EFFECT OF ADSORPTION COMPRESSION

ON CATALYTIC CHEMICAL REACTIONS

ARPA-E DE-AR0000708 2017 Final Report

Table of Contents

PART I. Adsorption Compression and Its Influence on Catalytic Chemical Reactions:

Theoretical Predictions, Modeling, and Comparisons with Experimental Data

1. Generalized Ono-Kondo Model and Its Application for Analysis of Adsorption

Compression ……………………………………………………………………………..3
1.1. Thermodynamic fundamentals of Ono-Kondo density functional model………… 3
1.2. Generalization of the Ono-Kondo techniques for data analysis……………………5
1.3. Application of the new Ono-Kondo techniques for adsorption of NO on exfoliated
graphite……………………………………………………………................................8
2. Application of the new Ono-Kondo techniques to determine adsorption compression for
NO in nano-porous zeolites………………………………………………………………9
2.1. NO on NaX………………………………………………………………………11
2.2. NO on CaA and on NaY…………………………………………………………14
2.3. NO on 5A………………………………………………………………………..17
2.4. N2 on ZSM-5 ..................................................…………………………………..21
3. Energies of adsorption compression for NO on zeolites………………………………..22
4. Thermodynamic capacity of catalysts………………………………………..………….24
4.1. Grand Canonical Model ………………………………………………………..24
4.2. Thermodynamic capacity versus geometrical (BET) capacity……………..……26
4.3. Influence of distribution of distances between active sites…………………………31
5. Correlation between reaction rate and adsorption compression…………………...........36
5.1. Correction of Langmuir-Hinshelwood model, taking into account adsorbate-
adsorbate interactions…………………………………………………………..36
6. Adsorption compression indicated by adsorption measurements
for NO on Cu-ZSM-5……………………………………………………………………41
7. Major Results Summary…………………………………………………………………44
8. Fundamental and Practical Implications…………………………………………………45
9. References to Part I……………………………………………………………….……..45

1 
 
PART II. Demonstration of Catalysis under Adsorption Compression

1. Introduction -------------------------------------------------------------------------------- 48

2. Experimental Setup --------------------------------------------------------------------- 49

3. Synthesis of Cu Substituted Zeolites ---------------------------------------------------------- 50

3.1.Material Characterizations and Results -------------------------------------------- 51

4. Catalytic Studies of Direct NO Decomposition and “Volcano Type” Dependence on

Temperature -------------------------------- ----------------------------------------------------- 53

5. Kinetics Studies of Direct NO Decomposition over Cu Substituted Zeolites ------------ 55

6. Isothermal NO Adsorption Measurements --------------------------------------------------- 56

7. Effect of Cu···Cu Dimer Distances on Catalytic Activity of NO Decomposition over Cu

Substituted Zeolites ------------------------------------------------------------------------------ 60

8. Preliminary Results for NO Decomposition over Cu/CeO2 --------------------------------- 61

9. Comparisons with experimental data on rate vs. temperature

for NO on Cu-ZSM-5------------------------------------------------------------------------------63

10. References to Part II ----------------------------------------------------------------------------- 68

2 
 
PART I. Adsorption Compression and Its Influence on Catalytic Chemical Reactions:
Theoretical Predictions, Modeling, and Comparisons with Experimental Data

1. Generalized Ono-Kondo Model and Its Application for Analysis of Adsorption

Compression

1.1. Thermodynamic fundamentals of Ono-Kondo density functional model

Classical Ono-Kondo model is a density functional theory relating density distribution in

adsorbed phase with the bulk density of fluid [1 – 3]. In the simplest versions of this model,
adsorbate is semi-infinite one-component lattice system with ε being the energy of interaction
between nearest neighbors. Consider taking an adsorbate molecule at a layer i and moving it to
an empty site in an infinitely distant layer in the bulk. This is equivalent to the exchange of a
molecule with a vacancy [4 – 5],
Mi + V → Vi + Mb (1)
where M is the adsorbate molecule, and V is the vacancy (empty site) that it fills (and vice versa).
If this exchange occurs at equilibrium, then:
∆H - T∆S = 0 (2)
where ∆H and∆S are the enthalpy and entropy changes, and T is the absolute temperature.
The value of ∆S can be represented in the form:
∆S = k lnW1 - k lnW2 (3)
where W1 is the number of configurations where site in the layer i is occupied by an adsorbate
molecule and the site in the infinitely distant layer is empty, W2 is the number of configurations
where the site in the infinitely distant layer is occupied by an adsorbate molecule and site in the
layer i is empty, and k is Boltzmann’s constant.
If the overall number of configurations for the system is W0, then:
W1/W0 = xi (1 – xb ) (4)
and
W2/W0 = xb (1 - xi ) (5)
where xi is the fraction of a layer i occupied by molecules of adsorbate, and xb is the fraction of
the bulk phase occupied by adsorbtive molecules.

3 
 
 Substituting equations (4) and (5) into equation (3) we obtain
∆S = k ln[xi (1 – xb )/(1 - xi )xb ] (6)
The change in enthalpy can be calculated in the mean-field approximation by considering
the number of neighboring sites that are occupied near the surface compared to the bulk:
∆H = - ε (z1 xi+1 + z2 xi + z1xi-1 - z xb ) (7)
where z1 is the number of bonds of a molecule with next layer, z2 is the coordination number in a
monolayer, and z is the coordination number in the bulk. From equations (2), (6), and (7) it follows
that for i≥2:
ln{[xi (1 - xb )] / [(1 - xi )xb]} + (ε /kT)[z1(xi+1 – xb ) + z2(xi – xb ) + z1(xi-1 – xb )] = 0 (8)
For i=1 , we have instead of equation (7):
∆H = εs - ε(z2 x1 + z1 x2 - z xb ) (9)
Combining equations (2), (6), and (9), we obtain:
ln{[x1 (1 - xb )] /[(1 - x1 )xb]} + εs /kT + (ε/kT)(z2 x1 + z1 x2 - z xb ) = 0 (10)
Equation (10) relates density distribution in adsorbed phase with the density in the bulk phase.
Various versions of equation (10) allow modeling of various types of adsorption behavior,
including adsorption hysteresis [6], order-disorder phase transitions [7], adsorption in micropores
[8], and adsorption of supercritical fluids [9]. Based on such modeling, new classification of
adsorption isotherms was developed [10].

Equations (8) and (10) are coupled and hence must be solved simultaneously. However,
for monolayer adsorption we have xi = xbfor i≥2. In this case equation (8) is:
ln{[x1 (1 - xb )] /[(1 - x1 )xb]} + (ε/kT)[z2 x1 - (z1 + z2) xb ] + εs /kT = 0 (11)
In the low concentration limit, equation (11) gives
x1 = xb exp(-εs /kT) (12)
which is Henry’s law.
When there are no adsorbate - adsorbate interactions, ε= 0, it follows from equation (11)
that
x1 = xb /[xb + (1 - xb )exp(εs /kT)] (13)
which is the Langmuir isotherm widely used in the classical theory of heterogeneous catalysis.
If xb << x1 then equation (11) gives
x1 = xb /[xb + (1 - xb )exp(εs /kT + z2 x1 ε/kT)] (14)

4 
 
 1.2. Generalization of the Ono-Kondo techniques for data analysis

Here we develop a new method allowing application of Ono-Kondo techniques for

microporous adsorbents. In addition, this new approach does not require knowing Henry’s
constants and it is not based on a (mean-field) model of enthalpy for adsorbed molecules.
As seen from equation (7), enthalpy term of Ono-Kondo model is based on lattice model
and mean-field approximation. To make Ono-Kondo model more general, consider ∆H in the
form of Taylor expansion in powers of density:
∆
⋯ (15)

where

(16)
!

and is the energy of intermolecular (adsorbate-adsorbate) interactions.

Plugging∆S = k ln[x1 (1 – xb )/(1 – x1 )xb ] (17)
and ∆H from equation (15) into equation (2) gives:
x1 (1  xb )  s
ln   A1 x1  A2 x12  A3 x13 ...  0
(1  x1 ) xb kT (18)
Comparison of equation (18) with equation (11) indicates that the coefficient A1 characterizes
energy of adsorbate-adsorbate interaction in the limit of small x1. In particular, for classical Ono-
Kondo model, it becomes (ε/kT)z2 .
Taking into account equation (16), equation (18) can be presented in the following form:
x1 (1  xb ) 
ln   s  x1 ( x1 )
(1  x1 ) xb kT (19)

where ⋯.
! !

As seen from equation (19), plotting

x1 (1  xb )
Y  ln
(1  x1 ) xb (20)

5 
 
as a function of x1 gives information about the sign and the magnitude of intermolecular
interactions in adsorbed phase. For example, if the initial slope is negative, then there are
repulsions between nearest neighbors at small x1; if this slope is positive, there are attractions
between nearest neighbors at small x1. For larger x1, the slope can vary and indicate changes of
adsorbate-adsorbate interactions.
Figure 1 illustrates plotting isotherms in coordinates Y vs. x1 . In Figure 1, predictions of
the classical Ono-Kondo equation are represented at 5 for two different values of :

first is at 2 (which is attraction) and second is at 2 (which is repulsion).

X1
1.0 1 
2 
0.8

0.6

0.4

0.2

0.0 Xb
0.0 0.2 0.4 0.6 0.8 1.0
Figure 1. Adsorption isotherms predicted by the classical Ono-Kondo model at 5
for two different values of : first is at 2 (which is attraction) and second is
at 2 (which is repulsion).

As seen from Figure 1, the difference between the two isotherms is not significant at these
parameters. To make this difference more pronounced (and measurable), we plot these isotherms
in coordinates Y (given by equation (20) vs. x1, as given by Figure 2.

6 
 
 Y

6
1 
5

4 2 

X1
0.2 0.4 0.6 0.8 1.0

Figure 2. Adsorption isotherms shown in Figure 2 in Ono-Kondo coordinates.

As seen from Figure 2, first isotherm (with attractive adsorbate-adsorbate interactions) has
positive slope in this coordinates, and second isotherm (with repulsive adsorbate-adsorbate
interactions) has negative slope in this coordinates. So, Ono-Kondo coordinates allow to
distinguish between attractive and repulsive interactions, and the absolute value of the slope
gives information about the magnitude of those interactions.

Figure 3 illustrates the proposed technique for the classical example of adsorption for
ethylene on molecular sieve 13X at 50 C [17]. In Figure 3, adsorption isotherm is plotted as Y
(given by equation (20)) vs x1.

14

12

10

0 X1
0.0 0.2 0.4 0.6 0.8 1.0

7 
 
 Figure 3. Adsorption isotherm for ethylene on molecular sieve 13X at 50 C in coordinates of
equation (19): Y (in units of RT) versus adsorbate density, X1. Experimental data from [17].

As seen from Figure 3, the slope at small x1 is negative and A11 is about 13RT, which
gives energy of adsorbate-adsorbate interactions about 35 KJ/mol. It is remarkable that the slope
of the curve on Figure 3 is negative starting from zero adsorption. Since negative slope indicates
repulsions, this result demonstrates that a nano-pore can “accommodate” only one molecule of
ethylene without compression. However, each neighbor will cause adsorbate-adsorbate
repulsions. Therefore, adsorption compression starts from low densities.

1.3. Application of the new Ono-Kondo techniques for adsorption of NO on exfoliated

graphite

To illustrate the difference between adsorption of NO on flat surface and in nano-porous

zeolite, consider example of NO adsorption on exfoliated graphite [22]. Figure 4 shows
experimental isotherm for this case at T = 77.3 K.
monolayer coverage
1.0

0.8

0.6

0.4

0.2

0.0 P,Torr
0.000 0.005 0.010 0.015 0.020 0.025 0.030

Figure 4. Adsorption isotherm for NO on exfoliated graphite at T = 77.3 K. Data from [22].

Figure 5 gives this isotherm in normalized coordinates.

8 
 
 X1
1.0

0.8

0.6

0.4

0.2

0.0 Xb
0 1. 10 8 2. 10 8 3. 10 8 4. 10 8

Figure 5. Adsorption isotherm for NO on exfoliated graphite at T = 77.3 K in normalized coordinates.

Data from [22].

Figure 6 shows data presented in Figures 4 and 5 in coordinates of equation (19).

20

15

10

0 X1
0.0 0.2 0.4 0.6 0.8 1.0

Figure 6. Adsorption isotherm for NO on exfoliated graphite at T = 77.3 K in coordinates of equation

(19). Data from [22].

9 
 
 As seen from Figure 6, at low densities the slope is positive which indicates adsorbate-
adsorbate attractions. This is consistent with a classical concept of adsorption on a flat surface.
However, at normalized density between 0.6 and 0.8, there is a negative slope indicating
repulsions. These results are consistent with the independent study by neutron scattering [22]
showing transition from two-dimensional liquid to two-dimensional solid.

2. Application of the new Ono-Kondo techniques to determine adsorption compression

for NO in nano-porous zeolites

Analysis of literature indicates that there are few publications reporting on adsorption
isotherms for nitric oxide (NO) on nano-porous zeolites. In recent papers by Yi, Deng et al [20,
21], the authors presented adsorption isotherms for NO on zeolites (NaX, NaY, CaA, and 5A)
and found significant deviations from Langmuir’s behavior (see, for example, Figure 3 in ref.
[20]). Since Ono-Kondo technique allows to analyze deviations from Langmuir’s behavior and
allows to determine energies of adsorbate-adsorbate interactions, we plotted these isotherms in
coordinates of equation (19).
The problem of using equation (19) for microporous adsorbents is finding adsorption
capacity for normalization of adsorption amount. However, BET method is not applicable for
microporous adsorbents (see, for example, IUPAC documents [23, 24]), and there is no reliable
method to determine surface areas for them. This is because BET model is set for macropores
with flat surfaces and adsorption compression changes area per molecule; in addition, surface
area measured at 77 K does not reflect adsorption capacity at catalytic conditions, such as T >
300 K. For this reason, we developed a new approach where knowing adsorption capacity is not
necessary to plot data in coordinates of equation (19). In this approach, we consider systems
where adsorbed amount is relatively small (micromoles per gram), which is typical for
adsorption of supercritical NO on zeolites (critical temperature for NO is 180 K). In this case,
≪1 (21)
and
≪1 (22)
With conditions (21) and (22), equation (19) can be transformed to the following:

10 
 
 x1 
ln   s  x1 A
xb kT (23)
Plugging
x1 = a/am (24)
in equation (23) gives:
a / am 
ln   s  (a / a m ) A
xb kT (25)
Since adsorption capacity is adsorption at the point where attraction to adsorbent is compensated
by repulsions from neighboring molecules of adsorbate, at this point ∆H = 0, which results in
∗
0 (26)
∗ ∗
where Erep is the energy of repulsion, is x1 at a = am, i.e. 1 , which gives instead of
equation (26):

0 (27)

Plugging in equation (25) and eliminating by using equation (27) gives:

a E rep E rep
ln  ln(a m )   a
xb kT a m kT (28)
As seen from equation (28), plotting ln(a/xb) vs a allows to get the slope, S, and the intercept, I,
which are:

(29)

ln (30)

Knowing S and I from plotting experimental data allows to solve equations (29) and (30) with

respect to and am. So, the proposed method gives the energy of adsorbate-adsorbate

intermolecular interactions without knowing adsorption capacity. In addition, this procedure

gives actual adsorption capacity, am. In fact, excluding from equations (29) and (30) results

in the following equation for determining am:

ln 0 (31)
Note that, in equation (31), units of am are the same as units of a.

11 
 
 2.1. NO on NaX
Figure 7 gives the adsorption isotherm for NO on zeolite NaX measured at T = 323 K by
static volumetric apparatus [20]. At this temperature, NO is supercritical, and at pressures below
70 kPa, adsorption amounts are very small (micromoles per gram) which allows using equation
(28).

a, mol g
140

120

100

80

60

40

20

0 P, kPa
0 10 20 30 40 50 60 70

Figure 7. Adsorption isotherm for NO on zeolite NaX measured by static volumetric apparatus at T = 323 K. Data
from [20].

Figure 8 shows the adsorption isotherm for NO on NaX (shown in Fig. 7) in coordinates
a [μmol/g] vs xb. This is necessary to plot this isotherm in coordinates of equation (28).

a, mol g

120

100

80

60

40

20

Xb
0.0001 0.0002 0.0003 0.0004 0.0005

Figure 8. Adsorption isotherm for NO on zeolite NaX at T = 323 K in coordinates a [μmol/g] vs xb.

12 
 
 Figure 9 gives adsorption isotherm, presented in Figure 8, in coordinates of equation (28),
i.e. Y = ln(a/xb) vs a. Lower frame shows this plot in a larger scale.

Y Ln a Xb
20

15

10

0 a, mol g
0 20 40 60 80 100 120 140
Y Ln a Xb
15

14

13

12

11

10 a, mol g
0 20 40 60 80 100 120 140

Figure 9. Adsorption isotherm for NO on zeolite NaX in coordinates of equation (28). Upper and lower frames show
the isotherm in different scales.

As seen from Figure 9, the slope is negative (repulsions) which indicates adsorption
compression. From Figure 9, the slope, S ≈ - 0.019 and the intercept, I ≈ 14.5. With these S and

I, equation (31) gives am ≈ 442.5 μmol/g. Then, energy of adsorption compression is ≈ 8.04

which corresponds to Erep ≈ 5.194 kCal/mol. Figure 10 shows the same isotherm in coordinates
of equation (19) without approximations (21) and (22), where adsorption capacity is am ≈ 442.5
μmol/g (determined from Figure 9).

13 
 
 Y
10
x1 (1  xb )
Y  ln
8 (1  x1 ) xb

0 X1
0.0 0.2 0.4 0.6 0.8 1.0

Figure 10. Adsorption isotherm for NO on zeolite NaX in coordinates of equation (19).

2.2. NO on CaA and on NaY

Figure 11 shows adsorption isotherms for NO on zeolites CaA and NaY measured by
static volumetric apparatus at T = 323 K. Figure 12 gives this adsorption isotherm in coordinates
a [μmol/g] vs xb. This is necessary to plot this isotherm in coordinates of equation (28).

a, mol g
140

120

100

80

60

40

20

0 P,kPa
0 10 20 30 40 50 60 70
Figure 11. Adsorption isotherms for NO on zeolites CaA (■) and NaY (●) measured
by static volumetric apparatus at T = 323 K. Data from [20].

14 
 
 a, mol g
140

120

100

80

60

40

20

Xb
0.0001 0.0002 0.0003 0.0004 0.0005
Figure 12. Adsorption isotherms for NO on zeolites CaA (■) and NaY (●) at T = 323 K in
coordinates a [μmol/g] vs xb.. Data from [20].

Figure 13 gives this adsorption isotherm in coordinates of equation (28), ln(a/xb) vs a.

Y Ln a Xb
15

14

13

12

11

10 a, mol g
0 20 40 60 80 100 120 140

Y Ln a Xb
15

14

13

12

11

10 a, mol g
0 20 40 60 80 100 120 140

Figure 13. Adsorption isotherms for NO on zeolites CaA (■) and NaY (●) in coordinates of equation (28).
Data from [20].

15 
 
 As seen from Figure 13 the slopes for both CaA and NaY are negative (repulsions) which
indicates adsorption compression. From Figure 13, the slopes are - 0.019 (for CaA) and - 0.04
(for NaY). The intercepts are 14.55 (for CaA) and 12.8 (for NaY). With these S and I, equation
(31) gives am ≈ 444.85 μmol/g (for CaA) and am ≈ 188.96 μmol/g (for NaY). Then, energy of

adsorption compression is ≈ 8.45 (for CaA) and ≈ 7.56 (for NaY). This corresponds to

Erep ≈ 5.46 kCal/mol (for CaA) and Erep ≈ 4.88 kCal/mol (for NaY). Figure 14 shows the same
isotherms in coordinates of equation (19) without approximations (21) and (22), where
adsorption capacities are am ≈ 444.85 μmol/g for CaA and am ≈ 188.96 μmol/g for NaY
(determined from Figure 13).
Y
10
x1 (1  xb )
Y  ln .
8 (1  x1 ) xb

0 X1
0.0 0.2 0.4 0.6 0.8 1.0

Y
10

x1 (1  xb )
8 Y  ln .
(1  x1 ) xb

0 X1
0.0 0.2 0.4 0.6 0.8 1.0

Figure 14. Adsorption isotherms for NO on zeolites CaA (■) and NaY (●) in coordinates of equation (19).

16 
 
 2.3. NO on 5A

Figure 15 gives the adsorption isotherms for NO on zeolite 5A measured by static

volumetric apparatus at various temperatures [21]: 323 K (●), 348 K (♦), and 363 K (■).
Figure 16 gives this adsorption isotherm in coordinates a [μmol/g] vs xb. This is necessary to plot
this isotherm in coordinates of equation (28).

a, mol g
140

120

100

80

60

40

20

0 P, kPa
0 10 20 30 40 50 60 70

Figure 15. Adsorption isotherms for NO on zeolite 5A measured by static volumetric apparatus at various
temperatures: 323 K (●), 348 K (♦), and 363 K (■). Data from [21].

a, mol g

150

100

50

Xb
0.0001 0.0002 0.0003 0.0004 0.0005 0.0006

Figure 16. Adsorption isotherms for NO on zeolite 5A in coordinates a [μmol/g] vs xb.: 323 K (●), 348 K (♦), and
363 K (■). Data from [21].

17 
 
 Figure 17 gives adsorption isotherms for NO on zeolite 5A at various temperatures in
coordinates of equation (28), ln(a/xb) vs a. As seen from Figure 17, the slopes are negative
(repulsions) which indicates adsorption compression. From Figure 17, the slopes are - 0.02 (for T
= 323 K), - 0.022 (for T = 348 K), and - 0.024 (for T = 363 K). The intercepts are 14.7 (for T =
323 K), 14.1 (for T = 348 K), and 13.8 (for T = 363 K). With these S and I, equation (31) gives
am ≈ 431.62 μmol/g (for T = 323 K), am ≈ 371.88 μmol/g (for T = 348 K), and am ≈ 332.995

μmol/g (for T = 363 K). Then, the energy of adsorption compression is ≈ 8.63 (for T = 323

K), ≈ 8.18 (for T = 348 K), and ≈ 7.99 (for T = 363 K). This corresponds to Erep ≈ 5.58

kCal/mol (for T = 323 K), Erep ≈ 5.69 kCal/mol (for T = 348 K), and Erep ≈ 5.8 kCal/mol (for T =
363 K). Figure 18 shows the same isotherms in coordinates of equation (29) without
approximations (21) and (22), where adsorption capacities are am ≈ 431.62 μmol/g for T = 323 K,
am ≈ 371.88 μmol/g for T = 348 K, and am ≈ 332.995 μmol/g for T = 363 K (determined from
Figure 17).

18 
 
 Y Ln a Xb
15

14 T = 323 K 

13

12

11

10 a, mol g
0 20 40 60 80 100 120 140

Y Ln a Xb
15

14
T = 348 K 
13

12

11

10 a, mol g
0 20 40 60 80 100 120 140

Y Ln a Xb
15

14 T = 363 K 

13

12

11

10 a, mol g
0 20 40 60 80 100 120 140

Figure 17. Adsorption isotherm for NO on zeolite 5A at T = 323 K, T = 348 K, and T = 363 K
in coordinates of equation (28).

19 
 
 Y
10
T = 323 K 
8

0 X1
0.0 0.2 0.4 0.6 0.8 1.0

Y
10

T = 348 K 
6

0 X1
0.0 0.2 0.4 0.6 0.8 1.0

Y
10

8
T = 363 K 

0 X1
0.0 0.2 0.4 0.6 0.8 1.0

Figure 18. Adsorption isotherms for NO on zeolite 5A at T = 323 K, T = 348 K, and T = 363 K
x1 (1  xb )
in coordinates of equation (19). In this Figure, Y  ln .
(1  x1 ) xb

20 
 
 2.4. N2 on ZSM-5

Measurements of adsorption isotherms for N2 on zeolite ZSM-5 were performed on BET

apparatus Micromeritics 2010. Zeolite sample ratio of Si/Al was 11.5, and sample weight 0.09 g.
Figure 19 shows adsorption isotherm for N2 on ZSM-5 at T = 77 K. This isotherm was
plotted in coordinates of generalized Ono-Kondo equation.
adsorption , cc g
300

250

200

150

100

50

0 P,mmHg
0 200 400 600 800

Figure 19. Adsorption isotherm for N2 on ZSM-5 (Si/Al = 11.5) obtained on BET apparatus at T = 77 K.

Figure 20 gives adsorption isotherm for N2 on zeolite ZSM-5 in coordinates of equation (28).

Y
20

15

10

0 a, cc g
0 50 100 150 200 250 300

Figure 20. Adsorption isotherm for N2 on ZSM-5 (Si/Al = 11.5) in coordinates of equation (28).

21 
 
 As seen from Figure 20, the slope is negative (repulsions) which indicates adsorption
compression. From Figure 20, the slope is – 0.12 and the intercept is about 26.5. With these S

and I, equation (31) gives am ≈ 180 cc/g. Then, energy of adsorption compression is ≈ 21.6

which corresponds to Erep ≈ 3.3 kCal/mol. Figure 21 shows the same isotherm in coordinates of
equation (19) without approximations (21) and (22), where adsorption capacity is am ≈ 180 cc/g
(determined from Figure 20). Note that these techniques allow processing of isotherms in the
original units of BET apparatus, adsorption in cc/g and pressure in mm Hg.

Y
20

15

10

0 X1
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4

Figure 21. Adsorption isotherm for N2 on ZSM-5 (Si/Al = 11.5) in coordinates of equation (19).

3. Energies of adsorption compression for NO on zeolites

Table 1 shows energies of adsorption compression for NO on various zeolites obtained

from Figures 7 – 18. Table 1 also shows information on pores sizes distribution known from
nitrogen adsorption and from small angle X-ray scattering (SAXS) [32].  

22 
 
Table 1. Energies of adsorption compression for NO on zeolites obtained from Figures 7 – 18.

Zeolites T, K Pore size distribution, Å * Energy of adsorption

compression

NaX 323 K 2 Å … 15 Å 5.2 kCal/mol

NaY 323 K 3 Å … 14 Å 4.88 kCal/mol

CaA 323 K 2 Å … 15 Å 5.46 kCal/mol

5A 323 K Narrow around 5 Å 5.6 kCal/mol

5A 348 K Narrow around 5 Å 5.7 kCal/mol

5A 363 K Narrow around 5 Å 5.8 kCal/mol

Table 2 shows activation enthalpies for NO oxidation on Silica, SIL-1, SIL-1D, BEA, and CHA
zeolites at 278 – 373 K [33].

Table 2. Activation enthalpies for NO oxidation on Silica, SIL-1, SIL-1D, BEA, and CHA zeolites
at 278–373 K [33].
Zeolites Activation enthalpies for NO on zeolites

Silica 7.42 ± 1.1 kCal/mol

SIL-1 8.56 ± 0.5 kCal/mol

SIL-1D 9.45 ± 0.19 kCal/mol

BEA 8.97 ± 2.97 kCal/mol

CHA 9.86 ± 0.33 kCal/mol

23 
 
As seen from comparisons of Tables 1 and 2, energies of adsorption compression are comparable
with activation enthalpies of zeolites.

4. Thermodynamic capacity of catalysts

4.1. Grand Canonical Model

As seen from the section on NO on zeolites, at T = 323 K estimated values of am for
NaX, NaY, and CaA are about 440 μmol/g. If area per one molecule of NO is 19 Å2, than the
surface area is about 50 m2/g which is at least an order of magnitude less than that of BET
surface areas for these adsorbents. This difference comes from the fact that the value of am in
equation (28) is adsorption at the point where attraction to the surface is compensated by
repulsions between adsorbate molecules. Therefore, it is an actual (thermodynamic) capacity, not
geometrical capacity which can be expected from BET measurements (not to mention that BET
model is incorrect for microporous adsorbents [23, 24]).

To illustrate the concept of thermodynamic capacity, consider a simplest situation where

adsorption compression is possible and which allows rigorous theoretical treatment [31]: two
active sites at various distances, as shown in Figure 22.

  A    B 
a 
 
 

  A    B 
b  
 

   B 
A   
c 
 
Figure 22. Molecules on active sites 
 

When the active sites are close and their attraction of adsorbate molecules is strong, both
sites can be occupied, but the adsorbed molecules will repel each other (Figure 22b). In this case,
attraction to the active sites must be stronger than repulsion between neighbors. Therefore,
adsorption of both molecules simultaneously is thermodynamically favorable; however, the
distance between such molecules is smaller than in a normal liquid and adsorbate molecules

24 
 
repel each other. This is the simplest case of the adsorption compression, and it can occur
whenever the lattice spacing for molecules in the adsorbent is smaller than the minimum in the
potential function for adsorbate – adsorbate interactions. At very small distances between active
sites (Figure 22c), this effect disappears: the A molecule blocks the neighboring active site
because repulsions between neighbors exceeds attractions to the active sites at this distance.
For the grand canonical ensemble, the variables are chemical potential, , number of
molecules on active sites, N, and absolute temperature, T. To calculate the grand canonical
partition function,  , we assume that a gas phase is in equilibrium with two active sites with εs
being the energy of molecule-active site interactions, and d being the distance between sites. For
this model, there are four different states:
- one state with both active sites empty; configurational energy of this state is zero;
- two states where only one of sites is occupied; configurational energy of each of these
states is εs ;
- one state where both sites are occupied and the interaction energy between molecules
sitting on these sites is  (d):
 
 (d )  4 [( )6  ( )12 ] (32)
d d
Configurational energy of this state is 2εs +(d).
The grand canonical partition function for this system can be written as:
E0  E1'  E '' 2  E2
  exp(  )  exp(  )  exp(  1 )  exp(  ) (33)
kT kT kT kT kT kT kT
where k is Boltzmann’s constant, and:
E0 = 0 (34)

E1'  E1''   s (35)

E2 = 2εs +(d) (36)

The average number of molecules, <<N>> , sitting on the two active sites is:
1  E1'  E1'' 2 E2
 N  [exp(  )  exp(  )  2 exp(  )] (37)
 kT kT kT kT kT kT
Plugging equations (33) – (36) into equation (37) gives:

25 
 
  s
2 2 s  (d )
2 exp(  )  2 exp[
 ]
 N  kT kT kT kT kT (38)
  2 2 s  (d )
1  2 exp(  s )  exp[   ]
kT kT kT kT kT
 

4.2. Thermodynamic Capacity versus Geometrical (BET) Capacity

Figure 23 gives adsorption isotherm predicted by equation (38) at ε/kT = -0.5, εs/kT = -
15, σ = 3.5 Å, and d = 2.8 Å.

Figure 23. Adsorption isotherm predicted by equation (38) at ε/kT = -0.5, εs/kT = -15, σ = 3.5 Å, and d = 2.8 Å.

As seen from Figure 23, at the distance of d = 2.8 Å, the first molecule blocks the second site,
and at μ/kT → 0 value of adsorption asymptotically approaches 1.

Figure 24 gives the same adsorption isotherm (as shown in Figure 23) in Ono-Kondo

coordinates, Y vs. xa, where and xa = <<N>>/2. As seen from equation (19), Y

= 0 defines the point where attraction to the surface is compensated by repulsions between
adsorbate molecules, i.e. where

s
 xa ( xa )  0
kT (39)

26 
 
Therefore, the point indicated by the arrow on Figure 24 is thermodynamic capacity,
corresponding to one molecule (0.5 * 2 = 1), though geometric capacity is 2 molecules.

thermodynamic capacity

Figure 24. Adsorption isotherm predicted by equation (38) at ε/kT = -0.5, εs/kT = -15, σ = 3.5 Å, and d = 2.8 Å in
coordinates Y vs. xa. Arrow indicates thermodynamic capacity.

Equations (28) and (31) are written for cases where energy is a linear function of
adsorption amount. Dashed line in Figure 24 shows a linear function Y(xa), with the same Y(0)
and same thermodynamic capacity (compared to the curve predicted by equation (38)). For this
dashed line, S = -30 and I = 15. Therefore, equation (31) becomes

ln 30 15 0 (40)

Solution of equation (40) gives am ≈ 0.52 which is very close to the thermodynamic capacity of
0.5 for this case.

Figure 25 gives adsorption isotherm predicted by equation (38) at ε/kT = -0.5, εs/kT = -
15, σ = 3.5 Å, and d = 2.9 Å.

27 
 
Figure 25. Adsorption isotherm predicted by equation (38) at ε/kT = -0.5, εs/kT = -15, σ = 3.5 Å, and d = 2.9 Å.

As seen from Figure 25, at the distance of d = 2.9 Å, the first molecule does not block the second
site (just hinders), and, at μ/kT → 0 , <<N>> ≈ 1.8 molecules which gives xa ≈ 1.8/2 = 0.9.

Figure 26 gives the same adsorption isotherm (as shown in Figure 25) in Ono-Kondo

coordinates, i.e. Y vs. xa, where and xa = <<N>>/2.

apparent   thermodynamic  
capacity  capacity 

Figure 26. Adsorption isotherm predicted by equation (38) at ε/kT = -0.5, εs/kT = -15, σ = 3.5 Å, and d = 2.9 Å in
coordinates Y vs. xa. Arrows indicate apparent and thermodynamic capacities.

28 
 
As seen from Figure 26, thermodynamic capacity is 0.9 which corresponds to 1.8 molecules on
Figure 25 at μ/kT → 0. At xa = 0.5, Figure 26 also indicates an apparent capacity of 0.5
corresponding to a plateau on Figure 25 at μ/kT between -12 and -4 where only one molecule can
be adsorbed.

For the case of Figure 26, equation 40 becomes:

ln 16.67 15 0 (41)

Solution of equation (41) gives am ≈0.905 which is very close to thermodynamic capacity of 0.9
for this case.

Figure 27 gives adsorption isotherm predicted by equation (38) at ε/kT = -0.5, εs/kT = -
15, σ = 3.5 Å, and d = 3.0 Å.

Figure 27. Adsorption isotherm predicted by equation (38) at ε/kT = -0.5, εs/kT = -15, σ = 3.5 Å, and d = 3.0 Å.

As seen from Figure 27, at d = 3.0 Å, for μ/kT → 0 we have <<N>> ≈ 2 which gives xa ≈ 1.

Figure 28 gives the same adsorption isotherm (as shown in Figure 27) in Ono-Kondo

coordinates, i.e. Y vs. xa, where and xa = <<N>>/2.

29 
 
 apparent   thermodynamic  
capacity  capacity 

Figure 28. Adsorption isotherm predicted by equation (38) at ε/kT = -0.5, εs/kT = -15, σ = 3.5 Å, and d = 3.0 Å in
coordinates Y vs. xa. Arrows indicate apparent and thermodynamic capacities.

As seen from Figure 28, thermodynamic capacity is 1 which corresponds to 2 molecules as μ/kT
→ 0. Figure 28 also indicates an apparent capacity of about 0.62 corresponding to a section on
Figure 27 at μ/kT between -13 and -10 (its slope is marked by the dashed line).

For the case of Figure 28, equation 40 becomes:

ln 15 15 0 (42)

Solution of equation (42) gives am =1 coinciding with the geometrical capacity.

30 
 
 4.3. Influence of distribution of distances between active sites

Consider normalized Gaussian distribution of distances between active sites:

exp (43)

where d0 is the average distance between active sites, c is the parameter characterizing width of
the distribution, and q is the normalizing factor. Figure 29 illustrates dependence of distribution
of distances for d0 = 3 Å and q =1 on parameter c.

P P
1.0 1.0

0.8 1  0.8 2 

0.6 0.6

0.4 0.4

0.2 0.2

0.0 d 0.0 d
0 1 2 3 4 5 6 0 1 2 3 4 5 6
P
1.0

0.8

3 
0.6

0.4

Figure 29. Distribution of distances, P(d), for d0 = 3 Å,

0.2
q =1, and different values of c: 0.01 (1), 0.001 (2), and
0.0001 (3).
0.0 d
0 1 2 3 4 5 6

Taking into account equation (43), equation (38) can be rewritten in the following form:
2 2
1 2 exp 2 exp
≫
2 2
1 2exp exp

(44)
Here <<Ndist>> is the number of molecules occupying the two sites averaged over distribution
of distances, and d1 and d2 are limits of possible distances between active sites.

31 
 
 Figure 30 shows adsorption isotherm predicted by equation (44) at εs /kT = -15, σ = 3.5,
d0 = 3 Å, ε/kT = -0.5, and c = 0.0001. Also shown in Figure 30 is adsorption isotherm without
distribution of distances (d = 3 Å).

a b

Figure 30. Adsorption isotherm predicted by equation (44) at εs /kT = -15, σ = 3.5 Å, d0 = 3 Å, ε/kT = -0.5, and c =
0.0001 (frame a). Frame (b) shows adsorption isotherm for the same parameters without distribution of distances
(prediction of equation (38) at d = 3 Å ).

As seen from Figure 30, narrow distribution (c = 0.0001) results in insignificant change of
adsorption isotherm.
Figure 31 shows adsorption isotherm predicted by equation (44) at εs /kT = -15, σ = 3.5,
d0 = 3 Å, ε/kT = -0.5, and c = 0.001. Also shown in Figure 31 is adsorption isotherm without
distribution of distances (d = 3 Å).

a b

Figure 31. Adsorption isotherm predicted by equation (44) at εs /kT = -15, σ = 3.5 Å, d0 = 3 Å, ε/kT = -0.5, and c =
0.001 (frame a). Frame (b) shows adsorption isotherm for the same parameters without distribution of distances
(prediction of equation (38) at d = 3 Å ).

32 
 
As seen from Figure 31, wider distribution at c = 0.001 results in a noticeable differences in
adsorption isotherm; though the main features (such as two steps and plateau between them)
remain well pronounced.
Figure 32 shows adsorption isotherm predicted by equation (44) at εs /kT = -15, σ = 3.5,
d0 = 3 Å, ε/kT = -0.5, and c = 0.01. Also shown in Figure 32 is adsorption isotherm without
distribution of distances (d = 3 Å).

a b

Figure 32. Adsorption isotherm predicted by equation (44) at εs /kT = -15, σ = 3.5 Å, d0 = 3 Å, ε/kT = -0.5, and c =
0.01 (frame a). Frame (b) shows adsorption isotherm for the same parameters without distribution of distances
(prediction of equation (38) at d = 3 Å ).

As seen from Figure 32, wide distribution at c = 0.01 still results in a step at μ/kT about -15.
However, the second step disappeared; instead, at μ/kT > -15, there is saturation, not reaching
full capacity of N = 2.
Figure 33 shows adsorption isotherms for two-site model in Ono-Kondo coordinates at εs
/kT = -15, σ = 3.5Å, d0 = 3Å, and ε/kT = -0.5: (c) with distribution of distances between active
sites, c = 0.001, and (d) with constant distance between active sites (d = 3 Å ). Also shown are
adsorption isotherms for two-site model at εs /kT = -15, σ = 3.5 Å, d0 = 3 Å, and ε/kT = -0.5: (a)
with distribution of distances between active sites, c = 0.001, and (d) with constant distance
between active sites (d = 3 Å ).

33 
 
 b
a

P
1.0

0.8

0.6

0.4

0.2

0.0 d
0 1 2 3 4 5 6

c d

P
1.0

0.8

0.6

0.4

0.2

0.0 d
0 1 2 3 4 5 6

thermodynamic capacity: 0.98*2=1.96  thermodynamic capacity 

Figure 33. Adsorption isotherms for two-site model in Ono-Kondo coordinates at εs /kT = -15, σ = 3.5Å, d0 = 3Å,
and ε/kT = -0.5: (c) with distribution of distances between active sites, c = 0.001; (d) constant distance between
active sites (d = 3 Å ). Also shown adsorption isotherms for two-site model at εs /kT = -15, σ = 3.5 Å, d0 = 3 Å, and
ε/kT = -0.5: (a) with distribution of distances between active sites, c = 0.001; (b) constant distance between active
sites (d = 3 Å ).

As seen from Figure 33, distribution of distances between active sites results in a noticeable
changes in adsorption isotherm. However, these changes are reflected in adsorption isotherms in
Ono-Kondo coordinates in a way allowing to get information about adsorption capacity. As seen
from Figure 33, frame b, adsorption capacity does not reach maximum capacity as μ/kT → 0, and
this is seen from the isotherm in Ono-Kondo coordinates as xa → 1.
Figure 34 shows adsorption isotherms for two-site model in Ono-Kondo coordinates at εs
/kT = -15, σ = 3.5Å, d0 = 3Å, and ε/kT = -0.5: (c) with distribution of distances between active
sites, c = 0.01, and (d) with constant distance between active sites (d = 3 Å ). Also shown are
adsorption isotherms for two-site model at εs /kT = -15, σ = 3.5 Å, d0 = 3 Å, and ε/kT = -0.5: (a)

34 
 
with distribution of distances between active sites, c = 0.01, and (d) with constant distance  
between active sites (d = 3 Å ).

a b

Pd
1.0

0.8

0.6

0.4

0.2

0.0
0 1 2 3 4 5 6
x  

c d

Pd
1.0

0.8

0.6

0.4

0.2

0.0 x
0 1 2 3 4 5 6

thermodynamic capacity: 0.85*2=1.7  thermodynamic capacity 

Figure 34. Adsorption isotherms for two-site model in Ono-Kondo coordinates at εs /kT = -15, σ = 3.5Å, d0 = 3Å,
and ε/kT = -0.5: (c) with distribution of distances between active sites, c = 0.01; (d) constant distance between
active sites (d = 3 Å ). Also shown adsorption isotherms for two-site model at εs /kT = -15, σ = 3.5 Å, d0 = 3 Å, and
ε/kT = -0.5: (a) with distribution of distances between active sites, c = 0.01; (b) constant distance between active
sites (d = 3 Å ).

For the case of Figure 34, equation 40 becomes ln 17.65 15 0. Solution of this
equation gives am ≈0.858 which is very close to thermodynamic capacity of 0.85 indicated in
Figure 34 at the point where Ydist = 0 (lower left frame). So, Ono-Kondo procedure provides
correct thermodynamic capacity even at a relatively wide distribution of distances between active
sites.

35 
 
 5. Correlation between reaction rate and adsorption compression

5.1. Correction of Langmuir-Hinshelwood model, taking into account adsorbate-

adsorbate interactions

Classical theory of heterogeneous catalytic reactions is based on Langmuir-Hinshelwood

(LH) mechanisms considering two major factors: a) adsorption from the gas phase, and b)
thermally activated reaction on the surface. In particular, for mono- and bimolecular reactions,
LH model results in the following equations for the reaction rate, r [25]:
45
46
Here, K is the kinetic constants of the surface reaction, and is the adsorption isotherm.
Kinetic constant can be determined by the Arrhenius term:

47

where E is the activation energy and K0 is the limit of the kinetic constant at high temperature.
Classical LH model is based on the following two assumptions:
(a) E is the minimum kinetic energy, Ekinetic, of the thermally activated molecules necessary
to overcome activation barrier of the reaction, i.e.
E = Ekinetic = E0 (48)
where E0 is the classical catalytic reaction barrier.
(b) adsorption isotherm can be described by Langmuir’s isotherm.
However, assumption (a) takes into account only kinetic energy of the molecules and
neglects their potential energy from lateral interactions. This can be fixed by including potential
energy of adsorbate-adsorbate interaction, :
(49)
For adsorption compression,
(50)
where 0 (repulsions).
Note that the values of Erep for NO on zeolites (obtained in this project and given in the
previous section of this report) are in the range from 4.88 kCal/mol to 5.8 kCal/mol. These

36 
 
values are comparable with the typical enthalpies for catalytic reactions for NO on zeolites [26],
which range from 7 kCal/mol to 9 kCal/mol.
Assumption (b) includes neglecting adsorbate-adsorbate interactions. However, our
results in Figures 3, 6, 9, 10, 13, 14, 17, 18, 20, and 21 show significant deviations from the
Langmuir model for all isotherms of adsorption for NO on zeolites. Straightening in Ono-Kondo
coordinates demonstrates that Ono-Kondo isotherm with repulsive energy of lateral interactions
is more adequate and more realistic representation of adsorption isotherms for NO on zeolites.
To generalize Langmuir-Hinshelwood model, we consider more general equation (49)
instead of equation (35) for the Arrhenius term and replace Langmuir’s isotherm by Ono-Kondo
model taking into account lateral interactions. These improvement allow capturing essential
physics of adsorbate-adsorbate interactions for catalytic reaction in adsorbed phase.
Plugging equations (49) and (50) in equations (46) and (47) gives for bimolecular
reaction:

51

where E0 is the catalytic reaction barrier.

Ono-Kondo isotherm corresponding to straight lines in Figures 3, 6, 9, 10, 13, 14, 17, 18,
20, and 21 can be written in the following form:
∗
∗ 52

where
∗
exp 53

Equations (51) – (53) determine reaction rate as a function of temperature and density of gas.
To analyze dependence of reaction rate on temperature, rewrite equation (53) in the
following form:

∗
(54)

Plugging kT from equation (54) into equation (51) gives:

∗
55
1

Equations (52), (54), and (55) represent rb(T, xb) in a parametric form with K* being a parameter.

37 
 
 For NO decomposition on Cu-ZSM-5, reaction activation energy is 19.5 kCal/mol and
enthalpy of adsorption of NO onto Cu+ is -34.1 kCal/mol [27]. Absolute value of K0 depends on
units of . To plot reaction rates as functions of temperature, we used normalized reaction rate
in units of conversion fraction which is equivalent to considering / as a function of T.
Figure 35 shows dependence of reaction rate as a function of the distance between active
sites, d, and temperature, T.

2.7 Å
600 K

3.2 Å

Figure 35 Dependence of reaction rate as a function of the distance between active sites, d, and temperature, T, at 
⁄RT = ‐6,  ⁄R= ‐12000,  ⁄R=‐9000, and  ⁄R = ‐500. 

As seen from Figure 35, optimal regime (with maximum rate) is limited to a narrow
range of parameters. Finding these optimal conditions is difficult without measurements of
adsorption compression.
Figure 36 shows reaction rate (upper frame) and energy of adsorption compression (lower
frame) as functions of the distance between active sites, d, and temperature, T, at  ⁄RT = ‐6, 
⁄R= ‐12000,  ⁄R=‐9000, and  ⁄R = ‐500. As seen from Figure 36, there is a correlation between 
maximum of reaction rate and maximum of the energy of adsorption compression.  

38 
 
 2.7 Å 1000 K
600 K

3.2 Å

Figure 37. Reaction rate (upper frame) and energy of adsorption compression (lower frame) as functions of the
distance between active sites, d, and temperature, T, at  ⁄RT = ‐6,  ⁄R= ‐12000,  ⁄R=‐9000, and  ⁄R = ‐500. 

39 
 
 Figure 37 shows dependence of the reaction rate on temperature and on average distance
between sites at ⁄RT = -1, ⁄R= -12000, ⁄R=-9000, and ⁄R = -500 for Gaussian distribution of the
distances between active sites.

1200 K

600 K
2.7 Å
3 Å

3.7 Å

Figure 37. Dependence of the reaction rate on temperature and on average distance between sites at ⁄RT = -1,
⁄R= -12000, ⁄R=-9000, and ⁄R = -500 for Gaussian distribution of the distances between active sites.

As seen from Figure 37, for Gaussian distribution of the distances between active sites,
maximum of reaction rate is not sharp with respect to the dependence on temperature. However,
it is still sharp with respect to the distance between active sites.

6. Adsorption compression indicated by adsorption measurements

for NO on Cu-ZSM-5

We have analyzed parameters of corrected LH model based on comparisons with

experimental data obtained in this project (presented separately in Part II). Table 3 summarizes
results on comparison of corrected LH model with experimental data shown in Figures 14 – 22
of Part II. As seen from Table 3, energy of adsorption compression goes down if content of Cu
goes down.

40 
 
 Table 3. Energy of adsorption compression as a function of copper content

Si/Al Cu/Al Cu/Total Energy of Adsorption

compression

11.5 0.55 0.044 9 kCal/mol

20 0.51 0.024 8.9 kCal/mol

0.31 0.015 8.7 kCal/mol
30 0.42 0.01355 8.6 kCal/mol
0.35 0.0113 8.4 kCal/mol
50 0.69 0.01353 8.4 kCal/mol
0.55 0.01 7.0 kCal/mol

Table 4 shows results of measurements of adsorbed amount for NO on Cu-ZSM-5 at 80oC.

Table 4. Adsorption isotherm for NO on Cu-ZSM-5 at 80oC.

Pressure 30 Pa 50 Pa 80 Pa 150

Normalized density, xb 0.24*10-6 0.4*10-6 0.64*10-6 1.2*10-6

Adsorbed amount 198.65 μmol/g 305.52 μmol/g 442 μmol/g 684.9 μmol/g

Table 5 shows these data in coordinates of Ono-Kondo equation for small xb.

Table 5. Adsorption isotherm for NO on Cu-ZSM-5 at 80oC in Ono-Kondo coordinates.

Pressure, P [Pa] Adsorbed amount, a [μmol/g] ln(a/P) ln(a/xb)

30 198.65 1.89 20.53

50 305.52 1.81 20.45

80 442 1.71 20.35

150 684.9 1.52 20.16

Figure 38 shows adsorption isotherm for NO on Cu-ZSM-5 at 80oC in Ono-Kondo

coordinates, ln(a/P) vs. a (upper frame) and ln(a/xb) vs. a (lower frame).

41 
 
 ln a P
2.2

2.0

1.8

1.6

1.4 a, mol g
0 100 200 300 400 500 600 700
ln a Xb
20.8

20.6

20.4

20.2

20.0 a, mol g
0 100 200 300 400 500 600 700

Figure 38. Adsorption isotherm for NO on Cu-ZSM-5 at 80oC in Ono-Kondo coordinates, ln(a/P) vs. a (upper
frame) and ln(a/xb) vs. a (lower frame).

As seen from Figure 38, the slope of this isotherm in Ono-Kondo coordinates is negative which
indicates adsorption compression. From the data in Table 5, we can determine the slope, S ≈
0.00075 and intercept, I ≈ 20.67. Then, the following equation
ln 0.00075 20.67 0
gives am ≈ 14437 μmol/g, which allows calculation of the (repulsive) energy of adsorption
compression, Erep /RT = amS= 14760.4* 0.00075 ≈ 11.07, which gives, at 80oC, Erep ≈ 7.8
kCal/mol. This result is consistent with substantial evidence that the decomposition of NO occurs
on closely located ions of copper [28 – 30].
Table 6 shows results of measurements of adsorbed amount for NO on Cu-ZSM-5 at
120oC.

42 
 
Table 6. Adsorption isotherm for NO on Cu-ZSM-5 at 120oC.
Pressure 80 Pa 150 Pa 200 Pa

Normalized density, xb 0.64*10-6 1.2*10-6 1.6*10-6

Adsorbed amount 352.6 μmol/g 562.8 μmol/g 685.4 μmol/g

Table 7 shows these data in coordinates of Ono-Kondo equation for small xb.

Table 7. Adsorption isotherm for NO on Cu-ZSM-5 at 120oC in Ono-Kondo coordinates.

Pressure, P [Pa] Adsorbed amount, a [μmol/g] ln(a/P) ln(a/xb)

80 352.6 1.483 20.13

150 562.6 1.322 19.97

200 685.4 1.232 19.88

Figure 39 shows adsorption isotherm for NO on Cu-ZSM-5 at 120oC in Ono-Kondo

coordinates, ln(a/P) vs. a (upper frame) and ln(a/xb) vs. a (lower frame).
ln a P
1.7

1.6

1.5

1.4

1.3

1.2 a, mol g
0 200 400 600 800
ln a Xb
20.4

20.3

20.2

20.1

20.0

19.9

19.8

19.7 a, mol g
0 200 400 600 800
43 
 
Figure 39. Adsorption isotherm for NO on Cu-ZSM-5 at 120oC in Ono-Kondo coordinates, ln(a/P) vs. a (upper
frame) and ln(a/xb) vs. a (lower frame).

As seen from Figure 39, the slope of this isotherm in Ono-Kondo coordinates is negative which
indicates adsorption compression. From the data in Table 7, we can determine the slope, S ≈
0.00075 and intercept, I ≈ 20.4. Then, the following equation

ln 0.00075 20.4 0
gives am ≈ 14430 μmol/g, which allows calculation of the (repulsive) energy of adsorption
compression, Erep /RT = amS=14430* 0.00075 ≈ 10.82, which gives, at 120oC, Erep ≈7.72
kCal/mol. This result is consistent with substantial evidence that the decomposition of NO occurs
on closely located ions of copper [28 – 30].

7. Major Results Summary

We have:
♦ Discovered a mechanism of adsorption compression in zeolites that is different from those
observed previously on flat surfaces (as we have studied and Ertl observed).
♦ Measured thermodynamic adsorption capacity, am, of zeolites (defined where attraction to the
surface is compensated by repulsions between adsorbate molecules). am gives a fraction of
geometric (BET) surface area which is active and actually covered by adsorbate. For zeolites am
≈ 10 % of geometric (BET) capacity. Our results indicate the necessity for a new standard for
surface areas of microporous catalysts.
♦ Measured adsorption isotherms for decomposition of NO on Cu-ZSM-5. Confirmed strong
adsorption compression of NO with energies in the range of 7 kCal/mol to 9 kCal/mol.
♦ Measured rates of NO decomposition on Cu-ZSM-5 as function of T in the range from 350oC
to 550oC. To analyze the results, we developed an Ono-Kondo correction to Langmuir-
Hinshelwood (OK-LH) model taking into account adsorbate-adsorbate interactions. Comparisons
of OK-LH model with our experimental data indicate that adsorption compression is a key factor
in catalytic decomposition of NO on Cu-ZSM-5.

44 
 
♦ Found fundamental correlation between reaction rate and adsorption compression for NO on
Cu-ZSM-5, consistent with recent findings of other authors that “NO-decomposition reaction
occurs on the Cu+…Cu+ pairs” [27 - 30].

8. Fundamental and Practical Implications

- We have developed a new understanding of catalytic reaction mechanisms: adsorption
compression is one of the key factors;
- We have proposed a new model for catalysis kinetics, OK-LH, which takes into account
adsorbate-adsorbate interactions and gives correlation between adsorption compression
and reaction rate;
- We have developed a new understanding of how to design and optimize catalysts;
adsorption compression can dramatically boost catalytic reactions.

9. References for Part I

1. Ono, S. and Kondo, S., "Molecular Theory of Surface Tension in Liquids", Encyclopedia
of Physics (ed. S. Flügge), vol. 10, p. 134, Springer - Verlag, Berlin - Gottingen -
Heidelberg (1960).
2. Rowlinson, J.S. and Widom, B., "Molecular Theory of Capillarity", Clarendon Press,
Oxford (1982); Lane, J.E. and Johnson, C.H.J., Aust. J. Chem. 20, 611 (1967).
3. Lane, J.E., Aust. J. Chem. 21, 827 (1968); Altenberger, A.R. and Stecki, J., Chem. Phys.
Letters. 5, 29 (1970).
4. Aranovich G.L. and Donohue M.D. "ANALYSIS OF ADSORPTION ISOTHERMS:
Lattice Theory Predictions, Classification of Isotherms for Gas-Solid Equilibria, and
Analogy in Adsorption Behavior between Gas and Binary Liquids", Journal of Colloid
and Interface Science, 1998, v.200, p.273-290.
5. Aranovich G.L. and Donohue M.D.”Phase Loops in Density Functional Theory
Calculations of Adsorption in Nano-Scale Pores”, Physical Review E, 1999, v.60, n.5,
p.5552-5560.

45 
 
 6. Sangwichien C., Aranovich G.L., and Donohue M.D., “Density Functional Theory
Predictions of Adsorption Isotherms with Hysteresis Loops”, Colloids and Surfaces A:
Physicochemical and Engineering Aspects, 2002, v. 206, p. 313-320.
7. Aranovich G.L. and Donohue M.D. “Lattice Density Functional Theory Predictions of
Order-Disorder Phase Transitions”,The Journal of Chemical Physics, 2000, v. 112 (5), p.
2361-2366.
8. Aranovich G.L. and Donohue M.D. “Vapor Adsorption on Microporous Adsorbents”,
Carbon, 2000, v.38, p. 701-708.
9. Aranovich G.L. and Donohue M.D. "Adsorption of Supercritical Fluids". Journal of
Colloid and Interface Science, 1996, v.180, p.537-541.
10. Donohue M.D., and Aranovich G.L., “A New Classification of Isotherms for Gibbs
Adsorption of Gases on Solids”, Fluid Phase Equilibria, 1999, v.158-160, p.557-563.
11. Aranovich G.L., and Donohue M.D. "Surface Compression in Adsorption Systems",
Colloids and Surfaces A, 2001, v.187-188, p.95-108.
12. Sircar, S. and Myers, A.L., AIChE Journal 1973, v.19, 159.
13. Nakahara, T., Hirata, M., and Omori, T., J. Chem. Eng. Data 1974, 19, 310.
14. Talu, O. and Zwiebel, I., AIChE Journal 1986, v.32, 1263.
15. Hyun, S.H. and Danner, R.P. J. Chem. Eng. Data 1982, v.27, 196.
16. Nolan, J.T., McKeehan, T.W., and Danner, R.P., J. Chem. Eng. Data 1981, v.26, 112.
17. Kaul, B.K., Ind. Eng. Chem. Res. 1987, v.26, 928-933.
18. Szepesy, L. and Ilies, V., Acta Chim. Hung. 1963, v.35, 37.
19. Valenzuela, D.P. and Myers, A.L. Adsorption Equilibrium Data Handbook; Prentice
Hall: New Jersey, 1989.
20. Yi, H., Deng, H., Tang, X., Yu, Q., Zhou, X., and Lju, H., Journal of Hazardous
Materials, 2012, v.203-204, 111-117.
21. Deng, H., Yi, H., Tang, X., Yu, Q., Ning, P., Yang, L., Chemical Engineering Journal,
2012, v. 188, 77-85.
22. Coulomb, J.P., Suzanne, J., Bienfait, M., Matecki, M., Thomy, A., Croset, B., and Marti,
C., J. Physique, 1980, v. 41, 1155-1164.
23. IUPAC Commission on colloid and surface chemistry including catalysis, Pure Appl.
Chem. 57 (1985) 603.

46 
 
 24. “Recommendations for the Characterization of Porous Solids”, IUPAC Commission on
Colloid and Surface Chemistry, Pure Appl. Chem. 66 (1994) 1739.
25. Davis, M. E.; Davis, R. J. Fundamentals of Chemical Reaction Engineering, Chapter 5
“Heterogeneous Catalysis”; McGraw Hill: New York, 2003.
26. Artioli, N., Lobo, R.F., and Iglesia, E., J. Phys. Chem. 2013, v.117, 20666-20674.
27. Lee, D.K., “Thermodynamic features of the Cu-ZSM-5 catalyzed NO decomposition
reaction”, Korean Journal of Chemical Engineering, 2006, v. 23 (4), 547 – 554.
28. Y. Kuroda, R. Kumashiro, T. Yoshimoto, and M. Nagao, Characterization of active sites
on copper ion-exchanged ZSM-5-type zeolite for NO decomposition, Phys. Chem. Chem.
Phys., 1999, 1, 649-656.
29. T. Yumura, S. Hasegawa, A. Itadani, H. Kobuyashi, and Y. Kuroda, Materials, 2010, 3,
2516-2535.
30. A. Itadani, Y. Koroda, M. Tanaka, M. Nagao, Unambiguous evidence supporting the
decomposition reaction of NO on two types of monovalent copper-ion in Cu-ZSM-5
zeolite, Microporous and Mesoporous Materials, 2005, 86, 159-165.
31. Aranovich G.L., and Donohue M.D., “ADSORPTION COMPRESSION: An Important
New Aspect of Adsorption Behavior and Capillarity”, Langmuir 19, 2722-2735 (2003).
32. Du, X. and Wu, E., “Porosity of microporous zeolites A, A, and ZSM-5 studied by small
angle X-ray scattering and nitrogen adsorption”, Journal of Physics and Chemistry of
Solids 68 (2007), pp. 1692-1699.
33. Nancy Artioli, Raul F. Lobo, and Enrique Iglesia, “Catalysis by Confinement: Enthalpic
Stabilization of NO Oxidation Transition States by Micropororous and Mesoporous
Siliceous Materials”, The Journal of Physical Chemistry C, 2013, v.117, pp.
20666−20674.

47 
 
PART II. Demonstration of Catalysis under Adsorption Compression

1. Introduction

Nitrogen oxides (NOx), which include nitrous oxide (N2O), nitric oxide (NO) and nitrogen
dioxide (NO2), are some of the most dangerous exhaust gases emitted from automobile engines
and industries. NOx is responsible for acid rain, photochemical smog and harmful effects on
human health. The large binding energy of N-O impedes the abatement of NOx exhaust to occur
at low temperature. Efficient catalysts for removal of NOx is pressingly needed for meeting the
future emission regulations.1

Copper zeolites have been investigated extensively for NO decomposition due to their
excellent activity and stabilities. Direct decomposition of NO to N2 and O2 is a robust strategy
for NOx removal because a co-reactant is not required. Ever since the superior activity of
CuZSM-5 in direct NO decomposition was discovered by Iwamoto and co-workers,2-3 great
efforts have been devoted to investigate a variety of Cu-containing zeolites, including CuFER,4
CuZSM-11,5-6 CuMOR,7 CuBEA,8 CuSSZ-13,9 as catalysts for NO decomposition. The active
sites on Cu-containing zeolites for direct NO decomposition and the associated catalytic
mechanisms have been comprehensively studied. According to previous literatures,10 one of the
most widely accepted mechanisms of direct NO decomposition over Cu exchanged zeolites
include two NO molecules first adsorbed on a Cu dimer site on CuZSM-5, two proximal NO
molecules then react to form the N-N bond and produce intermediate N2O, which is the rate-
limiting step. The N2O molecule reacts with the [Cu2+–O–Cu2+]2+ site and subsequently form N2
and O2 as final products. Therefore, the distance between two Cu active sites can be critical for
the formation of N-N bond. In Cu-containing zeolites prepared by ion exchange method, Cu ions
are expected to connect with framework Al sites, the distances between Al sites significantly
depend on Si/Al ratios or Al distributions in zeolites with different topologies. Therefore, as
expected, Cu-Cu distances are different in CuZSM-5 with different Si/Al ratios or other Cu-
containing zeolites. Meanwhile, it is well-known that the activity of NO decomposition on Cu-
exchanged catalysts exhibited interesting “volcano-type” dependence on temperature, which is
different from common reaction behavior (i.e., generally the conversion increased with
increasing temperature, then reached to constant at high temperature). So far, only few works
discussed this phenomenon. Iglesia et al.11 and Lee12 ascribed the decrease in NO decomposition

48 
 
rates at high temperatures to unfavorable NO adsorption thermodynamics. However, no
systematic investigation was conducted to explain the unconventional “volcano-type” behavior.
Furthermore, a new adsorption model is required.

Classical and most widely used adsorption model is based on Langmuir theory,13 which
only takes adsorbate-adsorbent interactions into account and ignores adsorbate-adsorbate
interactions. While adsorption compression theory and the Ono-Kondo coordinate developed by
two of the current authors provided us strong tools to understand the interactions between
neighboring molecules adsorbed on proximal active sites,13 e.g. NO molecules adsorbed on Cu
dimers in Cu-exchanged zeolites, which is the preliminary and rate-limitation step for NO
decomposition. Since compression between two adsorbates strongly relied on the distances
between them, by understanding the way that distance affecting the compression effect could
help us discover the influence of distances of proximal active sites on catalytic reactions and
hence develop efficient strategies to design Cu based catalysts for NOx removal.

Herein, this annual report summarizes the progress and outcome of the research project
funded by ARPA-E DE-AR0000708 in aim of investigating adsorption and compression effects
of Cu-exchanged zeolites in direct NO decomposition reaction. The contents include
establishment of catalytic reaction setup, synthesis and characterizations of Cu-exchanged
zeolites. Reaction performance and kinetics studies of direct NO decomposition, analysis of
adsorption isotherm for NO in Ono-Kondo coordinates and illustration of adsorption
compression effect in direct NO decomposition over Cu-exchanged zeolites. Inspired by the
study on adsorption compression of Cu dimers in Cu-exchanged zeolites, a robust Cu/CeO2
catalyst has been developed to achieve efficient abatement of NOx at low temperatures,
preliminary data of NO decomposition over this catalyst was also provided.

2. Experimental Setup

The catalytic evaluation system was established as shown in Fig. 1. NO decomposition and
isothermal NO adsorption were performed in this setup. In a typical experiment, different
reactant gases were adjusted at desired ratios with mass flow controller (MFC). Catalytic
reaction was conducted in the fixed-bed flow reactor at atmospheric pressure. The catalyst was

49 
 
loaded into a microflow quartz reactor (7 mm i.d.) and reaction temperature was monitored by a
Carbolite GERO vertical tube furnace system. To determine the conversions of reactants and the
formation of products, an FTIR equipped with a 5 m gas cell and a MCT detector (Nicolet 6700,
Thermo Electron Co.) as well as a gas chromatograph equipped with a BID detector (GC-2010
plus, Shimadzu) were employed. Particularly, NO, N2O, and NO2 were determined by FTIR at
1905 cm-1, 2237 cm-1 and 1630 cm-1 respectively with resolution of 8 cm-1. N2 was determined
by GC using HP-POLT Molesieve column.

Figure 1. Experimental setup of catalyst evaluation system.

3. Synthesis of Cu Substituted Zeolites

Three types of zeolites with different frameworks including Zeolite Socony Mobil–5 (ZSM-5),
Mordenite (MOR) and SSZ-13 were prepared. ZSM-5 was synthesized with three Si/Al ratios:
11.5, 20 and 30. Copper was exchanged into the zeolite frameworks with ion-exchange method.
The schematic illustration of zeolite preparation was shown in Fig. 2.

50 
 
 Figure 2. Schematic illustration of Cu-zeolites preparation.

The pristine ZSM-5 with different Si/Al molar ratios (Si/Al = 11.5, 20, 30, measured by X-

ray fluorescence spectrum) as well as SSZ-13 (Si/Al = 11) were synthesized by hydrothermal

method. Mordenite was purchased from Zeolyst Co. (CBV 21A, Si/Al = 10). The obtained

zeolite frames were first thoroughly exchanged with Na+ and then exchanged with Cu with liquid

phase ion-exchange method.

3.1. Material Characterizations and Results

Scanning electron microscopy (SEM) images were taken on a JEOL 6700F field emission

electron scanning microscope operating at 10.0 kV. The representative SEM image of each type

of zeolite was shown in Fig. 3. ZSM-5 exhibited rectangular parallelepiped shapes with the

length of 1.4 μm, and the width of 1 μm (Fig. 3a). Commercial mordenite (Si/Al = 10) exhibited

small particles with the diameter of 100-200 nm, and some big granules around 400 nm can also

be observed, which were aggregated by those small particles (Fig. 3b). For SSZ-13 (Si/Al = 11),

the particles were individual cube-shaped crystals measuring approximately 400 nm in size,

agglomerated by many smaller particles (Fig. 3c).

51 
 
 Figure 3. SEM images of (a) ZSM-5, (b) MOR-10 and (c) SSZ-13.

X-ray diffraction (XRD) patterns were obtained from a PANalytical X’Pert X-ray

diffractometer equipped with a Cu Kα radiation source (λ = 1.5406 Å). Comparing XRD patterns

of the samples of each zeolite with corresponding standard reference, the characteristic

reflections of each zeolite topology were demonstrated, i.e. MFI (JCDPS No. 37-0359) for ZSM-

5 (Fig. 4a), MOR (JCDPS No. 29-1257) for Mordenite (Fig. 4b), and CHA (JCDPS No. 47-

0762) for SSZ-13 (Fig. 4c). It’s also revealed in XRD patterns that ion exchanging with Cu2+

didn’t change the crystal structures of zeolites as the diffraction patterns of Cu exchanged

zeolites are almost identical to Na type zeolites. Additionally, no any Cu2O (JCDPS No. 34-

1354) or CuO (JCDPS No. 44-0706) peak can be distinguished from XRD patterns of Cu-

exchanged zeolites, which indicates Cu exist as pure ionic form in zeolite pores.

Figure 4. XRD patterns of (a) CuZSM-5 with different Si-Al ratios and Cu loadings, (b)

CuMOR-10 and (c) CuSSZ-13 with comparisons to the database.

52 
 
 Bulk Si/Al ratios of the prepared ZSM-5 and SSZ-13 samples and commercial Mordenite

were measured by X-ray fluorescence (XRF) on a Bruker-AXS S4. The Cu and Al contents were

determined by inductively coupled plasma mass spectrometry (ICP-MS) using a PerkinElmer

Elan DRC II Quadrupole ICP-MS after dissolution of the zeolites in HF. Nitrogen adsorption

measurements were measured on a Micromeritics ASAP 2010 instrument. The samples were

degassed under vacuum at 300°C for 4 h prior to the measurements. Specific surface area (SSA)

was calculated using the Brunauer-Emmett-Teller (BET) theory. Table 1 summarizes the Cu

content measured by ICP-MS and BET surface area measured by N2 adsorption-desorption

isotherm of each Cu zeolites.

Table 1. Summary of Cu Content and BET Surface Area of Cu-zeolites

Cu exchanged zeolitesa Cu content (wt%)b BET surface area (m2/g)

CuZSM-5-11.5 (0.51) 3.10 373
CuZSM-5-20 (0.51) 2.64 373
CuZSM-5-30 (0.42) 2.21 406
CuMOR-11 (0.50) 3.02 401
CuSSZ-13-11 (0.59) 3.98 593
a
The Si/Al ratio of each zeolite was determined by XRF.

b
The Cu content of each zeolite was measured by ICP-MS.

4. Catalytic Studies of Direct NO Decomposition and “Volcano Type” Dependence on

Temperature.

Catalytic direct NO decomposition reactions over Cu exchanged zeolites were conducted

in a pre-described flow reactor system at atmospheric pressure. 1 g of each Cu-zeolite (4060

53 
 
mesh) was loaded into a microflow quartz reactor as catalyst. The Cu-zeolite was pretreated at

500 oC in He before each reaction. The catalytic performance was evaluated using 500 ppm NO

balanced by He and at different temperatures range from 350 oC to 550 oC. NO conversion, N2

yield and rate of N2 formation was evaluated with every 50 oC increment. The reaction was

carried out at each temperature for at least 45 mins in order to allow reaction to reach

equilibrium and conversion of NO remain constant before the measurements were taken. The NO

conversion and N2 yield were calculated using the following equations:

[NO]inlet -[NO]outlet
NO conversion = ×100%
[NO]inlet

2[N2 ]outlet
N2 yield= ×100%
[NO]inlet -[NO]outlet

where [NO]inlet refers to the concentration of NO in the inlet, and [NO]outlet or [N2]outlet refers to

the concentration of NO or N2 in the outlet. NO conversion, N2 yield and rate of N2 formation

over each Cu-zeolite at the temperature range from 350 oC to 550 oC was shown in Fig. 5. As it’s

exhibited, direct NO decomposition over all Cu substituted zeolites shows typical volcano-type

behavior, as the NO conversion, N2 yield and rate of N2 formation all reach a maximum and then

decrease within the temperature range which was measured. The catalytic activity of all Cu

substituted zeolites in the present study follows the trend of CuZSM-5-11.5 (0.51) > CuZSM-5-

20 (0.51) > CuZSM-5-30 (0.42) > CuMOR-10(0.50) > CuSSZ-13-10 (0.59).

54 
 
Figure 5. Catalytic activities of direct NO decomposition over different Cu-zeolites at

temperatures from 350 oC to 550 oC. (a) NO conversions. (b) N2 yields. (c) Rate of N2 formation.

NO concentration: 5000 ppm, balance gas: He, contact time: 0.05 min*gcatal/ml.

5. Kinetics Studies of Direct NO Decomposition over Cu Substituted Zeolites

Fig. 6 showed the Arrhenius plot of NO decomposition over different Cu-exchanged

zeolites. The apparent activation energy (Eapp) could be obtained from the slope in the Arrhenius

plot where (lnK) was plotted against the reciprocal temperature (Fig. 6a). Comparing the

apparent activation energies over different Cu-exchanged zeolites, it is obvious that activation

energies in low temperature followed the sequence of CuZSM-5-11.5 (0.51) < CuZSM-5-20

(0.51) < CuZSM-5-30 (0.42) < CuMOR-10 (0.50) < CuSSZ-13-10 (0.59), which suggests that

adsorption-compression effect is beneficial to facilitate catalytic NO decomposition over Cu-Cu

dimers at low temperature (T < 400 oC). Meanwhile, the apparent activation energies at high

temperature followed the invertible sequence of CuZSM-5-11.5 (0.51) > CuZSM-5-20 (0.51) >

CuZSM-5-30 (0.42) > CuMOR-10 (0.50) > CuSSZ-13-10 (0.59), suggesting that adsorption-

compression plays a prohibitive role in catalytic direct NO decompositions at high temperature

(T > 400 oC).

55 
 
Figure 6. (a) Arrhenius plot of apparent activation energies of direct NO decomposition over

different Cu-zeolites. (b) Calculated apparent activation energies at low temperature (T<400 oC,

red) and high temperature (T>400 oC, blue) regimes of direct NO decomposition over different

Cu-zeolites.

6. Isothermal NO Adsorption Measurements

Isothermal NO adsorption on different Cu-containing zeolites catalysts was conducted in

the fixed-bed flow reactor at atmospheric pressure. The same pretreatment (He, 500 oC) was

implemented. After pretreatment, the temperature was cooled down to 80 oC under pure He

atmosphere and kept the temperature at 80 oC for adsorption measurement. The feed

concentrations of NO were adjusted to 300 ppm, 500 ppm, 800 ppm and 1500 ppm by mixing

pure He and 0.5% NO/He gas. The NO/He mixture (201 h-1) was fed to the catalyst. The

composition of the effluent stream was continuously monitored for the entire length of the

experiment until all concentrations of different gas species (NO, N2O, N2) were stable. The

detection system was the same experimental setup as above. The measured outlet concentration
56 
 
of NO, N2O and N2 in the time on stream during NO isothermal adsorptions of different NO

concentrations on CuZSM-5-11.5 (0.51) at 80 oC was chosen as an example and was shown in

Fig. 7. The amount of NO adsorbed on the catalyst at the end of the adsorption measurement was

evaluated by measuring the overall uptake of NO and the quantities of N2O, NO2 and N2

transiently produced, as follows:

NOads = NOuptake – 2N2Oprod – 2 N2, prod –NOhold up

where:

NOads represents NO totally adsorbed (as calculated from mass balance in the gas phase);

NOuptake represents the amount of NO consumed (directly measured in the gas phase);

N2Oprod represents the amount of N2O produced (directly measured in the gas phase);

N2, prod represents the amount of N2 produced (directly measured in the gas phase);

NOhold up represents the amount of NO necessary to fill the dead volumes of the experimental set-

up (evaluated in the opportune blank tests).

The amounts of these adsorbed species were estimated by integrating the corresponding

signals as a function of time; all these quantities are reported in Table 2 for CuZSM-5-11.5

(0.51).

57 
 
Figure 7. Outlet concentrations of NO, N2O and N2 during NO isothermal adsorptions of different
NO concentrations on CuZSM-5-11.5 (0.51) at 80 oC. (a) 300 ppm NO. (b) 500 ppm NO. (c) 800
ppm NO. (d) 1500 ppm NO.

Table 2. Amounts of Gaseous Species Evaluated by the Integration of the Signals during the
Adsorption of Different Concentrations of NO on CuZSM-5-11.5 (0.51) in Ono-Kondo
Coordinates.
NOads NOuptake N2Oprod N2prod NOhold up
Concentrations of NO
μmol g-1 μmol g-1 μmol g-1 μmol g-1 μmol g-1

300 ppm 199 527 125 36 6

500 ppm 305 703 154 40· 10

800 ppm 441 903 160 63 16

1500 ppm 686 1188 170 70 22

58 
 
 The compression energies can be calculated followed the principles introduced in the

previous section from the slopes in Fig. 8, which had been listed in Table 3 and related to the

average Cu-Cu distances of each Cu substituted zeolites measured by XAFS techniques in previous

literature with very similar Cu loadings and Si/Al ratios. It can be found that calculated

compression energy followed the sequence of CuZSM-5-11.5 (0.51) > CuZSM-5-20 (0.51) >

CuZSM-5-30 (0.42) > CuMOR-10 (0.50) > CuSSZ-13-10 (0.59), which agrees well with catalytic

activity at low temperature regime (T<400 oC).

Ec = 5.38 eV
Ec = 7.77 eV
20

Ecp = 6.02 eV
ln (a/xb)

Ec = 7.05 eV
19
CuZSM-5-11.5
CuZSM-5-20
Ec = 6.51 eV CuZSM-5-30
CuMOR-10
CuSSZ-13-11
180 200 400 600 800 1000
Absorbed NO amount (a, μmol g-1)

Figure 8. Adsorption isotherms for NO on Cu-Zeolites at 80 oC in Ono-Kondo coordinates.

Table 3. Summary of Compression Energy Calculated from Ono-Kondo Coordinates and Average Cu-Cu
Distances of Cu Dimers in Cu Zeolites in Present Study

Catalysts Cu···Cu distances (Å) Compression Energy (kCal mol-1)f

CuZSM-5-11.5 (0.51) 2.84a 7.77
CuZSM-5-20 (0.51) 2.94b 7.05
CuZSM-5-30 (0.42) 3.13c 6.51
CuMOR-11 (0.50) 3.34d 6.02

59 
 
 CuSSZ-13-11 (0.59) 3.76e 5.38

a
J. Am. Chem. Soc. 125 (2003) 7629-7640

b
J. Phys. Chem. 98 (1994) 10832-10846

c
Catal. Lett. 5 (1990) 189-196

d
J. Mater. Chem. 7 (1997) 1917–1923

e
J. Phys. Chem. C 116 (2012) 4809-4818

f
Obtained from NO isothermal adsorption measurements

7. Effect of Cu···Cu Dimer Distances on Catalytic Activity of NO Decomposition over

Cu Substituted Zeolites

By plotting the calculated compression energy at T < 400 oC versus Cu-Cu distance (Fig.

9a), it can be revealed that shorter Cu-Cu distance gave rise to larger compression energy in Cu-

exchanged zeolites. Also, by plotting the rate of N2 formation at T < 400 oC versus compression

energy (Fig. 9b), we can see the rate of N2 formation increases as compression energy increases.

Hence, it can be concluded that the calculated compression energy is bridged to the average Cu-

Cu distance of each Cu-zeolite, which is crucial for catalytic activity of direct NO decomposition.

60 
 
 Figure 9. (a) Dependence of compression energy on Cu-Cu distance. (b) Dependence of

rates of N2 formation at 400 oC on compression energy.

8. Preliminary Results for NO Decomposition over Cu/CeO2

Recently, a series of research works demonstrated that cerium oxide can provide oxygen

vacancies for absorption and release of oxygen as well as the regeneration of catalytic active

sites, which is ideal for NOx decomposition.15,16 Inspired by the conclusion of our studies on Cu-

zeolites and previous literatures, it is believed CeO2 can serve as a better support for Cu(I)

dimers, which can give rise to robust catalytic ability in NO decomposition. To the best of our

knowledge, no catalyst which can achieve direct NO decomposition at room temperature has yet

been reported. Here as shown in Fig. 10, we report the excellent performance of a 5% Cu/CeO2

catalyst which was capable of achieve direct NO decomposition at room temperature with almost

100% conversion and 100% selectivity to N2 and O2. The catalyst was active for more than 400

min. If CO as a reducing agent was introduced, the catalyst was able to reach steady state, 100%

conversion of NO and 100% selectivity to N2 was achieved at 100 oC (Fig. 11). Systematic

61 
 
investigation of this catalyst is undergoing, details of catalyst preparation and experiments were

kept confidential.

NOx concentrations (ppm) 500 NO

N2O 500 ppm NO 30 oC
400 N2
O2
300

200

100

0 200 400 600 800 1000

Time (min)

Figure 10. Performance of direct NO decomposition over 5% Cu/CeO2 at 30 oC.

100 100
NO conversion (%)

N2 selectivity (%)
80 80
500 ppm NO+ 1% CO
60 60

40 40

20 20

0 0
50 100 150 200
Reaction temperature (oC)
Figure 11. Performance of NO+CO reaction over 5% Cu/CeO2.

62 
 
 9. Comparisons with experimental data on rate vs. temperature for NO on Cu-ZSM-5

Figure 12 shows reaction rate as a function of temperature predicted by equations (51)-

(53) of Part I for E0 =19.5 kCal/mol, εs = 34.1 kCal/mol, xb = 0.001, z = 2, K0 = 3, and different
values of ε: 9 kCal/mol (1) and 0 (2). Also shown in Figure 12 are experimental points for NO
on zeolite ZSM-5 at Si/Al = 11.5 and Cu/Al = 0.55.

Figure 12. Reaction rate, rb, as a function of temperature predicted by equations (51)-(53) of Part I for E0 =19.5
kCal/mol, εs = 34.1 kCal/mol, xb = 0.001, z = 2, K0 = 3, and different values of ε: 9 kCal/mol (1) and 0 (2). Also
shown are experimental points for NO on zeolite ZSM-5 at Si/Al = 11.5 and Cu/Al = 0.55.

As seen from Figure 12, without adsorption compression (at ε = 0) reaction rate is very small.
However, at ε = 9 kCal/mol , reaction rate goes up dramatically and predictions of equations
(51)-(53) of Part I are in agreement with the experimental data.
Figure 13 shows reaction rate as a function of temperature predicted by equations (51)-
(53) of Part I for E0 =19.5 kCal/mol, εs =34.1 kCal/mol, xb =0.001, z =2, K0 =3, and ε = 8.9
kCal/mol. Also shown in Figure 13 are experimental points for NO on zeolite ZSM-5 at Si/Al =
20 and Cu/Al = 0.51.

63 
 
Figure 13. Reaction rate, rb, as a function of temperature predicted by equations (51)-(53) of Part I for E0 =19.5
kCal/mol, εs = 34.1 kCal/mol, xb = 0.001, z = 2, K0 = 3, and ε = 8.9 kCal/mol . Also shown are experimental points
for NO on zeolite ZSM-5 at Si/Al = 20 and Cu/Al = 0.51.

As seen from Figure 13, at ε = 8.9 kCal/mol predictions of equations (51)-(53) of Part I are in
reasonably good agreement with the experimental data for NO on zeolite ZSM-5 at Si/Al = 20
and Cu/Al = 0.51. The difference between Figures 12 and 13 can be explained by slight decrease
of distance between ions of Cu+ for Si/Al =11.5 and Cu/Al =0.55 compared to Si/Al =20 and
Cu/Al =0.55.
Figure 14 shows reaction rate as a function of temperature predicted by equations (51)-
(53) of Part I for E0 =19.5 kCal/mol, εs = 34.1 kCal/mol, xb = 0.001, z = 2, K0 = 3, and ε = 8.7
kCal/mol. Also shown in Figure 14 are experimental points for NO on zeolite ZSM-5 at Si/Al =
20 and Cu/Al = 0.31.

64 
 
Figure 14. Reaction rate, rb, as a function of temperature predicted by equations (51)-(53) for E0 =19.5 kCal/mol, εs
= 34.1 kCal/mol, xb = 0.001, z = 2, K0 = 3, and ε = 8.7 kCal/mol. Also shown are experimental points for NO on
zeolite ZSM-5 at Si/Al = 20 and Cu/Al = 0.31.

Figures 15 – 18 give reaction rates as functions of temperature predicted by equations

(51)-(53) of Part I for E0 =19.5 kCal/mol, εs = 34.1 kCal/mol, xb = 0.001, z = 2, and K0 = 3. In
Figure 15, ε = 8.6 kCal/mol and experimental points are for NO on zeolite ZSM-5 at Si/Al = 30
and Cu/Al = 0.42. In Figure 16, ε = 8.4 kCal/mol and experimental points are for NO on zeolite
ZSM-5 at Si/Al = 30 and Cu/Al = 0.35. In Figure 17, ε = 8.4 kCal/mol and experimental points
are for NO on zeolite ZSM-5 at Si/Al = 50 and Cu/Al = 0.69. In Figure 18, ε = 7.0 kCal/mol and
experimental points are for NO on zeolite ZSM-5 at Si/Al = 50 and Cu/Al = 0.55.

65 
 
Figure 15 (left frame). Reaction rate, rb, as a function of temperature predicted by equations (51)-(53) of Part I for
E0 =19.5 kCal/mol, εs = 34.1 kCal/mol, xb = 0.001, z = 2, K0 = 3 and ε = 8.6 kCal/mol. Also shown are experimental
points for NO on zeolite ZSM-5 at Si/Al = 30 and Cu/Al = 0.42.

Figure 16 (right frame). Reaction rate, rb, as a function of temperature predicted by equations (51)-(53) of Part I for
E0 =19.5 kCal/mol, εs = 34.1 kCal/mol, xb = 0.001, z = 2, K0 = ,3 and ε = 8.4 kCal/mol. Also shown are experimental
points for NO on zeolite ZSM-5 at Si/Al = 30 and Cu/Al = 0.35.

Figure 17 (left frame). Reaction rate, rb, as a function of temperature predicted by equations (51)-(53) of Part I for E0
=19.5 kCal/mol, εs = 34.1 kCal/mol, xb = 0.001, z = 2, K0 =3, and ε = 8.4 kCal/mol. Also shown are experimental
points for NO on zeolite ZSM-5 at Si/Al = 50 and Cu/Al = 0.69.

Figure 18 (right frame). Reaction rate, rb, as a function of temperature predicted by equations (51)-(53) of Part I for
E0 =19.5 kCal/mol, εs = 34.1 kCal/mol, xb = 0.001, z = 2, K0 = 3, and ε = 7.0 kCal/mol. Also shown are experimental
points for NO on zeolite ZSM-5 at Si/Al = 50 and Cu/Al = 0.55.

As seen from Figures 15 – 18, at low concentration of Cu, experimental points are shifted
toward lower temperatures. In our opinion, this is because, at lower concentration of copper,
some ions of copper are oxidized and become Cu2+, and this changes εs . Enthalpy of adsorption
onto oxidized Cu2+ is 27.8 kCal/mol (see Ref. [27] of Part I). If we use the average between 34.1

66 
 
kCal/mol (Cu+) and 27.8 kCal (Cu2+) (which gives about 31 kCal/mol), agreement between
theory and experiment becomes much better. This is illustrated on Figures 19 – 22 with
normalization factor K0 = 4.6).

Figure 19 (left frame). Reaction rate, rb, as a function of temperature predicted by equations (51)-(53) of Part I for
E0 =19.5 kCal/mol, εs = 31 kCal/mol, xb = 0.001, z = 2, K0 = 4.6, and ε = 8.6 kCal/mol. Also shown are experimental
points for NO on zeolite ZSM-5 at Si/Al = 30 and Cu/Al = 0.42.

Figure 20 (right frame). Reaction rate, rb, as a function of temperature predicted by equations (51)-(53) of Part I
for E0 =19.5 kCal/mol, εs = 31 kCal/mol, xb = 0.001, z = 2, K0 = 4.6, and ε = 8.4 kCal/mol. Also shown are
experimental points for NO on zeolite ZSM-5 at Si/Al = 30 and Cu/Al = 0.35.

Figure 21 (left frame). Reaction rate, rb, as a function of temperature predicted by equations (51)-(53) of Part I for E0
=19.5 kCal/mol, εs = 31 kCal/mol, xb = 0.001, z = 2, K0 =4.6, and ε = 8.4 kCal/mol. Also shown are experimental
points for NO on zeolite ZSM-5 at Si/Al = 50 and Cu/Al = 0.69.

Figure 22 (right frame). Reaction rate, rb, as a function of temperature predicted by equations (51)-(53) of Part I
for E0 =19.5 kCal/mol, εs = 31 kCal/mol, xb = 0.001, z = 2, K0 = 4.6, and ε = 7.0 kCal/mol. Also shown are
experimental points for NO on zeolite ZSM-5 at Si/Al = 50 and Cu/Al = 0.55.

67 
 
 10. References for Part II

1. National Archives and Records Administration, Fed. Regist. 2016, 81, 73478–74274.
2. Iwamoto, M.; Furukawa, H.; Mine, Y.; Uemura, F.; Mikuriya, S. I.; Kagawa, S., Copper(Ii)
Ion-Exchanged Zsm-5 Zeolites as Highly-Active Catalysts for Direct and Continuous
Decomposition of Nitrogen Monoxide. J Chem Soc Chem Comm 1986, (16), 1272-1273.
3. Iwamoto, M.; Yahiro, H.; Mizuno, N.; Zhang, W. X.; Mine, Y.; Furukawa, H.; Kagawa, S.,
Removal of Nitrogen Monoxide through a Novel Catalytic Process .2. Infrared Study on
Surface-Reaction of Nitrogen Monoxide Adsorbed on Copper Ion-Exchanged Zsm-5 Zeolites.
J Phys Chem-Us 1992, 96 (23), 9360-9366.
4. Attfield, M. P.; Weigel, S. J.; Cheetham, A. K., On the nature of nonframework cations in a
zeolitic deNO(x) catalyst - A synchrotron X-ray diffraction and ESR study of Cu-ferrierite. J
Catal 1997, 172 (2), 274-280.
5. Kustova, M. Y.; Rasmussen, S. B.; Kustov, A. L.; Christensen, C. H., Direct NO
decomposition over conventional and mesoporous Cu-ZSM-5 and Cu-ZSM-11 catalysts:
Improved performance with hierarchical zeolites. Appl Catal B-Environ 2006, 67 (1-2), 60-67.
6. Xie, P. F.; Ma, Z.; Zhou, H. B.; Huang, C. Y.; Yue, Y. H.; Shen, W.; Xu, H. L.; Hua, W. M.;
Gao, Z., Catalytic decomposition of N2O over Cu-ZSM-11 catalysts. Micropor Mesopor Mat
2014, 191, 112-117.
7. Granger, P.; Parvulescu, V. I., Catalytic NOx Abatement Systems for Mobile Sources: From
Three-Way to Lean Burn after-Treatment Technologies. Chem Rev 2011, 111 (5), 3155-3207.
8. Mihai, O.; Widyastuti, C. R.; Andonova, S.; Kamasamudram, K.; Li, J. H.; Joshi, S. Y.;
Currier, N. W.; Yezerets, A.; Olsson, L., The effect of Cu-loading on different reactions
involved in NH3-SCR over Cu-BEA catalysts. J Catal 2014, 311, 170-181.
9. Paolucci, C.; Parekh, A. A.; Khurana, I.; Di Iorio, J. R.; Li, H.; Caballero, J. D. A.; Shih, A. J.;
Anggara, T.; Delgass, W. N.; Miller, J. T.; Ribeiro, F. H.; Gounder, R.; Schneider, W. F.,
Catalysis in a Cage: Condition-Dependent Speciation and Dynamics of Exchanged Cu Cations
in SSZ-13 Zeolites. J Am Chem Soc 2016, 138 (18), 6028-6048.
10. Sajith, P. K.; Shiota, Y.; Yoshizawa, K., Role of Acidic Proton in the Decomposition of NO
over Dimeric Cu(I) Active Sites in Cu-ZSM-5 Catalyst: A QM/MM Study. Acs Catal 2014, 4
(6), 2075-2085.

68 
 
11. Moden, B.; Da Costa, P.; Fonfe, B.; Lee, D. K.; Iglesia, E., Kinetics and mechanism of steady-
state catalytic NO decomposition reactions on Cu-ZSM5. J Catal 2002, 209 (1), 75-86.
12. Lee, D. K., Thermodynamic features of the Cu-ZSM-5 catalyzed NO decomposition reaction.
Korean J Chem Eng 2006, 23 (4), 547-554.
13. Aranovich, G. L.; Donohue, M. D., Adsorption compression: An important new aspect of
adsorption behavior and capillarity. Langmuir 2003, 19 (7), 2722-2735.
14. Aranovich, G. L.; Donohue, M. D. Phase Loops in Density-Functional-Theory Calculations of
Adsorption in Nanoscale Pores. Phys. Rev. E - Stat. Physics, Plasmas, Fluids, Relat.
Interdiscip. Top. 1999, 60 (5), 5552–5560.
15. Zabilskiy, M.; Djinović, P.; Tchernychova, E.; Pintar, A. N2O Decomposition over
CuO/CeO2catalyst: New Insights into Reaction Mechanism and Inhibiting Action of H2O and
NO by Operando Techniques. Appl. Catal. B Environ. 2016, 197, 146–158.
16. Zabilskiy, M.; Djinović, P.; Tchernychova, E.; Tkachenko, O. P.; Kustov, L. M.; Pintar, A.
Nanoshaped CuO/CeO2 Materials: Effect of the Exposed Ceria Surfaces on Catalytic Activity
in N2O Decomposition Reaction. ACS Catal. 2015, No. October, 5357–5365.

69 
 
 70