 films used in DSSCs. Part (a) refers to the surface
of a layer made by a compression technique [47], and part (b) shows a
cross section through an as-deposited TiO2 films made by sputter depo-
sition at oblique incidence toward a rotating unheated substrate. The
nanostructure formed by the latter technique has been termed parallel
penniform [48]. Comparable efficiencies were observed on cells based on
these two types of TiO2 films. Nanowires may be used as an alternative
to the porous oxide films, but the efficiencies obtained so far have been
low. Plastic substrates may allow for cheap manufacturing, and DSSCs
of this type have reached η = 7.4% [49].
The TiO2 in the DSSC absorbs predominantly in the UV and hence can
convert only a tiny fraction of the solar energy into electricity. Virtually

© 2011 by Taylor and Francis Group, LLC

 Heat and Electricity 291

0.7 µm

(a) (b)

Figure 6.21  Scanning electron micrograph of a TiO2 layer made from

P25 nanoparticles using a standard procedure (a), and cross-sectional
transmission electron micrograph for a 4.2-μm-thick sputter-deposited
TiO2 film. (Panel a from H. Lindström et al., Solar Energy Mater. Solar
Cells 73 [2002] 91–101. With permission; panel b from M. Gómez et al.,
Solar Energy Mater. Solar Cells 62 [2000] 259–263. With permission.)

all of the solar absorption is in the dye, and very large research efforts
have gone into the development of dyes covering the solar spectrum, at
least for λ < 0.9 μm. Furthermore, the dye should be firmly grafted to the
TiO2 surface and inject electrons into the conduction band with a quan-
tum efficiency near unity and simultaneously be stable against degrada-
tion for long enough times. Dye development requires advanced organic
chemistry, and only some indications can be made here.
Polypyridyl complexes of ruthenium are particularly successful sen-
sitizers, with cis-RuL 2-(NCS)2 —where L stands for 2,2′-bipyridyl-4,4′-
dicarboxylic acid—is the premier sensitizer, known as the N3 dye. It
has been challenged by “black dye” more recently; this is tri(cyanato)-
2,2′,2′′-terpyridyl-4,4′,4′′-tri(carboxylate)Ru(II). Figure  6.22 shows the
spectral response of the photocurrent for DSSCs sensitized by N3 and
“black dye,” and it is apparent that the latter gives superior coverage of
the NIR and hence higher efficiency.
Plasmonic effects can increase the absorption in the dye and in prin-
ciple lead to larger efficiencies [50]. Nanoparticles of Ag or Au have
the desired plasmonic properties, but a problem with regard to DSSC
applications is that the I−/I3− system is extremely corrosive, even for Au.
A conceivable solution is to cover the nanoparticles with pinhole-free
TiO2 [51].
Another way to increase the efficiency is to increase the amount of
light trapping in the TiO2 , which can be made by adding scattering sub-
micron particles [44]. Clearly, this is the same strategy as the one used to
improve conventional thin-film solar cells via “milky” ZnO:Al transpar-
ent conductors (cf. Section 4.4.2). Figure 6.23 shows that when the haze

© 2011 by Taylor and Francis Group, LLC

 292 Green Nanotechnology

80

Black dye
60
N3
IPCE [%]

40

20

0
400 500 600 700 800 900 1000
Wavelength [nm]

Figure 6.22  Spectral incident photon to current conversion efficiency

(IPCE) for DSSCs sensitized by two different dyes. (From A. Hagfeldt,
M. Grätzel, Acc. Chem. Res. 33 [2000] 269–277. With permission.)

(ratio of diffuse to total optical transmittance) was enhanced by addition

of particles with a diameter of 0.4 μm, the photo-response is widened in
a manner analogous to the one shown in Figure 6.22. It is noteworthy
that even as little as 10% haze gives a substantial improvement.
The majority of the work on DSSCs is related to TiO2 , and this mate-
rial turns up in many green nanotechnologies. Widening the absorption
to include the luminous part of the spectrum is of great interest, and this
issue is discussed in Box 6.3.

6.2.4  Organic Solar Cells

Organic solar cells (OSCs) attract much interest today as power sources,
primarily for non demanding applications. Their development has largely
paralleled that of OLEDs (cf. Section 5.6), which use similar chemicals
and structures. But it remains to be seen whether OSC development will
follow that of OLEDs, which are already commercial and growing in
importance. Possibilities to achieve very low-cost solar cells through
solution processing, as well as flexibility and light weight, are strong
technology drivers.
OSCs can be designed according to several different principles and
involve many types of materials, but so-called bulk heterojunctions com-
prised of a conjugated polymer blended with a fullerene derivative have
been widely studied recently, especially with regard to the system poly(3-

© 2011 by Taylor and Francis Group, LLC

 Heat and Electricity 293

100

80

IPCE (%) 60
Haze 76%
40 Haze 60%
Haze 53%
Haze 36%
20 Haze 10%
Haze 3%

0
400 600 800 1000
Wavelength (nm)

Figure 6.23  Spectral incident photon to current conversion efficiency

(IPCE) for DSSCs containing light scattering particles. (From Y. Chiba
et al., Jpn. J. Appl. Phys 45 [2006] L638–L640. With permission.)

hexylthiophene) (P3HT) as conjugated polymer and [6,6]-phenyl-C61-

butyric acid methyl ester (PCBM) as fullerene derivative [54–56]. The
P3HT has a band gap of ~1.9 eV and enables efficient light absorption
for λ < 0.65 μm at film thicknesses as small as ~200 nm. OSCs of this
type have yielded η ≈ 5% [57], and even larger efficiencies, up to 6.5%,
have been reported recently for tandem cells incorporating alternative
conjugated polymers and fullerene derivatives separated by a Ti oxide
layer [58]. Figure 6.24 illustrates a typical device design with an active
composite between a TCO-coated substrate and an Al back electrode.
The mechanism for photoelectric conversion in the OSCs is compli-
cated but can be described in terms of four basic steps:

• Absorption of light in the conjugated polymer to generate exci-

tons (i.e., strongly bound electron–hole pairs, which are preva-
lent as a consequence of the low dielectric constant)
• Diffusion of these excitons
• Dissociation of the excitons with charge generation
• Charge transport in the polymer and in the fullerene derivative

These steps are illustrated in Figure 6.24. The exciton dissociation takes

place almost exclusively at the interfaces between the two materials,
with P3HT serving as electron donor and PCBM as electron acceptor.
This model is oversimplified, though, and does not explain why it is
energetically favorable for the exciton to dissociate and charge separate.
A detailed discussion of several different theoretical issues is given in the
literature [56].

© 2011 by Taylor and Francis Group, LLC

 294 Green Nanotechnology

Box 6.3 Modified TiO2 Which Absorbs at

Visible Wavelengths: A “Holy Grail” in
Solar-Energy-Related Materials Science
The “holy grail” in the tales of King Arthur was sought for its divine
and miraculous powers, but it also came to symbolize much wasted
effort and tragedy, with little or no reward. The words are some-
times used to describe an ultimate and problematic goal. Titanium
dioxide is an important core material related to solar energy and
energy efficiency and crops up as such in a range of applications
throughout this book. These many-fold uses stem from its being a
wide band gap semiconductor with very high refractive index and
UV photocatalytic capability. The latter includes the ability to gen-
erate hydrogen from water directly under solar irradiation, though
not efficiently enough to be commercially viable. TiO2 is transpar-
ent at visible wavelengths but absorbs solar light in the UV. If it
could be modified so that its electrical and photocatalytic prop-
erties were retained, but it absorbed at visible solar wavelengths,
then many exciting new opportunities would arise. The use of dyes
on the surface of TiO2 nanoparticles in the DSSC solar cells is the
only real useful advance to date in this field, but this approach has
severe limitations. Doing away with the dye would make life much
easier, but is it possible?
This prospect of visibly absorbing TiO2 has entranced and
enmeshed in unforeseen difficulties many scientists for nearly four
decades, with a long trail of claims of relevant breakthroughs that
either have never come to useful fruition or have proven to be
experimental artifacts. One wonders how many research dollars
(many millions probably) have been raised in the quest for this
technological grail. Many dopants have been tried, including Te,
Cr, N, C, Nb, and Eu, along with many doping techniques from
chemical to ion implantation. Another approach is to plasma treat
with hydrogen to create oxygen defects. Recent work has shown
[52,53] that nitrogen defects can introduce tails into the band gap
that extend into visible wavelengths. However, such tails can be
reminiscent of amorphous semiconductors, and associated elec-
trical characteristics are likely to be degraded. Nanostructures
may be worth some effort. For example core-shell systems with
interfaces between TiO2 and another nanothin visibly absorbing
material might deserve a look, and slight lattice distortions at such
interfaces can be of interest. Oxygen defects on some crystalline
planes of nanorods might be visibly active, as is found in ZnO

© 2011 by Taylor and Francis Group, LLC

 Heat and Electricity 295

nano-rods. Empirical work has not so far provided a solution,

so advanced theoretical band structure and interface studies are
urgently needed to see if any real progress is possible.
Visibly absorbing TiO2 has clearly proven to be a holy grail of
solar-energy-related materials science, with much to offer but little
to show for decades of effort. Will a solution eventually emerge?
Or will we be forever tantalized by an illusory prospect? Will good
theoretical science show the way forward, or draw an end to this
saga? Only the future, and possibly some inspired nanotechnology,
will tell!
We return to visibly absorbing TiO2 in the context of photoca-
talysis and air cleaning in Section 8.2.

TCO coated PEDOT Aluminum

substrate PCBM
:PSS Polymer electrode

Charge
separation
Charge
transport
Light

Exciton
generation
and diffusion

External current flow

O
O
S S
S

P3HT PCBM

Figure 6.24  Schematic rendition of a polymer-fullerene solar cell (upper

part) and of its molecular components (lower part). (From C. J. Brabec,
J. R. Durrant, MRS Bull. 33 [2008] 670–675. With permission.)

© 2011 by Taylor and Francis Group, LLC

 296 Green Nanotechnology

100 nm

Figure 6.25  Transmission electron micrograph depicting the nano-

structure of a 1:1 blend of P3HT and PCBM after annealing at 150°C
for ½ h. (From K. Sivula et al., J. Am. Chem. Soc. 128 [2006] 13988–
13989. With permission.)

The nanostructure of the active layer in the OSC is of the greatest

importance for achieving a high efficiency. Ideally, the nanostructure
should comprise a two-component mixture with both donor and accep-
tor forming continuous structures with maximum interfacial area and
a mean domain size commensurate with the exciton diffusion length,
which is 5 to 10 nm. In practice, judicious choices of external parameters
such as solvent type and annealing temperature are needed to reach the
desired nanoscale interpenetrating network for phase segregated P3HT
and PCBM. Figure 6.25 shows an example of the nanostructure that can
be accomplished experimentally [59].
The efficiency of the OSC hence depends critically on its manu-
facturing details, and structural control at the nanoscale is necessary.
Figure  6.26 is a direct proof of this and shows IPCE data on a 1:1
blend of P3HT and PCBM in its pristine state and after heat treatment.
Clearly, the heating has improved the properties significantly and η =
2.7% was recorded for this particular OSC [60]. Plasmonic effects can
be used to improve the light absorption on OSCs, as in other types of
solar cells [61,62].

© 2011 by Taylor and Francis Group, LLC

 Heat and Electricity 297

60
Pristine
50 Annealed

40
IPCE (%)

30

20

10

400 500 600 700 800

Wavelength (nm)

Figure 6.26  Spectral incident photon to current conversion efficiency

(IPCE) for an OSC in pristine state and after annealing at 120°C for 1 h.
(From X. Yang et al., Nano Lett. 5 [2005] 579–583. With permission.)

References

1. G. M. Morrison, Solar collectors, and solar water heating, in Solar

Energy: The State of the Art, edited by J. Gordon, James & James
Science Publishers, London, 2001, chap. 4 and 5, pp. 145–289.
2. Z.-Q. Yin, Development of solar thermal systems in China, Solar
Energy Mater. Solar Cells 86 (2005) 427–442.
3. T. P. Otanicar, P. E. Phelan, J. S. Golden, Optical properties of liquids
for direct absorption solar energy systems, Solar Energy 83 (2009)
969–977.
4. B. O. Seraphin, Spectrally selective surfaces and their impact on
photothermal solar energy conversion, in Solar Energy Conversion:
Solid-State Physics Aspects, Springer, Berlin, 1979; Topics in Applied
Physics Vol. 31, chap. 2, pp. 5–55.
5. G. A. Niklasson, C. G. Granqvist, Spectrally solar-absorbing surface
coatings: Optical properties and degradation, in Materials Science
for Solar Energy Conversion Systems, edited by C. G. Granqvist,
Pergamon, Oxford, U.K., 1991; pp. 70–105.
6. E. Wäckelgård, G. A. Niklasson, C. G. Granqvist, Selectively solar-
absorbing coatings, in Solar Energy: The State of the Art, edited by
J. Gordon, James & James Science Publishers, London, 2001, chap.
3, pp. 109–144.

© 2011 by Taylor and Francis Group, LLC

 298 Green Nanotechnology

7. N. Etherden, T. Tesfamichael, G. A. Niklasson, E. Wäckelgård, A theo-

retical feasibility study of pigments for thickness-sensitive spectrally
selective paints, J. Phys. D: Appl. Phys. 37 (2004) 1115–1122.
8. J. A. Thornton, J. L. Lamb, Thermal stability studies of sputter-
deposited multilayer selective absorber coatings, Thin Solid Films
96 (1982) 175–183.
9. C. G. Granqvist, Coatings of ultrafine chromium particles: Efficient
selective absorbers of solar energy, Phys. Scripta 16 (1977) 163–164.
10. E. Rephaeli, S. Fan S., Tungsten black absorber for solar light with
wide angular operation range, Appl. Phys. Lett. 92 (2008) 211107
1–3.
11. G. A. Niklasson, C. G. Granqvist, Optical properties and solar selec-
tivity of coevaporated Co-Al2O3 composite films, J. Appl. Phys. 55
(1984) 3382–3410.
12. T. Tesfamichael, E. Wäckelgård, Angular solar absorptance of
absorbers used in solar thermal collectors, Appl. Opt. 38 (1999)
4189–4197.
13. Å. Andersson, O. Hunderi, C. G. Granqvist, Nickel pigmented
anodic aluminum oxide for selective absorption of solar energy, J.
Appl. Phys. 51 (1980) 754–764.
14. E. Wäckelgård, A study of the optical properties of nickel-pigmented
anodic alumina in the infrared region, J. Phys: Cond. Matter 8 (1996)
5125–5138.
15. B. Carlsson, K. Möller, M. Köhl, U. Frei, S. Brunold, Qualification
test procedure for solar absorber surface durability, Solar Energy
Mater. Solar Cells 61 (2000) 255–275.
16. B. Carlsson, K. Möller, M. Köhl, M. Heck, S. Brunold, U. Frei, J.-C.
Marechal, G. Jorgenson, The applicability of accelerated life testing
for assessment of service life of solar thermal components, Solar
Energy Mater. Solar Cells 84 (2004) 255–274.
17. E. Wäckelgård, G. Hultmark, Industrially sputtered solar absorber
surface, Solar Energy Mater. Solar Cells 54 (1998) 165–170.
18. S. Zhao, E. Wäckelgård, Optimization of solar absorbing three-layer
coatings, Solar Energy Mater. Solar Cells 90 (2006) 243–261.
19. T. Boström, G. Westin, E. Wäckelgård, Optimization of a solution-
chemically derived solar absorbing spectrally selective surface, Solar
Energy Mater. Solar Cells 91 (2007) 38–43.
20. M. Lazarov, R. Sizmann, U. Frei, Optimization of SiO2-TiOxNy-Cu
interference absorbers: Numerical and experimental results, Proc.
Soc. Photo-Opt. Instrum. Engr. 2017 (1993) 345–356.
21. M. Lanxner, Z. Elgat, Solar selective absorber coatings for high ser-
vice temperatures, produced by plasma sputtering, Proc. Soc. Photo-
Opt. Instrum. Engr. 1272 (1990) 240–249.

© 2011 by Taylor and Francis Group, LLC

 Heat and Electricity 299

22. D. Zhu, S. Zhao, Chromaticity and optical properties of colored and

black solar-thermal absorbing coatings, Solar Energy Mater. Solar
Cells (2010) in press.
23. R. Kunič, M. Koželj, A. Šurca Vuk, A. Vilčnik, L. Slemenik Perše,
D. Merlini, S. Brunold, Adhesion and thermal stability of thickness
insensitive spectrally selective (TISS) polyurethane-based paint coat-
ings on copper substrates, Solar Energy Mater. Solar Cells 93 (2009)
630–640.
24. I. Jerman, M. Koželj, B. Orel, The effect of polyhedral oligomeric
silsesquioxane dispersant and low surface energy additives on spec-
trally selective paint coatings with self-cleaning properties, Solar
Energy Mater. Solar Cells 94 (2010) 232–245.
25. B. Orel, private communication (2009).
26. H. Akbari, S. Menon, A. Rosenfeld, Global cooling: Increasing
world-wide urban albedos to offset CO2, Climatic Change 94 (2009)
275–286.
27. I. Edmonds, G. Smith, Conversion efficiency and ground albedo
dependence of technologies for mitigating global warming, to be
published (2010).
28. A. L. Fahrenbruch, R. H. Bube, Fundamentals of Solar Cells:
Photovoltaic Solar Energy Conversion, Academic, New York,
1983.
29. R. H. Bube, Photovoltaic Materials, Imperial College, London,
1998.
30. M. A. Green, Solar Cells: Operating Principles, Technology, and System
Applications, 2nd edition, Bridge Printery, Sydney, Australia, 2002.
31. M. A. Green, Third Generation Photovoltaics: Advanced Solar
Energy Conversion, Springer, Berlin, Germany, 2003.
32. A. Goetzberger, C. Hebling, H.-W. Schock, Photovoltaic materials:
History, status and outlook, Mater. Sci. Engr. R 40 (2003) 1–46.
33. K. L. Chopra, P. D. Paulson, V. Dutta, (2004), Thin-film solar cells:
An overview, Progr. Photovoltaics: Res. Appl. 12 (2004) 69–92.
34. J. Berry, private communication (2009).
35. A. Romeo, M. Terheggen, D. Abou-Ras, D. L. Bätzner, F.-J. Haug,
M. Kälin, D. Rudmann, A. N. Tiwari, Development of thin-film
Cu(In,Ga)Se2 and CdSe solar cells, Progr. Photovoltaics: Res. Appl.
12 (2004) 93–111.
36. B. J. Stanbery, Copper indium selenides and related materials for
photovoltaic devices, Crit. Rev. Solid State Mater. Sci. 27 (2002)
73–117.
37. B. Tian, X. Zheng, T. J. Kempa, Y. Fang, N. Yu, G. Yu, J. Huang,
C. M. Lieber, Coaxial silicon nanowires as solar cells and nanoelec-
tronic power sources, Nature 449 (2007) 885–889.

© 2011 by Taylor and Francis Group, LLC

 300 Green Nanotechnology

38. H. R. Stuart, D. G. Hall, Absorption enhancement in silicon-on-insu-

lator waveguides using metal islands, Appl. Phys. Lett. 69 (1996)
2327–2329.
39. S. Pillai, K. R. Catchpole, T. Trupke, G. Zhang, J. Zhao, M. A. Green,
Enhanced emission from Si-based light emitting diodes using sur-
face plasmons, Appl. Phys. Lett. 88 (2006) 161102 1–3.
40. C. Hägglund, B. Kasemo, Nanoparticle plasmonics for
2D-photovoltaics: Mechanisms, optimization, and limits, Opt.
Express 17 (2009) 11944–11957.
41. K. R. Catchpole, S. Pillai, Absorption enhancement due to scatter-
ing by dipoles into silicon waveguides, J. Appl. Phys. 100 (2006)
044504 1–8.
42. K. R. Catchpole, A. Polman, Design principles for plasmon enhanced
solar cells, Appl. Phys. Lett 93 (2008) 191113 1–3.
43. B. O’Regan, M. Grätzel, A low-cost, high-efficiency solar cell based
on dye-sensitized colloidal TiO2 films, Nature 353 (1991) 737–740.
44. Y. Chiba, A. Islam, Y. Watanabe, R. Komiya, N. Koide, L. Han, Dye-
sensitized solar cells with conversion efficiency of 11.1%, Jpn. J.
Appl. Phys 45 (2006) L638–L640.
45. H. Tributsch, Dye sensitization solar cells: A critical assessment of
the learning curve, Coord. Chem. Rev. 248 (2004) 1511–1530.
46. A. Hagfeldt, M. Grätzel, Molecular photovoltaics, Acc. Chem. Res.
33 (2000) 269–277.
47. H. Lindström, E. Magnusson, A. Holmberg, S. Södergren, S.-E.
Lindquist, A. Hagfeldt, A new method for manufacturing nano-
structured electrodes on glass substrates, Solar Energy Mater. Solar
Cells 73 (2002) 91–101.
48. M. Gómez, E. Magnusson, E. Olsson, A. Hagfeldt, S.-E. Lindquist,
C. G. Granqvist, Nanocrystalline Ti-oxide-based solar cells made
by sputter deposition and dye sensitization: Efficiency versus film
thickness, Solar Energy Mater. Solar Cells 62 (2000) 259–263.
49. T. Yamaguchi, N. Tobe, D. Matsumoto, H. Arakawa, Highly effi-
cient plastic substrate dye-sensitized solar cells using a compression
method for preparation of TiO2 photoelectrodes, Chem. Commun.
(2007) 4767–4769.
50. C. Hägglund, M. Zäch, B. Kasemo, Enhanced charge carrier genera-
tion in dye sensitized solar cells by nanoparticle plasmons, Appl.
Phys. Lett. 92 (2008) 013113 1–3.
51. S. D. Standridge, G. C. Schatz, J. T. Hupp, Toward plasmonic solar
cells: Protection of silver nanoparticles via atomic layer deposition
of TiO2, Langmuir 25 (2009) 2596–2600.

© 2011 by Taylor and Francis Group, LLC

 Heat and Electricity 301

52. T. Lindgren, J. M. Mwabora, E. Avendaño, J. Jonsson, A. Hoel, C.

G. Granqvist, S.-E. Lindquist, Photoelectrochemical and optical
properties of nitrogen-doped titanium dioxide films prepared by
reactive DC magnetron sputtering, J. Phys. Chem. B 107 (2003)
5709–5716.
53. Y. Nosaka, M. Matsushita, J. Nishino, A. Y. Nosaka, Nitrogen-
doped titanium dioxide photocatalysts for visible response prepared
by using organic compounds, Sci. Technol. Adv. Mater. 6 (2005)
143–148.
54. C. J. Brabec, J. R. Durrant, Solution-processed organic solar cells,
MRS Bull. 33 (2008) 670–675.
55. B. R. Saunders, M. L. Turner, Nanoparticle–polymer photovoltaic
cells, Adv. Colloid Interface Sci. 138 (2008) 1–231.
56. B. C. Thompson, J. M. J. Fréchet, Polymer-fullerene composite solar
cells, Angew. Chem. Int. Ed. 47 (2008) 58–77.
57. W. Ma, C. Yang, X. Gong, K. Lee, A. J. Heeger, Thermally stable,
efficient polymer solar cells with nanoscale control of the inter-
penetrating network morphology, Adv. Funct. Mater. 15 (2005)
1617–1622.
58. J. Y. Kim, K. Lee, N. E. Coates, D. Moses, T.-Q. Nguyen, M. Dante,
A. J. Heeger, Efficient tandem polymer solar cells fabricated by all-
solution processing, Science 317 (2007) 222–225.
59. K. Sivula, C. K. Luscombe, B. C. Thompson, J. M. J. Fréchet,
Enhancing the thermal stability of polythiophene: Fullerene solar
cells by decreasing effective polymer regioregularity, J. Am. Chem.
Soc. 128 (2006) 13988–13989.
60. X. Yang, J. Loos, S. C. Veenstra, W. J. H. Verhees, M. M. Wienk,
J. M. Kroon, M. A. J. Michels, R. A. Janssen, Nanoscale morphol-
ogy of high-performance polymer solar cells, Nano Lett. 5 (2005)
579–583.
61. S.-S. Kim, S.-I. Na, J. Jo, D.-Y. Kim, Y.-C. Nah, Plasmon enhanced
performance of organic solar cells using electrodeposited Ag nano-
particles, Appl. Phys. Lett. 93 (2008) 073307 1–3.
62. D. Duche, P. Torchio, L. Escoubas, F. Monestier, J.-J. Simon, F. Flory,
G. Mathian, Improving light absorption in organic solar cells by
plasmonic contribution, Solar Energy Mater. Solar Cells 93 (2009)
1377–1382.

© 2011 by Taylor and Francis Group, LLC

 Chapter 7
Coolness
High-Albedo Surfaces and
Sky Cooling Devices

Cooling is as important as heating—at least. Cooling is needed to make

buildings comfortable, and it is necessary for the preservation of food
and for innumerable other purposes. Yet cooling is seldom talked about
as much as it deserves to be, and one problem in that respect is with lan-
guage. We provide warmth to heat a room, but what do we provide if we
want to make it cool? Is it “coolth”? This word is not entirely unknown
in English but is archaic and arcane at best. We have chosen to speak
of “coolness” in this chapter. Coolness can be obtained by many tech-
niques, and two are of particular interest for this book: the coolness con-
nected with prevention of excessive temperatures through reflection of
incoming solar radiation by surfaces having high albedo (“whiteness”),
and the coolness that can be captured under a clear sky—that is, what
we referred to as “sky cooling” in earlier chapters.
The energy needed for cooling increases in the world, and it does
so steeply [1,2]; it is becoming increasingly important, especially for the
peak demand of electricity. Solid numbers are not always easy to come
by, but it is known that the increase in the energy for air conditioning and
refrigeration has risen by an average of 17% per annum in the European
Union between 1995 and 2003 [3]. Such an increase, of course, is unsus-
tainable and must be curtailed. Another indication of the same growth
trend, from a different part of the world, is that more than half of the
electricity demand in some major Chinese cities is for air conditioning,
refrigeration, and water cooling [4].
Our emphasis in the past chapters has been on using and controlling
natural energy flows during the daytime when solar radiation is present.
We now also consider the natural bounty that the clear sky has to offer,
especially at night. To most people, including many engineers and scien-
tists, what we will reveal in this chapter at first seems too good to be true.

303

© 2011 by Taylor and Francis Group, LLC

 304 Green Nanotechnology

Figure 7.1  “Dew-rain” harvest system outside Ajaccio, Corsica,

France. The white foil is designed so as to collect water efficiently. (From
www.opur.fr.)

But large cooling capability without external power is almost effortless

once you know which materials and structures to use. This cooling is
almost “unreal” in both senses of the word—the literal “not believable”
and the more modern meaning “amazing.” Cooling of this kind can
lead to many new applications; one of them is illustrated in Figure 7.1,
which is a photo of a foil-type device capable of spontaneously reaching
subambient temperatures and going below the dew point so that water is
collected [5]. This “dew rain” can then be used for human consumption,
irrigation, etc. The energy for cooling the device is not visible; it consists
of flows of heat energy in the form of thermal radiation that is always
going out from the earth and coming in from the atmosphere, with more
going out than in during the night.
We cannot see the flows of infrared energy with our eyes, but infra-
red cameras and simple hand-held noncontact thermometers can detect
them. Point such a thermometer at the ground and then up into the
clear night sky! The measured temperatures are very different and can
be lower for the sky reading by tens of degrees Celsius. The thermom-
eter can measure the sky to be at −40°C when the ground is at around
+20°C. So the clear sky is very much colder than the air! But why are
they different? If there is a cloud about and you point the thermometer
at it, the cloud will appear much warmer than the clear sky.
These simple observations, whose origins were outlined in Chapter
2, hint at the possibility of cooling to temperatures well below those of

© 2011 by Taylor and Francis Group, LLC

 Coolness 305

the ambient without any need for external power, and they also empha-
size the great advantages of a clear sky over a cloudy sky. To understand
these ideas in a simple way, imagine for a moment that your eyes could
“see” in the infrared at night. Emerging from the earth would then be
all of the “IR colors” given off by most bodies at ambient temperature.
The radiation falling onto the earth from above, however, would look
distinctly different because one band of “color” that is very strong in
the outgoing radiation is weak and almost missing. This means that the
atmosphere is more or less unable to radiate some “colors” so a cooler
sky results. By the way, starlight would stand out with a distinct “color,”
as it is only for wavelengths where the incoming atmospheric radiation is
weak that IR “starlight” can easily get through to the earth’s surface.
Chapters 2 and 3 showed that it is the trace gases in the atmosphere
that cause these radiative properties; in particular the amounts of water,
CO2 , and ozone were found to be critical for getting the net outflow of
heat from the earth by thermal radiation just right for most forms of life.
We saw that the atmosphere is delicately tuned with regard to the amount
of radiant heat that gets through into space, and that much of this radia-
tion gets trapped and reradiated back to earth. The materials and struc-
tures discussed in this chapter are aimed at making maximum use of this
“sky window”—which is more “open” in dry geographic locations than
humid ones—while minimizing the impact of incoming radiation, which
includes both thermal radiation and daytime solar energy.

7.1  Two Cooling Strategies

Cooling can be achieved by use of two classes of surfaces that are dis-
tinctly different with regard to function and relevant spectral range.
Thus, cooling can be accomplished by

• Reflection of excessive solar energy

• Radiation of energy directly into outer space, where it is perma-
nently lost

This chapter will discuss both types of surfaces.

7.1.1  High-Albedo Surfaces

The first class of surfaces reduces solar heat gains by increasing the
albedo (i.e., by reflecting an increased amount of solar energy). This
works for an individual building and vehicle as well as for a whole urban
environment. In particular, there are very great benefits if the surfaces

© 2011 by Taylor and Francis Group, LLC

 306 Green Nanotechnology

within a global urban environment become more solar reflective. Roofs

occupy 20 to 25% of the urban areas while road surfaces occupy 30 to
44%, and in total they make up 55 to 65% of the urban area exposed
to the sun. The surfaces are huge in absolute numbers and amount to
~3.8 ×1011 and ~5.3 × 1011 m 2 for roofs and paved areas, respectively,
according to a recent study [6]. Typical roof materials absorb 75 to 90%
of the incident solar energy, and typical road surfaces absorb 70 to 90%,
whereas fresh bitumen and asphalt has a solar absorptance of ~95%.
The potential to reduce the solar heat gain is great for both elements,
and reasonable practical goals are that the albedos are increased by 0.25
and 0.15 for roofs and roads, respectively [6]. In combination, the aver-
age solar reflectance over all urban areas then rises by 0.1. The resulting
gain in the earth’s solar reflectance is 0.03% if this takes place world-
wide in villages, towns, and cities.
The reduction in heat gain, and hence in temperature rise due to
solar absorption, can be equated to an amount of additional incoming
heat from a rise in atmospheric greenhouse gases so that the net impact
on earth’s climate is zero. It has been estimated that a rise in urban
albedo across the globe by an average of 0.1 would offset the addition
of 44 Gt (gigatonnes) of CO2 to the atmosphere [6]. At current growth
rates in emissions, the change in albedo would offset about 11 years of
emissions increase, which would be highly beneficial by allowing new
energy technologies some breathing space to take hold. If carbon offsets
are worth US$ 25 per ton, then the effort is worth 1.1 trillion U.S. dol-
lars, which would pay for much of what is needed to create this impact.
Much of the albedo change could be achieved by normal activities, with
just a slight change in materials used. Some excellent coatings that are
well suited to increase the albedo will be analyzed in Section 7.3.
The mentioned 44 Gt savings in CO2 emissions is in essence a one-
off reduction in heat gain, but the use of high-albedo building materials
will also have a continuous effect and reduce the need for cooling of the
buildings by 20% or more. Some examples from roof albedo changes
on actual buildings (to be discussed shortly), in fact, show even larger
effects. The cooling energy benefit also extends to the whole city or town
when its average local air temperature falls. Urban “greening” from parks
and from gardens on rooftops, walls, balconies, and public and private
grounds can also save in both heating and cooling needs, in addition to
giving albedo gains. The connection between the urban microclimate,
energy use, and health benefits, it seems, is not widely recognized.
Like the adverse effects of the fluorescent lamp on daylight use from
the 1950s, which were noted in Chapter 5, air conditioning technologies
have sadly desensitized many of us to natural ways of thermal manage-
ment and, as with daylight, we then also lost many psychological and
health benefits.

© 2011 by Taylor and Francis Group, LLC

 Coolness 307

7.1.2  Sky Cooling

The second class of surface of main interest in this chapter pumps heat
away by radiative cooling to the atmosphere and dumps the heat into
space. The energy is removed directly by transmission through the sky
window and indirectly via absorption in the atmosphere and re-radia-
tion (about a third of which goes outward, as seen in Section 2.1). High
levels of cooling can be achieved with surfaces of this type as long as
there is no incident solar energy. But it is also possible to achieve useful
sky cooling in the daytime, though some special arrangements must then
be made as will be discussed later. A combination of high solar reflec-
tion and efficient sky cooling in one surface can lead to daytime cooling
and is thus of great relevance for this chapter; solar reflection keeps the
building cool, while sky cooling contributes to make its radiative output
surpass the solar heat gain so that subambient cooling starts earlier in
the afternoon than would be the case without sky cooling.
One can even go further with some surfaces designed for sky cool-
ing and combine heating and cooling in a single surface or single stacked
system having suitable spectral properties. This can be done sequentially
with daytime heating and night-time cooling but, most fascinating, is it
is even possible to simultaneously supply coolness and heat in a specially
designed single-stacked system. This idea can be varied to provide—
instead of the heat–cold combination—a joint supply of daylight and
cooling, or electric power and cooling.
To evaluate the benefits and relative costs of albedo increases, sky
cooling, and other novel cooling concepts, it is necessary to compare sav-
ings per dollar invested to that for advanced conventional systems and
other options for natural cooling. This is discussed further in Box 7.1.

7.2  City Heating, Global Cooling,

and Summer Blackouts

The world’s population does not stop increasing. At the moment (2010),
it amounts to about 6.8 billon people, and it is not expected to stabilize
until after 2200 when the population is just above 10 billion [9]. More
than half of the population now lives in cities; 670 million inhabit the 100
largest cities, and the mean area per person is 560 m 2 [6]. Naturally, this
agglomeration of people and activities influences the cities’ climates.

7.2.1  Urban Heat Islands

Decreasing the temperatures in the cities is particularly important

since they trap heat and hence get hotter than their surrounding nearby

© 2011 by Taylor and Francis Group, LLC

 308 Green Nanotechnology

Box 7.1 On the Efficiency of

Alternative Cooling Techniques
There are many ways to obtain cooling, and electrically powered
compressors, absorption cycle chillers, evaporative cooling sys-
tems, and thermoelectric cooling can all be adapted for use with
solar power or solar heat; it is also possible to have hybrid systems
combining sky cooling with more conventional methods. Many of
these options can cool without drawing electric power from the
grid. Some need no electric power at all apart, perhaps, for fans or
pumps, and some just increase the performance efficiency.
Electrically powered compressors for cooling can be driven by
solar cells, but this is very expensive. Solar heat can be used for
absorption cycle refrigeration or air conditioning, but these sys-
tems, which rely on heat-induced chemical changes, have much
lower efficiencies than electrically powered systems. A key concept
here is the coefficient of performance (COP), which measures how
much supplied energy Ein is needed to remove a given amount of
heat Q H. Clearly, COP = Q H /Ein should be as high as possible.
One should note that all of these systems dump an amount of heat
Qout equal to Q H + Ein into the local environment that is larger
than the heat Q H removed from a room or fridge. If the COP has
a value one, as is usually the case for absorption cycle cooling, the
thermal impact locally is severe and is more than twice the heat
that is pumped away. And if the exhaust heat is extracted by water
evaporation, then there is a high demand on water resources and
an elevation of local humidity, both of which are undesirable.
One of the authors (GBS) was involved in solar thermal absorp-
tion cycle air conditioning of a large medical center in a Sydney
suburb, plus a university-based facility, in the late 1970s [7]. The
system worked well but was far too expensive to run and maintain,
and it also used a lot of water. Other solar-based systems were
evaluated around that time and had typical COPs in the range 0.5
to 0.6 [8]. The COP of absorption chillers has not advanced very
much in the past 30 years. Modern two-stage absorption coolers
with >200 kW cooling capacity can have COP ≈ 1.1, but single-
stage coolers still only achieve COP ≈ 0.7.
Sky cooling, in contrast to the cooling discussed above, does
not have any adverse impact on the local environment and actu-
ally helps ameliorate the urban heat island effect discussed next,
whereas these other options—even if solar driven—will exacer-
bate it.

© 2011 by Taylor and Francis Group, LLC

 Coolness 309

33
Late Afternoon Temperature (°C)
32

31

30

Rural Suburban Commercial Downtown Urban Park Suburban Rural

Residential Residential Residential Farmland

Figure 7.2  Schematic of the urban heat island effect, showing tem-
perature rise for different types of urban landscape. (From http://www.
urbanheatislands.com. With permission.)
countryside. This is shown schematically in Figure  7.2, which indi-
cates that the temperature difference can be as large as ~4°C between
a downtown area with high-rise buildings and its rural setting.
Exceptionally, the energy increase can be even larger, and specific data
are available for a number of big cities [10]. This “urban heat island”
(UHI) effect has several reasons, the major ones being

• Low albedos of roofs and roads

• High thermal mass
• Commercial and industrial activity
• Pumping of heat from inside buildings to just outside them

For the specific case of Greater London, the air conditioning load was
estimated to be 25% above that of the surrounding rural region due to
the UHI effect [11].
As discussed above, increasing the albedos of roofs and roads will
lower the solar heat gain, and thereby the UHI. The reduced need for air
conditioning will lead to

• Reduction in local heating outside buildings and cars

• Power savings
• Reduced leakage of greenhouse refrigerant gases
• Less water needed for cooling towers

Cooler urban environments also dramatically reduce urban smog [12,13]

and thus have health benefits.

© 2011 by Taylor and Francis Group, LLC

 310 Green Nanotechnology

7.2.2  Global Cooling by Increased Albedo

The discussion in Section 7.1.1 showed that albedo changes can be

very important for offsetting CO2 emissions. Here, we analyze this
effect further, mainly by way of two illustrative examples presented in
Box 7.2. The first example deals with air cooling and leads to the main
conclusion that a 20% reduction globally of the power for air condi-
tioning would give the same effect on CO2 emissions as an increase
of the urban albedo effect discussed above by as little as 0.1%, or in
absolute numbers, an increase of the urban albedo by 0.0001. We will
show below that there are techniques that can raise the albedo much
further. The second example is related to power stations and demon-
strates again the overwhelming effect of even minor changes in the
urban albedo.
The very large effect of albedo changes may come as a surprise to
most people and, to get some perspective on them, let us consider the
other ways by which we heat up the atmosphere. All waste heat, which
is a by-product from power plants, industry, car and truck exhausts,
and air conditioning, can be reduced if each of these sources becomes
more efficient or, in some cases, if there is a switch to nonheat-gener-
ating processes. No doubt, efficiency gains will reduce the projected
heat burden in coming years. It is worth looking at this likely impact
on global cooling, relative to the albedo effect, to emphasize the lat-
ter’s magnitude because this raises a key issue in the way we humans
respond, and the need for new ways of thinking as we seek to com-
bat global warming. We sense the heat coming from air conditioner
compressors, car exhausts, steel plants, and foundries, and we see its
effects in the steam coming off the cooling towers of power plants.
These heat outlets are localized. In contrast, the impacts of raised solar
reflectance are subtle and disperse. It may thus be surprising that the
albedo effect on the UHI and on global cooling by far outweighs what
one might expect to gain from improved efficiencies in these more
obviously “hot” systems. Of course, both the one-off carbon offsets
and ongoing CO2 reductions from each of the “hot” sources are still
very beneficial.
These analyses irrevocably point at the major potential impact of
paints and surface coatings on global warming. Their joint direct and
indirect effects will, if implemented, outweigh those from wind, solar
cells, and solar thermal energy by an enormous margin for many years
to come. Add in the relative very low cost per unit of CO2 emissions,
both offset and cumulatively avoided, and surface coatings should
become an immediate high-priority focus area. Thus, Section 7.3 on
paints should be seen as a core section in this book.

© 2011 by Taylor and Francis Group, LLC

 Coolness 311

Box 7.2 Quantitative Albedo

Effects: Two Examples
The first example is in regard to air cooling. Typical cooling units
add additional heat to the neighborhood given by input energy Ein
= [(heat removed from a space)/COP] = Q H /COP, where COP is
the coefficient of performance. World air conditioning in 2003
needed 395 TWh per year, and by 2030 that is projected to rise to
1,269 TWh [14]. The additional external heat load globally due to
air conditioning is thus now around 45 GW at any one time and,
assuming an average COP equal to three (i.e., high performance
chillers), 180 GW is continuously heating up outside urban air.
This heat load depends on the season and is peaked in the middle
of the day, so instantaneous values per urban square meter can be
very much higher than the 24-hour average.
A reduction in the power for air conditioning by 20%, say, would
give a direct global cooling that is equivalent to a global CO2 off-
set by 40 Mt (megatonnes), which is 0.1% of the offset from the
albedo effect of 44 Gt discussed in Section 7.1.1. This 0.1% offset
may seem small, but it is still very worthwhile, and there are two
main benefits: locally, there is a positive effect on the UHI effect
and associated smog reduction, and globally there is a net build-
up of CO2 reductions relative to business-as-usual (BAU) due to
the associated reduction of cooling power needs of ~80 TWh each
year. Here, BAU corresponds to unchanged urban albedos. This
benefit is 80 Mt per annum, assuming coal fired power, so it takes
just 6 months to match and then surpass the environmental benefit
from the decrease in direct cooling. After 5 years, the accumu-
lated contribution to atmospheric cooling relative to BAU (a 440
Mt reduction) is still only 1% of the benefit given by an increase
of the albedo by 0.1. It should be remembered that if a 20% heat
output reduction today (2010) decreases the global and UHI heat
load by 79 TWh, such a percentage decrease in 2030 will diminish
the global heat load by 250 TWh, with a direct emission offset of
126 Mt of CO2 and an ongoing accumulating reduction relative to
BAU of 253 Mt of CO2 per year.
To emphasize the albedo effect once more, we now look at a
second example: power stations. The total atmospheric heat load
just from power stations—given typical efficiencies—is now prob-
ably around 35,000 TWh each year (~24,000 TWh from coal and
~11,000 TWh from oil and gas), or at any one time around 4 TW
of heat. This massive total is less than the direct heating reductions

© 2011 by Taylor and Francis Group, LLC

 312 Green Nanotechnology

that follow from raising the urban albedo by 0.1. A 10% reduction
in power station heat output worldwide—either from efficiency
gains, more gas power, replacement by renewables, or some combi-
nation of these—would offset the warming of nearly one extra Gt
of CO2 added to the atmosphere. To achieve an equivalent reduc-
tion from raised albedo would require lifting its average by a very
modest 0.0025! Note that, as in the example above of air condi-
tioning reductions, CO2 reductions from this combination of more
efficient power stations and renewables would continue to com-
pound year after year, and would surpass the 10% direct heating
offsets of ~1 Gt in 8 to 11 months. However, they would take over
10 years to surpass the benefits of raising average urban albedos
by 0.1! What is even more interesting is the relative financial ben-
efits, or return on investment, of the albedo approach versus the
renewables approach, both in terms of investment capital per ton
of CO2 reduction and relative accumulated returns over 10 years,
assuming a fixed price of CO2 per ton avoided or offset. These
financial considerations enhance the value of the albedo approach
even more.
Similar comparative analyses could be pursued in other exam-
ples applied to reductions of waste heat outputs from industrial
processes, especially from metal and ceramic processing, and from
vehicle exhaust emissions. The less heat pumped into the air, the
more CO2 can be added without a rise in temperature. The near-
to-medium term scope is thus probably up to 50 Gt of CO2 offsets,
with any further upside to come from the effect of an albedo rise
being too conservative.

7.2.3  Avoiding Summer Blackouts

The UHI effect is a growing problem. It not only detracts from the pleas­
ure of being outside but actually compounds the demands on cooling
systems as hot air infiltration adds to cooling loads and also degrades
chiller efficiencies. The more heat derived from cooling one can pump
into outer space the better, which is why it is so important to make more
use of solar reflectance and sky cooling.
The air conditioning demand reaches a peak when the weather is
at its hottest, near midday or early afternoon. Then, wholesale electric
power gets very expensive, inefficient generators are fired up, and black-
outs ensue when demand exceeds supply. Power supply from solar cells
or solar thermal power stations is advantageous for this situation, as
their output peaks when the load maximizes and traditional power costs

© 2011 by Taylor and Francis Group, LLC

 Coolness 313

are high. However, among all current options, it is even more energy
efficient and financially attractive to spread the load over time and use
conventional power stations at night for cooling. Chillers work more
efficiently at night, and one can further reduce demand by using supple-
mentary night sky cooling. A number of state-of-the-art air conditioning
systems cool glycol-water mixtures to near ice temperature overnight
and then use the stored cold fluid the next day in various ways such
as to cool overhead beams. The cooled thermal mass of the building
can involve using interior walls or floors, columns or beams filled with
chilled water-glycol, or suitably located phase change materials. The lat-
ter can be micro-encapsulated waxes. Greater use of night cooling with
conventional compressors plus storage is thus to be encouraged. This
also opens the way for hybrid operation with advanced night sky cool-
ing systems. These do not seem to have been considered before, but have
much to offer. They will reduce compressor power needs and also pump
much of the exhaust heat into space instead of into nearby air. Such
hybrid systems are well suited to both large buildings and to homes.
The sky cooling devices introduced in Section 7.4 may also be applicable
in homes for collecting and storing cold fluid overnight to fully supply
cooling needs the next day. The implication is that if cooling loads in
homes are not too high—which presumes good design—these systems
could completely eliminate the need for electrically powered cooling,
apart from a small amount of power for fans or pumps.
Other sources of natural cooling also need to be considered in
buildings, including natural ventilation, cool underground soil, and
evaporative cooling. In some parts of the world, snow and ice can be
accumulated during the winter and used for cooling during the summer.
But ultimately these sources still pump heat or water vapor into the local
environment, just as electrically powered cooling does. Night sky cool-
ing avoids this.

7.3  High-Albedo Paints for Cool Buildings

The thermal performance of façades and roofs in buildings is a key to

determine the need for heating or cooling energy and for the thermal
comfort of the occupants. The improvement potential is huge, and it has
been estimated that retrofitting of façades and roofs to make them more
thermally efficient can reduce today’s (2010) demand for heating and
cooling by 50 to 60% [14]. The need for action is urgent, particularly
in warm countries undergoing rapid development, since improved liv-
ing standards tend to rapidly accelerate the use of electrically powered
air conditioning [4]. Increased use of air conditioning in cars and other
forms of transport also emphasizes the need for better control of solar

© 2011 by Taylor and Francis Group, LLC

 314 Green Nanotechnology

heat gains. Both coatings and insulation materials are of importance for
improving the energy efficiency.
Chapters 4 and 5 demonstrated that windows and luminaires are sig-
nificant elements for the façade’s impact on the internal energy demand.
These are important for the building’s apertures but, of course, the
opaque parts of the building envelope matter very much, too, and, as we
will discuss here, new types of paints have much to offer and can be as
significant as good thermal insulation. Furthermore, as explained earlier
in this chapter, improved roof coatings can contribute substantially to
raising the urban albedo and hence offset large amounts of added CO2 ,
as well as contribute to demand savings and create nicer microclimates.
For a given color of the paint, different combinations of selective
coatings and insulation may yield the same overall improvement of the
thermal performance of a wall or roof of a building in a given geo-
graphic location. Therefore, the final decision of what paint to actually
use will come down to relative costs, availability of the different options,
aesthetic appeal, ease of application and installation, and whether the
building is under construction or being retrofitted. Routine maintenance
of painted surfaces may provide an opportunity for easy improvement
in performance once the technologies and approaches we address here
become more widely available or lower in cost. The number of white and
colored paint products being marketed as “solar” or “sun reflecting” has
increased greatly in recent times as people are becoming more interested
in green technology and aware that solar and thermal control can yield
large energy savings.

7.3.1  How Cool Can a Solar Exposed Roof Get?

A surface having a neutral color can range from very bright white with
Asol ~ 10%, through dull white and light gray with Asol between 40 and
60%, to near black with Asol ~ 90%. So, whitish-looking surfaces are not
necessarily very good solar reflectors and may absorb as much as half of
the incident solar energy in some cases. Also, dull white solar irradiated
surfaces with Asol ≈ 50% can get hot and yield significant internal heat
transfer to whatever is thermally connected to such a surface.
It is not uncommon that roofs reach 60°C to 75°C under clear skies,
so it is easily realized that roofs should have high solar reflectance com-
bined with high thermal emittance, at least for warm and hot climates.
The high emittance not only helps keep down daytime temperatures on
roofs and walls but allows the roof, and often the interior and building
mass, to cool to a temperature a few degrees below that of the ambient
at night. With highly solar reflecting roofs and walls it can take some
time after sunrise to overcome the stored coolness. This type of spectral

© 2011 by Taylor and Francis Group, LLC

 Coolness 315

selectivity minimizes solar heat gain and maximizes emitted radiation

and hence is the exact reverse of what is needed for the efficient solar
thermal collectors discussed in Section 6.1; the ideal spectrum for the
surfaces of present interest was introduced in Section 2.9.
Normal metal roofing is neither strongly solar reflecting nor strongly
thermally emitting, so it has far from ideal properties. A well-known
treatment for roofs involves coatings that are heavily pigmented with
TiO2 microparticles. This pigment has a high refractive index so the
paints backscatter strongly and hence reflect the solar heat, and suit-
able additional constituents can render them efficient thermal radiators.
A variety of such high-performance basic white paints are available.
Further inorganic and other additives can improve them optically and
thermally and sometimes also enhance their long-term durability; such
improvements can be especially important for roofs, given their high
solar exposures and large temperature changes.
Coatings with high solar reflectance and high thermal emittance,
and related pigmented polymer foils, have been discussed over the years
for both daytime use and for night sky cooling [15–18]. The best “cool”
roof paints combine a hemispherical solar reflectance of ~90% with a
blackbody thermal emittance of ~95%, and are able to maintain these
characteristics for many years with only slight changes; these paints
have a diffuse appearance, which is important since they do not produce
glare. The same coating will be mentioned as an excellent candidate for
day–night sky cooling in Section 7.4.
Glare can be disturbing, and there has been confusion on this issue,
which may have prevented the uptake of coatings with low solar absorp-
tance. In general, a highly diffuse but strongly reflective coating is much
less of a problem for glare than a partly smooth but much darker coating
(or even a plane glass sheet). In other words, glare from reflected solar
light is a large problem for surfaces with a moderate specular compo-
nent if they are viewed from a point to which the mirror component
of the reflected beam is directed. And it is difficult not to look at that
bright spot! A highly diffuse white surface, on the other hand, spreads
the reflected energy into the whole outgoing hemisphere, so viewing
from any one direction can be maintained with relative ease. Therefore,
bright white paint rarely causes glare unless it has a semisheen or glossy
surface, and a glossy black surface may cause more glare than a diffuse
white one.
The practical energy savings of a coating with high solar reflectance
and thermal emittance can be very large. Box 7.3 reports specific data
for three different commercial buildings, demonstrating that the power
for air conditioning can be decreased to a fraction of what is demanded
with standard roofing, and that the average roof temperature can be
decreased by 15°C to 20°C or even more.

© 2011 by Taylor and Francis Group, LLC

 316 Green Nanotechnology

Box 7.3  Practical Experience for

Buildings with “Cool” Painted Roofs
Figure B7.3.1 shows power for air conditioning, measured during
1.5 years for two nearly identical supermarkets located in the same
region and exposed to the same warm-to-hot climate [19]. When
the standard finish was used, the power reached 40 to 45 MWh per
month during several months. A “cool” paint had a dramatic effect
and limited the power to between 15 and 22 MWh per month.
50
45
40
35
MWh/month

30
25
20
15
10
May-07

May-08
Jul-06
Aug-06
Sep-06
Oct-06
Nov-06
Dec-06
Jan-07
Feb-07
Mar-07
Apr-07

Jun-07
Jul-07
Aug-07
Sep-07
Oct-07
Nov-07
Dec-07
Jan-08
Feb-08
Mar-08
Apr-08

Jun-08
Jul-08
Figure B7.3.1  Air conditioning power usage in two almost
identical nearby supermarkets near Brisbane, Australia, during
a 1.5-year-long measurement campaign. Upper and lower curves
refer to buildings having standard external finish and coated with
“cool” paint, respectively. (From J. Bell et al., Proceedings CD and
Summary Book of the CIB International Conference on Smart and
Sustainable Built Environment, Brisbane, Australia, [November
19–21, 2003] [ISBN: 1-74107-040-6]; Paper T606.)

The huge energy savings observed in Figure B7.3.1 cannot all be

attributed to direct reductions in heat gain. There are other ben-
efits, too, which are related to two main influences, namely

• Significantly improved microclimate, especially just at and

above the roof
• The “cool” coating’s ability to reach temperatures below
those of the ambient overnight

The cooler microclimate can increase the efficiency of roof-

mounted cooling units and/or provide cooler air-to-air exchange
systems. The quantitative levels of power savings are dependent on

© 2011 by Taylor and Francis Group, LLC

 Coolness 317

the installation of the cooling unit and how it is operated, the way
internal air exchange occurs, and the roof and ceiling insulation.
The conclusion is that the very large savings evident from Figure
B7.3.1 may not always occur, but at least half of the shown savings
can be expected in most cases.
Figure B7.3.2 reports additional data demonstrating the large
impact of a “cool” roof finish, here applied to a newspaper office/
production facility [20]. When a standard roof material was used,
the temperature frequently reached 50°C to 70°C. Replacing this
material with a “cool” white paint led to a very significant decrease
in the roof temperature: the temperature difference was ~17°C on
average and was exceptionally as large as ~25°C.
80

70

60
Temperature (°C)

50
Plain roof

40
SkyCool roof
30

20 Ambient
Before application After application
10
1/11/2007
3/11/2007
5/11/2007
7/11/2007
9/11/2007
11/11/2007
13/11/2007
15/11/2007
17/11/2007
19/11/2007
21/11/2007
23/11/2007
25/11/2007
27/11/2007
29/11/2007
1/12/2007
3/12/2007
5/12/2007
7/12/2007
9/12/2007
11/12/2007
13/12/2007
15/12/2007
17/12/2007
19/12/2007
21/12/2007
23/12/2007
25/12/2007
27/12/2007
29/12/2007
31/12/2007
2/01/2008
4/01/2008
6/01/2008
8/01/2008
10/01/2008
12/01/2008

Figure B7.3.2  Maximum roof temperatures on a newspaper

office/production facility in Sydney, Australia, during a 2.5-month-
long measurement campaign. Data are shown for a standard plain
roof, for a roof coated with “cool” white paint, and for the ambi-
ent. (From SkyCool Pty Ltd., Sydney, Australia [2007], private
communication.)

A third example of the great benefits of “cool” paints is a 35,000

m 2 roof on the Melbourne Airport. This project also involved
additional cooling at night by thermal storage in building elements
and contents, whose thermal mass then reduced the warming on
the following day. The project led to reductions in CO2 emission
exceeding 4,000 tons during 18 months.

© 2011 by Taylor and Francis Group, LLC

 318 Green Nanotechnology

The “cool” paints used on the supermarket, office/production facil-

ity, and airport discussed in Box 7.3 raised the roof albedo by some 0.4
to 0.5. The magnitude of the change makes it interesting to return to the
discussion in Section 7.1 about the possibility of using increased urban
albedos to offset CO2 emissions on a global scale. As argued there, an
overall change of the roof albedo by 0.25, together with a change of the
road albedo of 0.15, would lead to an offset of these emissions by an
amount corresponding to 11 years of expected increases. The conclusion
then is that the “cool” paints can have a significant impact on the global
CO2 emissions if they are implemented in the urbanscape on a large
enough scale.
The “cool” paints discussed above are visibly white. This is fine on
shopping malls, supermarkets, warehouses, factories, airport terminals,
some offices, and the like, but there are many roofs and walls that must
be colored for aesthetic diversity and appeal. The same applies to cars.
Is it still feasible to increase the albedo in this case? The answer is “yes”
and, as discussed below, it is possible to have visibly identical black paints
with a solar reflectance of 5 and 55%, that is, an albedo gain of 0.5 can
be achieved. All “cool” paints should have maximum NIR reflection,
and this property should be maintained while the visible reflectance
is altered as desired for color. Most standard paints do not have high
NIR reflectance, though, so we need to examine ways of achieving that.
Various nanostructures are one option, as discussed below.

7.3.2  Colored Paints with High Solar Reflectance

It is important to understand how the hemispherical solar reflectance

R H,sol can be divided into components for ultraviolet, luminous, and near-
infrared reflectance. Using the data in Chapter 2, it can be found that

R H,sol = 0.05R H,UV + 0.43R H,lum + 0.52R H,NIR (7.1)

that is, 52% of the reflectance is in the infrared. This shows that high
solar reflectance can only be obtained if the NIR reflectance is very large.
And if a distinct color is required, not just a pale one, then the luminous
reflectance should be confined to a narrow band. The latter feature can
be achieved with certain dyes, select pigments which may incorporate
nanofeatures, and some nanoparticles.
The ideal colored “cool” paint for warm or hot climates would thus
have, neglecting the UV, a maximum solar reflectance of 0.43R H,lum +
0.52. Thus, a black paint with R H,lum ~ 0.05 could still have R H,sol as
high as 0.55. But a conventional carbon black coating can achieve R H,sol
~ 0.06 at best. It then follows that a black roof, wall, or car body that

© 2011 by Taylor and Francis Group, LLC

 Coolness 319

100
90
80
Transmittance [%]
Lumogenblack
·
70
60
50
40
30
20
10 Carbon black
0
300 400 500 600 700 800 900 1000 1100 1200 Wavelength [nm]
UV Visible Near infrared (NIR)

Figure 7.3  Spectral transmittance for a “cool” black and a conven-

tional carbon black. (From B. Schuler, Proceedings of Convation06,
Sydney, Australia, 2006. With permission.)

reflects most of the NIR reflects almost ten times more solar energy than
a conventional black paint! This qualitative difference can lead to large
reductions in roof temperatures, from ~65°C to around 35°C, with ensu-
ing lower internal heat gains.
Figure 7.3 shows transmittance spectra of a “cool” black paint and
of a normal black paint based on carbon [21]. The “cool” paint is trans-
parent in the NIR and, if it is applied to a reflecting backing, the reflec-
tance spectrum corresponds to the transmittance spectrum of a coating
at double thickness. The conventional black, on the other hand, has a
low reflectance throughout the spectral range shown.
The data in Figure 7.3 do not extend beyond 1.2 μm in wavelength.
Most of the NIR solar energy lies at 0.7 < λ < 1.2 μm, so it is most impor-
tant that the reflectance is high in this range. Nevertheless, the optical
properties are of some interest also at longer wavelengths, since many
pigments and binders may have absorption features there. Figure  7.4
shows specific examples of “cool light brown” and “cool green” paints
[22]. Both paints have optical properties in good agreement with those
for their standard counterparts within the luminous range. Clearly,
the brown paint reflects moderately well in the entire NIR spectrum,
whereas the green paint displays unwanted low reflectance for 1.2 < λ
< 1.8 μm.

7.3.3  Mechanisms and Nanostructures for Colored “Cool” Paints

This section will discuss a number of strategies for combining color

and high NIR reflectance of the kind illustrated in Figures 7.3 and 7.4.
Nanostructures will be seen to play important roles for these strategies.
The focus will be on roof and wall paints, but it should be noted that

© 2011 by Taylor and Francis Group, LLC

 320 Green Nanotechnology

100 UV VIS NIR

Cool light brown
80
Reflectance (%)
60

40

20 Standard light brown

0
300 800 1300 1800 2300
Wavelength [nm]

100 UV VIS NIR

80
Reflectance (%)

Cool green Standard green

60

40

20

0
300 800 1300 1800 2300
Wavelength [nm]

Figure 7.4  Spectral luminous and NIR reflectance for two “cool” col-
ored paints and for their standard counterparts. (From A. Synnefa et al.,
Solar Energy 81 [2007] 488–497. With permission.)

white and colored paints with opposite infrared properties are of inter-
est for surfaces inside rooms, and they can reduce inward heat flows
from hot walls and roofs; these paints require a low emittance. Thus,
ideal paints for indoor and outdoor applications have radically different
thermal properties.
Oxide-coated metal flakes can be used for “cool” paints. Aluminum
flakes are added to a transparent binder to create a paint coating with a
cross section as illustrated in Figure 7.5a. Such paints would be neutral
in color and highly solar reflecting. A variety of colors can be accom-
plished by having each flake coated with an oxide of precisely controlled
thickness in the nano range, as seen in Figure 7.5b. The visual impres-
sion is created by optical interference, and different oxide thicknesses
in the range between 20 and 200 nm lead to a variety of colors. The
thin-oxide layers are largely transparent in the NIR, so reflection in that
range occurs from the Al flake. In practice, the coatings are produced
in a hot fluidized bed reactor using an iron-containing organometallic
compound so that each flake is coated with iron oxide [23,24].
A single-layer coating on the Al flakes gives colors which look similar
from most directions. With double layers, however, the perceived color

© 2011 by Taylor and Francis Group, LLC

 Coolness 321

Clear topcoat
Oxide
coated flakes
Initial coat
Metal substrate
(a)

(b)

Figure 7.5  Cross-sectional views of a colored “cool” paint layer con-

taining oxide-coated metal flakes (a) and of an individual flake with a
double-layer coating (b).
changes as the view direction is altered, while the high NIR reflection
is preserved. This latter application uses Al flakes that have been pre-
coated with nano-thin SiO2 layers via sol-gel coating before the iron oxide
layer is applied. This two-layer coating is notable for its use of two of the
earth’s most abundant oxides. A clear top overcoat, shown in Figure 7.5a,
imparts a high emittance—as is normally desired—but a different and
thinner top coat would make it possible to achieve a low emittance.
Coated metal flakes in paint binders have other applications as well,
and most surfaces, including those on wood and fabrics, can nowadays
be coated with a commercial paint which provides a low thermal emit-
tance of about 0.2 to 0.3. Metal flakes can also be incorporated directly
into or onto the polymer or on natural fibers used to weave fabrics. The
aim of the flakes, apart from decorative innovation and shiny fabrics,
can be to enhance the thermal insulation. For daytime cooling, such
a coating can be used on inner surfaces to reduce incoming heat and
on outer surfaces to reflect solar heat. At night, both types of surfaces
reduce heat loss. The metal flakes also reduce the wearer’s thermal image
at night, and military applications of this kind have probably been the
main interest to date.
Paints incorporating metal flakes can find additional applications for
solar thermal systems, as discussed in Section 6.1.4. They may also be of
interest as low-cost coatings combining high solar reflection with high-
performance angular selective sky cooling. The desired structure looks
like that in Figure 7.5, but the overall thickness should be up to 1.5 μm
(i.e., much larger than for a “cool” paint for which 0.2 μm suffices).
Inorganic pigments which transmit most of the NIR give another
option for colored “cool” paints. Here, one uses a strongly NIR-reflecting
substrate directly under a pigmented colored layer that is at least mod-
erately NIR transparent. This transparency is the main feature, and
the pigments are normally in nanoparticle form. A variety of special

© 2011 by Taylor and Francis Group, LLC

 322 Green Nanotechnology

inorganic pigments are available, and their visual appearance is close to

that of many standard paints that have much higher solar absorptance
[22]. Paints of the type indicated here clearly avoid the need for metal
flakes and can be used on any type of base surface including wood shin-
gles, ceramic tiles, polymer sheet, and metal sheet, so this paint can be
of great practical value considering the vast array of existing buildings
that need to be made more energy efficient easily and at low cost. The
under-layer or substrate for the paint can either be an undercoat or the
metal onto which the coating is applied.
Plasmonic nanoparticles that impart color but transmit most of the
NIR is the third and final option for colored “cool” paints to be consid-
ered here. This option has received less consideration than the other two
but has interesting features, including ease of incorporation into fabrics.
Select nanoparticles which display surface plasmon resonances can have
narrow and strong absorption bands in the visible. If they also transmit
at NIR wavelengths, then they can be used for “cool” paints. Obviously,
there are some analogies to the laminated windows discussed in Section
4.5.3. In an application, the nanoparticles are added into an other-
wise transparent thin overlayer, just as for the inorganic pigments. The
nanoparticles of practical interest must be very durable, which largely
excludes the noble metals and aluminum. Plasmonic nitrides and some
borides are robust, though, and thus have interesting possibilities. ZrN,
TiN, and HfN nanoparticles can impart color.
The best possibilities may be for core-shell nanoparticle systems in
which the shell or core is conducting, provided they can be mass-produced
at low enough cost. Again, the shell material at least must be robust. The
beauty of the core-shell nanoparticles, as explained earlier in Section 3.9,
is that they can be color tuned just by varying the core-shell size ratio.
One such system has been considered for nanoinks [25]. Figure 7.6 shows
spectral absorption and extinction for two systems: (a) Al cores and Si
shells, and (b) Ag cores and Si shells. Absorption dominates but there is
some scattering. A narrow absorption peak stands out clearly; its loca-
tion can be varied within the visible spectrum when the thickness of the
metal shell is changed. Scattering also peaks near the resonance.

7.4  Sky Cooling to Subambient Temperatures

This aspect of green nanotechnology has received attention only spasmod-

ically over a number of years [26], and the applications have emphasized
passive cooling of buildings and water collection. This is the demand or
efficiency side of the energy equation. But sky cooling also has poten-
tial as a widespread supply side or active cooling technology—not just
for passive removal of heat though that will be always be one aspect.

© 2011 by Taylor and Francis Group, LLC

 Coolness 323

5
4 nm

8 nm
4
12 nm
Cross-Sections

2 nm
2
18 nm

0
0.3 0.35 0.4 0.45 0.5 0.55 0.6 0.65 0.7 0.75 0.8
Wavelength [+m]
(a)

12
No coat
10 2 nm 4 nm
6 nm

8 10 nm
Cross-Section

0
0.3 0.35 0.4 0.45 0.5 0.55 0.6 0.65 0.7 0.75 0.8
Wavelength [µm]
(b)

Figure 7.6  Spectral absorption (solid curves) and extinction (dashed

curves) for shells of Si on (a) Al and (b) Ag cores of 44 nm diameter.
The Si shells have the thicknesses shown. (From M. R. Kuehnle, H.
Statz, Encapsulated nanoparticles for the absorption of electromagnetic
energy, U.S. Patent Publication US 2005/0074611 A1 [2005].)

© 2011 by Taylor and Francis Group, LLC

 324 Green Nanotechnology

“Active” means cooling and storing cold fluids, or hybrid integration into
active systems such as directly in the condenser circuits of normal chillers
or indirectly to improve external heat removal. An additional potential
that is not often considered is for electric power supply side, as discussed
later in Section 7.6. Cooling is needed in all electric power systems, as
outlined in Box 2.2, and this adds directly to global warming; sky cool-
ing can have major implications for which renewable sources to use and
where to locate them. Water collection from the atmosphere, and as an
aid to water condensation in distillation, is still another area where sky
cooling will have an impact [27]; this is discussed in Section 7.5.
But before one embarks on active sky cooling to maintain comfort-
able interiors in buildings, the base cooling load should be minimized by
having good thermal insulation, appropriate roof and façade technologies
(such as high-albedo paints, as discussed above), and advanced fenestra-
tion as discussed in Chapter 4. The overall design is also essential for new
or renovated buildings. Indeed, substantial reductions in cooling loads
can be obtained by diminishing direct and indirect gains from solar irra-
diation, and if this is done well, the desired cooling can be accomplished
by use of moderate areas for emitting radiation to the sky.
Why has the great potential of sky cooling not been successfully
exploited to date, despite our ready access to it? First, it is not a widely
understood or appreciated field, and few scientists are active in it. But
more important, there has been little effort to develop products based
on sky cooling, possibly apart from arrangements for water collection.
The diverse technological scope beyond these applications is not well
understood, and this “knowledge gap” has yet to be bridged.
Clearly, the field of sky cooling has so far fallen short of its potential
by a wide margin, but it has too much to offer to be neglected. Thus,
practical cooling at a low cost down to 15°C below the coldest ambient
temperature of the night has been demonstrated, and we will discuss
how even lower temperatures are achievable.

7.4.1  Sky Radiance

Chapter 2 introduced the transparency of the atmosphere as a function

of wavelength λ and zenith angle θz and emphasized the existence of a
sky window capable of yielding subambient cooling. The sky window
was found to gradually close as θz approached 90°. We now go into more
detail on the spectral and directional properties of the sky radiance.
This information is needed in order to devise optimum materials and
device designs for sky coolers, and for modeling cooling performance
for different atmospheric conditions. The sky radiance emerges from all
directions and is measured in units of Wm−2Sr−1.

© 2011 by Taylor and Francis Group, LLC

 Coolness 325

Outside the sky window, the atmosphere acts essentially as a black-

body radiator with regard to both wavelength and angle. Inside the sky
window, the radiance is weak for small zenith angles, which is a conse-
quence of the very low temperature of outer space. Nevertheless, there
is a small atmospheric absorptance within the sky window, which adds
somewhat to the zenith radiance at small zenith angles but adds a lot at
large zenith angles. This absorptance originates from a sharp absorp-
tion feature due to ozone as well as weak residual absorption from CO2
and water vapor, and it must be included in the analysis. The angular
dependence occurs because, effectively, the atmospheric layer thickness
d a contributing to the radiance is d a(θz) = d a(0)/cosθz. Hence, the atmo-
spheric transmittance in the sky window, denoted Ta,sw(θz,λ), drops for
increasing values of θz while there is an associated rise in the emittance
E a,sw in the sky window according to

Ea,sw (θz , λ) = 1 − Ta,sw (θz , λ) = 1 − [Ta,sw (0, λ)]1/cosθz (7.2)

This equation agrees with Equation 2.22. Further discussion on this

matter can be found elsewhere [28].
But how, in detail, does the intensity of the sky radiance depend on
atmospheric conditions? Figure 7.7 reports data obtained from compu-
tations based on an elaborate representation of the atmospheric con-
stituents and their height profiles, and illustrates a number of important
features: Part (a) reports the angular dependence of the sky radiance for
the case of a moderately dry atmosphere characteristic of a clear mid-
latitude summer climate, and shows results for 0 < θz < 90° [29]. The sky
window is seen to be gradually closed as θz goes up, and the variation is
consistent with Equation 7.2.
The role of atmospheric humidity, which is implicit in Figure 7.7a, is
elaborated in Figure 7.7b, which shows zenith radiance for three amounts
of precipitable water [30,31]. This parameter is the length a water col-
umn would have if all of the water along a straight path perpendicular
from the earth’s surface to the top of the atmosphere would condense.
Figure 7.7b shows that the sky window closes as the amount of precipi-
table water goes up. Hence, sky cooling is increasingly efficient as the
humidity drops. The lowest amount of precipitable water in the figure is
1.5 cm, but it can be as small as 0.3 to 0.4 cm under particularly favor-
able conditions. The sharp spectral feature at 9.7 μm, due to ozone, is of
particular interest as it gives the strongest contribution to the radiance in
the sky window under clear and dry conditions. This feature has implica-
tions for the selection of materials, and a material that absorbs entirely
within the sky window range, but not near 9.7 μm, can reach much lower
temperatures than one whose absorption band straddles the ozone peak.

© 2011 by Taylor and Francis Group, LLC

 0
5 10 15 20 25
Wavelength [µm]
Spectral Radiance of Clear Skies
(a)
Midlatitude Summer Atmosphere

326
10
10
90°
Water vapor [cm]
75° 1.5
8 3

Radiance [MW(Sr–1)m–3]
6
Radiance [MW(Sr–1)m–3]

Blackbody
60° 6
5

Green Nanotechnology
0° 4

0
0
5 10 15 20 25 5 10 15 20 25
Wavelength [µm] Wavelength [µm]
(a) (b)
  

10 Spectral radiance for a clear mid-latitude summer sky at different zenith angles including the vertical and
Figure 7.7 
the horizontal (part a) and spectralWater
zenith radiance
vapor [cm] from clear skies containing different amounts of precipitable water
(part b). (Part a from Berdahl, R. Fromberg,1.5 Solar Energy 29 [1982] 299–314. With permission. Part b from P. Berdahl,
M. Martin,8Proceedings of the Second National
3 Passive Solar Conference, edited by D. Prowler, I. Duncan, B. Bennett,
Radiance [MW(Sr–1)m–3]

American Section of the International Solar 6 Energy Society, Newark, DE [1978], Vol. 2, pp. 684–686, as replotted in
Blackbody
C. G. Granqvist, A. Hjortsberg, J. Appl. Phys. 52 [1981] 4205–4220. With permission.)
6

© 2011 by Taylor and Francis Group, LLC

 Coolness 327

In other words, spectral fine-tuning is of interest within the sky window

range. This is one aspect where nanoeffects will come into play and, as
we will see below, SiC nanoparticles are ideal from this perspective.
Figure 7.7a demonstrated that the atmospheric radiance has a strong
angular dependence, and it is worth exploring whether angular selec-
tive surfaces can enhance the cooling of an underlying material. Just
as for spectral selectivity the aim is to reach as low temperatures as
possible by further reducing the amount of absorbed radiation while
still emitting strongly for angles where it is of value to do so. Thus, an
angular selective surface devised for sky cooling should have very low
reflectance within the sky window toward the zenith, but the reflectance
should gradually rise as the angle of incidence goes up. This occurs for
many smooth surfaces, whose reflectance increases steeply at angles of
incidence above 60° and a similar dependence can take place in some
doped polymer coatings. Deliberate engineering of reflectance profiles
with two- or three-layer thin films is also feasible. For some situations it
is useful to have an angular-spectral profile for which the absorptance
is blackbody-like for zenith radiation and decreases continuously as θz
increases across the whole Planck spectrum.

7.4.2  Spectral Selectivity and Sky Cooling: Idealized Surfaces

When surfaces are exposed to the sky, they can experience cooling due
to a net loss of thermal radiation. This means that the thermal radiation
given off by the surface has to be larger than that coming in from the
atmosphere (and possibly other surrounding surfaces) and is absorbed,
rather than being reflected or transmitted. And absorption of incom-
ing radiation reduces the cooling potential. The incident radiation is
largely beyond our control, but it is possible to control how much of it is
absorbed, provided that we understand the spectral and angular distri-
bution of the sky radiance. This was discussed in the previous section.
The technologies addressed here aim to maximize the net radiative out-
put, which is the difference between the outgoing and the absorbed part
of the incoming radiation, and it is very instructive to consider outgoing
and incoming radiation relative to each other when the goal is to cool to
subambient temperatures. Cooling of hot objects, on the other hand, is
a lot different as the incoming radiation is then much weaker, relatively
speaking. A number of strategies for maximizing the difference will be
introduced, but the main emphasis is on the spectral properties of the
emitting surface.
Section 2.7 outlined the basic physics of thermal radiation from
surfaces at finite temperatures and especially introduced the blackbody
radiator, which emits with a spectral power density given by P(λ,τ). This

© 2011 by Taylor and Francis Group, LLC

 328 Green Nanotechnology

35 1
E (Black body)
30
Intensity (W/m2/µm) 0.8
25

Emittance
20 0.6
Planck spectrum 290 K
15
0.4
10
E (“Ideal” sky emitter) 0.2
5

0 0
0 5 10 15 20 25 30 35
Wavelength [µm]

Figure 7.8  Spectral emittance (or absorptance) of a blackbody and

ideal sky window emitter relative to the Planck radiation spectrum from
an atmosphere at 290 K.

is the Planck spectral radiant exitance, which we refer to for simplicity

as the “Planck spectrum.” The surface temperatures τs of interest in sky
cooling lie near or below those of the ambient atmosphere at τa , and
then P(λ,τs) ranges on average from around 2.5 to 35 μm in wavelength.
Chapter 2 also introduced the ideal spectral properties of a surface
capable of reaching very low temperatures at night (cf. Figure 2.15b). Its
spectral reflectance R(λ) was not that of a blackbody, which has R(λ) = 0
over the Planck spectrum. Instead, the ideal radiator only radiates with
100% efficiency, or equivalently has R(λ) = 0, at wavelengths where the
atmosphere is significantly transparent to thermal radiation, which is
the segment of the Planck spectrum we called the “sky window” from 8
to 13 μm. Over the remainder of the Planck spectrum, this ideal surface
reflects all of the incoming radiation, that is, it has R(λ) = 1.
Thus, the blackbody and this ideal sky window radiator represent
two spectral extremes, as shown in Figure 7.8. We shall see that in prac-
tice surfaces approaching both of these limits have roles to play, because
both radiate strongly within the sky window. Which is to be preferred
depends mainly on how far below τa one wants to cool, and also on
the details of the practical arrangement for cooling and its geographic
location. A blackbody is better just below ambient because it then gives
off a great deal more radiation than the selective emitter. It should be
kept in mind, however, that a surface approaching the ideal properties
of the sky window selective radiator, shown in Figure 7.8, always is able
to get coldest if nonradiative heat gains from the surrounds are kept
very small, as we shall soon see. Given that both spectral limits have
their uses, one also needs to ask whether it is also worth considering

© 2011 by Taylor and Francis Group, LLC

 Coolness 329

intermediate spectral properties. The answer is that some work quite

well, but not all.

7.4.3  Calculated Cooling for Ideal and Practical Materials

We will first consider various surfaces open to the sky and then surfaces
under practical covers designed so as to limit convection gains but trans-
mit most of the radiated heat. Suitable cover materials were introduced
in Section 3.5. The net radiated power Prad without covers is given by

 π /2 ∞ 
Prad = 

0
∫ ∫
d (sin θz ) d λP(λ, τ s )Es (θz , λ) 
2

0


(7.3)
 π /2 ∞ 

0
∫ 0
∫
−  d (sin θz ) d λP(λ, τ a )Ea (θz , λ)Es (θz , λ) 
2



It is important to understand each of the terms in this equation. The first

term is the outgoing radiation from the surface, which is at temperature
τs. The second term is the incoming atmospheric radiation that is not
reflected or transmitted and thus involves both the atmospheric emit-
tance E a and surface emittance or absorptance E s. If atmospheric tem-
perature and humidity are fixed, this second term does not change, but
as cooling proceeds and τs drops, the first term weakens until Prad = 0. At
this point τs is the radiative stagnation temperature τs,min. If the surface
happens to be colder than τs,min , then it will be warmed by atmospheric
radiation. The change in solid angle and intensity projection onto the
surface with angle of incidence are combined in the d(sin 2θz) term as
introduced in Section 2.4 for general diffuse beams.
How do different surface spectral properties influence the two terms
in Equation 7.3? The first term will be largest for a blackbody emitter,
so why bother with a sky window emitter? The ideal sky window emit-
ter has E s ~ 0.32, so it emits a lot less radiation than the blackbody. Its
value lies in the second term, which is very small for the ideal spectrum
of Figure 7.7 for clear dry skies. We shall now show that ultimately this
factor enables the sky window emitter to pump heat at a higher rate
and achieve much lower temperatures than the blackbody. But at near-
ambient temperature it does have a lower value of Prad. One should keep
in mind that when we get to real surfaces and systems, Equation 7.3
dictates a balance between the advantages of maximizing output (first
term) and minimizing input (second term), but since the second term is

© 2011 by Taylor and Francis Group, LLC

 330 Green Nanotechnology

fixed and the first term falls with increasing τs this balance gradually
shifts toward minimizing the second term. Thus, the preferred practical
spectrum depends on how far τs needs to be below ambient temperature.
The humidity also plays a significant role, as expected from Figure 7.7b.
Interestingly, the drier the atmosphere, the more the balance is shifted
toward broadband “blackish” emitters. This may seem counter-intuitive
but arises because the second term in Equation 7.3 is then weaker for all
surfaces, and hence it takes a lower magnitude of τs for the sky window
emitters to be superior. This influence of humidity is one example of a
broader principle that will come up again soon, namely, external factors
that act to reduce the incoming radiation enhance the relative merits of
a blackbody to those of a spectrally selective emitter.
A number of results based on Equation 7.3 have been presented in
the past for ideal and blackbody surfaces. The detailed results depend on
the model of the atmosphere. Figure 7.9 shows data from an application
of this model to both the blackbody and the ideal sky window emitting
surface of Figure 7.8 for hemispherical radiation in the case of various
climate zones ranging from tropical to very dry [32]. The highest cooling
performance occurs for the “U.S. standard atmosphere,” for which the
precipitable water is as small as 1.83 mm. The plots then steadily shift
down as humidity levels rise. A key result is that the ideal spectrally selec-
tive emitter can achieve, in the absence of nonradiative heat gains, much
lower stagnation temperatures than the blackbody. It will get to 50°C to
60°C below ambient temperature, while the blackbody reaches only 25°C
to 30°C. For temperatures above 10°C below ambient, the blackbody
is pumping more heat radiatively because the output term in Equation
7.3 then dominates, but by 15°C below ambient, the spectrally selective
radiator is superior because the input term is then much more important.
Nonradiative heat gains, which input heat according to Pnr = Unr(τa − τs),
where Unr is a constant, limit practical stagnation. The common line in
Figure 7.9 is for a very well-insulated system with Unr = 1 Wm−2K−1. The
stagnation points occur when Prad = Pnr. Average values of (τa − τs,min) for
both types of surfaces are closer together when Unr has a nonzero value,
but the spectrally selective surface still cools further by around 8°C to
10°C. The low-cost practical systems in the following have Unr values
that are higher and lie at 2 to 2.5 Wm−2K−1, so stagnation temperatures
around 15°C below the ambient are their limits. In this case, the spec-
trally selective surface still achieves a lower value of τs,min, but the black-
body reaches a temperature that is only two or three degrees higher.
Infrared transparent convection covers are needed in practice for
achieving a low value of Unr. This cover has a moderate impact on the sec-
ond term in Equation 7.3, but it does not influence the first term since the
power density leaving the surface depends only on τs. To find the incom-
ing power density before absorption one must replace P(λ,τa)E a(θz,λ) in

© 2011 by Taylor and Francis Group, LLC

 Coolness 331

120
Hemispherical radiance:
100 TROP
MS
80 MW
SS
60 Blackbody SW
surface US STD
40
Radiated Power, Prad [Wm–2]

20

0
10 20 30
100

80
Infrared-selective
60 surface

40

20

0
0 10 20 30 40 50 60 70
Temperature Difference [ºC]

Figure 7.9  Calculated relationships between sky cooling power and

temperature difference for infrared-selective and blackbody surfaces freely
exposed to atmospheres representing six different climatic conditions spe-
cifically being tropical (TROP), midlatitude summer (MS), midlatitude
winter (MW), subarctic summer (SS), subarctic winter (SW), and the U.S.
standard (US STD). The shaded areas are inaccessible for a device with a
nonradiative heat transfer coefficient of 1 Wm−2K−1. (From T. S. Eriksson,
C. G. Granqvist, Appl. Opt. 21 [1982] 4381–4388. With permission.)
Equation 7.3 by three terms: (i) Tc(θz,λ)P(λ,τa)E a(θz,λ), (ii) Rc(θz,λ)P(λ,τs)
E s(θz,λ), and (iii) Ec(θz,λ)P(τa ,λ). The coefficients Tc(θz,λ) and Rc(θz,λ)
are the directional spectral transmittance and reflectance of the cover,
respectively, and Ec(θz,λ) = 1 − Tc(θz,λ) − Rc(θz,λ) is the emittance of the
cover. The terms (i) – (iii) are easily found either directly from data on
the cover sheet or from the optical constants and thickness of the materi-
als, normally polymers, that are used. Term (i) amounts to a reduction in
the incoming radiation due to the cover, whereas terms (ii) and (iii) add
to it. Term (iii) is the contribution of thermal emission from the cover.
In practice, the polymer is always at a temperature very close to ambient

© 2011 by Taylor and Francis Group, LLC

 332 Green Nanotechnology

temperature, so one can assume that any slight tendency it has to warm
due to (weak) absorption of radiation emitted from the surface of inter-
est can be neglected.

7.4.4  Some Practical Surfaces for Sky Cooling: Bulk-Type Solids

Section 3.5 introduced ionic and molecular absorption processes that can
yield narrow absorption bands in the sky window range. The absorption
can ensue from the natural ionic or molecular vibration modes in some
materials comprised of light atoms including Si, Al, Mg, C, N, and O.
Practical experiments have been conducted on thin films of materials such
as silicon monoxide [31,33], silicon nitride [34], and silicon oxynitride
[35,36]. Other alternatives include a number of polymer layers (with poly
vinyl fluoride being especially good [15,37]) and layers of MgO and LiF
[38], as well as gaseous ammonia, ethylene, and ethylene oxide [39,40].
These materials yield little absorption across the Planck spectrum lying
outside the sky window range, and the desired selective reflectance pro-
file can be approximately reached provided that the thin-film or polymer
layer has an adequate thickness and is backed by a highly reflecting metal
such as an aluminum film. Some selectivity can be achieved also with
anodized alumina [41,42] and with alumina films [43], but the data are
not as good as for the previously mentioned alternatives.
Figure 7.10a illustrates spectral reflectance for an optimized silicon
oxynitride film [36]. High absorption in the sky window is due to molecu-
lar vibrations in tetrahedrally coordinated SiO4, SiO3N, SiO2N2 , SiON3,
and SiN4 structural units [36], and the high reflectance outside this win-
dow is caused by the underlying Al. Figure 7.10b shows spectral reflectance
of a thick ceramic MgO layer backed by metal [38]; the large reflectance
at λ > 13 μm is due to the Reststrahlen effect (cf. Section 3.5).

7.4.5 Nanotechnology for Optimum Sky Radiators:

Computed and Measured Data

A number of surfaces and materials have been used to demonstrate sky

cooling as discussed above. None of them was developed in order to ben-
efit from the possibilities offered by nanotechnology. However, nano-
structures can indeed add new sky cooling capabilities, raise performance
by doping in some existing materials, and lower costs. Specifically, the
phonon properties of the nanoparticles allow the infrared optical prop-
erties to be fine tuned to the sky window, and nanoparticles can serve as
convenient dopants for polymers or paint binders.

© 2011 by Taylor and Francis Group, LLC

 Coolness 333

100

Reflectance (%) 80

60

40

20

0
5 7 10 20 50
Wavelength [µm]
(a)

100

80
Reflectance (%)

60

40

20

0
5 8 13 20 30 40
Wavelength [µm]
(b)

Figure 7.10  Spectral reflectance of a 1.2-μm-thick film of SiO1.47 N0.54

made by sputtering onto Al-coated glass (a) and for a 1.1-mm-thick
MgO layer on a metal backing (b). (Panel a from T. S. Eriksson, C. G.
Granqvist, J. Appl. Phys. 60 [1986] 2081–2091. With permission; panel
b from P. Berdahl, Appl. Opt. 23 [1984] 370–372, as replotted in C. G.
Granqvist, T. S. Eriksson, Materials Science for Solar Energy Conversion
Systems, edited by C. G. Granqvist, Pergamon, Oxford, U.K. [1991], pp.
168–203. With permission.)

© 2011 by Taylor and Francis Group, LLC

 334 Green Nanotechnology

Nanoparticles of SiC and SiO2 turn out to be of special interest

[44,45]. Their optical properties are governed by surface resonances in
an ionic material, which has a negative dielectric constant over a lim-
ited frequency range between the transverse and longitudinal phonons at
the frequencies ωT and ωL , respectively, as discussed in Section 3.5. The
frequency range with ε1(ω) < 0 is called the Reststrahlen band. A nano-
sphere of an ionic material has a surface plasmon resonance centered at

ω T2 (2ε h ) + ω 2Lε(∞)
ω 2 SPR = (7.4)
2ε h + ε(∞)

where ε(∞) is the residue from high-frequency absorption and εh is the

dielectric constant of the host in which the nanospheres are embedded.
Equation 7.4 is obtained from the formulas given in Chapter 3, spe-
cifically Equation 3.16 for the polarizability of a sphere. The resonance
occurs when the dielectric constant of the nanoparticle fulfills εp = −2εh,
with εp given for ionic motion by Equation 3.14. Using the values of ωT,
ωL and ε(∞) appropriate for SiC, reported in Table 3.3, together with εh
= 1.0, leads to an absorption peak centered at 11.3 μm (i.e., in the middle
of the sky window). The relaxation rate ωτ is small for SiC, as also found
in Table  3.3, which makes the resonance very sharp and, in fact, SiC
nanoparticles have been put forward as a material where surface plas-
mon modes are particularly clear cut [46]. The desired optical properties
are not achieved if the SiC is in the form of a smooth continuous layer
or large particles; for those cases, the SiC would reflect or scatter in the
resonant range. However, embedding a SiC layer in certain specific mul-
tilayers can produce useful tuning of the phonon absorption, as will be
discussed below. The latter feature is analogous to what one can find in
thin silver layers, as discussed in Section 4.3.4. SiO2 nanoparticles can
serve as an alternative to SiC and are also of interest. They resonate in
the same way as SiC and have a narrow peak centered at 8.97 μm.
We note, in passing, that SiC nanoparticles have been widely studied
by astronomers [47,48]. Such nanoparticles are pervasive in interstellar dust
and leave a sharp spectral signature in the IR radiation from deep space.
This spectral feature lies where the atmosphere has its largest IR transmit-
tance, which clearly has aided ground-based IR astronomy. Nevertheless,
and despite the previous interest in SiC by astronomers, it seems that appli-
cations of this material to sky cooling are only very recent [44,45].
We now turn to detailed results for nanoparticles embedded in
10-μm-thick polyethylene foil and specifically illustrate transmittance
data for doping with 10 vol.% SiO2 , 10 vol.% crystalline SiC (c-SiC), and
20 vol.% amorphous SiC (a-SiC). Figure  7.11 shows calculated results
for which the effect of the nanoparticles was obtained from the models

© 2011 by Taylor and Francis Group, LLC

 Coolness 335

(a) 1
0.9
0.8
0.7
Transmittance

0.6
0.5
0.4 SiO2 nanoparticles
0.3 in PE

0.2
0.1
0
5 10 15 20 25 30
Wavelength [µm]
(b) 1
0.9
0.8
0.7
Transmittance

0.6

0.5
a-SiC nanoparticles
0.4 in PE
0.3
0.2

0.1 c-SiC nanoparticles

in PE
0
5 10 15 20 25 30
Wavelength [µm]

Figure 7.11  Spectral normal transmittance calculated for nanopar-

ticles of SiO2 (a), c-SiC (b), and a-SiC (b) embedded in polyethylene (PE)
foil. Topmost curves refer to undoped PE. Fine structure in the curves
emanate from optical interference. Vertical lines denote the sky window.
(From A. R. Gentle, G. B. Smith, Nano Lett. 10 [2010] 373–379. With
permission.)
for optical homogenization in Section 3.9 [44,45]. The optical constants
of SiC and SiO2 were obtained from the literature [49], whereas optical
data for the undoped foil material were determined separately. Sharp
absorption features due to surface plasmons stand out distinctly within
the 8 < λ < 13 μm sky window for SiO2 and c-SiC. However, a-SiC only
gives a very broad absorption feature, which is as expected, since that
material does not support surface phonons. The qualitative difference
between the data for SiO2 and c-SiC on one hand and a-SiC on the other

© 2011 by Taylor and Francis Group, LLC

 336 Green Nanotechnology

1.0
0.8
Reflectance

0.6
0.4
0.2
0.0
5
6 80 90
7
8 60 70
Wav 9 40 50 ce [°
]
elength 10 11 20 30
n c iden
[µm] I
12 10 le of
13 0 Ang

Figure 7.12  A color version of this figure follows page 200. Spectral
and angular-dependent reflectance for nanoparticles of SiO2 and c-SiC
embedded in polyethylene foil. (From A. R. Gentle, G. B. Smith, Nano
Lett. 10 [2010] 373–379. With permission.)

serves as a clear illustration of the importance of the Reststrahlen condi-

tion on ε(ω).
The SiO2 doping leads to a strong absorption centered at 8.97 μm,
which straddles the narrow but sharp absorption band due to atmo-
spheric ozone. There is also a resonant peak near 20 μm, but this feature
is less pronounced and lies near the tail of the Planck spectrum, so it
is not a major concern here. The c-SiC-doped foil has one very sharp
absorption peak around 11.3 μm, which lies at a wavelength where the
sky window has minimum absorption. Hence, c-SiC nanoparticles are
more efficient with regard to sky cooling than SiO2 nanoparticles.
Angular selectivity is of importance for sky cooling and suppresses
the effect of blackbody-like radiation emerging at large zenith angles, as
discussed previously. Combined angular and spectral selectivity can be
computed, and Figure 7.12 shows some initial results for a 10-μm-thick
polyethylene foil containing 5 vol.% SiO2 and 5 vol.% of c-SiC [45].
This foil is backed by nontransparent aluminum. The reflectance is given
in the 5 < λ < 13 μm range for the full angular interval from normal
to glancing. The low reflectance in the sky window is seen to prevail
for angles up to ~60°, whereas the reflectance rapidly approaches high
values beyond this angle. Interference features stand out more clearly
in Figure 7.12 than in Figure 7.11 as expected from the influence of the
aluminum layer.
We now turn to experimental data on nanoparticle-based coat-
ings. Figure 7.13 shows the structure of clustered 50-nm-diameter SiC
nanoparticles prepared by spray coating onto Al-coated glass. Similar

© 2011 by Taylor and Francis Group, LLC

 Coolness 337

Figure 7.13  Scanning electron micrograph of SiC nanoparticles made

by spray coating onto an aluminum substrate.

coatings, on larger areas of Al sheet, were used in outdoor sky cooling

tests. SiO2 nanoparticle coatings look very much the same.
As an alternative to having a doped polymer, the SiC or SiO2 nano-
particles can be embedded in, or positioned under, a vacuum-coated layer
that absorbs by normal phonon modes in a narrow range not covered by
the surface phonon resonances. Figure 7.14 refers to such a combination,
with SiO2 particles under a 264-nm-thick vacuum-coated SiO thin film
backed by aluminum. The reflectance data shown indicate that phonons
in SiO combine with surface phonons in SiO2 nanoparticles to cover the
sky window range [44]. The coverage would be even better for a thicker
SiO film. Some scattering can be seen at short wavelengths and is due to
clustering, but this effect has a very minor impact on the potential for
sky cooling.
Dispersions of SiC or SiO2 nanoparticles in appropriate silicon-based
or polymer-based paint binders (such as silicones, siloxanes, or polyvi-
nylidene fluorides) can be expected to yield results similar to those in
Figure 7.14 if applied to aluminum in a sufficiently thin layer. Different
mixtures may apply depending on the binder layer’s contribution to the
absorption in the sky window.

7.4.6  Practical Sky Cooling: Systems and Data

There are many design issues to consider when setting up practical sky
cooling devices. Among those already discussed we note local atmospheric

© 2011 by Taylor and Francis Group, LLC

 338 Green Nanotechnology

80
Scattering (due to
Sky window
Scattering nanoparticle clusters)
70

60
Specular Reflectance [%]

50
Dense silicon
40 monoxide
Second SiO2
30 resonance

20
SiO2 nanoparticle
10
surface phonon
resonance
0
3 6 9 12 15 18 21 24
Wavelength [µm]

Figure 7.14  Spectral specular reflectance of a SiO2 nanoparticle layer

under a continuous SiO film and backed by Al. Vertical lines denote the
sky window. (From A. R. Gentle, G. B. Smith, Proc. SPIE 7404 [2009]
74040J 1–8. With permission.)
conditions, average operating temperature relative to that of the ambient,
and spectral properties of the emitting surface. But there are more issues
such as the emitting area needed to provide the desired amount of heat
pumping, the possible use of heat mirrors to confine the radiation angles,
the housing in which the emitting surface sits, the properties of the IR
transparent cover sheet (introduced in Section 7.4.3), the specific location
(e.g., roof, façade, or on open ground), possible tilt to the horizontal, and
the purpose of the cooling. This purpose might require that the collected
coolness is used or stored away from the collecting unit, in which case a
heat exchange fluid and either pumping or thermo-siphoning is needed.
Of course, this fluid should not freeze and it must be in good thermal
contact with the emitting surface as it passes through.
Figure 7.15 illustrates a basic arrangement for practical sky cooling.
Other designs have been summarized elsewhere [26,50]. The coating
indicated in Figure 7.15 may have high thermal emittance or be infra-
red selective and adapted to the sky window. The IR transparent cover
can be UV-stabilized polyethylene. It is needed to reduce convection
exchange with ambient air and, along with the nontransparent insula-
tion, helps achieve a satisfactory nonradiative heat transfer. The particu-
lar design in Figure 7.15 also incorporates a plate with channels for a
heat exchange fluid.
One factor that comes into play for outdoor experiments is water
condensation, which occurs when surface temperature falls below the

© 2011 by Taylor and Francis Group, LLC

 Coolness 339

Coating IR transparent
cover (10 µm PE)

Insulation

Aluminum plate (with embedded

heat exchange channels)

Figure 7.15  Schematic of a basic sky cooling system.

dew point. Condensation thus happens first on the emitter surface,

which is coldest, and may sometimes occur as dew on the cover as well.
Water has high thermal emittance and is hence strongly IR absorbing.
It is essential that water is removed from the cover. The impact of water
condensation on the infrared emitting surface, on the other hand, is
usually not a problem for the cooling performance achievable in basic
set-ups like the one in Figure 7.15. Not even ice formation is usually a
problem. There are several strategies to reduce the effect of water drops
on the cover or IR emitter, and a small tilt of the collector can lead
to adequate drainage. The cooling performance is, of course, affected
by the tilt, and the quantitative effect can be calculated from formu-
las similar to Equations 7.3 and 7.5, to come. Excessive tilt should be
avoided in order to prevent thermal influx from surrounding surfaces.
As discussed in Section 7.5, condensation onto surfaces undergoing sky
cooling can be used for obtaining substantial amounts of water.
Figure 7.16 shows an example of some actual cooling results recorded
over a continuous clear period comprising three nights and two days [51].
Data were taken on two cooling devices placed side by side, both similar
to the one in Figure 7.15. One of the devices had a blackbody-like emit-
ter with a paint surface, and the other had a surface with spectral and
angular-dependent selectivity similar to the data in Figure  7.12. Both
surfaces had high solar reflectance and an emitting area of 0.08 m 2 .
The devices were shaded from direct solar radiation by a wall, but they
received diffuse solar radiation during the day.
The data in Figure  7.16 show many interesting features. It is evi-
dent that both surfaces are capable of significant cooling and reach tem-
peratures below 0°C and that the selective radiator clearly outperforms
the blackbody-like surface except during daytime. The selective surface
quickly gets to 13°C below ambient temperature once the solar intensity
begins to drop, and that temperature difference is then maintained dur-
ing the night. The only deviation from this pattern is during short peri-
ods when some water vapor, enclosed between the radiator surface and

© 2011 by Taylor and Francis Group, LLC

 340 Green Nanotechnology

20

15
Air temperature

10
Temperature [°C]

High emittance/
5 solar reflecting

–5
Selective
–10
Time

Figure 7.16  Cooling performance of two devices similar to the one

in Figure 7.15 during a period of three nights and two days. One of the
devices had a blackbody-like coating and the other a selectively emit-
ting surface. The temperature of the ambient air is shown for reference.
(From A. R. Gentle, G. B. Smith [2010] unpublished data.)

the polyethylene cover, condenses and forms ice; the cooling power is
then needed for the phase change. Once the water has condensed as ice,
it has no obvious effect on the cooling performance.
Figure  7.16 proves that net cooling prevails during the day. The
increase in temperature after sunrise mirrors the drop in net cooling
rate due to the onset of diffuse solar influx. Some sharp but short-lived
temperature peaks at night arise when clouds pass over and add some
incoming radiation. The cooling surface then responds rapidly to the
cloud cover and finds a new steady-state temperature.
The cooling devices giving the data in Figure  7.16 had a nonradi-
ative heat transfer as large as 2.5 to 3 Wm−2K−1. Scaling up and bet-
ter insulation should enable 2 Wm−2K−1 or less, and hence even lower
temperatures. Nevertheless, the transparent cover remains a significant
challenge as elaborated in Box 7.4.
Much further engineering development of sky cooling systems and
their interface with storage or other devices is clearly needed. However,
the potential for easily implemented, low-cost cooling for a variety of
applications is beyond doubt. We believe that sky cooling technology
will become increasingly important and possibly emerge as a centerpiece
for tackling a variety of environmental and lifestyle challenges that the
world will face in coming years.

© 2011 by Taylor and Francis Group, LLC

 Coolness 341

Box 7.4 Transparent Convection
Shields for Sky Coolers
Sky radiators reaching temperatures well below those of the ambi-
ence need good thermal insulation. This is easy for nontransparent
parts of the cooling device but a challenge for the upward-facing
part, which should transmit radiation at least in the 8 < λ < 13 μm
range while providing small non-radiative heat transfer, compris-
ing conduction and convection. The conductive component is rela-
tively straightforward and requires a sufficient distance between
the radiating surface and the shield. At first sight, the convection
might seem easy too for the case of a horizontal device, and one
could hope for a stagnant gas layer that is coldest at the bottom. In
reality, this is a problem, though, since wind blowing past the con-
vection shield will move it—as long as it is a thin foil or a series of
foils—so that forced convection transfers heat to the radiating sur-
face. The convection shield hence must have mechanical rigidity.
A first attempt at developing a material with the properties required
for sky cooling is illustrated in Figure B7.4.1. It shows a three-dimen-
sional material consisting of crossed layers of vee-corrugated high-
density polyethylene foils [52]. Measurements on a typical material
consisting of three components, each with a height h of 1.5 cm and
an apex angle θa of 45°, yielded Unr ≈ 0.9 Wm−2K−1, together with an
infrared transmittance (measured with an IR-imaging instrument) of
73%. The result is encouraging for further materials development,
which might involve other types of polymers in alternative configu-
rations, as well as fiber-based reinforcement. Thus, the transparent
convection shield is not an Achilles’ heel for sky cooling, but it is
definitely a material in need of more development!

h
Θa

Figure B7.4.1  IR-transparent convection shield comprised of cor-

rugated polymers foils characterized by height h and apex angle
θa. (From N. A. Nilsson et al., Solar Energy Mater. 12 [1985]
327–333. With permission.)

© 2011 by Taylor and Francis Group, LLC

 342 Green Nanotechnology

One should appreciate the principal difference from a flat-plate

solar collector, such as the device shown in Figure 6.2. In that case,
the convection shield needs to be transparent in the 0.3 < λ < 3 μm
solar range, which can be accomplished with a pane of iron-free glass
(cf. Figure 4.2) that has a thickness of several millimeters. The solar
wavelengths are shorter than those of typical lattice vibrations, and
a material serving as a transparent convection shield in the thermal
infrared must be judiciously chosen to avoid lattice vibrations in the 8
< λ < 13 μm range and/or have a very small mass. The solar transmit-
tance of this shield is also of concern in many cooling applications.

7.4.7  Amplifying Sky Cooling with Heat Mirrors

Spectral and angular selectivity are not the only strategies to reduce
incoming atmospheric radiation and hence amplifying the net radiative
output of a surface exposed to the sky. If part of the sky vault seen by the
emitter is replaced by a surface with much lower emittance than the sky
in that zone, then the incoming radiation from those solid angles must be
lower. The effect is as though the whole sky, on average, got a lot colder
[53]. The sky is most “black” at high angles to the zenith and most trans-
parent toward the zenith, so the obvious choice is to surround the emit-
ting surface with a heat mirror that leaves open a reasonable-size aperture
around the zenith direction. The heat mirrors must have a low emittance
in order to maximize this benefit, and thus can be made of bare aluminum
or a metallized surface (e.g., silver) with a thin protective oxide layer over
the metal. Such surfaces can have an emittance between 0.02 and 0.1,
whereas the sky at high zenith angles has an emittance approaching unity.
Figure 7.17 illustrates a few practical heat mirror arrangements for boost-
ing sky cooling. Some of these structures have features in common with
solar concentrators used, for example, in solar thermal power plants.
The optical design does not have to be particularly rigorous in order
to provide strong gains, and thus the structures can be low cost. But
it is important that the optical design ensures that incoming near-hor-
izontal rays are not reflected so that they reach the emitting surface.
An acceptance angle, denoted θz,max for the case of a radially symmetric
heat mirror, defines what radiation gets in and must exclude rays from
θz,max < θz < 90°. All emitted radiation, apart from what the heat mirror
absorbs, gets out. The incoming radiation is now strongly attenuated, so
one expects net cooling gains and lower stagnation temperatures than in
the absence of heat mirror arrangements.
Which type of radiator benefits most from a heat mirror arrange-
ment: the broad-band absorber with high emittance or the sky window

© 2011 by Taylor and Francis Group, LLC

 Coolness 343

Enclosed
cone

Parabola
Low emittance
or CPC
surfaces

Radiative coating
Sky window selective
or high emittance

Figure 7.17  Structures with heat mirrors blocking some or all of the
atmospheric radiation emerging from large zenith angles.
selective emitter with intermediate emittance? In other words, does
the evaluation regarding the benefits of the two types of surfaces in
Figure 7.9 change when angular restrictions are included in the analysis?
It is easy to model Prad with mirrors, which we will do, but some reason-
ing can also tell what will happen: the net output from the blackbody
falls off most rapidly as τs goes down, as seen in Figure 7.9, because this
surface absorbs most of the incoming radiation. Thus, the blackbody
stands to benefit most from using heat mirrors. Calculations of Prad for
an aperture that is symmetric about the zenith direction can be made by
replacing Equation 7.3 with

 π /2 ∞ 

0
∫ 0
∫
Prad =  d (sin2 θz ) d λP(λ, τ s )Es (θz , λ)  −
 

 θz ,max ∞ 


 0
∫ ∫
d (sin θz ) d λP(λ, τ a )Ea (θz , λ)Es (θz , λ)  −
2

0


(7.5)

 π /2 ∞ 

 ∫
 θz ,max
∫
d (sin θz ) d λP(λ, τ a )Em (θz , λ)Es (θz , λ) 
2

0



© 2011 by Taylor and Francis Group, LLC

 344 Green Nanotechnology

where Em is the emittance of the heat mirror and taken to be wave-

length independent.
Calculations based on Equation 7.5 were performed in the same way
as for obtaining the data in Figure 7.9 and using the U.S. standard model
atmosphere. The acceptance angle θz,min was put to 45° and 90°, where
the latter value represents free exposure to the sky. Figure  7.18 shows
that the data for θz,min = 45° yield a very large enhancement of Prad over
what was found for free exposure [53]. The enhancement is particularly
large for the blackbody surface, whose net radiative power exceeds that
for the infrared selective surface for (τa − τs) being as large as 50°C.
Including a realistic value of Unr then leads to the conclusion that the
blackbody surface is consistently superior to the infrared selective sur-
face. For θz,min = 90°, on the other hand, the results are very similar to
those earlier found in Figure 7.9. Figure 7.18 also shows two triangles,
which indicate the possible performance of nonideal, though carefully
chosen, surfaces; the upper triangle refers to a blackbody-like surface
with an emittance of 0.9 and θz,min = 45°, and the lower triangle refers
to an infrared selective surface with an emittance of 0.1 in the 8 < λ <
300
Ideal 90°
Ideal 45°
250
Black body 90°
Net Radiative Power [Wm-2]

Black body 45°

200
Es = 0.95, 45°(top)
As(sky window) = 0.90,
150 As(outside sky window) = 0.1
oSX (90°) 45° (bottom)

100

oSX (45°)
50

0
0 10 20 30 40 50 60 70 80 90 100
oa–os [°C]

Figure 7.18  Net radiative cooling power for ideal blackbody and
infrared selective surfaces exposed to the U.S. standard atmosphere
under condition so that the acceptance angle is limited to 45° and for
free exposure to the sky. The triangles on the vertical axis indicate cool-
ing powers for surfaces with non-ideal properties as discussed in the
main text. (From G. B. Smith, Solar Energy Mater. Solar Cells 93 [2009]
1696–1701. With permission.)

© 2011 by Taylor and Francis Group, LLC

 Coolness 345

13 μm range and 0.90 outside this range along with θz,min = 45°. Cooling
curves emanating from these triangles will run almost in parallel with
the pertinent curves for ideal surfaces.
The main result of Figure 7.18 is the prediction that a device with
a blackbody-like emitter and with highly reflecting surfaces confining
the radiative exchange to an angular range around the zenith direction
can lead to a very large cooling power and to temperatures lying far
below those of the ambience. For example, with Unr = 2 Wm−2K−1 one can
remove heat at 20°C below ambient temperature at 60 to 70 Wm−2 .

7.4.8  Impact of Solar Irradiance on Sky Cooling

Sky cooling can function both night and day. Cooling during the night
is most effective, but there are many possibilities to combine this with
cooling also in the day (though at a lower rate than in the night), heating
in the day (or during winter), daylight transmission, and photovoltaic
energy generation. Thus, it is of interest to consider the impact of solar
irradiance on sky cooling.
Even if night cooling is the sole concern, it is worthwhile to limit
solar heating during the day. This is so because the entire cooling device
might heat up, and then there is a surplus of energy that takes time
to remove by sky cooling before subambient temperatures are reached.
There are a number of ways to limit the solar heating: The simplest
may be to totally close off the system, or to put it away if it is portable.
Secondly, the cooling surface itself may combine its desirable IR prop-
erties with high solar reflectance or transmittance. A third approach,
which may be the best for most practical cooling devices, is to have a
convection-suppressing shield that reflects or back scatters solar radia-
tion while it transmits in the thermal infrared. Microparticles of ZnS
can be effective for the latter purpose, as we will see below in the dis-
cussion about foils for water condensation. Another option is nanosized
TiO2 incorporated in polyethylene [54].
It is now becoming clear that useful net heat pumping is possible
during the day, provided that the impacts of daytime solar radiation
are managed. It is very difficult to achieve net cooling in the presence of
direct solar irradiation, though it is physically possible in principle with
materials that have very small solar absorption and very large radia-
tion output. One such material might be sheet glass in which the iron
content is almost zero (cf. Figure 4.2). To have a worthwhile net cooling
all day or most of the day in direct sun, a rough guide is that the solar
heating of the cooling surface should be less than 40 Wm−2 . This means
a solar absorptance less than ~8%, along with high emittance across
the sky window. These properties are possible but may be difficult to

© 2011 by Taylor and Francis Group, LLC

 346 Green Nanotechnology

accomplish in practice, and an easier approach is to combine a signifi-

cant sky view with shading of the direct solar irradiance during most
of the day. A recent study with façade-mounted radiators, facing away
from solar incidence at midday, demonstrated useful net sky cooling in
the day and almost as large cooling as for horizontal mounting in the
night [51]. Continuous cooling during night and day was demonstrated
also in some early work with heat mirrors, one of which also acted as a
shield of direct solar radiation [41,42].

7.5  Water Condensation Using Sky Cooling

Surfaces which can pump heat at subambient temperatures can be used

to condense water vapor from the atmosphere; applications of this type
may have a venerable history and go back to the ancient Greeks [55].
Condensation can be made from other sources, such as the condensation
part of a solar desalination cycle for extracting fresh water from salty or
brackish water. The condensation rate from air depends on the relative
humidity or dew point temperature τdp and the surface temperature τs of
the radiating surface relative to the ambient temperature τa [27,56]. Air
is saturated when τa = τdp. Net radiative cooling still occurs normally
down to and below the dew point, but below τdp some of the pumped
heat is taken up by the condensation process, thus limiting the minimum
value of τs. A steady-state energy balance is reached at some rate of con-
densation Rd (in kgs –1m –2) such that

Prad = Unr (τa − τs) + RdLw (7.6)

where Lw = 2.26 × 106 Jkg−1 is the latent heat of water vaporization, that
is, the heat that must be pumped away each second to condense 1 kg of
water. For an open surface, RdLw depends on the convective exchange
between the air away from the surface and that near the radiating sur-
face. This is governed by differences in vapor pressures, which are higher
at τa than at τs, plus the convective exchange part of Unr according to
relations that are well known in atmospheric science [57]. The perti-
nent equations must be solved iteratively together with Equation 7.6 to
determine self consistently both τs and Rd for the particular relationship
between Prad and τs for a given experimental set-up.
Figure 7.19 shows nightly collected dew in Dodoma, Tanzania, ver-
sus relative humidity [58,59]. Both data sets are based on a very simple
system comprised of a 0.4-mm-thick polyethylene foil containing 15
vol.% of ZnS pigment. The solar reflectance was ~85 %, and the emit-
tance in the sky window was high as a consequence of the ZnS [18]. The

© 2011 by Taylor and Francis Group, LLC

 Coolness 347

0.20
Clouds 0/8–1/8

Total Collected Water [liter/m2]

Clouds ≥2/8
0.15
Lushiku and Kivaisi

0.10

0.05

0.00
70 75 80 85 90 95
Relative Humidity at 06:00 [%]

Figure 7.19  Water condensed on a pigmented polymer foil as a function

of relative humidity overnight in Dodoma, Tanzania, during November
1993. Data refer to different fractional cloud covers. The line indicates
averaged data taken during earlier measurements. (From T. Nilsson,
Solar Energy Mater. Solar Cells 40 [1996] 23–32. With permission.)
collected water can reach ~0.1 liter per m 2 but is often much less. Clouds
are found to deteriorate the dew collection.
Dew collection is currently being implemented for drinking water in
hot and arid India [60] and is studied elsewhere too [61]. The technique is
referred to as “dew-rain” and typically uses pigmented foils similar to the
one used for the experiments shown in Figure 7.19. Figure 7.1 depicted a
unit in France capable of producing significant amounts of water. Even
simple galvanized iron roofs are capable of collecting some dew, given
the proper climatic conditions [62]. The preliminary nature of the dew
collection performed thus far should be emphasized, and several of the
sky cooling systems discussed above could lead to superior results.
The performance of simple foil collectors does not only depend on
the radiative properties of the materials but also on the rate at which
water runs off the collecting surface (i.e., on its hydrophobic or hydro-
philic character). A good approach might be to let the water condense
inside a channel in a polymer or under a cooling plate within a heat
mirror aperture system. If the vapor is forced by a fan or by natural
convection through a channel under a plate, or through channels in open
IR-transparent polymer systems where the polymer is doped with sur-
face plasmon resonant nanoparticles such as SiC, then there will be a
thermal gradient across the channel as the vapor progressively falls off.
RdLw will vary with location and flow speed.

© 2011 by Taylor and Francis Group, LLC

 348 Green Nanotechnology

Dew formation can be beneficial but it can also be a nuisance, par-

ticularly if it is accompanied by frost formation. Thus, modern fenes-
tration with excellent thermal insulation can make it possible for the
outer window pane to drop below τdp so that water is condensed on sur-
faces exposed to the clear sky, especially during early autumn mornings
[63,64]. A striking illustration of this effect is given below in Figure 8.8.
The water usually disappears shortly after sunrise, so the problem is
seldom a very serious one.
An analogous situation is well known for cars that are parked out-
doors during clear winter nights in cold climates: their windscreens are
covered with adherent frost in the morning [65]. The condensation is
effectively removed under most, though not all, weather conditions if the
glass surface has a transparent low-emittance coating that diminishes
the sky cooling. Pyrolytically produced SnO2:F, discussed in Section 4.4,
is an excellent coating for this purpose. Another way to make the water
condensation inconspicuous, as long as the water is liquid, is to have a
thin photocatalytic and super-hydrophilic coating of TiO2 on the outer
glass surface (cf. Section 8.2).

7.6  A Role for Cooling and Waste Heat

in Electric Power Generation

Today’s power stations place great thermal and water resource burdens
on the environment. Cooling is an essential aspect of thermally driven
generation of electricity, and cooling towers are stark visual reminders
of this need. If local lakes and rivers are employed for cooling, the con-
sequence is degradation of the environment and reduction of natural
water flows for irrigation or human use. It is a fundamental aspect of
thermodynamics, as outlined in Chapter 2, that for engines utilizing
heat input at temperature τh the useful work available to drive a turbine
and produce power depends on the temperature τc at which the waste
heat is exhausted. The maximum available power is given by the Carnot
efficiency [1 − τc /τh]. Practical, or irreversible, heat engines have more
realistically efficiencies of around [1 − (τc /τh)1/2] [66], with the rate at
which heat can be exhausted by cooling having an important influence
on this efficiency. Thus, for every GW of normal electric power with τc
lying near the ambient temperature, around 1.8 GW of heat has to be
exhausted, often as saturated air at around 35°C.
Many emerging concepts for large-scale, low-cost renewable power
exploit vast thermal resources with lower temperatures than those
for the usual power stations. When differences between τh and τc are
small—as for power generation in OTEC (cf. Section 2.10), salt gradient
solar ponds and solar chimneys—low efficiencies for conversion of heat

© 2011 by Taylor and Francis Group, LLC

 Coolness 349

to mechanical work must thus result. But if the natural energy resource
is large enough and free at both input and output, these systems never-
theless are very attractive. Furthermore, the source of coolness is differ-
ent from the local ambient, which is a significant aspect. Any thermal
power system operating at τh below ~170°C can benefit significantly in
efficiency by having τc fall from, say, τc > 25°C, which would be typi-
cal for cooling by air or water, to a temperature of 5°C to 10°C. Such a
decrease of τc is achievable, and a number of power systems with such
characteristics are now projected.
Waste heat from conventional power stations and industries, using
high-temperature processes, can also be tapped for additional power.
Power from waste heat requires turbines working at low temperatures,
so cooling can add significantly to the power output. A viable and con-
servative goal for the world could be up to 1,000 TWh of power per
annum from a fraction of such resources operating at 10 to 15% con-
version efficiency. Costs will depend on the specific type of source but
should undercut solar thermal power generation if sufficient efforts are
put into their design. Savings in CO2 emissions would be up to 1 Gt
per annum. Thus, designers of new large-scale power stations, found-
ries, kilns, and steel plants should considering incorporating waste heat
capture into plant design for subsequent use in power production. Such
features are easier and much cheaper if they are put it in at the outset
rather than as add-ons later. The associated carbon credits would also
help offset the carbon liabilities for normal operation and hence add to
the economic viability of such additions.
Cooling, if sufficiently cheap and simple, can also be a useful adjunct
for boosting the output from renewable power systems, especially ther-
mal systems operating at low temperatures. Solar thermal power gen-
eration with an input temperature τh in the range 80°C to 100°C could
have a boost in efficiency of 50% to 100% if the engine condensation
cycle is done with coolness collected via sky cooling. This idea, which
could provide very low-cost power, is considered among our specula-
tions in Chapter 9. And cooling may also be an asset for some electronic
power systems, whose efficiency depends on junction temperatures or
thermal gradients. Large-scale photovoltaic generation systems do not
produce electricity at night, but they are commonly located in near-per-
fect locations for night sky cooling under clear skies and in dry air. Such
generation systems would benefit from the additional cooling by use of
night-cooled fluids, which may be able to decrease the temperatures of
the solar cells by ~5°C or more over what is possible with regular air
cooling during daytime. Stirling cycle engines, used in small solar ther-
mal systems, could also have their efficiencies raised and their environ-
mental impact diminished by using stored coolness, but the gains are
smaller than for solar cells.

© 2011 by Taylor and Francis Group, LLC

 350 Green Nanotechnology

Interest in thermoelectric power is growing. It can be solar driven

and would be much improved with overnight-generated coolness. Some
developments in this area involve new nanotechnology, and it has the
possibility of becoming a key source of highly efficient refrigeration in
the future, so it is addressed in detail next.

7.7  Electronic Cooling and

Nanotechnology

Cooling as well as electric power generation can be achieved by using

the mobile electrons in two different conducting solids connected at
one end and arranged with a common temperature difference Δτ across
each. Thermocouples using two different metals work on this principle
as do thermoelectric devices using n- and p-doped semiconductors as
the two materials. Bismuth telluride has been the best semiconductor
so far and has been available for this kind of devices for some time.
Solid-state energy conversion of heat to electricity, and using elec-
tricity to pump heat, would be ideal, but it is not efficient enough at
present for replacing compressor-based cooling systems or to use in solar
thermal systems for power generation. However, the area of electronic
cooling is one where future developments utilizing known nanoscience
could have a major impact and reduce the problems associated with
leaking refrigerants, which are very bad greenhouse gases. A practical
breakthrough could revolutionize many aspects of how we use and gen-
erate energy and lead to vast reductions in CO2 emissions. Transferring
scientific concepts into commercial devices appears to be the major hur-
dle. Figure 7.20 shows a standard commercial device configuration for
thermoelectric cooling [67].
There are three factors that reduce efficiency when electrons are
driven by an external voltage to carry heat from a cold junction to a
hot junction for cooling applications: First, the electrons lose energy
by normal resistive losses as they traverse the solid and, second, heat
can flow back through the device from the hot side. The third factor
affecting efficiency is intrinsic to the materials, namely the Seebeck
coefficient S (i.e., the voltage per unit of temperature difference for
each arm in the thermoelectric device). These three factors are lumped
together into a performance factor Z, which should be as large as pos-
sible. It is defined by

Z = σS 2 /kτ (7.7)

where σ is electrical conductivity and kτ is thermal conductivity. A

widely used figure of merit is then obtained by multiplying Z with the

© 2011 by Taylor and Francis Group, LLC

 Coolness 351

Heat
absorbed n-type semiconductor pellets

(Cold side) p-type

Ceramic semiconductor
substrate pellets

N
P
N N
P P
N N
P P
Positive (+)
Conductor
Heat
rejected Negative (–)

(Hot side)

Figure 7.20  Standard configuration for commercial p-n junction ther-

moelectric cooling modules. The two white layers are ceramic. (From
L. E. Bell, Science 321 [2008] 1457–1461. With permission.)

average temperature (in K). This quantity is labeled Zτ and is about 1.0
in the best currently available commercial modules. This is much too
small to match conventional cooling methods, but the compactness of
the thermoelectric devices has led to a number of valuable niche mar-
kets. A Zτ value of ~2 would open up the markets for domestic heating
and cooling.
There are many variations on the simple system in Figure 7.20 that
work better or cost less, and nanostructures can be used to boost the sys-
tems [67]. Nanostructures such as quantum wells can in principle help by
raising S and reducing kτ , and various promising related developments have
occurred, but these ideas remain to be made commercially attractive.
Power generation needs higher values of Zτ than those available
with present materials, and the requirements for cooling are even more
stringent. Cooling also needs high values of Δτ. A Zτ of around 1.25 will
open opportunities for widespread use of waste heat and solar radiation
for power. In practice, Z depends on τ, and lamination of materials with
different values of Z may be the best option [68].
It is very difficult to keep the cold side of a thermoelectric generator
at air temperature if the hot side is at ~400°C, say, as is appropriate in
a system operating with concentrated solar radiation. Clearly, cold-side
warming will reduce the power output a lot, so the application of stored
coolness from overnight cooling should raise the efficiency.

© 2011 by Taylor and Francis Group, LLC

 352 Green Nanotechnology

Very high-efficiency thermoelectric conversion may be possible if

two conditions are fulfilled, that is,

• The electrons are transported ballistically from cold to hot

(i.e., without the collisions which lead to zigzag paths and
cause resistance.)
• The physical structure itself limits heat flow.

Ballistic transport is possible, in principle, by thermionic emission or

tunneling across a vacuum gap that is less than 10 nm in width, and this
can lead to power generation with close to Carnot efficiency, because
thermal transport from hot to cold is limited to radiation [69]. It is a
very difficult technical challenge to maintain such a gap uniformly over
useful areas in the presence of thermal expansion and mechanical stress,
but serious commercial efforts are being made. Arrays of nanopores in
insulators and stacks of layers may be the solution. Keeping the heat
transport down then becomes the key issue. Another important aspect is
the ease of getting electrons to leave a surface. The potential barrier they
must overcome—that is, the work function—can be reduced by surface
nanostructure [70].
Cooling has been the dominant focus to date for applications of ther-
moelectric devices, and power generation has been confined mainly to
difficult locations, including in deep space where the devices are driven
by the heat supplied by small radioactive sources. Thermoelectric gener-
ators have proven their durability also in other harsh environments, such
as at the earth’s poles. The enhancements in Zτ for making power gen-
eration competitive are not as large as those needed for general cooling
applications, so it is possible that solar thermal plants using thermoelec-
trics could be one of the first renewables to benefit from improvements in
this technology, while energy efficiency will benefit through waste heat
becoming a viable source of extra power. A solar thermal plant not need-
ing fluids or steam would be much easier and cheaper to run and main-
tain than current plants of this type, so solar thermal electricity could
drop sufficiently in price to undercut coal- and nuclear-fired plants—if
only the thermoelectric components were not too expensive. An expand-
ing use at present is in hybrid and normal cars to provide extra electric
power from their hot exhaust systems.
In summary, electronic cooling and electronic thermal power gen-
eration is a tantalizing field with regard to potential. It is growing in
impact, but its true potential will only be realized if advances in nano-
technology and nanoscience open up ways to practical mass production
of materials and systems that exhibit a significant jump in S, Zτ, and
overall efficiency relative to today’s (2010) values. The required materi-
als also must be sufficiently abundant.

© 2011 by Taylor and Francis Group, LLC

 Coolness 353

7.8  Whither Cooling?

7.8.1  Some Environmental and Health-Related Benefits

New cooling techniques can reduce and offset CO2 emissions, as dis-
cussed above, but there are other major environmental and health ben-
efits as well. Thus, they can have an impact on water supply and water
quality for agriculture and human consumption, and in this context one
should note the necessity to have fresh water in order to diminish diar-
rheal diseases and other severe health threats in the third world. Other
outcomes of the new cooling techniques can be to allow long-term stor-
age of food and to cut down emissions of potent greenhouse gases.
Water condensation directly via sky cooling was discussed above,
but there are numerous other water-related applications. For example,
there are many areas in the world where it is necessary to intercept salty
groundwater before it enters major rivers, which are needed for irriga-
tion and human use. This salty water is typically pumped into evapo-
ration basins so both water and salt are wasted. But fresh water can
be extracted cheaply using the sky cooling from the air just above and
around such basins, or even from select estuaries, because this air has
elevated humidity whereas the sky above can be clear. Crops can be
grown starting with salt water and using special greenhouses. Extraction
of useable bacteria from water can be aided by using low-cost sky cool-
ing together with doped polymers or coated metals.
Consider keeping food fresh. Field-based, low-cost sky cooling sys-
tems, operating both day and night, can provide the temperature drop
that is needed to preserve fruit and vegetables long enough to avoid spoil-
age. Cooling hen houses cheaply in warm countries, perhaps mainly in
the third world, can enable higher egg and chick yields. And there are no
doubt many other niche examples.
Today’s refrigeration technologies use compressed gases that are
worse greenhouse gases than CO2 , and these gases may end up in the
atmosphere unless the scrapping of old refrigeration systems is done with
care. If the projected growth in conventional air conditioning by 2030 was
to occur, then refrigerant gases may eventually pose an equal or a worse
threat than that from fossil fuels. Thus, widespread use of sky cooling
and/or electronic cooling, along with conscious building design to reduce
cooling loads, should be a matter of urgency if this risk is to be averted.

7.8.2  Cooling Plus

Cooling technologies—whether based on electronic advances, radiation

to the night sky, or hybrids of the latter with existing methods—could

© 2011 by Taylor and Francis Group, LLC

 354 Green Nanotechnology

mean that we are at the threshold of a new technical era in cooling

and refrigeration, but much more research and development is clearly
needed. And this is a matter of urgency.
In a BAU scenario, the cooling demand is set to soon escalate into
a dominant problem for power supply and for the environment. And
these projections may be underestimates, as proceeding global warming
will lead to increasing demands on cooling. Seen in this perspective, it
is fortunate that there is an untapped and vast natural low-cost cool-
ing resource—the clear night sky—overhead. But sky cooling must be
combined with other measures in order to meet the growing cooling
demand with less environmental impact than at present: these measures
include energy-efficient building structures with much less pumping of
heat into local air, cooler urban microclimates from raised albedos via
new roof coatings and new road surfaces, and more urban “greening.”
Implementing these measures will make the urbanscapes—indoors and
outdoors—much more pleasant places for the billions who are destined
to reside in them in the years ahead.
There are other aspects of cooling that we have barely touched
because studies have yet to begin. In particular, spectrally selective
transmittance may enable cooling in the daytime jointly with various
uses of the transmitted solar energy, as noted earlier. The adjunct ability
to generate electric power in new low-cost ways with the aid of sky cool-
ing could ultimately also have a major influence. We might even go as
far as to say that it is not unreasonable to imagine a world where clean
power sources, using some combination of solar energy and sky cooling,
become the backbone of a low-carbon economy. The prospect then is not
only less pollution but, in due course, lower power cost. Further com-
ments on this issue follow in Chapter 9.

References

1. M. Bojić, F. Yik, Cooling energy evaluation for high-rise residential

buildings in Hong Kong, Energy Buildings 37 (2005) 345–351.
2. K. W. J. Barnham, M. Mazzer, B. Clive, Resolving the energy crisis:
Nuclear or photovoltaics? Nature Mater. 5 (2006) 161–164.
3. A. Jäger-Waldau, editor, REF-SYST Status Report 2004, EUR 21297
EN, Joint Research Center, Ispra, Italy (2004).
4. D. Brockett, D. Fridley, J.-M. Lin, J. Jin, A tale of five cities: The
China residential energy consumption survey, in Human and Social
Dimensions of Energy Use: Understanding Markets and Demands,
ACEEE Summer Study on Building Energy Efficiency, 2002, pp.
8.29–8.40.
5. www.opur.fr.

© 2011 by Taylor and Francis Group, LLC

 Coolness 355

6. H. Akbari, S. Menon, A. Rosenfeld, Global cooling: Increasing

world-wide urban albedos to offset CO2, Climatic Change 94 (2009)
275–286.
7. G. B. Smith, M. B. Riley, Solar absorption cycle cooling installa-
tions in Sydney, Australia, in Proceedings of the International Solar
Energy Society Silver Jubilee Congress, edited by K. W. Böer, B. H.
Glenn, Pergamon, Elmsford, NY, 1979, Vol. 1, pp. 729–733.
8. J. C. Hedstrom, H. S. Murray, J. D. Balcomb, Solar heating and cool-
ing results for the Los Alamos study center, in Proceedings of the
Conference on Solar Heating and Cooling Systems Operational
Results, Colorado Springs, CO, November 28–December 1, 1978;
Los Alamos Scientific Laboratories LA-UR-78-2588, 1978, pp. 1–7.
9. http://www.un.org/esa/population/publications/sixbillion/sixbil-
part1.pdf.
10. http://www.urbanheatislands.com.
11. M. Kolokotroni, Y. Zhang, R. Watkins, The London heat island and
building cooling design, Solar Energy 81 (2007) 102–110.
12. H. Taha, Episodic performance and sensitivity of the urbanized
MM5 (uMM5) to perturbations is surface properties in Houston
Texas, Boundary-Layer Meterol. 127 (2008) 193–218.
13. H. Taha, Urban surface modification as a potential ozone air-quality
improvement strategy in California: A mesoscale modeling study,
Boundary-Layer Meteorol. 127 (2008) 219–239.
14. Vattenfall AB (2007), Global Mapping of Greenhouse Gas
Abatement Opportunities up to 2030: Building Sector Deep-Dive,
June 2007, www.vattenfall.com/www/ccc/ccc/Gemeinsame_Inhalte/
DOCUMENT/567263vattenfall/P0272861.pdf.
15. A. Addeo, E. Monza, M. Peraldo, B. Bartoli, B. Coluzzi, V. Silvestrini,
G. Troise, Selective covers for natural cooling devices, Nuovo
Cimento C 1 (1978) 419–29.
16. A. Addeo, L. Nicolais, G. Romeo, B. Bartoli, B. Coluzzi, V. Silvestrini,
Light selective structures for large scale natural air conditioning,
Solar Energy 24 (1980) 93–98.
17. T. M. J. Nilsson, G. A. Niklasson, Optimization of optical properties
of pigmented foils for radiative cooling applications: Model calcula-
tions, Proc. Soc. Photo-Opt. Instrum. Engr. 1536 (1991) 169–182.
18. T. M. J. Nilsson, G. A. Niklasson, Radiative cooling during the day:
Simulations and experiments on pigmented polyethylene cover foils,
Solar Energy Mater. Solar Cells 37 (1995) 93–118.
19. J. Bell, R. Lehman, G. Smith, Advanced roof coatings: Materials
and their applications, in Proceedings CD and Summary Book of
the CIB International Conference on Smart and Sustainable Built
Environment, Brisbane, Australia, November 19–21, 2003 (ISBN:
1-74107-040-6), Paper T606.

© 2011 by Taylor and Francis Group, LLC

 356 Green Nanotechnology

20. SkyCool Pty Ltd, Sydney, Australia (2007), private communication.

21. B. Schuler, Innovation for construction and buildings by
BASF, in Proceedings of Convation06, Sydney, Australia, 2006
(unpublished).
22. A. Synnefa, M. Santamouris, K. Apostolakis, On the development,
optical properties and thermal performance of cool colored coatings
for the urban environment, Solar Energy 81 (2007) 488–497.
23. G. B. Smith, A. Gentle, P. D. Swift, A. Earp, N. Mronga, Coloured
paints based on coated flakes of metal as the pigment, for enhanced
solar reflectance and cooler interiors: Description and theory, Solar
Energy Mater. Solar Cells 79 (2003) 163–177.
24. G. B. Smith, A. Gentle, P. Swift, A. Earp, N. Mronga, Coloured paints
based on iron oxide and silicon oxide coated flakes of aluminium as
the pigment, for energy efficient paint: Optical and thermal experi-
ments, Solar Energy Mater. Solar Cells 79 (2003) 179–197.
25. M. R. Kuehnle, H. Statz, Encapsulated nanoparticles for the
absorption of electromagnetic energy, US patent publication US
2005/0074611 A1 (2005).
26. C. G. Granqvist, T. S. Eriksson, Materials for radiative cooling to
low temperatures, in Materials Science for Solar Energy Conversion
Systems, edited by C. G. Granqvist, Pergamon, Oxford, U.K., 1991,
pp. 168–203.
27. T. M. J. Nilsson, Optical Scattering Properties of Pigmented
Foils for Radiative Cooling and Water Condensation: Theory
and Experiment, Ph.D. thesis, Department of Physics, Chalmers
University of Technology, Göteborg, Sweden (1994).
28. X. Berger, J. Bathiebo, Directional spectral emissivities of clear skies,
Renewable Energy 28 (2003) 1925–1933.
29. P. Berdahl, R. Fromberg, The thermal radiance of clear skies, Solar
Energy 29 (1982) 299–314.
30. P. Berdahl, M. Martin, The resource for radiative cooling, in
Proceedings of the Second National Passive Solar Conference,
edited by D. Prowler, I. Duncan, B. Bennett, American Section of the
International Solar Energy Society, Newark, DE, 1978, Vol. 2, pp.
684–686.
31. C. G. Granqvist, A. Hjortsberg, Radiative cooling to low tempera-
tures: General considerations and application to selectively emitting
SiO films, J. Appl. Phys. 52 (1981) 4205–4220.
32. T. S. Eriksson, C. G. Granqvist, Radiative cooling computed for
model atmospheres, Appl. Opt. 21 (1982) 4381–4388.
33. M. Tazawa, H. Kakiuchida, G. Xu, P. Jin, H. Arwin, Optical constants
of vacuum evaporated SiO film and an application, J. Electroceram.
16 (2006) 511–515.

© 2011 by Taylor and Francis Group, LLC

 Coolness 357

34. Z. Liang, H. Shen, J. Li, N. Xu, Microstructure and optical proper-

ties of silicon nitride thin films as radiative cooling materials, Solar
Energy 72 (2002) 505–510.
35. T. S. Eriksson, C. G. Granqvist, Infrared optical properties of elec-
tron-beam evaporated silicon oxynitride films, Appl. Opt. 22 (1983)
3204–3206.
36. T. S. Eriksson, C. G. Granqvist, Infrared optical properties of silicon
oxynitride films: Experimental data and theoretical interpretation,
J. Appl. Phys. 60 (1986) 2081–2091.
37. P. T. Tsilingiris, The total infrared transmittance of polymerized
vinyl fluoride films for a wide range of radiant source temperature,
Renewable Energy 28 (2003) 887–900.
38. P. Berdahl, Radiative cooling with MgO and/or LiF layers, Appl.
Opt. 23 (1984) 370–372.
39. E. M. Lushiku, A. Hjortsberg, C. G. Granqvist, Radiative cooling
with selectively infrared-emitting ammonia gas, J. Appl. Phys. 53
(1982) 5526–5530.
40. E. M. Lushiku, C. G. Granqvist, Radiative cooling with selectively
infrared-emitting gases, Appl. Opt. 23 (1984) 1835–1843.
41. F. Trombe, A. Lè Phat Vinh, M. Lè Phat Vinh, Description des expéri-
ences sur le refroidissement des corps terrestres, J. Rech. CNRS 65
(1964) 563–580.
42. F. Trombe, Perspectives sur l’utilization des rayonnements solaires
et terrestres dans certaines régions du monde, Rev. Gén. Therm. 6
(1967) 1285–1314.
43. T. S. Eriksson, A. Hjortsberg, C. G. Granqvist, Solar absorptance
and thermal emittance of Al2O3 films on Al: A theoretical assess-
ment, Solar Energy Mater. 6 (1982) 191–199.
44. A. R. Gentle, G. B. Smith, Angular selectivity: Impact on optimised
coatings for night sky radiative cooling, Proc. SPIE 7404 (2009)
74040J 1–8.
45. A. R. Gentle, G. B. Smith, Radiative heat pumping from the earth
using surface phonon resonant nanoparticles, Nano Lett. 10 (2010)
373–379.
46. C. F. Bohren, D. R. Huffman, Absorption and Scattering of Light by
Small Particles, John Wiley & Sons, New York, 1983.
47. M. Dkaki, L. Calcagno, A. M. Makthari, V. Raineri, Infrared spec-
troscopy and transmission electron microscopy of polycrystalline
silicon carbide, Mater. Sci. Semicond. Proc. 4 (2001) 201–204.
48. T. Henning, H. Mutschke, Formation and spectroscopy of carbides,
Spectrochim. Acta, Part A 57 (2001) 815–824.
49. E. D. Palik, Handbook of Optical Constants of Solids, Academic,
Orlando, FL, 1985.

© 2011 by Taylor and Francis Group, LLC

 358 Green Nanotechnology

50. T. S. Eriksson, Surface Coatings for Radiative Cooling to Low

Temperatures, Ph.D. thesis, Department of Physics, Chalmers
University of Technology, Göteborg, Sweden, 1985.
51. A. R. Gentle, G. B. Smith (2010) unpublished data.
52. N. A. Nilsson, T. S. Eriksson, C. G. Granqvist, Infrared-transparent
convection shields for radiative cooling: Initial results on corrugated
polyethylene foils, Solar Energy Mater. 12 (1985) 327–333.
53. G. B. Smith, Amplified radiative cooling via optimised combina-
tions of aperture geometry and spectral emittance profiles of sur-
faces and the atmosphere, Solar Energy Mater. Solar Cells 93 (2009)
1696–1701.
54. Y. Mastai, Y. Diamant, S. T. Aruna, A. Zaban, TiO2 nanocrystal-
line pigmented polyethylene foils for radiative cooling applications:
Synthesis and characterization, Langmuir 17 (2001) 7118–7123.
55. V. S. Nikolayev, D. Beysens, A. Gioda, I. Milimouk, E. Katiushin, J.-P.
Morel, Water recovery from dew, J. Hydrology 182 (1996) 19–35.
56. D. Beysens, The formation of dew, Atmospheric Res. 39 (1995)
215–237.
57. J. L. Monteith, M. H. Unsworth, Principles of Environmental
Physics, 3rd edition, Academic, Burlington, MA, 2008.
58. E. M. Lushiku, R. T. Kivaisi, Applications of radiative cooling for
condensation irrigation, Proc. Soc. Photo-Opt. Instrum. Engr. 1149
(1989) 111–113.
59. T. Nilsson, Initial experiments on dew collection in Sweden and
Tanzania, Solar Energy Mater. Solar Cells 40 (1996) 23–32.
60. G. Sharan, H. Prakash, Dew condensation on greenhouse roof at
Kothara (Kutch), J. Agriculture Engr. 40 (4) (2003) 75–76.
61. M. Muselli, D. Beysens, J. Marcillat, I. Milimouk, T. Nilsson, A.
Louche, Dew water collector for potable water in Ajaccio (Corsica
Island, France), Atmospheric Res. 64 (2002) 297–312.
62. G. Sharan, D. Beysens, I. Milimouk-Melnytchouk, A study of dew
water yields on galvanized iron roofs in Kothara (north-west India),
J. Arid Environments 69 (2007) 256–269.
63. C. G. Granqvist, Transparent conductors as solar energy materi-
als: A panoramic review, Solar Energy Mater. Solar Cells 91 (2007)
1529–1598.
64. A. Werner, A. Roos, Simulations of coatings to avoid external conden-
sation on low U-value windows, Opt. Mater. 30 (2008) 968–978.
65. I. Hamberg, J. S. E. M. Svensson, T. S. Eriksson, C. G. Granqvist, P.
Arrenius, F. Norin, Radiative cooling and frost formation on sur-
faces with different thermal emittance: Theoretical analysis and
practical experience, Appl. Opt. 26 (1987) 2131–2136.
66. F. L. Curzon, B. Ahlborn, Efficiency of a Carnot engine at maximum
power output, Am. J. Phys. 43 (1975) 22–24.

© 2011 by Taylor and Francis Group, LLC

 Coolness 359

67. L. E. Bell, Cooling, heating, generating power, and recovering waste

heat with thermoelectric systems, Science 321 (2008) 1457–1461.
68. G. J. Snyder, Application of the compatibility factor to the design
of segmented and cascaded thermoelectric generators, Appl. Phys.
Lett. 84 (2004) 2436–2438.
69. Y. Hishinuma, T. H. Geballe, B. Y. Moyzhes, T. W. Kenny,
Refrigeration by combined tunneling and thermionic emission in
vacuum: Use of nanometer scale design, Appl. Phys. Lett. 78 (2001)
2572–2574.
70. A. Tavkhelidze, V. Svanidze, I. Noselidze, Fermi gas energetics in
low-dimensional metals of special geometry, J. Vac. Sci. Technol. B
25 (2007) 1270–1275.

© 2011 by Taylor and Francis Group, LLC

 Chapter 8
Supporting
Nanotechnologies
Air Sensing and Cleaning,
Thermal Insulation, and
Electrical Storage

This chapter brings together a range of important issues that impact on

energy use in general and, in particular, on the quality of spaces within
buildings. Specifically, we discuss sensing and purification of indoor air,
self-cleaning surfaces, thermal insulation, and storage of electricity. These
are huge topics in their own right, and our treatment of each of them
will take the form of an introduction, overview of key aspects, and some
recent progress. The reader who is interested in more detail could start
with the references provided. Figure 8.1 shows one of the materials we
will discuss below [1]. It is an aerogel comprised of silica nanoparticles
and has a porosity that can be as high as 99% or even above. This makes
silica aerogel the lightest solid material known today, and its density can
be only about three times that of air. Such “solid gases” are interesting for
transparent thermal insulation, as further discussed in Section 8.3.

8.1  Air Quality and Air Sensing

8.1.1  The Sick Building Syndrome

People in industrialized countries typically spend as much as 80 to 90%

of their time indoors, in buildings and vehicles [2]. Hence, it is obvious
that the indoor air quality is important for health and well-being and
for work productivity [3,4]. But the air quality experienced is often not
as good, and the “sick building syndrome” (SBS) [5] is a frequently used
term to describe situations in which the users of buildings experience

361

© 2011 by Taylor and Francis Group, LLC

 362 Green Nanotechnology

Figure 8.1  A color version of this figure follows page 200. Silica aero-
gel tile. (From http://www.airglass.se. With permission.)
acute health effects and discomfort seemingly connected with the time
spent in the building while no specific illness or cause can be pinned
down. As much as 30% of all new and refurbished buildings worldwide
may be the subject of complaints connected with indoor air quality. The
SBS is usually nonspecific with symptoms such as headache, eye, nose
or throat irritation, dry cough, itchy or dry skin, dizziness and nausea,
difficulty in concentrating, fatigue, and sensitivity to odors. The cause of
these symptoms is normally not known, but the syndrome is no longer
felt by persons who have left the “sick” building.
There are three main causes of the SBS:

• Inadequate ventilation can cause the SBS, which implies that

efforts to improve energy efficiency by cutting down on ventila-
tion may backfire and lead to health problems.
• Chemical contaminants are another cause and can originate
from indoor sources such as adhesives, carpeting and uphol-
stery, and manufactured wood products. All of these sources
may emit volatile organic compounds (VOCs), such as form-
aldehyde (HCHO), some of which are known carcinogens.
Tobacco smoke should be noted as it produces high levels of
VOCs, other toxic compounds, and respirable particulate mat-
ter. Inadequately vented heaters, stoves, and fireplaces can lead
to harmful combustion products such as carbon monoxide,
nitrogen dioxide, and respirable particles.
• Biological contaminants are the third major cause of SBS and
include bacteria, moulds, pollen, and viruses, which contribute

© 2011 by Taylor and Francis Group, LLC

 Supporting Nanotechnologies 363

microbial volatile organic compounds (MVOCs) and emanate

from stagnant water in ducts, drains, and humidifiers, and
through water ingress by way of leaks. The indoor bacterium
Legionella is a particularly well-known problem.

These three causes may act in combination and may supplement other
perceived problems such as inadequate lighting, temperature, or humid-
ity, or excessive noise. Radon and asbestos are not included in the causes
of the SBS. There may be a gender aspect in SBS, and women tend to be
more vulnerable than men.
There are many VOCs—such as decane, toluene, xylene, formal-
dehyde, tetrachlororethylene, ozone, and nitrogen dioxide—each with
specific threshold levels for causing problems [6]. Formaldehyde can be
particularly problematical, especially in new buildings [7], and has a
human sensory irritation threshold of 0.1 mg/m3 and an odor threshold
of 1 mg/m3. Inhaled formaldehyde may affect learning and memory, as
found from animal studies [8]. The concentrations of the VOCs are often
higher indoors than outdoors [9], and the same is true for MVOCs [10].
It is clearly important to be able to measure and survey the air quality
inside buildings. This requires good sensors, which must be low cost and
energy efficient in order to be of practical interest. It will be shown in the
following text that nanostructured oxides can be used for this purpose.
Removal of VOCs and MVOCs is also of great relevance and, as discussed
in Section 8.2, this can be done by photocatalysis using, again, nanostruc-
tured oxides. Interestingly, the very same transition metal oxides are rel-
evant for sensing and purification of air, and these oxides are of central
interest also for electrochromic window coatings (cf. Section 4.11) [11] and
solar reflective high emittance paints (cf. Section 7.3). Gas sensors enable
demand controlled ventilation, and photocatalytic air cleaning makes it
possible to keep the air quality at a desired level without excessive and
energy-consuming air exchange. The energy efficiency will be particularly
good if the photocatalysis is driven by direct solar irradiation.

8.1.2  Gas Sensing with Nanoporous Metal Oxides: General

It has long been known that semiconducting metal oxides can be used for
gas sensing. The underlying mechanism is chemical interaction of the gas
molecules with the semiconductor’s surface leading to changes in its elec-
trical conductivity [12–15]. Molecules in the gas phase, which can serve
as electron donors or acceptors, adsorb on the oxide and form surface
states, which then can exchange electrons with the oxide. If the molecule
is an acceptor it will extract electrons from the oxide and hence decrease
its conductivity (in the case of an n-type conductor); if the molecule is

© 2011 by Taylor and Francis Group, LLC

 364 Green Nanotechnology

a donor it will increase the conductivity. A space charge layer is thus

formed with a conductivity that is a function of the concentration and
type of the gas, and this chemically induced effect can be transformed
into an electrical signal via an appropriate electrode arrangement.
The influence of the gas absorption will be particularly strong if the
oxide is nanoporous and comprises nanoparticles that are weakly linked
to one another. Such a structure is shown in the upper part of Figure 8.2.
Also shown there is a simplified model of the electronic bands, illustrating

Contact Metal oxide

electrode Depletion
grain
layer

(a)
E

eVig eVge
Ecb
EF

(b)

Free
Adsorbed molecules
molecules
Catalyst

Electrodes
(c)
e– e– Metal oxide

Substrate C. Contacts
C S G C S. Surface
B B. Bulk
G. Grain
boundaries
(d)

Figure 8.2  Physical structure of a gas sensor with oxide grains between
electrodes (a) and a one-dimensional electronic structure showing inter-
grain and grain-electrode band bending (by eVig and eVge, respectively),
as well as the lower limit of the conduction band of the oxide (Ecb) and
its Fermi level (E F) (b). Also shown are processes upon gas adsorption in
the presence and absence of a catalyst with e – denoting electrons (c) and
an equivalent circuit of the various contributions to the overall conduc-
tivity (d). (After M. Graf et al., J. Nanoparticle Res. 8 [2006] 823–839.
With permission.)

© 2011 by Taylor and Francis Group, LLC

 Supporting Nanotechnologies 365

that inter-grain band bending leads to energy barriers whose height can
be influenced by gas adsorption. The lower part of Figure 8.2 shows how
adsorbed molecules provide charge to the oxide. This process is usually
not very species selective, but it can be made so if the metal oxide is
doped with a suitable catalyst, as also shown in Figure 8.2. The bottom
part of the figure contains a simple equivalent circuit of the sensor with
separate contributions due to the surface, bulk, and grain boundaries of
the nanoparticles and of the electrodes.
A good gas sensor is characterized by four main features, namely

• Sensitivity, so that the conductivity is strongly changed upon

gas exposure
• Selectivity, meaning that the sensitivity is large only for the
desired gas or gases
• Operating temperature, which typically is 200°C to 400°C and
requires a heating device that in practice should be small (known
as a “micro-hotplate”)
• Long-term stability

These features have a number of consequences for the implementation of

gas sensors in buildings. Of course, the sensors must be able to reliably
and unambiguously detect particularly relevant gases such as formalde-
hyde. Furthermore, it is desirable to allow operation at room tempera-
ture in order to avoid excessive energy consumption during long-term
operation, and a fire hazard.

8.1.3 Gas Sensing with Nanoporous Metal

Oxides: Illustrative Examples

Highly sensitive detection of gases is possible by use of a variety of

oxides, with SnO2 being a particularly well-known example [16]. WO3,
TiO2 , In 2O3, Ga2O3, SrTiO3, and many other oxides have been exten-
sively investigated as well. The films should be nanoporous in order
to allow a large contact area to the gas; such structures can be made
with a variety of the thin film techniques as discussed in Appendix 1.
Chemically prepared nanoparticulate layers [17], “nanosponges” [18],
and hybrid layers consisting of oxide and carbon nanotubes [19] are
some recently studied options.
Box 8.1 discusses gas sensing with WO3 -based layers composed of
well-characterized nanoparticles with and without the addition of cata-
lyst nanoparticles. The data show that high sensitivity can be combined
with selectivity and room-temperature operation.

© 2011 by Taylor and Francis Group, LLC

 366 Green Nanotechnology

Box 8.1  Gas Sensing with Nanoparticulate

WO3-Based Thin Films
Figure B8.1.1 shows the structure of two nanoparticulate WO3 -
based layers made by the advanced gas deposition (AGD) technique
introduced in Appendix 1 [20]. These films are somewhat substoi-
chiometric and contain particles with a tetragonal structure. The
images indicate agglomerations of nanoparticles with narrow size
distributions. The WO3 particles in Figure B8.1.1a are shown in
as-deposited state, whereas those in Figure B8.1.1b—referring to a
WO3:Pd mixed-particle film with a Pd:W atomic ratio of 0.5%—
have been sintered at 600°C in air for 1 hour. Clearly, the sinter-
ing has resulted in grain growth with preserved nanoporosity. The
AGD technique is unusual in that it separates nanoparticle nucle-
ation and growth from film growth—with both processes occur-
ring in a clean ambience—which enables the interactions between
neighboring particles to be fine tuned.

(a) (b)

100 100

Figure B8.1.1  Scanning electron micrographs of an as-depos-

ited film of WO3 (a) and of a WO3:Pd film sintered at 600°C (b).
The horizontal bars are 100 nm. (From A. Hoel et al., Sensors
Actuators B 105 [2005] 283–289. With permission.)

Figure B8.1.2 shows changes in the conductivity when unheated

films such as those in Figure B8.1.1 were exposed to different
amounts of H 2S gas in air. These changes are expressed in terms of
a sensitivity Σ defined by Σ = σgas/σair, where σgas is the electrical con-
ductivity of the sensor exposed to the pertinent gas and σair refers
to pure, dry, synthetic air, respectively. The H 2S was introduced
after 10 minutes and led to a gradual increase of the conductivity
in proportion to the H 2S content, as shown in Figure B8.1.2b for
the WO3:Pd film. After 10 minutes of exposure, the films were sub-
jected to a 10-minutes-long heating pulse at 200°C, which restored

© 2011 by Taylor and Francis Group, LLC

 Supporting Nanotechnologies 367

the original sensitivity. The sensitivity, recorded after 10 minutes

of gas exposure, depends strongly on the H 2S concentration, but
readily measurable effects can be seen at the 500 ppb level, and
Figure B1.2b indicates that Σ ≈ 10 then applies to the WO3:Pd film.
For the undoped WO3 film, the straight line indicates that the data
were consistent with Σ ≈ A C gasB , where C gas is the gas concentra-
tion and A and B are constants.

104
104 0.5 ppm WO3
1 ppm
2 ppm WO3: Pd
4 ppm
103
103 6 ppm
8 ppm
Sensitivity

Sensitivity
10 ppm
Heating 102
102 pulse 10 min

101 H2S in H2S out 101

100 100
0 5 10 15 20 25 30 1 10
Time (min) H2S Concentration (ppm)
(a) (b)

Figure B8.1.2  Room-temperature sensitivity versus time of a

nanoporous WO3:Pd film exposed to different concentrations of H 2S
in air (a), and room-temperature sensitivity versus H 2S concentra-
tion for sensors comprised of WO3 and WO3:Pd (b). The latter data
for WO3 fall on a straight line in the log–log plot. (From A. Hoel et
al., Sensors Actuators B 105 [2005] 283–289. With permission.)

Gas specificity is of obvious interest. This can be accomplished

in a number of different ways [21], the most straightforward being
by use of an array of sensors with each element optimized for a spe-
cific gaseous species. This is only one option, though. Nanoporous
oxides are usually sensitive to a range of different gases and, for
example, WO3 -based films can detect NO2 and CO for the former
gas down to ~1 ppm [22]. The sensitivity is peaked at different
operating temperatures for the various gases, and Figure B8.1.3
shows large differences for H 2S, NO2 , and CO for the case of
WO3:Al. If one allows sensor operation at an elevated temperature,
then it is clearly possible to use this effect to accomplish specificity.
Temperature-modulated resistance together with statistical analy-
sis using pattern recognition algorithms make it possible to reliably
detect as little as 20 ppb of H 2S and 200 ppb of ethanol [23,24].

© 2011 by Taylor and Francis Group, LLC

 368 Green Nanotechnology

5000 120
H2 S 100
4000 NO2
CO 80

Sensitivity
3000
60
2000
40
1000 20

0 0
300 400 500 600 700 800
Temperature (K)

Figure B8.1.3  Sensitivity versus operating temperature of a

WO3:Al film exposed to 10 ppm of H 2S (left-hand scale), 5 ppm of
NO2 and 100 ppm of CO (right-hand scale). (From L. F. Reyes et
al., Sensors Actuators B 117 [2006] 128–134. With permission.)

Electrical conductivity, as discussed in Box 8.1, is one of the many

properties that can be used for sensing purposes. Electromagnetic fluc-
tuation spectroscopy (“noise”) is another alternative that has attracted
attention recently [25]. This technique is interesting for buildings since it
can detect several gases simultaneously with a single sensor and since it
is not vulnerable to long-term drift of the dc electrical properties; on the
negative side, the noise detection requires rather advanced electronics.
Indoor air often contains formaldehyde, which is one of the prime
contaminants associated with the SBS, as pointed out above. Detecting
this gas is hence an important issue, but detector development is still in
its infancy. Sputter deposition of NiO (a p-type material) under condi-
tions giving submicrometer grains has been demonstrated to give films
that respond to formaldehyde exposure at levels of a few ppm when the
film is kept at 300°C [26]. Specifically, a change in the gas content by 5
ppm altered the resistance by ~1.5%. Recent work on NiO nanoparticle
films made by advanced gas deposition (cf. Appendix 1) showed good
sensor properties for H 2S and NO2 , particularly when the operating
temperatures were ~150 and ~100°C, respectively [27].

8.2  Photocatalysis for Cleaning

8.2.1  General

Photoexcitation can occur when semiconductor nanoparticles are exposed

to light with energy above the band gap energy. Figure 8.3 shows what

© 2011 by Taylor and Francis Group, LLC

 Supporting Nanotechnologies 369

– CB
Surface hν
recombination –++ hν

A + VB
– +
D
D+
B
C +
Volume
recombination
A– –++ D
–

Figure 8.3  Photoexcitation in a semiconductor particle by a photon

with energy hν lifts an electron from the valence band (VB) to the con-
duction band (CB). The electron-hole pair can then undergo several dif-
ferent processes. (From A. L. Linsebigler et al., Chem. Rev. 95 [1995]
735–758. With permission.)

goes on [28]: Photon absorption produces an exciton followed by charge

separation so that an electron-hole pair is created. Charge transport to
the particle surface along the paths C and D then leads to reduction
and oxidation processes at the surface, respectively. Other processes,
denoted A and B, end up with surface recombination and bulk recombi-
nation. The processes clearly show some resemblance to those in the dye
sensitized and organic solar cells discussed in Section 6.2.
The detailed physical and chemical processes occurring at the sur-
faces of the nanoparticles are surprisingly poorly understood. A discus-
sion of them would take us deeply into surface science even for the most
widely investigated photocatalyst, which is TiO2 [29], and this is beyond
the scope of this book. However, there is now convincing evidence of an
intimate coupling between nanoparticle morphology, exposed crystal-
line phases, and particle reactivity [30]. TiO2 has a number of crystalline
structures, with the anatase and rutile being the most common, and each
of these can form nanoparticles of different shapes. Figure 8.4 depicts
five different shapes of TiO2 nanoparticles present in samples made by
solution-based chemical preparation. The particle sizes are in the 10 to
70 nm range. Each of the faces corresponds to a specific crystallographic
index, as shown.
Figure 8.5 shows the normalized surface concentration as a function
of time for HCOOH and HCOO molecules absorbed on five different

© 2011 by Taylor and Francis Group, LLC

 370 Green Nanotechnology

AI AII AIII RI RII

(001) (001)
(101)
(001)
(101)

(100)
(112)
(011)
(101) (011) (110)

(100)
(101) (101)
(100)
(100) (111) (110)

Anatase Rutile

Figure 8.4  Compilation of most probable shapes of TiO2 nanopar-

ticles with anatase (A) and rutile (R) structures. The crystallographic
face indices are indicated.

P25
Normalized Surface Concentration

0.8
RII

0.6

AI, AII
0.4

0.2
AI
AIII
0
0 20 40 60 80 100 120 140
Time (min)

Figure 8.5  Normalized surface concentration of adsorbed HCOOH

and HCOO molecules on anatase (A) and rutile (R) nanocrystals versus
time for irradiation by simulated sunlight. Data are shown also for a
commercial photocatalyst (P25).

© 2011 by Taylor and Francis Group, LLC

 Supporting Nanotechnologies 371

samples comprised of TiO2 nanocrystals under simulated solar irradia-

tion. The data were obtained from in situ infrared spectroscopy [30]. The
samples are dominated by the nanoparticle shapes shown in Figure 8.4,
and results are given also for a material designated P25 which is used
extensively as a base line for work on photochemical and photoelectric
effects of TiO2-based materials (cf. Figure 6.21). The important conclu-
sion to be drawn from Figure  8.5 is that photocatalytic reaction rates
depend strongly on the type of nanoparticle that dominates the mate-
rial. Hence, one can boost the photocatalysis by controlling the sur-
face structure on the nanometer scale. The data in Figure  8.5 should
be regarded as indicative for a practical situation, and the presence of
water vapor, the temperature, and other parameters are important for
the performance of the photocatalyst. Another important aspect is that
residuals can build up on the surface during the photocatalysis and cause
its deactivation.
Photocatalysis is well known in oxides based on Cd, Ce, Fe, Sb,
Sn, Ti, W, Zn, and Zr. The most widely studied material is TiO2 with
anatase-type crystal structure [28,31,32]. Thin films of this material are
absorbing in the UV, whereas they are nonabsorbing in the luminous
and NIR ranges. Thus, only a few percent of the solar radiation is use-
ful for photocatalysis, and it is of interest to widen the active range so
as to include longer wavelengths, too (cf. Box 6.3). This widening can
be accomplished by doping the TiO2 with N, C, S, F, various metals,
or combinations of these dopants. Nitrogen doping to form TiO2−xNx
has been investigated in particular detail [33–36]. Figure 8.6 shows that
nitrogen doping extends the absorption to the 0.4 < λ < 0.5 μm range.
Large research activities are currently (2009) pursued in order to explore
how this correlates with photocatalytic activity and the general mecha-
nism outlined in Figure 8.3.
TiO2 has a high refractive index, and microparticles of this material
can backscatter strongly and hence be efficient in high-albedo paints (cf.
Section 7.3). Doping leads to some coloration, but this effect is confined to
a minor part of the solar spectrum so that TiO2–xNx pigments still make it
possible to combine overheat protection with efficient solar photocataly-
sis, as long as the paint binder does not cover the pigment surfaces.
Highly reactive species are usually created during the processes fol-
lowing photoexcitation, as shown in Figure 8.3. They can be used for
breakdown of organic molecules or microorganisms in contact with the
semiconductor surface. These processes can be used for numerous appli-
cations [32], such as

• Self-cleaning surfaces
• Air purification
• Self-sterilization

© 2011 by Taylor and Francis Group, LLC

 372 Green Nanotechnology

2-x x

Figure 8.6  Spectral optical absorption for films of TiO2 and TiO2–xNx.
(From G. Romualdo Torres et al., J. Phys. Chem. B 108 [2004] 5995–
6003. With permission.)
• Antifogging
• Heat transfer and heat dissipation
• Water purification
• Corrosion protection

A number of these applications are discussed next.

8.2.2  Self-Cleaning Surfaces

TiO2-coated ceramic tiles, cements, plastic foils, membrane materials

(for tents, etc.), and glass panes have been used for several years, espe-
cially for outdoor applications [31,32]. The self-cleaning is by far most
efficient in conjunction with water flow, such as natural rainfall, and is
associated with the superhydrophilic character of the TiO2 surface. This
surface is structurally modified and metastable, and its nature does not
seem to be fully understood [37]; it is not simply a very clean surface.
The water penetrates the molecular-level space between this surface and
the stain so that the latter can be washed away. Thus, cleaning can occur
even if the UV irradiation is not strong. Drying artifacts normally do
not remain after this cleaning. The cleaning may not be perfect, but
the superhydrophilicity increases the intervals between periodic clean-
ings and hence cuts down maintenance costs. So “easy-clean” may be a
better designation than “self-clean.” A number of striking illustrations
of the cleaning of objects that have been partly covered with TiO2 can
be found in the literature [31,32]. TiO2-based films also can serve as

© 2011 by Taylor and Francis Group, LLC

 Supporting Nanotechnologies 373

anticorrosion treatment, especially if they are deposited onto a layer of

phosphotungstic acid [38].
Figure  8.7 illustrates a number of features associated with super-
hydrophilic TiO2 surfaces [37]. The middle part shows the difference
between the wetting of an unirradiated film and of an UV irradiated,
superhydrophilic one. The wetting is characterized by a water contact
angle θwc , defined as the angle between the underlying surface and the
tangent of the liquid phase at the interface of the solid–liquid–gas phase
as shown in the upper part of Figure 8.7. Figure 8.7a shows that UV irra-
diation by 1.1 mW/cm 2 makes θwc decrease gradually as a function of the
irradiation time, and θwc ≈ 0 if this time is ~2 h, that is, the water then

Hydrophobic Hydrophilic
surface surface

θwc
θwc

UV

Dark

35 35
(a) (b)
30 30
Contact Angle (deg.)

Contact Angle (deg.)

25 25
20 20
15 15
10 10
5 5
0 0
0 40 80 120 0 500 1000 1500
UV Irradiation Time (min) Dark Storage Time (hour)

Figure 8.7  The upper part shows the definition of a water contact
angle θwc , which is large for a hydrophobic surface and small for a hydro-
philic surface. The middle part contains photographs of water coverage
on TiO2 surfaces kept in the dark and after UV irradiation. The lower
part shows water contact angle after UV irradiation (a) and after subse-
quent dark storage (b). (From K. Hashimoto et al., Jpn. J. Appl. Phys. 44
[2005] 8269–8285. With permission.)

© 2011 by Taylor and Francis Group, LLC

 374 Green Nanotechnology

covers the surface as a uniform thin layer. The superhydrophilicity is not

permanent, though, and storing the TiO2 in the dark for a few months
brings back the original wetting properties, as seen from Figure  8.7b.
Doped TiO2 films can show improved hydrophilicity [39].
Self-cleaning TiO2-based surfaces have been deposited onto float
glass in commercial products for several years; the coatings have been
made by spray pyrolysis as well as by reactive dc magnetron sputtering.
There are several issues with such products; one of them has to do with
the high refractive index of TiO2 , which leads to higher reflectance and
lower transmittance than for uncoated glass, and the other concerns the
loss of self-cleaning ability in the absence of significant solar irradiation.
As elaborated in Box 8.2, films consisting of TiO2 and SiO2 can improve
both of these issues.
The superhydrophilic surfaces spread water condensation so that
light scattering from droplets is avoided, that is, the surfaces are expe-
rienced as fog free. This is of obvious interest for windows and mirrors.
Figure 8.8 is a striking illustration and shows a thermally well-insulated
window in a private house in Uppsala, Sweden [43]. The data were taken
just after sunrise during a clear day in the fall. The right-hand part of
the window is covered by scattering droplets caused by dew formation,
which is associated with significant sky cooling (cf. Section 7.5). The
left-hand part of the same window has a commercial self-cleaning and
superhydrophilic TiO2-based coating on its outer surface. This coating
spreads the droplets and renders them invisible.
Antifogging can have wider applications than those discussed above
and can be useful to enhance boiling and evaporation heat transfer [44].
One application in this context is for cooling exterior walls of buildings
via falling water films in order to combat the urban heat island effect
(cf. Section 7.2) [37]. Another application of superhydrophilic surfaces
is in air coolers in which water bridges can be avoided between adjacent
cooling fins and therefore heat exchange efficiency can be maintained,
as well as air throughput. Superhydrophilicity is also useful for surfaces
employed in water supply by atmospheric water condensation via night
sky cooling.
Self-sterilization is a variety of self-cleaning and is of particular
interest for hospitals and facilities for elder care. Tiles covered with
TiO2:Cu and irradiated with UV have been shown to function well for
these applications [45].
Self-cleaning can be achieved not only by superhydrophilicity but
also by the reverse effect, superhydrophobicity. This is usually referred
to as the “lotus effect” and implies that water drops roll off the sur-
face of interest and then carry undesirable particulates with them.
Superhydrophobicity is at hand for θwc > 150° and leads to water
removal from surfaces at tilt angles that are only a few degrees from the

© 2011 by Taylor and Francis Group, LLC

 Supporting Nanotechnologies 375

Box 8.2  Enhanced Superhydrophilicity

in Films Comprising TiO2 and SiO2
Figure B8.2.1 shows spectral transmittance of glass coated with
a TiO2-SiO2 composite film in which submicrometer-size SiO2
and nanosize TiO2 particles were applied by stepwise electrostatic
deposition. These films serve as efficient antireflection coatings,
and it is seen that the transmittance can exceed 99% in the lumi-
nous range, while the self-cleaning properties are maintained [40].
Similar properties can be achieved in analogous double-layer TiO2-
SiO2 films [41].

100

98
Transmittance (%)

96

94

92
Glass

90

88
400 500 600 700 800
Wavelength (nm)

Figure B8.2.1  Spectral transmittance of glass with and without

a coating consisting of SiO2 particles covered with TiO2 nanopar-
tices as indicated in the inset sketch. (From X.-T. Zhang et al.,
Chem. Mater. 17 [2005] 696–700. With permission.)

TiO2-SiO2 films made by sol-gel technology can display natural

superhydrophilicity [42]. Figure B8.2.2 shows that θwc can remain
small for months in the dark, which is qualitatively different from
the behavior of pure SiO2 and TiO2 (cf. Figure 8.7). Much research
in self-cleaning oxides is ongoing, and the last word on this subject
is far from spoken.

© 2011 by Taylor and Francis Group, LLC

 376 Green Nanotechnology

35
30
SiO2

Contact Angle (deg.)

25
20

15

10
5 TiO2 – SiO2

0
0 10 20 30 40 50
Dark Storage Time (day)

Figure B8.2.2  Water contact angle for films of SiO2 and four dif-
ferent TiO2-SiO2 composites after dark storage. (From M. Houmard
et al., Surface Sci. 602 [2008] 3364–3374. With permission.)

Figure 8.8  Photograph of a window of a private house in Uppsala,

Sweden. Both parts of the window are triple glazed and have a cen-
ter-of-glass thermal conductance of 0.65 Wm−2K−1. The window is in a
northeast facing façade with an open view toward the sky. The left-hand
part of the window has an outer glass with a self-cleaning and super-
hydrophilic TiO2-based coating, and the right-hand part has standard
uncoated glass.

© 2011 by Taylor and Francis Group, LLC

 Supporting Nanotechnologies 377

horizontal. Detailed calculations have shown that θwc depends strongly

on the root-mean-square surface roughness amplitude but is rather
insensitive to other geometrical features [46]. Low tilt angles are associ-
ated with heterogeneous wetting, meaning that the fluid does not pen-
etrate into regions between the protrusions [47]. Surfaces meeting the
conditions for superhydrophobicity can be constructed in a variety of
ways and can consist of many materials, with “sponge-like” nanostruc-
tured TiO2 being one example of many [48].

8.2.3  Air Purification

An important application of photocatalysis is for decontamination,

deodorization, and disinfection of indoor air. The photocatalyst can be
attached to surfaces on walls, roofs, and floors, as well as to other sur-
faces such as on textiles or nonwoven fabrics. The photocatalytic sys-
tems are different from traditional filter-based air cleaners in that they
do not accumulate the pollutants and hence they avoid the risk of sec-
ondary pollution.
There is a large literature on air cleaning using devices incorporat-
ing UV sources. However, the natural solar radiation, as well as residual
UV light from fluorescent lights, can be used and offers advantages with
regard to energy use. Figure 8.9 shows the breakdown of acetone under
simulated sunlight in a reaction chamber containing surfaces of four
different commercial catalysts based on anatase TiO2 [49]. Clearly, some
of these catalysts can bring down the acetone concentration from 10 to
20% of the original content during 20 minutes of exposure.
One should note here that a regular window can be arranged as
a device for air cleaning if the outer glass is UV transparent (i.e., iron
free as discussed in Section 4.1) and a glass pane with a photocatalytic
surface is positioned so that it is irradiated by UV light while there are
provisions allowing the room air to circulate past the photocatalytic sur-
face [50]. Pigments of photocatalysts operating with visible light might
be used as paint materials on walls and ceilings in buildings equipped
with windows having normal float glass. However, one should be aware
of the fact that photocatalytic breakdown of VOCs may lead to the cre-
ation of noxious intermediate compounds such as formaldehyde and
acetaldehyde [51], and these aspects need further investigation.
Cementitious building materials are in themselves nanocomposites
[52]. They can be made photocatalytic by TiO2 admixture [53,54], and
such materials can be efficient for purification not only of indoor air
but also outdoors. This effect has been demonstrated convincingly with
regard to the removal of NOx caused by automobile exhausts. The puri-
fication can be particularly strong with concrete containing glass cullets

© 2011 by Taylor and Francis Group, LLC

 378 Green Nanotechnology

1 Hombikat UV100
Ishihara ST01
Ishihara ST21
Ishihara ST41
Acetone Concentration (c/c0)
0.8

0.6

0.4

0.2

0
0 20 40 60 80 100
Time (min)

Figure 8.9  Relative acetone concentration versus time as measured

in a reaction chamber containing four different commercial catalysts
and irradiated by artificial sunlight. The catalysts contain TiO2 anatase
nanoparticles with sizes in the 5 to 200 nm range and surface areas
between 300 and 10 m 2g−1. (From L. Österlund [2009], private commu-
nication. With permission.)
[55]. Air cleaning also can be achieved by inclusion of TiO2 in roadway
surfaces [32] and pavements [56]. The ensuing improvement in the air
quality will not only have beneficial health effects for persons being out-
side buildings, but the improved quality of the air that is let into the
building will diminish the need for indoor air treatment and hence be an
energy boon. Furthermore, the high refractive index of TiO2 can lead to
an increase of the albedo over that for common materials for buildings
and roads, thereby offsetting global warming, as further discussed in
Section 7.1.

8.3  Thermal Insulation with Nanomaterials

Thermal insulation materials for walls, ceilings, and roofs might be

thought of as based on mature technologies and hence are rather “low
tech.” However, the science behind even the most traditional porous
insulators is relatively complex, and the image of the thermal insulator
is about to change. Indeed, a leap forward is imminent and will involve

© 2011 by Taylor and Francis Group, LLC

 Supporting Nanotechnologies 379

nanostructures and vacuum. Most insulation is used in situations where

there is no light transmittance, but there are also applications where
insulation materials with good light transparency or translucency are of
interest. Clearly, transparent insulation is important for windows and
luminaires discussed in Chapters 4 and 5, and this issue is considered
separately at the end of this section.

8.3.1  Thermal Conductance of Porous and Nanoporous Materials

Good thermal insulation is essential for buildings in both cold and

warm climates and for refrigeration. It also plays an important role
in the efficiency of solar hot water collectors, high performance sky
cooling systems, hot water storage tanks, cold storage tanks, and for
transport and storage of fresh food and some medicines. Improvements
and wider use of good insulation can lead to very large energy sav-
ings, greater comfort, and less food and medicine spoilage. Codes for
energy-efficient buildings continue to require ever-higher standards of
thermal insulation, and this means ever-increasing thicknesses as long
as one is bound to traditional insulating materials. But larger thick-
ness can be awkward, and in many structures impossible, especially for
retrofitting of old buildings. Thus, the aim of the emerging insulation
materials addressed here is to achieve higher levels of performance in
thinner layers. The material property of interest is usually called ther-
mal conductivity or “kτ -value” [Wm−1K−1]. The use of the term thermal
conductivity, however, hides the fact that most good insulating mate-
rials—because they are porous—transport heat by three mechanisms:
conduction, convection, and radiation. Thus, kτ is, strictly speaking, an
effective thermal conductivity.
Thermal conductance U is given by U = kτ /di [Wm−2K−1], where di is
the thickness of the insulating layer, and the heat flow P across a tem-
perature drop Δτ is then P = UΔτ [Wm−2]. Many products are defined in
terms of their R-value or thermal resistance given by R = 1/U [W−1m 2]. As
a reference point for insulating ability, recall from Section 4.2 that mul-
tiglazed windows with good thermal insulation generally had a thermal
conductance between 1 and 2 Wm−2K−1, though the best achieved well
below 1 Wm−2 K−1 and, under special conditions, down to 0.2 Wm−2K−1
(the window in Figure 8.8 has 0.65 Wm−2K−1). How thick would a typi-
cal mineral wool insulating panel need to be to achieve such U values?
With kτ ~ 0.04 Wm−1K−1, a thickness of around 4 cm would provide U =
1 Wm−2K−1. However, most modern building codes presently require that
U be around 0.2 Wm−2 K−1or less, which means for the traditional prod-
ucts thicknesses of 20 cm or more. To put this in perspective we note

© 2011 by Taylor and Francis Group, LLC

 380 Green Nanotechnology

0.04 1 Glass fiber batt

2 Expanded polystyrene foam
0.035 3 Fluorcarbon expanded polyurethane foam
4 Gas-filled panel, argon
Conductivity (W/mK)
0.03 5 Gas-filled panel, krypton
6 Gas-filled panel, xenon
0.025 7 Powder vacuum panel
8 Aerogel vacuum panel
0.02 9 Glass fiber vacuum panel

0.015
0.01
0.005
0
1 2 3 4 5 6 7 8 9
Conventional Insulations Advanced Insulations

Figure 8.10  Thermal conductivity for conventional and advanced

materials. (From http://gfp.lbl.gov/.)
that for a 20 m 2 area of wall or ceiling and, say, Δτ = 15°C this means the
heat gain or heat loss falls from 300 to 60 W.
Conventional insulating materials are compared with one another
and with emerging ones in Figure 8.10 [57]. Traditional materials based
on glass and expanded polymer foam have 0.18 < kτ < 0.35 Wm−1K−1.
The best materials, on the other hand, utilize nanostructured aerogels
and vacuum and can achieve a kτ -value as low as 0.002 Wm−1K−1. These
two types of materials will be considered next.
The advantages of the nanoporous materials, such as inorganic aero-
gels and related polymeric materials, ensues from their favorable convec-
tive heat transfer. Convection arises because density gradients in a gas
exist across a space in which there is a thermal difference (i.e., molecules
spread out more in warmer gases). This fact, in addition to gravity, causes
hot gases to rise and cold to fall, and heat flow is then established. To
eliminate convection, one must ensure that there is not enough gas pres-
ent to generate such flows. There are two ways of doing this:

• To have vacuum
• To make the gas-filled pores small enough that they contain very
little gas even at normal pressure

A nanopore fulfills the second requirement, and convection is impossible

when the mean free path between molecule collisions becomes much
larger than the pore size. Nitrogen gas at room temperature, to take an
example, has a mean free path of around 100 nm.
Figure  8.11 illustrates convection prevention as the pore size is
decreased [58]. Once convection is under way, the molecular mass comes

© 2011 by Taylor and Francis Group, LLC

 Supporting Nanotechnologies 381

neon
Thermal conductivity(mW/°Cm)

nitrogen

CO2

log(pore diameter/mean free path)

Figure 8.11  Pore size versus onset of convection in nanoporous and

microporous insulating material incorporating three different gases.
(Replotted from data in J. C. Harper, A. F. El Sahrigi, Ind. Engr. Chem.
Fundamentals 3 [1964] 318–324.)
into play and lighter molecules are more able to transfer heat, as seen in
Figure  8.11. The residual thermal conductivity kτ ~ 0.009 Wm−1K−1 is
due to conduction along the thin skeleton material plus thermal radia-
tion across the pores.

8.3.2  Vacuum Insulation Panels

Vacuum insulation panels (VIPs) have become of interest in recent years

and can achieve kτ −values from 0.005 to 0.002 Wm−1K−1 (i.e., a U-value
as low as 0.1 Wm−2K−1). This would need a thickness of only 2 to 5
cm, which is suitable for retrofitting into many current wall cavities or
onto the internal wall surfaces without much loss of room space [59].
Figure 8.12 compares the thicknesses needed for the same level of insula-
tion between a VIP and a traditional insulating material [60]. A 20 m 2
wall area with Δτ = 15°C would pass less than 30 W of heat, and such
panels would also be very useful for sky cooling applications.
The VIP is normally based on a microporous material which is
“open” (i.e., without enclosed cavities), and vacuum is maintained by
aluminum foils which seal off the internal structure. Obviously, the VIPs
must be handled so that this seal is not damaged. Another practical
aspect is that VIPs may decrease the acoustic insulation. VIPs can be
used within door and window frames, on floors, under flat roofs, and on
terraces and internal floors. The insulating performance of doors, which

© 2011 by Taylor and Francis Group, LLC

 382 Green Nanotechnology

Figure 8.12  Thicknesses for traditional glass wool (left) and a VIP
(right), both having the same U-value. (From R. Baetens et al., Energy
Buildings 42 [2010] 147–172. With permission.)

is a weak spot in many well-insulated buildings, is typically increased

by 50% [59].
Air and moisture ingress slowly degrade the vacuum in VIPs, and
hence their U-value will increase over time. This loss of performance
may eventually be manageable with gettering and desiccants, but that
remains for future study. Thus establishing the relative merits of VIPs
and nanofoams, characterized by closed nanocavities, must await fur-
ther development and testing. The main hold-up for the widespread
use of VIPs seems to be the current manufacturing methods, which are
labor intensive.

8.3.3  Silica Aerogel

Silica aerogel is a nanoporous translucent material that can be used

in fenestration of various types [61–63]. A photo of this material was
shown in Figure 8.1. It consists of an open structure of interconnected
SiO2 nanoparticles with diameters down to 1 nm. The pore size is 1
to 100 nm, and the overall porosity can exceed 99%. The material is
mechanically fragile, though it can be strong enough to be load bearing

© 2011 by Taylor and Francis Group, LLC

 Supporting Nanotechnologies 383

100

Transmittance (%)
50

0
0.5 1 1.5 2 2.5 3
Wavelength (µm)

Figure 8.13  Spectral transmittance through a 4-mm-thick silica aero-

gel tile. (From C. G. Granqvist, Materials Science for Solar Energy
Conversion Systems, edited by C. G. Granqvist, Pergamon, Oxford,
U.K., 1991, pp. 106–167. With permission.)

in evacuated devices for transparent insulation. It is normally produced

as tiles or granules via supercritical drying.
Figure 8.13 shows the spectral transmittance through an aerogel tile
[64,65]. The transmittance lies above 80% for 0.6 < λ < 2.2 μm but
drops sharply at shorter wavelengths as a result of diffuse scattering
due to density fluctuations. This scattering leads to a bluish appearance
of thick aerogel tiles, as apparent from Figure 8.1. The aerogels can be
used in VIPs and yield superior thermal insulation, especially if radiative
transfer across the aerogel is prevented by low-emittance surfaces.

8.4  Green Energy Storage

8.4.1  Energy Storage: Survey of a “Missing Link”

Energy storage is a vital aspect of systems for renewable energy and

improved energy efficiency. It is sometimes referred to as the “missing
link” for such systems [66]. Storage is essential if the renewables compo-
nent of our energy supply is to be capable of providing base-load power;
without this capability the nonhydro contribution will be capped, proba-
bly well below 50%. One attraction of renewable power sources to some
people, especially those in remote locations, is that they can create an
independence from external grid-based supplies or diesel fuel deliveries.
For many of the millions who make up the world’s nonurban poor, these

© 2011 by Taylor and Francis Group, LLC

 384 Green Nanotechnology

sources can liberate them from the grinding task of collecting waning
supplies of firewood.
It is also attractive from an energy security viewpoint to have local
supplies, backed by moderate-size storage, in addition to external sup-
plies. This means, for example, that power is available both during a
blackout and during an extended lack of local solar energy due to bad
weather. Energy security is also a national issue as outside resources
become scarce, too expensive, or are cut off as a result of political,
industrial, or other events. Thus, fires in gas processing plants, or pipe-
line problems, have been known to severely reduce gas or oil supplies
in some regions for many months on end. Large local and community-
scale storage of energy will also become increasingly attractive if severe
weather events become more common, which is predicted to happen as
the impact of global warming grows. Thus, renewables combined with
advanced storage technologies will not only help to reduce the extent of
global warming but they also hold much promise as measures for adap-
tation to global warming.
Energy can be stored in many forms: gravitational, electrical, mag-
netic, thermal, chemical, and mechanical. Volume, weight, cost, leak-
age, internal loss, and ease and speed of charging and discharging must
all be considered. The more energy one can easily pack into and remove
from a given volume or mass the better. Mechanical energy storage is
not just in kinetic form, as in a flywheel, but also in stationary form as
elastic strain or compressed air. Coal, oil, and natural gas are stores of
chemical energy laid down by natural processes over millennia. The sun
is a store of nuclear fusion fuel. Each form has a major role in renew-
able energy systems, including dams (gravitational), hydrogen storage
(chemical splitting of water), hot water tanks, massive building compo-
nents, deep underground rocks heated by natural nuclear decay, phase
change materials (thermal and geothermal), and superconducting coils
(magnetic). Thermal storage is important and a natural adjunct to solar
thermal systems and night sky cooling. The key aspect is what material
to use, and this depends on the temperature at which the thermal store
operates. For cooling purposes one needs to store energy from ~0 to
~20°C, for domestic hot water from around 60 to 80°C, and for solar
thermal power at hundreds of degrees. Water, water-glycol mixtures,
and phase change materials are ideal for temperatures from 100°C and
down to about 0°C. For very high temperatures, one can use molten
salts, solid rocks, or large blocks of graphitic carbon. Hot rock and car-
bon block storage is, in essence, man-made geothermal storage.
The focus for the discussion below is on advanced electric energy
storage in batteries and capacitors utilizing nanostructured materi-
als. The likely eventual shift to an all-electric or hybrid-electric vehi-
cle fleet seems to be the main driving force at present in the quest for

© 2011 by Taylor and Francis Group, LLC

 Supporting Nanotechnologies 385

improvements in electrical storage and electrical power management,

with batteries and capacitors being the key components. Better batteries
are also needed for remote and stand-alone renewable energy systems.
High-performance capacitors, which can charge and discharge quickly,
are very important for vehicle acceleration and power switching, as well
as for information, communication, and display technologies.
Nearly all of the major advances in compact electrical storage are
being driven by progress in nanomaterials, and the main attraction is in
their exceptionally large surface-to-volume ratio. Surfaces and interfaces
are where the electrical energy is stored in advanced capacitors, and it is
also where charging and discharging take place in batteries.

8.4.2  Electrical Storage Using Electrochemistry

Most of us carry around batteries in a variety of small devices such as

mobile phones, laptops, digital cameras, calculators, etc. The majority
of these devices are rechargeable, sometimes by use of small solar cells.
For larger power, lead acid car batteries have long been the stalwarts of
vehicle electrical systems and of solar power storage. These batteries are
heavy, include poisonous or otherwise undesired materials, and often
occupy too much space to be convenient for solar power systems.
Today’s electronic systems can also rely on small, or even tiny, devices
for temporary storage of energy (i.e., capacitors). Battery and capaci-
tor technologies have evolved extensively in recent years, and further
large improvements are expected. Many of these are due to nanoscience.
Developments over the last two decades have begun to fill in missing
parts of the “electrical storage landscape,” which traditional batteries
and capacitors did not cover. We first need to become familiar with this
“landscape” via its “map” so as to fully appreciate what nanoscience has
recently added and has to offer in the future.
Batteries and capacitors have to both store charge and convert their
stored energy into electrical power. The latter is a time-dependent issue.
If a lot of energy has to go in or out in a short time, then power flow is
the main issue. And if a lot of energy has to be stored, then high energy
capacity is needed. Capacitors are good at the former, batteries at the
latter. Storage capacity is important for systems which deliver power
steadily, such as solar systems for homes or buildings that are not grid
connected. Having both high power flow and high energy capacity in
one device would be ideal for electric cars, but it is not easily accom-
plished. Cars need a lot of storage and also high power for acceleration,
and ideally high power flow for fast charging ability. The latter aspect
is the least developed one, but some nanostructured batteries make it a
genuine possibility to fully charge a car at a “battery power station” in

© 2011 by Taylor and Francis Group, LLC

 386 Green Nanotechnology

the same time it takes to fill today’s car with petrol at a gas station. At
the moment, though, an electric car needs all-night charging, or “most-
of-the-day” or “while-at-work-or-shopping” charging. An alternative
might be a quick change-over battery pack from specialized “battery
stations,” and such facilities are starting to appear to service the growing
electrical vehicle fleet.
The power delivered per unit mass in W/kg is called the power den-
sity, and the energy stored per unit mass in Wh/kg or J/kg is called
the energy density. The “electrical storage landscape” involves a “map”
where these two functions are plotted against each other in a Ragone
chart such as the one shown in Figure 8.14 [67].
Traditional capacitors can deliver 104 to 107 W/kg but can only
store 0.01 to 0.05 Wh/kg. Lead acid batteries can deliver, at best, a
much more leisurely 100 W/kg but have high storage, between 20 Wh/
kg to 8 Wh/kg. What about the storage “territory” between 0.05 and
10 Wh/kg with good power? Can we exceed 20 Wh/kg storage and/or
achieve much better power rates in batteries other than lead-acid ones?

10000
4s 40s
6 min
1000 Electrochemical 1h
Power Density (W/kg)

capacitor
10 h
100
Pb-Acid Zinc-Air
Ni-Cd Ni/MH 100 h
Alkaline Primary
Li-ion
Li
10
Fuel cell
1000 h

1
1 10 100 1000 10000
Energy Density (Wh/kg)

Figure 8.14  Ragone chart of power density versus energy density for
electrochemical capacitors, batteries, and fuel cells. MH denotes metal
hydride. Traditional capacitors are not included as they would require an
extension of the chart by two orders of magnitude down for the x-axis
and three orders up for the y-axis. Diagonal lines represent charge/dis-
charge times. (From Y. Zhang et al., Int. J. Hydrogen Energy 34 [2009]
4889–4899. With permission.) The solid circle has been added to show
the measured performance of a nanostructured virus-based battery in
order to indicate what is becoming possible (cf. Box 8.3 for details).

© 2011 by Taylor and Francis Group, LLC

 Supporting Nanotechnologies 387

Definitely! The question now is how far we can go to cover the map in
Figure 8.14. A combination of nanoscience and advances in electrochem-
istry have allowed the filling in of many gaps with batteries based on
lithium chemistry, Ni-Cd, and Ni metal hydrides as well as with more
recent devices such as fuel cells and electrochemical capacitors similar to
the one discussed shortly.
With the advent of very large scale photovoltaic systems, a totally
new class of batteries is being considered for stationary applications;
they need both very large storage capacity and ability for ultra large cur-
rent flows upon charging and discharging. These batteries utilize two
liquid metals separated by a liquid salt, which means they have to be
maintained at ~700°C. The underlying idea is that ions can move much
more easily in liquids than in solids. This work has only just begun, but
is of high interest [68]. It builds upon the well-known high-temperature
electrochemical process for producing aluminum metal in which alumi-
num oxide is first dissolved in molten salt cryolite (Na3AlF6) at 900°C to
enable extraction and reduction of the Al3+ ions to Al metal.

8.4.3  Electrochemical Super- and Ultracapacitors

The structure and chemistry of electrodes and electrolytes determine

power flows and can also influence storage capacity. Electrochemical
capacitors include super- and ultracapacitors, and their scientific designa-
tion “electrical double-layer capacitors” indicates how they work. Here,
a charge that is opposite to that on the electrode builds up alongside it in
the electrolyte, not on the opposite electrode as in a conventional capaci-
tor. The thickness ddc of this double charge layer is only about 0.5 to 1
nm. Using the basic relation for capacitance, C = ε0εA/ddc where A is the
area, now yields that C is 10 to 20 microfarad per cm 2 or, in mass units,
of the order of 100 farad per gram of electrode (for nanoporous carbon
electrodes). The voltage U ranges from around 5 to 20 V, and the stored
energy, given by ½CU2 , is about 10 −3 J/cm 2 . No chemical reaction occurs
at the electrode, only charge separation, and this means that many more
charge–discharge cycles can occur than in a battery over the lifetime of
the device.
The capacitance is very high in supercapacitors and ultracapacitors,
and ranges from ~ 0.1 to more than 20 F, which implies that the active
electrode area lies between ~1 and 200 m 2 . How can such a large area
occur in a compact device? The reason is that the conducting electrodes
are highly nanoporous. The lower size limit of the pores, which then
dictates the maximum effective area, is set by the requirement that ions
in the solution should be able to enter them and is estimated to be about
2 nm [67]. Nanoporous carbon, which exists in many forms, has been

© 2011 by Taylor and Francis Group, LLC

 388 Green Nanotechnology

Electrolyte, separator
Electrolyte, active layer

Current collector + – Current collector

+ –
+ –
+ –
+ + – –
+ –
+ –
+ –

Carbon particles in
contact with an
electrolyte film
+ –
+ + + – – –
+ – –
+ + – –
+ –

+ – + –
+ – + –
+ – + –
+ – + –
1–4 V
+ – + –
+ – + –
+ – + –

0.2–1 nm

Figure 8.15  Basic double-layer electrochemical cell using carbon par-

ticles packed around each electrode. An inert, porous separator serves as
a pure ion conductor and prevents any flow of electrons from one side to
the other. The carbon particles are charged oppositely on the two sides.
The lower plot shows the double-layer on each particle and the associ-
ated voltage profile on pairs of particles on opposite sides of the separa-
tor. (From R. Kötz, M. Carlen, Electrochim. Acta 45 [2000] 2483–2498.
With permission.)
the mainstay of this technology, and Figure 8.15 outlines the principles
and basic structure of electrochemical capacitors based on this mate-
rial [69]. Devices can involve packed nanoparticle powders, as well as
activated carbon, which is a highly porous form that is usually prepared
by pyrolysis. Carbon nanotubes and nanofibers have been used recently
[67,70] and lead to fast charge/discharge rates though rather low energy
density. Mesoporous and nanoporous gold is also of interest for electro-
chemical capacitors and enables moderate storage with fast discharge at
high power [71].
Recent interest in electrochemical storage has been directed at the
new highly conducting carbon-based nanomaterial known as “gra-
phene.” It essentially comprises single layers or sheets of carbon atoms

© 2011 by Taylor and Francis Group, LLC

 Supporting Nanotechnologies 389

as in graphite, and the structure was shown in the right-hand part of

Figure  4.27 [72]. In a study of their potential as ultracapacitors, the
graphene sheets agglomerated partly into particles but had sections
with many free individual sheets [73]. Capacitance values as large as
135 and 99 F/g were measured with aqueous and organic electrolytes,
respectively. Given their scope for improvements, these nanomaterials
may even ultimately play a role in energy storage rivaling some batter-
ies. In the future, commercial carbon-based electrochemical capacitors
may store 20 to 40 Wh/kg [70]. Looking at Figure 8.14, this puts them
in the domain of many of today’s batteries, but with longer lifetimes and
higher powers if required.
The need for both power and sufficient storage has led to hybrid
systems. These are based on a variety of the electrochemical capacitors,
called pseudo-capacitors, in which the electrode interacts chemically to
some extent with the electrolyte (i.e., it is part capacitor and part bat-
tery). The electrodes in this case are conducting oxides or conducting
polymers whose functionality again relies on nanoporosity. One of the
cheapest options is porous MnO2 , which can supply up to 260 F/g at
0.8 V. More complex and expensive oxide and polymer materials under
development can achieve up to ~1000 F/g at around 1 V [67,74].

8.4.4  Nanomaterials for Advanced Batteries

Developments in nanostructured materials for battery electrodes and

electrolytes are both needed if major advances are to occur, and the main
focus is currently on Li+ ion-based batteries. The goals are faster charg-
ing and discharging and better storage, along with extended lifetime.
Anode materials may involve nanostructured alloys and composites
which store Li+ —that is, intercalation hosts for Li+ such as mixtures of
Li and TiO2 nanowires that can accommodate enough lithium to form
Li0.91TiO2 [74]. Various oxides with internal nanostructures that let them
take up Li+ are other alternatives. Rod-shaped nanoparticles of lithium
iron phosphate coated with thin layers of a conducting polymer such
as polypyrrole, and a mix of polypyrrole and polyethylene glycol are of
current interest for electric cars but need further development. The aim
of coating the nanoparticles with a polymer conductor is to minimize
the charge transfer resistance between the LiFePO4 particles and the
polymer electrolyte [75]. Recently, LiFePO4 nanoparticles, devised for
battery electrodes, with slightly shifted stoichiometry have been found
to conduct ions along the particle surface at high rates. A full battery
discharge in 10 to 20 s seems to be achievable with electrodes based
on these nanoparticles mixed with carbon. The addition of carbon is
needed to allow fast electron flow in order to match the fast ion flow

© 2011 by Taylor and Francis Group, LLC

 390 Green Nanotechnology

[76]. Clearly, much new science is needed for the processes that occur at
nanoelectrodes of different materials before one can proceed with confi-
dence to commercial systems in this field. It is an area of much technical
promise that needs urgent and accelerated research and development.
Nanocathodes are less developed than nanoanodes. The reaction
with the electrolytes can be facilitated with nanoparticles, but this may
also lead to overheating and safety issues. Nanopillars of various oxides
are of concern for cathodes, and such pillars of V2O5 inside porous alu-
mina are of special interest. The V2O5 can also be in nanoporous form.
There is much interesting science to sort out in order to understand how
the physical and lattice structures of these and other cathode materials
interact with the electrolyte and how they intercalate lithium ions [74].
Nanocomposites of ceramic nanoparticles embedded in solid poly-
meric ion conductors—such as polyethylene oxide (PEO) combined
with a lithium salt—show much promise for future improvement in the
performance of electrolytes. They have already led to large enhance-
ments compared to the starting materials. Significant ionic conductivity
in polymeric conductors was only thought to happen at high tempera-
tures, above the glass transition temperature where the polymer chains
become disordered. But crystalline polymer compounds have now
been found in which the Li+ ions conduct and reside in tunnels formed
among the polymer chains, as seen in Figure  8.16. These structures
facilitate fast ion transport. The same electrolyte materials are also of
value for supercapacitors.
We end with two examples of batteries devised according to princi-
ples rooted in biology. Thus, recent work on an all-polymer paper-based
battery-employed cellulose fibers of algal origin (Cladophora) coated
with 50-nm-thick layers of polypyrrole [77]. The specific surface area
of this cellulose is exceptionally large—much larger than for the cellu-
lose typically used in the paper industry—which allows large charging
capacity at a high rate. It is interesting to note that Cladophora is usually
considered an environmental pollutant, which here it is put to good use
via nanotechnology.
Another example of a new nanomaterials-based approach to high-
performance batteries touches on an area of materials development that
is in its infancy but has much to offer. It uses genetic engineering, which
now has reached a point at which it is feasible to implement nanoscale
electrical wiring based on biological principles. This is further elabo-
rated in Box 8.3.

© 2011 by Taylor and Francis Group, LLC

 Supporting Nanotechnologies 391

(a) (b)

Figure 8.16  A color version of this figure follows page 200. Structure
of the nanostructured polymer ion conductor PEO6:LiAsF6 as seen look-
ing into the chains and tunnels (a) and along them (b). The Li+ sits inside
the tunnels formed by the PEO. The atoms are fluorine, carbon, and
oxygen. (From A. S. Aricò et al., Nature Mater. 4 [2005] 366–377. With
permission.)

Box 8.3 A Virus-Based Nanobattery

Multifunctional viruses can be used as scaffolds for the synthesis
of high-performance nanostructured electrodes in batteries [78].
A two-gene virus enabled linking of amorphous a-FePO 4 ·H 2O
nanowires to single-wall carbon nanotubes (SWNTs) and made
it possible to engineer very high performance nanoelectrodes that
combine a large storage capacity with high power delivery, as
apparent from the Ragone chart in Figure B8.3.1a. The cycling
stability was good, which can be inferred from Figure B8.3.1b. The
nanostructure resulting from the two-gene tethering of SWNTs
to a-FePO 4 ·H 2O is depicted in Figure B8.3.2 at high and medium
magnifications. The performance of these batteries is remarkable,
as can be seen by comparison to other battery technologies in
Figure  8.14, and points at the opportunities that nanomaterials
will provide in order to meet future needs for electrical storage.

© 2011 by Taylor and Francis Group, LLC

 392 Green Nanotechnology

600 200
500
400

Specific Capacity
Specific Energy

300 150
(Wh/Kg)

(mAh/g)
200

EC#2
100 EC#2 50
EC#1 EC#1
E4 E4

101 102 103 104 10 20 30 40 50

Specific Power (W/Kg) Cycle Number
(a) (b)

Figure B8.3.1  Ragone chart of energy storage versus power

capability for batteries based on one-gene (E4) and two-gene
(EC#1 and EC#2) tethering (a), and cycling stability of these sys-
tems (b). (Modified image from Y. J. Lee et al., Science 324 [2009]
1051–1055. With permission.)

a-FePO4

SWNTs

100 nm 4 nm

(a) (b)

Figure B8.3.2  Nanostructure of a-FePO4 grown on the two-

gene virus system EC#2 and tethered to single-walled carbon nano-
tubes (SWNTs). Panel (a) shows a low-magnification transmission
electron micrograph (30,000x) with faint images of SWNTs in the
upper part and a-FePO4 attached to EC#2 in the lower part. Panel
(b) is a very high magnification image (800,000x) showing tether-
ing of a single SWNT to an a-FePO4 particle. (From Y. J. Lee et al.,
Science 324 [2009] 1051–1055. With permission.)

© 2011 by Taylor and Francis Group, LLC

 Supporting Nanotechnologies 393

References

1. http://www.airglass.se.
2. J. A. Leech, W. C. Nelson, R. T. Burnett, A. Aaron, M. E. Raizenne, It’s
about time: A comparison of Canadian and American time-activity
patterns, J. Exposure Anal. Environm. Epidem. 12 (2002) 427–432.
3. T. Godish, Indoor Environmental Quality, CRC Press, Boca Raton,
FL, 2001.
4. M. J. Mendell, W. J. Fisk, K. Kreiss, H. Levin, D. Alexander, W. S. Cain,
J. R. Girman, C. J. Hines, P. A. Jensen, D. K. Milton, L. P. Rexroat,
K. M. Wallingford, Improving the health of workers in indoor
environments: Priority research needs for a national occupational
research agenda, Am. J. Public Health 92 (2002) 1430–1440.
5. M. Murphy, Sick Building Syndrome and the Problem of Uncertainty:
Environmental Politics, Technoscience, and Women Workers, Duke
University Press, Durham, NC, 2006.
6. P. Wolkoff, C. K. Wilkins, P. A. Clausen, G. D. Nielsen, Organic com-
pounds in office environments: Sensory irritation, odor, measurements
and the role of reactive chemistry, Indoor Air 16 (2006) 7–19.
7. H. Guo, N. H. Kwok, H. R. Cheng, S. C. Lee, W. T. Hung, Y. S. Li,
Formaldehyde and volatile organic compounds in Hong Kong homes:
Concentrations and impact factors, Indoor Air 19 (2009) 206–217.
8. Z. Lu, C. M. Li, Y. Qiao, Y. Yan, X. Yang, Effect of inhaled form-
aldehyde on learning and memory of mice, Indoor Air 18 (2008)
77–83.
9. B. M. Eklund, S. Burkes, P. Morris, L. Mosconi, Spatial and tempo-
ral variability in VOC levels within a commercial retail building,
Indoor Air 18 (2008) 365–374.
10. B. Wessén, K.-O. Schoeps, Microbial volatile organic compounds:
What substances can be found in sick buildings? Analyst 121 (1996)
1203–1205.
11. C. G. Granqvist, A. Azens, P. Heszler, L. B. Kish, L. Österlund,
Nanomaterials for benign indoor environments: Electrochromics
for “smart windows,” sensors for air quality, and photo-catalysis for
air cleaning, Solar Energy Mater. Solar Cells 91 (2007) 355–365.
12. G. Eranna, B. C. Joshi, D. P. Runthala, R. P. Gupta, Oxide materials
for development of integrated gas sensors: A comprehensive review,
Crit. Rev. Solid State Mater. Sci. 29 (2004) 111–188.
13. M. Graf, A. Gurlo, N. Bârsan, U. Weimar, A. Hierlemann,
Microfabricated gas sensor systems with sensitive nanocrystalline
metal-oxide films, J. Nanoparticle Res. 8 (2006) 823–839.
14. N. Barsan, D. Koziej, U. Weimar, Metal oxide-based gas sensor
research: How to? Sensors Actuators B 121 (2007) 18–35.

© 2011 by Taylor and Francis Group, LLC

 394 Green Nanotechnology

15. M. Tiemann, Porous metal oxides as gas sensors, Chem. Eur. J. 13

(2007) 8376–8388.
16. N. Barsan, M. Schweizer-Berberich, W. Göpel, Fundamental and
practical aspects in the design of nanoscaled SnO2 gas sensors: A
status report, Fresenius J. Anal. Chem. 365 (1999) 287–304.
17. S. Pokhrel, C. E. Simion, V. S. Teodorescu, N. Barsan, U. Weimar,
Synthesis, mechanism, and gas-sensing application of surfactant
tailored tungsten oxide nanostructures, Adv. Functional Mater. 19
(2009) 1767–1774.
18. A. S. Zuruzi, N. C. MacDonald, M. Moskovits, A. Kolmakov, Metal
oxide “nanosponges” as chemical sensors: Highly sensitive detec-
tion of hydrogen using nanosponge titania, Angew. Chem. Int. Ed.
46 (2007) 4298–4301.
19. E. H. Espinosa, R. Ionescu, E. Llobet, A. Felten, C. Bittencourt, E.
Sotter, Z. Topalian, P. Heszler, C. G. Granqvist, J. J. Pireaux, X.
Correig, Highly selective NO2 gas sensors made of MWCNTs and
WO3 hybrid layers, J. Electrochem. Soc. 154 (2007) J141–J149.
20. A. Hoel, L. F. Reyes, S. Saukko, P. Heszler, V. Lantto, C. G. Granqvist,
Gas sensing with films of nanocrystalline WO3 and Pd made by
advanced reactive gas deposition, Sensors Actuators B 105 (2005)
283–289.
21. P. Heszler, R. Ionescu, E. Llobet, L. F. Reyes, J. M. Smulko, L. B.
Kish, C. G. Granqvist, On the selectivity of nanostructured semicon-
ductor gas sensors, Phys. Stat. Sol. B 244 (2007) 4331–4335.
22. L. F. Reyes, A. Hoel, S. Saukko, P. Heszler, V. Lantto, C. G. Granqvist,
Gas sensor response of pure and activated WO3 nanoparticle films
made by advanced reactive gas deposition, Sensors Actuators B 117
(2006) 128–134.
23. R. Ionescu, A. Hoel, C. G. Granqvist, E. Llobet, P. Heszler, Ethanol
and H2S gas detection in air and in reducing and oxidising ambi-
ence: Application of pattern recognition to analyse the output from
temperature-modulated nanoparticulate WO3 gas sensors, Sensors
Actuators B 104 (2005) 124–131.
24. R. Ionescu, A. Hoel, C. G. Granqvist, E. Llobet, P. Heszler, Low-level
detection of ethanol and H2S with temperature-modulated WO3
nanoparticle gas sensors, Sensors Actuators B 104 (2005) 132–139.
25. L. B. Kish, Y. Li, J. L. Solis, W. H. Marlow, R. Vajtai, C. G. Granqvist,
V. Lantto, J. M. Smulko, G. Schmera, Detecting harmful gases using
fluctuation-enhanced sensing with Taguchi sensors, IEEE Sensors J.
5 (2005) 671–676.
26. C.-Y. Lee, C.-M. Chiang, Y.-H. Wang, R.-H. Ma, A self-heating gas
sensor with integrated NiO thin-film for formaldehyde detection,
Sensors Actuators B 122 (2007) 503–510.

© 2011 by Taylor and Francis Group, LLC

 Supporting Nanotechnologies 395

27. C. Luyo, R. Ionescu, L. F. Reyes, Z. Topalian, W. Estrada, E. Llobet,

C. G. Granqvist, P. Heszler, Gas sensing response of NiO nanopar-
ticle films made by reactive gas deposition, Sensors Actuators B 138
(2009) 14–20.
28. A. L. Linsebigler, G. Lu, J. T. Yates, Jr., Photocatalysis on TiO2 sur-
faces: Principles, mechanisms, and selected results, Chem. Rev. 95
(1995) 735–758.
29. T. L. Thompson, J. T. Yates, Jr., Surface science studies of the photo-
activation of TiO2: New photochemical processes, Chem. Rev. 106
(2006) 4428–4453.
30. L. Österlund, Structure-reactivity relationships of anatase and rutile
TiO2 nanocrystals measured by in situ vibrational spectroscopy,
in Solid-State Chemistry and Photocatalysis of Titanium Dioxide,
edited by M. K. Nowotny and J. Nowotny, Trans Tech Publ., Stafa-
Zürich, Switzerland, 2010, Solid State Phenomena, 162, 203–219.
31. A. Fujishima, K. Hashimoto, T. Watanabe, TiO2 Photocatalysis:
Fundamentals and Applications, Bkc, Tokyo, Japan, 1999.
32. A. Fujishima, X. Zhang, D. A. Tryk, TiO2 photocatalysis and related
surface phenomena, Surf. Sci. Rep. 63 (2008) 515–582.
33. R. Asahi, T. Morikawa, T. Ohwaki, K. Aoki, Y. Taga, Visible-light
photocatalysis in nitrogen-doped titanium oxides, Science 293
(2001) 269–271.
34. T. Lindgren, J. M. Mwabora, E. Avendaño, J. Jonsson, A. Hoel,
C. G. Granqvist, S.-E. Lindquist, Photoelectrochemical and opti-
cal properties of nitrogen doped titanium dioxide films prepared
by reactive DC magnetron sputtering, J. Phys. Chem. B 107 (2003)
5709–5716.
35. G. Romualdo Torres, T. Lindgren, J. Lu, C. G. Granqvist, S.-E.
Lindquist, Photoelectrochemical study of nitrogen-doped tita-
nium dioxide for water oxidation, J. Phys. Chem. 108 (2004)
5995–6003.
36. J. M. Mwabora, T. Lindgren, E. Avendaño, T. F. Jaramillo, J. Lu, S.-E.
Lindquist, C. G. Granqvist, Structure, composition, and morphol-
ogy of photoelectrochemically active TiO2-xNx thin films deposited
by reactive DC magnetron sputtering, J. Phys. Chem. B 108 (2004)
20193–20198.
37. K. Hashimoto, H. Irie, A. Fujishima, TiO2 photocatalysis: A histori-
cal overview and future prospects, Jpn. J. Appl. Phys. 44 (2005)
8269–8285.
38. P. Ngaotrakanwiwat, S. Saitoh, Y. Ohko, T. Tatsuma, A. Fujishima,
TiO2-phosphotungstic acid photocatalyst systems with an energy
storage ability, J. Electrochem. Soc. 150 (2003) A1405–A1407.

© 2011 by Taylor and Francis Group, LLC

 396 Green Nanotechnology

39. Y. W. Sakai, K. Obata, K. Hashimoto, H. Irie, Enhancement of vis-

ible light-induced hydrophilicity on nitrogen and sulfur-doped TiO2
thin films, Vacuum 83 (2009) 683–687.
40. X.-T. Zhang, O. Sato, M. Taguchi, Y. Einaga, T. Murakami, A.
Fujishima, Self-cleaning particle coating with antireflection proper-
ties, Chem. Mater. 17 (2005) 696–700.
41. X.-T. Zhang, A. Fujishima, M. Jin, A. V. Emeline, T. Murakami,
Double-layered TiO2-SiO2 nanostructured films with self-clean-
ing and antireflective properties, J. Phys. Chem. B 110 (2006)
25142–25148.
42. M. Houmard, D. Riassetto, F. Roussel, A. Bourgeois, G. Berthomé,
J. C. Joud, M. Langlet, Enhanced persistence of natural super-hydro-
philicity in TiO2-SiO2 composite thin films deposited via sol-gel
route, Surface Sci. 602 (2008) 3364–3374.
43. A. Roos, A. Werner, Formation of external condensation on low
U-value windows with and without hydrophilic coatings, in
Proceedings 7th International Conference on Coatings on Glass and
Plastics, edited by C. I. M. A. Spee, J. L. B. de Groot, H. A. Meinema,
J. Pütz, TNO Science and Industry, Eindhoven, the Netherlands,
2008; pp. 295–298.
44. Y. Takata, S. Hidaka, J. M. Cao, T. Nakamura, H. Yamamoto, M.
Masuda, T. Ito, Effect of surface wettability on boiling and evapora-
tion, Energy 30 (2005) 209–220.
45. T. N. Rao, D. A. Tryk, A. Fujishima, Applications of TiO2 photo-
catalysis, in Encyclopedia of Electrochemistry: Semiconductor
Electrodes and Photoelectrochemistry, edited by A. J. Bard, M.
Stratmann, S. Licht, Wiley-VCH, Weinheim, Germany, 2002; Vol. 6,
pp. 536–561.
46. C. Yang, U. Tartaglino, B. N. J. Persson, Influence of surface rough-
ness on superhydrophobicity, Phys. Rev. Lett. 97 (2006) 116103
1–4.
47. A. Marmur, The Lotus effect: Superhydrophobicity and metastabil-
ity, Langmuir 20 (2004) 3517–3519.
48. Y. Lai, C. Lin, J. Huang, H. Zhuang, L. Sun, T. Nguyen, Markedly
controllable adhesion of superhydrophobic spongelike nanostruc-
ture TiO2 films, Langmuir 24 (2008) 3867–3873.
49. L. Österlund (2009), private communication.
50. C.G. Granqvist, Materials for good day-lighting and clean air: New
vistas in electrochromism and photocatalysis, in Society of Vacuum
Coaters 50th Annual Technical Conference Proceedings, Society
of Vacuum Coaters, Albuquerque, NM, 2007; pp. 561–567; also
in Spring Bulletin, Society of Vacuum Coaters, Albuquerque, NM,
2008; pp. 48–53.

© 2011 by Taylor and Francis Group, LLC

 Supporting Nanotechnologies 397

51. A. T. Hodgson, H. Destaillats, D. P. Sullivan, W. J. Fisk, Performance

of ultraviolet photocatalytic oxidation for indoor air cleaning appli-
cations, Indoor Air 17 (2007) 305–316.
52. Z. Bittnar, P. M. J. Bartos, J. Nĕmecĕk, V. Šmilauer, J. Zeman, Eds.,
Nanotechnology in Construction: Proceedings of the NICOM3,
Springer, Berlin, Germany, 2009.
53. L. Cassar, Photocatalysis of cementitious materials: Clean buildings
and clean air, MRS Bull. (May) (2004) 328–331.
54. B. Ruot, A. Plassais, F. Olive, L. Guillot, L. Bonafous, TiO2-containing
cement pastes and mortars: Measurements of the photocatalytic effi-
ciency using a rhodamine B-based colourimetric test, Solar Energy
83 (2009) 1794–1801.
55. J. Chen, C.-S. Poon, Photocatalytic activity of titanium diox-
ide modified concrete materials: Influence of utilizing recycled
glass cullets as aggregates, J. Environ. Management 90 (2009)
3436–3442.
56. M. M. Ballari, M. Hunger, G. Hüsken, H. J. H. Brouwers,
Heterogeneous photocatalysis applied to concrete pavement for air
remediation, in Nanotechnology in Construction: Proceedings of
the NICOM3, edited by Z. Bittnar, P. M. J. Bartos, J. Nĕmecĕk, V.
Šmilauer, J. Zeman, Springer, Berlin, Germany, 2009; pp.409–414.
57. http://gfp.lbl.gov/.
58. J. C. Harper, A. F. El Sahrigi, Thermal conductivities of gas-filled
porous solids, Ind. Engr. Chem. Fundamentals 3 (1964) 318–324.
59. R. Baetens, B. P. Jelle, J. V. Thue, M. J. Tenpierik, S. Grynning, S.
Uvsløkk, A. Gustavsen, Vacuum insulation panels for building appli-
cations: A review and beyond, Energy Buildings 42 (2010) 147–172.
60. H. Cauberg, M. Tenpierik, Vacuümisolatiepanelen en andere
noviteiten, in Praktijkhandboek Duurzaam Bouwen, edited by D. W.
Dicke, E. M. Haas, Weka Publishers, Amsterdam, the Netherlands;
pp. 4.1 VAC 1–19.
61. J. Fricke, Ed., Aerogels, Springer Proceedings in Physics, Vol. 6,
Springer, Berlin, Germany, 1986.
62. M. Reim, W. Körner, J. Manara, S. Korder, M. Arduini-Schuster,
H.-P. Ebert, J. Fricke, Silica aerogel granulate material for thermal
insulation and daylighting, Solar Energy 79 (2005) 131–139.
63. A. Soleimani Dorcheh, M. H. Abbasi, Silica aerogel: Synthesis,
properties and characterization, J. Mater. Proc. Technol. 199 (2008)
10–26.
64. P. H. Tewari, A. J. Hunt, J. G. Lieber, K. Lofftus, Microstructural
properties of transparent silica aerogels, in Aerogels, edited by J.
Fricke, Springer Proceedings in Physics, Springer, Berlin, Germany,
1986; Vol. 6, pp. 142–147.

© 2011 by Taylor and Francis Group, LLC

 398 Green Nanotechnology

65. C. G. Granqvist, Energy-efficient windows: Present and forthcom-

ing technology, in Materials Science for Solar Energy Conversion
Systems, edited by C. G. Granqvist, Pergamon, Oxford, U.K., 1991,
pp.106–167.
66. G. Marsh, RE storage, the missing link, Refocus 3(2) (2002)
38–42.
67. Y. Zhang, H. Feng, X. Wu, L. Wang, A. Zhang, T. Xia, H. Dong,
X. Li, L. Zhang, Progress of electrochemical capacitor elec-
trode materials: A review, Int. J. Hydrogen Energy 34 (2009)
4889–4899.
68. D. Sadoway, Liquid-battery: Research in energy storage (2009);
http://mit.edu/dsadoway/www/.
69. R. Kötz, M. Carlen, Principles and applications of electrochemical
capacitors, Electrochim. Acta 45 (2000) 2483–2498.
70. P. Simon, A. Burke, Nanostructured carbons: Double-layer capaci-
tance and more, Electrochem. Soc. Interface (Spring) (2008) 38–43.
71. M. B. Cortie, A. I. Maaroof, G. B. Smith, Electrochemical capaci-
tance of mesoporous gold, Gold Bull. 38 (2005) 14–22.
72. A. K. Geim, K. S. Novoselov, The rise of graphene, Nature Mater. 6
(2007) 183–191.
73. M. D. Stoller, S. Park, Y. Zhu, J. An, R.S. Ruoff, Graphene-based
ultracapacitors, Nano Lett. 8 (2008) 3498–3502.
74. A. S. Aricò, P. Bruce, B. Scrosati, J.-M. Tarascon, W. van Schalkwijk,
Nanostructured materials for advanced energy conversion and stor-
age devices, Nature Mater. 4 (2005) 366–377.
75. A. Fedorková, H.-D.Wiemhöfer, R. Oriňáková, A. Oriňák, M. C.
Stan, M. Winter, D. Kaniansky, A. N. Alejos, Improved lithium
exchange at LiFePO4 cathode particles by coating with composite
polypyrrole–polyethylene glycol layers, J. Solid State Electrochem.
13 (2009) 1867–1872.
76. B. Kang, G. Ceder, Battery materials for ultrafast charging and dis-
charging, Nature 458 (2009) 190–193.
77. G. Nyström, A. Razak, M. Strømme, L. Nyholm, A. Mihranyan,
Ultrafast all-polymer paper-based batteries, Nano Lett. 9 (2009)
3635–3639.
78. Y. J. Lee, H. Yi, W.-J. Kim, K. Kang, D. S. Yun, M. S. Strano, G.
Ceder, A. M. Belcher, Fabricating genetically engineered high-power
lithium-ion batteries using multiple virus genes, Science 324 (2009)
1051–1055.

© 2011 by Taylor and Francis Group, LLC

399
400 Green Nanotechnology
 Conclusions 401
402 Green Nanotechnology
 Conclusions 403
404 Green Nanotechnology
 Conclusions 405
406 Green Nanotechnology
 Conclusions 407
408 Green Nanotechnology
 Conclusions 409
410 Green Nanotechnology
 Conclusions 411
413
414 Appendix 1: Thin-Film Deposition
 Appendix 1: Thin-Film Deposition 415
416 Appendix 1: Thin-Film Deposition
 Appendix 1: Thin-Film Deposition 417
418 Appendix 1: Thin-Film Deposition
 Appendix 1: Thin-Film Deposition 419
420 Appendix 1: Thin-Film Deposition
 Appendix 1: Thin-Film Deposition 421
422 Appendix 1: Thin-Film Deposition
 Appendix 1: Thin-Film Deposition 423
424 Appendix 1: Thin-Film Deposition
 Appendix 1: Thin-Film Deposition 425
426 Appendix 1: Thin-Film Deposition
 Appendix 1: Thin-Film Deposition 427
 Appendix 2: Abbreviations,
Acronyms, and Symbols
Abbreviations, acronyms, and symbols used only in the immediate vicin-
ity of where they are defined are not given here.

A Absorptance
A Surface area
AGD Advanced gas deposition
A H Hemispherical absorptance
A lum Luminous absorptance
AM Air mass
Aq Oscillator strength
Asol Solar absorptance
ac Alternating current
α Absorption coefficient
α Deposition angle
αpol Polarizability

BAU Business as usual

BIPV Building integrated photovoltaics
β Column inclination angle

C Capacitance
C Cost
C abs Absorption cross section
CB Conduction band
CFL Compact fluorescent light

429

© 2011 by Taylor and Francis Group, LLC

 430 Appendix 2: Abbreviations, Acronyms, and Symbols

COP Coefficient of performance

C scatt Scattering cross section
CVD Chemical vapor deposition
c Speed of light

D Dimension
DSSC Dye-sensitized solar cell
d Film thickness
dc Critical thickness (for percolation)
dc Direct current
ddc Thickness of double layer
deq Equivalent film thickness
di Insulator thickness
Δt Time correction
Δτ Temperature difference

E Electric field strength

E Emittance
E a Atmospheric emittance
EAA Ethylene acrylic acid
E appl Applied electric field strength
E a,sw Atmospheric emittance in the sky window
EBA Ethylene butyl acrylate
EC Electrochromic
Ecb Conduction band energy
E g Semiconductor band gap
E H Hemispherical emittance
Ein Input energy
Em Emittance of heat mirror
EMA Effective medium approximation
EMA Ethyl methacrylate
EMMA Ethylene methacrylic acid
E s Surface (substrate) emittance
ESL Electron-stimulated luminescence
ETFE Ethylene tetra fluoro ethylene
EVA Ethylene vinyl acetate
e Electron charge
e-beam Electron-beam (deposition)

© 2011 by Taylor and Francis Group, LLC

 Appendix 2: Abbreviations, Acronyms, and Symbols 431

ε Dielectric function (= ε1 + iε2)

ε* Effective dielectric function
εh Dielectric function of host material
εp Dielectric function of particle
ε0 Permittivity of free space
ε∞ High-frequency dielectric constant

f Filling factor
f Frequency
fc Critical filling factor
Φ Energy (or light) flux

GBO Giant birefringent optics

g Acceleration due to gravity
ΓL Luminous intensity
ΓR Radiant intensity

H Height
H Magnetic field strength
HOMO Highest occupied molecular orbital
h Planck’s constant
hcp Hexagonal close packed
hν Photon energy
η Efficiency
ηCA Curzon–Ahlborn efficiency
ηCarnot Carnot efficiency

I Intensity
I D Diffuse intensity
Ipb Parallel beam intensity
IPCE Incident photon to current conversion efficiency
Iref Reflected intensity
IR Infrared

© 2011 by Taylor and Francis Group, LLC

 432 Appendix 2: Abbreviations, Acronyms, and Symbols

Isol Solar intensity

ITO Indium tin oxide (In 2O3:Sn)

J (Julian) day

K Luminous efficacy
k Extinction coefficient
k* Effective extinction coefficient
k B Boltzmann’s constant
kτ Thermal conductivity
k0 Wave vector

L Depolarization factor
L Lifetime
L Longitude
LCD Liquid crystal display
LD Lorentz–Drude
LED Light-emitting diode
LL Luminance
Lloc Local longitude
LOR Light output ratio (for luminaires)
LR Radiance
LSC Luminescent solar concentrator
LSP Localized surface plasmon
LUMO Lowest unoccupied molecular orbital
Lw Latent heat of water
L x,Ly,Lz Depolarization factors for axes x,y,z
λ Wavelength
λc Critical wavelength
λP Plasma wavelength

M Photon exitance
MG Maxwell Garnett

© 2011 by Taylor and Francis Group, LLC

 Appendix 2: Abbreviations, Acronyms, and Symbols 433

MMA Methyl methacrylate

MVOC Microbial volatile organic compound
MW Molecular weight
m* Effective electron mass
μ Magnetic permeability
μ0 Permeability of free space

N Generalized refractive index

NA Numerical aperture
Ne Electron density
NIR Near-infrared
n Refractive index
n* Effective refractive index
nh Refractive index of host material
np Refractive index of particle
ν Particle volume

OLED Organic light emitting diode

OSC Organic solar cell
OTEC Ocean thermal energy conversion
Ω Solid angle
ω Angular frequency
ωL Longitudinal phonon frequency
ωP Plasma frequency
ωP* Shielded plasma frequency
ωPH Phonon frequency
ωSPR Surface plasmon resonance frequency
ωT Transverse phonon frequency
ωτ Damping frequency

P Dipole moment
P Power
P(λ,τ) Planck radiant exitance
PC Polycarbonate
PCMB [6,6]-phenyl-C61-butyric acid methyl ester

© 2011 by Taylor and Francis Group, LLC

 434 Appendix 2: Abbreviations, Acronyms, and Symbols

Pd Dipole moment
PEN Poly ethylene naphthalate
PEO Poly ethylene oxide
PET Poly ethylene terephthalate
Pin Input power
PMMA Poly methyl methacrylate
Pout Output power
PV Photovoltaic
PVB Polyvinyl buteral
PVD Physical vapor deposition
P3HT poly(3-hexylthiophene)
p Aspect ration (of light pipe)
p Polarization direction
p(ω) Oscillating dipole moment
pa Per annum
pair Air pressure
Ψ Rotation angle
ϕ Angle (azimuthal)

Q Heat energy
QE Quantum efficiency
QH Heat removal

R Radius of random unit cell

R Reflectance
R Thermal resistance
Rb Reflectance from back side
Rd Rate of condensation
Rf Reflectance from front side
R g Universal gas constant
R H Hemispherical reflectance
R lum Luminous reflectance
R sol Solar reflectance
Rtherm Thermal reflectance
RUC Random unit cell
R◽ Sheet resistance
r Radius

© 2011 by Taylor and Francis Group, LLC

 Appendix 2: Abbreviations, Acronyms, and Symbols 435

ρ Resistivity
ρ(ω) Dynamic resistivity

S Seebeck coefficient
S(λ) Solar spectrum
SBS Sick building syndrome
SFR Skylight to floor area ratio
SPP Surface plasmon polariton
SPR Surface plasmon resonance
SSL Solid-state lighting
SWNT Single-walled carbon nanotube
s Polarization direction
sw Sky window (subscript)
Σ Gas sensitivity
σ Electrical conductivity
σ* Effective electrical conductivity
σair Electrical conductivity in air
σgas Electrical conductivity in gas
σSB The Stefan–Boltzmann constant

T Transmittance
Ta Atmospheric transmittance
Ta,sw Atmospheric transmittance in the sky window
TCO Transparent conducting oxide
TIR Total internal reflectance
Tlum Luminous transmittance
TRIMM Transparent index matched microparticle
Tsol Solar transmittance
t Time
τ Temperature
τa Ambient (atmospheric) temperature
τc Critical temperature
τc Output temperature (cold)
τdp Dew point temperature
τe Carrier lifetime
τh Input temperature (hot)
τs Surface (substrate) temperature

© 2011 by Taylor and Francis Group, LLC

 436 Appendix 2: Abbreviations, Acronyms, and Symbols

τs,min Radiative stagnation temperature

θ Angle (polar)
θcrit Critical angle
θwc Water contact angle
θz Zenith angle
θz,max Acceptance value for zenith angle

U Heat transfer coefficient (U-value)

U Voltage
UHI Urban heat island
Unr Nonradiative heat transfer coefficient
UV Ultraviolet

V Volume
VB Valence band
VIP Vacuum insulation panel
VOC Volatile organic compound
υF Fermi velocity

WOLED White organic light emitting diode

Y Admittance
Y(λ) Spectral response of the eye
Y0 Admittance of free space

Z Thermoelectric performance factor

z Distance
ζ Number of reflections

© 2011 by Taylor and Francis Group, LLC

 Temperature (°C)
0.5 (a) Global mean temperature 14.5
(°C)

0.0 14.0
400
Trend=R1.4
–0.5 13.5 350

(mm/day)
50 (b) Global average sea level 300
0 250
(mm)

–50
–100 200
–150
150
1850 1900 1950 2000 1960 1970 1980 1990 2000
Year Year

(a) (b)
62

60

58
Lattitude (°N)

56 B
A

54 C

52

50
–4 –2 0 2 4 6 8
(c)
Longitude (°E)

(d)

(e) (f )

Figure P1  Panorama over climate changes and some of their impacts. Panel (a) shows
increasing global mean temperature and associated rise of global average sea level since
1850. (See text for full caption and credits.)

© 2011 by Taylor and Francis Group, LLC

 (a) (b)

(c) (d)

(e) (f )

Figure P2  Panorama over some aspects of green nanotechnologies. Panel (a) shows a
low-energy commercial building in San Francisco with energy-efficient glazing and natu-
ral ventilation. (See text for full caption and credits.)

© 2011 by Taylor and Francis Group, LLC

 Figure 1.2  Images at successive magnifications of a green leaf. The linear scale in
neighboring images falls by a factor of ten, starting at 1 cm and ×10 magnification and
ending at 1 nm and ×100 million magnification. The last three images are in the domain
of nanotechnology, with the final one showing molecular orbitals. (From http://micro.
magnet.fsu.edu/primer/java/scienceopticsu/powersof10/index.html. Reprinted with per-
mission from Dr. Mike Davidson, Florida State University National Magnet Laboratory.)

Figure 1.3  Examples of brilliantly colored nanoporous systems whose internal pattern
of holes and solid matter leads to unique optical effects. The photos show the external color
on the left and the associated nanostructure on the right; they refer to “peacock eye” but-
terfly wings (upper) and a famous opal gem, “the flame queen” (lower). (Images used with
permission from the following sources: Upper left: The Peacock Eye butterfly © Copyright
Lynne Kirton [image licensed under the Creative Commons Licence. http://creativecom-
mons.org/licenses/by-sa/2.0/]; upper right: H. Ghiradella, Appl. Opt. 30, 3492–3500
(1991); lower panel: (c) en.wikipedia.org/wiki/Flame_Queen_Opal [image licensed under
Creative Commons ShareAlike 3.0 en.wikipedia.org/wiki/Creative_Commons].)

© 2011 by Taylor and Francis Group, LLC

 (a) World investment in renewable energy
each year from 1995 to 2004
35
30
25

Billion Dollars
20
15
10
5

1995 1996 1997 1998 1999 2004 2004 2004 2004 2004

(b) Composition and capacity of renewable power

in the world, developing world, Europe and
countries which have installed the most
175
150
Solar PV (grid)
125 Geothermal
Biomass
Gigawatts

100
Wind
75 Small Hydro
50
25
0
World Developing EU-25 China Germany United Spain Japan
World States

Figure 1.5  Panel (a) reports annual world investment (in billions of U.S. dollars) in
renewable energy since 1995. Panel (b) shows renewable power capacities in GW for the
world and for select groups of nations and individual countries in 2004. (Adapted from
REN21 [Renewable Energy Policy Network for the 21st Century], Renewables 2005: Global
Status Report, The Worldwatch Institute, Washington, DC, 2005; http://www.ren21.net.)

Figure 2.1  This oil painting on canvas by Prince Eugen of Sweden (1865–1947) dates
from 1895 and is known as The Cloud. It is in the collection of the Gothenburg Museum
of Art in Sweden. (From http://www.waldemarsudde.se/xsaml_molnet_g.html. Photo by
Lars Engelhardt. Reprinted with permission.)

© 2011 by Taylor and Francis Group, LLC

 0.75 μm 2.5 μm 30 μm 8.0 μm 13.0 μm
100%
100%
Sky
UV UV
window

Visible Visible

Reflectance
Reflectance

Thermal radiation Thermal radiation

Solar Solar

NIR NIR

0% 0%
Wavelength
(a) (b)

Ideal cool paint Ideal solar absorber

0.75 μm 2.5 μm 30 μm
100%

UV

Visible
Reflectance

Thermal radiation
Solar

NIR

0%
(c)

Colored paint Colored paint

warm climate cold climate

Figure 2.15  Ideal spectral reflectance for opaque surfaces. Panel (a) refers to a solar
absorber and a surface (paint) for cooling applications; the sharp transition is at ~2.5
μm in wavelength. Panel (b) shows ideal spectral reflectance for maximizing radiation
loss to the clear sky; the reflectance is 100% except where the atmosphere is transpar-
ent (between 8 and 13 μm). Panel (c) applies to colored surfaces (paints) that look the
same but have widely different thermal performance; the properties are ideal for warm
and cold climates. Ranges for ultraviolet (UV), visible, near-infrared (NIR), and thermal
radiation are shown.

© 2011 by Taylor and Francis Group, LLC

 0.75 μm 2.5 μm 30 μm 0.75 μm 2.5 μm 30 μm
100% 100%

UV UV
Visible Visible

Transmittance

Transmittance
Thermal radiation Thermal radiation
Solar Solar

NIR R≈1 NIR R≈1

E≈0 E≈0
0% 0%
Wavelength
(a) (b)

Figure 2.16  Ideal spectral transmittance for windows in (a) cold and (b) hot climates;
the dotted option at visible wavelengths is for glare reduction. Ranges for ultraviolet (UV),
visible, near-infrared (NIR), and thermal radiation are shown.

(a) (b)

Figure 3.1  Two pieces of glassware, both doped with gold nanoparticles. Part (a) is the
Roman Lycurgus cup from the British Museum and part (b) is an old Victorian vase owned
by one of the authors (GBS). [Part (a) is from http://www.britishmuseum.org/explore/high-
lights/highlight_objects/pe_mla/t/the_lycurgus_cup.aspx. Reproduced with permission.]

N
θ S

29°

100 mm

(a) (b)

Figure B3.1.1  Magnetochromics based on nanoparticle arrays and magnetically

rotatable photonic microspheres. Part (a) shows the experimental set-up, and part (b)
illustrates the color changes that occur as the angle θ has different values. (From J. Ge et
al., J. Am. Chem. Soc. 131 [2009] 15687–15694. With permission.)

© 2011 by Taylor and Francis Group, LLC

 Figure 4.41  Scanning near-field micrographs of 4 x 4 μm areas of VO2
Figure 4.34  Color shifts, caused by a range of viewing angles, in reflected films as meas­u red at the shown temperatures. Dark blue represents an insu-
light from a photonic crystal such as the one depicted in Figure 4.33. Data are lating state and light blue, green, and red colors represent metallic conduc-
shown for different particle sizes. (R. J. Leyrer, personal communication to one tivity. (From M. M. Quazilbash et al., Science 318 [2007] 1750–1753. With
of the authors [GBS]. Figure courtesy of BASF—The Chemical Company.) permission.)

© 2011 by Taylor and Francis Group, LLC

 Figure 5.4  Concentrated illuminance from clear skylights in a supermarket aisle
(left-hand panel) and uniform illuminance in another supermarket with diffuse sky-
lights (right-hand panel). The skylights are most easily seen for diffuse lighting. (After J.
McHugh et al., Visible light transmittance of skylights, California Energy Commission,
PIER, 2004; Contract Number 400-99-013. With permission.)

c
b
a

3.0 µm

Figure 5.16  The left-hand panel shows a cross section of a one-dimensional photo-
nic polymer crystal constructed from alternating layers; dark-colored layers are PMMA
and light-colored layers are (birefringent) PET. The right-hand panel shows relative light
transport along mirror light pipes of aspect ratio 17 including a foil of the shown material
(a) and including specular reflecting aluminum (b) and silver (c) films. (From M. F. Weber
et al., Science 287 [2000] 2451–2456. With permission.)

© 2011 by Taylor and Francis Group, LLC

 Illumination
spot

Edge exit
light

Figure 5.25  Light output from a TRIMM-doped PMMA sheet.

Figure 5.26  Flexible transparent polymer light guide containing transparent index
matched microparticles. An integrating sphere (cf. Section 2.4.2) is used to measure the
total light output at different segments along the guide.

Figure 5.28  Two applications of colored LEDs. (From K. Dowling, LED Essentials,
Department of Energy, Webinar October 10 [2007]; http://apps1.eere.energy.gov/build-
ings/publications/pdfs/ssl/webinar_2007-10-11.pdf.)

© 2011 by Taylor and Francis Group, LLC

 (a)

(b) (c)

Figure 6.1  Photos illustrating building-integrated devices for photothermal and

photoelectric conversion of solar energy. Panel (a) shows a roof-mounted solar collec-
tor arrangement from Sweden, used to heat a number of private houses; panel (b) is a
solar-cell-covered façade at Solar-Fabrik in Freiburg, Germany; and panel (c) is a solar
cell installation forming an architectural element on a building belonging to Fraunhofer
Institute of Solar Energy Systems in Freiburg, Germany.

100 100
Asol = 0.86 Asol = 0.81
80 Etherm = 0.04 80 Etherm = 0.06
Reflectance (%)

Reflectance (%)

60 60
575 nm
40 683 nm 40

20 20

0 0
0.2 0.5 1 2 5 10 20 50 0.2 0.5 1 2 5 10 20 50
Wavelength (µm) Wavelength (µm)
100 100
Asol = 0.92 Asol = 0.86
80 Etherm = 0.08 80 Etherm = 0.08
Reflectance (%)

Reflectance (%)

60 60
40 515 nm 40 463 nm
20 20

0 0
0.2 0.5 1 2 5 10 20 50 0.2 0.5 1 2 5 10 20 50
Wavelength (µm) Wavelength (µm)

Figure 6.12  Spectral reflectance for TiAlOxNy films made by sputter deposition onto
Al. Different colors (shown) were obtained by selecting film thicknesses leading to reflec-
tance maxima at the wavelengths shown in the luminous wavelength range. Corresponding
values of A sol and E therm are stated. (From D. Zhu, S. Zhao, Solar Energy Mater. Solar
Cells [2010] in press. With permission.)

© 2011 by Taylor and Francis Group, LLC

 0.4
60 nm

Plasmon induced aborption

40 nm
0.3 20 nm
10 nm
5.0 nm
0.2
2.5 nm

0.1

0
0.001 0.01 0.1 1
PV layer extinction coefficient

Figure 6.19  Absorption induced in various thicknesses of very thin Si by neighboring

Ag nanoparticles (peaked curves to right). The x-axis is the absorption coefficient of the
Si material. The particles themselves absorb according to the plots with crosses. (From C.
Hägglund, B. Kasemo, Opt. Express 17 [2009] 11944–11957. With permission.)

1.0
Reflectance

0.8
0.6
0.4
0.2
0.0
5
6 80 90
7
8 60 70
Wav
elen
9
10 40 50 nce (°)
30 de
gth ( 11 20 Inci
µm) 12
13 0 10
n g le of
A

Figure 7.12  Spectral and angular-dependent reflectance for nanoparticles of SiO2 and
c-SiC embedded in polyethylene foil. (From A. R. Gentle, G. B. Smith, Nano Lett. 10
[2010] 373–379. With permission.)

© 2011 by Taylor and Francis Group, LLC

 Figure 8.1  Silica aerogel tile. (From http://www.airglass.se. With permission.)

(a) (b)

Figure 8.16  Structure of the nanostructured polymer ion conductor PEO6:LiAsF6 as

seen looking into the chains and tunnels (a) and along them (b). The Li+ (blue) sits inside
the tunnels formed by the PEO. The atoms are fluorine (magenta), carbon (green), and oxy-
gen (red). (From A. S. Aricò et al., Nature Mater. 4 [2005] 366–377. With permission.)

© 2011 by Taylor and Francis Group, LLC