The formation and hydrolysis of isocyanic acid during the reaction of NO, CO, and H2 mixtures on supported platinum, palladium, and rhodium

Article abstract of DOI:10.1006/jcat.2001.3359

The extent to which isocyanic acid (HNCO) is formed during the reaction of NO/CO/H₂ mixtures over silica-supported Pt, Rh, and Pd was studied with the subsequent hydrolysis of HNCO on oxide systems placed downstream. HNCO formation was a characteristic feature of the NO + CO + H₂ reaction over silica-supported Pt, Rh, and Pd. Platinum produced the largest quantity in two stages, i.e., from H₂ and then using NH₃ being formed as a coproduct. With Pd, HNCO arose largely from NH₃ alone because H₂ was totally removed by reaction with NO at low temperature. Rhodium gave rise to the least HNCO. Formation was confined to a narrow temperature area due to the coincident consumption of H₂ and NO, which precluded NH₃ reaction with CO and NO. Hydrolysis of HNCO to NH₃ and CO₂ was appreciable on SiO₂ alone and faster when a metal was present. Other oxide systems gave complete hydrolysis to the limit of the water present and total reaction with even small excesses of water. The possible presence of HNCO in vehicle exhaust was not an issue since the presence of a vast excess of steam and an active washcoat in three-way converters would ensure complete hydrolysis. However, the latter process might contribute to ammonia emissions at moderate temperatures under conditions where CO is still present.

Full text of DOI:10.1006/jcat.2001.3359

Journal of Catalysis 204, 11–22 (2001)
doi:10.1006/jcat.2001.3359, available online at http://www.idealibrary.com on
The Formation and Hydrolysis of Isocyanic Acid during the
Reaction of NO, CO, and H2 Mixtures on Supported
Platinum, Palladium, and Rhodium
Dean C. Chambers, Dennys E. Angove,1 and Noel W. Cant2
Department of Chemistry, Macquarie University, Sydney, New South Wales 2109, Australia
Received December 6, 2000; revised July 20, 2001; accepted July 20, 2001
pores of three-way vehicle catalysts during warm-up, the rapidity of
hydrolysis on oxide washcoats with water in large excess is so great
that no HNCO would ever emerge. Only the hydrolysis products,
The extent to which isocyanic acid (HNCO) is formed during
the reaction of NO/CO/H2 mixtures over silica-supported Pt, Rh,
and Pd has been investigated together with the subsequent hydrol-
ysis of HNCO on oxide systems placed downstream. The yield of
HNCO from NO is highest for Pt/SiO2 exceeding 35% at 315 C
with a standard 2800/3400/1200 ppm NO/CO/H2 mixture. Hydro-
gen consumption is complete at 220 C with ammonia and water
as major products, but above 260 C HNCO becomes the favoured
product with some arising from NH3. Hydrolysis of HNCO to
NH3 and CO2 takes over once all NO has been consumed. The
extent of hydrolysis is increased somewhat if additional SiO2 is
placed downstream. Other oxide systems—CeO2/SiO2, BaO/SiO2,
CeO2/Al2O3, and CeO2-ZrO2—give complete hydrolysis to the
extent of the available water at 315 C, and no HNCO remains if
additional water is included in the feed. Hydrogen consumption
during the NO + CO + H2 reaction commences at the lowest tem-
perature on the Pd/SiO2, and for 100 C from the onset tempera-
ture of 130 C the reaction can be largely described as the sum of
the NO + H2 and NO + CO ones. Formation of HNCO commences
at 235 C, with a maximum yield of 20% at 300 C. It appears to
arise solely through utilisation of NH3 made as a side-product to
the NO + H2 reaction. Rh/SiO2 is much less active than Pt/SiO2 and
Pd/SiO2 for the NO + H2 reaction, but more active for the NO + CO
and NO + CO + H2 ones. The latter exhibits a small sharp peak in
HNCO formation, but the maximum yield is only 14% and this co-
incides with total consumption of NO. Considerable ammonia is
formed at higher temperatures, even though none is produced dur-
ing the NO + H2 reaction. HNCO is believed to arise on each metal
through the combination of surface hydrogen atoms with metal-
bound NCO groups which exist in small numbers when N atoms
are located adjacent to adsorbed CO molecules. The differences in
behaviour between the metals can be rationalised in terms of the rel-
ative strengths of adsorption of CO and NO, and the temperature
difference between total consumption of H2 and NO. If the differ-
ence is large, then HNCO can be produced from ammonia as well
as hydrogen. A general conclusion is that, although some HNCO
might be generated on platinum group metal particles within the
c
2001 Academic Press
NH3 and CO2, will be seen.
Key Words: isocyanic acid; HNCO; Pt/SiO2; Rh/SiO2; Pd/SiO2;
hydrolysis; CeO2; BaO; CeO2–ZrO2.
INTRODUCTION
The reduction of NO by CO is the key reaction in con-
ventional three-way catalysts used for emission control on
motor vehicles (1). The reaction has been extensively stud-
ied, especially over rhodium, the most active platinum
group metal (PGM). The most recent surface science stud-
ies using rhodium single crystals demonstrate that N2O and
N2 are initially formed in parallel (2–5) by the reactions
2NO + CO → N2O + CO2
2NO + 2CO → N2 + 2CO2,
[1]
[2]
although there have been claims that N2O and N2 are pro-
duced sequentially on supported rhodium (6, 7). The gen-
erally accepted mechanism with Rh, Pt, and Pd is that NO
dissociates to form adsorbed N and O adatoms, the lat-
ter reacting with CO to form CO2 while the former either
dimerise to N2 or react with adsorbed NO to yield N2O (8).
For supported PGMs the reaction is accompanied by the
accumulation of isocyanate species (9–11), now known to
be largely on the support (12, 13). Metal-bound isocyanates
are also well established (14–17) and have been included in
some reaction schemes (18), but the general belief is that
they are too unstable to play a major role during steady-
state catalysis.
Real exhaust contains other reducing agents, notably H2
at approximately one-third the concentration of CO as set
by the water–gas shift equilibrium at high temperature.
Despite the importance of reductant mixtures, there have
been few detailed studies of such systems for supported
metals since the original comparative work of Kobylinski
and Taylor (19) using percent level concentrations. They
1 Present address: CSIRO Division of Energy Technology, P.O. Box 136,
North Ryde, NSW 1670, Australia.
2 To whom correspondence should be addressed. Fax: 61-2-9850-8313.
E-mail: noel.cant@mq.edu.au.
11
0021-9517/01 $35.00
c
Copyright
2001 by Academic Press
All rights of reproduction in any form reserved.

12
CHAMBERS, ANGOVE, AND CANT
showed that the NO + H2 reaction is faster than the ferred washcoat for three-way catalysts, and other oxides,
NO +CO one on alumina-supported Pt, Pd, and Rh, the dif- such as BaO, are of interest as the basis for the nitrate stor-
ference in activity being least for Rh. Activity for the age catalysts developed for use with intermittent lean burn
NO + CO + H2 reaction is intermediate between the in- engines (32).
dividual ones but closer to that of the NO + CO one due to
CO inhibition of hydrogen adsorption. Ammonia is formed
as an undesired product, especially with platinum and palla-
dium (20), which is partly responsible for the unsatisfactory
performance of Pt alone as a three-way catalyst.
EXPERIMENTAL
The catalytic reactions were carried out as outlined pre-
viously (28). Samples of catalyst (weighing 75 mg), and
Discussion of the reaction of NO/CO/H2 mixtures over downstream oxides (also 75 mg if present), were held be-
supported PGMs in emission control catalysts has generally tween quartz wool plugs in a 4-mm i.d. Pyrex tube in a
been on the basis that the NO +H2 and NO +CO reactions tube furnace. The test mixtures were made up by blend-
proceed competitively with the individual characteristics ing analysed mixtures (BOC Aust.) using electronic mass
observed for binary mixtures. However, work by Voorho- flow controllers (Brooks 5850TR). The outlet stream was
eve and Trimble (21) more than two decades ago showed periodically sampled by a micro gas chromatograph (MTI
that the reaction of a CO-rich ternary mixture over unsup- Instruments model M200), for analysis of H2 and N2, and
ported PGMs gives high yields of isocyanate compounds then flowed through heated 1/16-in. o.d. stainless-steel tub-
in trapped products, predominantly ammonium cyanate ing to a 16-cm pathlength infrared cell in a box maintained
(NH4NCO) with Pt and Rh but isocyanic acid (HNCO) at 90 C. The concentrations of CO, CO2, NO, N2O, NH3,
with Pd and Ir (21–25).
and H2O were obtained using the software program MALT
1
More recently we have demonstrated, through online (33) to process spectra, recorded as 64 scans at 0.25-cm
analysis using FTIR, that HNCO is a major gaseous product resolution on a Nicolet Magna 550 FTIR with an MCT de-
when NO, CO, and H2 are reacted together over Pt/SiO2 tector, against a set of synthetic spectra constructed from
(26–28). The mechanism inferred was that a small num- the HITRAN database (34). This procedure is in principle
ber of Pt-bound NCO groups existed in equilibrium with N an absolute method if absorbances of the target compounds
adatoms and adsorbed CO, and that these Pt–NCO groups are within appropriate bounds, as was confirmed by flow-
gave rise to HNCO by reaction with H adatoms. No HNCO ing single and multicomponent mixtures of known concen-
was seen when Al2O3 was placed downstream of Pt/SiO2 tration. The concentration of HNCO was calculated from
1
or with a Pt/Al2O3 catalyst. Nonetheless, combined in situ the absorbance of its band at 2284 cm as described pre-
FTIR/isotope studies using Pt/Al2O3 showed that support- viously (28), but with a slightly revised calibration factor
1
bound NCO groups were present and that approximately (ppm = 3470 absorbance at 2284 cm after correction
60% of all CO2 formed passed through these NCO species for CO2). The analyses for carbon and nitrogen-containing
when the feed contained some water (29). The conclusion compounds are believed accurate to better than 10% since
was that HNCO was a significant short-lived gas-phase in- the corresponding mass balances were seldom outside this
termediate, desorbing from Pt and then undergoing facile limit. Hydrogen and water analyses are somewhat less ac-
hydrolysis to CO2 and NH3 after adsorption on alumina. curate due to the low sensitivity of the gas chromatograph
Tests with more complex simulated exhaust mixtures over detector for H2 and hold-up of water on the catalyst and
Pt/Al2O3 and Pt/CeO2/Al2O3 also indicate that the ammo- system walls.
nia formed in such systems during light-off under near sto-
ichiometric conditions is derived by that route (30).
The Pt/SiO2 and Pd/SiO2 catalysts were from the se-
ries prepared and characterised in detail by Uchijima et al.
The purpose of the present work was twofold. First, we (35) and Pitchai et al. (36). The Pt/SiO2 (1.1 wt% Pt,
sought to establish if isocyanic acid is formed during the 40% dispersion, designated 40–SiO2–PtCl–L in Ref. (35))
NO + CO + H2 reaction over silica-supported Pd and Rh was prepared by impregnation of silica gel (Davison
in the same way as with Pt/SiO2, given that the alumina- grade 62, 285 m2/g) with chloroplatinic acid and Pd/SiO2
supported metals differ considerably in their tendency to (0.49 wt% Pd, 79% dispersion, batch 79.1-Pd/SiO2–IV in
make ammonia (19). The only previous study here is that Ref. (36)) by ion exchange with Pd(NH3)4(NO3)2. The
of Hecker and Bell (31) for Rh/SiO2 who observed the for- Rh/SiO2 used the same base silica and was made here by
mation of solid urea, an isomer of ammonium isocyanate, impregnation to incipient wetness with a solution of RhCl3
with a reduced catalyst and suspected, but were unable to to a metal content of 0.50% , dried overnight at 140 C, and
prove, the formation of HNCO when the catalyst was pre- reduced in 10% H2/He on a 1 C/min ramp ending with 2 h at
oxidised. Our second purpose was to determine if isocyanic 350 C. The dispersion, as determined by H2 chemisorption
acid undergoes hydrolysis to CO2 and NH3 as easily and in a flow system, was 38% which agreed with estimates from
completely on other oxides as it does on alumina. This transmission electron microscopy showing rounded metal
is of some importance since CeO2–ZrO2 is now the pre- particles of diameter 3 nm. Thus the number of surface

FORMATION AND HYDROLYSIS OF HNCO
13
metal atoms in the three catalysts on a weight of catalyst
basis is in the approximate ratio Pd : Pt : Rh = 2.0 : 1.2 : 1.0.
The 10 wt% BaO/SiO2 sample was made as for Rh/SiO2
with impregnation of a nitrate solution followed by drying
and calcination in air at 640 C with exposure to 2000 ppm
CO2/He for 40 min at 400 C during cooling to minimise car-
bon dioxide uptake in subsequent tests. The CeO2/Al2O3
(22 wt% CeO2, surface area 140 m2/g) and CeO2–ZrO2
(Ce : Zr mol ratio 1 : 1, 143 m2/g) samples were commer-
cial materials used in the manufacture of three-way cata-
lytic converters and provided by Johnson–Matthey Cata-
lytic Systems Division.
Testing of each sample of catalyst was commenced by
ramping the temperature up and down in the standard
NO + CO mixture ( 2900 and 3400 ppm, respectively)
until behaviour was stable. The NO +CO, NO +H2 ( 2000
and 1100 ppm, respectively), and NO +CO +H2 ( 2900,
3400, and 1100 ppm, respectively) reactions were then
carried out in turn by ramping the catalyst at 2 C/min with
analyses every 10 min. Mass balance calculations indicated
that the ramp rate was sufficiently slow to avoid transient
adsorption effects for all substances present except water.
RESULTS
Reactions on Pt/SiO2
The two individual reactions were investigated first. As il-
lustrated by Fig. 1A, the NO + H2 reaction occurs at a much
lower temperature than the NO + CO one on Pt/SiO2 with
complete H2 consumption by 80 C. At low temperature the
dominant product is N2O,
FIG. 1. Concentrations versus temperature for binary reactions over
75 mg Pt/SiO2. (A) 1900 ppm of NO with 1100 ppm H2. (B) 3000 ppm NO
with 3400 ppm CO and a total flowrate of 100 cm3/min.
reactant concentrations and CO2 shown in Fig. 2A and all
products, including CO2, on an expanded scale in Fig. 2B.
Carbon monoxide greatly inhibits the reaction of hydrogen
in the NO + CO + H2 system with an onset temperature
of 150 C for formation of N2O and H2O, compared with
near room temperature for the NO + H2 reaction on its
own (Fig. 1A). The amount of CO2 made throughout the
consumption of NO and H2 is several times greater than ob-
served for the NO + CO reaction. Thus the presence of H2
promotes CO consumption as noted previously (39). For-
mation of ammonia approximates that of CO2 while H2 is
being consumed, but then passes through a maximum. Iso-
cyanic acid is measurable from the onset of reaction with a
shoulder in its concentration near 250 C. This is followed
by a steep rise to a maximum of 1020 ppm at 316 C (a yield
relative to the input NO of 36% ) that coincides with a
minimum in NH3 concentration. At higher temperatures
HNCO production declines in concert with a fall in H2O
concentration, and a rise in NH3 and CO2 formation, as
expected for hydrolysis:
2NO + H2 → N2O + H2O,
[3]
which matches the NO/H2 ratio of two in the feed. After H2
consumption is complete, formation of NH3 increases with
temperature so that it becomes the major product between
150 and 200 C. At higher temperatures still, formation of
NH3 declines to zero, and N2 and N2O grow with the for-
mer taking over as the major product, especially so above
350 C. It should be noted here that the product distribu-
tion is highly dependent on the input NO/H2 ratio. With
2000 ppm NO and 5000 ppm H2 (not shown), NH3 was
the dominant product at all temperatures above 100 C in
accord with the literature (37, 38).
The NO + CO reaction (Fig. 1B) commences at 250 C
on Pt/SiO2 with complete conversion of NO by 380 C.
Nitrous oxide is the dominant nitrogen-containing prod-
uct at intervening temperatures with peak production of
1220 ppm at 330 C and a N2O selectivity (N2O/(N2O +N2))
of 82% . Nitrogen takes over as the major product above
360 C and becomes the sole one above 390 C.
Figure 2 shows the exit gas concentrations during the
reaction of NO + CO + H2 reaction over Pt/SiO2 with
HNCO + H2O → NH3 + CO2.
[4]

14
CHAMBERS, ANGOVE, AND CANT
distinct shoulder in CO removal, followed by a rise in N2
production, indicates the presence of a slower second-stage
reaction between N2O and CO.
The behaviour of the ternary mixture over Pd/SiO2 is
shown in Fig. 4. Hydrogen consumption requires a higher
temperature than for the corresponding NO + H2 reac-
tion, but the elevation in temperature is much less than for
Pt/SiO2 so inhibition by CO is less. As with the NO + H2
reaction alone, nitrous oxide is formed initially with N2 and
NH3 increasing once H2 consumption is complete. Forma-
tion of CO2 is similar to that for the NO + CO reaction on
its own. HNCO is not evident until 235 C, 75 C above the
temperature at which H2 consumption is complete, and ex-
hibits a peak of 550 ppm near 300 C (a yield of 20% ). Thus
the combined system appears to comprise a rapid NO +H2
reaction, which will be increasingly confined to the front of
the bed as temperature rises, a slower parallel NO + CO
reaction, and an even slower process generating HNCO
above 230 C.
The standard concentrations used above were chosen in
part to maximise the formation of HNCO with Pt/SiO2
FIG. 2. Concentrations versus temperature for the reaction of
2800 ppm NO, 3400 ppm CO, and 1200 ppm H2 over 75 mg Pt/SiO2 with a
total flowrate of 100 cm3/min. (A) Reactants plus CO2. (B) All products.
The general pattern for the NO + CO + H2 reaction is sim-
ilar to that seen previously with a reactant mixture with
approximately two-thirds the concentrations (26, 27), but
the HNCO concentrations are considerably higher. More
accurate analyses, better overall with hydrogen and water
for the first time, also allow a more detailed definition of
the key processes as explained later.
Reactions on Pd/SiO2
Figures 3 and 4 show the corresponding results for the
three reactions over Pd/SiO2. Its activity for the NO + H2
reaction (Fig. 3A) is slightly less than that of Pt/SiO2. Again,
N2O formation by reaction [3] dominates initially with
nitrogen taking over above 200 C. Formation of ammo-
nia is less than with Pt/SiO2 but, as in that system, it in-
creased greatly in experiments with H2 in 2.5-fold excess
(not shown). The activity of Pd/SiO2 for the NO + CO
reaction (Fig. 3B) is greater than that of Pt/SiO2 with a
somewhat different temperature profile. Formation of N2O
peaks at a lower concentration (960 ppm, selectivity 62% )
FIG. 3. Concentrations versus temperature for the binary reactions
over 75 mg Pd/SiO2. (A) 2000 ppm NO with 1000 ppm H2. (B) 2900 ppm
but is spread out over a wider range of temperature. A NO with 3450 ppm CO and a total flowrate of 100 cm3/min.

FORMATION AND HYDROLYSIS OF HNCO
15
FIG. 5. Dependence of HNCO formation on individual concentra-
tions over 75 mg of Pd/SiO2 at 225 C when remaining concentrations are
fixed (NO at 2900 ppm, CO at 3400 ppm, and H2 at 1200 ppm).
FIG. 4. Concentrations versus temperature for the reaction of
2900 ppm NO, 3400 ppm CO, and 1200 ppm H2 over 75 mg Pd/SiO2 with a
total flowrate of 100 cm3/min. (A) Reactants plus CO2. (B) All products.
based on previousmeasurementsofthe pressure dependen-
cies in that system (29). However, such concentrations are
not necessarily optimal for the production of HNCO with
the other PGMs. The pressure dependencies of HNCO pro-
duction over Pd/SiO2 at 282 C, when the NO conversion is
80% with the standard feed, are plotted in Fig. 5. HNCO
formation is favoured by higher H2 and CO concentrations,
as with Pt/SiO2 (29), and exhibits a maximum for NO con-
centrations somewhat below the standard one.
Reactions on Rh/SiO2
Data for the individual reactions over Rh/SiO2 are shown
in Fig. 6. The NO + H2 reaction (Fig. 6A) is much slower
than for Pt/SiO2 or Pd/SiO2 and produces N2O and N2 in the
approximate ratio of 4 : 1 during H2 consumption. Ammo-
nia is not a product under these conditions with NO in ex-
cess but, as with the other two PGMs, considerable amounts
were formed when usinga mixture with H2 in substantialex-
cess (not shown). The onset temperature for the NO + CO
reaction (Fig. 6B) is similar to that for the NO + H2 one and
FIG. 6. Concentrations versus temperature for the binary reactions
over 75 mg Rh/SiO2. (A) 1950 ppm NO with 1000 ppm H2. (B) 2900 ppm
much lower than that required for Pt/SiO2 and Pd/SiO2. The NO with 3500 ppm CO and a total flowrate of 100 cm3/min.

16
CHAMBERS, ANGOVE, AND CANT
reaction profile for Rh/SiO2 is almost identical to that for
Pt/SiO2, but displaced to lower temperature by 90 C. The
maximum concentration of N2O is 1120 ppm at 240 C
with a selectivity of 79% , very similar to that for the
NO + H2 reaction. Nitrogen formation takes over above
280 C with complete disappearance of N2O by 330 C.
When NO, CO, and H2 are reacted together over Rh/SiO2
(Fig. 7), allthree reactantsare removed simultaneouslyover
a narrow temperature interval (180 to 220 C). Production
of HNCO shows a sharp peak at 226 C, but the maximum
concentration (400 ppm, 14% yield), and the temperature
range over which it is seen, is much less than for Pt/SiO2 or
Pd/SiO2. Unlike the NO + H2 reaction over Rh/SiO2, the
ternary reaction yields considerable NH3 which accounts
for 25% of the NO reacted between 250 and 300 C. For-
mation of N2O passes through a maximum in a similar way
to that for the NO + CO reaction. A minor feature of the
ternary reaction over Rh/SiO2 alone is the reformation of
some hydrogen above 360 C, presumably by the water–gas
shift reaction between residual CO and product water.
FIG. 8. Dependence of HNCO formation on individual concentra-
tions over 75 mg of Rh/SiO2 at 225 C when remaining concentrations are
fixed (NO at 2900 ppm, CO at 3200 ppm, and H2 at 1200 ppm).
Figure 8 shows how the HNCO concentration changed
when the concentration of each reactant was varied in turn
while the other concentrations were held constant at a
temperature (225 C), giving 60% NO conversion under
standard conditions. Production of HNCO is increased if
the concentrations of NO and CO are reduced relative to
the standard ones or if higher H2 concentrations are used.
However, this result is likely to be dependent on the tem-
perature as well, and no attempts have been made to max-
imise HNCO concentrations due to the extensive experi-
mentation needed for optimisation in a four-dimensional
system.
Hydrolysis of HNCO on Oxides
Our previous work on the further hydrolysis of HNCO
on alumina placed downstream of Pt/SiO2 was carried out
at 230 C (27). As may be seen from Fig. 2, this temperature
falls within a range in which production of water exceeds
that of HNCO over Pt/SiO2. Thus complete hydrolysis of
HNCO according to Eq. [4] is feasible as was observed in
practice both with Al2O3 downstream of Pt/SiO2 and, sepa-
rately, with a Pt/Al2O3 catalyst. However, data in Fig. 2 also
show that the H2O/HNCO ratio is less than unity at tem-
peratures above 275 C. The hydrolysis/adsorption activity
of other oxides has now been investigated in this regime.
Figure 9 compares HNCO, H2, and NH3 concentrations
during the NO + CO + H2 reaction in two-random-order
experiments, one using Pt/SiO2 alone, the other a sequential
bed of Pt/SiO2 followed by CeO2–ZrO2. The presence of
FIG. 7. Concentrations versus temperature for the reaction of
2800 ppm NO, 3450 ppm CO, and 1500 ppm H2 over 75 mg Rh/SiO2 with
a total flowrate of 100 cm3/min. (A) Reactants plus CO2. (B) All products.

FORMATION AND HYDROLYSIS OF HNCO
17
tems (Fig. 10). Thus BaO/SiO2 is able to hydrolyse HNCO
to the limit of the water produced by the NO + CO + H2
reaction. When additional water is introduced HNCO dis-
appears altogether. The general behaviour of the Pt/SiO2 +
CeO2/Al2O3 system is similar except that there is an ini-
tial delay in the formation of HNCO, with elevated NH3
formation, that is attributable to HNCO hydrolysis on
residual hydroxyl groups. The steady-state concentration of
HNCO also appearsslightlylower than that with BaO/SiO2,
which may reflect some accumulation of surface isocyanate
groups. This would explain why the complete hydrolysis of
HNCO that occurs on subsequent introduction of water
gives rise to a higher concentration of ammonia than with
BaO/SiO2.
Similar experiments have been carried out on the ef-
fect of added water with other ceria-containing mixed ox-
ides downstream of Pt/SiO2. The behaviour of CeO2/SiO2
(28) is very similar to that of BaO/SiO2 in Fig. 10 while
that of CeO2–ZrO2 closely resembled that of CeO2/Al2O3
(Fig. 11), except that the time needed for HNCO to reach
a steady-state concentration when first placed on stream
FIG. 9. Comparison of HNCO, NH3, and H2O concentrations during
the NO + CO + H2 reaction in one- and two-bed systems: open symbols,
75 mg of Pt/SiO2 alone; closed symbols, 75 mg of Pt/SiO2 followed by 75 mg
of CeO2–ZrO2. Reactant concentrations and flowrates are the same as for
Fig. 2.
CeO2–ZrO2 reduces the concentration of HNCO and H2O
with corresponding increases in formation of NH3 (and also
of CO2, not shown). Indeed, since the H2O concentration is
zero with CeO2–ZrO2 present, reaction [4] has proceeded
to the fullest extent possible at all temperatures. Also, since
the reduction in HNCO concentration is largely balanced
by the formation of the corresponding amount of addi-
tional NH3, and CO2, one can conclude that loss of HNCO
by other processes, such as adsorption on the downstream
oxide, is small.
On the basis of Fig. 9, one would expect more hydroly-
sis if additional water was available. Figures 10A and 10B
compare the effect of adding 1200 ppm water at 315 C
when using Pt/SiO2 alone with that using Pt/SiO2, followed
by an equal mass of SiO2. With both systems the introduc-
tion of water is followed by stepwise loss of HNCO, 25%
for Pt/SiO2 alone but 50% for Pt/SiO2 + SiO2, with cor-
responding rises in the formation of NH3 (and of CO2, not
shown). Thus SiO2 alone has significant activity for HNCO
hydrolysis. This explains the lower concentration of HNCO
seen with the two-bed system compared to the single-bed
one prior to the addition of water.
The results of similar experiments using sequential beds
with BaO/SiO2 and CeO2/Al2O3 placed downstream of
Pt/SiO2 are shown in Fig. 11. Prior to water addition, the
product stream with BaO/SiO2 is completely dry. Corre-
spondingly, the concentration ofHNCO islower, and that of
FIG. 10. Effect ofadding 1200 ppm water duringthe NO + CO + H2
reaction at 315 C. (A) For 75 mg of Pt/SiO2 alone. (B) With 75 mg of
Pt/SiO2 followed by 75 mg of SiO2. Reactant concentrations and flowrates
NH3 ishigher, than with the Pt/SiO2 and Pt/SiO2 + SiO2 sys- are the same as for Fig. 2.

18
CHAMBERS, ANGOVE, AND CANT
standard concentrations). Comparison in terms of turnover
frequencies at a fixed temperature is impractical because
of the large extrapolations that would be needed to reach a
common temperature. The activity order is Rh
for the NO + CO reaction, but Pt Pd
Pd
Pt
Rh for the
NO + H2 one. This is the same as for -Al2O3-supported
metals (40). The only difference for -Al2O3 is that Pd is
slightly more active than Pt for the NO + H2 reaction (19,
41). Although it has been suggested that the NO + H2 re-
action proceeds by hydrogen-assisted dissociation (42, 43),
the most recent evaluation of single crystal data concludes
that both reactions can be adequately explained by an ini-
tial direct dissociation of NO followed by removal of the
fragments by adsorbed reductant (8), i.e.,
for the NO + H2 reaction with S representing a surface
site
NOads + S → Nads + Oads
Nads + NOads → N2O(g) + 2S
Nads + Nads → N2(g) + 2S
[5]
[6]
[7]
*
)
Nads + 3Hads
NHads + 2Hads + S
*
)
NH2ads + Hads + 2S → NH3(g) + 4S [8]
Oads + 2Hads → H2O(g) + 3S
[9]
and for NO + CO, reactions [5]–[7] plus
Oads + COads → CO2(g) + 2S.
[10]
FIG. 11. Effect of adding 1200 ppm water during the NO + CO +
H2 reaction at 315 C. (A) For 75 mg of Pt/SiO2 followed by 75 mg of
BaO/SiO2. (B) For 75 mg of Pt/SiO2 followed by 75 mg of CeO2/Al2O3.
Reactant concentrations and flowrates are the same as for Fig. 2.
The activity pattern for the two reactions can be inter-
preted in terms of these steps as follows. On rhodium, NO
adsorbs in preference to CO and readily dissociates with
creation of a surface that, depending on temperature, may
contain both adsorbed NO and nitrogen atoms under reac-
tion conditions (5). The rate is limited by the availability of
sites for NO dissociation and nitrous oxide is the favoured
product. As long as any NO remains, these processes are
little affected by the reductant. Thus the two binary re-
actions can be expected to exhibit rather similar rates (and
selectivities to N2O) on rhodium as observed here (Table 1).
with the dry feed was much shorter. CeO2–ZrO2 seemed
especially active for hydrolysis since even a few fine par-
ticles remaining on the walls of a used reactor were suffi-
cient to cause complete hydrolysis in subsequent tests with
Pt/SiO2 alone.
It is noticeable that conversions of both NO and CO in-
creased after water was introduced for all four of the ex-
periments illustrated in Figs. 10 and 11. This has nothing to
do with the downstream oxides since it occurs to a similar
degree with Pt/SiO2 alone (Fig. 10A). As explained later, it
appears to arise from a reaction of some of the additional
NH3, produced when H2O is present, with NO and CO.
TABLE 1
Comparison of Activities for Reactions over Pt/SiO2,
a
Pd/SiO2, and Rh/SiO2
Reaction
Pt/SiO2
Pd/SiO2
Rh/SiO2
DISCUSSION
NO + H2
60 C
285 C
183 C
96 C
261 C
127 C
202 C
213 C
196 C
NO + CO
Reactions on the Metals
NO + CO + H2
Table 1 compares the activity of the three metals in terms
of the temperatures required to give a fixed turnover fre-
quency of 0.02 molecules (NO) per surface atom per sec-
ond (which corresponds to conversions of 13–37% with the 2800 ppm for the NO + CO and NO + CO + H2 reactions.
a Temperatures giving a turnover frequency of 0.02 molecules(NO)/
surface atom/second with standard conditions of CO = 3400 ppm, H2
1000 ppm, and with NO = 2000 ppm for the NO + H2 reaction, but
=

FORMATION AND HYDROLYSIS OF HNCO
19
On platinum, NO is adsorbed less strongly and displaced by (50, 51). These form either by direct spillover or perhaps by
CO (44). Hence the NO +CO reaction is strongly inhibited the short-range transport of HNCO generated using the hy-
by CO and slower than on rhodium. The NO + H2 reac- drogen in neighbouring hydroxyl groups (52). In the ternary
tion is free of this inhibition. Dissociation of NO is possible reaction system, adsorbed hydrogen and nitrogen atoms are
at room temperature on some crystal planes of platinum continuously created, and this provides the driving force for
(8), and formation of N2O by the N + NO step is possi- the generation of HNCO as a gaseous product.
ble at temperatures as low as 100 K (45). This allows the
The amount of HNCO formed is much higher with
NO + H2 to be much faster on Pt than Rh. Palladium is Pt/SiO2 than either Pd/SiO2 or Rh/SiO2. This is consistent
closer to Pt in respect of CO adsorption, which is inhibiting with the especially strong adsorption of CO on Pt which
for the NO +CO reaction (46), but like Rh, the coverage by maximisesthe concentration ofCO in proximityto N atoms,
nitrogen atoms may be significant (47), and hence it exhibits thus favouring formation of metal-bound NCO and then
intermediate behaviour.
HNCO. However, consideration of product distributions at
It should be noted that the above considerations for different temperatures reveals a number of complexities in
rhodium apply strictly only when NO is in stoichiometric the Pt/SiO2 system. For example, at 236 C when the conver-
excess and no ammonia is formed, as is the case for the data sion of H2 is already complete, the amount of CO2 produced
in Fig. 6A. With hydrogen in excess the rate of the NO +H2 by the NO + CO + H2 reaction is much greater than that
reaction becomes faster than the NO + CO one (42) with for the NO + CO reaction (Table 2). Likewise the amount
considerable formation of ammonia, as found here when of NH3 made is much greater than that observed for the
using a 2 : 5 NO/H2 ratio. With a 2 : 1 mixture, formation of NO + H2 reaction and is similar in concentration to that
ammonia was confined to Pt/SiO2 and Pd/SiO2 at temper- of CO2. It seems probable that, despite the relatively low
atures above those giving complete hydrogen conversion temperature, they are both arising from the same source:
(Figs. 1A and 3A) and may be a consequence of mass trans- the hydrolysis of HNCO. The overall product distribution
fer restrictions (i.e., the initial step, [5], occurs rapidly in a is best described as the sum of
narrow section at the front of the bed where diffusion of
NO + CO + 1.5H2 → HNCO + H2O
[14]
NO to the metal surface is insufficient to maintain high NO
coverage). This would allow build-up of nitrogen adatoms
which can then be hydrogenated to NH3 using hydrogen
adatoms made available through the higher diffusion rate
of H2 compared with NO. The fall-off in NH3 production
above 200 C in these systems may be a partial consequence
of secondary reactions in the rear of the bed, for example,
and
HNCO + H2O → NH3 + CO2,
with a smaller contribution from the NO + H2 reaction
2NO + H2 → N2O + H2O.
[4]
[3]
6NO + 4NH3 → 5N2 + 6H2O,
[11]
This agrees with previous conclusions (27).
Above 260 C, formation of CO2 increases steeply while
that of NH3 (and also H2O) declines (Fig. 2). It is in this
region that HNCO production grows most strongly to a
maximum at 315 C which is close to a minimum in NH3
formation. It seems highly unlikely that the extra HNCO
arises directly from input H2, which must be completely
consumed towards the front of the bed. Likewise it can-
not be attributed to H2 arising from the water–gas shift
which is relatively fast on platinum at these temperatures
(48).
Reactions of the Ternary H2 + NO + CO Mixture
The ternaryreactionsexhibit a number ofunique features
not predictable from the behaviour of the binary reactions.
Most notable is the formation of HNCO in each system.
The most probable route is via the combination of adsorbed
hydrogen atoms with small numbers of isocyanate species
which exist on the metals in equilibrium with N atoms and
adsorbed CO, that is,
TABLE 2
Product Distributions for the Binary and Ternary Reactions
over Pt/SiO2 at 236 Ca
*
)
Nads + COads
NCOads + S
[12]
[13]
Product concentrations (ppm)
NCOads + Hads → HNCO(g) + 2S.
Reaction
NO + H2
N2O
N2
NH3
H2O
CO2
HNCO
Metal-bound isocyanate groups are known for all three
293
24
115
417
4
5
—
1145
—
—
74
—
—
317
PGMs from infrared studies (16). They can be observed NO + CO
NO + CO + H2
30
434
544
382
during the NO + CO reaction on Rh (15, 49), but not with
Pt or Pd. Nonetheless, their existence in small numbers
can be inferred from the rapid production of more sta-
a Data for the NO + CO + H2 reaction taken from Fig. 2, with the
NO + H2 and NO + CO reactions by interpolation between data points
ble support-bound NCO groups under reaction conditions on Figs. 1A and 1B, respectively.

20
CHAMBERS, ANGOVE, AND CANT
TABLE 3
amounts of NH3 arising directly from the NO + H2 reac-
tion at the front of the catalyst bed. Data for Pd/SiO2 in
Table 3 show that this is feasible. Again HNCO concentra-
tions start to decline at the point where NO consumption is
complete.
Formation of HNCO by Reaction of NO and CO with NH3
a
Compared with H2
Input concentrations, ppm
Temperature
C
HNCO,
ppm
The behaviour of Rh/SiO2 for the NO + CO + H2 reac-
tion (Fig. 6) exhibits a third pattern. The maximum yield
of HNCO is even less than with Pd/SiO2 and confined to
a narrow interval near 225 C. Formation of NH3 increases
above that temperature to a maximum of 700 ppm near
260 C despite the fact that ammonia is not a product of the
NO + H2 reaction with the standard 2 : 1 mixture (Fig. 7).
Most significantly, the peak in HNCO concentration co-
incides with total NO consumption while the increase in
NH3 is matched by a drop in water formation. The implica-
tion is that HNCO is still being made, by reaction [14], at
temperatures above its maximum yield, but is being hy-
drolysed to NH3, reaction [4], which cannot be recycled to
HNCO for lack of NO. Thus the primary reason for the low
yield of HNCO with Rh/SiO2 under the conditions used
here is that total consumption of H2 and NO is near coinci-
dent (because NO is simultaneously consumed by reaction
with CO). With Pt/SiO2 and Pd/SiO2, the NO + CO reac-
tion is much slower than the NO +H2 one which provides a
wide gap in temperature between the total consumption of
H2 and that of NO, a gap in which NO, CO, and NH3 may
coexist and generate HNCO via the secondary reaction,
Eq. [15].
Catalyst
Pt/SiO2
NO
CO
H2
NH3
315
2800
2800
2800
3100
3100
3100
1160
none
none
none
1140
245
1100
1200
650
Pd/SiO2
Pd/SiO2
296
282
2900
2900
3300
3300
1180
none
none
1170
600
800
1000
2400
none
400
300
a Using 75 mg of catalyst with total flowrate of 100 cm3/min.
reaction since estimates based on the data of Grenoble
et al. (53) for a Pt/SiO2 catalyst with similar dispersion indi-
cate that the rate of formation will be orders of magnitude
too slow under the present conditions. The clear implica-
tion is that much of the HNCO must be formed by reuse of
NH3 formed towards the front of the bed through reactions
[14] and [4].
Table 3 shows the results of several tests of this hypothe-
sis. With both Pt/SiO2 and Pd/SiO2, 1150 ppm of ammonia
is slightly more efficient in forming HNCO than a equiva-
lent concentration of hydrogen. With a lower NH3 concen-
tration of 245 ppm, Pt/SiO2 generates 650 ppm HNCO, a
remarkably high hydrogen selectivity of 88% . Since ther-
modynamics do not allow the formation of significant
amounts of HNCO from NH3 and CO alone, the process
must be driven by the dissociation of endogonic NO. The
most efficient stoichiometry is
The above considerations suggest that HNCO formation
will be very dependent on reaction conditions. Hecker and
Bell (31), using a reduced 4.6% Rh/SiO2 catalyst and 4- to
20-fold higher pressures than here, observed ammonia (in
19% yield) at 180 C plus a solid product, identified as urea
(NH2)2CO, in 12% yield. They concluded from mass bal-
ance calculations that the corresponding preoxidised cata-
NH3 + 2NO + 5CO → 3HNCO + 2CO2.
[15]
As one would then anticipate, and seen in Fig. 2, HNCO lyst gave a different solid product, with a C/N ratio of 1.0, in
formation peters out in the NO + CO + H2 system once a 42% yield which they suspected, but were unable to prove,
all NO is consumed at 315 C. At higher temperatures was HNCO. The present results demonstrate that gaseous
still, HNCO hydrolysis dominates and NH3 concentrations HNCO can indeed be formed. In our experience conver-
grow since recycling back to HNCO becomes impossible. sion of HNCO to solid products, by polymerisation or re-
It may be noted that reaction [15], coming on top of reac- action with ammonia, is difficult to avoid unless the entire
tions [14] and [4], represents an amplification in the HNCO analysis system is heated, with a temperature of 90 C being
concentration by a factor of 3, which does approximate the satisfactory for the concentrations produced here. The urea
increase in HNCO yield between the shoulder in its forma- observed by Hecker and Bell (31) probably arose through
tion at 250 C and the maximum yield at 315 C.
combination of HNCO with ammonia, being produced in
The characteristics of the NO + CO + H2 reaction on excess, in cooler sections of their system.
Pd/SiO2 are rather different from those for Pt/SiO2. The
According to Voorhoeve and Trimble (21), urea will be
reaction of H2 is much less inhibited by CO (Table 1), formed if the condensation is carried out at an elevated
and from 100 to 230 C the NO + H2 proceeds largely in temperature, but ammonium cyanate (NH4OCN) is formed
parallel with the slower NO + CO reaction (Fig. 4) as instead at room temperature. The products they report for
noted earlier. Formation of HNCO, commencing at 235 C, unsupported PGMs, HNCO with Pd but NH4OCN with Pt
is accompanied by a reduction in the amount of NH3. As and Rh (21–24) conform to the tendency to form ammonia
with the second stage of HNCO formation over Pt/SiO2, observed here (i.e., least with Pd/SiO2). However, they ob-
this is best explained in terms of a reaction between NO, served much higher yields: 60% HNCO with Pd and >90%
CO, and NH3, Eq. [15], but in this case using the small NH4OCN (i.e., 45% isocyanate) for Pt and Rh. The lower

FORMATION AND HYDROLYSIS OF HNCO
21
yields here can be partly attributed to the effect of hydroly- not exceed that required for removal of all carbon monox-
sis on the metal/support combination. While reaction with ide (29). Ammonia, probably derived from isocyanic acid,
NO and CO does allow some reconversion of ammonia pro- is also seen during light-off with complex mixtures, that are
duced by hydrolysis of HNCO, reaction [15], it necessarily marginally lean overall, on alumina-supported PGMs (30).
requires dissociation of NO which increases the scope for The pattern of ammonia formation for the individual metals
diversion of nitrogen atoms to other products. Voorhoeve resembles that for HNCO here. However, HNCO forma-
and Trimble also used higher temperatures which, coupled tion does require significant CO coverages, which means
with much higher CO pressures (50,000 ppm vs 3400 ppm that it cannot form when CO is fully consumed with oxy-
here), is sure to aid in the capture of N atoms as metal- gen in excess or at temperatures where equilibrium CO
bound NCO groups and then HNCO.
coverages are low. This probably confines the isocyanic
acid-to-ammonia route to <400 C on vehicles. In any event,
isocyanic acid itself is highly unlikely to exist in exhaust
under any circumstances since the high water content will
HNCO Hydrolysis on Oxides
It is apparent from Fig. 10 that SiO2 alone has some inevitably ensure complete hydrolysis on the active oxide
modest activity for the hydrolysis of HNCO. The amount washcoats used in catalytic converters.
of HNCO produced from NO + CO + H2 over Pt/SiO2 is
reduced when SiO2 is placed downstream and, with or with-
CONCLUSIONS
out this SiO2 present, it falls further when water is added.
However, the conversion to NH3 and CO2 is only partial
even at 315 C. The implication of the above discussion con-
cerning ammonia formation is that the PGM/silica catalysts
are more active than SiO2 alone for hydrolysis. This appears
Pt produces the largest amount, seemingly in two stages,
especiallyso for Rh/SiO2 where the generation ofNH3 from
The formation of HNCO is a characteristic feature of the
NO + CO + H2 reaction over silica-supported platinum,
palladium, and rhodium. Under the conditions used here,
firstly from H2 and then using NH3 being formed as a co-
product. With Pd, HNCO arises largely from NH3 alone
NO + CO + H2, but not from NO + H2, at temperatures
above 230 C can be explained in terms of HNCO hydrol-
because H2 is totally removed by reaction with NO at low
ysis. Higher hydrolysis activity is not unreasonable since
temperature. Rhodium gives rise to the least HNCO. For-
HNCO formed at the front of the bed may undergo dis-
mation is confined to a narrow temperature region due
sociation on metal particles downstream, a known process
to the coincident consumption of NO and H2 which pre-
on PGMs (14), thus facilitating contact between isocyanate
cludes reaction of NH3 with NO and CO. Hydrolysis of
groups and adjacent water molecules or surface hydroxyl
HNCO to NH3 and CO2 is appreciable on SiO2 alone and
groupsin a processwhich isadditionalto that on silica alone.
faster when a metal is present. Other oxide systems, both
Water may undergo dissociation on the metal, thus facilitat-
acidic and basic, give complete hydrolysis to the limit of the
ing hydrolysis of support-bound isocyanates which is also
water present and total reaction with even small excesses of
much faster on Pt/SiO2 than on SiO2 (52).
water. The possible presence of HNCO in vehicle exhaust
is not an issue since the presence of a vast excess of steam
It is clear that other oxide supports are much more ac-
tive than silica for HNCO hydrolysis. As shown by Figs. 9
and an active washcoat in three-way converters will ensure
and 11, CeO2–ZrO2, CeO2/Al2O3, and BaO/SiO2 give com-
complete hydrolysis. However, the latter process may con-
plete reaction to the limit of the water available. The same is
tribute to ammonia emissions at moderate temperatures
true for CeO2/SiO2 (28) and for Al2O3 alone (27). Hydrol-
under conditions where CO is still present.
ysis of isocyanic acid may be either acid- or base-catalysed
(54) and the ease with which it occurs on the range of
ACKNOWLEDGMENTS
oxides tested here indicates that weak sites of either type
will suffice.
This work has been supported by grants from the Australian Research
Several recent studies (55–57) have shown that mod-
ern vehicles emit significant quantities of ammonia with
concentrations up to 1,400 ppm being found in one study
(57). Formation of ammonia is of environmental impor-
tance since it contributes to the formation of atmospheric
haze through combination with NOx and SOx to make
ammonium salts (56). The present results indicate that the
generation of isocyanic acid as an intermediate is a proba-
ble contributor to ammonia formation with three-way cata-
lysts. It isincorrect to assume that HNCO cannot form when
oxygen is present. Previous work shows that isocyanic acid
formation continues as long as the amount of oxygen does
Council. The Pt/SiO2 and Pd/SiO2 catalysts used here were a past gift from
Professor R. Burwell.
REFERENCES
1. Heck, R. M., and Farrauto, R. J., “Catalytic Air Pollution Control:
Commercial Technology.” Van Nostrand Reinhold, New York, 1995.
2. Peden, C. H. F., Belton, D. N., and Schmieg, S. J., J. Catal. 155, 204
(1995).
3. Permana, H., Ng, K. Y. S., Peden, C. H. F., Schmieg, S. J., and Belton,
D. N., J. Phys. Chem. 99, 16344 (1995).
4. Permana, H., Ng, K. Y. S., Peden, C. H. F., Schmieg, S. J., Lambert,
D. K., and Belton, D., J. Catal. 164, 194 (1996).

22
CHAMBERS, ANGOVE, AND CANT
5. Herman, G. S., Peden, C. H. F., Schmeig, S. J., and Belton, D. N., 33. Griffith, D. W. T., Appl. Spectrosc. 50, 59 (1996).
Catal. Lett. 62, 131 (1999).
34. Rothman, L., Gamache, R., Tipping, R., Rinsland, C., Smith, M.,
6. Cho, B. K., J. Catal. 138, 255 (1992).
7. Cho, B. K., J. Catal. 148, 697 (1994).
Benner, D., MalathyDevi, V., Flaud, J. M., Camy-Peyret, C., Perrin, A.,
Goldman, A., Massie, S., Brown, L., and Toth, R., J. Quant. Spectrosc.
Radiat. Transfer 48, 469 (1992).
8. Nieuwenhuys, B. E., Adv. Catal. 44, 259 (2000).
9. Unland, M. L., Science 179, 567 (1973).
10. Unland, M. L., J. Catal. 31, 459 (1973).
11. Unland, M. L., J. Phys. Chem. 77, 1952 (1973).
12. Dalla Betta, R. A., and Shelef, M., J. Mol. Catal. 1, 431 (1975).
35. Uchijima, T., Herrman, J. M., Inoue, Y., Burwell, R. L., Butt, J. B., and
Cohen, J. B., J. Catal. 50, 464 (1977).
36. Pitchai, R., Wong, S. S., Takahashi, N., Butt, J. B., Burwell, R. L., and
Cohen, J. B., J. Catal. 94, 478 (1985).
13. Solymosi, F., Vo¨lgyesi, L., and Rasko´, J., Z. Phys. Chem. 120, 79 (1980). 37. Bauerle, G. L., and Nobe, K., Ind. Eng. Chem. Prod. Res. Dev. 13, 185
14. Solymosi, F., and Ba´nsa´gi, T., J. Phys. Chem. 83, 552 (1979).
15. Hecker, W. C., and Bell, A. T., J. Catal. 85, 389 (1984).
16. Solymosi, F., and Rasko´, J., Appl. Catal. 10, 19 (1984).
17. Kiss, J., and Solymosi, F., J. Catal. 179, 277 (1998).
(1974).
38. Mergler, Y. J., and Nieuwenhuys, B. E., Appl. Catal. B 12, 95 (1997).
39. Du¨mpelmann, R., Cant, N. W., and Trimm, D. L., Catal. Lett. 32, 357
(1995).
18. Matyshak, V. A., and Krylov, O. V., Catal. Today 25, 1 (1995).
19. Kobylinski, T. P., and Taylor, B. W., J. Catal. 33, 376 (1974).
40. Shinjoh, H., Muraki, H., and Fujitani, Y., Stud. Surf. Sci. Catal. 30, 187
(1987).
20. Shelef, M., and Gandhi, H. S., Ind. Eng. Chem. Prod. Res. Dev. 11, 393 41. Taylor, K. C., and Klimisch, R. L., J. Catal. 30, 478 (1973).
(1972). 42. Hecker, W. C., and Bell, A. T., J. Catal. 92, 247 (1985).
21. Voorhoeve, R. J. H., Trimble, L. E., and Freed, D. J., Science 200, 759 43. Burch, R., and Watling, T. C., Catal. Lett. 37, 51 (1996).
(1978).
44. Brown, M. F., and Gonzalez, R. D., J. Catal. 44, 477 (1976).
45. Wang, H., Tobin, R. G., DiMaggio, C. L., Fisher, G. B., and Lambert,
D. K., J. Chem. Phys. 107, 9569 (1997).
46. Holles, J. H., Switzer, M. A., and Davis, R. J., J. Catal. 190, 247
(2000).
22. Voorhoeve, R. J. H., and Trimble, L. E., J. Catal. 53, 251 (1978).
23. Voorhoeve, R. J. H., and Trimble, L. E., Science 202, 525 (1978).
24. Voorhoeve, R. J. H., and Trimble, L. E., J. Catal. 54, 269 (1978).
25. Voorhoeve, R. J. H., and Trimble, L. E., in “Chemical Industries:
Catalysis of Organic Reactions,” Vol. 5, p. 289. Dekker, New York, 47. Rainer, D. R., Vesecky, S. M., Koranne, M., Oh, W. S., and Goodman,
1981. D. W., J. Catal. 167, 234 (1997).
26. Du¨mpelmann, R., Cant, N. W., and Trimm, D. L., Appl. Catal. B 6, 48. Otto, K., Shelef, M., and Kummer, J. T., J. Phys. Chem. 74, 2690 (1970).
L291 (1995).
49. Chafik, T., Kondarides, D. I., and Verykios, X. E., J. Catal. 193, 303
(2000).
27. Du¨mpelmann, R., Cant, N. W., and Trimm, D. L., J. Catal. 162, 96
(1996).
28. Cant, N. W., Chambers, D. C., Cowan, A. D., Liu, I. O. Y., and Satsuma,
A., Top. Catal. 10, 13 (2000).
29. Du¨mpelmann, R., Cant, N. W., and Cowan, A. D., Stud. Surf. Sci. Catal.
101, 1175 (1996).
50. Chang, C. C., and Hegedus, L. L., J. Catal. 57, 361 (1979).
51. Almusaiteer, K., and Chuang, S. S. C., J. Catal. 184, 189 (2000).
52. Lorimer, D’A., and Bell, A. T., J. Catal. 59, 223 (1979).
53. Grenoble, D. C., Estadt, M. M., and Ollis, D. F., J. Catal. 67, 90
(1981).
30. Cant, N. W., Angove, D. E., and Chambers, D. C., Appl. Catal. B 17, 54. March, J., “Advanced Organic Chemistry,” p. 786. Wiley, New York,
63 (1998).
1985.
31. Hecker, W. C., and Bell, A. T., J. Catal. 88, 289 (1984).
32. Takahashi, N., Shinjoh, H., Iijima, T., Suzuki, T., Yamazaki, K., Yokota,
K., Suzuki, H., Miyoshi, N., Matsumoto, S.-I., Tanizawa, T., Tanaka,
T., Tateshi, S.-S., and Kasahara, K., Catal. Today 27, 63 (1996).
55. Laurikko, J., Int. J. Vehicle Des. 20, 159 (1998).
56. Fraser, M. P., and Cass, G. R., Environ. Sci. Technol. 32, 1053 (1998).
57. Baum, M. M., Kiyomiya, E. S., Kumar, S., and Lappas, A. M., Environ.
Sci. Technol. 34, 2851 (2000).