Surface chemistry of CN bond formation from carbon and nitrogen atoms on Pt(111)

Article abstract of DOI:10.1021/jp052434s

The mechanism of CN bond formation from CH₃ and NH₃ fragments adsorbed on Pt(111) was investigated with reflection absorption infrared spectroscopy (RAIRS), temperature-programmed desorption (TPD), and X-ray photoelectron spectroscopy (XPS). The surface chemistry of carbon-nitrogen coupling is of fundamental importance to catalytic processes such as the industrial-scale synthesis of HCN from CH₄ and NH₃ over Pt. Since neither CH₄ nor NH₃ thermally dissociate on Pt(111) under ultrahigh vacuum (UHV) conditions, the relevant surface intermediates were generated through the thermal decomposition of CH ₃I and the electron-induced dissociation of NH₃. The presence of surface CN is detected with TPD through HCN desorption as well as with RAIRS through the appearance of the vibrational features characteristic of the aminocarbyne (CNH₂) species, which is formed upon hydrogenation of surface CN at 300 K. The RAIRS results show that HCN desorption at a??500 K is kinetically limited by the formation of the CN bond at this temperature. High coverages of C_ads suppress CN formation, but the results are not influenced by the coadsorbed I atoms. Cyanide formation is also observed from the reaction of adsorbed N atoms and carbon produced from the dissociation of ethylene. ? 2005 American Chemical Society.

Full text of DOI:10.1021/jp052434s

17560
J. Phys. Chem. B 2005, 109, 17560-17566
Surface Chemistry of CN Bond Formation from Carbon and Nitrogen Atoms on Pt(111)
Eldad Herceg and Michael Trenary*
Department of Chemistry, UniVersity of Illinois at Chicago, 845 West Taylor Street,
Chicago, Illinois 60607-7061
ReceiVed: May 10, 2005; In Final Form: July 26, 2005
The mechanism of CN bond formation from CH3 and NH3 fragments adsorbed on Pt(111) was investigated
with reflection absorption infrared spectroscopy (RAIRS), temperature-programmed desorption (TPD), and
X-ray photoelectron spectroscopy (XPS). The surface chemistry of carbon-nitrogen coupling is of fundamental
importance to catalytic processes such as the industrial-scale synthesis of HCN from CH4 and NH3 over Pt.
Since neither CH4 nor NH3 thermally dissociate on Pt(111) under ultrahigh vacuum (UHV) conditions, the
relevant surface intermediates were generated through the thermal decomposition of CH3I and the electron-
induced dissociation of NH3. The presence of surface CN is detected with TPD through HCN desorption as
well as with RAIRS through the appearance of the vibrational features characteristic of the aminocarbyne
(CNH2) species, which is formed upon hydrogenation of surface CN at 300 K. The RAIRS results show that
HCN desorption at ∼500 K is kinetically limited by the formation of the CN bond at this temperature. High
coverages of Cads suppress CN formation, but the results are not influenced by the coadsorbed I atoms. Cyanide
formation is also observed from the reaction of adsorbed N atoms and carbon produced from the dissociation
of ethylene.
Introduction
coupling reaction of surface C and N atoms.5 The surface CN
is then hydrogenated to HCN. However, in another study they
speculate that surface CHx and NHy may be involved in a surface
reaction to form a CNHz species (z g 2), which is then
dehydrogenated to HCN.4 Therefore, to definitely establish the
key reaction steps in catalytic HCN synthesis, identification of
the surface species that combine to form the C-N bond is of
great importance.
The C-N coupling reaction to form CN on the Pt(111)
surface is directly relevant to the use of platinum gauze catalysts
in the industrial synthesis from CH4 and NH3 of hydrogen
cyanide, an important chemical intermediate for products such
as nylon.1 In the absence of oxygen (Degussa process), the
catalytic reaction is strongly endothermic and the heat must be
supplied (∼1500 K) externally by burning fuel, typically more
CH4 in another chamber. In contrast, addition of oxygen
(Andrussow process) gives a net exothermic reaction. This later
process has been used industrially to synthesize HCN since the
1940s.2 In this paper we focus on the system without the
presence of coadsorbed oxygen.
The synthesis of HCN over polycrystalline platinum has been
the subject of several laboratory-scale studies at both atmo-
spheric and reduced pressure.3-6 Given that several competing
reactions, such as NH3 decomposition to N2 and H2, would seem
to be favored by equilibrium thermodynamics, it is surprising
that the reaction can occur at all, particularly at total pressures
as low as 1 Torr. The kinetics of HCN formation in both the
absence4,5 and the presence6 of O2 was studied in detail by
Hasenberg and Schmidt. They reached several important
conclusions. First, the reaction is definitely metal catalyzed, as
opposed to occurring in the gas phase from reactive intermedi-
ates produced by the heated surface. Second, the rate law
indicates a simple Langmuir-Hinshelwood mechanism, which
implies that the key steps in the reaction occur between adsorbed
species and therefore can be mimicked in UHV studies. Third,
the coverage of surface carbon must be high enough to suppress
NH3 decomposition to N2 and H2 but not so high as to poison
the surface. Fourth, the same basic mechanism applies to the
reaction over both Pt and Rh. In one model reactor study they
provide indirect evidence for the formation of CN through the
Secondary ion mass spectrometry (SIMS) has been used to
detect C-N coupling reactions on Rh(111)7 and Rh-loaded
CeOx(111).8 However, to our knowledge, there are no previous
reports of this reaction on a platinum surface. The fact that the
actual catalyst for HCN synthesis is a 90% Pt-10% Rh gauze
makes a study on a platinum surface more relevant to the
catalytic reaction. Since Pt has a relatively high ionization
potential, which significantly lowers the probability of secondary
ion formation, static SIMS is not a good choice for observing
surface processes on Pt(111). In addition, the detection efficiency
is substantially lower for relatively high mass clusters in the
case of Pt compared to secondary ions that come from the first-
or second-row transition metals. Finally, SIMS is a destructive
technique that is difficult to quantify. Therefore, in this study
we use TPD of HCN (m/e ) 27) as a means to detect CN bond
formation. In the absence of hydrogen, CN desorbs as cyanogen
(C2N2) in the range of 600-800 K,9 but cyanogen desorption
was not detected here. Surface CN can also be sensitively
detected through its hydrogenation to the aminocarbyne species
(CNH2), which has a strong and characteristic RAIR spec-
trum.10,11 If CN bond formation occurs below ∼350 K (the onset
of H2 desorption) and there is hydrogen on the surface, then
we should expect to see the CNH2 spectrum. If however the
formation temperature is above 350 K, annealing a mixed CHx/
NHy overlayer to various temperatures, cooling to 300 K, and
then exposing to H2 will permit the detection of any CN on the
surface through the appearance of the CNH2 spectrum.
* Corresponding author. E-mail: mtrenary@uic.edu.
10.1021/jp052434s CCC: $30.25 © 2005 American Chemical Society
Published on Web 08/27/2005

Surface Chemistry of CN Bond Formation
J. Phys. Chem. B, Vol. 109, No. 37, 2005 17561
Both CH4 and NH3 have very small dissociative sticking
probabilities on Pt(111).12-14 Therefore, the reactive surface
species that form during catalytic processes involving CH4 and
NH3 must be generated by other means. Methyl iodide (CH3I)
provides a convenient means to generate surface CHx (x ) 0-3)
species. The weak C-I bond dissociates below 200 K to yield
surface methyl (CH3) groups, which then further decompose
stepwise to finally give surface C atoms.15,16 The coadsorbed
iodine does not strongly influence the surface chemistry of
hydrocarbons except for a site-blocking effect at higher cover-
ages, as shown by comparing CH3 on Pt(111) deposited from
the gas phase with CH3 produced from the dissociation of
CH3I.17,18 Similarly, we find that the presence of coadsorbed
iodine has minimal influence on the C-N coupling reaction.
Ammonia does not dissociate on Pt(111) at low pressures
and temperatures. It desorbs molecularly in two stages: a lower
binding energy state â (second layer) that is populated at higher
coverages and desorbs at 120-180 K and an R state representing
molecules directly bonded to the platinum surface and desorbing
over the range of 200-450 K.13,14 However, it has been reported
that NH3 can decompose on Pt surfaces under conditions other
than UHV and low temperature. Vajo and co-workers19 found
ready decomposition under a steady flux at a pressure of ∼1 ×
10-6 Torr and temperatures of 300-600 K. On the other hand,
adsorbed NH3 on Pt(111) is readily dissociated through electron
irradiation to form N, NH, and NH2 species on the surface.20
Our present apparatus is arranged such that with the sample in
the IR position it can be exposed to an electron beam from our
LEED optics, which can be defocused to cover a large area. In
this way we were able to induce NH3 decomposition and
generate surface NHy (y ) 0-2) species and study the reaction
of these species with CHx (x ) 0-3) to form CN. We reported
our basic observation of C-N coupling in this system in a brief
earlier publication.21 Here we provide additional details and
insights into the reaction.
Figure 1. (a, b, d, and f) TPD spectra for m/e ) 27 (HCN), 28 (N2/
CO), 16 (CH4), and 2 (H2), respectively, obtained after exposing Pt-
(111) to 0.2 L of CH3I and 0.2 L of NH3 at 85 K followed by 100 eV
e- beam bombardment. (c and g) TPD spectra for m/e ) 28 (N2/CO)
and 2 (H2) obtained after 0.2 L of NH3 has been e- beam irradiated at
85 K. (e and h) TPD spectra for m/e ) 16 (CH4) and 2 (H2) obtained
after 0.2 L of CH3I has been e- beam irradiated at 85 K.
reference spectrum was also taken at 85 K. A series of TPD
experiments have been performed to correlate the exposures and
establish the reproducibility between the two chambers and
establish if the same surface conditions prevailed during the
TPD, RAIRS, and XPS experiments. The Pt(111) surface was
cleaned and judged free of impurities by a standard procedure
described earlier.10 The methyl iodide (CH3I) was purchased
from Alfa Aesar Co. with a quoted purity of 99.5% and further
purified by transferring it to a glass bulb and subjecting it to
five freeze-pump-thaw cycles using a liquid nitrogen bath. It
was then placed in a dry-ice-acetone bath and condensed into
another glass bulb cooled by liquid nitrogen. Finally, it was
shielded from the light because of its tendency to photolyze.
NH3, H2, and C2H4 were purchased from Matheson Tri-Gas Inc.
with quoted purities of 99.9992%, 99.9999%, and 99.99%,
respectively, and used without further purification. A series of
TPD experiments have been performed to correlate exposures
with absolute coverages. Exposures of ∼0.2, 2.5, and 2.0 L are
necessary to saturate the first layer of NH3, CH3I, and C2H4,
respectively. These coverages are correlated with the absolute
coverages by comparison with data available in the litera-
ture.14,15,24 We used electron exposures of ∼2 × 1015 electrons
per cm2 in all experiments reported here.
Experimental Section
The results were obtained in two separate UHV chambers.
The XPS results were obtained in a chamber (chamber 1) with
a base pressure of ∼2 × 10-10 Torr. The system has been
described in detail elsewhere.22 In brief, the UHV chamber is
equipped with low-energy electron diffraction (LEED), an X-ray
photoelectron spectroscopy (XPS) system, a quadrupole mass
spectrometer (QMS) for TPD, and a Fourier transform infrared
(FTIR) spectrometer for RAIRS studies. The XPS system
consists of a VG CLAM2 hemispherical analyzer and X-ray
source. Mg KR radiation was used for the XPS study, and the
spectrometer was calibrated with the Pt 4f7/2 peak at a binding
energy of 71.2 eV. All TPD and RAIRS experiments were
performed in a second chamber (chamber 2) with a base pressure
of ∼1 × 10-10 Torr. A detailed description of this system can
be found elsewhere.23 In brief, it consists of a stainless steel
UHV chamber equipped for Auger electron spectroscopy (AES),
LEED, and TPD experiments with a QMS. It is also coupled to
a commercial FTIR spectrometer (Bruker IFS 66v/S). The IR
beam enters and exits the UHV chamber through differentially
pumped O-ring sealed KBr windows and passes through a
polarizer before reaching an IR detector. To achieve maximum
sensitivity in the NH and CH stretch regions, all RAIR spectra
reported here were obtained with a liquid nitrogen cooled InSb
detector with a low wavenumber cutoff near 1950 cm-1 and a
tungsten IR source. In cases where the sample was annealed to
a temperature above 85 K, the sample was then cooled back to
85 K before the IR spectrum was acquired. The background
Results
(1) Detection of CN Bond Formation. (i) Temperature-
Programmed Desorption. Figure 1 shows TPD traces of the most
relevant desorption products in the course of CN bond formation
on Pt(111) from CH3 and NH3 fragments. The fact that the CN
group can be formed under UHV on Pt(111) is demonstrated
by the observation of HCN desorption in Figure 1a. After
exposing the surface to 0.2 L of CH3I and 0.2 L of NH3 at 85
K followed by bombardment by 100 eV electrons, HCN (m/e
) 27) is observed at 497 K with a high-temperature shoulder
at 587 K. Following the direct adsorption of HCN on Pt(111),
HCN desorbs at a lower temperature of ∼460 K.21,25 Various
control experiments were performed that proved that the HCN

17562 J. Phys. Chem. B, Vol. 109, No. 37, 2005
Herceg and Trenary
is produced only when NH3, CH3I, and electron irradiation are
used.21
For the same conditions that lead to desorption of HCN
(Figure 1a), Figure 1 also shows results for CO/N2 (m/e ) 28)
(Figure 1b), CH4 (m/e ) 16) (Figure 1d), and H2 (m/e ) 2)
(Figure 1f). These results are compared with the cases of
irradiation with 100 eV electrons of surfaces on which only
NH3 (Figure 1c and 1g) or only CH3I (Figure 1e and 1h) is
adsorbed. Although in principle the m/e ) 28 signals in Figure
1c and 1b could be due to either desorption of N2 or CO
adsorbed from the background, we established that the peaks
at 443 K in b and c are due to CO with the higher temperature
features due to N2. This is based on the fact that CO is known
to desorb from the clean surface at ∼440 K26 and by comparison
of the m/e ) 28 traces in Figure 1b and 1c with the m/e ) 12
and 14 signals (not shown), which are predominantly due to
CO and N2, respectively. The decrease in the N2 desorption at
∼500 K in Figure 1b (CH3I + NH3 + e-) compared to Figure
1c (NH3 + e-) is likely due to consumption of surface N atoms
by the CN bond formation reaction.
Figure 2. RAIR spectra obtained after exposing 0.2 L of CH3I and
0.2 L of NH3 to the sample at 85 K followed by e- beam irradiation
with subsequent heating to the indicated temperatures and exposure to
10 L of H2 at 300 K.
Desorption of H2 corresponding to the conditions for HCN
formation in Figure 1a is shown in Figure 1f. The H2 peak
observed at 457 K is attributed to the decomposition of surface
NH and/or CH and is seen at essentially the same temperature
of 452 K for electron irradiation of NH3 adsorbed alone (Figure
1g) and at 464 K for electron irradiation of CH3I alone (Figure
1h). The lower temperature H2 desorption peaks in Figure 1f,
1g, and 1h correspond to H2 recombinative desorption from
hydrogen adsorbed from the background and from hydrogen
produced through the dehydrogenation of NH3 and CH3.27,28
The lower amount of H2 desorption at 457 K in Figure 1f
compared with the peaks at 452 and 464 K in Figure 1g and 1h
is attributed to loss of hydrogen through HCN desorption.
Figure 3. N 1s XPS peak obtained after (a) exposing 2.0 L of HCN
to the sample at 85 K, followed by annealing to 500 K, (b) annealing
to 500 K a surface that had been exposed to 0.2 L of CH3I, 0.2 L of
NH3, and 100 eV electrons at 85 K, and (c) annealing to 500 K a surface
that had been exposed to 0.2 L of NH3 and 100 eV electrons at 85 K.
Methane (m/e ) 16) desorption is not observed (Figure 1e)
following a 0.2 L CH3I exposure at 85 K and electron irradiation
unless NH3 is present (Figure 1d). Comparison with the thermal
decomposition15,29 of CH3I shows that exposure to the electron
beam does not exert a strong influence on the surface chemistry
of this molecule. It therefore follows that electron irradiation is
needed only to induce NH3 dissociation. This is confirmed by
a series of TPD experiments (not shown here) that show little
or no difference in the m/e ) 27 trace when both NH3 and CH3I
are e- irradiated compared to when only NH3 is irradiated and
then CH3I adsorbed afterward. At the low initial CH3I coverages
corresponding to the 0.2 L exposures used in Figure 1,
dehydrogenation dominates the thermal chemistry when CH3I
is adsorbed alone and no CH4 desorption is observed, which is
in agreement with previous studies.29 However, the electron-
induced dissociation makes additional hydrogen available, so
that hydrogenation of the CH3 groups formed from the low-
temperature breaking of the C-I bond of CH3I can compete
with the CH3 dehydrogenation reaction, and methane desorption
is observed. This interpretation is supported by the presence of
a small valley at 268 K in the H2 desorption shown in Figure
1f.
with an NH symmetric stretch in the range of 3365-3370 cm-1
.
The 450 K anneal yields an intense peak at 3312 cm-1, which
grows in intensity and shifts to 3315 cm-1 after H2 exposure.
This peak is due to the NH species, and we described its
properties in detail elsewhere.27 There is no peak due to CNH2
for the 450 K anneal nor for any annealing temperatures below
450 K. The 500 K anneal eliminates the ν(NH) peak of the NH
species, which reappears at 3311 cm-1 after H2 exposure due
to rehydrogenation of surface N atoms. In addition, ν(NH) of
the CNH2 species now appears at 3370 cm-1, demonstrating
that the 500 K anneal leads to formation of CN on the surface.
This peak is still present after a 550 K anneal but has a lower
intensity.
(iii) X-ray Photoelectron Spectroscopy. In Figure 3 we
compare XPS N 1s binding energies for CN produced by HCN
dissociation (Figure 3a) with CN produced by the C-N coupling
reaction (Figure 3b). In both cases peaks are centered at 398.4
eV, a N 1s binding energy that is characteristic of the CN
group.30 Figure 3c is a control experiment showing the N 1s
binding energy of N atoms adsorbed on Pt(111). The N-covered
surface was produced by a 0.2 L NH3 exposure at 85 K,
irradiation with 100 eV electrons to dissociate the ammonia,
and annealing to 500 K to dissociate any ammonia dissociation
fragments and desorb hydrogen. The surface was then cooled
back to 85 K, where the spectrum was obtained. The results
provide further proof of the formation of CN from the surface
C1 and N1 species present. Although the signal-to-noise ratio is
(ii) Reflection Absorption Infrared Spectroscopy. The use of
RAIRS to detect the C-N coupling reaction is demonstrated
in Figure 2. After exposing the Pt(111) surface to 0.2 L of CH3I
and 0.2 L NH3 at 85 K, it was irradiated with 100 eV electrons
and annealed to 450, 500, and 550 K. After each anneal the
sample was cooled to 300 K, exposed to 10 L of H2, and then
cooled to 85 K, where the spectra were acquired.
As previous work has shown,10,11 if there is CN on the Pt-
(111) surface, then H2 exposure at 300 K should produce CNH2

Surface Chemistry of CN Bond Formation
J. Phys. Chem. B, Vol. 109, No. 37, 2005 17563
Figure 4. (a) HCN yield as a function of CH3I coverage for a fixed initial NH3 coverage of ∼0.5 ML. (b) HCN yield as a function of NH3
coverage for a fixed initial CH3I coverage of ∼0.1 ML. The HCN yield is equal to the area below the TPD curve for m/e ) 27.
Figure 5. (a) TPD spectra for m/e ) 2 (H2) and 27 (HCN) obtained after exposing Pt(111) to 0.1 L of C2H4 and 0.2 L of NH3 at 85 K followed
by 100 eV e- beam bombardment. (b) RAIR spectra obtained after exposing 0.1 L of C2H4 and 0.2 L of NH3 to the sample at 85 K followed by
e- beam irradiation with subsequent heating to the indicated temperatures and exposure to 10 L of H2 at 300 K.
low, comparison of the N 1s shape in the three spectra suggests
a shoulder on the low binding energy side of the peak in Figure
3b, consistent with the presence of both CN molecules and N
atoms, as indicated by the RAIRS results.
(3) C-N Coupling in the Absence of Iodine. Although CH3I
is a convenient source of C1 hydrocarbons, it is desirable to
verify the benign role of coadsorbed iodine by using a different
source of carbon in the C-N coupling reaction. In a separate
study of submonolayer coverages of C2H4 on Pt(111), we have
shown that C-C bond scission is feasible at temperatures above
450 K.31 Therefore, surface carbon atoms can be produced from
ethylene decomposition well below the CN bond formation
temperature. After the surface was exposed to 0.1 L of C2H4
and 0.2 L of NH3 at 85 K and then exposed to the electron
beam, HCN (m/e ) 27) is observed to desorb with a peak
temperature of 516 K in Figure 5a, demonstrating that C-N
coupling occurs under these circumstances. The fact that the
HCN TPD peak in Figure 5a is 19 K above what was seen in
Figure 1 when CH3I was used as the source of carbon atoms
may be due to the HCN desorption being limited by C-C bond
scission.
In addition to TPD, RAIRS also reveals the presence of the
CN coupling reaction when ethylene is used as the carbon
source. The spectra in Figure 5b were obtained after the surface
was first exposed to 0.1 L of C2H4 and 0.2 L of NH3 at 85 K
and then exposed to the electron beam. Annealing to 510 K
yields only one peak at 2964 cm-1, which is assigned to ν(CH)
of the CH species. Further details on the formation and
properties of CH formed from C2H4 can be found elsewhere.31
Upon hydrogen exposure at 300 K two new peaks are observed
with the same assignments as for Figure 2d. The 3370 cm-1
peak of CNH2 demonstrates that the CN bond is formed,
although the amount appears to be less based on the lower peak
intensity than that observed for CH3I, even though the overall
carbon coverage was the same in the two cases. Figure 6 shows
the HCN yield as a function of initial ethylene coverage for an
initial ammonia coverage of 0.5 ML. A similar dependence of
(2) Coverage Dependence. Temperature-Programmed De-
sorption. Figure 4 shows the dependence of HCN yield from a
series of TPD experiments where we varied the CH3I coverage
for a fixed initial NH3 coverage and vice versa. The HCN yield
is defined as the area of the m/e ) 27 TPD peak. A monolayer
(ML) is defined here as the saturated first layer of NH3 (R-
NH3) or CH3I. By this definition, 1 ML of ammonia corresponds
to one NH3 molecule per four surface Pt atoms14 while for
methyl iodide it corresponds to one CH3I molecule per five
surface atoms.15
The initial NH3 coverage of approximately 0.5 ML was
achieved by a 0.1 L exposure of NH3 at 85 K. The surface was
then bombarded with 100 eV electrons and exposed to different
amounts of CH3I. The HCN yield reaches a maximum for 0.10-
0.15 ML of CH3I, after which it drops significantly, implying
that either CH3I itself or one of it dissociation products causes
the quenching of the C-N coupling reaction. Figure 4b shows
the HCN yield vs NH3 coverage for a fixed initial CH3I coverage
of ∼0.1 ML for an experiment in which 0.2 L of CH3I was
exposed at 85 K followed by NH3 exposure and e- beam
irradiation. The HCN yield increases up to 0.2 ML of NH3,
after which it remains essentially constant up to 1.0 ML. Up to
1.0 ML the HCN yield behavior is consistent with the reaction
of all of the surface carbon by nitrogen with the plateau in HCN
yield reached when all of the carbon has been consumed. The
decrease in the amount of HCN for NH3 coverages above 1
ML can be attributed to the build-up of second- (â-NH3) and
higher order layers13,14 of NH3 that shield the R state of NH3
from the electrons, thus preventing dissociation.

17564 J. Phys. Chem. B, Vol. 109, No. 37, 2005
Herceg and Trenary
Figure 9. (a) RAIR spectrum acquired at 85 K after exposing the
sample to 0.2 L of CH3I at 85 K, annealing to 800 K for 60 s, cooling
back to 85 K, exposing to 0.2 L of NH3, e- beam irradiating, and
annealing to 500 K. (b) RAIR spectrum acquired at 85 K after exposing
10 L of H2 at 300 K.
Figure 6. HCN yield from the integrated area of the m/e ) 27 TPD
peak as a function of C2H4 coverage for a fixed initial NH3 coverage
of ∼0.5 ML. A ML of C2H4 corresponds to one C2H4 molecule per
four Pt atoms.24
react with nitrogen to form cyanide. The surface was prepared
by exposing to 0.2 L of CH3I, annealing to 800 K, cooling to
85 K, exposing to 0.2 L of NH3, followed by electron irradiation.
The m/e ) 2 TPD trace in Figure 8a has a shape very similar
to the one in Figure 1g. This is expected since annealing to
800 K decomposes all CH3 groups to surface C,15,16,29 and in
both Figures 8a and 1g hydrogen desorption is mainly due to
the electron-induced decomposition of NH3. However, no HCN
desorption is observed (Figure 8b), indicating that surface CN,
if present, is not detectable by this method. In contrast, the
RAIRS data in Figure 9 do reveal that the C-N coupling
reaction has taken place. The surface was prepared in the same
way as for Figure 8. No peaks are present in the spectrum after
the 500 K anneal. Upon exposure to 10 L of H2 at 300 K (Figure
9b), two peaks are observed with the same assignments as for
Figure 2d. Although development of the CNH2 peak at 3370
cm-1 confirms the presence of CN, the intensity of the peak is
about one-half of that in Figure 2d. This is most likely due to
conversion of some of the carbon into an unreactive graphitic
form upon annealing to 800 K.15,33
Figure 7. (a) I 3d XPS spectrum obtained after (a) exposing 0.2 L of
CH3I to the sample at 85 K, (b) annealing the sample in part a to 500
K, and (c) annealing the sample in part b to 800 K.
Discussion
According to model-reactor studies by Schmidt and Hick-
man,34 the highly endothermic formation of HCN from CH4
and NH3 should not occur at all at temperatures below ∼1000
K. However, desorption of HCN at 497 K in Figure 1a is
consistent with CN bond formation at or even below this
temperature. We have shown earlier that the high-temperature
shoulder at 587 K is not specific to the C-N coupling reaction
but rather is related to HCN desorption from the step sites.21
Moreover, RAIRS data in Figure 2e and 2f further confirm that
there is no significant C-N coupling at temperatures above 500
K. This is probably due to the fact that above 500 K there are
fewer N atoms available on the surface, as confirmed by the
N2 desorption trace in Figure 1c. In addition, surface C atoms
gain higher mobility, which could lead to formation of unre-
active graphitic islands. The appearance of a CNH2 peak at 3370
cm-1 in Figure 2d, after hydrogenation of the surface at 300 K
following the 500 K anneal, reveals that CN bond formation
likely coincides with the HCN desorption peak centered at 497
K (Figure 1a).
Figure 8. TPD results for m/e ) 2 (H2) (a) and 27 (HCN) (b) after
the surface had been exposed to 0.2 L of CH3I at 85 K, annealed to
800 K for 60 s, cooled to 85 K, exposed to 0.2 L of NH3, and then
exposed to 100 eV electrons.
HCN yield on C2H4 coverage is observed as in Figure 4a,
indicating that a low carbon coverage is most favorable for HCN
production.
Another way to study the reaction in the absence of iodine is
to perform the experiment after annealing the CH3I-covered Pt-
(111) surface to 800 K for 60 s, which previous studies have
suggested desorbs the iodine.15,16 The XPS spectra of the I 3d
region as a function of annealing temperature shown in Figure
7 verify this. The I 3d5/2 and 3d3/2 peaks at 621.6 and 634.0
eV, respectively, for 0.2 L of CH3I adsorbed at 85 K shift to
620.9 and 632.6 eV but do not change intensity after a 500 K
anneal, reflecting the fact that although the molecule dissociates
at around 200 K, the iodine is still present at 500 K. These
values in Figure 7b are typical of atomic iodine on surfaces.15,32
However, the iodine peaks disappear completely after a 800 K
anneal.
The appearance of the aminocarbyne peak at 3370 cm-1 in
Figure 9b for the case when CH3 is completely dehydrogenated
to C before the experiment was performed indicates that a
hydrogenated form of carbon is not needed for the C-N
coupling reaction to occur. The TPD results in Figure 1c and
1g showing N2 and H2 desorption peaks at 500 and 452 K,
respectively, together with the disappearance of the NH stretch
The TPD results in Figure 8 were obtained to determine if
the carbon remaining on the surface after iodine desorption could

Surface Chemistry of CN Bond Formation
J. Phys. Chem. B, Vol. 109, No. 37, 2005 17565
coverage that completely poisons the reaction for C2H4 is about
one-half of that of CH3I. This strongly suggests that surface
carbon is responsible for deactivation of the surface, in
agreement with the kinetic studies of Hasenberg and Schmidt.5
They inferred that every C atom blocks three sites on their
polycrystalline surface, and if we assume that this number is
approximately the same for Pt(111), then the free sites on the
surface necessary for C-N bond formation to occur could be
completely blocked at higher carbon coverages. Furthermore,
graphitic islands are much more easily formed at higher than
lower carbon coverages. It then follows that at higher carbon
coverages N atoms are able to react readily only with C atoms
outside of ordered islands or at island edges, which would
explain the inability to convert all of the carbon present on the
surface to CN.
Comparison of Figures 4a and 6 shows that the maximum
HCN yield is higher by a factor of ∼3 for ethylene than for
CH3I. However, RAIRS experiments (Figures 2d and 5b) show
that the intensity of the aminocarbyne ν(NH) peak in the case
of C2H4 is actually smaller than that obtained from CH3I,
suggesting that less CN is formed from C and N atoms in the
former case. This seemingly contradictory observation can be
readily explained by the fact that more hydrogen is available at
500 K when C2H4 is used as a source of C atoms. Previous
studies35 of ethylene decomposition on Pt(111) have shown that
hydrogen desorption can be easily observed up to ∼600 K. This
also can be seen in Figure 5a for the m/e ) 2 trace. In separate
experiments of hydrogen desorption we determined that the
amount of hydrogen evolving from the decomposition of C2H4
at 500 K is much higher than that for the CH3I. It follows then
that most of the CN molecules formed from C2H4 are im-
mediately hydrogenated and desorb as HCN. However, in the
case of CH3I, even if more C-N bonds are formed, not all of
the CN formed can be hydrogenated because of the lack of
hydrogen atoms on the surface.
Figure 10. Proposed reaction mechanism of CN bond formation from
CH3I and NH3 on Pt(111).
peak at 3315 cm-1 in the RAIR spectrum of Figure 2c after
annealing to 500 K indicate that by 500 K only N atoms remain
on the surface following electron-induced dissociation of NH3.
It therefore follows that the nitrogen-containing species in the
C-N coupling reaction are adsorbed N atoms. These conclu-
sions are summarized in Figure 10. Once formed at around 500
K, the surface CN can be hydrogenated to CNH2 if the surface
is cooled to 300 K. The thermal dissociation of CNH2 occurs
between 400 and 500 K and leads to desorption of HCN in the
range of 450-600 K. Since dehydrogenation of the NH species
is complete well before 500 K, the source of the hydrogen for
the HCN desorption must originate from the dissociation
products of CH3I. We have direct RAIRS evidence31 that the
CH species is stable up to 500 K, and in Figure 1h the
dissociation of CH gives rise to the H2 desorption peak at 464
K, which has a high-temperature tail that extends to about 550
K. Therefore, we conclude that the hydrogen in the HCN
desorption observed in Figure 1 originates from the dissociation
of CH.
From kinetic measurements of the gas-phase products in
model reactor studies of the catalytic synthesis of HCN from
CH4 and NH3 it was inferred that the reaction is feasible when
the surface contains less than a monolayer of carbon and that it
is strongly inhibited and even totally quenched by the presence
of a carbon multilayer.5 Our experiments provide more direct
information on the influence of carbon coverage on HCN yield.
At the simplest level we would expect the amount of CN formed
on the surface to be limited by whichever is lower, the N
coverage or the C coverage. If the initial coverage of NH3 is
one molecule per eight surface atoms, then the amount of HCN
formed should increase steadily up to a coverage of one
molecule of CH3I per eight Pt atoms. However, this is not the
case, as shown in Figure 4a. Initially, the HCN yield increases,
but then it suddenly drops after ∼0.1 ML of CH3I. This coverage
corresponds to 1 molecule of CH3I for every 50 surface atoms,
which is below what is implied by the reaction stoichiometry
by a factor of ∼6. At 500 K C and I atoms from CH3I
dissociation are present on the surface and, as already shown
in Figure 9, the coupling reaction occurs even in the absence
of Iads. In other words, surface iodine is not directly involved
in the coupling reaction, and therefore, it is highly unlikely that
Iads is the CH3I dissociation product that is responsible for the
quenching, especially in the low coverage regime (1 Iads/50 Pt
atoms). Additional evidence can be found in Figure 6, where a
molecule that does not contain iodine, C2H4, showed similar
coverage dependence for HCN yield. Not surprisingly, the
Finally, there are various ways to extract kinetic parameters
from thermal desorption data, provided certain assumptions are
valid. Hagans et al.36 reported HCN TPD results on Pt(111)
and found that the HCN desorbs with a constant peak temper-
ature at ∼460 K as a function of HCN exposure. They labeled
the desorption at this temperature as the â state and attributed
it to recombination of Had and CNads. Although recombinative
desorption should lead to second-order kinetics, desorption at
a constant peak temperature is a clear indication of first-order
kinetics. We also observed HCN desorption at a constant
temperature in the range of 460-500 K as a function of
increasing HCN exposure to the clean Pt(111) surface and at
various CH3I and NH3 exposures in the case of the C-N
coupling reaction. If the rate of the â HCN desorption is
determined by the rate of aminocarbyne decomposition, then
first-order desorption kinetics would be expected. Since RAIRS
observations of the stability of aminocarbyne show that it
decomposes in the temperature range of 400-500 K, it is
reasonable to assume that aminocarbyne decomposition is rate
determining for HCN desorption in this temperature range.
For first-order desorption with a peak temperature of 470 K
and an assumed preexponential factor of 1013 s-1, the Redhead
equation37 gives an activation energy of 128 kJ/mol, whereas
assumed preexponentials of 1015 and 1011 s-1 yield activation
energies of 146 and 109 kJ/mol, respectively. A method
described by Chan, Aris, and Weinberg (CAW)38 permits both
the preexponential and activation energy to be obtained from
measurements of the peak temperature and peak width. By the
CAW method we obtain a preexponential of 5 × 1013 s-1 and

17566 J. Phys. Chem. B, Vol. 109, No. 37, 2005
Herceg and Trenary
III; Schmidt, L. D. Appl. Catal. A: Gen. 2000, 180, 287. Bodke, A. S.;
Olschki, D. A.; Schmidt, L. D. Appl. Catal. A: Gen 2000, 201, 13.
(4) Hasenberg, D.; Schmidt, L. D. J. Catal. 1985, 91, 116.
(5) Hasenberg, D.; Schmidt, L. D. J. Catal. 1986, 97, 156.
(6) Hasenberg, D.; Schmidt, L. D. J. Catal. 1987, 104, 441.
(7) van Hardeveld, R. M.; Schmidt, A. J. G. W.; Niemantsverdriet, J.
W. Catal. Lett. 1996, 41, 125. van Hardeveld, R. M.; van Santen, R. A.;
Niemantsverdriet, J. W. J. Phys. Chem. B 1997, 101, 7901. van Hardeveld,
R. M.; Borg, H. J.; Niemantsverdriet, J. W. J. Mol. Catal. A: Chem. 1998,
131, 199.
(8) Mullins, D. R.; Zhang, K. J. Phys. Chem. B 2001, 105, 1374.
(9) Hoffmann, W.; Bertel, E.; Netzer, F. P. J. Catal. 1979, 60, 316.
(10) Jentz, D.; Celio, H.; Mills, P.; Trenary, M. Surf. Sci. 1995, 341, 1.
(11) Jentz, D.; Mills, P.; Celio, H.; Trenary, M. Surf. Sci. 1996, 368,
354.
(12) Ukraintsev, V. A.; Harrison, I. Surf. Sci. 1993, 286, L571. Meixner,
D. L.; George, S. M. Surf. Sci. 1993, 297, 27. Alberas-Sloan, D. J.; White,
J. M. Surf. Sci. 1996, 365, 212.
(13) Sexton, B. A.; Mitchell, G. E. Surf. Sci. 1980, 99, 523.
(14) Fisher, G. B. Chem. Phys. Lett. 1981, 79, 452.
(15) Zaera, F.; Hoffmann, H. J. Phys. Chem. 1991, 95, 6297.
(16) Henderson, M. A.; Mitchell, G. E.; White, J. M. Surf. Sci. Lett.
1987, 184, L325.
an activation energy of 128 kJ/mol. Since the HCN desorption
that accompanies the C-N coupling reaction reported here
occurs at a higher temperature than for HCN desorption
following HCN exposure, it is reasonable to assume that the
rate of HCN desorption here is limited by the rate of CN
formation. With this assumption the rate constant for CN
formation can be obtained. The CAW method applied to the
HCN desorption peak at 497 K in Figure 1 yields a preexpo-
nential of 8 × 1010 s-1 and an activation energy of 113 kJ/mol,
whereas the Redhead equation with an assumed preexponential
of 1013 s-1 yields an activation energy of 132 kJ/mol. A review39
of the effect of lateral interactions on kinetic parameters
extracted from TPD data notes that both the Redhead and CAW
methods, although simple and easy to use, are invalid in the
presence of lateral interactions except in the limit of zero
coverage. For HCN desorption from Pt(111), the invariance of
the peak position with coverage suggests that lateral interactions
are small. The kinetic parameters obtained here from TPD data
can be compared with the preexponential of 1011(1 s-1 and
activation energy of 210 ( 15 kJ/mol for the C-N coupling
reaction on Rh(111) obtained from temperature-programmed
SIMS data.7 The higher activation energy is associated with the
higher HCN desorption temperature of ∼600 K on Rh(111).
(17) Fairbrother, D. H.; Peng, X. D.; Viswanathan, R.; Stair, P. C.; Fan,
J.; Trenary, M. Surf. Sci. 1993, 285, L455.
(18) Fan, J.; Trenary, M. Langmuir 1994, 10, 3649.
(19) Vajo, J. J.; Tsai, W.; Weinberg, W. H. J. Phys. Chem. 1986, 90,
6531.
(20) Burns, A. R.; Stechel, E. B.; Jennison, D. R.; Li, Y. S. J. Chem.
Phys. 1994, 101, 6318. Sun, Y.-M.; Sloan, D.; Ihm, H.; White, J. M. J.
Vac. Sci. Technol. A 1996, 14, 1516. Bater, C.; Campbell, J. H.; Craig, J.
H., Jr. Surf. Interface Anal. 1998, 26, 97.
Summary
(21) Herceg, E.; Trenary, M. J. Am. Chem. Soc. 2003, 125, 15758.
(22) Kang, D. H.; Trenary, M Surf. Sci. 2000, 470, L13.
(23) Brubaker, M. E.; Trenary, M. J. Chem. Phys. 1986, 85, 6100.
(24) Mitchell, I. V.; Lennard, W. N.; Griffiths, K.; Massoumi, G. R.;
Huppertz, J. W. Surf. Sci. 1991, 256 (1-2), L598.
(25) Celio, H.; Mills, P.; Jentz, D.; Pae, Y. I.; Trenary, M. Langmuir
1998, 14, 1379.
(26) Norton, P. R.; Goodale, J. W.; Selkirk, E. B. Surf. Sci. 1979, 83,
189.
(27) Herceg, E.; Mudiyanselage, K.; Trenary, M. J. Phys. Chem. B 2005,
109, 2828.
We have demonstrated that the carbon-nitrogen coupling
reaction that underlies the catalytic synthesis of HCN from
ammonia and methane over platinum can be induced on the
Pt(111) surface under UHV conditions. The experiments show
that the carbon-containing and nitrogen-containing species
involved in the coupling reaction are individual C and N atoms
rather than hydrogen-containing CHx or NHy species. The
formation of the CN species was detected both by desorption
of HCN and by observation of the aminocarbyne species, CNH2,
with RAIRS following hydrogen exposure. The experiments
further show that C-N bond formation occurs at ∼500 K.
Electron-induced dissociation of ammonia was used to produce
surface N atoms, whereas both CH3I and C2H4 were used to
produce surface C atoms. Formation of surface CN is strongly
dependent on carbon coverage, in agreement with previous
kinetic studies.
(28) Herceg, E.; Celio, H.; Trenary, M. ReV. Sci. Instrum. 2004, 75 (8),
2545.
(29) Henderson, M. A.; Mitchell, G. E.; White, J. M. Surf. Sci. 1991,
248, 279.
(30) Sexton, B. A.; Hughes, A. E. Surf. Sci. 1984, 140, 227. Lindquist,
J. M.; Ziegler, J. P.; Hemminger, J. C. Surf. Sci. 1989, 210, 27. Serafin, J.
G.; Friend, C. M. J. Phys. Chem. 1988, 92, 6694.
(31) Deng, R.; Herceg, E.; Trenary, M Surf. Sci. 2004, 573 (2), 310.
(32) Zaera, F. Surf. Sci. 1989, 219, 453.
(33) Land, T. A.; Michely, T.; Behm, R. J.; Hemminger, J. C.; Comsa,
G. J. Chem. Phys. 1992, 97 (9), 6774. Smirnov, M. Yu.; Gorodetskii, V.
V.; Cholach, A. R.; Zemlyanov, D. Yu. Surf. Sci. 1994, 311, 308.
(34) Schmidt, L. D.; Hickman, D. A. In Catalysis of Organic Reactions;
Kosak, J. R., Johnson, T. A., Eds.; Marcel Dekker: New York, 1994; p
195.
Acknowledgment. We gratefully acknowledge support by
a grant from the National Science Foundation (CHE-0135561).
(35) Salmero´n, M.; Somorjai, G. A. J. Phys. Chem. 1982, 86, 341.
(36) Hagans, P. L.; Guo, X.; Chorkendorff, I.; Winkler, A.; Siddiqui,
H.; Yates, J. T., Jr. Surf. Sci. 1988, 203, 1.
References and Notes
(1) Satterfield, C. N. In Heterogeneous Catalysis in Practice; McGraw-
Hill: New York, 1980.
(37) Redhead, P. A. Vacuum 1967, 12, 203.
(2) Satterfield, C. N. In Heterogeneous Catalysis in Industrial Practice,
2nd ed.; McGraw-Hill: New York, 1991.
(3) Koberstein, E. Ind. Eng. Chem. Process Des. DeV. 1973, 12, 444-
448. Sua`rez, M. P.; Lo¨ffler, D. G. J. Catal. 1986, 97, 240. Dietz, A. G.,
(38) Chan, C.-M.; Aris, R.; Weinberg, W. H. Appl. Surf. Sci. 1978, 1,
360.
(39) Nieskens, D. L. S.; van Bavel, A. P.; Niemantsverdriet, J. W. Surf.
Sci. 2003, 546, 159.