Gas-phase reactions of rhenium-oxo species ReOn+, n = 0, 2 6, 8, with O2, N2O, CO, H2O, H2, CH4 and C2H4

Article abstract of DOI:10.1039/b100538n

The reactions of ReO_n⁺, n = 0, 2-6, 8, with the small molecules O₂, N₂O, CO, H₂O, H₂, CH₄ and C₂H₄, are studied by Fourier transform ion cyclotron resonance (FT-ICR) mass spectrometry under single collision conditions on the timescale of seconds. The ReO_n⁺ species are produced by laser vaporization of solid rhenium and pulsed supersonic expansion in a helium-oxygen mixture into high vacuum. A wide variety of reactions are observed, including methane activation and epoxidation reactions. Re⁺ reacts with ethylene by sequential dehydrogenation. ReO₂⁺ and ReO₄⁺ exhibit the most diverse reaction pathways, while ReO₅⁺ almost exclusively undergoes ligand exchange. ReO₆⁺ and ReO₈⁺ are largely unreactive, the only efficient reactions are observed are with ethylene and water. Both molecules seem to be able to directly attack the dioxygen ligands. The observed chemistry is governed by a fine interplay between available coordination sites and thermochemistry.

Full text of DOI:10.1039/b100538n

View Article Online / Journal Homepage / Table of Contents for this issue
Gas-phase reactions of rhenium-oxo species ReO ‘, n = 0, 2–6, 8,
n
with O , N O, CO, H O, H , CH and C H
2
2
2
2
4
2 4
Martin K. Beyer, Christian B. Berg¤ and Vladimir E. Bondybey*
Institut fur Physikalische und T heoretische Chemie, T echnische Universitat Munchen,
Ž Ž
Ž
L ichtenbergstrae 4, 85747 Garching, Germany
Received 15th January 2001, Accepted 20th March 2001
First published as an Advance Article on the web 6th April 2001
The reactions of ReO `, n \ 0, 2È6, 8, with the small molecules O , N O, CO, H O, H , CH and C H , are
n
2
2
2
2
4
2 4
studied by Fourier transform ion cyclotron resonance (FT-ICR) mass spectrometry under single collision
conditions on the timescale of seconds. The ReO ` species are produced by laser vaporization of solid rhenium
n
and pulsed supersonic expansion in a heliumÈoxygen mixture into high vacuum. A wide variety of reactions
are observed, including methane activation and epoxidation reactions. Re` reacts with ethylene by sequential
dehydrogenation. ReO ` and ReO ` exhibit the most diverse reaction pathways, while ReO ` almost
2
4
5
exclusively undergoes ligand exchange. ReO ` and ReO ` are largely unreactive, the only efficient reactions
6
8
observed are with ethylene and water. Both molecules seem to be able to directly attack the dioxygen ligands.
The observed chemistry is governed by a Ðne interplay between available coordination sites and
thermochemistry.
oxygen atoms.3 Much less extensively explored were species
with more than one oxo-ligand, and the e†ects of multiple oxo
ligands upon the metal reactivity. Irikura and Beauchamp10
have produced ionic fragments by electron impact ionisation
of neutral osmium tetroxide, and investigated their reactions
with a variety of small molecules. Cassady and McElvany11
1
. Introduction
FT-ICR mass spectrometry is an extremely convenient and
useful tool for investigating ionÈmolecule reactions, and in the
last decade hundreds of such studies have been reported.1h4
Such reactions in the nearly collision free gas-phase environ-
ment proceed under conditions which di†er drastically from
those of condensed phase processes, and one thus has to exer-
cise considerable caution when trying to transfer and apply
the results to more conventional, solution phase or high pres-
sure gas-phase reactions.4 On the other hand, mass spectrom-
etry coupled with external ion sources makes it possible to
generate solvated ions, and investigate their reactions and the
e†ects of ligands upon the ion reactivity.5 One advantage of
the absence of the condensed phase environment is that it may
provide insight into which reactions are solvent induced, and
which are due to the intrinsic properties of the reacting ion
itself.5h8 The gas phase studies also permit unambiguous iden-
tiÐcation of the individual reactant ions, quantitative determi-
nation of reaction rates and branching ratios, and clear
distinction of primary products from those which are due to
consecutive reactions.
Particularly extensively studied were ions and ion clusters
involving transition metals.4,9 The interest in these species is
motivated by their importance in catalysis. Most of these ele-
ments are distinguished by a complex electronic structure with
a number of oxidation states, a multitude of low-lying elec-
tronic states, and many unpaired electrons. Their ability to
accept or donate electrons in a large variety of redox pro-
cesses explains their efficiency in catalysing chemical reactions.
Transition metals and their compounds are not only the most
frequently used catalysts in chemical industry, but they are
also of paramount importance in a whole variety of key bio-
logical processes.
explored the reactions of the MoO` and MoO ` cations with
2
small hydrocarbons. They considered the possibility that
MoO ` could have a peroxo structure, but had to reject this
2
idea based on the observed reactions. Also in the case of the
OsO ` ions, a peroxo structure is, based on the known struc-
2
ture of the parent OsO , unlikely.
4
Even though there is considerable practical interest in the
oxides of rhenium, in particular in view of the recent demons-
tration that methyltrioxorhenium (MTO) is quite an efficient
and selective epoxidation catalyst,12h14 little information is
available about the existence of rhenium oxides in the gas
phase. The Re O `, Re O `, Re O `, ReO ` and ReO `
2
7
2 6
2 5
3
2
ions were detected in the products resulting from electron
impact ionization of sublimed Re O ÈReO oxide
a
2
7
3
mixture.15 It was also observed that interaction of impurities
with a heated rhenium Ðlament in mass spectrometers yields
the ReO`, ReO ` and ReO `,16 as well as ReO ~ and
2
3
2
ReO ~ ions.17 The formation of the anionic species was also
3
reported in the reaction of the hot Ðlament with water, cataly-
sed by rare earth oxides.18 In all the studies it was assumed
that the ions are fragments of the stable oxides characterised
in the solid, but no studies of their structure were carried out.
Work in our laboratory has demonstrated that, using a versa-
tile ion source combining laser vaporization with supersonic
expansion which we constructed and interfaced to our
FT-ICR mass spectrometer, metal cations with up to a dozen
oxygen atoms are easily generated. In a recent investigation19
we could show that with the help of collision induced disso-
ciation (CID) results, one can get useful information about the
thermochemistry of rhenium oxide cations, and even gain
some insight into their structures. In order to gain some
further insight into the rhenium cation chemical properties
and how they are a†ected by the ligands, in the present work
While there are numerous studies of transition metal oxide
ions, most of these involve small species with one or two
¤
Current address: Bruker Daltonics, Inc., Manning Park, Billerica,
MA 01821, USA.
1840
Phys. Chem. Chem. Phys., 2001 3, 1840È1847
This journal is ( The Owner Societies 2001
DOI: 10.1039/b100538n

View Article Online
we investigate reactions of the rhenium oxide cations, ReO `,
3
. Results and discussion
n
n \ 0, 2È6 and 8, with a variety of small molecules. We
present the results in tabular form, and interpret and discuss
them in terms of possible reaction mechanisms, and ionic
structures.
The results of the reactivity studies are summarized in Tables
È4. These give the reaction rate constants kabs, efficiencies
U \ kabs/k with the collision rate k as derived from the
1
ADO
ADO
average dipole orientation theory,22 as well as, where multiple
products are observed, the branching ratios. Also given is the
accuracy class as letters A, B and C. For those reactions which
are relatively efficient and where the products are observed
with good signal to noise ratio, the rate constants are in
general reproducible to ^10% and labeled class A. In view of
the possible systematic errors in pressure calibration and
other factors we conservatively estimate the absolute accuracy
of these results to be ^20È30%. For some reactions where the
product formation is slower, or where secondary processes or
reactions with impurities interfere, the accuracies are lower,
and we indicate such values in the table by the letter B. Those
denoted by C refer either to species whose intensities are low,
or to tertiary or further sequential reaction products. Mass
spectrometry naturally gives no direct information about the
neutral species produced in the reactions, but they can usually
be unambiguously identiÐed from the mass balance of the
reaction. It is in general assumed that the products in most
cases should be thermodynamically stable molecules, whose
formation and stabilization does not require complex or
unusual mechanisms. Thus, for instance, when three oxygen
2
. Experimental details
The experiments were performed on a modiÐed Spectrospin
CMS47X FT-ICR mass spectrometer described in detail else-
where.9 Rhenium cations ligated with oxygen of the nominal
composition [Re, O ]`, n \ 0È10,19 were produced in a disk
n
type laser vaporization source. Only traces of the n \ 1, 7 and
n [ 8 were present, and these species were therefore not inves-
tigated. The target material was pure rhenium, the helium
carrier gas (Helium 4.6, Messer Griesheim) at 10 bar was
seeded with 0.2 mbar O (Sauersto† 4.5, Messer Griesheim).
The Nd : YAG laser Quanta Ray GCR3, Spectra Physics, was
2
operated at a pulse energy of 12 mJ at 532 nm with a repeti-
tion rate of 25 Hz, and focused to a 1 mm spot on the target.
The metal plasma was entrained in the carrier gas pulse,
cooled by Ñowing through a 10 mm long conÐning channel
and by subsequent supersonic expansion into high vacuum of
ca. 1 ] 10~4 mbar. The ligated rhenium cations formed were
transferred through several stages of di†erential pumping into
the high-Ðeld region of the superconducting magnet by a
system of electrostatic lenses, and stored inside the ICR cell in
the presence of the reaction gas. Ions from 50 injection cycles
were accumulated to improve the signal to noise ratio. The
reaction gas O (Sauersto† 4.5), N O (Stickoxydul 2.0), CO
atoms are lost, it is assumed that the products are O and an
2
O atom, rather than ozone, O .
3
Reactions with molecular oxygen
Out of all the species studied, only the dioxide, ReO `,
2
2
2
(
Kohlenmonoxid 4.7), H (Wassersto† 5.0), CH (Methan 2.5)
exhibits an observable reaction with O , Table 1. It is, with
2
4
2
fairly high collision efficiency, oxidized to ReO `, as already
3
and C H (Ethylen 3.5, all Messer Griesheim) was introduced
into the ultrahigh vacuum region via a needle valve at a
typical nominal pressure of 1.0 ] 10~8 mbar, as measured at
the Balzers TPG 300 ion gauge. De-ionized H O was intro-
2
4
reported by Irikura et al.23 They give for the rate constant of
this reaction a value of 20 ] 10~12 cm3 s~1, more than a
factor of two lower than observed here, but without discussing
the structures of the ions. The thermochemistry of the reaction
will obviously depend strongly on the assumed speciÐc
geometry of both the reactant and the product ion. If one
assumes that the reactant ion is a rhenium cation complex,
2
duced via the same needle valve after three freezeÈpumpÈthaw
cycles, the water vapor pressure was kept constant by a bath
heated to 40 ¡C. The nominal pressure was corrected to an
absolute pressure by the relative sensitivity of the ion gauge20
and a geometry factor of 3.7 derived from ionÈmolecule reac-
tions whose rate constants are accurately known.21 We esti-
mate the accuracy of the absolute pressures so determined to
be about ^25%. Ion selection was achieved by applying
single frequency shots, each 8 ms long with a peak-to-peak
Re(O )`, the reaction (1) leading to a trioxorhenium cation
2
is, according to DFT calculations,19 exothermic with
*H B [300 kJ mol~1, while for the covalent Re(V) com-
pound, ReO `, it is computed to be endothermic, *H B ]40
2
kJ mol~1. The efficient oxidation would appear to strongly
voltage of V \ 2.7 V. To avoid o†-resonant excitation of
the selected ions, nearby peaks, especially the second rhenium
favor the complex ion, which can easily be formed in our
supersonic expansion source. This could also explain the
seeming discrepancy between our result and that of the pre-
vious study: even though our ions are surely cold, we observe
a much higher rate than Irikura et al., who generate the ions
by direct laser vaporization without any cooling, and hence
they are likely to be produced hot. Since in their source cova-
lent ions are more likely to be produced, the discrepancy dis-
php
isotope, were not ejected. The reaction was monitored from 0
s until Ðnal products were reached, and up to 20 mass spectra
were taken for the intensityÈtime proÐle. In most cases, the
branching of the reaction could be identiÐed by analyzing the
reaction kinetics, and products could be unambiguously
assigned with their absolute mass. In some cases, products
were conÐrmed by ejecting one of the two rhenium isotope
peaks. The product then also lacked the respective isotope.
appears: their hot ReO ` ions can undergo endothermic
2
reaction to ReO `.
3
Table 1 Reactions with oxidising or reducing agents O , N O and CO. kabs denotes the absolute rate constant, U the efficiency of the reaction,
2
2
and the accuracy class is given to indicate the reliability of the rate constant, where A is excellent, B good, and C semi-quantitative
Branching
ratio (%)
kabs/
Accuracy
class
Reaction
10~12 cm3 s~1
U (%)
ReO ` ] O ] ReO ` ] O
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
100
100
100
100
100
100
100
100
100
50
751
148
834
6
331
242
433
4
9
100
22
100
9
50
37
A
A
A
A
C
A
A
A
C
2
2
3
ReO ` ] N O ] ReO ` ] N
2
2
3
2
ReO ` ] N O ] ReO ` ] N ] O
4
2
3
2
2
ReO ` ] N O ] ReO N O` ] O
5
2
3
2
2
2
ReO ` ] N O ] ReO N O` ] O
6
2
4 2
ReO ` ] CO ] ReO ` ] CO
3
2
2
2
ReO ` ] CO ] ReO ` ] CO
4
3
ReO ` ] CO ] ReO CO` ] O
66
0.6
5
3
2
2
ReO ` ] CO ] ReO CO` ] O
6
4
Phys. Chem. Chem. Phys., 2001, 3, 1840È1847
1841

View Article Online
The behavior of our cold ions can also easily be under-
ReO `. Both of these reactions are computed to be strongly
3
stood, if they predominantly have the Re(O )` complex struc-
exothermic, with *H B [638 and [429 kJ mol~1, respec-
tively. It appears likely that structurally this product is the
covalent trioxorhenium cation. While the DFT calculations
suggest that the reactions would be exothermic even if one
2
ture. Collision with an additional O molecule results in the
2
formation of an excited ReV(O ) ` bis-peroxo-complex. This
2
2
can decompose again with the loss of one of the O ligands,
2
but the energy stored in the ion can also help surmount the
assumed the alternative ReO(O )` product, in view of the
2
barrier to breakage of one of the OÈO bonds. The Re(O )`
high energy content of the reaction complex intermediate, the
2
entity isomerizes to the covalent ReO `, which is then oxi-
isomerization to the more stable covalent product appears
more probable. This is also supported by the fact that no trace
of the ligand exchange products Re(N O)` or Re(O )(N O)`
2
dized by the second O ligand to the trioxorhenium cation.
2
The rhenium atom thus reaches its preferred oxidation state of
2
2
2
VII, and an oxygen atom is formed as a neutral product. We
will see that such highly exothermic isomerization of peroxo
ligands with formation of covalent Re2O double bonds is
apparently instrumental in numerous reactions to be dis-
cussed below.
was detected. The fact that no reaction is observed for ReO `
3
provides strong support for this ion produced in our source
being almost exclusively the covalent oxide of Re(VII), rather
than a ReO(O )` complex.
2
In contrast with the smaller oxides, both ReO ` and
5
ReO ` react with carbon monoxide and nitrous oxide by
6
For none of the higher oxide ions was any detectable reac-
tion with oxygen found. It would naturally be very interesting
simply exchanging one of the O molecules for a CO or N O
2
2
to investigate reactions with isotopic 18O , and search for
ligand. In either case the reaction is efficient for the smaller
2
ligand exchange reactions, which could provide further insight
ReO ` ion, but the rate drops by more than an order of mag-
5
into the structure of the ions. In recent studies of [Cr,O ]`
nitude for ReO `. This decrease in the rate probably reÑects
2
6
the complex Cr(O )` was unambigously identiÐed by this type
of experiment.24
the increasing difficulty of the N O reactant to penetrate the
solvation shell and interact with the metal as the number of
2
2
oxygen ligands is increased. We have discussed this behavior
Reactions with carbon monoxide and nitrous oxide
of peroxo-complex cations in our previous work with N and
2
CO .6 This interpretation Ðnds further conÐrmation in the
Carbon monoxide reduces ReO ` to ReO ` and Re(O ) ` to
2
3
2
2 2
fact that ReO `, where the Ðrst solvation shell is presumably
completed and closed, does not react at all with either CO or
ReO `, in both cases with nearly 50% collisional efficiency,
8
3
and is presumably oxidized to CO in the process. Both reac-
tions are computed to be exothermic, with *H B [69 and
630 kJ mol~1, respectively. It is interesting to note that the
2
N O.
2
[
Reactions with molecular hydrogen and water
latter reaction is slower, even though its exothermicity is
nearly a factor of ten larger. This may reÑect the increasing
difficulty of the CO reactant to interact with the metallic core
as the number of oxygen atoms is increased, or a high activa-
tion barrier for breaking the OÈO bond and isomerizing the
peroxo ligand to two covalently bound oxygen atoms. It
Water impurities are always present in the residual gas in our
spectrometer, and we have therefore investigated its reactions,
as well as those with molecular hydrogen. All of the oxides
studied react with water, albeit the smallest species, ReO `
2
and ReO ` only rather slowly, with rate constants shown in
3
should be noted that the product ReO ` is di†erent from the
Table 2. It is interesting that for each of the ions one of the
2
Re(O )` species presumably produced in the laser vapor-
2
Ðnal products of the reaction of water is the ReO H ` ion,
4 2
ization source. There may be another, reaction with CO invis-
which in fact appears to be a very ubiquitous product in the
ible in mass spectrometry: As in the case of Cr(O )`,4,24
reactions of the rhenium oxides, and is apparently thermody-
namically strongly favored. In fact, among some twenty reac-
2
Re(O )` may isomerize to ReO ` upon collision with CO,
2
2
instead of undergoing a ligand exchange with the very
common carbonyl ligand CO.
tions of the oxides with water and H which we investigated,
2
thirteen form either this product ion directly, or an interme-
Also interesting, but di†erent are the nitrous oxide reac-
tions. While ReO ` does not react with N O at all, both
diate, which then in turn reacts to ReO H `. While this might
4
2
be formulated as a rhenic acid cation, ReO (OH) `, a water
3
2
2 2
complex ReO (H O)` appears to be a likelier structure. The
ReO ` and ReO ` do, with the former being oxidized and
2
4
3 2
free rhenic acid H ReO is an unstable compound but the
the latter reduced, resulting in the same, unique product,
2
4
Table 2 Reactions with H and H O
2
2
Branching
ratio (%)
kabs/
Accuracy
class
Reaction
ReO ` ] H O ] ReO ` ] H
10~12 cm3 s~1
U (%)
(10)
(11)
(12)
(13)
(14)
(15)
(16)
(17)
(18)
(19)
(20)
(21)
(22)
(23)
(24)
(25)
(26)
(27)
(28)
(29)
100
100
61
25
14
100
100
33
67
65
33
48
2
3
C
B
B
B
B
C
A
B
B
C
B
C
C
A
A
B
A
A
C
B
2
2
3
2
ReO ` ] H O ] ReO H `
3
2
4 2
ReO ` ] H O ] ReO H ` ] O
922
369
205
1130
1850
462
924
1270
693
1840
1500
81
69
25
144
50
18
53
21
12
65
100
27
53
73
40
100
86
54
46
16
96
33
12
3
4
2
4 2
]
]
ReO H` ] OH
4
ReO H ` ] O
3
2
2
ReO H` ] H O ] ReO H ` ] OH
4
2
4 2
ReO ` ] H O ] ReO H ` ] O
5
2
4
2
2
2
ReO ` ] H O ] ReO H ` ] O
6
2
5 2
]
ReO H ` ] O ] O
4
2
2
2
ReO ` ] H O ] ReO H ` ] O
8
2
7 2
]
ReO H ` ] 2O ] O
35
4
2
2
ReO H ` ] H O ] ReO H ` ] O ] O
100
100
54
46
11
66
23
100
100
7
2
2
5
4
2
ReO H ` ] H O ] ReO H ` ] 2H O
5
4
2
4
2
2
ReO ` ] H ] ReO H `
3
2
3 2
]
ReO ` ] H O
2
2
ReO ` ] H ] ReO H` ] H
4
2
4
]
]
ReO H` ] OH
3
ReO ` ] H O
3
2
ReO H` ] H ] ReO H ` ] H
4
2
4 2
ReO ` ] H ] ReO H ` ] O
5
5
2
3
2
2
1842
Phys. Chem. Chem. Phys., 2001, 3, 1840È1847

View Article Online
Table 3 Reactions with CH
4
Branching
ratio (%)
kabs/
Accuracy
class
Reaction
10~12 cm3 s~1
U (%)
ReO ` ] CH ] ReO CH ` ] H
(30)
(31)
(32)
(33)
(34)
(35)
(36)
(37)
(38)
(39)
100
100
100
58
42
27
40
33
100
100
607
679
20
351
251
150
226
188
1200
504
62
70
2
36
26
15
23
19
100
52
A
B
C
A
A
B
B
B
B
A
2
4
2
2
2
ReO CH ` ] CH ] ReO (CH ) ` ] H
2 2
ReO (CH ) ` ] CH ] ReO (CH ) ` ] H
2
2
4
2
2
2
2
2 2
4
2
2 3
ReO ` ] CH ] ReO CH ` ] H
3
4
3
2
2
]
ReO H ` ] CO
2
4
ReO ` ] CH ] ReO H ` ] CH
4
4
4
2
2
]
]
ReO H` ] CH
4
3
ReO H ` ] CH O
3
2
2
3
2
ReO H` ] CH ] ReO H ` ] CH
4
4
4 2
ReO ` ] CH ] ReO CH ` ] O
5
4
3
4
water complex structure is strongly supported by its highly,
Reactions with CH and methane activation
4
1
00%, efficient formation in the reaction of ReO `, an ion
5
whose propensity for ligand exchange is evidenced in all of its
Methane is an abundant raw material, which is, however,
perhaps of all hydrocarbons the least reactive. Activation of
its bonds, ““C1ÏÏ chemistry, and conversion of methane into
methanol and other compounds is often viewed as one of the
major challenges facing chemistry and chemical catalysis. Effi-
cient formation of higher hydrocarbons from methane would
permit easy use of natural gas as a fuel. This naturally
enhances the interest in the studies of methane chemistry, and
its reactions with metal atoms or ions. Previous studies have
reactions. It should be noted, that the ReO H ` also appears
4
2
fairly prominently among the products of the rhenium oxide
reactions with methane and ethylene, which will be discussed
later.
It is interesting to note that ReO `, besides reacting
4
directly to give the above ReO H ` ion, also yields ReO H`
4
2
4
and ReO H ` products. The neutral product in the formation
of the former must be OH radical, the ion then reacts further
3
2
with a second water molecule yielding again ReO H `, and
concluded that MoO` and MoO` do not react with methane
at all.11 The corresponding osmium oxides were found to par-
tially dehydrate up to three methane molecules, with addition
4
2
2
another OH radical. The latter, ReO H ` ion also appears as
3
2
product in several reactions with hydrogen and methane.
ReO `, ReO ` and ReO ` all exhibit ligand exchange of O
of methylene, CH being also accompanied by a reduction of
5
6
8
2
2
molecule for water, resulting in ReO H `, ReO H ` and
5 2
ReO H `, respectively. The last ion in a two-step process
the metal oxide and elimination of water.10 The observed
chemistry of rhenium oxides with methane observed in the
present work is summarized in Table 3.
4
2
7
2
then yields again ReO H ` as a Ðnal product.
Irikura and Beauchamp in their study observed that
osmium oxide cations are reduced by hydrogen according to
4
2
While the largest, n \ 6 and 8, oxides do not react with
methane all the smaller species studied do. ReO ` is found to
5
simply exchange the O ligand for methane, resulting in a
2
the reaction OsO ` ] H ] OsO
] H O, for n \ 1È3,
n
2
n~1
2
while for n \ 4 the reaction results in OsO H` and a hydro-
ReO CH ` complex. Also the reactions of ReO ` are simple,
4
4
3
4
2
gen atom.10 In the case of ReO ` the ReO H` ion is only a
the oxide dehydrogenates partially three consecutive methane
4
minor product, with loss of OH and formation of ReO H` or
molecules, forming ReO (CH ) `, n \ 1È3 species. The second
3
reduction to ReO ` being the preferred channels. We also
3
2
2 n
step is somewhat faster than the Ðrst, the third one is then
considerably slower. Unfortunately, we could only speculate
about the structure of the products: are there three discrete
methylene groups, or do CÈC bonds form, yielding ethylene,
Ðnd that both ReO ` and ReO ` yield the ReO H ` ion as a
3
5
3 2
major product. The ReO ` reaction appears to be simply a
5
ligand exchange. The efficient stabilization of the small
ReO H ` product ion in high vacuum in a two-body collision
or perhaps a C hydrocarbon? While CID might provide
some insights into the product structures, only optical spec-
3
2
3
between ReO ` and H is interesting. The di†erence between
3
2
the observed chemistry of osmium and rhenium oxide ions is
probably due to their di†erent structures: while the osmium
troscopic studies would provide unambiguous information.
The fragments CH and CH are formed in the reaction of
3
2
ions formed by fragmentation of neutral OsO are
ReO ` with methane. This formation of highly energetic
4
4
undoubtedly covalent species with single oxo ligands, the cold
radical products is undoubtedly made possible by the large
rhenium ions formed in supersonic expansion have peroxidic
structures.
amount of energy which becomes available by the isomer-
ization of the two peroxo ligands. A third parallel channel
Fig. 1 Proposed mechanism for methane activation by Re(O ) `. The primary reaction step is insertion of the metal into the CÈH bond,
2
2
followed by a-hydrogen migration.
Phys. Chem. Chem. Phys., 2001, 3, 1840È1847
1843

View Article Online
results in formaldehyde or, less probably, in the CO ] H
2
neutral molecules. A possible reaction mechanism explaining
all the products formed in a few simple steps is shown sche-
matically in Fig. 1. The key Ðrst step proposed there is reac-
tion of the ion core with methane, and insertion of the metal
into one of the CH bonds. In the next step the hydrogen atom
drifts from the metal to one of the oxygen atoms, resulting in
breakage of one of the O ligands and OH group formation.
2
The CH group can then either be lost as free methyl, or alter-
3
natively can interact with the remaining peroxoligand produc-
ing either CH or formaldehyde products.
2
Particularly puzzling are the two parallel reactions of
methane with ReO `, whose kinetics are shown in Fig. 2.
3
While partial dehydrogenation with a formation of the
ReO CH ` product can easily be understood, the parallel
3
2
reaction leading to oxidation of methane to CO, and resulting
Fig. 2 Kinetics of the reaction of ReO ` with CH . The data unam-
in an ReO H ` ion with uncertain structure is difficult to
3 4
biguously show that both ReO H ` and ReO CH ` are primary
2
4
understand. It is interesting to note also that several reactions
of ethylene, which will be discussed in the next section
produce ions of the same elemental composition.
2
4
3
2
products. In the Ðrst case, the methane molecule is totally dehydroge-
nated, forming a CO molecule as the neutral product.
Reactions with ethylene, and its partial dehydrogenation
Perhaps not surprisingly, of all the compounds studied, ethyl-
ene exhibits the highest reactivity towards the rhenium oxide
cations, as can immediately be seen by examining the
observed reactions summarized in Table 4. It is the only reac-
tant which not only reacts with all of the oxide cations studied
but also with bare Re`, and it produces the richest and most
complex chemistry. The bare ion reacts in sequential steps
with up to seven ethylene molecules, each time with partial
dehydrogenation and formation of products of the
Re(C H ) ` elemental composition. However, only the Ðrst
2
2 n
four reaction steps follow pseudo-Ðrst order kinetics, and only
these rate constants are given in Table 4. This process appears
analogous to the previously studied reactions of Nb` and
W`,9,25 which similarly seem to add up to six and nine C H
2
2
Table 4 Reactions with C H
2
4
Branching
ratio (%)
kabs/
Accuracy
class
Reaction
10~12 cm3 s~1
U (%)
Re` ] C H ] ReC H ` ] H
(40)
(41)
(42)
(43)
(44)
(45)
(46)
(47)
(48)
(49)
(50)
(51)
(52)
(53)
(54)
(55)
(56)
(57)
(58)
(59)
(60)
(61)
(62)
(63)
(64)
(65)
(66)
(67)
(68)
(69)
(70)
(71)
(72)
(73)
(74)
(75)
(76)
(77)
(78)
(79)
100
100
100
100
83
17
10
64
26
100
100
9
91
100
66
34
42
58
69
7
190
658
372
95
801
166
69
442
180
884
207
36
354
53
601
313
159
216
497
50
20
68
38
10
83
17
7
46
19
92
22
4
B
B
C
C
A
A
B
B
B
B
B
C
C
C
A
B
C
C
A
C
C
C
C
C
A
A
A
C
B
B
C
A
B
C
C
A
A
B
C
C
2
4
2
2
2
ReC H ` ] C H ] Re(C H ) ` ] H
2 2
Re(C H ) ` ] C H ] Re(C H ) ` ] H
2 2 2 3
Re(C H ) ` ] C H ] Re(C H ) ` ] H
2 3 2 4
2
2
2
4
2
2
2
2
2
2
4
2
2
2
4
2
ReO ` ] C H ] ReO C H ` ] H
2
2
4
2 2
2
2
]
ReOC H ` ] H O
2
2
2
ReO C H ` ] C H ] ReO (C H ) ` ] H
2
2
2
2
4
2
2
2 2
ReO(C H ) ` ] H O
2 2
ReO C H ` ] C H
2
]
]
2
2
2
2
4
2 2
ReOC H ` ] C H ] ReO(C H ) ` ] H
2 2
ReO (C H ) ` ] C H ] ReO(C H ) ` ] H O
2 2 2 3
ReO(C H ) ` ] C H ] ReO(C H ) CH ` ] CH
2
2
2
4
2
2
2
2
2
4
2
2
2
2 2
2
4
2
2 2
ReO(C H ) ` ] H
2 3
2
2
]
37
6
2
2
2
ReO(C H ) ` ] C H ] ReO(C H ) ` ] H
2
2 3 2 4
2
4
2
ReO ` ] C H ] ReO H ` ] C H
63
33
17
23
52
5
4
7
7
55
44
12
36
2
16
8
2
32
4
3
2
4
3
2
2 2
]
ReO CH ` ] CH O
2
2
2
ReO CH ` ] C H ] ReO C H ` ] H
2 3
ReO C H ` ] CH
2
2
2
4
4
2
]
2
2
2
4
ReO ` ] C H ] ReO H ` ] C H
4
2
4
4
2
2
2
2
]
]
]
]
ReO C H ` ] O
2
2 4
ReO H` ] C H ] OH
6
9
9
41
69
69
3
2 2
ReO ` ] C H O
3
2 4
ReO H ` ] C H ] O
2
2
2
2
2
2
2
ReO H ` ] C H ] ReO C H ` ] H
2 2
100
47
13
40
8
62
30
4
525
414
111
345
19
152
72
15
303
39
2
2
2
4
4
ReO ` ] C H ] ReO C H ` ] O
5
2
4
3 2 4
]
]
ReO H ` ] C H O
4
2
2
2
ReO CH ` ] CO ] O
2
4
2
ReO C H ` ] C H ] ReO C H ` ] H
3
2
4
2
4
3 4
6
2
]
]
ReO C H ` ] H O
2
4
6
2
ReO C H ` ] C H O
2
2
4
2 4
ReO ` ] C H ] ReO H ` ] C H O
6
2
4
5
2
2 2
]
]
]
ReO H ` ] C H ] O
72
9
4
2
2 2
2
ReO CH ` ] CO ] O
3
4
2
ReO H ` ] 2CO
15
100
65
12
23
63
37
65
7
2
4
2
ReO H ` ] C H ] ReO C H ` ] 2H
606
433
82
156
398
234
63
46
9
16
41
24
2
4
2
4
2 2 2
4
ReO ` ] C H ] ReO H ` ] C H ] 2O
8
2
4
4
2
2
2
2
]
]
ReO CH ` ] CO ] 2O
3
4
2
ReO H ` ] 2CO ] O
2
4
2
2
2
ReO H ` ] C H ] ReO C H ` ] H
2 2
2
4
2
4
6
]
ReO C H ` ] 2H
2
2
4
2
1844
Phys. Chem. Chem. Phys., 2001, 3, 1840È1847

View Article Online
Fig. 3 Proposed mechanisms for the possible epoxidation of ethylene. (a) External nucleophilic attack. (b) Initial p-bond and formation of a
Ðve-membered ring. (c) Epoxidation vs. aldehyde formation from the MTO-analogue ReO C H ` through a four-membered ring.
3
2 4
entities, respectively. In spite of the deceptively simple appear-
ance of the reaction scheme, closer examination of the results
suggests that it must in reality be relatively complex.
H O appear most frequently, for the higher oxides, loss of
2
molecular O is more common. However, acetylene, methane,
2
CO, CO , formaldehyde (or CO ] H ), ethylene oxide,
2
2
In the case of Re` we Ðnd that at least three distinct ““ÐnalÏÏ
products which do not seem to react further are formed:
Re(C H ) ` with n \ 4, 5 and 7. Since the formation of the
C H OÈpossibly ketene, as well as O atoms, OH radicals
and methylene are also identiÐed among the reaction pro-
ducts.
2
2
2
2 n
n \ 5 and 7 species must proceed via a n \ 4 intermediate, it
As for the bare Re`, dehydrogenation of ethylene to form
acetylene is also the most common reaction for the oxide reac-
tions. For instance, for ReO `, ReO(C H )` and
is clear that there must be at least two distinct Re(C H ) `
2
2 4
species: the ““ÐnalÏÏ product, and the intermediate which dehy-
drogenates further ethylene molecules. While, in principle,
they could di†er only in their ““temperatureÏÏ, that is in their
energy content, it is more likely, that several structurally dif-
ferent isomers are involved. In investigations of the Ti` cation
reactions in a Ñow reactor it was suggested that poly-
merization of the ethylene takes place.26 Irion and coworkers
have presented evidence for selective formation of benzene
2
2 2
ReO (C H )` are the primary products, with molecular H O
2
2 2
2
and H respectively being eliminated. In contrast with the
atomic ion, where the C H entity always remains attached to
the ion, in the case of a number of oxide reactions the ion is
hydrogenated, with molecular acetylene leaving as a free
2
2
2
gaseous
product.
Thus
the
reaction
ReO `
n
] C H ] ReO H ` ] C H is the major primary channel
2
4
n 2 2 2
upon reaction of ethylene with the Fe ` cations,27 and we in
for both ReO ` and ReO `. For the n \ 6 and 8 oxides the
4
3
4
our laboratory have reported similar C H formation on
W`.25
major primary reaction step is formation of free acetylene, but
in this case molecular oxygen is also eliminated:
6
6
Unlike the reactions of bare Re`, which all exhibit very
ReO ` ] C H ] ReO H ` ] C H ] (n [ 2)O ;
simple stoichiometry, elimination of molecular H , and
2n
2 4
4 2
2 2
2
2
attachment of a C H group to the ion, the reactions of the
(
n \ 3, 4)
2
2
oxide cations are characterized by a considerable complexity
and invariably by branching between a number of parallel
reaction channels. The multitude of products undoubtedly
reÑects the high exothermicity of the processes, which makes
more product channels energetically accessible, so that the
reaction does not have to proceed through the lowest energy
path. The complexity of the reaction is also reÑected in the
variety of neutral molecules appearing between the products.
While for the lower oxides and speciÐcally ReO `, H
Epoxidation reactions and methyltrioxorhenium (MTO)
Two of the observed reactions lead to a product of C H O
2
4
composition, possibly ethylene oxide. In the chemical liter-
ature it has repeatedly been reported and proposed that MTO
can be used as an efficient and selective epoxidation cata-
lyst,12h14 and it is therefore of interest to consider the
2 and
2
Phys. Chem. Chem. Phys., 2001, 3, 1840È1847
1845

View Article Online
observed reactions from this point of view. The ReO ` cation
4
yields C H O as one of the primary, albeit minor, reaction
2
4
products. One of the products of the ReO ` reaction is
ReO (C H )`, presumably the result of an exchange of a
peroxo-O ligand for ethylene. This ion, it could be argued, is
related to MTO, in that it has ethylene in the place of the
5
3
2 4
2
Fig. 4 Formation of formaldehyde through a four-membered ring
from C H and the coordinatively unsaturated ReO ` ion.
methyl ligand. One of the products of the secondary reaction
of this ion with an additional ethylene molecule is then again
C H O, presumably again ethylene oxide.
2
4
3
however, we cannot exclude that the neutral products could
2
4
Several competing mechanisms were proposed for epoxida-
tion by MTO.28 In the Ðrst step of the catalytic cycle MTO
reacts with a suitable oxidiser, for instance hydrogen peroxide,
generating a peroxo group on the metal. In two of three com-
peting proposed mechanisms, the ethylene interacts with
oxygen atoms of the peroxo group, the third mechanism
involves Ðrst binding of the oleÐne on the metal, followed by
its insertion into one ReO bond and formation of the epoxide
by break-up of the cyclic intermediate. The two possibilities,
attack on the oxygen vs. attack on the metal, can easily be
be two separate fragments, CO ] H .
2
4. Conclusions
The reactivity of rhenium oxide cations seems to be strongly
inÑuenced by several factors: (1) Space availability in the
coordination sphere, (2) the ability of additional reactants to
interact with the metal, (3) the energy available by the
rearrangement of a peroxo-ligand to rhenium with two doubly
bonded oxygen ligands, and (4) the activation energy required
transferred to the present case of ReO `, as shown schemati-
4
for this isomerization to occur. In both Re(O )` and
cally in Fig. 3(a) and (b). In aqueous solutions, i.e. catalysis by
2
Re(O ) ` the peroxo substituents leave enough room for
2 2
MTO, preference seems, in general, to be given to the former
mechanisms. The fact that in the gas phase, i.e. this study,
additional ligands to interact with rhenium. The former ion
reacted preferentially by dehydrogenating hydrocarbon reac-
tants. In the latter ion, the OÈO bonds in the peroxo ligands
appear to be weaker, and the barrier to its isomerization is
lowered. This results in high exothermicity of its reactions,
and is reÑected in their high efficiency, and the diversity of
their products.
Additional ligands seem to restrict or prevent direct inter-
action of the reaction partner with the Re` ionic core. Fur-
thermore, part of the available energy is carried away by the
direct epoxidation is detected with the ReO ` ion but not
4
with the higher oxides might be taken as an argument for the
latter mechanism. In the n \ 5 and higher oxides, crowding of
the ligands may make direct attack of ethylene on the metal
difficult, and thus block the epoxidation process. A possible
mechanism for the formation of the C H O product in the
2
4
secondary reaction of ReO `, assuming that the ion formed in
5
the primary reaction contains simply intact ethylene p-bonded
to the rhenium, is also shown in Fig. 3(c). This mechanism is
“
“evaporatingÏÏ excess O ligands, and the overall exother-
similar to that already proposed by Schro
Ž
der et al. for the
2
micity of the reactions is thus reduced. In the case of
reaction of the MTO cation in the gas phase.29 It should,
however, be noted that we have no direct structural evidence
for the product being ethylene oxide, and cannot exclude the
possibility that it is acetaldehyde or another C H O isomer.
Re(O ) `, the coordination shell appears to be completely
2 4
closed, so that the ion cannot exchange ligands and is unre-
active towards both CO and N O. Where reactions do occur,
2
2
4
they are apparently initiated by attack on the peroxo-ligands,
rather than on the metal.
Oxidation of ethylene by the higher rhenium oxide cations
ReO ` is, similar to transition metal monoxides, able to
oxidize hydrocarbon reactants. Of interest is its ability to bind
and coordinate ligands in a two-body high vacuum collision,
without loss of a neutral product, and without an apparent
3
The reactions of the higher oxides are apparently strongly
exothermic, which is, as mentioned above, evidenced by the
multitude of parallel channels, and the abundance of oxygen
may result in complete ““burningÏÏ of the hydrocarbon. Thus
for both ReO ` and ReO ` one of the primary reaction
need for collisional stabilization. In ReO `, or better
5
ReO (O )`, the Ðrst reaction step is invariably exchange of
3
2
6
8
the neutral reactant for the peroxoligand, which is lost and
mostly does not participate in the reaction.
channels completely oxidizes ethylene, yielding apparently
2
CO molecules, and forming an ReO H ` cation product.
2
2 4
This might be formulated as either H Re(OH) ` or
Re(H O) `, but the fact that a secondary reaction step with
an additional ethylene molecule results in dehydrogenation
with no loss of water ligand, but loss of one or two molecules
2
2
References
2
2
1
2
3
K. Eller and H. Schwarz, Chem. Rev., 1991, 91, 1121.
D. C. Parent and S. L. Anderson, Chem. Rev., 1992, 92, 1541.
of H , makes the latter structure quite unlikely.
D. Schro
Ž
der and H. Schwarz, Angew. Chem. Int. Ed. Engl., 1995,
2
34, 1973.
A minor primary channel in the reactions of both ReO `
5
4
5
V. E. Bondybey and M. K. Beyer, J. Phys. Chem. A, 2001, 105,
and ReO ` appears to lead to a neutral product of C H O
6
2 2
951.
elemental composition, possibly ketene. One could imagine
formation of ketene via a mechanism where the acetylene
formed in the dehydrogenation is inserted into an ReÈO bond,
resulting in a cyclic intermediate whose decomposition and
rearrangement then yields ketene. It should, however be noted
that the reaction exothermicity is probably high enough to
M. Beyer, C. Berg, H. W. Gorlitzer, T. Schindler, U. Achatz, G.
Ž
Albert, G. Niedner-Schatteburg and V. E. Bondybey, J. Am.
Chem. Soc., 1996, 118, 7386.
6
7
8
M. Beyer, C. Berg, G. Albert, U. Achatz, S. Joos, G. Niedner-
Schatteburg and V. E. Bondybey, J. Am. Chem. Soc., 1997, 119,
1
466.
C. Berg, U. Achatz, M. Beyer, S. Joos, G. Albert, T. Schindler, G.
Niedner-Schatteburg and V. E. Bondybey, Int. J. Mass Spectrom.
Ion Processes, 1997, 167/168, 723.
allow for the neutral products to be CH ] CO fragments.
2
The situation for the reaction of ReO ` with ethylene is
3
similar, a minor (34%) channel leads to a neutral CH O
2
U. Achatz, S. Joos, C. Berg, T. Schindler, M. Beyer, G. Albert, G.
Niedner-Schatteburg and V. E. Bondybey, J. Am. Chem. Soc.,
product, an elemental composition suggesting formaldehyde.
1
998, 120, 1876.
In analogy to the mechanism proposed in the literature for
formaldehyde formation in the reaction of CrO` with ethyl-
ene,30 one could propose the mechanism shown schematically
in Fig. 4. Here a four-membered ring is formed by addition of
the ethylene across the Re2O double bond, whose rearrange-
ment and break-up then results in formaldehyde. Again,
9
0
C. Berg, T. Schindler, G. Niedner-Schatteburg and V. E. Bond-
ybey, J. Chem. Phys., 1995, 102, 4870.
1
K. K. Irikura and J. L. Beauchamp, J. Am. Chem. Soc., 1989, 111,
7
5.
11 C. J. Cassady and S. W. McElvany, Organometallics, 1992, 11,
2367.
1846
Phys. Chem. Chem. Phys., 2001, 3, 1840È1847

View Article Online
1
1
1
2
3
4
W. A. Herrmann, R. W. Fischer, W. Scherer and M. U. Rauch,
Angew. Chem. Int. Ed. Engl., 1993, 32, 1157.
22 T. Su and M. T. Bowers, J. Chem. Phys., 1973, 58, 3027.
23 K. K. Irikura and J. L. Beauchamp, J. Phys. Chem., 1991, 95,
W. A. Herrmann, R. W. Fischer, M. U. Rauch and W. Scherer, J.
Mol. Catal., 1994, 86, 243.
8344.
24 M. K. Beyer, C. B. Berg, U. Achatz, S. Joos, G. Niedner-
Schatteburg and V. E. Bondybey, Mol. Phys., in press.
25 C. Berg, S. Kaiser, T. Schindler, C. Kronseder, G. Niedner-
Schatteburg and V. E. Bondybey, Chem. Phys. L ett., 1994, 231,
139.
J. Rudolph, K. L. Reddy, J. P. Chiang and K. B. Sharpless, J. Am.
Chem. Soc., 1997, 119, 6189.
H. B. Skinner and A. W. Searcy, J. Phys. Chem., 1973, 77, 1578.
M. A. Grayson, Int. J. Mass. Spectrom. Ion Phys., 1979, 30, 383.
W. J. Bouma, P. D. Burke and K. R. Jennings, Int. J. Mass
Spectrom. Ion Phys., 1980, 36, 117.
J. E. Delmore, J. Phys. Chem., 1987, 91, 2883.
M. Beyer, C. Berg, S. Joos, U. Achatz, W. Hieringer, G. Niedner-
Schatteburg and V. E. Bondybey, Int. J. Mass Spectrom., 1999,
1
1
1
5
6
7
26 B. C. Guo and A. W. Castleman Jr., J. Am. Chem. Soc., 1992,6,
114, 6152.
1
1
8
9
27 P. Schnabel, M. P. Irion and K. G. Weil, J. Phys. Chem., 1991, 95,
9688.
28 A. M. Al-Ajlouni and J. H. Espenson, J. Am. Chem. Soc., 1995,
1
85–187, 625. The level of theory used in the DFT calculations
117, 9243.
was B3LYP/LanL2DZ/D95(d).
29 D. Schroder, W. A. Herrmann, R. W. Fischer and H. Schwarz,
Ž
2
2
0
1
R. L. Summers, NASA Technical Note TN D-5285, National
Aeronautics and Space Administration, Washington, DC, 1969.
T. Schindler, C. Berg, G. Niedner-Schatteburg and V. E. Bondy-
bey, Ber. Bunsen-Ges. Phys. Chem., 1992, 96, 1114.
Int. J. Mass Spectrom. Ion Proc., 1992, 122, 99.
30 H. Kang and J. L. Beauchamp, J. Am. Chem. Soc., 1986, 108,
5663.
Phys. Chem. Chem. Phys., 2001, 3, 1840È1847
1847