Chromium dioxide cation OCrO+ in the gas phase: Structure, electronic states, and the reactivity with hydrogen and hydrocarbons

Article abstract of DOI:10.1021/ja960157k

The potential-energy surface of the triatomic cation [Cr,O₂]⁺ has been examined by means of ion-cyclotron resonance and sector-field mass spectrometry as well as high-level theoretical methods. Chromium(V) dioxide cation OCrO⁺ can be generated by electron ionization of CrO₂Cl₂ and exhibits a doublet ground state. The experimentally determined IE(CrO₂) of 9.7 ± 0.2 eV leads to ΔH(f)(OCrO⁺) = 209 ± 12 kcal/mol which compares well with a theoretical prediction of 217 kcal/mol. The high-valent chromium(V) dioxide OCrO⁺ slowly reacts with H₂ to form CrO⁺ and H₂O as products. Activation of saturated and unsaturated hydrocarbons, including methane and benzene, is much more efficient and involves C-H as well as C-C bond activation. In the reaction of OCrO⁺ with CH₄ the by-product OCr(OCH₂)⁺ points to the operation of a stepwise mechanism in that initially a single oxo unit in OCrO⁺ is activated in terms of a [2 + 2] cycloaddition of methane across the Cr-O double bond. A structurally different [Cr,O₂]⁺ cation can be generated by chemical ionization of a mixture of Cr(CO)₆ and O₂. The experimental findings together with the computations propose that the ions formed consist of a mixture of doublet and quartet states of the dioxide OCrO⁺ and the dioxygen complex Cr(O₂)⁺ (⁶A''); in the latter the dioxygen unit is end-on coordinated to the metal. Due to spin conservation, the direct formation of the doublet ground state OCrO⁺ (²A₁) from the ground state reactants Cr⁺ (⁶S) and O₂ (³Σ(g)^-) is not possible; rather, two curve sings from the sextet via the quartet to the doublet surface in the sequence ⁶Cr⁺ + ³O₂ → ⁶Cr(O₂)⁺ → ⁴OCrO⁺ → ²OCrO⁺ are suggested.

Full text of DOI:10.1021/ja960157k

J. Am. Chem. Soc. 1996, 118, 9941-9952
9941
Chromium Dioxide Cation OCrO+ in the Gas Phase: Structure,
Electronic States, and the Reactivity with Hydrogen and
Hydrocarbons1
Andreas Fiedler, Ilona Kretzschmar, Detlef Schro1der,* and Helmut Schwarz*
Contribution from the Institut fu¨r Organische Chemie der Technischen UniVersita¨t Berlin,
Strasse des 17, Juni 135, D-10623 Berlin, Germany
ReceiVed January 17, 1996X
Abstract: The potential-energy surface of the triatomic cation [Cr,O2]+ has been examined by means of ion-cyclotron
resonance and sector-field mass spectrometry as well as high-level theoretical methods. Chromium(V) dioxide cation
OCrO+ can be generated by electron ionization of CrO2Cl2 and exhibits a doublet ground state. The experimentally
determined IE(CrO2) of 9.7 ( 0.2 eV leads to ∆Hf(OCrO+) ) 209 ( 12 kcal/mol which compares well with a
theoretical prediction of 217 kcal/mol. The high-valent chromium(V) dioxide OCrO+ slowly reacts with H2 to form
CrO+ and H2O as products. Activation of saturated and unsaturated hydrocarbons, including methane and benzene,
is much more efficient and involves C-H as well as C-C bond activation. In the reaction of OCrO+ with CH4 the
by-product OCr(OCH2)+ points to the operation of a stepwise mechanism in that initially a single oxo unit in OCrO+
is activated in terms of a [2 + 2] cycloaddition of methane across the Cr-O double bond. A structurally different
[Cr,O2]+ cation can be generated by chemical ionization of a mixture of Cr(CO)6 and O2. The experimental findings
together with the computations propose that the ions formed consist of a mixture of doublet and quartet states of the
dioxide OCrO+ and the dioxygen complex Cr(O2)+ (6A′′); in the latter the dioxygen unit is end-on coordinated to
the metal. Due to spin conservation, the direct formation of the doublet ground state OCrO+ (2A1) from the ground
state reactants Cr+ (6S) and O2 (3Σg-) is not possible; rather, two curve crossings from the sextet via the quartet to
6
3
6
4
2
the doublet surface in the sequence Cr+ + O2 f Cr(O2)+ f OCrO+ f OCrO+ are suggested.
Introduction
studied by Walba et al. using flow-tube techniques,8 and it was
reported that instead of epoxidation O-atom transfer to the alkene
most likely leads to the thermodynamically more stable acetal-
dehyde. Later, Kang and Beauchamp9 applied ion-beam
techniques in their detailed examination of the reactions of bare
CrO+ with alkanes and alkenes and found that CrO+ behaves
somewhat different as compared to the other monoxide cations
of first-row transition metals.10 Recently, O-atom transfer from
CrO+ to benzene (most likely to yield phenol) has been
examined by ion-cyclotron resonance;11 however, in this study
a contribution of electronically excited CrO+ could not be ruled
out. In addition, the interactions of neutral and charged
chromium atoms with dioxygen in solution,12 in oxygen-doped
rare gas matrices13 as well as in the gas phase,14,15 and also
with respect to theoretical aspects15b,16 have attracted consider-
able attention.
Oxidation reactions with high-valent chromium compounds
are widely applied in organic synthesis.2 Besides classical
reactions, such as the oxidation of alcohols to the corresponding
aldehydes, ketones, or acids, most recently, chromium oxo-
species have been used to bring about stoichiometric C-H bond
activation of alkanes in the gas phase.3,4 Further, high-valent
compounds of the group 6 elements Cr, Mo, and W are
frequently used as catalysts for O-atom transfer from peroxides
to organic substrates, e.g., in epoxidation reactions.5 As to
catalytic oxidations, particularly the use of molecular oxygen
as a terminal oxidant would be chemically attractive, and,
therefore, the generation of high-valent transition-metal oxides
using oxygen is of prime interest.6
Despite the relevance of transition metals in oxidation
processes, the mechanistic details of metal-mediated reactions
in the condensed phase remain often unclear, and gas-phase
experiments can serve as a helpful complement for mechanistic
studies, as they probe directly at a molecular level the elementary
steps.7 The gas-phase oxidation of ethene by CrO2Cl+ has been
(7) For recent reviews of the chemistry of transition-metal ions in the
gas phase, see: (a) Eller, K.; Schwarz, H. Chem. ReV. 1991, 91, 1121. (b)
Eller, K. Coord. Chem. ReV. 1993, 126, 93. (c) Freiser, B. S. Acc. Chem.
Res. 1994, 27, 353.
(8) Walba, D. M.; DePuy, C. H.; Grabowski, J. J.; Bierbaum, V. M.
Organometallics 1984, 3, 498.
(9) Kang, H.; Beauchamp, J. L. J. Am. Chem. Soc. 1986, 108, 5663,
7502.
(10) Schro¨der, D.; Schwarz, H. Angew. Chem., Int. Ed. Engl. 1995, 34,
1973.
(11) Ryan, M. F.; Sto¨ckigt, D.; Schwarz, H. J. Am. Chem. Soc. 1994,
116, 9565.
X Abstract published in AdVance ACS Abstracts, October 1, 1996.
(1) Dedicated to Prof. Jo¨rn Mu¨ller on the occasion of his 60th birthday.
(2) (a) Cainelli, G.; Cardillo, G. Chromium Oxidations in Organic
Chemistry; Springer: Berlin, 1984. (b) Muzart, J. Chem. ReV. 1992, 92,
113.
(3) (a) Cook, G. K.; Mayer, J. M. J. Am. Chem. Soc. 1994, 116, 1855;
1995, 117, 7139. See, also: (b) Gandson, K. A.; Mayer, J. M. Science 1995,
269, 1849.
(12) (a) Scott, S. L.; Bakac, A.; Espenson, J. H. J. Am. Chem. Soc. 1991,
113, 7787. (b) Bakac, A.; Espenson, J. H. Acc. Chem. Res. 1993, 26, 519.
(c) Espenson, J. H.; Bakac, A.; Janni, J. J. Am. Chem. Soc. 1994, 116,
3438. (d) Bakac, A.; Scott, S. L.; Espenson, J. H.; Rodgers, K. R. J. Am.
Chem. Soc. 1995, 117, 6483.
(13) For matrix-isolation spectroscopy of [Cr,O2] and related compounds,
see: (a) Almond, M. J.; Downs, A. J. J. Chem. Soc. Dalton Trans. 1988,
809. (b) Almond, M. J.; Hahne, M. J. Chem Soc., Dalton Trans. 1988,
2255.
(4) Ziegler, T.; Li, J. Organometallics 1995, 14, 214.
(5) (a) Jørgensen, K. A. Chem. ReV. 1989, 89, 431. (b) Jørgensen, K.
A.; Schiott, B. Chem. ReV. 1990, 90, 1483. (c) Murahashi, S.-I. Angew.
Chem., Int. Ed. Engl. 1995, 34, 2616.
(6) (a) Elstner, E. F. Der Sauerstoff: Biochemie, Biologie und Medizin;
Wissenschaftsverlag, Mannheim, 1990. (b) Barton, D. H. R.; Martell, A.
E.; Sawyer, D. T. The ActiVation of Dioxygen and Homogeneous Catalytic
Oxidation; Plenum Press: New York, 1993.
(14) Parnis, J. M.; Mitchell, S. A.; Hackett, P. A. J. Phys. Chem. 1990,
94, 8152.
S0002-7863(96)00157-6 CCC: $12.00 © 1996 American Chemical Society

9942 J. Am. Chem. Soc., Vol. 118, No. 41, 1996
Fiedler et al.
16.5 eV)23 and toluene cation radical (Qmin ) 15.7 eV)21a and applied
the multiplicative correction method.21a All spectra were accumulated
and on-line processed with the AMD/Intectra data system; 5 to 40 scans
were averaged to improve the signal-to-noise ratio. Note that chromyl
chloride is not only Very poisonous but also heaVily oxidizing. Besides
appropriate safety precautions, a pretreating of the inlet system with
ozone or chlorine is indicated in order to achieve maximal ion currents.
In this article, we present a combined experimental and
theoretical study of the [Cr,O2]+ system in the gas phase,
namely, the chromium(V) dioxide cation OCrO+ and the
cationic dioxygen complex Cr(O2)+. Along with the formation,
energetics and electronic states of cationic [Cr,O2]+, we discuss
the reactivity of chromium dioxide cation with hydrocarbons
and the formation of OCrO+ from the ground states of Cr+ and
O2.17,18 Some aspects of the CrO2- anion and the neutral species
as well as the dication will be shortly addressed in the context
of the cationic species.
FTICR.24 These experiments were performed with a Spectrospin
CMS 47X FTICR mass spectrometer. [Cr,O2]+ cations were generated
either by EI of CrO2Cl2 in the external ion source of the instrument or
by an ion/molecule reaction of mass-selected Cr(CO)+ with oxygen
inside the ICR cell; Cr(CO)+ was formed by external EI of Cr(CO)6.
Externally produced ions were transferred into the analyzer cell which
is located within a superconducting magnet (max. field strength 7.05
Tesla).25 Subsequently, the ions of interest were mass-selected,
thermalized by repeated pulses of argon (>1000 collisions), and mass-
selected once more prior to further ion/molecule reactions. Branching
ratios and rate constants were derived from the pseudo first-order
kinetics and are reported with an error of (40%. A detailed report
for the complete procedure of calibration of the pressure measurement
of the ion gauge forms the subject of a forthcoming paper;26 in brief,
we used a thermal rate constant of 8.9 × 10-10 cm3 molecule-1 s-1 for
the reaction of Ar•+ with H2 as the central reference.27 Under FTICR
conditions thermalization of ions generated by EI is much more difficult
to achieve as compared to softer ionization methods as CI, FAB, glow
discharge, or laser desorption. Asides thermalization with pulsed-in
argon, additional quenching with pulsed-in methane (ca. 50 collisions)
turned out to be helpful to remove excited [Cr,O2]+ cations; however,
this procedure is associated with considerable losses of intensity,
because OCrO+ reacts with methane (see below). As a reasonable
compromise between ion intensity and thermalization, quenching with
methane prior to the second mass-selection was only performed in the
most crucial cases, i.e., the reactions of OCrO+ with H2, CH4, and
benzene as well as in the bracketing experiments for determining the
ionization energy. Thermalization was assumed to be complete when
additional collisions with argon or methane did neither effect rate
constants nor branching ratios of the ion/molecule reactions under study.
Reagent gases were introduced into the FTICR cell via leak valves at
typical pressures of 4-40 × 10-9 mbar. As will become obvious
further below, generation of [Cr,O2]+ from Cr(CO)6/O2 leads to a
mixture of different isomers and states of various internal energy
contents, and we refrain from a quantification of branching ratios and
rate constants for these [Cr,O2]+ ions. All data were accumulated and
on-line processed using an ASPECT 3000 minicomputer; depending
on the ion intensity up to 200 scans were averaged. Due to the high
cost of 18O2 (Campro Scientific), this reagent was only used in the
pulsed-valve experiments; therefore, in the comparison of 16O/18O
exchanges with dioxygen (see below), we reacted [Cr,16O2]+ from EI
of chromyl chloride with pulsed-in 18O2, but [Cr,18O2]+ from Cr(CO)+
and pulsed-in 18O2 with leaked-in 16O2.
Experimental and Computational Details
The experiments were performed in a tandem mass-spectrometer
(Sector-MS) and a Fourier transform ion cyclotron resonance (FTICR)
mass spectrometer. Because the instrumentation has been described
in detail previously, we limit ourselves to the essential aspects.
Sector-MS.19 A modified four-sector tandem mass-spectrometer of
BEBE configuration (B stands for magnetic and E for electric sector)
was used, in which MS-I is a VG ZAB-HF-2F and MS-II an AMD
604 double focusing mass spectrometer. [Cr,O2]+ cations were
generated by either electron ionization (EI) of gaseous chromyl chloride
CrO2Cl2 or chemical ionization (CI) of Cr(CO)6 with oxygen as reagent
gas (ratio ca. 1:50); the latter mixture was also used to generate [Cr,O2]-
anions in the negative CI mode. The ions of interest, having 8 keV
translational energy, were mass-selected by means of B(1)/E(1) at a
mass resolution of m/∆m ≈ 4000, and the unimolecular or collision-
induced reactions (collision gas: oxygen at 80% transmission, T)
occurring in the field-free region between E(1) and B(2) were recorded
by scanning B(2). Neutralization-reionization (NR)20 experiments were
performed by colliding B(1)/E(1) mass-selected ions with xenon (80%
T), deflecting the remaining ions by applying a potential of 1 kV,
reionizing the beam of fast neutrals by collision with oxygen (80% T),
and detecting the cationic products by scanning B(2). For the B(1)/
E(1) mass-selected [Cr,O2]- anion, -NR+ was performed using oxygen
in both collision cells (80% T each). Dications were generated by
colliding the monocations with oxygen (50-80% T). The energy
necessary to remove an electron from a fast moving projectile is taken
from its kinetic energy and results in a shift of the dication signal in
the kinetic energy scale from the expected E/2 value to a slightly lower
one; this energy difference is referred to as the Qmin value.21 The Qmin
values were determined by using two different scan modes22 in which
energy scans were done (i) with E(1) for B(1) mass-selected ions and
(ii) with E(2) using B(1)/E(1)/B(2) for mass-selection. The reported
values are averages of several independent measurements, and within
experimental error both methods gave identical results. As references
for the calibration of the energy scale,21a we used Cr+ cation (Qmin
)
Calculations. The ground state and the first excited states of CrO+
have been calculated previously in our group28 using a density functional
theory approach.29 Ab initio MO calculations of the [Cr,O2]+ system
were performed similar to our previous study of the [Fe,O2]+ system.18b
Initially, geometry optimizations of the inserted dioxides 2OCrO+ and
(15) For [Cr,On]- (n ) 1-5), see: (a) Hop, C. E. C. A.; McMahon, T.
B. J. Am. Chem. Soc. 1992, 114, 1237. (b) Hachimi, A.; Poitevin, E.; Krier,
G.; Muller, J. F.; Ruiz-Lopez, M. F. Int. J. Mass Spectrom. Ion Processes
1995, 144, 23.
(16) Schwarz, K.; Sorantin, P. I. Inorg. Chem. 1992, 31, 567.
(17) For a preliminary communication of part of these results, see:
Schro¨der, D.; Fiedler, A.; Herrmann, W. A.; Schwarz, H. Angew. Chem.,
Int. Ed. Engl. 1995, 34, 2517.
(23) Moore, C. E. Atomic Energy LeVels, National Standard Reference
Data Series, National Bureau of Standards, NSRDS-NBS 35, Washington
D.C. 1971.
(24) (a) Eller, K.; Schwarz, H. Int. J. Mass Spectrom. Ion Processes 1989,
93, 243. (b) Eller, K.; Zummack, W.; Schwarz, H. J. Am. Chem. Soc. 1990,
112, 621.
(18) For the related [Fe,O2]+/0/- system, see: (a) Helmer, M.; Plane, J.
M. C. J. Chem. Soc. Faraday Trans. 1994, 90, 395. (b) Schro¨der, D.; Fiedler,
A.; Schwarz, J.; Schwarz, H. Inorg. Chem. 1994, 33, 5094. (c) Fan, J.;
Wang, L.-S. J. Chem. Phys. 1995, 102, 8714.
(19) (a) Srinivas, R.; Su¨lzle, D.; Weiske, T.; Schwarz, H. Int. J. Mass
Spectrom. Ion Processes 1991, 107, 368. (b) Srinivas, R.; Su¨lzle, D.; Koch,
W.; DePuy, C. H.; Schwarz, H. J. Am. Chem. Soc. 1991, 113, 5970.
(20) (a) McLafferty, F. W. Science 1990, 247, 925. (b) Goldberg, N.;
Schwarz, H. Acc. Chem. Res. 1994, 27, 347.
(21) (a) Review: Lammertsma, K.; Schleyer, P. v. R.; Schwarz, H.
Angew. Chem., Int. Ed. Engl. 1989, 28, 1321. (b) For recent applications
in transition-metal chemistry, see: McCullough-Catalano, S.; Lebrilla, C.
B. J. Am. Chem. Soc. 1993, 115, 1441. (c) Dai, P. Q.; McCullough-Catalano,
S.; Bolton, M.; Jones, A. D.; Lebrilla, C. B. Int. J. Mass Spectrom. Ion
Processes 1995, 144, 67.
(25) Technical note: When performing EI in the external ion source of
the Spectrospin CMS 47X FTICR mass spectrometer used here, the
parameter XDFL of the transfer optics should not exceed an absolute value
of (20 V, because this potential is used for deflection of the ions during
mass-selection and the subsequent ion/molecule reactions. Otherwise
externally generated ions may enter the ICR cell during the duty cycle and
produce false results.
(26) Schro¨der, D.; Schwarz, H.; Clemmer, D. E.; Chen, Y.; Armentrout,
P. B.; Baranov, V. I.; Bohme, D. K, Int. J. Mass Spectrom. Ion Processes,
in press.
(27) Anicich, V. G. J. Phys. Chem. Ref. Data 1993, 22, 1469.
(28) Fiedler, A.; Schro¨der, D.; Shaik, S.; Schwarz, H. J. Am. Chem. Soc.
1994, 116, 10734.
(22) Heinemann, C.; Schro¨der, D.; Schwarz, H. J. Phys. Chem. 1995,
99, 16195.

Chromium Dioxide Cation OCrO+ in the Gas Phase
J. Am. Chem. Soc., Vol. 118, No. 41, 1996 9943
6
4OCrO+ as well as the Cr(O2)+ complex were performed using the
approximate density functional theory with the DGAUSS program30
employing the LCGTO-LSD approach (linear combination of gaussian
type orbitals, local spin density); this will be referred to as LSD. All
stationary points were characterized as minima or transition structures
(TS) on the potential-energy surface by evaluating vibrational frequen-
cies and normal modes using analytical first derivatives and a finite
difference scheme for the force-constant matrices. Subsequently, the
low lying roots in all spatial and spin symmetries derived from the
corresponding fragment asymptotes were computed at the CASPT2D
level of theory using atomic natural orbital (ANO) basis sets31
for oxygen (14s9p4d3f)/[5s4p3d]31a and chromium (17s12p9d4f)/
[7s6p4d3f];31b these basis sets will be referred to as BS-I. Then, the
energies of the lowest lying states for all structures were refined at the
CASPT2D level using very large ANO basis sets for oxygen (14s9p4d3f)/
[5s4p3d2f]31a and chromium (21s15p10d6f4g)/[8s7p6d4f2g];31c these
will be referred to as BS-II. In order to confirm the symmetry breaking
in the quartet states of OCrO+ (see below), these geometries were
reoptimized at the CASSCF level of theory using BS-I. Similarly, the
Cr-O-O bond angle of the end-on complex Cr(O2)+ (6A′′), which is
a transition structure with LSD, was scanned in a stepwise manner (10°
steps) using the CASPT2D approach with BS-II, and the end-on
complex turned out as a minimum at this level of theory. The active
molecular orbital space consists mostly of the 4s and 3d orbitals of the
metal atom and the 2p orbitals of oxygen (13 electrons in 12 orbitals).
All electrons were correlated at the CASPT2D level of theory. The
CAS calculations were performed using the MOLCAS (versions 2 and
3) package.32 Bond dissociation energies (BDEs) were derived from
the sum of the energies of the isolated molecules without any further
corrections. All computations were performed on either IBM/RS 6000
workstations or a CRAY-YMP computer.
Figure 1. (a) CA mass spectrum (O2, 80% T) of [Cr,O2]+ obtained by
electron ionization of CrO2Cl2 and (b) NR mass spectrum (Xe 80% T;
O2 80% T).
Results and Discussion
This part is divided into four sections. First, we address the
generation and thermochemistry of chromium dioxide cation.
Then, we examine the structure of [Cr,O2]+ for different spin
multiplicities by means of ab initio MO calculations and assign
the global minimum and electronic ground states of OCrO+ and
Cr(O2)+. Next, the reactivity of the chromium dioxide cation
toward hydrogen and some representative alkanes and alkenes
as well as benzene is presented. Finally, we discuss the possible
generation of other [Cr,O2]+ cations and the dissociation of
[Cr,O2]+ into Cr+ and O2 in the context of a chromium-catalyzed
oxidation of hydrocarbons with dioxygen as terminal oxidant.
In view of the fact that neutral and charged chromium
monoxides have formed the subject of several experimen-
tal9,11,33,34 and theoretical investigations,28,34,35 and neutral and
anionic chromium dioxide have also been examined repeat-
edly,13-16,36 we adopt the energetics and electronic states of these
species from the literature.37
Scheme 1
Generation and Thermochemistry of Chromium Dioxide
Cation. The [Cr,O2]+ cation can be generated by electron
ionization (EI) of chromyl chloride CrO2Cl2. Figure 1 shows
the collisional activation (CA) and the neutralization reionization
(NR) mass spectra of these [Cr,O2]+ ions. Both spectra are in
line with the structure of an inserted chromium(V) dioxide 1
(Scheme 1):18b Upon collisional activation (Figure 1a) sequen-
tial losses of oxygen atoms are observed, i.e., CrO+:Cr+ ≈ 2:1,
•+
while formation of O2 is almost negligible. In addition,
2+
CrOn dications (n ) 0-2) are formed in a charge-stripping
process;21a note, that the signal for CrO22+ is much weaker than
those for CrO2+ and Cr2+ (see below). The NR-spectrum
(Figure 1b) is dominated by signals for CrOn+ (n ) 0-2), and
the intense recovery signal for [Cr,O2]+ indicates that the
corresponding neutral is quite stable toward dissociation.20
(29) (a) Parr, R. G.; Yang, W. Density-Functional Theory of Atoms and
Molecules; Oxford University Press: Oxford, 1989. (b) Ziegler, T. Chem.
ReV. 1991, 91, 651.
(30) Andzelm, J.; Wimmer, E.; Salahub, D. R., Cray Research Inc.:
Pittsburgh, PA, 1991.
(31) (a) Widmark, P.-O.; Malmqvist, P.-A.; Roos, B. O. Theor. Chim.
Acta 1990, 77, 291. (b) Pierloot, K.; Dunez, B.; Widmark, P.-O.; Roos, B.
O., unpublished results. (c) Pou-Amerigo, R.; Merchan, M.; Widmark, P.-
O.; Roos, B. O.; unpublished results.
(32) Andersson, K.; Fu¨lscher, M. P.; Lindh, R.; Malmqvist, P.-Å.; Olsen,
J; Roos, B. O.; Sadlej, A. J. MOLCAS version 2; University of Lund,
Sweden. Widmark, P. O. IBM Sweden, 1992.
(33) Fisher, E. R.; Elkind, J. L.; Clemmer, D. E.; Georgiadis, R.; Loh,
S. K.; Sunderlin, L. S.; Armentrout, P. B. J. Chem. Phys. 1990, 93, 2676.
(34) For photoelectron spectroscopy of CrO+ and the ground-state
assignment, see: (a) Dyke, J. M.; Gravenor, B. W. J.; Lewis, R. A.; Morris,
A. J. Chem. Soc. Faraday Trans. 2 1983, 79, 1083. (b) Harrison, J. F. J.
Phys. Chem. 1986, 90, 3313. (c) Jasien, P. G.; Stevens, W. J. Chem. Phys.
Lett. 1988, 147, 72.
(35) (a) Carter, E. A.; Goddard, W. A., III J. Phys. Chem. 1988, 92,
2109. (b) Shaik, S.; Danovich, D.; Fiedler, A.; Schro¨der, D.; Schwarz, H.
HelV. Chim. Acta 1995, 78, 1393.
•+
Nevertheless, the signal for O2 in Figure 1b indicates that a
certain fraction of the [Cr,O2]+/0 species underwent dissociation
to atomic chromium and dioxygen either in the neutral or the
cationic stage via the dioxygen complexes 2 or 3.
Not surprisingly, the [Cr,O2]+ ions generated by EI (70 eV)
of CrO2Cl2 contain a considerable amount of internal energy,
and proper thermalization is a prerequisite for the examination
(36) For previous experimental work, see: Rudny, E. B.; Sidorov, L.
N.; Kuligina, L. A.; Semenov, G. A. Int. J. Mass Spectrom. Ion Processes
1985, 64, 95.
(37) If not stated otherwise, all thermochemical data were taken from:
Lias, S. G.; Bartmess, J. E.; Liebman, J. F.; Holmes, J. L.; Levin, R. D.;
Mallard, W. G. Gas Phase Ion and Neutral Thermochemistry, J. Phys. Chem.
Ref. Data, Suppl. 1 1988, 17.

9944 J. Am. Chem. Soc., Vol. 118, No. 41, 1996
Fiedler et al.
Table 1. Total Energies (Hartree)a of Different Electronic States
of [Cr,O2]+ Using BS-I and the LSD-Optimized Structures 1 and 2a
of the thermochemistry of [Cr,O2]+ under FTICR conditions.
In order to confirm thorough thermalization of the OCrO+ ions
by multiple collisions with argon and/or methane (see Experi-
mental Section), we determined the ionization energy (IE) of
[Cr,O2] by the bracketing method in that thermalized [Cr,O2]+
is reacted with neutral molecules of known IEs.37 The observa-
tion of an efficient electron transfer from 1,2,3,4-tetrafluoroben-
zene (IE ) 9.5 eV) to [Cr,O2]+ and the absence of electron
transfer in the presence of hexafluorobenzene (IE ) 9.9 eV)
imply that IE(CrO2) ) 9.7 ( 0.2 eV.38 This value is slightly
lower than an older value of IE(CrO2) ) 10.3 ( 0.5 eV as
determined by appearance energy measurements.39 However,
the reasonable agreement within the error bars together with
the fact that our bracketed IE is lower than the previously
determined one confirm that the [Cr,O2]+ ions are sufficiently
thermalized in our experiments. Combining these findings with
the experimental number for the heat of formation (∆Hf) of
neutral CrO2 (-18 ( 10 kcal/mol),40 we arrive at ∆Hf(OCrO+)
) 206 ( 15 kcal/mol, and on the basis of the results of the
ion/molecule reactions which will be described below, this figure
can be further refined to ∆Hf(OCrO+) to 209 ( 12 kcal/mol.
Finally, the practical absence of 16O/18O isotopic exchange of
18O2 with the [Cr,O2]+ cation generated from CrO2Cl2 is in
keeping with a dioxide structure. Moreover, even when the
structure rCr-O(1) rCr-O(2) rO(1)-O(2) state
CASSCF
CASPT2D
1 (C2V)
1.55
1.55
2.51
2A1 -1192.94363 -1193.82319
2B1 -1192.84518 -1193.70362
2B2 -1192.92351 -1193.81571
2A2 -1192.87642 -1193.76548
4A1 -1192.81601 -1193.71668
4B1 -1192.80719 -1193.71598
4B2 -1192.83206 -1193.71676
4A2 -1192.85907 -1193.75827
6A1 -1192.70319 -1193.58776
6B1 -1192.65876 -1193.57995
6B2 -1192.70759 -1193.61124
6A2 -1192.70261 -1193.60764
4A′ -1192.87175 -1193.74844
4A′′ -1192.88169 -1193.77139
4A1 -1192.81967 -1193.66003
4B1 -1192.85590 -1193.72743
4B2 -1192.81291 -1193.65586
4A2 -1192.75986 -1193.59598
6A1 -1192.84087 -1193.71045
6B1 -1192.86617 -1193.69704
6B2 -1192.91705 -1193.77036
6A2 -1192.85636 -1193.71365
8A1 -1192.56307 -1193.45607
8B1 -1192.40694 -1193.27279
8B2 -1192.82806 -1193.70108
1 (Cs)
1.55
1.92
1.66
1.92
2.69
1.29
2 (C2V)
8A2 -1192.43665 -1193.27280
•+
ions were not thermalized at all, electron transfer to yield 18O2
-
O2
1.22
3Σg
-149.76068 -150.12478
prevails, while 16O/18O-exchange was observed for less than
1% of the ions, no matter which thermalization procedures were
applied.
Cr+
6S -1043.05930 -1043.64300
a 1 Hartree ) 627.51 kcal/mol.
As far as the connectivity of [Cr,O2]+ is concerned, we
interpret these findings in that EI of CrO2Cl2 and subsequent
ion thermalization lead to the inserted chromium(V) dioxide
cation 1 in its electronic ground state (see below). In the
following, we will refer to the ion formed from CrO2Cl2 as
chromium dioxide cation OCrO+.
Table 2. Total Energies (Hartree) of the Low-Lying States of
[Cr,O2]+ Using BS-II for the LSD-Optimized Structures 1-3 and
4
the CASSCF/BS-I-Optimized Geometry of OCrO+ (4A′ and A′′)
structure geometry rCr-O(1) rCr-O(2) rO(1)-O(2) state
CASPT2D
1 (C2V)
1 (Cs)
LSD
LSD
1.55
1.55
1.55
2.51
2A1
2B2
4A2
-1194.00127
-1193.99664
-1193.93611
Computational Study of the [Cr,O2]+ System. In order to
examine the structures of conceivable [Cr,O2]+ isomers and the
corresponding electronic ground states, we performed a com-
putational study using a combined density functional and ab
initio MO approach. As far as connectivities are concerned,
the structures 1-3 are feasible, and the spin multiplicities to
be considered may range from doublet, quartet, sextet, up to
octet states.41 In general, geometries were optimized using the
LSD approach, and the energies were then calculated using the
CASPT2D method with the reasonably large basis set BS-I and
refined using BS-II. If not stated otherwise, we will refer to
the CASPT2D/BS-II//LSD results.
1.66
1.81
1.82
1.92
3.08
2.69
3.04
3.07
1.29
1.24
1.22
4A′′ -1193.94846
4A′′ -1193.95048
CASSCFa 1.55
CASSCFa 1.55
4A′
-1193.94423
-1193.94802
2 (C2V)
LSD
LSDb
LSD
1.92
1.84
6B2
3 (C∞V
O2
)
6A′′ c -1193.96295
3Σg
6S
-150.17523
-1043.77258
-
Cr+
a Geometry optimization at the CASSCF/BS-I level of theory. b In
order to evaluate the stability of structure 3, the Cr-O-O bond angle
was reoptimized at the CASPT2D level of theory using BS-II, while
the bond lengths were fixed to the LSD-geometry of end-on complex
Cr(O2)+; see text. c The 6A′′ state transforms to 6B2 in C2V or to either
6
6Φ or Π in C∞V
.
Qualitatively, chromium dioxide cation can be derived by
perfect pairing of CrO+ with O (3P). For CrO+, the 4Σ and 4Π
states are nearly degenerated and only differ in the occupation
3
of mainly nonbonding orbitals.28,34,35 Together with the P
ground state of oxygen, we can expect doublet, quartet, and
sextet states for the inserted OCrO+ cation, and inspection of
Tables 1 and 2 reveals that perfect pairing of CrO+ + O to
doublet states of OCrO+ is energetically favored. LSD geom-
etry optimization in the doublet state results in a C2V-symmetrical
structure for 1 (Figure 2), and the two short Cr-O distances of
(38) In these particular cases the error of the bracketing method is
assumed to be quite low, since no competing reactions occur, i.e., for
tetrafluorobenzene only electron transfer takes place and hexafluorobenzene
Figure 2. Geometries of the dioxides OCrO+ (2A1) and OCrO+ (4A′′)
and the dioxygen complexes Cr(O2)+ (6B2) and Cr(O2)+ (6A′′) calculated
at the LSD/DZP level of theory (bond lengths in Å and bond angles in
degree).
is essentially unreactive toward OCrO.+
.
(39) Grimley, R. T.; Burns, R. P.; Inghram, M. G. J. Chem. Phys. 1961,
34, 664.
(40) Chase, M. W., Jr.; Davies, C. A.; Downey, J. R., Jr.; Frurip, D. J.;
McDonald, R. A.; Syverud, A. N. J. Phys. Chem. Ref. Data 1985, 14
(JANAF Tables), p 937.
(41) (a) For a theoretical study of the dioxides of 4d metals, see:
Siegbahn, P. E. M. J. Phys. Chem. 1993, 97, 9096. (b) See also: Bytheway,
I.; Hall, M. B. Chem. ReV. 1994, 94, 639.
1.55 Å are in line with a description of OCrO+ (2A1) as an
chromium(V) dioxide cation with two Cr-O double bonds. The
driving force for bending to C2V can be attributed to the more
favorable interaction between the oxygen lone pair electrons

Chromium Dioxide Cation OCrO+ in the Gas Phase
J. Am. Chem. Soc., Vol. 118, No. 41, 1996 9945
and the half-filled 3d-orbitals of the metal as compared to a
linear geometry.41
in which two electrons, one of Cr and one of O2, undergo spin
coupling. States of other spatial symmetries as well as the octet
and quartet states are much higher in energy and can therefore
not account for complex formation between Cr+ and O2.
However, at the CASPT2D/BS-I//LSD level of theory even for
Cr(O2)+ (6B2) the dissociation energy of the Cr+-O2 bond is
calculated to be as small as 1.7 kcal/mol and decreases to only
0.1 kcal/mol with BS-II (Table 2), a result which is not
reasonable for the complexation of the polarizable oxygen
molecule by a metal cation. The small imaginary frequency of
the end-on structure 3 indicates a flat potential-energy surface
for the complex and led us to examine the end-on structure at
a higher level of theory. Actually, using the CASPT2D/BS-
II//LSD approach the end-on complex 3 is calculated to be 9.5
kcal/mol more stable than the side-on species 2. In addition, a
scan of the Cr-O-O bond angle at this level of theory reveals
the linear end-on complex 3 as the minimum, even though the
Inserted chromium dioxide cation 1 in the low-spin coupled
2
doublet electronic ground state A1 corresponds to the global
minimum of the [Cr,O2]+ potential-energy surface. However,
the 2B2 state is located only slightly higher in energy, such that
the accuracy of our computational approach does not allow for
a definitive assignment of the ground state symmetry. For the
sake of simplicity, in the following we refer to 2A1 as the ground
state of OCrO+. The stability of OCrO+ (2A1) with respect to
dissociation into ground state Cr+ (6S) and O2 (3Σg-) is
calculated as 33.6 kcal/mol. Combining this figure with the
experimental heats of formation of Cr+ and O2, we arrive at a
theoretically predicted ∆Hf (OCrO+) of 217 kcal/mol, which is
in good agreement with the experimental figure of 209 ( 12
kcal/mol derived above.
The quartet and sextet states of chromium dioxide cation are
6
potential of the A′′ surface is quite flat with respect to the
4
much higher in energy. The lowest quartet state, A′′, is 33.2
bending mode. These results can be rationalized by considering
the electronic prerequisites: Ground state Cr+ (6S) has a 3d5
high-spin configuration and an empty 4s shell. Formation of a
covalent bond is most favored through spin-pairing of a Cr+
3d electron with the in-plane π* electron of side-on dioxygen
leading to the sextet Cr(O2)+ with structure 2. However, the
small radial extent of the compact 3d shell results in an small
overlap with O2 and the large exchange-energy losses upon
distortion of the half filled high-spin 3d5 configuration of Cr+
(6S) disfavor this type of bonding. Consequently, simple
complex formation by donation of charge from the O2 ligand
to Cr+ favors the end-on structure 3 due to the larger polariz-
ability of dioxygen along the O-O bonding axes. Actually,
both types of bonding are involved to some extent in 6Cr(O2)+
as indicated by (i) the sextet multiplicity and (ii) the relatively
weakly bound end-on structure with a small bending force
constant. In fact, as the LSD level of theory is known to
underestimate exchange-energy losses and to overestimate the
covalent contributions,43 in a borderline case like 6Cr(O2)+ the
wrong side-on structure is predicted to correspond to the
minimum, in contradiction to the CASPT2D results. Thus, even
though LSD geometries are usually reasonable starting points
for energy calculations, caution is indicated in problematic cases
bearing low-lying frequencies.
kcal/mol less stable than OCrO+ (2A1) and 0.4 kcal/mol below
the Cr+ + O2 asymptote; all sextet states are very high in energy
and were not pursued any further. The origin of the stability
difference between the doublet and quartet states of OCrO+
arises from the breaking of one double bond by decoupling one
electron pair. Qualitatively, the energy difference between these
states can be traced back to the difference between ionization
of the triplet ground state14 of neutral CrO2 in a bonding
π-orbital of a Cr-O double bond and removal of a nonbonding
electron from the metal center. Interestingly, the LSD-optimized
geometry of OCrO+ (4A′′) is asymmetric with two different
Cr-O bond lengths, a typical short one of 1.55 Å and a long
one of 1.66 Å. In terms of molecular orbital theory, symmetry
breaking from C2V to Cs is somewhat unexpected, because in
the C2V point group there exist no degenerate orbitals which
could give rise to spontaneous symmetry breaking, e.g., Jahn-
Teller or Rener-Teller effects. Although we cannot exclude that
the asymmetry of OCrO+ (4A′′) is an artifact of our computa-
tional approach, it has also been argued42 that excited SO2
undergoes the transition C2V f Cs. Notwithstanding these
arguments, the Cs structure of OCrO+ (4A′′) is perfectly in line
with the valence bond description of this isomer by localized
single and double bonds. In order to verify that the LSD
computed geometries are not an artifact of the single reference
approach, these quartet states were also geometry optimized
using CASSCF/BS-I. At this level of theory, again bent Cs-
symmetrical structures were obtained, and the differences of
the Cr-O bond lengths are even more pronounced, i.e., ∆rCr-O
) 0.26 Å for both states. Using these geometries for CASPT2D/
BS-II calculations has only minor influence on the computed
energetics in that OCrO+ (4A′′) is 1.7 kcal/mol lower in energy
than separated Cr+ and O2 and still favored over OCrO+ (4A′).
Reactivity of Ground State OCrO+. Formally, chromium
2
dioxide cation OCrO+ represents a high-valent chromium(V)
compound. Therefore, it is not unexpected that the thermalized
OCrO+ cation generated by EI of CrO2Cl2 oxidizes hydrocar-
bons quite efficiently. In fact, under FTICR conditions even
dihydrogen is activated by OCrO+ to yield CrO+ and water
(reaction 1).
OCrO+ + H2 f CrO+ + H2O
(1)
Quite a different order of stabilities for the various spin
multiplicities is found for the side-on and end-on Cr(O2)+
complexes (structures 2 and 3). In general, ground state
chromium cation (6S) and ground state dioxygen (3Σg-) can give
rise to octet, sextet, and quartet states for Cr(O2)+. The LSD
geometry optimization for the sextet complex reveals a side-on
minimum with structure 2, whereas the end-on structure 3
displays an imaginary frequency of i143 cm-1 for bending of
the O2 unit. The energetically lowest-lying state is Cr(O2)+ (6B2)
According to the thermochemistry of 2OCrO+ outlined above,
the exothermicity of reaction 1 amounts to ca. 43 kcal/mol.
However, the experimentally measured rate constant of kR )
9.8 × 10-12 cm3 molecule-1 s-1 is rather low as compared to
the capture rate constant kC ) 1.5 × 10-9 cm3 molecule-1 s-1
as derived from gas kinetic theory.44 Thus, the reaction
efficiency φ (defined as φ ) kR/kC) amounts to only about 0.007
for the reaction of OCrO+ with hydrogen and 0.004 for
deuterium, respectively. In order to ensure ion thermalization,
argon and methane were pulsed-in repeatedly prior to mass-
selection of OCrO+, and its reaction with hydrogen was followed
(42) (a) Herzberg, G. Molecular Spectra and Molecular Structure; Krieger,
Malabar, 1991; Vol. 3, p 605. (b) For a related example of Cs symmetry in
the [Cu,O2] system, see: Bauschlicher, C. W., Jr.; Langhoff, S. R.; Partridge,
H.; Sodupe, M. J. Phys. Chem. 1993, 97, 856. (c) See also: Wu, H.; Desai,
S. R.; Wang, L.-S. J. Chem. Phys. 1995, 103, 4363. (d) Also the excited
quartet state W(O)2+ (4A′) has Cs-symmetry: Ferhati, A.; Ohanessian, G.,
Paris, personal communication.
(43) Ziegler, T.; Li, J. Can. J. Chem. 1994, 72, 783.
(44) Su, T.; Bowers, M. T. Int. J. Mass Spectrom. Ion Phys. 1973, 12,
347.

9946 J. Am. Chem. Soc., Vol. 118, No. 41, 1996
Fiedler et al.
Scheme 2
Scheme 3
quartet ground state in analogy to Cr(CH3)2+ (4B2).49 Thus, if
spin-inversion takes place in reaction 1, it is likely to occur at
until almost complete conversion (i.e., >99%) in the presence
of a 4-fold excess of argon as a buffer gas. In these experiments,
we found no deviation from strict first-order kinetics, even
though at high conversions of OCrO+ also many thermalizing
collisions with hydrogen occurred prior to each reactive one.
As discussed in a previous publication,45 for reactive metal-
oxide cations the determination of very low rate constants can
be associated with experimental errors, in particular with respect
to the purity of the reagents and background contaminants. In
fact, we observe that (i) the CrO+ cation formed in reaction 1
is slowly converted into Cr+, although thermalized CrO+ does
not react with H2,9,46 and (ii) small amounts of other products
appear at long reaction times, e.g. 1% [Cr,C2,H4,O]+ at 99.5%
conversion of the OCrO+ precursor. Thus, we conclude that
background hydrocarbons are present in the vacuum system,
and, therefore, the rate constant of 9.8 × 10-12 cm3 molecule-1
s-1 for reaction 1 as determined from the disappearance of
OCrO+ can be slightly too high. Further, these sources of error
as well as the accuracy of the pressure measurement26 do not
allow for a precise quantitative analysis of the different
efficiencies for H2 and D2 in terms of an intermolecular kinetic
isotope effect, except for the qualitative finding that the isotope
effect is quite small.47 Despite these uncertainties, the fact that
reaction 1 occurs for thermalized ions is beyond any doubt.
The ion/molecule reaction of OCrO+ with hydrogen is also
of interest from a fundamental point of view: As derived above,
OCrO+ exhibits a doublet ground state. However, the quoted
enthalpy of reaction 1 would require that CrO+ is formed in its
quartet ground state (either 4Σ+ or 4Π),34 while the spin-allowed
formation of CrO+ (2∆) is ca. 30 kcal/mol less exothermic.35b
Though the spin-allowed formation of CrO+ (2∆) would still
be thermochemically possible, the low efficiency of reaction 1
and the small kinetic isotope effect imply that the reaction
involves curve crossing from the doublet to the quartet surface
via spin-orbit coupling as a central step in the reaction coordinate
en route to product formation.35b This analysis is analogous to
the recently studied oxidation of molecular hydrogen by bare
FeO+ cation.26,28,47,48
2
2
+
4
+
this point, i.e., OCrO+ + H2 f Cr(OH)2 f Cr(OH)2
f
4
4CrO+ + H2O. Subsequently, isomerization leads to (H2O)-
CrO+ and finally spin-allowed dissociation into the ground state
4
products CrO+ (4Σ+ or Π) and H2O (1A1).50
As compared to reaction 1, the oxidation of methane with
OCrO+ is more efficient (φ ) 0.12), i.e., kR ) 1.2 × 10-10 cm3
molecule-1 s-1, kC ) 1.0 × 10-9 cm3 molecule-1 s-1. The
OCrO+/CH4 couple necessarily involves spin inversion en route
to product formation, because most of the products (Scheme 3)
have ground states of higher multiplicities with rather large
excitation energies to the doublet surface, e.g. 3.8 eV for the
excitation Cr+ (6S) f Cr+ (2D). The losses of neutral H2O,
CH2O, and “CH4O2” (most probably in form of H2O + CH2O)
are not surprising and can be explained to proceed via the
bisligand complex (H2O)Cr(OCH2)+ as a common intermediate
(see below). The key product is the loss of dihydrogen to yield
the [Cr,C,H2,O2]+ product ion. Because previous studies have
demonstrated that formaldehyde may undergo Cr+-mediated
dehydrogenation to the corresponding carbonyl complex,51a
one is tempted to assign an analogous (H2O)Cr(CO)+ structure
to this product ion and attribute its formation to a subsequent
degradation of the formaldehyde ligand in (H2O)Cr(OCH2)+.
However, the secondary reaction of the [Cr,C,H2,O2]+ product
ion with background water rules out this conjecture in that
formation of [Cr,H2,O2]+ takes place; a process which formally
represents an exchange reaction in which water substitutes a
formaldehyde ligand in [Cr,C,H2,O2]+. Further, irrespective of
the isotopic composition of the [Cr,C,H2-n,Dn,O2]+ ions (n )
0-2) generated in the reactions of OCrO+ with [D2]- and [D4]-
methane, [Cr,H2,O2]+ is the only secondary product and the
complete label is lost in the neutral. These findings suggest
that the [Cr,C,H2,O2]+ ion corresponds to a formaldehyde
complex of chromium oxide, i.e., OCr(OCH2)+. At first sight,
the formation of this ion is quite surprising, because it combines
a reactive metal-oxo moiety with an oxidizable aldehyde ligand.
However, similar reactions of only one of two oxo-ligands have
+
With respect to the reaction mechanism, two different
scenarios are conceivable (Scheme 2): After formation of the
adduct complex [OCrO+ • H2], either addition of the H-H unit
across a single Cr-O unit ([2 + 2] pathway) or a simultaneous
reaction involving both oxygen termini ([3 + 2] pathway) can
occur.4 The [2 + 2] pathway leads to OCr(H)(OH)+ which
can then undergo H-migration from the metal to the oxygen
been previously proposed in the reactions of OsO2 with
methane and CrO2Cl+ with ethylene.8,52 Finally, the very minor
product channel to yield CrO+ and methanol is of interest for
thermochemical reasons, because its occurrence can be used
for the refinement of the ∆Hf(OCrO+) figure as mentioned
above.
It is now well-accepted10 that in the C-H bond activation of
an alkane R-H by a reactive metal oxide [M]O ([M] ) ligated
or bare transition metal) the formation of an O-H bond provides
the driving force for further reactions such that R is initially
attached to [M]. This conjecture is also in line with simple
thermochemical considerations, since the strength of an O-H
bond exceeds that of an O-C bond, and vice versa, for
coordinatively unsaturated species often [M]-C bonds are
+
atom to yield Cr(OH)2 or to the hydroxy group to form the
oxo complex (H2O)CrO+. The chromium dihydroxide cation
+
Cr(OH)2 may serve as an intermediate in both pathways and
most probably, this formal chromium(III) compound exhibits a
(45) Ryan, M. F.; Fiedler, A.; Schro¨der, D.; Schwarz, H. Organometallics
1994, 13, 4072.
(46) An alternative explanation for the subsequent formation of Cr+ may
assume that CrO+ is formed in an excited state, e.g. CrO+ (2∆), which
reacts with hydrogen. Alternatively, due to the large reaction exothermicity,
the CrO+ ions may be formed in excited rovibronic states. As we cannot
discriminate between these two possibilities, we have not further attempted
to resolve this issue.
(49) Rosi, M.; Bauschlicher, C. W., Jr.; Langhoff, S. R.; Partridge, H. J.
Phys. Chem. 1990, 94, 8656.
(50) For the reaction of H2O2 with Cr+, see: Schalley, C. A.; Wesendrup,
R.; Schro¨der, D.; Schwarz, H. Organometallics 1996, 15, 678.
(51) (a) Schalley, C. A.; Wesendrup, R.; Schro¨der, D.; Schwarz, H. J.
Am. Chem. Soc. 1995, 117, 7711. (b) Wesendrup, R.; Schalley, C. A.;
Schro¨der, D.; Schwarz, H. Chem. Eur. J. 1995, 1, 608.
(47) Schro¨der, D.; Fiedler, A.; Ryan, M. F.; Schwarz, H. J. Chem. Phys.
1994, 98, 68.
(48) Clemmer, D. E.; Chen, Y.-M.; Khan, F. A.; Armentrout, P. B. J.
Phys. Chem. 1994, 98, 6522.
(52) Irikura, K. K.; Beauchamp, J. L. J. Am. Chem. Soc. 1989, 111, 75.

Chromium Dioxide Cation OCrO+ in the Gas Phase
J. Am. Chem. Soc., Vol. 118, No. 41, 1996 9947
Scheme 4
stronger than [M]-H bonds.37,53 Thus, an addition of R-H
across an [M]-O bond is much more likely to yield an insertion
intermediate R-[M]-OH than the isomeric H-[M]-OR spe-
cies; both can then undergo either reductive elimination of the
corresponding alcohol ROH or further reactions. The formation
of OCr(OCH2)+ from OCrO+/CH4 is also telling with regard
to the reaction mechanism: In fact, dehydrogenation is hardly
conceivable involving an initial formation of H3C-Cr(O)-OH+,
4, (Scheme 4), because once a (strong) O-H bond is formed,
it is unlikely to be broken again in favor of the dehydrogenation
process. However, when C-H bond activation of methane by
OCrO+ involves H-Cr(O)-OCH3+, 5, as the first intermediate,
â-hydrogen transfer from the methoxy group together with loss
of molecular hydrogen becomes a feasible option.7a,54,55 While
this type of addition has rarely been observed for transition-
metal monoxide cations,10 a similar reversal of the regioselec-
tivity for R-H addition across the Fe-O bond has been reported
in the reaction of the formal iron(IV) compound OFeOH+ with
methane.56 A different behavior of OCrO+ as compared to other
transition-metal monoxide cations is also in keeping with the
observation that the intramolecular kinetic isotope effect as-
sociated with C-H/D bond activation of CH2D2 is rather small
(kH/kD ≈ 1.6)57 as compared to cationic transition-metal
monoxides.10 Further, the involvement of 5 as an intermediate
suggests that, at least in part, bond activation of methane by
OCrO+ involves only a single O-atom in terms of a [2 + 2]
cycloaddition, and similar arguments apply for the formation
of CrO+. There also exists an electronic reasoning for the [2
+ 2] pathway in the reactions of 2OCrO+. The initial insertion
products 4 and 5 are likely to exhibit doublet ground states,
because four of the five 3d electrons of Cr+ are involved in
covalent bonds such that one uncoupled electron is left behind.
Thus, the [2 + 2] route conserves the spin of the reactants. In
contrast, the [3 + 2] pathway leads directly to 6 which is likely
to exhibit a quartet ground state. Hence, in this route the bond
formation and curve crossing must occur simultaneously along
the reaction coordinate. Finally, the formaldehyde ligand in
OCr(CH2O)+ may lower the splitting of 1.2 eV35b between the
doublet and quartet states of CrO+, such that the OCr(OCH2)+
Scheme 5
product might retain doublet spin. As a consequence, spin-
conservation would allow for the formation of OCr(CH2O)+,
while the other product channels have to circumvent the spin-
inversion bottleneck; while this is less effective, it certainly takes
place, e.g. for thermochemical reasons the formation of Cr+
(2D) is impossible.
The other reaction products can be formed by either CH3-
transfer from the metal to the oxygen (4 f 6), likewise via
H-transfer (5 f 6), or sthough less likelysby an addition of
the R-H bond to the oxygen termini of OCrO+ in terms of a
[3 + 2] cycloaddition which leads directly to 6. The latter ion
6 then serves as an intermediate which can undergo hydrogen
transfer from the methoxide unit to the hydroxy ligand (either
via the metal or directly to the oxygen atom)55 to yield the
bisligated complex (H2O)Cr(OCH2)+ 7 from which formalde-
hyde and/or water are lost eventually.
Not surprisingly, the reaction efficiencies φi of OCrO+ with
higher alkanes increase as compared to those for dihydrogen
and methane (i.e., φ (C2H6) ) 0.5, φ (C3H8) ) 0.7, φ (i-C4H10)
) 0.6). In general, the product patterns of the ion/molecule
reactions of OCrO+ with larger alkanes are more complex. This
is due to the fact that the more C-H bonds and possible
intermediates become available, the more reaction pathways
compete with each other.10 For the sake of clarity, we only
discuss those pathways in some more detail in which principally
new features or reaction mechanisms evolve. Further, not all
product ions were structurally characterized, and the formulas
given should only be viewed as reasonable suggestions based
on similar reactions of other transition-metal ions.7,10 Let us
illustrate the decrease in selectivity by discussing the reaction
of OCrO+ with ethane (Scheme 5): (i) The losses of H2, H2O,
C2H4O, C2H6O, and “C2H6O2” are analogous to the processes
observed in the OCrO+/CH4 couple and can be explained by
invoking the higher homologues of 4-7. (ii) Homolytic bond
(53) Martinho Simo˜es, J. A.; Beauchamp J. L. Chem. ReV. 1990, 90,
629.
(54) (a) Carlin, T. J.; Sallans, L.; Cassady, C. J.; Jacobson, D. B.; Freiser,
B. S. J. Am. Chem. Soc. 1983, 105, 6320. (b) Cassady, C. J.; Freiser, B. S.;
McElvany, S. W.; Allison, J. J. Am. Chem. Soc. 1984, 106, 6125. (c) Fiedler,
A.; Schro¨der, D.; Schwarz, H.; Tjelta, B. L.; Armentrout, P. B. J. Am. Chem.
Soc., 1996, 118, 5047.
•
cleavages lead to the eliminations of H• as well as CH3 , most
(55) Most likely, the hydrogen atom from the methoxy group is directly
transferred to the other hydrogen in 5 without involving a dihydrido-metal
intermediate; see: (a) Holthausen, M. C.; Fiedler, A.; Schwarz, H.; Koch,
W. Angew. Chem., Int. Ed. Engl. 1995, 34, 2282. (b) Holthausen, M. C.;
Fiedler, A.; Schwarz, H.; Koch, W. J. Phys. Chem. 1996, 100, 6236.
(56) Schro¨der, D.; Schwarz, H. Angew. Chem., Int. Ed. Engl. 1991, 30,
991.
likely from the intermediate [HO-Cr-OC2H5]+ 10, to form
the corresponding aldehyde complexes of CrOH+. (iii) A new
pathway which is not possible in the reactions of H2 and CH4
leads to the generation of Cr(OH)2+ concomitant with the loss
of neutral alkene. The latter channel can proceed via initial
C-H bond activation of ethane to yield the insertion intermedi-
ate 8, subsequent transfer of a â-hydrogen atom from the ethyl
group to the OCrOH unit (8 f 9) and terminating loss of the
ethene ligand (Scheme 6). Alternatively, 10 may undergo C-O
bond cleavage and electron transfer (ET) in the complex to an
(57) The isotope effect was determined for the OCrO+/CH2D2 couple
as an average of the isotopologous products formed according to Scheme
3. A precise evaluation of the individual effects for losses of water and
formaldehyde was hampered by (i) rapid H/D-exchange of Cr(OD2)+ and
Cr(ODH)+ with background water and (ii) accidental mass-overlap of Cr-
(OCD2)+ with the OCrO+ precursor ion.

9948 J. Am. Chem. Soc., Vol. 118, No. 41, 1996
Fiedler et al.
Scheme 6
Scheme 8
Scheme 7
mechanisms proposed for the activation of methane and ethane.
The reduced complexity of the product distribution as compared
to the OCrO+/C2H6 couple does, however, not imply an increase
of intrinsic selectivity, rather, it is a consequence of the
predominating channels leading to either Cr(OH)2+ or carboca-
tions. Finally, the formation of C3H7+ in the reaction of OCrO+
with isobutane indicates that besides formal H--transfer from
the alkane to OCrO+ also H3C--transfer is thermochemically
allowed.
In order to examine the reactivity of OCrO+ with unsaturated
hydrocarbons, we have chosen ethene, propene, and benzene
as model compounds. In the reaction of ethene with OCrO+
(φ ) 0.5) three main products are formed and two minor
intermediate that formally contains a carbocation (10 f 11)
and which can lead to Cr(OH)2 via proton transfer; note that
+
proton transfer from Cr(OH)2+ to ethylene (proton affinity, PA
) 163 kcal/mol)58 to yield neutral OCrOH and an ethyl cation
is not observed experimentally. In the context of a recent study
of the related ion/molecule reaction of Cr+ with dimethylper-
oxide,51 we argued that the absence of C1-products such as losses
of CHnO (n ) 2-4) indicates that direct C-C bond cleavage
in terms of a [3 + 2] cycloaddition to yield the insertion
intermediate (H3CO)2Cr+ does not occur in the OCrO+/C2H6
couple. In contrast, a bismethoxy chromium cation constitutes
the central intermediate en route to the Cr+-mediated decom-
position of dimethylperoxide. Thus, we conclude that OCrO+
exhibits a strong preference for C-H bond activation of alkanes
and that C-C bond activation via the [3 + 2] route does not
take place.
•
channels (<5%) lead to losses of CH3 and CH3CO• radicals
(Scheme 8). The major reaction products can be rationalized
by initial attack of OCrO+ to the C-C double bond in terms of
a [2 + 2] cycloaddition to form the metallaoxetane 12 (Scheme
9).5b Subsequently, the primary intermediate 12 can either
rearrange to the five-membered ring 13 which then undergoes
loss of one or two formaldehyde units.60 In competition, a
metal-assisted 1,2-hydrogen migration5b,9,61 may lead to the
acetaldehyde/CrO+ complex 14, which subsequently splits off
•
CH3 , CH3CO•, and C2H4O; the latter most likely corresponds
to acetaldehyde. Although we cannot rigorously exclude that
a [3 + 2] cycloaddition leads directly to 13, the generation of
methyl and acetyl radicals provides strong evidence for the
operation of a [2 + 2] mechanism, because the [3 + 2] pathway
should lead to almost exclusive elimination of one or two
formaldehyde molecules from 13 without any further rearrange-
ment prior to fragmentation. Further, the preference for the [2
+ 2] type of addition is also in accord with the findings outlined
above for the reactions of the alkanes.62 The reaction of OCrO+
with propene occurs close to the collision limit (φ ) 0.8), and
a large variety of products is formed which we have not
attempted to characterize further. The complexity of reaction
products can be attributed to competition between activation
of the C-C double bond of the alkene and C-H bond activation
in the allylic position; the latter process is particularly demon-
strated by formal hydride transfer from propene to OCrO+ to
yield the allyl cation concomitant with neutral OCrOH. In fact,
the decrease in selectivity in going from ethene to propene
reveals a general dilemma in the oxidation of substituted alkenes
In the OCrO+/C3H8 couple (Scheme 7) most of the pathways
observed for OCrO+/C2H6 are by and large suppressed in favor
of two major product channels to yield Cr(OH)2+, together with
neutral C3H6, and C3H7+ concomitant with neutral OCrOH. As
to the reaction mechanism the formation of C3H7+ can be traced
back to the stability of the isopropyl cation. Due to the similar
+
+
intensities of Cr(OH)2 and C3H7 as well as with respect to
the fact that these two channels are conceivably connected via
a proton transfer, we describe the product distribution in terms
+
of an acid-base equilibrium between Cr(OH)2 + C3H6 and
+ 59
OCrOH + C3H7
.
To a first approximation, we estimate that
the PA of neutral oxo-chromium hydroxide OCrOH is similar
to that of propene (PA ) 180 kcal/mol). This interpretation is
substantiated by the reaction of OCrO+ with isobutane, in which
the formation of tert-butyl cation and neutral OCrOH prevails
+
by and large, while the complementary products Cr(OH)2
+
C4H8 are much less abundant. This finding agrees well with
the proposal that PA(OCrOH) is smaller than that of isobutene
(PA ) 196 kcal/mol). The other products in the reactions of
OCrO+ with propane and isobutane are in keeping with the
(60) We cannot exclude that CH2O loss leads to OCr(CH2)+ instead of
the proposed structure Cr(OCH2)+ and have not attempted to further
characterize this product.
(61) (a) Fisher, E. R.; Armentrout, P. B. J. Phys. Chem. 1990, 94, 1674.
(b) Schro¨der, D.; Schwarz, H. Angew. Chem., Int. Ed. Engl. 1990, 29, 1431.
(62) For recent discussions of [2 + 2] versus [3 + 2] pathways in the
reactions of osmium(VIII) compounds with alkenes, see: (a) Veldkamp,
A.; Frenking, G. J. Am. Chem. Soc. 1994, 116, 4937. (b) Norrby, P.-O.;
Kolb, H. C.; Sharpless, K. B. J. Am. Chem. Soc. 1994, 116, 8470.
(58) Proton affinities were taken from: Lias, S. G.; Liebman, J. F.; Levin,
R. D. J. Phys. Chem. Ref. Data 1984, 13, 695.
(59) For a related behavior of Fe+ complexes, see: Schro¨der, D.;
Wesendrup, R.; Schalley, C. A.; Zummack, W.; Schwarz, H. HelV. Chim.
Acta 1996, 79, 123.

Chromium Dioxide Cation OCrO+ in the Gas Phase
J. Am. Chem. Soc., Vol. 118, No. 41, 1996 9949
Scheme 9
Scheme 10
Scheme 11
by oxidants on the basis of high-valent transition-metal oxides,
in that the latter are often too reactive to allow for a discrimina-
tion between activation of a C-C double bond and allylic C-H
bonds.63 Obviously, this lack of reactivity also holds true for
OCrO+, and only appropriate tuning of the reactivity by means
of ligand effects may bring about more selective pathways for
the oxidation of alkenes.64
Figure 3. Calculated potential-energy surface for [Cr,O2]+ at the
CASPT2D/BS-II//LSD/DZP level of theory. The exit channels leading
In the reaction of OCrO+ with benzene (φ ) 1.0) dehydration
to yield [Cr,C6,H4,O]+ dominates, while formation of Cr+, CrO+,
and electron transfer from benzene (IE ) 9.25 eV) to OCrO+
can hardly compete (Scheme 10). Considering the IE(CrO2)
) 9.7 ( 0.2 eV as derived above, the low intensity of electron
transfer with benzene demonstrates once more that C-H bond
activation by OCrO+ is very efficient and can almost suppress
electron transfer. We can merely speculate about the reaction
mechanism,65 but the intense dehydration process demonstrates
that OCrO+ can bring about the C-H bond activation of
benzene, although C-H bonds of arenes are among the strongest
ones.37 A number of structures are conceivable for the
[M,C6,H4,O]+ species,66 and as one possibility we propose that
the formed [Cr,C6,H4,O]+ ion exhibits the benzo-metallaoxetane
structure 15 (Scheme 11); this would be in keeping with the
experimental findings that (i) in the secondary reaction of
[Cr,C6,H4,O]+ with benzene, clustering to [Cr,C12,H10,O]+ is
observed while ligand-exchange does not occur, and (ii) a slow
addition of background water leads to [Cr,C6,H6,O2]+ which
subsequently forms an adduct [Cr,C12,H12,O2]+ with benzene
instead of exchanging the water ligand for benzene.
6
to CrO+ (4Π/4Σ+and Π/6Σ+) + O (3P) and to Cr+ (6S) + O2 (1∆g)
were taken from refs 33, 35b, and 37. Crosses denote excitation energies
calculated using CASPT2D/BS-I for the various states at the LSD-
geometries of the corresponding ground states.
exhibits a clear preference for C-H versus C-C bond activa-
tion; the latter is in fact only observed for alkenes. Finally, the
product distributions indicate that the activation of hydrocarbons
by the dioxide OCrO+ involve a single oxo-ligand in the initial
step ([2 + 2] pathway), similar to the reactions of metal
monoxides, while there is no direct evidence for a simultaneous
addition of the substrate to both oxygen atoms in terms of a [3
+ 2] cycloaddition.
Dissociation of [Cr,O2]+ and Chromium-Mediated Activa-
tion of Dioxygen. In this section we address the dissociation
of OCrO+ (2A1) into the ground-state asymptote Cr+ (6S) + O2
(3Σg-) and vice versa, i.e., the formation of chromium dioxide
from this entrance channel via bond activation of molecular
oxygen by Cr+. Let us first consider the theoretical aspects:
For either the 2A1 or 2B2 states, dissociation of OCrO+ into the
ground states Cr+ (6S) and O2 (3Σg-) is spin-forbidden, and the
energetically lowest-lying allowed pathway leads to the forma-
tion of excited Cr+ (4D) + O2 (3Σg-). Nevertheless, spin-orbit
coupling (SOC) may render dissociation into the ground states
possible via curve crossing. In order to construct a rather
qualitative potential-energy surface for the [Cr,O2]+ system, we
computed several vertical excitations of the ground states of
structures 1-3 into different multiplicities (Figure 3). Inspection
of the surface reveals that a direct activation of O2 (3Σg-) by
ground-state Cr+ (6S) to form OCrO+ (2A1) cannot occur at
thermal energies, even if spin-orbit coupling (SOC) would
be very efficient. This conclusion is based on the following
arguments: (i) Even upon reduction of the symmetry group to
Cs, the ground states of the isomers OCrO+ and Cr(O2)+
transform to two different hypersurfaces (A1 f A′ and B2 f
A′′, respectively); note, however, that this argument does not
Concluding this section, we state that thermalized OCrO+ is
capable to activate hydrocarbons, including methane and
benzene, with moderate to large efficiencies. Three aspects of
the rich chemistry of OCrO+ should be pointed out.67 At first,
thermochemical arguments imply that changes in multiplicity
take place during the oxidation reactions. Secondly, OCrO+
(63) Blumenberg, B. Nachr. Chem. Techn. Lab. 1994, 42, 480.
(64) (a) Sto¨ckigt, D.; Schwarz, H. Chem. Ber. 1994, 127, 2499. (b)
Sto¨ckigt, D.; Schwarz, H. Liebigs Ann. 1995, 429.
(65) Becker, H.; Schro¨der, D.; Zummack, W.; Schwarz, H. J. Am. Chem.
Soc. 1994, 116, 1096.
(66) Schro¨der, D.; Zummack, W.; Schwarz, H. Organometallics 1993,
12, 1079.
(67) OMoO+, the higher homologue of OCrO+, is much less reactive
toward hydrocarbons; see: Cassady, C. J.; McElvany, S. W. Organome-
tallics 1992, 11, 2367.

9950 J. Am. Chem. Soc., Vol. 118, No. 41, 1996
Fiedler et al.
1:2 as compared to ca. 2:1 in Figure 1a. The most notable
difference, however, concerns the dication signals in that these
are more intense and CrO22+ is much larger for the ion generated
by CI of the Cr(CO)6/O2 mixture than for that obtained by EI
of CrO2Cl2. Similarly, in the NR spectrum (Figure 4b) the
signals for Cr+ and O2•+ are much more abundant than in Figure
1b, while the intensity of CrO+ decreases. In addition, the
recovery signal for OCrO+ is less intense in Figure 4b, which
implies that for this [Cr,O2]+ ion the efficiency of the NR
process is smaller as compared to doublet states of OCrO+.
These differences in intensities suggest that other isomers of
[Cr,O2]+ are formed in the CI experiment.
Besides these qualitative aspects, there exists a quantitative
difference between the [Cr,O2]+ ions formed by either EI of
CrO2Cl2 or CI of Cr(CO)6/O2 with respect to the second
ionization energies as expressed in terms of Qmin values (see
experimental details):21 For the chromium dioxide cation
generated from CrO2Cl2 we determined a Qmin value of 19.2 (
0.7 eV, while for the ion formed from the Cr(CO)6/O2 mixture,
it is significantly lower (Qmin ) 15.7 ( 1.3 eV).70 This
observation proves that at least two different states and/or
structures of [Cr,O2]+ were sampled in the charge-stripping
experiment, i.e., the ground state OCrO+ (2A1) exhibits a high
Qmin to form the corresponding chromium dioxide dication
CrO22+, while the [Cr,O2]+ ion from CI of Cr(CO)6/O2 is either
excited by ca. 3.5 eV as compared to OCrO+ (2A1) and/or the
lower Qmin is due to the formation of a [Cr,O2]2+ dication with
a different structure than that of the dioxide.71 The latter
conjecture is supported by the recent proposal that the dicationic
complex Cr2+-O2 can be formed in the reaction of aqueous Cr2+
with dioxygen.12
Figure 4. (a) CA mass spectrum (O2, 80% T) of [Cr,O2]+ obtained by
chemical ionization of Cr(CO)6 with O2, (b) NR mass spectrum (Xe
80% T; O2 80% T).
Although these two [Cr,O2]+ ions were generated under quite
different conditions, i.e., EI and CI, the differences in the spectra
are not likely to be due to internal energy effects, but rather
suggest that different isomers or states of [Cr,O2]+ were formed
in both experiments. While for the ion derived from CrO2Cl2
we are quite certain that it corresponds to a doublet state of
OCrO+ (2A1 or 2B2), an assignment for the ion generated from
Cr(CO)6/O2 is more difficult: On the one hand, the intense CrO+
signal upon collisional activation is unexpected for the weakly
bound Cr(O2)+ and, further, the observation of a recovery signal
in Figure 5b indicates the presence of an inserted structure. On
the other hand, the relatively more abundant signals for Cr+
apply for 2B2 as the ground state of OCrO+. (ii) SOC can only
occur between surfaces which differ in spin multiplicity by a
maximum of two (one spin inversion), such that SOC does not
allow for a direct connection of the sextet and doublet surfaces
but rather has to involve an intermediate quartet state.68,69 (iii)
The potential intermediate OCrO+ (4A′′) is energetically very
close to the Cr+ + O2 entrance channel, such that even low
barriers associated with stepwise SOC from the sextet to the
doublet would prevent formation of OCrO+ (2A1) from the
thermalized reactants. At elevated energies, OCRO+ (4A′′) may
be formed in a spin-allowed path (Figure 3), but one SOC to
the doublet ground state remains.
On the one hand, Figure 3 implies that under thermal
conditions a Cr+-catalyzed oxidation using molecular oxygen
as terminal oxidant cannot be realized in the gas phase. On
the other hand, the potential-energy surface leads to the proposal
that other states or isomers but doublet OCrO+ exist and may
be generated and characterized experimentally. To this end,
we subjected a mixture of Cr(CO)6 and O2 to chemical ionization
(CI) and recorded the CA and NR mass spectra of the [Cr,O2]+
species formed.18 Both spectra reveal substantial differences
as compared to the corresponding spectra of the OCrO+ ion
generated by EI of CrO2Cl2 (Figure 1). In the CA spectrum
(Figure 4a) the ratio CrO+:Cr+ is reversed and amounts to ca.
•+
and O2 in the CA and NR spectra, respectively, point to the
presence of the dioxygen complex Cr(O2)+ which may also
survive a NR experiment.14
One way to clarify this issue is to examine in more detail
the formation of [Cr,O2]+ from Cr(CO)6 and oxygen. Under
CI conditions [Cr,O2]+ is likely to be formed in a ligand
substitution of Cr(CO)+ with molecular oxygen. Therefore, we
first examined the carbonyl complex [Cr,O2•CO]+ which is
formed when a mixture of Cr(CO)6, O2, and CO (ratio ca. 1:50:
50) is subjected to chemical ionization. In principle, these
[Cr,O2,CO]+ ions can bear three different connectivities, i.e.,
(OC)Cr(O)2+, (OC)Cr(O2)+, or (O2C)CrO+. Unimolecular
dissociation of the metastable ions (MI) or collisional activation
(CA) of [Cr,O2,CO]+ ions leads to four fragmentations (Scheme
12).
6
3
(68) One reviewer has argued that formal spin-pairing of Cr+ and O2
should lead to a quartet state of [Cr,O2]+ without involving curve crossing
while formation of 6Cr(O2)+ should require a spin-flip. Therefore, it seems
worthy to clarify that simple spin counting, i.e., Cr+ (vvvvv) + O2 (vv) f
Cr(O2)+ (vvvvvvv) and Cr+ (vvvvv) + O2 (VV) f Cr(O2)+ (vvvVvVv), is not at all
appropriate for the evaluation of spin conservation rules. Rather, the possible
spin couplings of two separated fragments must be derived from a vector
model (ref 69) using a multireference approach. In fact, it turns out that
(70) The relatively large error in the determination of the Qmin value is
due to a sensitive dependence of the data on ion source conditions rather
than too low ion intensities.
(71) Preliminary calculations using the LSD approach and CASPT2D/
BS-I demonstrate that bent Cr(O2)2+ (5A′′) is in fact more stable than CrO2
2+
(1A1). In terms of second IEs, this difference translates to 16.1 eV for Cr-
(O2)+ versus 17.7 eV for OCrO+. However, a more detailed theoretical
treatment is indicated: Kretzschmar, I.; Fiedler, A.; Schro¨der, D.; Schwarz,
H., unpublished results.
6
3
6
the process Cr+ + O2 f Cr(O2)+ does not require curve crossing.
(69) Herzberg, G. Molecular Spectra and Molecular Structure; Krieger,
Malabar, 1989; Vol. 1, p 319.

Chromium Dioxide Cation OCrO+ in the Gas Phase
J. Am. Chem. Soc., Vol. 118, No. 41, 1996 9951
These findings are in marked contrast to the fact that OCrO+
generated by EI of CrO2Cl2 does hardly undergo any 16O/18O
exchange with 18O2 (<1%) and does not react with H2O except
for a slow association to form (H2O)Cr(O)2+. According to
the calculated potential-energy surface (Figure 3), exchange of
the 18O2 ligand by 16O2 or water point to the formation of the
dioxygen complex Cr(O2)+ (6A′′), while the occurrence of
sequential exchanges of the 18O-label indicate the presence of
a chromium dioxide cation other than OCrO+ (2A1). In
principle, these findings would offer the possibility to examine
state-selective reactions under FTICR conditions, once ther-
malization of the precursor ions can be warranted. However
in practice, thorough thermalization of [Cr,O2]+ generated from
Cr(CO)6/O2 is not only associated with substantial losses of
intensity (>90%), but further, these [Cr,O2]+ ions exhibit a
similar reactivity toward ethene and propane as compared to
OCrO+ (2A1), and the only significant difference is a shift of
the product distribution toward more energetic products (e.g.
more Cr+). The most intriguing finding is that after even more
careful thermalization the [Cr,O2]+ ions generated from Cr(CO)+
and O2 do not undergo charge transfer with hexafluorobenzene
(IE ) 9.9 eV). The lowest excited state OCrO+ (4A′′) is ca.
1.5 eV more energetic than OCrO+ (2A1) (IE ) 9.7 ( 0.2 eV),
and both ions can lead to neutral CrO2 in its triplet ground
state.14 Therefore, the absence of electron transfer suggests that
the fraction of ions (ca. 5%) which survives thermalization in
the presence of hexafluorobenzene corresponds to OCrO+ in
its doublet ground state.
Let us summarize these findings: First, CA and NR spectra
demonstrate that CI of Cr(CO)6/O2 gives rise to different
[Cr,O2]+ ions than does EI of CrO2Cl2. Secondly, the experi-
ments with the [Cr,O2•CO]+ complex as well as the 16O/18O
exchanges of [Cr,O2]+ under FTICR conditions indicate that a
mixture of 6Cr(O2)+, 4OCrO+, and 2OCrO+ has been generated.
However, either in the course of ion/molecule reactions or as a
consequence of the thermalizing collisions, this mixture ends
up in the ground state ion OCrO+ (2A1). Hence, the fascinating
perspective to examine reactions of state-selected transition-
metal ions other than bare atoms was not pursued further,75
because under FTICR conditions it would suffer from the pitfall
that discrimination between kinetic, rovibronic, and electronic
excitation is impossible.
Figure 5. Partial 16O/18O exchange in the reaction of [Cr,18O2]+ with
16O2 (reaction time 5 s, pO2 ) 5 × 10-9 mbar). [Cr,18O2]+ was generated
by reacting mass-selected Cr(CO)+ with pulsed-in 18O2.
Scheme 12
The relative abundances for the loss of CO versus O2 (Scheme
12) provide evidence against the formation of the dioxygen
complex Cr(O2)+ (6A′′) from [Cr,O2,CO]+, because the large
difference of the binding energies of O2 and CO to Cr+ cation
(9.5 kcal/mol versus 21.4 kcal/mol72 ) suggests a reversed ratio
+
for (OC)Cr(O2)+; consequently, the structure (OC)Cr(O)2 is
implied for the precursor ion and that of OCrO+ for the
product.73 For a further characterization of the [Cr,O2]+ ion
formed by unimolecular loss of CO from [Cr,O2,CO]+, the
product ion was subjected to a further collision experiment. In
the corresponding MI/CA spectrum, the Cr+:CrO+ ratio amounts
to 1.5:1 which is close to the 2:1 ratio in Figure 4a, but rather
different from the 1:2 ratio for the genuine OCrO+ species
(Figure 1a). Following the arguments concerning the competi-
tive elimination of CO versus O2, a Cr(O2)+ species cannot be
formed, and hence we conclude that another state than OCrO+
(2A1) must have been formed in the unimolecular dissociation.
In addition, loss of CO2 is in line with the formation of (OC)-
Cr(O)2+, but also with (O2C)CrO+. However, the relatively
large abundance of bare Cr+, even for the metastable ions, is
somewhat surprising for (OC)Cr(O)2+; rather, it points to
sequential losses of the carbonyl and oxygen ligands from (OC)-
Cr(O2)+. Thus, it seems as if CI of Cr(CO)6/O2 gives rise to a
mixture of different [Cr,O2]+ ions.
Finally, two different routes for the activation of dioxygen
by atomic chromium should be considered. While in the neutral
[Cr,O2] system the situation seems to be quite similar to the
cation surface,14 the anion and dication surfaces arise from lower
spin states which may help to circumvent the spin-inversion
problem, we encounter in the [Cr,O2]+/0 systems. Due to the
quartet ground state15b of 4CrO2- direct association of 5Cr- anion
3
The possible formation of other ions than doublet OCrO+
was further examined under FTICR conditions by reacting
[Cr,18O2]+, formed from Cr(CO)+ and pulsed-in 18O2 with 16O2
(Figure 5). The observation of a weak signal for [Cr,16O,18O]+
and a more intense one for [Cr,16O2]+ suggest that 16O/18O
exchange can either occur sequentially or involves the complete
oxygen ligand.74 Further, exchange of the oxygen ligand with
background water to yield (unlabeled) Cr(OH2)+ takes place.
with O2 is spin-allowed, but not very promising, because the
electron affinity of Cr is low (0.7 eV), such that electron
detachment is likely to occur. However, [Cr,O2]- anions are
easily accessible by CI of Cr(CO)6/O2 in the negative ion mode.
Similar to the related [Fe,O2]- system,18b,c the -NR+ spectrum
of [Cr,O2]- is in keeping with an inserted chromium(III) dioxide
•+
anion, i.e.: OCrO+ (40%), CrO+ (100%), Cr+ (90%), O2
-
(<1%). However, the mechanism for the formation of CrO2
from this mixture is different from the desired combination of
atomic chromium with oxygen, because also oxidations of the
(72) Armentrout, P. B.; Kickel, B. L., In Organometallic Ion Chemistry;
Freiser, B. S., Ed.; Kluwer: Dordrecht, 1996, p 1.
(73) This argument is based on the kinetic method; for a review, see:
Cooks, R. G.; Patrick, J. S.; Kotiaho, T.; McLuckey, S. A. Mass Spectrom.
ReV. 1995, 13, 287.
(74) The reaction kinetics were complicated and were not resolved,
because several processes compete, e.g. dissociation and 16O/18O-exchange
of Cr(O2)+ (6A′′), stepwise isotopic exchange and quenching of OCrO+
(4A′′) to the OCrO+ (2A1) ground state; the latter ion is unreactive toward
O2.
carbonyl ligands take place,15a e.g., (CO)nCr- + O2
f
(CO)n-1CrO- + CO2. Another route to chromium dioxide via
the dicationic surface has been reported by Bakac and co-
workers12 to play a role in the condensed phase. Under strictly
bimolecular collisions in the gas phase, this pathway for dioxide
(75) Review: Armentrout, P. B. Science 1991, 251, 175.

9952 J. Am. Chem. Soc., Vol. 118, No. 41, 1996
Fiedler et al.
formation may involve the sequence 5Cr2+ + 3O2 f 5Cr(O2)2+
f 3Cr(O2)2+ f 1CrO22+, and the close energy spacing of these
states71 suggests that this route may indeed lead to chromium
dioxide. Unfortunately, the intensities of the [Cr,O2]2+ dications
were too low in our experiments in order to allow for further
characterization of the ion as well as an exploration of this
attractive perspective.
of the [Cr,O2]+ system indicates that the reaction Cr+ (6S) +
O2 (3Σg-) a Cr(O2)+ (6A′′) a OCrO+ (2A1) is hampered by
both spin and symmetry restrictions and cannot occur at thermal
energies. This finding accounts for the experimental observation
that neither Cr+ nor its complexes Cr(L)+ with oxidizable
ligands L react with dioxygen in the sense of oxidation
processes, while Fe+ complexes do.76 Further, a detailed
examination of the [Cr,O2]+ system demonstrates that also the
more energetic species 6Cr(O2)+ and 4OCrO+ are experimentally
accessible. However, collisional cooling is associated with
dissociation and quenching to the ground state OCrO+ (2A1),
such that the examination of state-selective reactions of [Cr,O2]+
by means of FTICR technology is not indicated.
Conclusions
Chromium dioxide OCrO+ can be generated by electron
ionization of chromyl chloride and characterized by collision
spectroscopy. According to high-level ab initio calculations
OCrO+ exhibits two closely spaced doublet ground states (2A1
and 2B2), while states of higher multiplicities and the dioxygen
complex Cr(O2)+ are much higher in energy. The high-valent
2OCrO+ reacts with hydrocarbons including methane and
benzene via O-atom transfer from the metal to the substrate.
Remarkable is the initial C-H bond activation, which involves
a single oxo ligand of the dioxide via a formal [2 + 2] pathway
with a different regioselectivity as compared to that of transition-
metal monoxide cations.
Acknowledgment. This research has been sponsored by the
Deutsche Forschungsgemeinschaft, the Volkswagen-Stiftung,
and the Fonds der Chemischen Industrie. We thank Dr. Gilles
Ohanessian for helpful discussions and comments.
JA960157K
(76) (a) Schro¨der, D.; Schwarz, H. Angew. Chem., Int. Ed. Engl. 1993,
32, 1420. (b) Wesendrup, R.; Diplomarbeit, Technische Universita¨t Berlin,
1994. (c) Boissel, P.; Marty, P.; Klotz, A.; de Parseval, P.; Chaudret, B.;
Serra, G. Int. J. Mass Spectrom. Ion Processes 1995, 242, 157.
With respect to catalytic oxidation and the activation of
molecular oxygen by Cr+, the computed potential-energy surface