2
706 Organometallics, Vol. 23, No. 11, 2004
Chen et al.
higher distortion from the spheric symmetry of the
electric field gradient generated by its orbital electrons.
In our previous work, we were not able to observe the
Co NMR spectrum of Co2(CO)6[P(n-C4H9)3]2 either.
The absence of a 59Co resonance at δ -2200 ppm in
Calcd for C48
H
24
F
18
P
2
O
6
Co
2
: C, 47.31; H, 1.99; F, 28.06; P, 5.08;
3
1
Co, 9.67; O, 7.88. The compound exhibits a P resonance at δ
8.6 ppm (in toluene, 85% aqueous H PO at 0 ppm). Note on
6
3
4
5
9
the elemental analysis: incomplete liberation of fluorine was
encountered in the initial analysis by oxygen flask combustion.
The correct fluorine analysis was obtained by a more vigorous
decomposition technique, pyrolysis at 1100 °C. The analyses
were carried out by Galbraith Laboratories, Inc., Knoxville,
TN.
Figure 5 and in Figure b1-b3 in Figure 7 clearly
indicates the absence of Co2(CO)8 in these solutions. It
follows then that HCo(CO)4 (5) is also absent, or is a
very minor species, in these solutions, since Co2(CO)8
and 5 are in equilibrium under these reaction condi-
Solu bility of Co (CO) [P (p-CF C H ) ] in scCO . This
experiment was carried out to assess the density of scCO
2
6
3
6
4
3
2
2
2
5
9
tions. These results support our assignment of the Co
needed to dissolve a sufficient amount of 1 at ∼100 °C to study
the hydroformylation reaction. Preliminary experiments in-
resonance at δ -3010 ppm in Figure 5a to 4 rather than
dicated that 1 has limited solubility in scCO
of CO . The high-pressure cell, 9.5 mL in volume, was loaded
with 40 mg (33 µmol) of 1 and 275 mg (330 µmol) of
P(p-CF (L), the latter being added for calibration and
for suppression of its dissociation from 1. The cell was charged
with 7.5 atm of CO and 7.5 atm of CH at 35 °C, followed by
2
at a low density
5
. Compounds 4 and 5 happen to have nearly identical
2
chemical shifts.
Co2(CO)6[P(p-CF3C6H4)3]2 was found to catalyze the
hydroformylation of ethylene and propylene. In a solu-
tion containing a large excess of P(p-CF3C6H4)3 (62 mM),
the compound has a catalytic activity which is about
3
6 4 3
C H )
4
the addition of CO2 to bring the total pressure to 86.3 atm.
Methane was added for use in shimming the magnet. The
solubilities of the cobalt dimer at 80-130 °C were obtained
by comparing its integrated areas against that of free L after
equilibrium was reached. The results are shown in Figure 1.
1
/10 of that of Co2(CO)8. However, it is more stable
under low CO pressures, and it allows the hydroformyl-
ation reaction to be carried out to the complete depletion
3
1
59
of CO. The P and Co NMR spectra of the solutions
clearly indicate that HCo(CO)3[P(p-CF3C6H4)3] (4) is the
predominant, if not the only, cobalt complex present
under the reaction conditions. However, due to the
lability of all of these cobalt complexes, it is not possible
to rule out a contribution from the more reactive HCo-
Lin e Wid th of th e 31P Reson a n ce of P (p-CF
C
H
)
.
3
6
4
3
3
Early experiments indicated that the line width of the 1P
resonance of L changed with the concentration of 1. A set of
four experiments were carried out to correlate the line width
31
of the P resonance of L with the concentration of 1. The cell
(10.0 mL) was loaded with 154 ( 1 mg (33.0 mM) of L and
(CO)4 simply because it was not observed. Our attempts
the desired amounts of 1, which were 3.5 mg (0.29 mM), 8.9
mg (0.73 mM), 22.0 mg (1.81 mM), and 40.0 mg (3.28 mM),
respectively. Carbon monoxide and methane were introduced
at room temperature to give each a pressure of about 6.8 atm.
Carbon dioxide was then added to give a total pressure of ∼120
atm. The pressure was adjusted to 306 atm by the addition of
to establish the case for 4 as the catalyst for the
hydroformylation of propylene by the altered selectivity
of n-butyl aldehyde/isobutyl aldehyde turned out to be
negative. The n-butyl aldehyde/isobutyl aldehyde ratio
for 1 is similar to that obtained for Co2(CO)8. It is
possible that the unsubstituted and the phosphine-
substituted catalyst systems have similar product se-
lectivities, but it is also possible that the main effect of
the addition of the phosphine ligand is simply the
reduction of the concentration of the more active HCo-
more CO
2
after the cell was equilibrated at 100 °C. As a
background check, we have shown that, in the absence of 1,
3
1
the line width of the P resonance of L is 20 ( 1 Hz,
independent of the temperature and the density of CO
H yd r ogen a t ion of Co (CO) [P (p -CF (1) in
scCO . The high-pressure cell containing 21 mg of Co (CO)
[P(p-CF C H ) ] and 37.5 mg of P(p-CF C H ) was charged
2
.
2
6
3 6 4 3 2
C H ) ]
2
2
6
-
(CO)4.
3
6
4
3
2
3
6
4 3
2 2
with 7.9 atm of CO and 8.2 atm of H at 35 °C; CO was then
added to bring the pressure to 168 atm. The temperature was
then raised to 100 °C, at which point the pressure reached
470 atm. The 31P NMR spectra of the solution at 100, 80, 60,
and 40 °C, shown in Figure 4, were recorded after the solution
was allowed to reach equilibrium at the respective tempera-
tures.
The 59Co NMR spectrum was also recorded at 100 °C. For
comparison, a 59Co NMR spectrum was recorded for a reaction
carried out under different reaction conditions without added
CO. These spectra are shown in Figure 5.
Exp er im en ta l Section
Gen er a l Con sid er a tion s. A General Electric GN300/89
NMR spectrometer was used for the in situ NMR studies. The
high-pressure probe used in this study has been described
1
0
previously. Carbon monoxide, hydrogen, ethylene, propylene,
and carbon dioxide were of research grades from AGA Spe-
cialty Gases and Equipment, and were used as received.
Dicobalt octacarbonyl and tris(p-(trifluoromethyl)phenyl)phos-
phine from Strem Chemicals were used without further
purification.
Hyd r ofor m yla tion of Eth ylen e a n d P r op ylen e Ca ta -
Co
sized by the reaction of Co
amount of P(p-CF in toluene. To a flask containing
10 mg (0.61 mmol) of Co (CO) and 569 mg (1.22 mmol) of
P(p-CF under a nitrogen atmosphere was added 12.0
mL of deaerated toluene at ambient temperature. Stirring of
the solution caused immediate gas evolution, and a dark brown
solution was obtained. Dark brown precipitate gradually fell
out of the solution during the following 1 h. The solution was
heated at 60 °C for 1 h and then cooled for the product to
crystallize out of the solution. The precipitate was collected
on a sintered-glass filter, washed with petroleum ether, and
dried. Yield: 350 mg. The crude product was further purified
by recrystallization from toluene. It is important to filter the
solution to remove any metallic cobalt produced in the process.
Anal. Found: C, 47.89; H, 2.36; F, 27.82; P, 5.00; Co, 9.30.
2
(CO)
6
[P (p-CF
3
C
6
H
4
)
3
]
2
(1). The compound was synthe-
lyzed by Co
pressure cell containing 40 mg (33 µmol) of 1 and 275 mg (590
µmol) of P(p-CF was charged with 21.7 atm of H , 23.1
atm of CO, and 20.5 atm of C at 35 °C in the order listed,
followed by the addition of CO to give a total pressure of 178
atm. The cell was then heated to 100 °C and the reaction
2 6 3 6 4 3 2 2
(CO) [P (p-CF C H ) ] (1) in scCO . A high-
2
(CO) with the stoichiometric
8
1
6
C
6
H
4
)
3
3
C
6
H
4
)
3
2
3
2
2
8
2 4
H
3
C
6
H
4
)
3
2
3
1
1
followed by recording one P and one H spectrum every 2 h.
Integration of the 1H resonances of H
and C gives the
initial mole ratio of C /H ) 1.85. This higher ratio, as
compared to that calculated from the pressure increments in
2
2 4
H
2
H
4
2
1
3
2 4
the loadings, is due to the compressibility of C H . A C NMR
spectrum was recorded at the end of the reaction.
The hydroformylation of propylene was carried out in a
similar fashion. The high-pressure cell containing 12 mg (9.9
µmol) of 1 and 275 mg (590 µmol) of P(p-CF
3
C
6
H
4
)
3
was
/CO
charged with 9.9 atm of C and 47.8 atm of a 1:1 H
3
H
6
2