A comprehensive study of the reactions of methyl fragments from methyl iodide dissociation on reduced and oxidized silica-supported copper nanoparticles

Article abstract of DOI:10.1021/ja9633833

Transmission infrared spectroscopy and temperature programmed desorption have been used to investigate the chemistry of CH₃I adsorbed on silica-supported copper nanoparticles. The following three factors affect the chemistry of CH₃I on Cu/SiO₂: (i) the oxidation state of the copper, (ii) the hydroxyl group coverage on the silica support, and (iii) the surface roughness of the copper particles. These three factors can be controlled by sample preparation. On reduced-Cu/SiO₂ samples, C-I bond dissociation in adsorbed methyl iodide results in the formation of adsorbed methyl groups on the copper surface. The frequency of the symmetric stretch of adsorbed methyl groups is at 2913 cm^-1 for copper nanoparticles with atomically smooth surface morphology and at 2924 cm^-1 for copper nanoparticles with atomically rough surface morphology. In addition to methyl groups adsorbed on the copper nanoparticles, for reduced Cu/SiO₂ samples with Si-OH groups present in close proximity to the copper particles, methyl groups can spill over on to the silica support and react with OH groups to form SiOCH₃. The silica hydroxyl coverage also plays a role in methyl reactions on the copper particles as hydroxyl groups provide a source of hydrogen atoms. Methane and ethane are the predominant reaction products for reduced Cu/SiO₂ samples with high hydroxyl group coverage whereas methane, ethane, and ethylene form on samples with low silica hydroxyl group coverage. The copper particle morphology may also play a role in the chemistry of adsorbed methyl groups as there is some evidence for slower methyl reaction kinetics on the copper nanoparticles with rough surface morphology. C-I bond dissociation in adsorbed methyl iodide occurs on oxidized-copper nanoparticles as well. The infrared spectrum taken after adsorption of CH₃I on oxidized-Cu/SiO₂ is consistent with the presence of adsorbed methoxy groups and bidentate formate on the oxidized-copper particles. The chemistry of methyl groups on oxidized-copper particles is similar to that of methanol on oxidized-copper particles. Finally, the use and complexities of characterizing these samples and copper catalysts in general with CO adsorption in conjunction with infrared spectroscopy are discussed.

Full text of DOI:10.1021/ja9633833

J. Am. Chem. Soc. 1997, 119, 1697-1707
1697
A Comprehensive Study of the Reactions of Methyl Fragments
from Methyl Iodide Dissociation on Reduced and Oxidized
Silica-Supported Copper Nanoparticles
M. D. Driessen and V. H. Grassian*
Contribution from the Department of Chemistry, UniVersity of Iowa, Iowa City, Iowa 52242
X
ReceiVed September 26, 1996
Abstract: Transmission infrared spectroscopy and temperature programmed desorption have been used to investigate
the chemistry of CH3I adsorbed on silica-supported copper nanoparticles. The following three factors affect the
chemistry of CH3I on Cu/SiO2: (i) the oxidation state of the copper, (ii) the hydroxyl group coverage on the silica
support, and (iii) the surface roughness of the copper particles. These three factors can be controlled by sample
preparation. On reduced-Cu/SiO2 samples, C-I bond dissociation in adsorbed methyl iodide results in the formation
of adsorbed methyl groups on the copper surface. The frequency of the symmetric stretch of adsorbed methyl groups
is at 2913 cm for copper nanoparticles with atomically smooth surface morphology and at 2924 cm-1 for copper
nanoparticles with atomically rough surface morphology. In addition to methyl groups adsorbed on the copper
nanoparticles, for reduced Cu/SiO2 samples with Si-OH groups present in close proximity to the copper particles,
methyl groups can spill over on to the silica support and react with OH groups to form SiOCH3. The silica hydroxyl
coverage also plays a role in methyl reactions on the copper particles as hydroxyl groups provide a source of hydrogen
atoms. Methane and ethane are the predominant reaction products for reduced Cu/SiO2 samples with high hydroxyl
group coverage whereas methane, ethane, and ethylene form on samples with low silica hydroxyl group coverage.
The copper particle morphology may also play a role in the chemistry of adsorbed methyl groups as there is some
evidence for slower methyl reaction kinetics on the copper nanoparticles with rough surface morphology. C-I
bond dissociation in adsorbed methyl iodide occurs on oxidized-copper nanoparticles as well. The infrared spectrum
taken after adsorption of CH3I on oxidized-Cu/SiO2 is consistent with the presence of adsorbed methoxy groups and
bidentate formate on the oxidized-copper particles. The chemistry of methyl groups on oxidized-copper particles is
similar to that of methanol on oxidized-copper particles. Finally, the use and complexities of characterizing these
samples and copper catalysts in general with CO adsorption in conjunction with infrared spectroscopy are discussed.
-1
Introduction
details of the adsorption and decomposition of methanol on
model copper catalysts in an attempt to characterize the surface
reaction mechanisms which form methanol from CO and
The chemistry of composite systems can be quite complex
and often it is not easy to delineate the factors that affect the
chemistry of these systems. Oxide-supported metal catalysts
are such systems that are not easily understood because both
the oxide support and the metal particles can be involved in
1
3-15
H2.
Other studies have examined the adsorption of alkyl
fragments on single crystal copper surfaces in order to gain
insight into Fischer-Tropsch processes and other catalyzed
1
6-18
coupling reactions.
1-4
the chemistry.
In this study, the surface chemistry of copper
As is evident in the literature, the results obtained on high
nanoparticles supported on silica has been examined. Because
copper catalysts are the basis for several modern industrial
processes, there is a great deal of interest in them. Methanol
synthesis and hydrocarbon coupling reactions are two important
examples of copper catalyzed processes which are still not fully
surface area copper catalysts vary greatly depending upon
sample preparation methods and processing conditions.1
9-21
In
this study, we will show that the chemistry of Cu/SiO2 exhibits
both similarities and differences compared to single crystal
copper surfaces and that sample preparation plays an important
role in the surface chemistry of Cu/SiO2. The following three
factors affect the chemistry of the simplest C1 hydrocarbon
fragment, methyl, on Cu/SiO2 samples: (i) the oxidation state
5
-12
understood.
Several studies have been reported on the
*
Author to whom correspondence should be addressed.
Abstract published in AdVance ACS Abstracts, January 1, 1997.
X
(1) Stevenson, S. A.; Dumesic, J. A.; Baker, R. T. K.; Ruckenstein, E.,
Eds. Metal-Support Interactions in Catalysis, Sintering and Redispersion;
Van Nostrand Reinhold: New York, 1987.
(13) Clarke, D. B.; Lee, D.-K.; Sandoval, M. J.; Bell, A. T. J. Catal.
1994, 150, 81.
(
(
(
(
(
2) Haller, G. L.; Resasco, D. E. AdV. Catal. 1989, 36, 173.
3) Driessen, M. D.; Grassian, V. H. J. Catal. 1996, 161, 810.
4) Driessen, M. D.; Grassian, V. H. Langmuir 1995, 11, 4213.
5) Klier, K. AdV. Catal. 1982, 31, 243.
(14) Millar, G. J.; Rochester, C. H.; Waugh, K. C. J. Chem. Soc., Faraday
Trans. 1991, 87, 2795.
(15) Sakata, Y.; Domen, K.; Maruya, K.; Onishi, T. Appl. Surf. Sci. 1988/
89, 35, 363.
(16) Xi, M.; Bent, B. E. J. Am. Chem. Soc. 1993, 115, 7426.
(17) Chiang, C.-M.; Wentzlaff, T. H.; Jenks, C. J.; Bent, B. E. J. Vac.
Sci. Technol. A 1992, 10, 2185.
(18) Lin, J.-L.; Chiang, C.-M.; Jenks, C. J.; Yang, M. X.; Wentzlaff, T.
H.; Bent, B. E. J. Catal. 1994, 147 , 250.
6) Bowker, M.; Hadden, R. A.; Houghton, H.; Hyland, J. N. K.; Waugh,
K. C. J. Catal. 1988, 109, 263.
7) Chinchen, G. C.; Denny, P. J.; Jennings, J. R.; Spencer, M. S.; Waugh,
K. C. Appl. Catal. 1988, 36, 1.
8) Burch, R.; Golunski, S. E.; Spencer, M. S. J. Chem. Soc., Faraday
Trans. 1990, 86, 2683.
(
(
(19) Shimokawabe, M.; Takezawa, N.; Kobayashi, H. Appl. Catal. 1982,
2, 379.
(9) Dry, M. E. Catal.: Sci. Technol. 1981, 1, 159.
(
10) Anderson, R. B. The Fischer-Tropsch Synthesis; Academic Press:
(20) Nomura, M.; Kazusaka, A.; Kakuta, N.; Ukisu, Y.; Miyahara, K. J.
Chem. Soc., Faraday Trans. 1 1987, 83, 1227.
(21) VanDerGrift, C. J. G.; Elberse, P. A.; Mulder, A.; Geus, J. W. Appl.
Catal. 1990, 59, 275.
New York, 1984.
(
(
11) Rofer-DePoorter, C. K. Chem. ReV. 1981, 81, 447.
12) Biloen, P.; Sachtler, W. M. H. AdV. Catal. 1981, 30, 165.
S0002-7863(96)03383-5 CCC: $14.00 © 1997 American Chemical Society

1698 J. Am. Chem. Soc., Vol. 119, No. 7, 1997
Driessen and Grassian
of the copper, (ii) the hydroxyl group coverage on the silica
support, and (iii) the surface roughness of the copper particles.
Methyl iodide is used as the precursor for the formation of
adsorbed methyl groups in this study. In general, it has been
previously shown that adsorbed alkyl groups can be produced
Experimental Section
The IR experiments described here were performed in an infrared
cell which has been described previously.3
5,36
The cell consists of a
2
.75-in. stainless steel cube with two differentially pumped barium
fluoride windows and a sample holder through which thermocouple
and power feedthroughs are connected to a tungsten sample grid. The
temperature is monitored by spot welding thermocouple wires to the
top of the sample grid. The sample holder design is such that the
sample may be cooled to near liquid nitrogen temperatures and heated
resistively up to 1300 K. The cell is attached to an all stainless steel
vacuum chamber through a 2-ft bellows hose. The vacuum system is
rough pumped using a turbomolecular pump and then pumped with an
on single crystal metal surfaces under ultrahigh vacuum
conditions2
2-24
through the thermal or photochemical dissocia-
tion of an alkyl halide precursor (eq 1).
RX(a) f R(a) + X(a) (X ) Cl, Br, or I)
(1)
In the past ten years, studies of the surface chemistry of these
fragments have led to a better understanding of how alkyl
fragments bond and react on clean single crystal metal sur-
8
0 L/s ion pump.
Samples are made by spraying a slurry of copper(II) nitrate trihydrate
2
2
5
(Strem Chemicals, 99.999%) and silica (Cabosil, M-5, 200 m /g)
suspended in acetone and water onto a tungsten grid (Buckbee-Mears).
A template is used to mask one half of the tungsten grid, allowing one
faces.
As there is great interest in the selective partial
oxidation of alkyl groups through reaction with adsorbed oxygen
on metal surfaces,2
6-29
there have also been several recent
side to be coated with Cu/SiO
2
, and the other side is either left blank
studies on the reactions of alkyl groups and adsorbed oxygen.
Bol and Friend have shown that both ethyl and 2-propyl iodide
can react with an oxygen covered rhodium surface to selectively
afford oxygenated hydrocarbon fragments.26 Among the several
products reported, acetaldehyde from ethyl iodide and acetone
from 2-propyl iodide were formed when reacted with surface
(for detecting gas-phase species) or coated with pure silica (to monitor
the reactions of the silica support without the presence of the metal
particles). Metal loadings on the order of 15% by weight are typically
used in these experiments. The copper surface area is approximately
13% of the total catalyst surface area.
After the sample is prepared, it is mounted inside of the infrared
cell, wrapped in heating tape, and processed using one of three methods.
The three different processing procedures will be denoted as reduced-
27
oxygen. Bol and Friend and separately Solymosi and co-
workers2 have investigated the partial oxidation of CH2 groups,
formed from the dissociative adsorption of diiodomethane to
produce formaldehyde on an oxygen covered Rh(111) surface.
Zhou et al. have also reported the oxidation of adsorbed
8
Cu/SiO
throughout the manuscript. The reduced-Cu/SiO
Cu/SiO (673 K) are samples that have been reduced under different
2
(473 K), reduced-Cu/SiO
2
(673 K), and oxidized-Cu/SiO
2
2
(473 K) and reduced-
2
conditions. The processing for all three samples begins with a 12 h,
473 K bakeout. The copper is then reduced with hydrogen. Hydrogen
methylene to produce formaldehyde on a Pt(111) crystal using
chloroiodomethane as a methylene precursor.29 Despite the
(
4
Air Products, Research Grade) is introduced into the sample cell in
00-Torr quantities for 15 min followed by a 15-min evacuation.
Hydrogen is introduced for increasingly longer periods of time (30,
large number of studies concerning alkyl fragment adsorption
on single crystal metals, and the vast literature on hydrocarbon
adsorption on supported metal catalysts, there have been very
few which have examined the adsorption of alkyl fragments on
supported metal catalysts.3
6
1
0, and then 120 min), and each reduction period is followed by a
5-min evacuation. Following this initial reduction, the samples are
,4,30-34
then oxidized in 100 Torr of oxygen for 10 min followed by evacuation
and a 30-min reduction in hydrogen. This oxidation/reduction cycle
is repeated if necessary to remove residual organics. Nothing further
In this study, the adsorption and reaction chemistry of methyl
groups, from methyl iodide dissociation, on Cu/SiO2 samples
have been investigated with transmission infrared spectroscopy
in conjunction with temperature programmed desorption (TPD).
As discussed above, reaction of methyl fragments from methyl
iodide dissociation on Cu/SiO2 is dependent upon the sample
preparation conditions employed. Although Cu/SiO2 catalysts
have been known to be difficult to characterize, we will use
both the frequency of the IR absorption band of adsorbed CO
on the copper particle surface and the products formed from
CO adsorption to deduce the surface morphology and oxidation
state of the copper particles. TEM measurements are also used
to determine particle size. The information obtained from the
sample characterization will be used to aid in the interpretation
of the subsequent reactivity of Cu/SiO2 samples prepared under
different conditions.
2
is done for the reduced-Cu/SiO (473 K) sample but preparation of the
reduced-Cu/SiO
reduced-Cu/SiO
2
(673 K) sample continues from this point. The
(673 K) sample is then resistively heated to 673 K in
2
the presence of 400 Torr of hydrogen for periods of 12 h or longer
-1
until the IR frequency of adsorbed CO is near 2100 cm . Preparation
of oxidized-Cu/SiO
the reduced-Cu/SiO
2
samples follow the exact same procedure as for
(473 K) samples; however, after the last reduction
2
the sample is then oxidized in 5 Torr of oxygen at 473 K for 120 min.
After processing, the IR cell is placed on a linear translator inside
the spectrometer sample compartment. Either the Cu/SiO
or blank) can be translated into the infrared beam for data acquisition.
This design enables us to examine the chemistry of both Cu/SiO and
SiO (or gas-phase species) during the course of a particular experiment.
2 2
or the SiO
(
2
2
A Mattson RS-1 FT-IR spectrometer equipped with a narrow-band MCT
detector was used for the infrared measurements. Spectra were recorded
-
1
by averaging 1000 scans at an instrument resolution of 4 cm
Absorbance spectra shown represent single beam spectra referenced
to the appropriate single beam spectrum of the Cu/SiO or SiO sample
prior to reaction. The transmission range of SiO goes down to
.
(
27.
22) Zhou, X.-L.; Zhu, X.-Y.; White, J. M. Acc. Chem. Res. 1990, 23,
3
2
2
(
(
(
(
(
(
(
23) Zaera, F. Acc. Chem. Res. 1992, 25, 260.
24) Lin, J.-L.; Bent, B. E. J. Vac. Sci. Technol. A 1992, 10, 2202.
25) Bent, B. E. Chem. ReV. 1996, 96, 1361.
26) Bol, C. W. J.; Friend, C. M. J. Phys. Chem. 1995, 99, 11930.
27) Bol, C. W. J.; Friend, C. M. J. Am. Chem. Soc. 1995, 117, 8053.
28) Solymosi, F.; Klivenyi, G. J. Phys. Chem. 1995, 99, 8950.
29) Zhou, X.-L.; Liu, Z.-M.; Kiss, J.; Sloan, D. W.; White, J. M. J.
2
-
1
approximately 1300 cm
.
TPD experiments were performed in a high vacuum chamber which
consists of a 400 L/s ion pump, sample cell, and quadrupole mass
spectrometer (DetecTorr II, UTI Instruments). The temperature was
Am. Chem. Soc. 1995, 117, 3565.
ramped by interfacing a programmable power supply to a PC. A
(
(
(
30) Rasko, J.; Bontovics, J.; Solymosi, F. J. Catal. 1993, 143, 138.
31) Rasko, J.; Solymosi, F. J. Catal. 1995, 155, 74.
32) McGee, K. C.; Driessen, M. D.; Grassian, V. H. J. Catal. 1996,
heating rate of 1 K/s was used. A maximum of twelve different masses
were monitored simultaneously. The TPD data have been corrected
for fragmentation of other species using the cracking fragmentation
1
59, 69. McGee, K. C.; Driessen, M. D.; Grassian, V. H. J. Catal. 1996,
62, 151.
1
(
33) McGee, K. C.; Driessen, M. D.; Grassian, V. H. J. Catal. 1995,
57, 730.
34) Driessen, M. D.; Grassian, V. H. J. Phys. Chem. 1995, 99, 16519.
(35) Basu, P.; Ballinger, T. H.; Yates, J. T., Jr. ReV. Sci. Instrum. 1988,
59, 1321.
(36) Fan, J.; Yates, J. T., Jr. J. Phys. Chem. 1994, 98, 10621.
1
(

Reactions of CH3 on Si-Supported Cu Nanoparticles
J. Am. Chem. Soc., Vol. 119, No. 7, 1997 1699
Figure 1. A compilation of the wide range of values reported in the
literature for the frequency and assignment of CO adsorbed on various
copper surfaces that differ in surface morphology and oxidation state.
There is significant overlap in the frequency range for CO adsorbed
on reduced and oxidized copper making it difficult to use the frequency
of the CO absorption band as the only measure of the oxidation state
of the copper surface atoms. See text for further discussion.
Figure 2. The infrared spectra of CO adsorbed on (a) reduced-Cu/
SiO (673 K), (b) reduced-Cu/SiO (473 K), and (c) oxidized-Cu/SiO
2 2 2
in the presence of an equilibrium pressure of CO (7.8 Torr). Gas-phase
contributions have been subtracted from the spectra shown above.
-
1
well agreed upon that IR bands between ∼2065 and 2110 cm
are indicative of CO adsorbed on reduced copper
However, the region between 2110 and 2200
cm is not as clearly defined. In some cases, bands in the
patterns given in ref 37. After introduction of high pressures (ap-
proximately 15 Torr) of CH I in the sample cell, the cell was pumped
3
with a turbomolecular pump for 2 h before opening the valve to the
mass spectrometer chamber and acquiring TPD data.
38-42,46,47
particles.
-
1
-
1
2115-2130 cm region have been assigned to CO adsorbed
3
8,40,42
-1
CH
Laboratories) both had a purity of 99.5%. CH
transferred into glass sample bulbs and subjected to several freeze-
pump-thaw cycles prior to use. CH OH (EM Science, 99.8%) was
3
I (Aldrich Chemical Co.) and CD
3
I (Cambridge Isotopes
on Cu2O/SiO2
and bands in the 2125-2200 cm region
3
9,42-46
3
I and CD I were
3
to CO adsorbed on CuO/SiO .
2
In other cases, bands near
-1
2
130 cm have been assigned to CO adsorbed on atomically
3
38,39
rough, reduced-copper particles.
It has also been suggested
also transferred into a glass sample bulb and subjected to several
freeze-pump-thaw cycles prior to use. CO was obtained from
Matheson Gas Products and used as received with a purity of 99.99%.
that it is not the location of the CO absorption band which is
the critical factor, but the stability of the species under vacuum
at room temperature and the formation of CO2 or surface
carbonate from reaction with surface oxygen which indicates
Results
3
9,42
whether the sample is oxidized or reduced.
Catalyst Characterization. The characterization of oxide-
supported metal catalysts remains a challenge, especially for
oxide-supported copper catalysts. The methods traditionally
used by surface scientists are not usually adaptable to these
catalysts or the environment in which they are studied. Adsorp-
tion of CO is one of the most widely used methods of
characterizing the oxidation state of Cu/SiO2 catalysts because
of its ease and adaptability for making in-situ measurements.
However, there is a considerable amount of controversy in the
literature concerning the assignment of the CO stretch in
adsorbed CO on the surface of copper in its three oxidation
Here we characterize the three different Cu/SiO2 catalysts
prepared in this study using CO adsorption in conjunction with
infrared spectroscopy. Three criteria are considered in char-
acterizing these samples: i) the frequency of the CO absorption
band, (ii) the stability of the species under vacuum (on reduced
samples CO readily desorbs upon evacuation whereas on
39
oxidized samples it does not), and (iii) whether or not there
is formation of CO2 and/or surface carbonate from CO adsorp-
tion. The adsorption of CO was done by introducing CO into
the infrared cell at 298 K and a spectrum was taken for both
the SiO2 and Cu/SiO2 halves of the sample (Pequilibrium ) 7.8
Torr). Since CO does not adsorb to any appreciable extent on
SiO2 under these conditions, the spectrum taken of the SiO2
sample represents gas-phase CO. The gas-phase spectrum is
then subtracted from the Cu/SiO2 spectrum to yield a difference
spectrum which shows the infrared spectrum of CO adsorbed
on the surface of the copper particles.
3
8-47
states: Cu(0)-Cu metal, Cu(I)-Cu2O, and Cu(II)-CuO.
In addition to the great variance in values for the CO frequency
of CO adsorbed on oxidized and reduced copper, it has also
been observed that the frequency of adsorbed CO is highly
38,39
sensitive to surface structure.
Figure 1 displays a compila-
tion of band frequencies and assignments from the literature
for CO adsorbed on different copper surfaces.3
8-47
It is fairly
After room temperature adsorption of CO on reduced-Cu/
SiO2(673 K), the sample that is reduced for many hours at 673
K, the IR spectrum exhibits an asymmetric absorption band with
(37) CRC Atlas of Spectral Data and Physical Constants for Organic
Compounds; Grasseli, J. G., Ritchey, W. M., Eds.; CRC Press: Boca Raton,
FL, 1975; Vol. III.
-
1
-1
a peak near 2100 cm and a shoulder near 2074 cm (see
Figure 2). Upon evacuation this band decreases in intensity
(38) Van der Grift, C. J. G.; Wielers, A. F. H.; Joghi, B. P. J.; Van
Beijnum, J.; DeBoer, M.; Versluijs-Helder, M.; Geus, J. W. J. Catal. 1991,
1
31, 178.
(
73.
39) DeJong, K. P.; Geus, J. W.; Joziasse, J. Appl. Surf. Sci. 1980, 6,
(43) London, J. W.; Bell, A. T. J. Catal. 1973, 31, 32.
(44) Gardner, R. A.; Petrucci, R. H. J. Am. Chem. Soc. 1960, 82, 5051.
(45) De Jong, K. P.; Geus, J. W.; Joziasse, J. J. Catal. 1980, 65, 437.
(46) Sheppard, N.; Nguyen, T. T. In AdVances in Infrared and Raman
Spectroscopy; Clark, R. J. H., Hester, R. E., Eds.; Heyden: London, 1978;
Vol. 5.
2
(
40) Millar, G. J.; Rochester, C. H.; Waugh, K. C. J. Chem. Soc., Faraday
Trans. 1991, 87, 1467.
(
41) Higgs, V.; Pritchard, J. Appl. Catal. 1986, 25, 149.
(42) Kohler, M. A.; Cant, N. W.; Wainwright, M. S.; Trimm, D. L. J.
Catal. 1989, 117, 188.
(47) Xu, X.; Goodman, D. W. J. Phys. Chem. 1993, 97, 683.

1700 J. Am. Chem. Soc., Vol. 119, No. 7, 1997
Driessen and Grassian
and after 20 min has lost more than 99% of its intensity. There
is an additional band (very weak) observed after evacuation near
-
1
2
000 cm , which we have been unable to assign. The bands
-
1
near 2100 and 2074 cm are clearly assigned in the literature
to fully reduced copper surfaces.3
8-42,46,47
We therefore con-
clude that this sample is indeed reduced copper and the surface
is of an atomically smooth morphology consisting of Cu(111)
and Cu(100) planes.
After room temperature adsorption of CO on reduced-Cu/
SiO2(473 K), the infrared spectrum exhibits an asymmetric
-
1
absorption band near 2145 cm . After 5 min under vacuum
this band disappears to a negligible intensity (<0.25% of original
band). The frequency of this band falls into the frequency range
for CO adsorbed on CuO/SiO2 or atomically rough but reduced
Cu/SiO2 as discussed above and shown in Figure 1. Because
there is no formation of CO2 or surface carbonate and the CO
was easily removed upon evacuation of the gas phase, we
conclude that the copper is reduced and therefore assign the
-
1
2
145-cm band to CO adsorbed on reduced but atomically
Figure 3. Infrared spectra taken after the adsorption of 15.0 Torr of
rough copper particles.
3
CH I for 1 h on (a) SiO
2
, (b) reduced-Cu/SiO
2
(673 K), (c) reduced-
The absorption band observed in the IR spectrum of CO
adsorbed on oxidized-Cu/SiO2 is symmetric with a maximum
2
Cu/SiO (473 K), and (d) oxidized-Cu/SiO
2
. The infrared cell was
evacuated for 5 min prior to recording each spectrum. The OH stretching
region for the spectrum labeled c is shown in the inset. The negative
feature shown in the inset indicates there is a decrease in SiOH groups
-1
near 2116 cm (Figure 2). Gas-phase carbon dioxide was also
observed after introduction of CO into the sample cell. Upon
evacuation of the gas phase, the CO absorption band shifted to
3 2
upon adsorption of CH I on reduced-Cu/SiO (473 K).
-
1
2
108 cm and an additional, weak band became apparent in
-
1
the spectrum near 1986 cm . The integrated area of the CO
absorption band decreased by 60% from its original value after
evacuation for 5 min and 75% after 60 min. The frequency of
approximately 1 h at room temperature followed by evacuation.
It has been previously shown that the surface becomes saturated
with surface-bound reaction products after reaction times of
-
1
3
the 2116-cm band is low for oxidized copper, but it is close
approximately 1 h. It is obvious upon initial inspection that
-1
to the band observed near 2118 cm by Millar, Rochester, and
Waugh after the surface oxidation of Cu/SiO2 using N2O to form
each of the three different copper samples exhibits the formation
of unique products on the catalyst surface after the room
temperature adsorption of methyl iodide. It is also clear that
pure silica does not react with CH3I under these conditions as
there are no new absorptions after exposure of CH3I to SiO2
(Figure 3a). Therefore, it can be concluded that the adsorption
of CH3I on each of the three samples is a result of the presence
of the copper particles.
3
9
Cu2O/SiO2. In addition, those samples oxidized with N2O
form gaseous CO2 upon reaction with CO and the CO remains
present for longer periods of time on the copper surface after
-
1
evacuation of the gas phase. Therefore, the band at 2116 cm
is assigned to CO adsorbed on oxidized copper, most probably
Cu2O/SiO2.
Transmission electron micrographs were taken of the Cu/
SiO2 samples using a Hitachi H-600 electron microscope
operating at an acceleration voltage of 100 kV. The particles
of these samples were found to have diameters between 2 and
The spectrum taken after methyl iodide adsorption on the
reduced-Cu/SiO (673 K) sample (Figure 3b) is characterized
2
by absorption bands in the C-H stretching region near 2913
-
1
and 2805 cm . Both the reduced-Cu/SiO (473 K) (Figure 3c)
2
5
nm. For reduced-Cu/SiO2(673 K) samples, the copper
and the oxidized-Cu/SiO (Figure 3d) spectra contain multiple
2
particles ranged in size from 3 to 5 nm in diameter. The average
diameter, obtained from a sampling of 50 particles, was 3.7
nm. For reduced-Cu/SiO2 (473 K) samples, the copper particles
ranged in size from 2 to 5 nm in diameter and the average
diameter was 3.4 nm.
bands in the C-H stretching region, although the frequencies
of the bands and the intensity patterns are quite distinct between
the two samples. The reduced-Cu/SiO (473 K) spectrum
2
(Figure 3c) exhibits absorption bands in the region between 2800
-
1
-1
and 3000 cm near 2985, 2954, 2924, and 2852 cm and one
-
1
In summary, these three samplessreduced-Cu/SiO2(673 K),
reduced-Cu/SiO2(473 K), and oxidized-Cu/SiO2sare composed
of copper nanoparticles between 2 and 5 nm in diameter. The
samples reduced in H2 for long times and at high temperatures,
reduced-Cu/SiO2(673 K) samples, consist of Cu nanoparticles
that are reduced and have surfaces that are atomically smooth
consisting of (111) and (100) planes. Reduced-Cu/SiO2(473
K) samples consist of Cu nanoparticles that are reduced and
have surfaces that are atomically rough. Oxidized-Cu/SiO2
samples are oxidized surfaces with a surface stoichiometry of
Cu2O; the surface morphology of these samples is unknown.
in the C-H deformation region near 1462 cm . The inset in
Figure 3 displays a loss in the OH stretching region for the
SiOH groups after reaction with CH3I, indicating consumption
of OH groups during adsorption of CH3I onto reduced-Cu/SiO2-
(473 K). The oxidized-Cu/SiO2 spectrum (Figure 3d) exhibits
infrared bands near 2962, 2927, 2895, 2861, and 2814 cm in
the C-H stretching region. Several bands are also observed
for the oxidized-Cu/SiO2 sample in the region extending from
-
1
-1
1300 to 1600 cm with frequencies of 1584, 1570, 1530, 1468,
-
1
1446, 1428, 1388, and 1357 cm . The 1584-, 1446-, 1428-,
-
1
and 1388-cm bands are very weak and are not labeled in
Figure 3d. Again, it is clear that different surface species form
on each of the Cu/SiO2 samples.
Room Temperature Adsorption of CH3I on Reduced-Cu/
SiO2(673 K), Reduced-Cu/SiO2(473 K), and Oxidized-Cu/
SiO2. The infrared spectra shown in Figure 3 were recorded
for four different sampless(a) SiO2, (b) reduced-Cu/SiO2(673
K), (c) reduced-Cu/SiO2(473 K), and (d) oxidized-Cu/SiO2safter
The data presented in the next two sections will show that
the following reactions (eqs 2-4) take place upon room
temperature adsorption of CH3I on reduced-Cu/SiO2(673 K),
reduced-Cu/SiO2(473 K), and oxidized-Cu/SiO2 to yield dif-
1
5 Torr of CH3I had been introduced into the infrared cell for

Reactions of CH3 on Si-Supported Cu Nanoparticles
J. Am. Chem. Soc., Vol. 119, No. 7, 1997 1701
Table 1. Vibrational Assignment of Surface Species Formed from Reaction of CH
3
I on Silica-Supported Copper
I. Methyl Adsorbed on Copper
mode descriptiona
Cu/SiO
(673 K)b
Cu/SiO
(473 K)b
Cu(100)c
Cu(111)d
2835
Pt(111)e
2879
2
2
ν
s
(CH
ν(CH)
(CH
3
)
2913
2805
2924
2915
2760
1430
1150
δ
δ
a
3
)
s
(CH
3
)
1200
II. Methyl Adsorbed on Silica-SiOCH
mode descriptionf
3
spilloverb
CH
OH-SiO g
CH
OH-SiO
h
3
2
3
2
ν
ν
a
a
(CH
(CH
3
)
)
2985
2954
2930 (sh)
2852
2998
2961
2930 (sh)
2860
1465
3000
2958
2928 (sh)
2858
1464
3
i
overtone or combination band
ν
s
3
(CH )
δ(CH
3
)
1462
III. Methoxide Adsorbed on Copper
mode description j
b
Cu/SiO k
l
Cu(100)m
Cu(110)n
2940
Cu/SiO
2
2
Cu/SiO
2
ν
ν
ν
as(CH
as(CH
3
)
)
)
2927
2895
2814
1468
2920
2880
2824
2930
2901
2861
2787
3
s
(CH
3
2821
1443
2840
1460
3
δ(CH )
IV. Formate Adsorbed on Copper
mode descriptiono
Cu/SiO b
Cu/SiO p
Cu/SiO q
Cu(110)r
sodium formates
2
2
2
combination band [νas(COO) + δ(CH)]
(CH)
combination band [ν
as(COO)
δ(CH)
(COO)
2958
2867
2718
1564
1369
1360
2936
2856
2935
2851
2960
2920
2953
2841
2720
1567
1377
1366
ν
s
t
s
(COO) + δ(CH)]
ν
1579
1358
1553
1351
1560
ν
s
1360
a
Mode description taken from refs 18 and 48. b This work. c Reference 18. d Reference 48. e Reference 49. Mode description taken from ref
f
h
i
j
k
5
2. g This worksfrom methanol adsorption on SiO
. Reference 51. sh ) shoulder. Mode description taken from refs 53 and 54. Reference 13.
2
p r
l
Reference 14. Reference 53. Reference 54. Mode description taken from ref 55. Reference 56. q Reference 11. Reference 54. s Reference
m
n
o
t
5
7. Only observed at high coverages.
cm- may be due to the symmetric C-H stretch of methyl
1
ferent surface-bound products.
fragments adsorbed on different sites or to an overtone of the
asymmetric deformation mode.
The spectrum obtained after reaction of CH3I on reduced-
Cu/SiO2(473 K) (Figure 3c) is much more complex than the
spectrum for reduced-Cu/SiO2(673 K) (Figure 3b). This
spectrum has been assigned previously to two different surface
species, Cu-CH3 and SiOCH3 (see Table 1). The band near
924 cm is assigned to methyl groups adsorbed on the surface
of Cu particles with atomically rough morphology and is
approximately 10 cm higher in frequency than that for methyl
reduced-Cu/SiO (673 K)
24,49
2
CH3I
8 CH -Cu + I-Cu
(2)
3
reduced-Cu/SiO (473 K)
2
CH3I
CH3I
8 CH -Cu + I-Cu + Si-OCH
3
3
(
3)
4)
3
,4
oxidized-Cu/SiO2
-1
2
8 CH O-Cu + I-Cu + HCO -Cu
3
2
(
-1
These reactions of CH3I to form adsorbed methyl groups on
groups adsorbed on atomically smooth copper surfaces (Figure
reduced copper particles (CH3-Cu) and the silica support
3b). The remaining bands in the reduced-Cu/SiO (473 K)
2
(
SiOCH3) and to form adsorbed oxygenates (CH3O-Cu and
spectrum are assigned to SiOCH groups formed from the
3
50
HCO2-Cu) on oxidized copper particles are discussed in detail
below.
reaction of methyl groups, from methyl spillover, with surface
hydroxyl groups. It has been previously shown that hydroxyl
groups present near the Cu particles can react with methyl
Identification of Surface Species on Reduced Cu/SiO2
Sample. The spectrum recorded after reaction of CH3I on
reduced-Cu/SiO2(673 K) is simple and consists of two bands
3
,4
groups, from methyl spillover, to form SiOCH3.
The iden-
tification of SiOCH is made possible by forming this species
3
-
1
51
with frequencies of 2913 (s) and 2805 (w) cm . The simple
spectrum suggests there is a single species adsorbed on the
surface. The spectrum is not consistent with that of adsorbed
methyl iodide which has infrared absorption bands near 3026,
through the reaction of methanol and silica. The spectrum of
SiOCH from reaction of methanol with silica is shown in Figure
3
4a and the spectrum of reduced-Cu/SiO (473 K) after methyl
2
iodide reaction is shown in Figure 4b. There is fairly good
-
1
48
2
910, and 1420 cm when adsorbed on a metal surface. The
agreement between the frequencies and the intensity patterns
observed in the two spectra shown in Figure 4 (a and b).
An important experiment that clearly demonstrated the role
of silica hydroxyl groups in the chemistry of methyl iodide
2
spectrum is more consistent with adsorbed methyl groups
formed from the dissociation of the relatively weak C-I bond.
-
1
The 2913-cm
band is close in frequency to the C-H
-1 18
symmetric stretch of adsorbed methyl on Cu(100) (2915 cm ).
On Pt(111), the symmetric C-H stretch in adsorbed methyl has
been reported near 2879 cm . The assignment of adsorbed
methyl groups is given in Table 1. The weaker band at 2805
began with a reduced-Cu/SiO (473 K) sample that was then
(49) Fan, J.; Trenary, M. Langmuir 1994, 10, 3649.
-
1 49
(50) Spillover is a well documented phenomenon on supported metal
catalysts especially for hydrogen. For a general review see: Connor, W.
C., Jr.; Falconer, J. L. Chem. ReV. 1995, 95, 765.
(48) Lin, J.-L.; Bent, B. E. Chem. Phys. Lett. 1992, 194, 208.
(51) Morrow, B. A. J. Chem. Soc., Faraday Trans. 1 1974, 1527.

1702 J. Am. Chem. Soc., Vol. 119, No. 7, 1997
Driessen and Grassian
after room temperature reaction of CH3I on an oxidized-Cu/
SiO2 sample (Figure 3d) correlate well with bands observed
for methoxide (OCH3) on copper surfaces, as has been discussed
3
4
-1
previously. The bands at 2927, 2895, 2814, and 1468 cm
are assigned to methoxide adsorbed on the copper particles (see
Table 1). It is well-known that methoxide can decompose to
1
3-15
form surface formate.
A number of the bands that do not
correlate with methoxide in the CH3I-oxidized Cu/SiO2 spectrum
do match those of adsorbed formate. The bands at 2958, 2867,
-1
2
718, 1564, 1369, and 1360 cm can be assigned by compari-
son with literature spectra to bidentate formate adsorbed on the
copper particles. (The frequency and intensity pattern of these
bands are coverage dependent. The highest coverage spectrum
was used for the bidentate formate assignmentssee spectrum
1
5,55-58
labeled T ) 473 K, Figure 9.)
The shoulder at 1584
-
1
-1
cm is near the 1583-cm band observed for unidentate
formate adsorbed on an oxidized Cu/SiO2 catalyst. Although
59
distinct bands in the C-H stretching region for a unidentate
-
1
formate species (2978 and 2904 cm ) are not observed, these
bands could be easily obscured by the intense absorption bands
for adsorbed methoxy and bidentate formate. A possible
Figure 4. Infrared spectra of (a) SiO
SiOCH groups (b) reduced-Cu/SiO (473 K) after adsorption of methyl
iodide for 1 h and subsequent evacuation. The two spectra labeled a
and b are quite similar and indicate that SiOCH is formed on reduced-
Cu/SiO samples from CH I adsorption. One difference between the
two spectra is the band near 2924 cm which is assigned to CH
2 3
after reaction of CH OH to form
3
2
-
1
3
assignment for the shoulders observed at 1446 and 1428 cm
2
3
is that of the δ(CH3) of methoxy groups adsorbed on different
surface sites. In summary, the bands in the oxidized-Cu/SiO2
spectrum have been identified as CH3O-Cu (2927, 2895, 2814,
-
1
3
-1
adsorbed on the copper surface. The assignment of the 2924 cm band
is confirmed by the spectrum labeled c. The spectrum labeled c was
-1
and 1468 cm ) and HCOO-Cu (2958, 2867, 2718, 1564, 1369,
taken following adsorption of CH
sample that had been heated to 673 K to decrease the coverage of SiOH
groups on the support (this sample is denoted as reduced-Cu/SiO (473
Kf673 K)). The IR data show that SiOCH formation is dependent
on the silica hydroxyl group coverage. The lone band in the spectrum
3 2
I on a reduced-Cu/SiO (473 K)
-
1 34
and 1360 cm ). The assignments of these bands are given
-
1
in Table 1. The band at 1530 cm does not correlate with
any of the above species and is left unassigned.
2
3
Temperature-Dependent Infrared and TPD Results for
CH3I on Reduced Cu/SiO2 Samples. The reaction chemistry
of the surface species on reduced Cu/SiO2 catalysts as a function
of temperature was further investigated. Infrared experiments
were done by heating the samples in increments of 50 K for 60
s then cooling the sample down to room temperature after which
an infrared spectrum was recorded. During these experiments,
the infrared cell containing the sample was isolated from the
pumping system in order to observe the formation of gas-phase
species. Complementary TPD data were collected in separate
experiments.
After adsorption of CH3I on the reduced-Cu/SiO2(673 K)
sample and subsequent evacuation of the gas phase, the sample
was warmed to elevated temperatures as described above and
then cooled to room temperature before a spectrum was taken.
Figure 5 shows the temperature-dependent infrared data for a
reduced-Cu/SiO2(673 K) sample. As the data show, the bands
-1
at 2922 cm is due to methyl fragments adsorbed on the copper surface.
See text for further details.
heated to 673 K for 1 h under vacuum. This treatment affected
the hydroxyl group coverage of the silica support. By integrat-
ing the area of the bands in the silica hydroxyl group region
before and after heating to 673 K it is estimated that ap-
proximately 25% of the Si-OH groups have been removed due
to heating. It should be noted that the frequency of adsorbed
-
1
CO on the copper surface only shifted by 4 cm from 2147
-
1
-1
cm before heating to 2143 cm after heat treatment. This
small shift in the adsorbed CO frequency demonstrates that the
copper surface itself has not changed significantly after the 1 h
heating process under vacuum, in contrast to the large change
in surface morphology after heating to 673 K for 24 h, as
previously discussed. After preparation of the Cu/SiO2 sample
as detailed above (denoted as reduced-Cu/SiO2(473 Kf673 K)),
the sample was then exposed to 15.0 Torr of methyl iodide for
a period of 1 h. Following evacuation, an infrared spectrum
was taken, as shown in Figure 4c. As evidenced by the infrared
-
1
near 2913 and 2805 cm begin to decrease in intensity near
423 K and are completely gone by 473 K. Concomitant with
the decrease in the 2913- and 2805-cm bands is the growth
-
1
-1
-1
absorption near 2922 cm , only the Cu-CH3 species is formed
on these samples as is observed on reduced-Cu/SiO2(673 K)
samples. The conclusion drawn from these results is that heating
to 673 K removes a portion of the SiOH groups but does not
significantly change the surface of the copper particles and the
adsorption of CH3I and CH3 on the Cu particles is the same.
However, SiOH groups that play a role in adsorption of CH3I
on reduced-Cu/SiO2(473 K), most probably those in close
proximity to the metal particles, have been removed and do
not participate in the reaction.
of a band near 3017 cm , which is characteristic of gas-phase
methane (see inset, Figure 5). This reaction temperature near
423-473 K is also near the temperature where methyl fragments
react to form methane, ethane, and ethylene on single crystal
1
7,18,24,60
copper surfaces.
The TPD data corroborate the infrared results. Figure 6
shows the TPD spectrum after reaction of reduced-Cu/SiO (673
2
(
55) Hayden, B. E. In Vibrational Spectroscopy of Molecules on Surfaces;
Yates, J. T., Jr., Madey, T. E., Eds.; Plenum: New York, 1987; p 303.
56) Millar, G. J.; Newton, D.; Bowmaker, G. A.; Cooney, R. P. Appl.
(
Identification of Surface Species on Oxidized Cu/SiO2.
Several of the infrared bands observed in the spectrum obtained
Spectrosc. 1994, 48, 827.
(57) Ito, K.; Bernstein, H. J. Can. J. Chem. 1956, 34, 170.
58) Hayden, B. E.; Prince, K. C.; Woodruff, D. P.; Bradshaw, A. M.
(
(
52) Pelmenschikov, A. G.; Morosi, G.; Gamba, A.; Zecchina, A.;
Surf. Sci. 1983, 133, 589.
(59) Millar, G. J.; Rochester, C. H.; Waugh, K. C. J. Chem. Soc., Faraday
Trans. 1991, 87, 1491.
(60) Chiang, C.-M.; Wentzlaff, T. H.; Bent, B. E. J. Phys. Chem. 1992,
96, 1836.
Bordiga, S.; Paukshtis, E. A. J. Phys. Chem. 1993, 97, 11979.
(
(
66.
53) Ryberg, R. Phys. ReV. B 1985, 31, 2545.
54) Sexton, B. A.; Hughes, A. E.; Avery, N. R. Surf. Sci. 1985, 155,
3

Reactions of CH3 on Si-Supported Cu Nanoparticles
J. Am. Chem. Soc., Vol. 119, No. 7, 1997 1703
Figure 5. Infrared spectra taken as a function of temperature after the
adsorption of CH I on reduced-Cu/SiO (673 K). The 298 K spectrum
3 2
is identical to the one shown in Figure 3b. The sample was then heated
to the indicated temperatures and cooled down prior to recording a
spectrum. The infrared cell was isolated from the pumping system so
that gas-phase products from the desorption/decomposition could be
monitored. The inset displays an expanded view of the spectrum labeled
4
73 K to show the presence of a band near 3017 cm-1, which is assigned
Figure 7. The temperature-dependent infrared spectra recorded after
to gas-phase methane.
3 2
adsorption of CH I on reduced-Cu/SiO (473 K). The 298 K spectrum
is identical to the one shown in Figure 3c. The sample was then heated
to the indicated temperatures and cooled down prior to recording a
spectrum. The infrared cell is isolated from the pumping system in
order to monitor gas-phase products. The inset displays an expanded
-
1
view of the spectrum labeled 773 K to show the band near 3016 cm
which is associated with gas-phase methane.
,
Reduced-Cu/SiO2(473 K) samples exhibit more complicated
infrared and TPD spectra following methyl iodide adsorption.
As discussed above, two surface species have been identified
on reduced-Cu/SiO2(473 K): methyl groups adsorbed on the
copper particles and SiOCH3 formed from methyl spillover.
These two species differ in thermal stability. Figure 7 displays
the temperature-dependent infrared spectra after reaction of CH3I
on reduced-Cu/SiO2(473 K) at room temperature. As the data
-
1
show, the band near 2924 cm decreases in intensity as the
temperature is raised to 423 K concomitant with the growth in
-
1
a band near 3016 cm , attributed to gas-phase methane
expanded spectrum shown in inset). The remaining bands are
(
-
1
unaffected. The bands near 2985, 2954, and 2852 cm begin
Figure 6. TPD following adsorption of CH
3
I on reduced-Cu/SiO
2
-
4
to decrease near 573 K and do not fully disappear until 873 K
(
(
673 K) for 1 h. Three desorption products were detected in TPD: CH
m/e 16), C (m/e 27), and C (m/e 30).
-1
(
not shown). Gas-phase CO2 (2349 cm ) and CO adsorbed
2
H
4
2
H
6
-1
on the copper surface (2132 cm ) grow in intensity as the
temperature is raised above 473 K. These data further support
the conclusion that there are two species present on the
surface: one that is chemically similar to the species formed
on reduced-Cu/SiO2(673 K), i.e. Cu-CH3, and one that remains
on the surface to higher temperatures, i.e. SiOCH3.
K) with CH3I. There is a peak near 425 K for CH4 (m/e 16)
which is consistent with the temperature-dependent infrared data.
In addition to methane, the TPD data show that both ethane
(
m/e 30) and ethylene (m/e 27) evolve from the sample near
70 and 390 K, respectively. The TPD data are nearly identical
with the results obtained on single crystal copper sur-
3
The TPD data corroborate the temperature-dependent infrared
data. Figure 8 shows that two sets of TPD peaks are observed,
those with desorption rate maxima below 550 K and those above
550 K. A TPD peak is evident at 440 K for methane (m/e 16)
desorption and another peak is observed at 410 K for ethane
(m/e 30) desorption. Methane and ethane evolution from
reduced-Cu/SiO2(473 K) have desorption rate maxima that are
approximately 20-25 K higher than those observed for the
reduced-Cu/SiO2(673 K) samples. Importantly, there is no TPD
1
7,18,24,60
faces.
Bent and co-workers have reported the reaction
of methyl groups on copper single crystals to form methane,
ethane, and ethylene. Methane desorbs with a desorption rate
24
60
maximum near 450 K on Cu(111) and 470 K on Cu(110)
18
and Cu(100). Ethane desorbs with a desorption rate maximum
near 400 K from Cu(110)60 after adsorption of CH3I and
ethylene formation and desorption occurs near 470 K on Cu-
(
100).18 White and co-workers observe similar desorption
(61) Roop, B.; Zhou, Y.; Liu, Z.-M.; Henderson, M. S.; Lloyd, K. G.;
6
1
products from methyl reaction on polycrystalline Cu films.
Campion, A.; White, J. M. J. Vac. Sci. Technol. A 1989, 7, 2121.

1704 J. Am. Chem. Soc., Vol. 119, No. 7, 1997
Driessen and Grassian
Figure 8. TPD following the adsorption of CH
473 K). Desorption peaks were seen in several mass channels including
O (m/e 18), CO (m/e 44), C (m/e 30), CO (m/e 28) and CH
m/e 16). Desorption peaks below 550 K are due to CH reactions on
the Cu particles. No desorption peaks were seen for C (m/e 27)
not shown). Desorption peaks above 550 K are due to the decomposi-
tion of SiOCH groups.
3 2
I on reduced-Cu/SiO -
(
H
2
2
H
2 6
4
(
3
2 4
H
(
3
Figure 9. The temperature-dependent infrared spectra recorded after
reaction of CH I on oxidized-Cu/SiO . The 298 K spectrum is identical
3 2
peak for ethylene desorption from the reduced-Cu/SiO2(473 K)
sample (not shown) as there was for the reduced-Cu/SiO2(673
K) sample.
TPD peaks evident above 550 K are due to the decomposition
of SiOCH3. Desorption of CH4, CO and CO2 occurs near 615
K. Characteristic IR bands for these species are observed in
the temperature-dependent infrared spectra (see Figure 7).
Water evolution (m/e 18) is also observed over a broad
temperature range, as the silica support becomes dehydroxylated
at higher temperatures.
to the one shown in Figure 3d. The sample was then heated to the
indicated temperatures and cooled down prior to recording a spectrum.
As in previous experiments, the infrared cell was isolated from the
pumping system in order to observe gas-phase species produced upon
heating of the sample.
growth in intensity of a band centered at 2349 cm- (gas-phase
CO2) after heating to 423 K and above. The bands formed at
higher temperatures can be attributed to a combination of surface
COx species which are formed from the decomposition of
adsorbed formate. Frequency ranges for adsorbed COx species
1
Temperature-Dependent Infrared and TPD Results for
CH3I on Oxidized-Cu/SiO2. After adsorption of methyl iodide
on oxidized-Cu/SiO2, a complex infrared spectrum is seen
6
4
are as follows: bidentate carbonate (CO3), 1590-1630 and
-
1
65 a
1
1
1
260-1270 cm ; carboxyl (CO2),
1350-1420 and 1550-
(Figure 3d) which is assigned to two adsorbed oxygenates on
-1
64
630 cm ; organic-like carbonate (CO(O)2), 1870-1750 and
the copper surface: methoxide and bidentate formate. The
temperature-dependent infrared data (Figure 9) show the forma-
tion and desorption of several species which have also been
observed in the decomposition reaction of methanol on sup-
-
1
65b
-1
250-1280 cm ; and bridged CO, 1700-1900 cm . It is
most likely that it is a CO2 species which is responsible for the
-
1
bands at 1625 and 1429 cm in the spectrum recorded after
heating to 623 K, since it is the initial dehydrogenation product
1
3-15
ported copper catalysts.
98 and 473 K the bands near 2927, 2895, 2814, and 1468 cm-
decrease in intensity concomitant with the growth of bands near
538 (not shown), 2958, 2867, 2718, 1564, 1369, and 1360
As the sample is heated between
1
of bidentate formate. Upon further heating, the adsorbed CO2
2
may disproportionate to yield bridged CO and adsorbed CO3.
-
1
The 1893-cm band cannot be firmly assigned to a specific
CO species since any other associated bands may be masked
3
-
1
cm . Upon warming to 523 K, these bands decrease sharply
as bands near 3608 (not shown), 1663, 1615, and 1429 cm
develop.
x
-
1
-
1
by silica absorptions (below 1300 cm ).
TPD results, presented in Figure 10, provide confirmation
of the infrared results and the gas-phase species observed in
the IR spectra. Three desorption peaks are evident in the
spectrum. H2O (m/e 18) and CO2 (m/e 44) desorption occur
over broad temperature ranges from 400 to 800 K and 500 to
700 K, respectively. Formaldehyde, which was not detected
in the IR data, desorbs with a desorption rate maximum of 450
The hydrocarbon bands which grow in intensity between 323
and 473 K are attributed to bidentate formate (CH(O)2) on the
copper surface due to both the frequencies of the bands and the
fact that they reach maximum intensity and then decrease
sharply near temperatures where bidentate formate has been seen
to decompose (>425 K) on oxidized Cu/SiO213 and Cu(100).
The band which develops (along with the bidentate formate
62
66
K. Formaldehyde is a product in methanol decomposition on
bands) at 3538 cm- is assigned to ν(OH) of Cu-OH. The
bands which have been attributed to unidentate formate either
decrease or become obscured by the intense bidentate formate
bands above 373 K.
1
63
copper surfaces as well.
13
(64) Little, L. H. Infrared Spectra of Adsorbed Species; Academic: New
York, 1966.
(65) Davydov, A. A. Infrared Spectroscopy of Adsorbed Species on the
Heating to 673 K evolved yet another band (not shown) near
Surface of Transition Metal Oxides; Wiley and Sons: Chichester, 1990;
(a) p 38, (b) p 118.
-
1
1
893 cm . Also observed is the development and constant
(66) Although formaldehyde has the same mass as ethane (m/e 30), the
(
62) Sexton, B. A. Surf. Sci. 1979, 88, 319.
observed fragmentation pattern for this peak is only consistent with
formaldehyde and not ethane.
(63) Ellis, T. H.; Wang, H. Langmuir 1994, 10, 4083.

Reactions of CH3 on Si-Supported Cu Nanoparticles
J. Am. Chem. Soc., Vol. 119, No. 7, 1997 1705
Scheme 1. Reactions of Methyl Iodide and Methyl
Fragments on Cu/SiO2
Figure 10. TPD following adsorption of CH
following adsorption of CH I. Desorption peaks were evident for H
CO (m/e 30), CO (m/e 44), and H O (m/e 18).
3
I on oxidized-Cu/SiO
2
3
2
-
2
2
In addition to the experiments discussed above using copper
nitrate as the copper precursor, we have also prepared Cu/SiO2
samples using copper acetate monohydrate as the copper
precursor. These experiments were done to probe the effect of
the copper precursor on the chemistry of methyl fragments. Data
obtained from the copper nitrate precursor samples are repro-
duced when methyl iodide is adsorbed onto the Cu/SiO2 samples
made using copper acetate monohydrate, indicating little effect
of the copper precursor at least in terms of reaction products
formed. Samples with lower copper metal loading were also
examined. The reaction products were the same but the band
intensities of surface bound species decreased with metal loading
(between 15 and 4%).
Discussion
In general, oxide-supported metal particle catalysts are
complex and it is often times difficult to determine the important
parameters that affect the reactivity of these systems. In this
study, we have identified at least three important factors that
control the surface chemistry of methyl groups, formed from
the dissociative adsorption of methyl iodide, on Cu/SiO2
samples. These three factors are (i) the oxidation state of the
copper particles, (ii) the hydroxyl group coverage on the surface
of the silica support, and (iii) the surface roughness of the copper
particles. The effects that these three factors have on the
chemistry of methyl groups on Cu/SiO2 are discussed in detail
below.
After CH3I adsorption on reduced-Cu/SiO2(473 K), methyl
groups are present not only on the copper particles but also on
the silica support as SiOCH3 groups. The SiOCH3 groups
-
1
Surface Chemistry of CH3I on Reduced- and Oxidized-
Cu/SiO2 Samples. Similar to the single crystal surface chem-
istry of methyl iodide, C-I bond dissociation occurs to give
methyl and adsorbed iodine on the copper particle surface of
the reduced Cu/SiO2 samples (Scheme 1). Although we have
not detected adsorbed iodine in our experiments, we will assume
that, similar to what occurs on single crystal surfaces, the iodine
atom strongly bonds to the copper surface. On reduced-Cu/
SiO2(673 K), Cu-CH3 is the only species formed upon
adsorption of CH3I. This species is characterized by infrared
exhibit infrared bands near 2985, 2954, 2852, and 1462 cm
(Table 1). Because there is no direct reaction with methyl iodide
andthe silica surface under the conditions employed in this study,
the formation of SiOCH3 is a consequence of C-I bond
dissociation on the copper particles, followed by reaction of
methyl with the surface hydroxyl groups.
The number of SiOH groups that react with methyl fragments
per copper particle on reduced-Cu/SiO2(473 K) has been
calculated from the average particle diameter from TEM
measurements, the mass of both copper and silica in the sample,
an initial hydroxyl group coverage on the silica of 5 OH groups/
-1
bands near 2913 and 2805 cm (weak). The infrared spectrum
of adsorbed methyl groups on reduced-Cu/SiO2(473 K) samples
2
64
100 Å , and the percentage of SiOH groups lost after
adsorption of methyl iodide. This calculation yields ap-
proximately 25 OH groups reacting per copper particle in the
sample. Hydrogen atoms produced in the spillover process are
-
1
is characterized by an absorption band near 2924 cm that is
assigned to the symmetric C-H stretch of adsorbed methyl
groups on atomically rough copper particles.

1706 J. Am. Chem. Soc., Vol. 119, No. 7, 1997
Driessen and Grassian
available for reaction with methyl groups adsorbed on the Cu
particles as discussed in the next section.
There are two possible mechanisms for this reaction. One
is a stepwise mechanism (eq 5), similar to that observed for
hydrogen spillover, and has been discussed previously.
K) samples. To determine if there is another source of hydrogen
atoms, TPD data for deuterated methyl iodide on reduced-Cu/
SiO2(473 K) were obtained. If CD3H evolves in TPD, then it
is clear there is another source of hydrogen atoms available for
the formation of methane. However, if CD3 groups undergo
R-elimination to give CD2 and D atoms then there should be
exclusively CD4. In experiments with CD3I on reduced-Cu/
SiO2(473 K), a significant amount of CD3H evolution is
observed in TPD indicating there is another source of hydrogen
atoms available for reaction with CH3 groups. Hydrogen atoms
from SiOH plus methyl reaction (eq 5b) is the most likely source
of hydrogen available for reaction with methyl fragments.
The chemistry of CH3 fragments on reduced samples is
summarized in Scheme 1. For reduced-Cu/SiO2(473 K) samples,
the chemistry of SiOCH3 groups from methyl spillover is
included in the reaction scheme.
Role of Surface Roughness on the Surface Chemistry of
Methyl Groups. As discussed previously, the two reduced Cu/
SiO2 samples show differences in the copper particle surface
morphology. The long reduction at higher temperature (reduced-
Cu/SiO2(673 K)) has an annealing effect on the surface of the
copper particles. Similar to adsorbed CO, the frequency of the
symmetric stretch in adsorbed methyl is blue-shifted on reduced-
Cu/SiO2(473 K) samples that have particles of rough surface
morphology compared to reduced-Cu/SiO2(673 K). The shift
to higher frequency for the rougher surfaces is attributed to
adsorption on unsaturated copper atoms.
Cu + CH I f Cu-CH + Cu-I
(5a)
(5b)
3
3
SiOH + Cu-CH f SiOCH + H
3
3
Alternatively, the reaction may be a concerted one occurring at
the Cu-silica interface. As the C-I bond dissociates and forms
an I-Cu bond, the CH3 group can simultaneously react with
SiOH to form SiOCH3. In contrast to hydrogen spillover, which
can occur over macroscopic distances,5 the CH3 spillover
appears to be limited to OH groups in close proximity to the
metal particles. The fact that spillover is limited for CH3
suggests that the concerted mechanism may be operative in this
case.
0
On oxidized-Cu/SiO2 samples, the C-I bond in adsorbed
CH3I dissociates as it does on the reduced samples, however,
the predominant products that form on the oxidized copper
surface are methoxy and bidentate formate (Scheme 1). The
infrared spectra of these two surface bound species are assigned
by direct comparison to literature spectra (Table 1).
Surface Chemistry of Methyl Groups on Reduced Cu/
SiO2 Samples and the Role of Silica Hydroxyl Group
Coverage. The chemistry of methyl groups adsorbed on the
surface of the copper particles as a function of temperature can
be discerned from the IR and TPD data. For reduced-Cu/SiO2-
More subtle differences may also be discerned from the data.
The TPD data show different desorption rate maxima on
reduced-Cu/SiO2(673 K) (Figure 6) and reduced-Cu/SiO2(473
K) (Figure 8). The reduced-Cu/SiO2(673 K) shows a desorption
rate maximum near 370 K for ethane desorption compared to
(673 K) samples, identical reaction products are detected in TPD
to those produced on a single crystal Cu(110) surface. There-
fore, we use the reaction sequence determined for methyl
reactions on Cu(110) to explain the results for methyl reactions
on reduced-Cu/SiO2(673 K) samples (eqs 6-9).
4
10 K for ethane desorption from the reduced-Cu/SiO2(473 K)
sample. This difference in the desorption temperature for the
two reduced Cu/SiO2 samples represents a change in the reaction
kinetics for methyl fragments adsorbed on copper due to a
change in the surface morphology of the copper particles.
Diffusion of methyl fragments may be more facile on the
smoother copper surfaces.
CH f CH + H
(6)
(7)
(8)
(9)
3
2
H + CH f CH (g)
3
4
Surface Chemistry of Methyl Groups on Oxidized-Cu/
SiO2. The surface-bound products that are formed at room
temperature during the adsorption of methyl iodide on the
oxidized-Cu/SiO2 sample can be attributed to the formation and
partial dehydrogenation of methoxy groups adsorbed on the
copper surface. Scheme 1 depicts the initial formation of
methoxy from the dissociative adsorption of the CH3I and its
reaction with adsorbed oxygen. A portion of the methoxy
groups dehydrogenate to produce formate and hydroxyl groups
adsorbed on the copper surface. Both methoxy and formate
are observed at room temperature after adsorption of methyl
iodide. Upon warming, the remaining methoxy groups dehy-
drogenate to increase the concentration of bidentate formate and
to form water through the addition of a hydrogen to Cu-OH
groups. It is also known from the TPD experiments that
formaldehyde is produced in the temperature range where the
infrared spectra show methoxy groups dehydrogenating and the
bidentate formate coverage increasing. At higher temperatures,
above 523 K, bidentate formate undergoes further reaction to
produce gaseous CO2 and adsorbed carbonate. The chemistry
of methyl iodide on oxidized-Cu/SiO2 is identical to that of
methanol adsorbed on oxidized-Cu/SiO2.
CH + CH f C H f C H (g) + H
2
3
2
5
2
4
CH + CH f C H (g)
3
3
2
6
Methane can only form if there are available hydrogen atoms
for reaction with methyl. It has been shown that the rate
determining step in methyl reactions on Cu(110) is R-elimination
to yield methylene and hydrogen atoms (eq 6). H atoms then
react with methyl groups to yield methane (eq 7). Methylene
groups can insert into the Cu-CH3 bond to give ethyl which
rapidly undergoes â-hydride elimination to give ethylene and
adsorbed hydrogen atoms (eq 8). At high coverages methyl
groups can also couple to evolve ethane (eq 9).
The different product distribution for reduced-Cu/SiO2(473
K) samples compared to reduced-Cu/SiO2(673 K) and Cu(110)
surfaces suggests that there is another reaction mechanism
occurring for methyl groups adsorbed on reduced-Cu/SiO2(473
K). On reduced-Cu/SiO2(473 K), there is no evidence for
ethylene formation and there is an increase in the amount of
methane produced relative to ethane. From the integrated area
of the TPD peaks, it was determined that there was nearly an
eight-fold increase in the ratio of CH4:C2H6 on reduced-Cu/
6
7
SiO2(473 K) compared to reduced-Cu/SiO2(673 K) samples.
(67) The relative yields of methane, ethane, and ethylene were determined
by calibrating the mass spectrometer with pure gases and correcting for
the different ion-gauge sensitivities. The CH4:C2H6 yield was calculated to
be 0.16 for CH3I reaction on Cu/SiO2(673 K) and 1.20 for CH3I reaction
on Cu/SiO2(473 K).
The increased methane production and the absence of ethylene
suggest that there is another source of hydrogen besides from
R-elimination in adsorbed methyl for the reduced-Cu/SiO2(473

Reactions of CH3 on Si-Supported Cu Nanoparticles
J. Am. Chem. Soc., Vol. 119, No. 7, 1997 1707
In order to determine if the C-I bond in CH3I breaks on the
adsorbed oxygen to form methoxy or if methyl groups are
perhaps formed on vacant copper sites and subsequently react
with nearby adsorbed oxygen to form Cu-OCH3, an experiment
was done where a fraction of a monolayer of methyl fragments
on a reduced-Cu/SiO2(673 K) sample was prepared. Oxygen
was then added in an attempt to observe the formation of
methoxy on the copper surface. Even with gentle heating in
oxygen (below the desorption temperature of CH3), the methyl
fragments showed no conversion to methoxy. This demonstrates
that the methyl fragments, if formed, are not stabilized on the
copper surface prior to their reaction with the adsorbed oxygen.
This is similar to results observed by Bol and Friend in an
analogous experiment on an oxygen pre-covered rhodium
formed on the copper particle surface of reduced Cu/SiO2
samples after the dissociative adsorption of methyl iodide. The
frequency of the C-H symmetric stretch is dependent on surface
morphology. For smooth surfaces, the frequency is near 2913
-
1
cm and for rough surfaces the frequency blue shifts to near
-
1
2924 cm . On samples that are heated to a maximum
temperature of 473 K, the hydroxyl group coverage on the silica
support is high enough such that some of the methyl fragments
migrate to and react with SiOH groups to form SiOCH and
3
hydrogen atoms. These hydrogen atoms can then react with
methyl groups to form gaseous methane at elevated tempera-
tures. Oxygenated hydrocarbon fragments are formed on the
copper particle surface on samples that have been oxidized prior
to methyl iodide adsorption. These fragments include methoxy
and bidentate formate on the copper surface. In general,
characterization of Cu/SiO2 samples is difficult. The samples
characterized in these studies may provide for a better under-
standing of the reactivity of Cu/SiO2 catalysts and the results
may be applicable to other alkyl fragments from the dissociative
adsorption of alkyl halides and other chemical systems, e.g.
methanol synthesis.
2
6
surface.
Another important observation is that on the oxidized-Cu/
SiO2 samples, no formation of SiOCH3 is observed. This is
significant because it shows that the spillover process is inhibited
by the presence of surface oxygen.
Conclusions
The chemistry of methyl groups on Cu/SiO2 from methyl
iodide dissociation has been investigated in this study. It has
been shown that the C-I bond in methyl iodide dissociates on
Cu/SiO2 samples at room temperature. The subsequent reaction
chemistry of methyl groups on Cu/SiO2 is dependent on sample
preparation. Different sample preparations lead to surfaces with
different surface morphology, oxidation state, and hydroxyl
group coverage on the silica support. Methyl fragments are
Acknowledgment. The authors gratefully acknowledge the
donors of the Petroleum Research Fund, administered by the
American Chemical Society, for partial support of this research
and the National Science Foundation (Grants CHE-9300808 and
CHE-9614134).
JA9633833