RESEARCH ARTICLE | FEBRUARY 20 2020

Surface science of shape-selective metal nanocrystal

synthesis from first-principles: Growth of Cu nanowires and
nanocubes 
Special Collection: 30 years of the Nellie Yeoh Whetten Award — Celebrating the Women of the AVS

Kristen A. Fichthorn ; Zihao Chen

J. Vac. Sci. Technol. A 38, 023210 (2020)

https://doi.org/10.1116/1.5141995
 CHORUS


View Export
Online Citation

29 August 2024 16:10:58

 ARTICLE avs.scitation.org/journal/jva

Surface science of shape-selective metal

nanocrystal synthesis from first-principles: Growth
of Cu nanowires and nanocubes
Cite as: J. Vac. Sci. Technol. A 38, 023210 (2020); doi: 10.1116/1.5141995
Submitted: 9 December 2019 · Accepted: 31 January 2020 · View Online Export Citation CrossMark
Published Online: 20 February 2020

Kristen A. Fichthorna) and Zihao Chen

AFFILIATIONS
Department of Chemical Engineering, The Pennsylvania State University, 403 Chemical and Biomedical Engineering Bldg.,
University Park, Pennsylvania 16802

Note: This paper is part of the Special Topic Collection on 30 years of the Nellie Yeoh Whetten Award — Celebrating the Women
of the AVS.
a)
Electronic mail: [email protected]

ABSTRACT
The authors present results from first-principles density functional theory aimed at understanding the aqueous solution-phase growth of

29 August 2024 16:10:58

fivefold twinned copper nanowires and single-crystal nanocubes capped by hexadecylamine (HDA). The role of solution-phase chloride,
present in the Cu salt or as an additive, is emphasized. Using ab initio thermodynamics, the authors delineate the range of solution-phase
conditions, characterized by the chemical potentials of chloride and HDA, under which Cu nanowires and nanocubes can be grown. The
authors discuss the likelihood of thermodynamic and/or kinetic nanostructures for various solution-phase concentrations. Their results
are in good agreement with experiments and indicate that methods and insights developed for surface science in gas-phase or vacuum
conditions can yield much insight into liquid-phase systems.

Published under license by AVS. https://doi.org/10.1116/1.5141995

I. INTRODUCTION increasingly rapid pace, with numerous recent reports on the syn-
A significant challenge in the development of functional nano- thesis of various beneficial nanocrystal morphologies.19–22
materials is understanding the growth and transformations of col- Despite ample demonstrations that it can be highly beneficial
loidal metal nanocrystals. From a practical perspective, knowledge to tune nanocrystal morphologies for specific applications and
of how to selectively synthesize desired metal nanocrystal sizes and despite the tremendous strides made in nanocrystal synthesis
shapes would benefit numerous applications. For example, catalyst science, it is still difficult to achieve high, selective yields in most
nanocrystal synthesis science is playing an increasingly prominent synthesis protocols. Many fundamental aspects of these complex
role in electrocatalysis and photocatalysis.1–7 Metal nanocrystals syntheses remain poorly understood and fundamental studies can
figure prominently in solar-cell technologies as nanometallic plas- be beneficial. Since the shapes of metal nanocrystals are largely
monic structures to enhance the light absorption and efficiency of governed by phenomena occurring at their surfaces, fundamental
photovoltaics,8 as nonplasmonic photosensitizers for light-energy studies based on principles in surface science can be particularly
conversion,9 and as nanowire-based transparent conducting useful.
films.10–13 Metal nanocrystals are also featured in emerging applica- Theory and simulation can aid in efforts to understand these
tions targeting energy efficiency, such as photothermic desalination syntheses. The chemical complexities in these systems practically
based on plasmonic nanocrystals,14,15 wearable triboelectric genera- mandate that theoretical descriptions should be based on quantum
tors,16 flexible energy storage devices,17 and electrochromic smart mechanics. Indeed, quantum mechanical studies, particularly those
windows.18 For these and many other applications, the science of based on density functional theory (DFT), have been useful in
shape-selective nanocrystal synthesis has been advancing at an understanding many aspects of these syntheses,23–31 despite the

J. Vac. Sci. Technol. A 38(2) Mar/Apr 2020; doi: 10.1116/1.5141995 38, 023210-1
Published under license by AVS.
 ARTICLE avs.scitation.org/journal/jva

fact that such calculations are done at zero Kelvin and in vacuum present the results of new calculations, which more tightly delineate
while experiments occur in the liquid phase at room temperature the range of solution-phase conditions under which Cu nanowires
and above. can be grown. We extend our results to explain the experimentally
Below, we discuss our theoretical, DFT-based studies of the observed growth of Cu nanocubes under similar synthesis condi-
growth of fivefold twinned Cu nanowires and nanocubes. Fivefold tions.34 Based on our results, we gain insight into other possible
twinned metal nanowires are fascinating objects that, because of mechanisms that might be exploited to gain control of nanocyrstal
their high conductivity, optical transparency, flexibility, and their morphology and we discuss these.
ability to be grown with aspect ratios ranging from around 100–
1000, are currently components in commercial touch screens.10 II. BACKGROUND
Figure 1 depicts the fivefold twinned geometry observed for various
We consider the growth of Cu nanowires in aqueous solution
fcc metal nanowires. These nanowires possess a high
containing alkylamines, and HDA in particular, as capping agents.
surface-area-to-volume ratio, so their shape is not favored thermo-
When Cu(NO3)2 salt is used as the source of Cu, nanowires do not
dynamically and they are referred to as kinetically controlled
form. However, nanowires form when the salt is CuCl2 and they
shapes. For pentagonally twinned nanowires of Cu, Ag, and Au
also form when Cu(NO3)2 is the salt and NaCl is added.30 In this
with {100} facets on their sides and {111} facets on their ends,
latter case, the synthesized morphology is a function of the NaCl
researchers have hypothesized that organic capping agents direct
(Cl−) concentration: If the Cl− concentration is too low, spherical
anisotropic growth by selectively inhibiting atomic addition to the
nanoparticles form, and if the Cl− concentration is too high, the
{100} facets.32 However, there is little experimental evidence to
nanowires grow laterally and achieve a nonuniform cross section,
support the hypothesis that organic capping agents bind to {100}
creating javelinlike wires.30 Nanowires with high aspect ratios form
facets with a sufficient selectivity to create nanowires and recent
over an intermediate range of Cl− concentrations.
theoretical calculations concluded that this scenario is unlikely.33 In
Electrochemical experiments on single-crystal Cu(111) and Cu
addition, this hypothesis does not explain why most syntheses
(100) surfaces indicate that there is little difference in the rates at
require the presence of halide ions to grow Cu, Ag, and Au nano-
which Cu ions are reduced on the two Cu surfaces at low concen-
wires with high aspect ratios.30
trations of chloride—both reduction rates are low. At intermediate
Our recent work has focused on resolving the mechanisms of
Cl− concentrations, where nanowires are observed, the rate of
fivefold twinned nanowire growth using first-principles DFT calcu-
Cu-ion reduction is larger on Cu(111) than on Cu(100). High
lations.30 These studies, in conjunction with nanocrystal synthesis
Cu-ion reduction rates are observed on both surfaces at high chlo-
and electrochemical experiments on single-crystal surfaces, demon-

29 August 2024 16:10:58

ride concentrations, where the wires take on a javelinlike morphol-
strate that there are synergistic interactions between adsorbed Cl
ogy. Thus, experiments identify that chloride is needed for the
and capping agent hexadecylamine (HDA) that can direct most of
growth of Cu nanowires, its concentration has to fall within a
the solution-phase Cu-ion flux to the ends of a growing Cu nano-
certain window for nanowire growth to occur, and the resulting
wire to enhance the aspect ratio.30 Below, we review this work and
morphology can be connected to reduction rates observed in
single-crystal electrochemical experiments.30

III. METHODS
Our DFT calculations can explain most of the above experi-
mental trends. All DFT calculations were performed using the
Vienna Ab initio Simulation Package (VASP)35–37 with the projec-
tor augmented-wave method.38,39 The exchange-correlation func-
tional by Perdew, Burke, and Ernzerhof40 was employed for all
calculations. For structural optimizations, convergence criteria of
10−6 eV for energy and 0.02 eV/Å for forces were used along with
an energy cutoff of 450.0 eV. The DFT-D2 method by Grimme41
was applied to account for van der Waals (vdW) interactions, with
parameters for Cu suggested by Ruiz et al. to include the screening
effects of bulk Cu.42
A periodic unit cell, which consists of a Cu surface modeled
by a six-layer slab and adsorbates on the top-most Cu layer, was
used to study the coadsorption of Cl− and HDA on the Cu sur-
faces. The lower three layers were fixed at bulk positions and all
other atoms were fully relaxed. To avoid interactions between peri-
odic, neighboring unit cells, a vacuum space greater than 35 Å was
introduced and a dipole correction was applied along the surface
normal direction. Convergence tests were performed with respect
FIG. 1. Fivefold twinned nanowire geometry with {111} end facets (dark) and
{100} side facets (light).
to the k-point mesh, the energy cutoff, and the vacuum space.
Details regarding convergence tests and the unit cells used with

J. Vac. Sci. Technol. A 38(2) Mar/Apr 2020; doi: 10.1116/1.5141995 38, 023210-2
Published under license by AVS.
 ARTICLE avs.scitation.org/journal/jva

corresponding k-point meshes can be found in the supporting per Cu atom, NCl is the number of adsorbed Cl atoms, μCl
is the
information of Ref. 30. chemical potential of aqueous chloride, NHDA is the number of
adsorbed HDA molecules, μHDA is the chemical potential of
IV. RESULTS aqueous HDA, and Asurf is the surface area of the slab. γ fixed
Cu is the
surface energy of a bare Cu surface slab with atoms fixed at the
Our DFT calculations indicate that in the absence of adsorbed bulk coordinates, which we subtract as a result of performing our
Cl, HDA forms self-assembled monolayers (SAMs) that bind to calculations with halogen and HDA adsorbed on only one side of
Cu(100) slightly more strongly (by 0.12 eV/molecule) than to the slab.
Cu(111).27,43 However, this difference in binding energies renders In the absence of Cl, HDA forms passivating self-assembled
the surface energies of Cu(100) and Cu(111) virtually equal27 and monolayers on both surfaces,27,43,44 as shown in Figs. 2(a) and 2(b).
is an insufficient thermodynamic driving force for anisotropic The DFT surface energies for HDA-covered Cu(100) and
growth. In addition, the calculated binding free energy of HDA to Cu(111) are 0.087 and 0.079 eV/Å2, respectively. Using these
both Cu surfaces in solution has been estimated to be ∼60 kT, surface energies, we predict that the Wulff shape of an
meaning that the probability for spontaneous desorption of HDA HDA-functionalized Cu nanocrystal is a truncated octahedron,45
is vanishingly small. It is unclear how Cu nanowires can grow if consistent with the experimental observation of spherelike parti-
strongly adsorbed HDA passivates Cu surfaces and prevents elec- cles in the absence of Cl.30
trochemical reactions from occurring, as has been observed experi- At a Cl coverage of 0.25 ML, HDA chemically adsorbs to both
mentally.44 Thus, our calculations indicate that HDA, alone, cannot Cu surfaces. As we see in Figs. 3(a) and 3(b), HDA forms SAMs, in
mediate the growth of Cu nanowires. which it binds to both Cu(100) and Cu(111) via the nitrogen atom
We considered a possible synergistic interaction between chlo- in the amine group. Figure 2 shows that in DFT calculations of
ride and HDA in promoting the growth of Cu nanowires. To do HDA SAMs in the absence of adsorbed Cl, N prefers to bind to the
this, we focused on delineating the surface environments that are atop sites on both surfaces and this preference is maintained in the
likely to be present on Cu(100) and Cu(111) for various solution presence of adsorbed Cl. Increasing the Cl coverage to 0.33 ML, we
concentrations (chemical potentials) of Cl− and HDA and connect- see in Fig. 3(c) that Cu(100) can accommodate both Cl and HDA,
ing these to experiments. with HDA residing near atop sites. In contrast, Fig. 3(d) shows that
The preferred surface configuration is characterized by its HDA moves away from Cu(111), likely due to the short-range
surface energy, where the surface energy for a Cu slab containing repulsion that occurs as strongly bound Cl atoms “crowd out”
Cl and HDA γ Cu
Cl
HDA is given by HDA.30 HDA forms a physically adsorbed layer above the Cl

29 August 2024 16:10:58

surface layer in this configuration, in which the amine is attracted
(ECu
Cl
HDA
NCu ECu
bulk

NHDA μHDA
NCl μCl
)
γ Cu
Cl
HDA ¼ to the substrate by vdW interactions. Though the HDA layer stays
Asurf near the surface in the 0 K vacuum environment of DFT calcula-

γ fixed tions, we expect these weakly bound molecules to leave the surface,
Cu : (1)
perhaps forming a micelle, at nonzero temperatures and in the
Here, ECu
Cl
HDA is the DFT total energy of an optimized Cu solution phase. Finally, at a Cl coverage of 0.50 ML, the HDA
bulk
surface slab containing Cl and/or HDA, NCu , and ECu represent adlayers become weakly physisorbed on both surfaces, as shown in
the number of Cu atoms in the surface slab and the bulk energy Figs. 3(e) and 3(f ).

FIG. 2. Side view and top-down view of nitrogen atoms in HDA SAMS with a coverage of 0.25 ML on (a) Cu(100) and (b) Cu(111).

J. Vac. Sci. Technol. A 38(2) Mar/Apr 2020; doi: 10.1116/1.5141995 38, 023210-3
Published under license by AVS.
 ARTICLE avs.scitation.org/journal/jva

29 August 2024 16:10:58

FIG. 3. Snapshots of HDA SAMs: Side and top-down views (showing the N atoms only) for 0.25 ML HDA and Cl on (a) Cu(100) and (b) Cu(111); 0.33 ML HDA and Cl on
(c) Cu(100) and (d) Cu(111); and 0.25 ML HDA and 0.5 ML Cl on (e) Cu(100) and (f ) Cu(111). N atoms in chemisorbed HDA are dark and N atoms in physisorbed HDA
are light.

J. Vac. Sci. Technol. A 38(2) Mar/Apr 2020; doi: 10.1116/1.5141995 38, 023210-4
Published under license by AVS.
 ARTICLE avs.scitation.org/journal/jva

Using Eq. (1), or ab initio thermodynamics,46,47 we can TABLE I . Regions in Fig. 4 identified with fractional coverages of Cl and HDA on
delineate the range of solution-phase Cl− and HDA chemical Cu(100) and Cu(111).
potentials for which various {100} and {111} surface environ-
ments would be expected. We obtain the energetic terms for Cu(100) Cu(111)
Eq. (1) from DFT total energies for various coadsorbed configura- Region θHDA (ML) θCl (ML) θHDA (ML) θCl (ML)
tions of Cl− and HDA on the Cu surfaces, as discussed above.
1 0.0 0.0 0.0 0.0
Equation (1) yields the surface energy as a function of the
2 0.0 0.5 0.0 0.0
solution-phase chemical potentials of Cl− and HDA for each
3 0.25 0.25 0.0 0.0
coadsorbed configuration of Cl− and HDA. Considering various
4 0.25 0.0 0.0 0.0
coadsorbed configurations, we identify the configuration that
5 0.0 0.5 0.0 0.5
yields the lowest surface energy for each pair of solution-phase
6 0.25 0.25 0.0 0.5
chemical potentials. We conduct this procedure for both the Cu
7 0.33 0.33 0.0 0.5
(100) and Cu(111) surfaces and we construct a phase diagram by
8 0.25 0.25 0.25 0.25
identifying various unique chemical-potential regions of surface
9 0.25 0.25 0.25 0.0
configurations on the two surfaces.
10 0.25 0.0 0.25 0.0
Figure 4 shows the phase diagram of possible combinations of
surface compositions on Cu(100) and Cu(111)—surface structures
with the lowest surface energies given by Eq. (1)—as a function of
the solution-phase chemical potentials of Cl− and HDA. For the In Fig. 4 and Table I, we see that Cl and HDA are absent from
HDA chemical potential, we plot ΔμHDA ¼ μHDA
EHDA,opt
DFT
, where Cu(100) and Cu(111) at low chemical potentials (concentrations)
DFT
EHDA,opt is the DFT energy of an HDA SAM with the surface of the solution-phase species—this is region 1 in Fig. 4. Similarly,
configuration at different coverages, but in the absence of a solid in regions 2 (4, 10), we have a low chemical potential for HDA
surface. By plotting ΔμHDA , we remove the substantial influence of (chloride) and a high one for chloride (HDA), so that at least one
vdW interactions between alkyl tails in the HDA SAM. Since HDA of the surfaces is covered by chloride (HDA) alone. In region 8, we
should exist in a similar configuration within micelles in the solu- have moderate chemical potentials for both HDA and chloride and
tion phase, we approximately cancel the surface and solution-phase the surfaces contain both of these species. Regions 5, 6, and 7 rep-
HDA interactions by plotting ΔμHDA . Table I summarizes the frac- resent conditions for which adsorbed chlorine displaces HDA from
tional surface coverages of HDA and Cl on these surfaces in the a Cu surface, as we discussed for Fig. 3. Finally, from the point of

29 August 2024 16:10:58

various numbered regions in Fig. 4. view of nanowire growth, regions 3, 4, 6, and 7 are interesting
because HDA covers Cu(100) but not Cu(111) in these regions.
This selective HDA coverage of Cu(100) would allow for selective
Cu deposition at the {111} nanowire ends.
In experiments, the concentration of HDA is constant, while
the concentration of chloride is varied.30 To be consistent with
experimental observations, the chemical potential (concentration)
of HDA should be higher than the value delineated by the dashed
line in Fig. 4. The HDA concentration is not sufficiently high for it
to adsorb and protect both Cu surfaces at low concentrations of
chloride for HDA chemical potentials lower than those at the line.
We also note that the left end of the x axis in Fig. 4 (almost)
defines the upper limit of the chloride chemical potential, as CuCl2
salt is predicted to form at/above μCl
¼
2:43 eV.
Figure 5 shows a blow-up of the region for which the HDA
chemical potentials are higher than those delineated by the dashed
line in Fig. 4, along with experimental scanning electron micro-
scope (SEM) images, taken from Ref. 30, which could coincide with
the surface morphology in each region. In the experimental image
associated with region 10, the experimental concentration of chlo-
ride is zero. The rest of the experimental images indicate the
solution-phase concentration of chloride, which increases as μCl

increases.
FIG. 4. Phase diagram delineating regions of HDA and chloride chemical From Fig. 5, we see that there are several, experimentally con-
potentials for which various surface environments occur. For HDA, we plot sistent pathways that could be followed as μCl
increases at constant
ΔμHDA , as explained in the text. The fractional surface coverages of HDA and ΔμHDA. All pathways begin with region 10, where the concentration
Cl on Cu(100) and Cu(111) for each numbered region are given in Table I. of chloride is very low (or zero), there is no Cl adsorbed on the
Chemical-potential values for HDA that are greater than those delineated by the
surface, and both surfaces are covered with HDA. In this region,
dashed line are consistent with experiment.
Cu spheres are observed experimentally and the Cu-ion reduction

J. Vac. Sci. Technol. A 38(2) Mar/Apr 2020; doi: 10.1116/1.5141995 38, 023210-5
Published under license by AVS.
 ARTICLE avs.scitation.org/journal/jva

FIG. 5. Blow-up of the region for

which the HDA chemical potentials are
higher than those delineated by the
dashed line in Fig. 4, along with experi-
mental SEM images that we predict to
approximately coincide with the surface
morphology in each region. SEM
images are reprinted (adapted) with
permission from Kim et al., J. Am. Chem.
Soc. 140, 14740 (2018). Copyright 2018,
American Chemical Society.

rate is negligible—presumably due to the difficulty of Cu-ion trans- Figure 6 shows that the minimum-energy shape of a single-
port through the HDA SAM.30 From region 10, we progress to crystal Cu nanocrystal in the absence of HDA protection and
regions 9 and 8, where the concentration of chloride is low, both Cl− adsorption (region 1 in Fig. 4) is a truncated octahedron. For
surfaces are protected by HDA, and at least one surface also has moderate chloride chemical potentials and relatively high HDA
adsorbed chlorine. Here, we see the formation of both wires and chemical potentials (regions 6 and 7 in Fig. 4), truncated cubes are
particles in experiments, while the reduction rates on the single- predicted. In this region, Cu(100) is protected by HDA, but HDA
crystal Cu surfaces are still low. From region 8, one possibility is to is absent from Cu(111). Thus, in regions 6 and 7, it is possible that,

29 August 2024 16:10:58

access region 6, where Cu(100) is protected by HDA but Cu(111) is in addition to gaining a thermodynamic benefit for cube formation,
not (see Table I), upon further increasing the chloride concentra- there is the kinetic benefit that the Cu-ion flux to the {111} facets is
tion. This is the region for which wires are expected experimentally unimpeded, while that to the {100} facets is significantly hindered
and the reduction rate is high on Cu(111), but low on Cu(100) in
single-crystal experiments. By further increasing the concentration/
chemical potential of chloride, we reach region 5, where the con-
centration and surface coverage of Cl is high, HDA is displaced
from both surfaces, the Cu-ion reduction rates are observed to be
high on both surfaces, and javelinlike wires occur in experiments.
Two other possible pathways that could produce the same mor-
phology sequence as the solution-phase chloride concentration is
increased are 10 → 9 → 8 → 6→ 7 → 5 and 10 → 9 → 8 → 7 → 5.
Fivefold twinned nanowires grow as kinetic structures from deca-
hedral seeds.48 If the initial seed is a single crystal, then single-crystal
structures will form. For example, Jin et al. synthesized HDA-protected
Cu nanocubes beginning with single-crystal seeds.34 Thermodynamic
Wulff shapes can be predicted for single crystals based on their
surface energies.49 Since we obtain surface energies in our calcula-
tions, we are also able to predict single-crystal Wulff shapes.
Figure 6 shows a heat map of the ratio of surface energies of
Cu(111) to Cu(100) (i.e., γ {111} /γ {100} ) as a function of the chloride
and HDA chemical potentials. The surface-energy ratios were
obtained from the minimum surface energies, which led to the def-
inition of the ten regions in Fig. 4. Figure 6 also shows a few repre-
sentative Wulff shapes at various surface-energy ratios. In Fig. 6, FIG. 6. Ratios of the surface energies of Cu(111) to Cu(100) (i.e., γ {111} /γ {100} ),
the surface energies of Cu(100) and Cu(111) are the same in green given by the scale bar, as a function of the chemical potentials of chloride and
(medium gray) regions of the heat map, γ {111} . γ {100} in green to HDA. For HDA, we plot ΔμHDA , as explained in the text. Predicted Wulff shapes
are shown for selected values of γ {111} /γ {100} , where aqua (light) denotes {111}
red regions (left side and bottom-middle of Fig. 6) and γ {111} , γ {100}
facets and purple (dark) denotes {100} facets.
in green to purple regions (right and upper regions of Fig. 6).

J. Vac. Sci. Technol. A 38(2) Mar/Apr 2020; doi: 10.1116/1.5141995 38, 023210-6
Published under license by AVS.
 ARTICLE avs.scitation.org/journal/jva

by the HDA SAM. Molecular dynamics simulations aimed at lead to nanowires with aspect ratios (length-to-width ratios)
understanding the solution-phase growth of Ag nanocubes capped around 100.33 In this prior work, we predicted that nanowires can
with polyvinylpyrrolidone (PVP) demonstrated that small differ- grow (with a constant diameter) to a length at which the time scale
ences in the PVP binding affinity (surface density) on Ag(100) and for interfacet transport exceeds the deposition time. After this
Ag(111) could lead to kinetic nanocube formation.50–52 In the case point, the nanowires are predicted to grow thicker and this would
of HDA-capped Cu cubes grown in the presence of chloride, we most likely occur at the center of the nanowire. The javelinlike
predict that Cu(111) should contain little or no HDA, while Cu nanowire morphologies observed for high chloride concentrations
(100) should be fully protected, which is a much stronger kinetic in Ref. 30 (also see the inset to Fig. 5) are consistent with those
driving force than was observed for the Ag-PVP system. Thus, our that could occur in the model developed in Ref. 33.
calculations predict that cubes can also be formed, depending on the Thus, we use first-principles DFT calculations to provide
seed from which crystal growth originates, as is seen experimentally.34 insight into liquid-phase synthesis. Our results are in good agree-
ment with experiments and indicate that methods and insights
developed for surface science in gas-phase or vacuum conditions
V. DISCUSSION AND CONCLUSIONS can yield much insight into liquid-phase systems.
In this work, we used ab initio thermodynamics to predict the
surface environment of Cu(100) and Cu(111) facets during ACKNOWLEDGMENTS
solution-phase growth. Our calculations are consistent with single-
crystal electrochemical studies and synthesis experiments for Cu This work was supported by the Department of Energy, Office
nanowire growth in the presence of solution-phase chloride. of Basic Energy Sciences, Materials Science Division (No.
These experiments indicate that when Cu(100) and Cu(111) are DE-FG02-07ER46414). Z.C. also acknowledges training provided
both well-protected by HDA (at low solution-phase chloride con- by the Computational Materials Education and Training (CoMET)
centrations), spherical nanoparticles are synthesized and the NSF Research Traineeship (No. DGE-1449785).
reduction rate of Cu ions is similar (and low) on both surfaces.30
Increasing the chloride concentration, the reduction rate of Cu is REFERENCES
higher on Cu(111) than that on Cu(100) and high-aspect-ratio 1
P. Kamat, J. Phys. Chem. C 111, 2834 (2007).
nanowires are produced. Increasing the chloride concentration 2
S. Linic, P. Christopher, and D. Ingram, Nat. Mater. 10, 911 (2011).
even further leads to javelinlike nanowires with a nonuniform 3
H. Tong, S. Ouyang, Y. Bi, N. Umezawa, O. Mitsutake, and J. Ye, Adv. Mater.
cross section. In this regime, the reduction rate is similar (and 24, 229 (2012).

29 August 2024 16:10:58

4
high) on both surfaces. U. Aslam, S. Chavez, and S. Linic, Nat. Nanotechnol. 12, 1000 (2017).
5
When the concentration (chemical potential) of HDA is suffi- E. Kazuma and Y. Kim, Angew. Chem. Int. Ed. 58, 4800 (2019).
6
ciently high, our DFT calculations predict that both Cu(111) and M. Luo et al., Adv. Mater. 30, 1705515 (2018).
7
Cu(100) are covered by HDA SAMS at low chloride concentrations P. Strasser, M. Gliech, S. Kuehl, and T. Moeller, Chem. Soc. Rev. 47, 715
(2018).
(regions 10, 9, and 8 in Figs. 4 and 5). At intermediate chloride 8
H. Atwater and A. Polman, Nat. Mater. 9, 205 (2010).
concentrations, Cu(100) is covered by HDA SAMs while Cu(111) 9
M. A. Abbas, P. V. Kamat, and J. H. Bang, ACS Energy Lett. 3, 840 (2018).
is devoid of SAMs—regions 6 and 7 in Figs. 4 and 5. Both these 10
S. Ye, A. R. Rathmell, Z. Chen, I. E. Stewart, and B. J. Wiley, Adv. Mater. 26,
results are consistent with experimental electrochemical studies.30 6670 (2014).
Our DFT results can also explain why there is a similar (and high) 11
F. S. F. Morgenstern, D. Kabra, S. Massip, T. J. K. Brenner, P. E. Lyons,
Cu reduction rate on both Cu surfaces at high chloride and HDA J. N. Coleman, and R. H. Friend, Appl. Phys. Lett. 99, 183307 (2011).
12
concentrations, as HDA is absent from both Cu surfaces at these B. Li, S. Ye, I. E. Stewart, S. Alvarez, and B. J. Wiley, Nano Lett. 15, 6722
concentrations (region 5 in Figs. 4 and 5). (2015).
13
Considering nanowire morphologies, our results can explain F. Cui, Y. Yu, L. Dou, J. Sun, Q. Yang, C. Schildknecht, K. Schierle-Arndt, and
P. Yang, Nano Lett. 15, 7610 (2015).
why (at high HDA concentrations) spherical nanoparticles are 14
L. Zhou, Y. Tan, J. Wang, W. Xu, Y. Yuan, W. Cai, S. Zhu, and J. Zhu, Nat.
experimentally observed for low chloride concentrations and nano- Photonics 10, 393 (2016).
wires occur for intermediate chloride concentrations. Regarding the 15
M. Gao, P. K. N. Connor, and G. W. Ho, Energy Environ. Sci. 9, 3151 (2016).
formation of single-crystal morphologies, the predicted Wulff 16
C. Wu, T. W. Kim, F. Li, and T. Guo, ACS Nano 10, 6449 (2016).
shapes shown in Fig. 6 for regions 6 and 7 in Fig. 4 are truncated 17
T. Cheng, Y. Z. Zhang, J. P. Yi, L. Yang, J. D. Zhang, W. Y. Lai, and
cubes—although the absence of HDA on the {111} facets could W. Huang, J. Mater. Chem. A 4, 13754 (2016).
18
lead to “pointy” cubes due to kinetic effects. This morphology is S. Lin, X. Bai, H. Wang, H. Wang, J. Song, K. Huang, C. Wang, N. Wang,
also seen experimentally.34 Thus, we can resolve most regimes in B. Li, M. Lei, and H. Wu, Adv. Mater. 29, 1703238 (2017).
19
which Cu nanowire or nanocube morphology is governed by ther- K. D. Gilroy, A. Ruditskiy, H. C. Peng, D. Qin, and Y. Xia, Chem. Rev. 116,
modynamics, kinetics, or both influences. One exception is the 10414 (2016).
20
H. Wang, S. Zhou, K. D. Gilroy, Z. Cai, and Y. Xia, Nano Today 15, 121
mechanism by which javelin-shaped wires emerge at high chloride
(2017).
concentrations. 21
J. Qian et al., Mater. Today 21, 834 (2018).
In addition to deposition, surface diffusion can play a role in 22
S. Zhou, M. Zhao, T. H. Yang, and Y. Xia, Mater. Today 22, 108 (2019).
shape control and it seems likely that surface diffusion plays a role 23
D. S. Kilin, O. V. Prezhdo, and Y. Xia, Chem. Phys. Lett. 458, 113 (2008).
in the Cu-HDA-Cl− system. Our recent theoretical studies of the 24
W. A. Al-Saidi, H. Feng, and K. A. Fichthorn, Nano Lett. 12, 997 (2012).
25
growth of Ag nanowires show that surface diffusion alone could W. A. Saidi, H. Feng, and K. A. Fichthorn, J. Phys. Chem. C 117, 1163 (2013).

J. Vac. Sci. Technol. A 38(2) Mar/Apr 2020; doi: 10.1116/1.5141995 38, 023210-7
Published under license by AVS.
 ARTICLE avs.scitation.org/journal/jva

26
S.-H. Liu, W. A. Saidi, Y. Zhou, and K. A. Fichthorn, J. Phys. Chem. C 119, Kristen Fichthorn is the Merrell Fenske
11982 (2015). Professor of Chemical Engineering and a
27
S. H. Liu, T. Balankura, and K. A. Fichthorn, Phys. Chem. Chem. Phys. 18, Professor of Physics at the Pennsylvania State
32753 (2016).
28 University. She received a B.S. degree in
N. Almora-Barrios, G. Novell-Leruth, P. Whiting, L. M. Liz-Marzán, and
Chemical Engineering from the University of
N. López, Nano Lett. 14, 871 (2014).
29
Y. Zhou, W. A. Saidi, and K. A. Fichthorn, J. Phys. Chem. C 117, 11444 (2013). Pennsylvania in 1985 and a Ph.D. degree in
30
M. J. Kim, S. Alvarez, Z. Chen, K. A. Fichthorn, and B. J. Wiley, J. Am. Chem. Chemical Engineering from the University of
Soc. 140, 14740 (2018). Michigan in 1989. She spent one year as an
31
W. Gao, Y. Hou, Z. D. Hood, X. Wang, K. More, R. Wu, Y. Xia, X. Pan, and IBM Postdoctoral Fellow in the Department
M. Chi, Nano Lett. 18, 7004 (2018). of Chemical Engineering at the University of California at Santa
32
Y. Sun, B. Mayers, T. Herricks, and Y. Xia, Nano Lett. 3, 955 (2003). Barbara before joining the Department of Chemical Engineering at
33
X. Qi, Z. Chen, T. Yan, and K. A. Fichthorn, ACS Nano 13, 4647 (2019). Penn State as an Assistant Professor in 1990.
34
M. Jin, G. He, H. Zhang, J. Zeng, Z. Xie, and Y. Xia, Angew. Chem. Int. Ed. Professor Fichthorn’s research is primarily in multiscale simu-
50, 10560 (2011).
35 lation of fluid-solid interfaces, in which she applies theoretical tech-
G. Kresse and J. Hafner, Phys. Rev. B 47, 558 (1993).
36
G. Kresse and J. Hafner, Phys. Rev. B 49, 14251 (1994).
niques ranging from quantum density functional theory to
37
G. Kresse and J. Furthmueller, Phys. Rev. B 54, 11169 (1996). molecular dynamics, Monte Carlo methods, and continuum
38
P. E. Bloechl, Phys. Rev. B 50, 17953 (1994). theories to a diverse array of applications in nanoscale materials,
39 thin-film and crystal growth, colloidal assembly, and wetting. In
G. Kresse and D. Joubert, Phys. Rev. B 59, 1758 (1999).
40
J. P. Perdew, K. Burke, and M. Ernzerhof, Phys. Rev. Lett. 77, 3865 (1996). addition to being recognized by Penn State for her outstanding
41
S. Grimme, J. Comput. Chem. 27, 1787 (2006). research and teaching, she is the recipient of the NSF Presidential
42
V. G. Ruiz, W. Liu, E. Zojer, M. Scheffler, and A. Tkatchenko, Phys. Rev. Lett. Young Investigator Award (1990), an Alexander von Humboldt
108, 146103 (2012). Research Fellowship (1998); she is also a Fellow of the American
43
S. H. Liu and K. A. Fichthorn, J. Phys. Chem. C 121, 22531 (2017). Physical Society (2011), a Fellow of the American Institute of
44
M. J. Kim, S. Alvarez, T. Yan, V. Tadepalli, K. A. Fichthorn, and B. J. Wiley, Chemical Engineers (2017), and a recipient of the Nanoscale
Chem. Mater. 30, 2809 (2018).
45 Science and Engineering Forum Award of the American Institute
X. Qi and K. A. Fichthorn, Nanoscale 9, 15635 (2017).
46 of Chemical Engineers (2019).
K. Reuter and M. Scheffler, Phys. Rev. B 68, 1 (2003).
47 When asked the question “What would I tell my 16-year-old
A. Soon, M. Todorova, B. Delley, and C. Stampfl, Phys. Rev. B 75, 1 (2007).
48
K. D. Gilroy, H. C. Peng, X. Yang, A. Ruditskiy, and Y. Xia, Chem. Comm. 53,
self?” the answer is “Nothing that I would have listened to as a 16

29 August 2024 16:10:58

4530 (2017). year old. However, perhaps I would have told myself that I had the
49
C. Herring, Phys. Rev. 82, 87 (1951). talent to succeed as a professor, that I could have a career more
50
X. Qi, T. Balankura, Y. Zhou, and K. A. Fichthorn, Nano Lett. 15, 7711 (2015). exciting than I could imagine, with so many privileges, and that I
51
T. Balankura, X. Qi, Y. Zhou, and K. A. Fichthorn, J. Chem. Phys. 145, 144106 should do something to nurture that talent. I should look at the
(2016). bigger picture of life and not make decisions based in the
52
X. Qi, T. Balankura, and K. A. Fichthorn, J. Phys. Chem. C 122, 18785 (2018). moment.”

J. Vac. Sci. Technol. A 38(2) Mar/Apr 2020; doi: 10.1116/1.5141995 38, 023210-8
Published under license by AVS.