Photochemical studies of Co2(CO)6(acetylene) complexes and their phosphine derivatives in frozen Nujol matrices

Article abstract of DOI:10.1016/S0022-328X(00)00243-6

Photolysis of Co₂(CO)₆(alkyne), where alkyne=H₂C₂ (I), (C₂H₅)₂C₂ (II) (C₆H₅)₂C₂ (III) and C₆H₅C₂H (IV) in frozen Nujol at ca. 90 K results in loss of CO to give Co₂(CO)₅(alkyne) in which CO-loss appears to be from the axial position. Complete conversion to a second isomer, presumed to have a vacancy in an equatorial position, or to starting material is observed for all compounds upon annealing the matrices to ca. 140 K. The second isomer is also formed by Co₂(CO)₅[(C₆H₅)₂C₂] upon photolysis at λ<280 nm. Photolysis of Co₂(CO)₆As₂ (V), gave photochemical and annealing products whose spectra were analogous to those of the alkyne compounds. Photolysis of phosphine substituted derivatives such as axial-Co₂(CO)₅(PR₃)(C₂R′₂), where R=Bu; R′=H (VI), R=Ph; R′=H (VII), R=OPh, R′=H (VIII), R=Ph; R′=Ph (M) yields two isomeric CO-loss products. Annealing samples of V-IX to ca. 140 K results in simple reaction reversal. Solution photolysis of IX with PPh₃ gives Co₂(CO)₄(PPh₃)₂(Ph₂C₂). Photolysis of Co₂(CO)₄(PPh₃)₂(H₂C₂) (X) yields two products. Annealing the sample reforms the starting material and also generates a new species. Photolysis of Co₂(CO)₄(Ph₂P)₂CH₂][(C₂H₅)₂C₂] (XI) forms a single product that reforms XI upon annealing.

Full text of DOI:10.1016/S0022-328X(00)00243-6

www.elsevier.nl/locate/jorganchem
Journal of Organometallic Chemistry 605 (2000) 7–14
Photochemical studies of Co2(CO)6(acetylene) complexes and their
phosphine derivatives in frozen Nujol matrices
Thomas E. Bitterwolf *, W. Bruce Scallorn, Callie A. Weiss
Department of Chemistry, Uni6ersity of Idaho, Moscow, ID 83844-2343, USA
Received 15 October 1999; received in revised form 4 April 2000; accepted 24 April 2000
Abstract
Photolysis of Co2(CO)6(alkyne), where alkyne=H2C2 (I), (C2H5)2C2 (II) (C6H5)2C2 (III) and C6H5C2H (IV) in frozen Nujol at
ca. 90 K results in loss of CO to give Co2(CO)5(alkyne) in which CO-loss appears to be from the axial position. Complete
conversion to a second isomer, presumed to have a vacancy in an equatorial position, or to starting material is observed for all
compounds upon annealing the matrices to ca. 140 K. The second isomer is also formed by Co2(CO)5[(C6H5)2C2] upon photolysis
at uB280 nm. Photolysis of Co2(CO)6As2 (V), gave photochemical and annealing products whose spectra were analogous to those
of the alkyne compounds. Photolysis of phosphine substituted derivatives such as axial-Co2(CO)5(PR3)(C2R%2), where R=Bu;
R%=H (VI), R=Ph; R%=H (VII), R=OPh, R%=H (VIII), R=Ph; R%=Ph (M) yields two isomeric CO-loss products.
Annealing samples of V–IX to ca. 140 K results in simple reaction reversal. Solution photolysis of IX with PPh3 gives
Co2(CO)4(PPh3)2(Ph2C2). Photolysis of Co2(CO)4(PPh3)2(H2C2) (X) yields two products. Annealing the sample reforms the
starting material and also generates a new species. Photolysis of Co2(CO)4(Ph2P)2CH2][(C2H5)2C2] (XI) forms a single product that
reforms XI upon annealing. © 2000 Published by Elsevier Science S.A. All rights reserved.
Keywords: Photochemical studies; Acetylene; Phosphine derivatives
are approximately co-linear with the CoꢀCo ‘bent’
bond, while the remaining equatorial carbonyl ligands
are coplanar with the acetylene. Hoffmann has reported
molecular orbital calculations for these compounds car-
ried out at the extended Hu¨ckel level [3] while Van
Dam obtained similar results in an ab initio SC MO
calculation [4]. In Hoffmann’s model, the HOMO a2
orbital corresponds to alkyne to metal back-bonding,
while the HOMO-1 a1 orbital corresponds to the
CoꢀCo ‘bent’ bond. The LUMO, b2, and LUMO+1,
a1, levels are anti-bonding counterparts of the alkyne to
metal bonds. The relative order of the a2 and a1,
bonding orbitals determined by Hoffmann have been
confirmed by photoelectron spectroscopy [5].
1. Introduction
Compounds of the Co2(CO)6(alkyne) series are
among the earliest known organometallic compounds
[1] and their application to organic chemistry has been
well studied [2]. With symmetric alkynes the molecules
have been shown to have C26 symmetry with distorted
octahedral symmetry about the individual cobalt
atoms. As shown in Fig. 1, two axial carbonyl ligands
Thermal reactions in which one or more carbon
monoxide ligands may be replaced by phosphines to
yield stable derivatives have been reported with all
simple phosphines giving mono and disubstituted prod-
ucts in which the phosphine appears in the axial posi-
tion [6]. Bidentate ligands such as (Ph2P)2CH2 and
(Ph2As)2CH2 have been found to bridge equatorial
positions of the cobalt atoms [7]. Heck examined the
Fig. 1. Geometry of Co2(CO)6(alkyne).
* Corresponding author. Tel.: +1-208-8856426; fax: +1-208-
8856173.
E-mail address: bitterte@uidaho.edu (T.E. Bitterwolf).
0022-328X/00/$ - see front matter © 2000 Published by Elsevier Science S.A. All rights reserved.
PII: S0022-328X(00)00243-6

8
T.E. Bitterwolf et al. / Journal of Organometallic Chemistry 605 (2000) 7–14
Table 1
2. Results and discussion
Electronic spectral data for Co2(CO)6(alkyne) derivatives and
Co2(CO)6As2 in petroleum ether
Photochemical studies were carried out using a glass
cryostat as previously described [15]. Samples are dis-
solved in Nujol and frozen to ca. 90 K with liquid nitro-
gen. A 350 W high pressure mercury lamp was the
primary light source with wavelength selectivity being
achieved through the use of optical filters.
Alkyne
u (nm)
m (M−1 cm−1
)
C2H2 (I)
515
405
342
540
400
350
530
428
358
560
435
360
460
377
188
530
2700
252
1024
4640
86
834
3220
702
1140
3420
Not determined
Not determined
(C2H5)C2 (II)
2.1. Photolysis of Co2(CO)6(acetylene) (I–IV) and
Co2(CO)6As2 (V)
(C6H5)C2H (III)
(C6H5)2C2 (IV)
Co2(CO)6As2 (V)
Electronic spectra of compounds I–V exhibit two
weak bands (about 530 and 420 nm), a strong band
(about 350 nm) attributed to a MLCT transition, and an
intense ligand field band at uB290 nm (Table 1). IR
spectra of these compounds have four or five bands in
the carbonyl stretching region consistent with their C26
symmetry [16,17]. Photolysis at wavelengths greater than
450 nm resulted in no change in the IR spectra as deter-
mined by difference spectral methods. IR spectral
changes were noted for uB450 nm. Incremental de-
creases in wavelength using a series of graduated filters
established that photochemistry was associated with ex-
citation of either the MLCT or ligand field bands.
Photolysis of compounds I–V in the range 250BuB
380 nm resulted in the decrease in intensity of the bands
of the starting material and growth of new bands. This is
illustrated in Figs. 2 and 3 for II. IR data are presented
in Table 2. Fig. 2 presents 1:1 subtraction spectra while
in Fig. 3 the bands of starting material are nulled thus re-
vealing bands that overlap with those of the starting ma-
terial. As observed by Gordon [14], some product bands
overlap with those of the starting material thus it was
necessary to examine difference spectra with the starting
material bands canceled out to locate all product bands.
Small shifts of as little as 2 cm−1 between spectra occa-
sionally result in artificial splitting of the product bands.
The band at 2131 cm−1 observed in all cases is associ-
ated with ‘free’ CO indicating that in each case CO is lost
from the molecule. Loss of CO from these C26 molecules
results in reduction of symmetry to give either a Cs (axial
loss) or C1 (equatorial loss). In either case five IR active
carbonyl bands are expected and observed for I–V. Both
I and V were found to have six distinct photoproduct
bands in the carbonyl stretching region. In both cases the
extra band was at the high frequency end. A possible ex-
planation for this effect is discussed below.
kinetics of the reactions of I and II with triphenylphos-
phine in toluene and found the reactions to be first order
in cobalt compound. I was found to have activation
parameters of DE*=17.0 kcal mol−1 and DS* = −
25
23.0 cal mol−1 deg−1. The negative value for the en-
tropy of activation prompted Heck to suggest a
symmetric transition state with a bridging carbonyl [8].
Photochemical substitution of Co2(CO)6(RC2H),
where R=Ph, n-Bu, CH3, CH2, CH3O2C, or CH3CO,
by PPh3 has been reported by Anderson using the 351.1,
351.4 and 363.8 nm lines from an Ar ion laser [9].
Thermal loss of CO is postulated as a first step in the
Pauson–Khand synthesis of cyclopentenone derivatives
[10] and as the initiating step in alkyne trimerization [11].
The Pauson–Khand reaction has been observed to pro-
ceed stoichiometrically under photochemical conditions
[12] and recently Livinghouse has described visible light
conditions for a catalytic Pauson–Khand reaction [13].
Gordon et al., have photolyzed Co2(CO)6(C6H5C2H)
(IV), in argon, mixed argon–nitrogen, and nitrogen
matrices at 12 K and have observed the formation of
Co2(CO)5(C6H5C2H) (Ar matrix) or Co2(CO)5(N2)-
(C6H5C2H) (Ar–N2 or N2 matrices). Trace amounts of
what appeared to be a second species were also observed
in the argon matrix [14].
As part of an extended investigation into the photo-
chemistry of tetrahedrane complexes in frozen Nujol
matrices we have reexamined the photolysis of IV, three
other members of this class of compounds and the
analogous arsenic compound (V). The high quality spec-
tra obtained from this method and the ease with which
samples may be annealed to examine possible thermal re-
actions allow new details to be extracted from the data
that bear on the identity of the photoproducts. Further,
the excellent solubility of the phosphine derivatives in
Nujol allowed investigation of their photochemistry as
well.
Annealing each sample to ca. 140 K followed by its
return to ca. 90 K for recording of its spectrum was
found to result in complete loss of the bands associated
with the photoproducts and some reduction of the ‘free’
CO band. In addition to bands of the starting materials
additional bands were observed to grow in. In the case
of II, Figs. 2c and 3c clean conversion of the photoly-
sis product to the anneal product occurred without

T.E. Bitterwolf et al. / Journal of Organometallic Chemistry 605 (2000) 7–14
9
concurrent appearance of starting material or change of
CO. Back photolysis at u\500 nm results in simple
reversal of the forward photolysis as observed by Gor-
don [14]. In the case of IV it was found that photolysis
in the range 285BuB380 nm produced the expected
photoproduct while expanding this range to 250BuB
380 nm gave rise to the species observed upon anneal-
ing as well as the primary photoproduct.
that all compounds undergo CO loss to form an unsat-
urated species (A), of similar structure, and that all of
these photoproducts undergo isomerization upon
warming to form a thermodynamically preferred species
(B). Four of the five carbonyl bands in the IR spectra
of species A and B are similar in both position and
relative intensity, but the spectra of species A are
characterized by a medium intensity carbonyl band at
about 1970 cm−1, while the lowest frequency bands of
the B species are weak and are found at about 2000
The strong similarities in spectra of photoproducts
and anneal products for all five compounds indicate
Fig. 2. Nujol matrix photolysis of Co2(CO)6(3-hexyne). (a) Starting material; (b) 30 min photolysis (250BuB380 nm) minus starting material;
(c) 10 min anneal minus starting material.
Fig. 3. Nujol matrix photolysis spectra of Co2(CO)6(3-hexyne). (a) Starting material; (b) 30 min photolysis (250BuB380 nm) with starting
material bands nulled out; (c) 10 min anneal with starting material bands nulled out. The peak at 2025 cm−1 in 3(b) is a subtraction artifact as
is the broad feature at 1900 cm−1 in 3(c).

10
T.E. Bitterwolf et al. / Journal of Organometallic Chemistry 605 (2000) 7–14
Table 2
As noted, the pentacarbonyl fragments may have Cs
(axial CO loss) or C1 (equatorial CO loss) symmetry
with theory predicting five carbonyl stretching bands
for each of these isomers. There is a strong similarity
between the spectra of the A species and those of axial
IR spectral data for Co2(CO)6(alkyne) derivatives and Co2(CO)6As2
(cm−1) in frozen Nujol at ca. 90 K a
Alkyne
Parent
spectrum
Photoproduct
(A)
Anneal product
(B)
Co2(CO)5(PR3)(alkyne).
For
example,
axial-
C2H2 (I)
2099 (w)
2058 (s)
2032 (s)
2027 (sh) 2019 (m)
2016 (w)
2084 (w)
2080 (w)
2034 (s)
2084 (w)
2043 (s)
2019 (m)
2014 (sh)
2000 (w)
Co2(CO)5(PBu3)(C2H2) has bands in frozen Nujol at
2064 (m), 2008 (s), 2001 (s), 1989 (m), and 1966 (m)
cm−1
,
while the bands of the
A
species of
2011 (sh)
1980 (m)
2070 (w)
2018 (s)
2004 (m)
1999 (sh)
1966 (m)
2077 (w)
2029 (s)
Co2(CO)5(C2H2) are at 2084 (w), 2080 (w), 2034 (s),
2019 (m), 2011 (sh), and 1980 (m) cm−1. It is generally
observed that carbonyl loss fragments have the same
pattern of carbonyl bands as the corresponding phos-
phine derivatives with the bands of the carbonyl loss
(C2H5)2C2 (II)
2088 (w)
2047 (s)
2023 (s)
2013 (m)
2002 (w)
2095 (w)
2057 (s)
2030 (s)
2073 (w)
2032 (s)
2006 (m)
2001 (sh)
1988 (w)
2080 (w)
2042 (s)
2018 (m)
1998 (w)
2076 (w)
2042 (s)
2017 (s)
2012 (sh)
species shifted to higher frequency by 20–40 cm−1
.
(C6H5)C2H (III)
(C6H5)2C2 (IV)
Keeping with this pattern we have assigned A as the
axial-CO loss species.
Assuming the above analysis to be correct, the most
likely identity of the B species would be the correspond-
ing equatorial species, however, all thermal and photo-
2015 (m)
2026 (sh) 1976 (m)
2091 (w)
2056 (s)
2028 (m)
2025 (sh) 2005 (sh)
1975 (m)
2073 (m)
2028 (s)
2013 (m)
chemical
reactions
Of
Co2(CO)6(alkyne)
and
Co2(CO)6As2 compounds with simple phosphine lig-
ands yield products in which the phosphine is in the
axial position. It is possible that an equilibrium between
the two structural isomers occurs in solution and that
reaction between ligands and the axial isomers is fa-
vored for steric reasons. It should be noted that photol-
ysis of Fe2(CO)6(SR)2 compounds in frozen Nujol
yields a pentacarbonyl intermediate with an IR spec-
trum essentially identical to those of the A species
described here [18]. In the case of the iron compounds
there is no evidence of a thermal rearrangement.
The additional carbonyl bands observed for I and V
cannot be explained by the formation of simple isomers
since in both cases annealing results in the formation of
a clear second species. We propose that the additional
bands arise from splitting of the highest energy sym-
metric stretch as the result of vibrational coupling
between these bands and the Co2(C2H2) or Co2As2
tetrahedrane unit. For compounds II–IV, with alkyl or
aryl substituents on the acetylene group these vibra-
tional couplings may be damped.
Co2(CO)6As2 (V) 2094 (m)
2082 (m)
2074 (w)
2035 (s)
2030 (sh)
2017 (m)
1990 (m)
2078 (w)
2047 (s)
2025 (m)
2017 (w)
2058 (s)
2038 (s)
2031 (m)
a Free CO band at 2032 cm−1 observed in all cases.
cm−1. A careful search was made for additional bands
associated with bridging carbonyl groups, but no addi-
tional bands were found. Symmetry arguments provide
no clues as to the likely structural assignment of the
two species. The carbonyl band at 1970 cm−1 is well
above the region associated normally with semi-bridg-
ing carbonyl groups, thus there is no experimental
support for the bridging carbonyl intermediate species
proposed by Heck.
2.2. Photolysis of Co2(CO)5(PR3)(alkyne) (VI–IX),
Co2(CO)4(PBu3)(H2C2) (XI) and
Co2(CO)4(DMPM)(C2Et2) (XII)
Multiple substitution of carbon monoxide by phos-
phines is well known for Co2(CO)6(alkyne) compounds
[6]. We sought to establish whether phosphine sub-
stituted compounds undergo photolysis and CO-loss in
a manner similar to that of the parent carbonyl
compounds. Table 3 presents electronic spectral data
for these substituted compounds and Table 4 presents
IR data of the compounds and photoproducts.
Compounds VI–IX are strongly colored as a result of a

T.E. Bitterwolf et al. / Journal of Organometallic Chemistry 605 (2000) 7–14
11
Table 3
P(OPh)3\PPh3\PBu3 paralleling the decrease in p-
acceptor capacity of the ligand.
Electronic spectral data for phosphine and phosphite derivatives in
petroleum ether
Photolysis into the high energy electronic bands of
these compounds resulted in IR bands of the starting
materials decreasing in intensity while new bands grew
in. In each case a band at 2131 cm−1 associated with
‘free’ CO was observed while five or six well-defined
band maxima were found. Asymmetries in the shapes
of some of these bands suggest that there might be
additional, unresolved bands. Annealing of the pho-
tolyzed sample of VIII resulted in reduction of bands at
2047, 1986, and 1962 cm−1 and growth of bands at
2052, 2046, 2016, and 1990 cm−1 as well as bands of
the starting material. Annealing of photolyzed samples
of VI, VII and IIX results in reversal of the photochem-
ical reaction.
Compound
u (nm)
m (M−1 cm−1
)
Co2(CO)5(PBu3)(C2H2) (VI)
460
370
369
420
355
565
382
480
372
315
525
325
1010
8500
2330
838
4350
1880
27 600
2560
22 800
13 600
1100
Co2(CO)5(PPh3)(C2H2) (VII)
Co2(CO)5[P(OPh)3(C2H2) (VIII)
Co2(CO)5(PPh3)(C2Ph2) (IX)
Co2(CO)4(PBu3)2(C2H2) (XI)
Co2(CO)4(DPPM)(C2Et2) (VI)
24 700
Photolysis of mono-phosphine compounds, VII–IX,
results in CO-loss and the appearance of a complex set
of carbonyl bands. Although these compounds have
three unique CO groups, axial-Co(CO)3, equatorial-
Co(CO)3, and equatorial-Co(CO)2(PPh3), evidence from
thermal reactions suggest that addition of a second
phosphine to Co2(CO)5(PR3)(alkyne) dominantly or ex-
clusively occurs on the unsubstituted cobalt atom. It is
MLCT band in the region of 370 nm that tails into the
visible region. A very strong ligand field absorption is
found at uB300 nm. The band maxima for the MLCT
transition were found to be sensitive to the electron
donor ability of the phosphorus ligand with the order
being CO\P(OPh)3\PPh3\PBu3. The carbonyl
stretching bands within this series decrease in the order
Table 4
IR spectral data (cm−1) for Co2(CO)5(L)(alkyne), Co2(CO)4(PBu3)2(C2H2) and Co2(CO)4(DPPM)[(C2H5)2C2] in frozen Nujol at ca. 90 K a
Compound
Parent spectrum
Photoproduct
Anneal products
Co2(CO)5(PBu3)(C2H2) (VI)
2064 (w)
2008 (s)
2001 (m)
1989 (w)
1966 (s)
2067 (s)
2014 (s)
2005 (s)
1993 (w)
1970 (w)
2075 (s)
2025 (s)
2012 (s)
2000 (w)
1985 (w)
2053 (w)
2034 (s)
1988 (s)
1971 (m)
1948 (m)
2053 (w)
2038 (s)
1994 (s)
1977 (m)
1954 (m)
2047 (w)
2016 (s)
2004 (s)
1993 (sh)
1990 (m)
1962 (m)
2051 (m)
2029 (s)
1990 (s)
1979 (w)
1966 (m)
1951 (m)
1992 (m)
1976 (m0
1937 (s)
1920 (w)
2005 (s)
1955 (s)
1931 (s)
Co2(CO)5(PPh3)(C2H2) (VII)
Co2(CO)5[P(OPh)3](C2H2) (VIII)
2052 (w)
2046 (w)
2016 (s)
1990 (m)
Co2(CO)5(PPh3)(Ph2C2) (IX)
2062 (s)
2016 (s)
2002 (m)
1994 (w)
1970 (w)
Co2(CO)4(PBu3)2(C2H2) (XI)
2019 (s)
1969 (m)
1960 (s)
1997 (m)
1949 (m)
1905 (w)
Co2(CO)4(DPPM)[(C2H5)2C2] (XII)
2020 (m)
1990 (s)
1964 (m)
1945 (w)
a Free CO band at 2132 cm−1 is observed in all photolyses.

12
T.E. Bitterwolf et al. / Journal of Organometallic Chemistry 605 (2000) 7–14
DPPM ligand of XI is known to span equatorial
positions.
reasonable to propose that carbonyl groups may be lost
from the Co(CO)3 unit to give rise to both possible
isomeric axial (C), and equatorial (D), CO-loss photo-
products. This implies that substitution of axial car-
bonyl by phosphine on one cobalt atom either
facilitates CO-loss from both the equatorial and axial
positions of the second Co(CO)3 unit, or lowers the
thermal barrier to rearrangement so that an equilibrium
mixture is formed. In no instance was the phosphorus
ligand lost upon photolysis.
Photolysis of X was found to result in decreases in
the intensities of its carbonyl bands, and the appear-
ance of new bands at 2131, 1991, 1976, 1937, and 1918
(sh) cm−1. Annealing of this sample resulted in loss of
‘free’ CO and regeneration of X and also appearance of
new bands at 1997, 1949, and 1905 cm−1
Photolysis of
.
X
demonstrates that even
a
Co(CO)2(PR3) unit may lose CO. The resulting species,
Co2(CO)3(PBu3)2(H2C2), would be expected to have
three distinct carbonyl stretching bands. The small
shoulder observed at 1918 cm−1 may be evidence for a
rearrangement of phosphine to an equatorial position
analogous to the axial to equatorial shift proposed for
A to B. This latter possibility is reinforced by the
appearance of a new species upon annealing that may
be a structural isomer of VIII.
To explore the effect of phosphine position on the
photochemistry we have examined the photolysis of
Co2(CO)4(DPPM)(C2Et2) (XI) in frozen Nujol. The IR
spectrum of the photolyzed sample exhibits reduction in
the intensity of the bands of the starting material and
growth of new bands at 2131, 2005, 1955, and 1930
cm−1. Annealing the sample results in simple reversal
of the photoreaction.
The bidentate DPPM ligand in XI is known to bridge
equatorial positions on the two cobalt atoms, thus
photolysis could result in loss of CO from either an
axial or equatorial position. Photolysis appears to re-
sult in only one of the two possible carbonyl-loss
isomers being formed.
In the specific case of VIII, annealing clearly results
in the conversion of one photoproduct isomer to the
other. It is significant that the isomer which is depleted
upon warming is the one with the lowest energy CO
stretching band (1962 cm−1). This parallels the behav-
ior of the photoproducts from I–V, and argues that the
symmetries of the pairs of species A and C, and B and
D are the same. It is also important to note that the
photochemistry and annealing behavior of the P(OPh)3
derivative is most like the parent carbonyl derivatives,
consistent with the p-acid character of phosphites.
To test the possibility of carrying out photochemical
substitutions on monophosphine compounds, a sample
of Co2(CO)5(PPh3(H2C2) was photolyzed in heptane
with one equivalent of PPh3. IR spectra of the reaction
mixture indicated that Co2(CO)5(PPh3)(H2C2), had been
formed. Concentration of the reaction mixture and
recrystallization from dichloromethane–heptane gave
the bis(triphenylphosphine) derivative in 10% yield.
As noted above, there was no evidence to suggest
phosphine or phosphite loss from the cobalt com-
pounds upon photolysis, nor was there any evidence
that CO had been lost from the Co(CO)5(PR3) unit. To
examine whether the presence of a phosphine sup-
presses CO loss we photolyzed two derivatives, contain-
ing two phosphine ligands. The phosphine ligands in X
are known to occupy axial positions while the bidentate
3. Conclusions
Photolysis into the MLCT or ligand field bands
of cobalt tetrahedrane compounds results in loss of
CO to form one or two species in which the tetra-
hedrane core of the molecule is retained. Hexacarbonyl
compounds appear to generate one species upon
photolysis and a second upon annealing of the sample.
Pentacarbonyllphosphine compounds appear to form
two isomers upon photolysis while the tetracar-
bonylldiphosphine compounds give one dominant
product.
These matrix observations indicate that the cobalt
tetrahedrane compounds have a rich, mostly untapped
photochemistry that may be useful in preparing deriva-
tives with mixed ligands or intermediates for Pauson–
Khand and acetylene trimerization reactions. As an
illustration we have photochemically prepared the pre-
viously unknown tetrahydrothiofuran derivative,
Co2(CO)5(SC4H8)(C2Et2). A detailed description of syn-
thetic applications of photochemistry to cobalt tetrahe-
drane compounds is in preparation.

T.E. Bitterwolf et al. / Journal of Organometallic Chemistry 605 (2000) 7–14
13
4.2. Synthesis of Co2(CO)4(PPh3)2(Ph2C2)
4. Experimental
Complex VII (32 mg, 59 mmol) and PPh3 (31 mg, 118
mmol), were placed in a 50 ml microscale quartz photo-
chemical cell outfitted with a cold finger. The cell was
evacuated and flushed with N2 and 20 ml of degassed
benzene was added. The sample was irradiated for 30
min with a 350 W high pressure Hg lamp and filter such
that 240BuirrB380 nm. A water filter was used to
remove infrared light to prevent warming of the sam-
ple. After photolysis the solvent as removed and the
sample recrystallized from dichloromethane–heptane to
yield the bis-phosphine derivative as red crystals. 4.7
mg, 10%. IR: (CH2Cl2) 2019, 1971 (sh), 1960, 1935
cm−1 [6b].
Matrix studies were performed using a liquid nitro-
gen cooled glass cryostat and techniques described pre-
viously [15]. Spectra are recorded on a Perkin–Elmer
Spectrum 1000 FTIR Spectrometer at 4 cm−1 resolu-
tion. Compounds were prepared by literature proce-
dures. Nujol (light mineral oil) was purchased from
Aldrich and used as received.
4.1. Nujol matrix experiments
Compounds to be studied are dissolved in Nujol
by grinding a 10 mg sample with 0.5 ml Nujol
followed by centrifugation for at least ten min to
ensure that all particles have been removed from the
solution.
Nujol solutions are applied to the CaF2 windows of
the sample cell and the cell assembled. The cell is
screwed onto the inner reservoir of the glass cryostat
and the cryostat is assembled. Dry nitrogen is flushed
through the cell to remove water vapor, then liquid
nitrogen is introduced into the reservoir and the cell is
evacuated using a vacuum line equipped with an oil
diffusion pump. Ultimate vacuums of 4×10−3 torr are
desired to avoid frosting of the exterior windows during
the experiments.
Acknowledgements
T.E.B. thanks Research Corporation for a generous
Research Opportunity Award for the purchase of an
FTIR Spectrometer, and the State of Idaho for an
SBOE Research Grant.
References
Photolyses are carried out using a 350 W high pres-
sure mercury lamp with optical filters used to control
the irradiation wavelengths. Band pass filters have a
roughly bell shaped window with the transmittance at
the band center of about 63% and a band width of
970 nm. In some cases, particularly in the ultraviolet,
pairs of filters are used together to narrow the actual
irradiation wavelengths.
[1] H. Greenfield, H.W. Sternberg, R.A. Friedel, J.H. Wotiz, R.
Markby, I. Wender, J. Am. Chem. Soc. 78 (1956) 120.
[2] (a) A.G. Melikyan, K.M. Nicholas, in: P.J. Stang, F. Diederrich
(Eds.), Modern Acetylene Chemistry, VCH-Weinheim, Ger-
many, 1995, pp. 99–138. (b) M.J. Went, Adv. Organomet.
Chem. 41 (1997) 69. (c) R.D.W. Kemmitt, D.R. Russell, in: G.
Wilkinson, F.G.A. Stone, E.W. Able (Eds.), Comprehensive
Organometallic Chemistry, vol. 5, Pergamon, Oxford, 1982
(Chapter 34). (d) R.S. Dickson, P.J. Fraser, Adv. Organomet.
Chem. 12 (1974) 323.
After recording an initial spectrum the sample is
photolyzed at uirr\600 nm. While it is rare to observe
photochemistry at such low energies we have found
that samples undergo minor photoannealing that may
shift vibrational bands by 1–2 cm−1. By conducting a
long wavelength photoannealing step we avoid prob-
lems associated with these band shifts. In a typical
photochemical series, samples are photolyzed for 30
min followed by recording of a spectrum. In the current
studies photolyses were carried out at (1) uirr\600, (2)
uirr=550970, (3) uirr=500970, (4) uirr=450970,
(5) uirr=400970, (6) 350BuirrB400, (7) 250BuirrB
380 nm.
[3] D.M. Hoffman, R. Hoffmann, C.R. Fisel, J. Am. Chem. Soc.
104 (1982) 3858.
[4] H. Van Dam, D.J. Stufkens, A. Oskem, M. Doran, I.H. Hillier,
J. Electron Spectrosc. Relat. Phenom. 21 (1980) 47.
[5] R.L. DeKock, T.V. Lubben, J. Hwang, T.P. Fehiner, Inorg.
Chem. 20 (1981) 1627.
[6] (a) E. Sappa, G. Predieri, L. Marko, I. Chem. Acta 228 (1995)
147. (b) J.J. Bonnet, R. Mathieu, Inorg. Chem. 17 (1978) 1973.
(c) G. Wradi, A. Vizi-Orosz, S. Vastag, G. Pa´lyi, J. Organomet.
Chem. 108 (1976) 225. (d) L.S. Chia, W.R. Cullen, M. Franklin,
A.R. Manning, Inorg. Chem. 14 (1975) 2521. (e) M. Basato, A.J.
Poe¨, J. Chem. Soc. Dalton Trans. (1974) 456.
[7] (a) P.H. Bird, A.R. Fraser, D.N. Hall, Inorg. Chem. 16 (1977)
1923. (b) T. Fukumoto, Y. Matsumura, R. Okawara, J.
Organomet. Chem. 69 (1974) 437.
[8] R.F. Heck, J. Am. Chem. Soc. 85 (1963) 657.
Annealing experiments involve removing the liquid
nitrogen from the reservoir and allowing the cell to
warm for about 10 min during which time the cell
temperature rises to about 140 K. The sample is re-
turned to 90 K for recording the spectrum.
[9] J.C. Anderson, B.F. Taylor, C. Viney, G.J. Wilson, J.
Organomet. Chem. 519 (1996) 103.
[10] J.A. Casalnuovo, N.E. Schore, Organometallic cycloaddition
reactions of acetylenes, in: P.J. Stang, F. Diederich (Eds.),
Modern Acetylene Chemistry, VCR-Weinheim, Germany, 1995,
p. 139.

14
T.E. Bitterwolf et al. / Journal of Organometallic Chemistry 605 (2000) 7–14
[11] S.W. Brown, P.L. Pauson, J. Chem. Soc. Perkin Trans. 1 (1990)
1205.
[15] T.E. Bitterwolf, K.A. Lott, A.J. Rest, J. Mascetti, J. Organomet.
Chem. 419 (1991) 113.
[12] R. Kru¨erke, W. Hu¨bel, Chem. Ber. 94 (1961) 2829. (b) R.
Giordano, E. Sappa, G. Prediri, Inorg. Chem. Acta 228 (1995)
139. (c) G. Gervasio, E. Sappa, L. Marko´, J. Organomet. Chem.
444 (1993) 203.
[13] B.L. Pagenkopf, T. Livinghouse, J. Am. Chem. Soc. 118 (1996)
2285.
[16] (a) G. Bor, S.F.A. Kettle, P.L. Stanghellini, Inorg. Chem. Acta
18 (1976) L18. (b) S.F.A. Kettle, P.L. Stanghellini, Inorg. Chem.
16 (1977) 75. (c) G. Bor, Chem. Ber. 96 (1963) 2644.
[17] (a) A.S. Foust, C.F. Campana, J.D. Sinclair, L.F. Dahl, Inorg.
Chem. 18 (1979) 3047. (b) A.S. Foust, M.S. Foster, L.F. Dahl, J.
Am. Chem. Soc. 91 (1969) 5633.
[14] C.M. Gordon, M. Kiszka, I.R. Dunkin, W.J. Kerr, J.S. Scott, J.
Gebicki, J. Organomet. Chem. 554 (1981) 147.
[18] T.E. Bitterwolf, J.T. Bays, W.B. Scallorn, C.A. Weiss, J. John-
son, Manuscript in preparation, 1999.
.