Download FT-IR study of nitric-oxide chemisorbed on Rh/Al2O3

J . Phys. Chem. 1985, 89, 5840-5845
5840
activation energy converted in translational energy is given by26a
TPS = T Ps + T Ps + T Ps =
that the reason for this behavior is the low value of the vibrational
frequencies in the transition state of this step, 7. It appears as
well that the high frequencies in the reactant structure favor a
fast rate.23 Thus, in reactions where rearrangements occur before
groups are lost, Le., in multiple-step processes, it is difficult to
perform any RRKM calculation giving reliable information
without having an accurate knowledge of the potential surface,
and the related parameters.
Kinetic Energy Release. In a unimolecular dissociation, the
kinetic energy release (KZR) may have two contribution^:^^ Te,
the KER due to a potential energy barrier (reverse activation
energy), and T*,the KER associated with the nonfixed energy,
or excess energy. Baer et aI.l3 have obtained at a photon energy
of 13.6 eV (3.8 eV above the products 4) an average kinetic energy
release of 0.74 eV, which corresponds to 19.5% of the total energy.
A statistical distribution of the rvailable energy among the vibrational, rotational, and translational degrees of freedom can
be calculated by25
X
+ k T ( R - 1 ) / 2 + C h v , [ e x p ( h v , / k T )- 11-'
+ (UyPSURC)2 + (24:s
RC
1
(4)
where PS refers to three independent products separation motions,
and R C refers to the direction of the reaction coordinate in the
transition-state geometry. The uRCvector corresponds to the
mass-weighted atomic displacements vector associated with the
imaginary frequency of the transition state. By this technique,
the calculated relation between F and the reverse activation energy
is 47%, at 13.6 eV; the available energy for the products 4 is 2.41
eV, where 0.51 eV is the reverse activation energy and 1.90 eV
is the nonfixed energy. The translational energy coming from
the nonfixed energy can be estimated by partitioning the energy
among all degrees of freedom.26b The three translational degrees
would mean a fraction 3/N, N being the number of internal
degrees of freedom of the reactant. In this way, the global
translational energy is 0.41 eV, 17.1% of the available energy,
which agrees well with the experimental relation, 19.5%. The
absolute value is lower than the experimental value, but it is
probably due to the fact that our reverse activation energy is too
low, since the relative energy of the products seems to be overOn the other hand, in this reaction the proportion
of reverse activation energy converted in kinetic energy calculated
by the Derrick method is probably exaggerated. The two error
factors compensate, giving almost the experimental result.
In summary, the M I N D 0 / 3 method gives a good description
of the reaction C6H@H+' C5H6+' C o . The RRKM/QET
rate constants and the kinetic energy release can be calculated
from the information provided by the theoretical study, leading
to the conclusion that the determining step is the isomerization
from the phenol ion to the keto form, owing to both energetic and
vibrational factors. It can be inferred that, in theoretical studies,
attention has to be paid to both aspects, since cases exist where
the kinetic analysis may invert the predictions made from the
potential energy surface. Nevertheless, the results presented here
should not be extrapolated to other similar radical cations without
cautions. A theoretical analysis in other keto-enol isomerizations
and fragmentations are now under way.
(3)
,=l
where R and {v,]are the number of rotational degrees of freedom
and the vibrational frequencies of the products, respectively. 'The
translational energy is just k T , where T i s calculated from eq 3
from an available energy E above products 4. The statistical
distribution obtained in this way gives 5.3% of the available energy.
The result is in agreement with Baer's calculation, with assumed
vibrational frequencies for the products (CSH6+'+ c o ) . This
value is much lower than the experimental one (19.5%), showing
that in reaction 1 kinetic energy release is not distributed statiscally.
The reproduction of the KER would be important since it would
be a test of the transition-state and the product structures. Recently, an approximate method to determine the KER in terms
of the transition-state reaction coordinate has been developed by
Derrick et
In this method, the proportion of the reverse
-
(23) To obtain insight into the role of the vibrational frequencies in the
rate constant, we are carrying out a study on the system C6H5X+./C6DSX+'
C6H5+/C6D5+ X , where the relative rates will be mainly determined by
vibrational effects.
(24) R. G. Cooks, J. H. Beynon, R. M. Caprioli, and G. R. Lester,
"Metastable Ions", Elsevier, Amsterdam, 1973.
(25) C. E. Klots, J . Chem. Phys., 64, 4269 (1976).
(26) (a) J. R. Christie, P. J. Derrick, and G. J. Richard, J . Chem. Sac.,
Faraday Trans. 2, 74, 304 (1978); (b) N . W. Cole, G. J. Richard, J. R.
Christie, and P. J. Derrick, Org. Mass Spectram.. 1, 337 (1979); (c) Ibid..
J . Am. Chem. SOC.,100, 2904 (1978).
-
z
(U,PSURC)Z
n
E = kT
Y
+
+
Acknowledgment. Computer time was provided by the C.P.D.
of the Ministerio de Educaci6n y Ciencia, made available through
the terminal of the Centre de Cglcul de la Universitat PolitEcnica
de Catalunya and by the Centre de Cglcul of the Universitat de
Barcelona.
Registry No. C6H,0H+', 40932-22-7.
FT-IR Study of Nitric Oxide Chemisorbed on Rh/AI,O,
Jim Liang,* H. P. Wang, and L. D. Spicer*+
Department of Chemistry, University of Utah, Salt Lake City, Utah 84112 (Received: May 20, 1985)
Chemisorption of NO on Rh/AI2O3surfaces has been examined by FT-IR. The spectra are assigned to two forms of Rh(N0)
as well as the Rh(N0)2 species. Apparent interconversion of the linear nitrosyl and dinitrosyl complexes is readily observed
at room temperature. The dinitrosyl complex is characterized both by an invariant ratio of 1743- and 1825-cm-' asymmetric
and symmetric stretch bands with coverage and by isotopic data in 15N0 and the mixed 14N0 and 15N0systems. Force
constants for NO stretching motions and for NO/NO ligand interactions on Rh(N0)2 have been used to successfully calculate
the experimentally observed spectrum for the mixed isotope, dinitrosyl species. Thermal desorption data and displacement
of adsorbed NO with CO are also reported.
Introduction
The characterization of molecules chemisorbed on transition
metals by vibrational spectroscopy has recently attracted renewed
+Present address: Departments of Biochemistry and Radiology, Duke
University, Durham, NC 27710.
0022-3654/85/2089-5840$01 SO10
attention due largely to the availability of FT-IR and computerized grating IR. By use of these techniques difference spectra
for surfaces with and without adsorbed gas can be obtained with
high sensitivity. These spectra typically reveal even very weak
features and in addition permit one to follow the progress of
reactions "in situ" using multiple
0 1985 American Chemical Society
The Journal of Physical Chemistry, Vol. 89, No. 26, 1985 5841
Nitric Oxide Chemisorbed on Rh/A1203
,
IR Call for Spectroscopic Studies of Adsorbed Species
+k-
TEMPERATURE REGULATION (BOK -7WK )
MOLECULAR DOSING
I
RmLOlutlon = 4 crn-’
TC FEEDTHRU
\
-
T O PUMP
U
W
0
2
a
m
a
am
a
/\
SAMPLE
!
4
Figure 1. High-vacuum cell for infrared spectroscopy of adsorbed species.
In this laboratory a systematic FT-IR study of bonding associated with molecular adsorption on rhodium metal and supported
rhodium has been initiated. Rhodium has been widely used in
catalytic applications, and adsorption of small molecules on
rhodium has attracted considerable interest. Adsorption of
molecules like C0,l4 N2,5and
on supported rhodium has
been studied extensively in the past partially due to their importance in catalytic treatment of automotive exhaust streams.
Particular attention has been given N O where three forms of
chemisorbed NO on Rh/A1203are classified: Rh-NOd+, Rh/NO,
and Rh-NO*-. However, Lunsford and his co-workers13 have
shown that a gem-dinitrosyl complex, Rh(N0)2, exists in the
adsorbed NO/Rh-Y zeolite system. Hyde et a1.I0 subsequently
assigned 1830- and 1740-cm-’ bands to R h ( N 0 ) 2 in the study
of NO/Rh/A1203, but this assignment requires further proof to
be firmly established. In order to conclusively demonstrate the
existence of R h ( N 0 ) 2 and to characterize NO adsorption at low
surface coverage on supported rhodium, results of FT-IR studies
of N O chemisorption on Rh/A1203 are reported here.
Experimental Section
The preparation of supported Rh samples has been described
by Yang et al.I4 and Yates et al.’ Briefly, a mixture of Rh”’C13.3H20 dissolved in HzO, high surface area A1203(Degussa
A1203-C, 100 m2 g-l), and acetone (spectroscopic grade) was
sprayed with an atomizer onto a CaF2 disk (32 mm in diameter)
maintained at 350 K to flash evaporate solvents. Following reduction in H2(g) and evacuation of the cell, the sample was placed
(1) J. T. Yates, Jr., T. M. Duncan, S. D. Worley, and R. W. Vaughan,
J . Chem. Phys., 70, 1219 (1979).
(2) J. T. Yates, Jr., S. D. Worley, T. M. Duncan, and R. W. Vaughn, J .
Chem. Phys., 70, 1255 (1979).
(3) J. T. Yates, Jr., T. M. Duncan, and R. W. Vaughan, J . Chem. Phys.,
71, 3908 (1979).
(4) R. R. Cavanagh and J. T. Yates, Jr., J . Chem. Phys., 74,4150 (1981).
(5) H. P. Wang and J. T. Yates, Jr., J . Phys. Chem., 88, 852 (1984).
(6) P. Gelin, A. R. Siedle, and J. T. Yates, Jr., J . Phys. Chem., 88, 2978
(1984).
(7) H. Arai and H. Tominaga, J . Catal., 43, 131 (1976).
(8) F. Solymosi and J. Sarkany, Appl. Sur!. Sci., 3, 68 (1979).
(9) B. J. Savatsky and A. T. Bell, ACSSymp. Ser., No.178, 105 (1982).
(10) E. A. Hyde, R. Rudham, and C. H. Rochester, J. Chem. SOC.,
Faraday Trans. I, 80, 531 (1984).
(11) V. Rives-Arnau and G. Munuera, Appl. Surf. Sci., 6, 122 (1980).
(12) J. C. Conesa, M. T. Sainz, J. Soria, G. Munuera, V. Rives-Arnau,
and A. Munoz, J . Mol. Catal., 17, 231 (1982).
(13) T. Iizuka and J. H. Lunsford, J. Mol. Caral., 8, 391 (1980).
(14) A. C. Yang and C. W. Garland, J . Phys. Chem., 61, 1504 (1957).
21 )O
lsbo
16m
I700
1600
WAVENUMBERS (crn-ll
I&
1400
Figure 2. Infrared spectra for I4NO adsorbed on Rh/AI,O, as a function
of increasing I4NO coverage.
in the FT-IR spectrometer. The final “density” of Rh on the
to 8.7 X
A1203support was 8.3 X
g/(cm2 of disk) at
the 2.2 wt 7% loading employed here.
The IR cell’5 with a CaF2 window used in this study is shown
in Figure 1. A dosing tube is pointed toward the center of the
sample disk from an oblique angle, thus allowing the dosing gas
to interact with the rhodium surface directly and evenly.
Nitric oxide with a stated purity of 99.0% from Matheson was
further purified by trapto-trap distillation. Matheson O2(99.98%)
and C O (99.99%) were used directly from cylinders as was H2
(99.999%) obtained from Airco Co. The RhCl3.3H20was purchased from Pressure Chemical Co., Pittsburgh, PA.
Infrared spectra were recorded with a Nicolet 7199 series
Fourier transform IR spectrometer in the 4000-400-cm-’ range
at 4-cm-I resolution. A KBr beam splitter and MCT detector were
used, and each sample was scanned 400 times. Measurement
capability below the
absorbance level is achieved with good
signal-to-noise ratios and resolution by using the multiple scan
data accumulation method. The subtraction technique incorporated in the computerized data-processing package was used
throughout this study. All the spectra reported here are difference
spectra which are obtained by subtracting the IR spectra with
and without a given amount of adsorbate. The catalysis cell was
securely mounted and remained stationary throughout the course
of data acquisition, thus minimizing artifacts in the subtraction
technique. Pressure was monitored with MKS Baratron capacitance manometers covering a range from
to lo3 torr.
Results and Discussion
Spectra of Chemisorbed NO. We have studied N O adsorption
at 301 K on a reduced, A1203 supported Rh surface over the
to 5 X
torr. By incrementally
pressure range of 3 X
increasing the pressure and carefully examining the spectral
changes as the coverage is increased, the development of characteristic infrared bands and consequently the chemistry which
(15) The design of this cell was kindly provided by Prof. J. T. Yates, Jr.,
Department of Chemistry, University of Pittsburgh, Pittsburgh, PA.
5842
4
Liang et al.
The Journal of Physical Chemistry, Vol. 89, No. 26, 1985
I
T=300K
i
a i
I
,
I
,
0
10
I
I
30
20
50
40
60
torr)
P,,(Io-~
Figure 3. Peak absorbance ratio of the Rh(NO), doublet during adsorption of I4NO on Rh/AI,O,.
TABLE I: Infrared Frequencies and Calculated Bond Angles for
Dinitrosyl Species Adsorbed on Various Supports
dinitrosyl species
Rh(NO),/Y
zeolite
Rh(NO),/Y
zeolite
Mo(N0)2/A1203
Rh(N0)2/AI,O,
Cr"'(NO),/SiO,
Fe(NO),/Y zeolite
cm-'
LN-M-N,
deg
94"
1860
cm-'
1780
1848
1771
1817
1825
104
120
1875
1713
1743
1745
1917
1815
145"
wSym,
waaym,
96*
135.2
ref
13
16
18
present work
19
22
"Calculated from the published spectra
00
produces the spectra could be monitored. It is well-known that
nitric oxide dissociates on and oxidizes rhodium surfaces at room
temperature and low coverages. Thus, the rhodium surfaces
studied here are expected to be similar to preoxidized ones in their
adsorptive properties.
Figure 2 shows that as very small quantities of N O gas were
added prior to successive measurements, four bands began to grow
at 1912, 1825, 1743, and 1648 cm-I. The two broad bands
centered at 1912 and 1648 cm-' develop first and can be assigned
to Rh-N06+ and Rh-N=O
(bent NO) species, respecitvely,
in agreement with the results of Arai et a].?
At higher coverage, another pronounced spectral feature is a
doublet with components at 1825 and 1743 cm-I. This doublet
increases in intensity during the entire course of adsorption without
shifting frequency and has been attributed to Rh(NO), by Hyde
et a1.I0 Earlier investigators,l6 however, did not detect these
features, and thus confirmation of the assignment is important.
To be consistent with a dinitrosyl species, an invariant ratio of
the doublet peaks should exist over the pressure range below site
saturation. Figure 3 shows that this is true in the data represented
in Figure 2. This invariant ratio confirms that the doublet bands
at 1743 and 1825 cm-I represent the dinitrosyl species, and these
bands can be assigned as the asymmetric and symmetric stretching
modes, respectively.17 A dinitrosyl species has also been reported
in the study of N O adsorbed on molybdena-alumina,'s on chromia-~ilica,"-~' on Fe-Y zeolites,22 and on Rh-Y ze01ites.I~
The constant ratio of ~3 between the two peaks also indicates
that the angle between the N O groups is approximately 120'. In
(16) H . Arai. Ind. Eng. Chem. Prod. Res. Deu., 19, 507 (1980).
(17) J. R. Pearce, D. E. Sherwood, M. B. Hall, and J. H. Lunsford, J .
Phys. Chem., 84, 3215 (1980).
(18) J. Valyon and W. K. Hall, J . Catal., 84, 216 (1983).
(19) E. L. Kuger, R. J. Kodes, and J. W. Gryder, J . Catal., 36, 142 (1975).
(20) A. Zecchina, E. Garrone, C. Morterra, and S . Coluccia, J . Phys.
Chem., 79,978 (1975).
(21) G. Ghiotti, E. Garrone, G . D. Gatta, B. Fubini, and E. Giamello, J .
Card., 80, 249 (1983).
(22) K. Segawa, Y. Chen, J. E. Kubsh, W. N. Delgass, J. A. Dumesic, and
W. K. Hall. J . Card., 76, 112 (1982).
Figure 4. FT-IR spectra of isotopically substituted adsorbed NO species
on Rh/A1203at room temperature: (a) I4NO;(b) 14N0+ ISNO(1:l
ratio); (c) ISNO.
Table I we compare our result with related dinitrosyl species
adsorbed on different supported metal surfaces to illustrate consistency of this finding with literature reports.
In order to further confirm the existence of Rh(N0)2, studies
of isotopically substituted'N0 adsorbed on Rh were carried out.
Identical Rh/AI2O3 surfaces were exposed to saturation coverage
of 1 5 N 0 and a 1:l ratio of I4NO and I5NO, respectively. The
spectra for these surface species are shown in Figure 4 along with
the spectrum of adsorbed I4NO from Figure 2. It can readily be
observed that the bands at 1825 and 1743 cm-' from I4NO are
shifted to 1784 and 1698 cm-I in the presence of pure I5NO. The
1:1 mixed system provides conclusive evidence for the dinitrosyl
moiety. If a mononitrosyl species were responsible for the 1743
cm-' peak, then the IR spectrum for the 1:l mixture of I4NO and
15N0 would simply be a linear and equal combination of these
two bands at 1743 and 1698 cm-]. On the other hand, if a
dinitrosyl species were responsible for the band at 1743 cm-' in
the system ''NO/Rh/A1203, then three bands should be expected
in this region for the 1:l mixture of I4NO and ISNO, Le., a band
at 1743 cm-' due to Rh(I4N0)*, a band at 1698 cm-I due to
Rh(15NO)z,and an intense band between 1743 and 1698 cm-I
due to Rh(l4NO)(ISNO). Moreover, statistically the relative
intensities for the three peaks should be in the ratio 1:2:1. The
middle spectrum of Figure 4 shows this result with an intense peak
at 1722 cm-I located between 1743 and 1698 cm-I and with two
incompletely resolved smaller bands at 1743 and 1698 ~ m - ' . ~ ~
This confirms the assignment of the band at 1743 cm-I in the I4NO
adsorption to a dinitrosyl species. Indeed, because the intensities
of the bands at 1825 and 1743 cm-' are linearly correlated over
(23) Since the observed half-width of the band centered at 1722 cm-' is
70 cm-l, clear resolution of the triplet in Figure 4b is not likely. Deconvolution
can be used to verify the observation, but a more definitive method which we
are currently pursuing is to study the process with 15N180-labeledspecies at
reduced temperature.
The Journal of Physical Chemistry, Vol. 89, No. 26, 1985 5843
Nitric Oxide Chemisorbed on Rh/A1203
TABLE 11: infrared Frequencies and Force Constants for Isotopically
Substituted Dinitrosyl Species Adsorbed on Rh/Al,OB Surfaces
species
vSymrcm-I vaSym, cm-'
Rh('4N0)2
Rh(lSN0)2
Rh('4N0)(15NO) (obsd)
Rh(14N0)('5NO)
1825
1784
1807
1807
1743
1698
1722
1719
k , , dyn
cm-l
10" k2
1.402
1.385
0.064
0.068
1.395
0.066
(calcd)
I
I
I
I
I
1743
a
W
0
z
a
m
Lz
0
v,
m
a
21 10
I900
1800
1700
1600
WAVENUMBERS
1500
1400
Figure 5. Development of infrared spectra for I4NO adsorbed on Rh/
Al2O3at low temperature as a function of pressure: (a) PN0 = 2.8 X lo-'
torr; (c) PNO = 8.4 X
torr; (d) PNO =
torr; (b) PNo L: 5.5 X
85.9 X
torr.
the full coverage range, the band at 1825 cm-' can also be assigned
to the same dinitrosyl species.
Additional proof of the existence of the dinitrosyl species comes
from an analysis of the normal modes of the isotopically substituted
Rh(N0)2, analogous to that carried out for the R h ( C 0 ) 2 speci e ~ . The
~ ~ frequencies
, ~ ~
can be obtained by solving the secular
equation
where pi = reciprocal mass of the ith atom (1 = l6O; 2 = I4N;
3 = I6O; 4 = I5Nfor Rh('4NO)('5NO) species), v = observed
frequencies in units of s-l, k l = force constant for N-0 stretch,
and k2 = force constant for N O / N O coupling.
For the Rh('4N0)2 and R h ( 1 5 N 0 ) 2species by use of the observed asymmetric and symmetric frequencies, values of the k ,
and k2 were calculated and are tabulated in Table 11. As can
be seen, the agreement between k l and k2 for these two independent measurements is good. Based on these values for kl and
k2, vam and vSymfor Rh('4NO)('5NO) were calculated to compare
with the observed frequencies. Results indicate excellent agreement between predicted and measured spectral frequencies based
on assignment as a dinitrosyl surface species.
Adsorption of N O on Rh/A1203at low temperature (161 K)
was also studied. The spectra observed are shown in Figure 5
where all three surface species, Rh-NO*+, Rh-N=O
(bent
NO), and Rh(N0)2, are clearly represented, and the ratio of
intensities for the dinitrosyl peaks at 1743 and 1825 cm-' is
between 3 and 4. It should be noted there appears to be no
physisorption of NO on Rh/A1203at this temperature.
There are two differences as one compares the spectra in Figure
5 of N O adsorption at low temperature with those spectra in
(24) H. Knozinger, E. W. Thornton, and M. Wolf J . Chem. SOC.,Faraday
Trans. 1 , 75, 1888 (1979).
(25) J. T. Yates, Jr., and K. Kolasinski, J . Chem. Phys., 79, 1026 (1983).
-
Figure 2 measured at room temperature. First, the Rh-NO*+
band at 19 12 cm-l in Figure 5 appears to be suppressed, and
second, the development of the dinitrosyl doublet at 1825 and 1743
cm-' in Figure 5 is expedited at low coverage relative to the same
features at equivalent N O coverage in Figure 2a-c.
Several factors may account for the first difference. Kinetic
control via differences in activation energies for the adsorption
process for the different species is possible. From the relative
intensities of the peaks in Figure 5, one could conclude that the
activation energies for the linear and bent species are higher than
for the dinitrosyl species, and the overall order would be &,(linear)
> ,!?,,,(bent) > E,,,(dinitrosyl). Since the band at 1912 cm-I is
suppressed at 161 K, the activation energy for this adsorption
process, ,!?,,,(linear), may be greater than k T at this temperature.
Another plausible explanation for suppression of the 1912-cm-'
band at the lower temperature is that dissociation for NO is much
slower at 161 K and the preoxidized surface cannot readily form.
The third possible explanation for the observed 1912-cm-' peak
intensities at the two temperatures comes from metal-support
interaction properties and their effect on bonding at the surface.
Our results26in fact show that the 1912-cm-' band can be used
as a probe to gauge the interaction between metal and support.
The second difference is illustrated by comparing the spectrum
of Figure 5c with that in Figure 2c where the NO pressure is close
to 8.0 X
torr in both cases. Figure 5c shows a 1743-cm-'
absorbance of 0.09A whereas Figure 2c shows only 0.01A. We
believe this ninefold difference in absorbance cannot be accounted
for simply by sample thickness. Comparable intensities for this
doublet are found in Figures 5c and 2b, leading to the conclusion
that lowering the temperature of adsorption is equivalent to increasing the pressure of the adsorbate gas at room temperature?'
Thermal Desorption of N O and the Apparent Interconversion
ofRh(NO), and Rh(N0). Thermal desorption of N O on Rh/
A1203was also studied in this important system by heating a fully
covered surface and observing the spectra as a function of temperature. Results indicate clearly that NO could be reversibly
adsorbed and desorbed. The adsorbed species are very stable under
vacuum at room temperature as evidenced by no detectable
changes in the infrared spectra over extended periods. When the
surface is warmed from 301 to 469 K, the 1743- and 1825-cm-I
peaks disappear and the 1912-cm-' peak gains intensity as illustrated in Figure 6A. Over this range in temperature it should
be noted that the vibration frequencies of the components of the
R h ( N 0 ) 2 doublet are invariant at all stages of desorption.
When the system is heated over 469 K, the 1912-cm-' peak also
starts to diminish as shown in Figure 6B. Upon reaching 573 K,
a complete desorption spectrum of the NO/Rh/A120, is observed,
as seen in Figure 6C which is a mirror image of the adsorption
spectrum of N O on Rh/AI2O3at room temperature. This textbook
example of image formation clearly indicates that no additional
surface reactions occur during the entire course of thermal desorption.
It should be noted that during the development of the NO
adsorption spectrum, as the pressure was increased to above
torr, the 1 8 2 5 , 1743-, and -1640-cm-I bands gained intensity
while the 1912-cm-' band did not. Upon detailed investigation
of this spectrum, one finds that the intensity of the 1912-cm-' band
in fact actually decreases as shown in Figure 7. A likely explanation is that some of the Rh-N06+ species is interconverted
to form a dinitrosyl complex via
As illustrated in Figure 6, however, upon desorption of the dinitrosyl species, R h ( N 0 ) is formed. While this apparent backreaction of eq 2 could occur directly, it is more likely to originate
from a two-step process involving liberation of N O gas and adsorption of the gas on the oxidized Rh sites present. This result
also indicates that the bond strength of the Rh-NO*+ species is
~
~~~
( 2 6 ) J. Liang, H. P. Wang, and L. D. Spicer, manuscript in preparation.
(27) R. J. Madix, Surf. Sci., 89, 540 (1979).
5844
Liang et al.
The Journal of Physical Chemistry, Vol. 89, No. 26, 1985
r
I
I
I
I
1
I
I\
Tz300K
Resolution = 4 cm-'
4k-
I\
A A = 0,0°7
~
t
0
i
-
i\
I900
I800
IiOO
1600
WAVENUMBERS (cm-9
I500
1400
Figure 7. Difference spectrum obtained from spectrum 2h minus spectrum 2e shown in Figure 2.
I
1
2OW
1900
,v,
le00
1700
C
IS00
,
1500
,
WAVENUMBERS
Figure 6. Spectral changes during desorption of I4NO species. (A)
Desorption involving interconversion of surface species Rh(NO)z and
R h ( N 0 ) at different temperatures: (a) 379, (b) 414, and (c) 469 K. (B)
Desorption of all surface species at different temperatures: (d) 469, (e)
553, and (0 573 K (C) Mirror images of adsorption and desorption.
greater than that of the Rh(NO)2 species. If the bond of RhNO*+ were weaker than that of the R h ( N 0 ) 2 species, the consumption of R h ( N 0 ) 2 would not result in the net increase in
intensity of the 1912-cm-' band as shown in Figure 6A.
Figure 8 shows the desorption spectra of the 1:l mixture of
14N0and 15N0. This study further verifies the existence of the
dinitrosyl species on rhodium and its interconversion to R h ( N 0 )
by clearly showing that the broad triplet at 1722 cm-' is converted
to both Rh-I4NO and Rh-15N0 characteristic adsorptions at 1913
and 1874 cm-', respectively.
Interaction of CO with Adsorbed NO on RhIA120,. Carbon
monoxide adsorption on supported rhodium surfaces has been
clearly documented t o produce Rh'(CO), by Y a t e s and his cow o r k e r ~ .In
~ ~an effort to gain insight into the relative stabilities
and displacement kinetics of adsorbed CO and NO on rhodium,
the ligand-exchange process was studied. Ten torr of CO gas was
introduced into the sample cell at room temperature after the
rhodium surface had been saturated by N O adsorption. Essentially
complete displacement of adsorbed N O species was observed
within 10 min as shown in the spectrum of Figure 9. Since
desorption of NO species is relatively slow at ambient temperatures, active displacement must occur in this process rather than
sequential site occupation.
After complete C O exchange for adsorbed N O species was
accomplished, the sample cell was evacuated thoroughly and
NO(g) was readmitted to study the reverse process. The spectra
under these conditions clearly indicate that NO(g) does not
-008 2000
1900
le00
1700
1600
WAVENUMBERS(cm")
1500
Figure 8. IR difference spectrum showing interconversion of the isotopically substituted surface species Rh(NO)z and R h ( N 0 ) .
I
T-300 K
a
w
z
0
a
a
m
wa
m
2
3
2100
2000
1900
I800
1700
WAVENUMBERS (cm-')
1600
1500
I
Figure 9. Interaction of CO(g) with the adsorbed I4NO on Rh/AI,O,.
J. Phys. Chem. 1985, 89, 5845-5849
completely displace the Rh(C0)2, thus indicating that the dicarbonyl species is more stable than the dinitrosyl species on this
supported rhodium surface. Figure 9 also indicates very little
detectable linear CO or bridged CO on rhodium crystallite sites
(Rh,), and in fact, the major surface species is Rh'(C0)2. This
suggests that Rh' sites are the predominant sites represented on
the surface.
ConcIusion
The chemisorption of N O on A1203 supported Rh has been
studied using FT-IR spectroscopy. Based on the data presented,
the following conclusions are reached: (1) The N O species on
the surfaces are linear NO, bent NO, and gem-dinitrosyl complexes, and each is distinguishable on Rh/AI2O3 surfaces by
FT-IR. (2) An invariant ratio for the doublet IR band at 1825
and 1743 cm-' and studies of isotopic N O confirm the existence
of Rh(NO)z species. (3) There is no evidence for interaction of
5845
the dinitrosyl complexes with neighboring N O molecules as
coverage is increased, and the characteristic doublet represents
stretching modes for this species. (4) Adsorption and desorption
studies clearly show that as the dinitrosyl species is formed, the
concentration of linear species is reduced and as the dinitrosyl
species is desorbed, the linear R h ( N 0 ) is formed. Desorption
studies also show the ligand bond of the Rh-N06+ is stronger
than those in either R h ( N 0 ) 2 or Rh-N=O
(bent NO). (5)
Rapid displacement of CO with NO(ads) occurs for all of the
adsorbed N O species.
Acknowledgment. We acknowlege helpful comments and
suggestions by Dr. Max Matheson and one of the referees.
Support for this research by the United States Department of
Energy under Contract No. DE-AC02-76ER02190 is gratefully
acknowledged.
'Registry No. NO, 10102-43-9; Rh, 7440-16-6.
+ CH3CI(H20),
An MO Study of S,2 Reactions in Hydrated Gas Clusters: (H,O),OHHOCH3 CI( n m)H2O
+ + +
+
Katsuhisa Ohta and Keiji Morokuma*
The Institute for Molecular Science, Myodaiji, Okazaki 444, Japan (Received: June 21, 1985)
-
Potential energy surfaces are calculated with the ab initio MO method for the SN2reaction (H,O),OH- + CH3C1(H,0),
HOCH, + C1- + ( n + m)H20, where the reactants are complexed with up to two water molecules. When the hydroxide
ion is solvated by water molecules, the reaction takes place through the first step of reactant complex formation, followed
by inversion of the methyl group. The migration of water molecules from the hydroxide side to the chloride side is not involved
in the rate-determiningprocess. In the case of (n,m) = ( 2 , O ) the transition state for methyl inversion has an energy comparable
to that of the reactants.
Introduction
(H,O),CI-
Chemical reactions in gas-phase clusters are attracting considerable attention, as they provide information filling a wide gap
between reactions in the gas phase and in solution. sN2 reactions
of type X- CH3Y XCH, Y- have been studied extensively
in both solution and the gas phase.'-, The rate constant in the
gas phase has been found to be up to lozotimes faster than in
solution. To explain this large difference, Brauman and cc-workers
have suggested for the gas-phase reaction a double well potential
curve, which gradually changes to a single barrier upon stepwize
solvation of reactants.' Bohme and co-workers actually have
investigated the sN2 reaction in gas-phase hydrated clusters by
the flowing afterglow
For the reaction
+
OH-(H20),
-
+
+ CH3C1
-
CH30H
+ C1- + n H 2 0
(1)
they have reported the kinetics for the hydration number from
n = 0 to 3.3c The rate becomes slower in clusters for n = 1, 2,
and 3 by ca. 1.6, 500, and 1000 times, respectively, than that for
n = 0. A drastic decrease in rate upon going from n = 1 to 2 has
been interpreted as an indication that the overall barrier has
become positive at n = 2.
In order to provide information concerning the potential energy
surfaces in hydrated clusters, we have previously carried out ab
initio MO calculations for the following symmetric SN2 reaction^:^
(1) Olmstead, W. N.; Brauman, J. I. J . Am. Chem. SOC.1977, 99, 4219.
Pallerite, M. J.; Brauman, J. I. J . Am. Chem. SOC.1980, 102, 5993.
(2) Tanaka, K.; Mackay, G. I.; Payzant, J. D.; Bohme, D. K. Can. J .
Chem. 1976, 54, 1643.
(3) (a) Mackay, G. I.; Bohme, D. K. J. Am. Chem. SOC.1978, 100, 327.
(b) Bohme, D. K.; Mackay, G. I. J . Am. Chem. SOC.1981, 103, 978. (c)
Bohme, D. K.; Raksit, A. B. J . Am. Chem. SOC.1984, 106, 3447.
+ CH3Cl-
ClCH3
+ Cl-(H,O),
n = 0, 1, 2
(2)
- -
Our findings, shown in Figure 1, can be summarized as follows.
(i) The most favorable reaction path for n = 1 is reactants
reactant complex transition state for CH3 inversion transfer
of H 2 0 from the left (the newly formed CH3C1 side) to the right
product complex
products.
(the newly formed C1- side)
Since the system is symmetric, the process by which H 2 0 transfer
takes place before CH, inversion is equally favorable. The path
of simultaneous C H 3 inversion and H 2 0 transfer is both energetically and entropically unfavorable.
(ii) For n = 2 the most favorable path is reactants reactant
complex transfer of one H 2 0molecule from the left (the C1side) to the right (the CH3Cl side) CH, inversion transfer
of the other HzO molecule from left (the newly formed CH3C1
side) to the right (the newly formed C1- side). Having one water
molecule on each chlorine atom is the best way to stabilize the
intrinsically symmetric transition state for CH, inversion of reaction 2 .
(iii) The transfer of H 2 0from one side to the other takes place
with little or no barrier, via an intermediate having a bent C1C - C l configuration.
Henchman et al. have analyzed the product ions for the reaction
-
-
-
OH-(H,O),
-
+ CH3Br
-
CH30H
-
+ Br- + n H 2 0
(3)
and found the product bromide ion is actually unsolvated, probably
(4) Morokuma, K. J . Am. Chem. SOC.1982,104, 3732. Morokuma, K.;
Kato, S.;Kitaura, K.; Obara, S.;Ohta, K.; Hanamura, M. In 'New Horizons
of Quantum Chemistry", Lowdin, P. O., Pullman, B., Eds.; Reidel: Dordrecht,
1982; p 221.
0022-3654/85 /2089-5845%01.50/0 0 1985 American Chemical Societv