The surface chemistry of propylene, 1-iodopropane, and 1,3-diiodopropane on MoAl alloy thin films formed on dehydroxylated alumina

Article abstract of DOI:10.1021/jp060188u

The adsorption of C₃ hydrocarbons propylene, 1-iodopropane, and 1,3-diiodopropane is studied in ultrahigh vacuum on a molybdenum-aluminum alloy formed by molybdenum hexacarbonyl reaction with a planar alumina film grown on a Mo(100) substrate. Carbon-iodine bond scission occurs below ～200 K to deposit iodine, and form propyl species from 1-iodopropane and a C₃ metallacycle from 1,3-diiodopropane. Propyl species either undergo β-hydride elimination to yield propylene or hydrogenate to form propane. Propylene adsorbs as both π- and di-σ-bonded species, and the di-σ form hydrogenates to yield propane, where the addition of the first hydrogen to form propyl species is slower than the second hydrogenation step to yield propane. Propylene also thermally decomposes on the surface to desorb hydrogen and deposit carbon where the methylyne group is the most, and the methyl group the least reactive. The metallacyclic intermediate reacts to give an allylic intermediate, which forms propylene, but also decomposes by C-C bond cleavage to evolve ethylene and deposit methylene species on the surface. This is a key step in the mechanism proposed for heterogeneously catalyzed olefin metathesis and this is the first time that this chemistry has been directly identified in ultrahigh vacuum.

Full text of DOI:10.1021/jp060188u

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
J. Phys. Chem. B 2006, 110, 12555-12571
12555
The Surface Chemistry of Propylene, 1-Iodopropane, and 1,3-Diiodopropane on MoAl Alloy
Thin Films Formed on Dehydroxylated Alumina
Feng Gao, Yilin Wang, and W. T. Tysoe*
Department of Chemistry and Biochemistry, and Laboratory for Surface Studies,
UniVersity of Wisconsin-Milwaukee, Milwaukee, Wisconsin 53211
ReceiVed: January 10, 2006; In Final Form: March 27, 2006
The adsorption of C3 hydrocarbons propylene, 1-iodopropane, and 1,3-diiodopropane is studied in ultrahigh
vacuum on a molybdenum-aluminum alloy formed by molybdenum hexacarbonyl reaction with a planar
alumina film grown on a Mo(100) substrate. Carbon-iodine bond scission occurs below ∼200 K to deposit
iodine, and form propyl species from 1-iodopropane and a C3 metallacycle from 1,3-diiodopropane. Propyl
species either undergo â-hydride elimination to yield propylene or hydrogenate to form propane. Propylene
adsorbs as both π- and di-σ-bonded species, and the di-σ form hydrogenates to yield propane, where the
addition of the first hydrogen to form propyl species is slower than the second hydrogenation step to yield
propane. Propylene also thermally decomposes on the surface to desorb hydrogen and deposit carbon where
the methylyne group is the most, and the methyl group the least reactive. The metallacyclic intermediate
reacts to give an allylic intermediate, which forms propylene, but also decomposes by C-C bond cleavage
to evolve ethylene and deposit methylene species on the surface. This is a key step in the mechanism proposed
for heterogeneously catalyzed olefin metathesis and this is the first time that this chemistry has been directly
identified in ultrahigh vacuum.
1. Introduction
reverse of the insertion reaction forms metathesis products. Such
a mechanism was first proposed by He´risson and Chauvin,23
and was recently confirmed in ultrahigh vacuum on model
molybdenum carbide catalysts.24 As a precursor to studying the
metathesis chemistry, we focus here on the surface chemistry
of propylene, 1-iodopropane, and 1,3-diiodopropane (DIP) to
investigate the behavior of C3 species, as a prerequisite to
understanding the metallacycle formation pathway.
The formation and characterization of MoAl alloy films has
been extensively described.18 Briefly, they are formed by
adsorbing 5000 L of Mo(CO)6 at 700 K onto a dehydroxylated
alumina film grown on a Mo(100) substrate to generate a
molybdenum carbide film. This is then annealed to 1500 K to
generate a surface MoAl alloy. Alloy formation is confirmed
by X-ray photoelectron and Auger spectroscopies where it is
found that the binding energy of the deposited molybdenum,
following annealing, is lower than that of pure metallic
molybdenum.17,18 The thickness of the alloy film, based on
Auger electron probe depth analysis, is estimated to be 3-4
atomic layers22 with a Mo/Al atomic ratio of ∼1.6.
Mo(CO)6 has been extensively used as a catalyst precursor1-12
and more recently to generate model catalysts grown on planar
alumina thin films in ultrahigh vacuum, where surface-sensitive
analytical techniques can be used.13-16 Our recent studies have
shown that reacting Mo(CO)6 with aluminum17 and dehydroxyl-
ated and hydroxylated alumina18,19 thin films at 700 K deposits
molybdenum carbide incorporating a small amount of oxygen.
Reaction of low exposures of Mo(CO)6 with an alumina thin
film results in the formation of small molybdenum carbide
particles on the surface, while higher exposures (∼5000 L of
Mo(CO)6) lead to a thin carbide film that completely covers
the surface.18 Annealing these films to ∼1500 K causes CO
desorption through alumina reduction by the carbidic carbon,
and results in MoAl alloy film formation.17-19
Recently, studies of the chemistry of CH3I,20 CH2I2,21 and
ethylene and C2H5I22 on the MoAl alloy surface in ultrahigh
vacuum showed that besides hydrogenation, dehydrogenation,
and dissociation, extensive methylene migratory insertion20,21
and H-D exchange reactions20-22 occurred. It appears that the
catalytic properties of the alloy are different from metallic
molybdenum, but rather similar to those of late transition metals.
This has been rationalized by electronic effects where molyb-
denum gains electrons by alloying with aluminum.
Since MoAl alloy surfaces are extremely effective for
methylene insertion into surface-carbon single bonds to form
higher, singly bonded metal-alkyl groups,20,21 this implies that
they may also be similarly active for methylene insertion into
the carbon-surface bond of di-σ bonded olefin species to yield
metallacycles. Metallacycles have been proposed to be the
central intermediates in olefin metathesis catalysis, where the
2. Experimental Section
Temperature-programmed desorption (TPD) data were col-
lected in an ultrahigh vacuum chamber operating at a base
pressure of 8 × 10-11 Torr that has been described in detail
elsewhere13,17,18 where desorbing species were detected with a
Dycor quadrupole mass spectrometer placed in the line of sight
of the sample. Temperature-programmed desorption spectra
were collected at a heating rate of 10 deg/s. This chamber was
also equipped with a double-pass, cylindrical-mirror analyzer
for Auger spectroscopy, where the spectra are typically collected
by using an electron beam energy of 3 keV and the first
derivatives are obtained by numerical differentiation.
* To whom correspondence should be addressed. Phone: (414) 229-
5222. Fax: (414) 229-5036. E-mail: wtt@uwm.edu.
10.1021/jp060188u CCC: $33.50 © 2006 American Chemical Society
Published on Web 06/08/2006

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
12556 J. Phys. Chem. B, Vol. 110, No. 25, 2006
Gao et al.
X-ray photoelectron spectra (XPS) were collected in another
chamber operating at a base pressure of 2 × 10-10 Torr, which
was equipped with an X-ray source and double-pass cylindrical
mirror analyzer. Spectra were typically collected with an Mg
KR X-ray power of 250 W and a pass energy of 50 eV. The
alumina substrate was sufficiently thin that no charging effects
were found and the binding energies were calibrated by using
the Mo 3d5/2 feature (at 227.4 eV binding energy) as a
standard.17,18 Temperature-dependent XP and Auger spectra
were collected by heating the sample to the indicated temper-
ature for 5 s, allowing it to cool to 150 K, following which the
spectrum was recorded.
temperature decrease. At propylene exposures of 2 L and above,
the H2 yield saturates and the desorption temperature remains
constant at ∼370 K. The hydrogen desorption profile following
hydrogen adsorption on the MoAl alloy resembles that shown
in Figure 1b, indicating that propylene has decomposed at lower
temperatures. Note that the 2-amu signal below 200 K at
propylene exposures of 2 L and above is due to fragmentation
of molecular propylene. Shown in Figure 1c, at propylene
exposures from 2 to 7.5 L, weak yet detectable methane
desorption is found at ∼325 K and this increases to ∼350 K at
a propylene exposure of 10 L. The 16-amu signal at low
temperatures is due to fragmentation of molecular propylene
and CO contaminant. Figure 1d plots the 29-amu (propane)
desorption profiles, which is a very weak fragment of propylene
but the most intense for propane so that these traces represent
predominantly propane desorption. It is found, at propylene
exposures of 0.5 and 1 L, that weak yet detectable desorption
appears at ∼250 K. At propylene exposures of 2 and 5 L, the
intensity of this desorption state increases and the desorption
peak maximum also increases slightly to ∼270 K. Apparently,
this desorption state is due to propylene self-hydrogenation
where surface hydrogen originates from propylene dissociation
and background H2 adsorption. Meanwhile, sharp features
appear below 200 K at propylene exposures of 1 L and above.
Since the desorption temperature and line-shape of this state
resemble that of the 42 amu state (Figure 1a), it is assigned to
fragmentation of the parent molecule. As will be shown below,
no propane formation occurs at this temperature.
Infrared data were collected with a Bruker Equinox infrared
spectrometer equipped with a liquid nitrogen cooled, mercury
cadmium telluride detector operated at 4 cm-1 resolution and
data were typically collected for 1000 scans. The complete light
path was enclosed and purged with dry, CO2-free air.15,16
The Mo(100) substrate (1 cm diameter, 0.2 mm thick) was
cleaned by using a standard procedure, which consisted of argon
ion bombardment (2 kV, 1 µA/cm2), and any residual contami-
nants were removed by briefly heating to 2000 K in vacuo. The
resulting Auger spectrum showed no contaminants. Aluminum
was deposited onto Mo(100) from a small heated alumina tube,
which was enclosed in a stainless steel shroud to minimize
contamination of other parts of the system.25 The alumina thin
film is formed by cycles of aluminum deposition-water vapor
oxidation-annealing, until the Mo(100) XPS or Auger features
are completely obscured,18 yielding a film thickness of ∼2 nm.
Molybdenum hexacarbonyl (Aldrich, 99%), propylene (Mathe-
son, 99.5%), d6-propylene, CD2dCH-CH3, CH2dCH-CD3
(Cambridge Isotope, g99% D), 1-iodopropane, and 1,3-di-
iodopropane (Aldrich, 99%) were transferred to glass vials,
connected to the gas-handling line of the chamber, and purified
by repeated freeze-pump-thaw cycles, followed by distillation,
and their purities were monitored by mass spectroscopy. These
were dosed onto the surface via a capillary doser to minimize
background contamination. The exposures in Langmuirs (1 L
) 1 × 10-6 Torr s) are corrected by using an enhancement
factor determined by temperature-programmed desorption (see
ref 13 for a more detailed description of this procedure). H2
and D2 (Matheson, g99.5%) were used without further purifica-
tion.
It appears that two propylenic species form on the alloy
surface, one that desorbs at below 200 K, and another that
persists to much higher temperatures (Figure 1a). To establish
the nature of these two species, reflection-absorption infrared
spectroscopy (RAIRS) experiments were conducted and the
results are displayed in Figure 2, where 10 L of propylene was
adsorbed on the alloy surface at 80 K, and subsequently annealed
to higher temperatures. Following each annealing step, the
sample was allowed to cool to 80 K before each spectrum was
taken. It should be mentioned that propylene has a number of
infrared modes below 1000 cm-1, especially the generally most
intense ν(C-CH3) mode at ∼910 cm-1. Unfortunately, these
are obscured by an intense alumina LO mode.26,27 Following
adsorption at 80 K, features at 3072, 3056, 2977, and 2940 cm-1
are found in the C-H stretching region. In the low-frequency
region, relatively intense modes are detected at 1645, 1452, and
1435 cm-1. Comparing these features with solid/gas-phase
propylene (Table 128,29) immediately suggests that at least a
portion of the adsorbed propylene adopts a π-bonded conforma-
tion on the surface at 80 K. Annealing to 150 K causes the
disappearance of the 3072- and 3056-cm-1 features and a drastic
attenuation of the 2977-cm-1 peak, indicating the desorption
of π-bonded propylene. The CdC vibrational mode at 1645
cm-1 is still detectable, suggesting a portion of π-bonded
propylene still stays on the surface. In the meantime, the relative
intensity of the 1435-cm-1 feature increases compared with that
at 1452 cm-1. Annealing to 180 K substantially decreases the
intensity of the CdC vibrational mode, and results in a further
increase in intensity of the 1435-cm-1 feature compared to that
at 1452 cm-1. On heating to 200 K, the only detectable features
are a C-H stretching mode at ∼2923 cm-1 and a CH2
deformation mode at ∼1433 cm-1, which are assigned to di-
σ-bonded propylene.30,31 No features are detected above the
noise level at 220 K and higher. Experiments were also
performed by adsorbing propylene on the alloy surface at 300
K and no detectable features were found (not plotted), indicating
3. Results
3.1. Propylene on a MoAl Alloy. The surface chemistry of
propylene on a MoAl alloy surface was investigated with
temperature-programmed desorption (TPD). Figure 1 displays
a number of TPD profiles as a function of propylene exposure,
monitoring desorption at 42 (C3H6), 2 (H2), 16 (CH4), and 29
(C3H8) amu. As shown in Figure 1a, essentially no propylene
desorption is found at an exposure of 0.2 L, suggesting complete
dissociation of adsorbed propylene. At a propylene exposure
of 0.5 L, two weak desorption peaks are found at ∼210 and
∼340 K, respectively. At an exposure of 1 L, the intensity of
both states increases, and the desorption temperature of the low-
temperature state decreases to ∼195 K. Upon further increasing
the exposure to 2 L, an extra low-temperature state develops at
∼175 K, and the desorption temperature maximum of the high-
temperature state also decreases. The intensity of the 175 K
state continues to increase at higher exposures and saturates at
a propylene exposure of ∼7.5 L (not shown). Figure 1b depicts
the H2 desorption profiles. At the lowest propylene exposure
(0.2 L), H2 desorbs at ∼390 K. The H2 yield increases with
increasing propylene exposure, accompanied by a desorption

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Adsorption of C3 Hydrocarbons
J. Phys. Chem. B, Vol. 110, No. 25, 2006 12557
Figure 1. Temperature-programmed desorption spectra of propylene adsorbed on a MoAl alloy at 150 K as a function of propylene exposure,
where the exposures are marked adjacent to the corresponding spectrum, monitoring (a) 42 (propylene), (b) 2 (hydrogen), (c) 16 (methane), and (d)
29 (propane) amu. The data were collected at a heating rate of 10 deg/s.
no propylidyne formation on the surface. The RAIRS results,
together with TPD data shown in Figure 1a, suggest that both
π- and di-σ-bonded propylene adsorb on the alloy surface and
that the former desorbs below 200 K. Efforts were also made
to adsorb lower exposures of propylene on the alloy surface at
80 K to explore whether di-σ-bonded propylene forms prior to
π-bonded propylene, as seen on other surfaces,30,31 but no signal
above the noise level could be identified. In fact, even when
10 L of propylene was adsorbed at 80 K where relatively strong
features can be detected, the quality of the infrared data is
substantially worse than that obtained following propylene
adsorption on a Pd(111) single crystal30 with use of identical
experimental apparatus.
Data shown in Figure 1d indicate that di-σ-bonded propylene
hydrogenates to form propane where surface hydrogen originates
from both propylene dissociation (self-hydrogenation) and
background H2 adsorption. Since background H2 always adsorbs
on the surface during sample cooling, and since it is relatively
difficult to choose the proper mass to monitor propane formation
(29 and 43 amu contain a contribution from molecular propyl-
ene, at least at large propylene exposures, and 44 amu contains
a contribution from background CO2 adsorption), propylene self-

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
12558 J. Phys. Chem. B, Vol. 110, No. 25, 2006
Gao et al.
Figure 2. Reflection-absorption infrared spectra (RAIRS) following
the adsorption of 10 L of propylene on a MoAl alloy as a function of
annealing temperature, where the annealing temperatures are displayed
adjacent to the corresponding spectrum.
TABLE 1: Comparison of the Vibrational Frequencies
(cm-1) of Propylene Adsorbed on a MoAl Alloy at 80 K with
the Vibrational Frequencies of Solid or Gas-Phase Propylene
10 L C3H6/
solid/gas-phase
propyleneb
MoAl at 80 Ka
assignmentc
3072
3056
2977
-/3081
νa (CH2)
overtone
νs (CH2)
νa (CH3)
overtone
ν (CdC)
δa (CH3)
δ (CH2)
3064/3067
2970/2979
-/2960
2940
1646
1452
1435
2934/2942
1647/1647
1453/1448
1437/1416
aThis work. b Taken from refs 26 and 27. c ν, stretching; δ, deforma-
tion; a, asymmetric; s, symmetric.
hydrogenation and hydrogenation was explored by using C3D6
and D2 and the results are displayed in Figure 3, where
desorption at 52 amu (C3D8) was monitored. Figure 3a depicts
propane formation through self-hydrogenation. Almost no C3D8
desorption is found at a C3D6 exposure of 0.5 L, while at
exposures of 1 L and above, C3D8 desorbs at ∼260 K, a temp-
erature similar to that seen in Figure 1d. When the alloy surface
is predosed with 20 L of D2 prior to C3D6 adsorption, the C3D8
yield is much larger (Figure 3b). Note that even at a propylene
exposure of 0.5 L, where no propane desorption is found from
the bare alloy surface, extensive propane desorption occurs from
the deuterium-covered surface, at ∼260 K. The relative yields
of propane (measured by integrating the desorption peak areas)
are plotted as an inset to Figure 3b as a function of propylene
exposure to both bare and deuterium-covered surfaces. In both
cases, the propane yield saturates at a propylene exposure of 4
L, while the saturation yield of propane on the deuterium-
covered surface is ∼3 times larger than that on the bare surface.
Note also that no propane formation is found below 200 K.
H-D exchange reactions were conducted on the alloy surface
with TPD to monitor the desorption of propylene and propane
isotopomers. Figure 4a displays the desorption profiles between
29 and 32 amu following a 10 L C3H6 exposure to a surface
precovered by 20 L of D2 where these masses are due mainly
to C2 fragments of propane isotopomers. Note that the low-
Figure 3. Temperature-programmed desorption spectra of C3D6
adsorbed on (a) a clean MoAl alloy surface and (b) an alloy surface
precovered by 20 L of D2, monitoring 52 amu (propane) as a function
of C3D6 exposure, where the exposures are marked adjacent to the
corresponding spectrum. Shown as an inset is a plot of the propane
yield as a function of propylene exposure on the clean and deuterium-
covered alloy surface.
temperature desorption at 29 and 30 amu is due to fragmentation
of molecular C3H6 (Figure 1d). These desorption profiles clearly
indicate that H-D exchange does occur on the surface and that
the propane isotopomers desorb at ∼260 K, indicating that only
di-σ-bonded propylene undergoes H-D exchanges, as suggested
previously.32 Figure 4b displays the corresponding spectra
between 43 and 52 amu. Masses above 48 amu can be assigned
unambiguously to propane isotopomers; in particular the 51 and
52 amu signals are assigned to C3HD7 and C3D8, respectively.
However, the desorption profiles at from 43 to 48 amu may
contain contributions from both propylene and propane isoto-