Ligand exchange processes of OsHCl(CO)(L)(PR3)2 (L = vacant, H2, R'CN, O2; R = Cy, i-Pr)

Article abstract of DOI:10.1016/S1381-1169(98)00005-3

The reactivity of complexes formed by the addition of O₂, H₂ and R'CN to OsHCl(CO)(PR₃)₂ (1a:R = Cy; 1b:R = i-Pr) has been examined. Under 24 bar H₂ and 65°C, the dioxygen ligand of OsHCl(COXO₂XPR₃)₂ (2a,b) is displaced to yield the trans-hydridodihydrogen complexes OsHCl(η²-H₂)(CO)(PR₃)₂ (3a,b). Measurements of the equilibrium constant, K(H₂) = [3a]/[1a][H₂], for the direct addition of H₂ to 1a yield ΔH°= -49.1 ± 2.4 kJ/mol and ΔS°= -95.7 ± 7.9 J/mol K. 1a,b react reversibly with aryl and alkyl nitriles to produce the isolable complexes, OsHCl(CO)(R'CN)(PR₃)₂ (4a,b). The phosphine ligands of 1a,b and 3a,b exchange with unbound, bulky alkyl phosphines at a rate that is slow relative to the NMR timescale. In the presence of excess PCy₃, complex 3b yields the exchange products OsHCl(η₂-H₂)(CO)(Pi-Pr₃)(PCy₃) and 3a. While a tris-phosphine complex cannot be detected, limited kinetic data characterize the exchange as associative process.

Full text of DOI:10.1016/S1381-1169(98)00005-3

Ž
.
Journal of Molecular Catalysis A: Chemical 135 1998 285–293
ž /ž /ž
/
Ligand exchange processes of OsHCl CO L PR3 2
X
ž
/
Lsvacant, H2, R CN, O2; RsCy, i-Pr
J. Scott Parent a,1, Neil T. McManus , Garry L. Rempel a,), William P. Power ,
a
b
b,2
Todd B. Marder
a
Department of Chemical Engineering, UniÕersity of Waterloo, Waterloo, ON, Canada N2L 3G1
b
Department of Chemistry, UniÕersity of Waterloo, Waterloo, ON, Canada N2L 3G1
Received 19 September 1997; accepted 23 December 1997
Abstract
X
Ž
.Ž .
. Ž
The reactivity of complexes formed by the addition of O2, H2 and R CN to OsHCl CO PR3 1a:RsCy; 1b:Rsi-Pr
2
Ž
.Ž .Ž
. Ž
.
has been examined. Under 24 bar H2 and 658C, the dioxygen ligand of OsHCl CO O2 PR3 2a,b is displaced to yield
2
2
Ž
.Ž .Ž
.
Ž
.
the trans-hydridodihydrogen complexes OsHCl h -H2 CO PR3
3a,b . Measurements of the equilibrium constant,
2
w
x w xw
x
KH s 3a r 1a H2 , for the direct addition of H2 to 1a yield D H8sy49.1"2.4 kJrmol and DS8sy95.7"7.9 Jrmol
2
X
Ž
.Ž
.Ž
. Ž
.
K. 1a,b react reversibly with aryl and alkyl nitriles to produce the isolable complexes, OsHCl CO R CN PR3 4a,b . The
2
phosphine ligands of 1a,b and 3a,b exchange with unbound, bulky alkyl phosphines at a rate that is slow relative to the
2
Ž
.Ž .Ž
.Ž
.
NMR timescale. In the presence of excess PCy3, complex 3b yields the exchange products OsHCl h -H2 CO Pi-Pr3 PCy3
and 3a. While a tris-phosphine complex cannot be detected, limited kinetic data characterize the exchange as associative
process. q 1998 Elsevier Science B.V. All rights reserved.
Keywords: Osmium; Hydrogenation; High pressure NMR; Polymer modification; HNBR
1. Introduction
dienes through the activation of molecular hy-
w
x
w
x
drogen 3–8 or by hydrogen transfer 9,10 . We
have recently discovered the exceptional effi-
ciency with which complexes of the form Os-
While the catalytic potential of mononuclear
osmium complexes may have been overlooked
in the past, new research has demonstrated their
ability to facilitate a variety of reactions 1 .
These include the hydrosilylation 2 of alkynes
Ž
.Ž .Ž
.
Ž
HCl CO L PR3 2 1: Lsvacant; 2: LsO2;
w x
.
a: RsCy; b: Rsi-Pr catalyze the selective
hydrogenation of acrylonitrile-butadiene copoly-
w x
as well as the hydrogenation of alkenes and
w
x
mers 11 . Remarkably, these complexes are
more active at industrial conditions than the
current generation of Ru, Rh and Pd systems
)
Corresponding author.
Present address: Department of Chemical Engineering,
1
w
x
12 . Nevertheless, this class of hydrogenation
catalysts have received limited attention, espe-
cially at the more severe conditions employed in
Queen’s University, Kingston, Ontario, Canada, K7L 3N6.
2 Present address: Dept. of Chemistry, University of Durham,
South Road, Durham, England DH1 3LE, UK.
1381-1169r98r$ - see front matter q 1998 Elsevier Science B.V. All rights reserved.
Ž
.
PII: S1381-1169 98 00005-3

(
)
286
J.S. Parent et al.rJournal of Molecular Catalysis A: Chemical 135 1998 285–293
Ž
.
the HNBR industry P)20 bar, T)1008C
gen adducts 2a,b when used as catalyst precur-
sors as well as the propensity of nitrile to
coordinate with 1a,b. Furthermore, the PCy3
analogue of 3b remains uncharacterized in spite
of its superior catalytic activity. In this short
paper we present high-pressure 31P and 1H NMR
data that furthers our knowledge of these mat-
ters. Preliminary studies of the lability of phos-
phine ligands within the 3a,b system have helped
to clarify their role in the hydrogenation pro-
cess.
w
x
13 .
Owing to their coordinative unsaturation, 1a,b
readily add small Lewis bases to form stable,
six coordinate complexes. Moers 14 , Moers
and Langhout 15 , and Moers et al. 16,17
have demonstrated the coordinative capacity of
pyridine, SO2, CS2 and CO to 1a. In addition to
the CO analogue of 1b, Esteruelas and Werner
18 have characterized the products of PMe3
and P OMe addition. It is also known that O2
w
x
w
x
w
x
w
x
Ž
.
3
adds to 1a,b to form the corresponding dioxy-
w
x
gen adducts 2a,b 19,5 , which possess superior
stability as solids. While unactivated olefins
2. Experimental
w
x
such as ethylene coordinate only weakly 19 ,
Ž
electron withdrawing substituents CH2 5CHR;
.
Rs –CN, –CO2 Me or –COR stabilize the
Os–olefin bond sufficiently to yield isolable
Manipulation of solvents, reagents and sam-
ples was carried out with rigorous exclusion of
air using either a Vacuum Atmospheres dry box
w
x
octahedral complexes 18,20 .
Ž
Complex 1b forms a dihydrogen complex
or standard Schlenk techniques. H2 99.9% pure,
3b through the h2-coordination of H2
-1 ppm O2 from Linde-Union Carbide, puri-
Ž
w
.
.
x
5,21,22 . Consistent with the addition reactions
described above, dihydrogen coordinates trans
to the hydride to create the Os II complex
Scheme 1 . Since the removal of an H2 atmo-
fied monochlorobenzene and toluene from Fis-
cher Chemicals and deuterated solvents from
Cambridge Isotopes were used as received.
Benzonitrile, PCy3, Pi-Pr3, PCp3 and Pt-Bu3
from Strem Chemicals were used without fur-
ther purification. Fully saturated acrylonitrile–
butadiene copolymer was prepared by hydro-
genating Krynac 38.50 from Bayer 38% acryl-
onitrile, Mn s70,000, polydispersitys3.6 us-
ing 2a as a catalyst.
Variable-temperature NMR spectra were
recorded on a Bruker AC-200 NMR spectrome-
ter operating at 80.9 MHz and 200.1 MHz for
31P and 1H, respectively. 1H chemical shifts
were referenced to TMS using the residual
methyl protons of toluene-d8 or the aromatic
resonance of benzene-d6. 31P signals were refer-
enced using 85% H3PO4 as an external stan-
dard. Temperatures above ambient were cali-
brated against an 80% solution of ethylene gly-
col in DMSO while those below 258C were
measured using a 4% solution of methanol in
CD3OD. High-pressure spectra were recorded
using a prototype of the apparatus developed by
Ž .
Ž
.
sphere rapidly transforms 3b into its five-coor-
dinate precursor, this complex has yet to be
isolated. However, the structure and dynamics
of 3b in solution have been studied in detail by
Bakhmutov et al. 21 using NMR. In addition
to a rapid exchange of free and bound H2, 3b
undergoes fast hydrido-dihydrogen ligand ex-
change relative to the NMR timescale at tem-
peratures greater than 108C.
At present, our understanding of the reactions
that constitute an acrylonitrile–butadiene
copolymer hydrogenation remains incomplete,
particularly at high pressures and temperatures.
Outstanding issues include the fate of the dioxy-
Ž
w
x
.
w
x
Roe 23 . This unit, consisting of a 5 mm sap-
Scheme 1.

(
)
J.S. Parent et al.rJournal of Molecular Catalysis A: Chemical 135 1998 285–293
287
1
Ž
.
Ž .
phire tube epoxied to a titanium valve assembly,
was inserted directly into a standard NMR probe.
Found for 4b: H NMR C6D6 : d y12.6 t ,
JP – H s18.5 Hz; 31P H NMR: C6D6 : d 24.1
s ; IR C6H6 : n CO 1885 cm , n CN 2232
2
1
Ä
4
Ž
.
y
1
Ž .
Ž
.
Ž
.
Ž
.
1
cmy
.
2.1. Preparation of known compounds and com-
plexes
2.3. Measurement of equilibrium constants, KH
and KCN
2
To synthesize PMet-Bu2, a slight excess of
BuLi in hexane was added dropwise to a THF
solution of PHt-Bu2 at 08C. CH3I in THF was
then added slowly to generate the product in
good yield. This THF solution of PMet-Bu2
was washed with water, decanted and dried with
MgSO4 before removing the solvent by distilla-
tion.
The methodology used to measure K
for
H2
the 1a–3a equilibrium was developed by
w
x
Bakhmutov et al. 21 . To similarly characterize
the 1a–4a equilibrium, a 31P NMR variation of
the technique was employed. A monochloroben-
zene solution containing 0.0127 M of 1a and
X
0.0119 M of R CN from hydrogenated acryloni-
Complexes 1a and 1b were prepared by re-
trile–butadiene copolymer was charged to a
sapphire NMR tube and sealed under N2. The
Ž
.
fluxing OsCl3 P3H2O Strem Chemicals with
the required phosphine in methoxyethanol and
ethanol, respectively, and isolated by the proce-
w
x w x
ratio, Xs 4a r 1a was determined from the
chemical shift of the exchanged averaged phos-
phine resonance according to:
w
x
dure of Esteruelas and Werner 18 . Found for
1
31
1
Ž
.
Ž
.
Ä
4
1a: H NMR C6D6 : d y32.6 br. t ; P H
Ž
.
Ž .
Ž
.
Ž
.
4a
1a
d avg yd 1a
Ž . Ž
NMR: C6D6 : d 36.7 s ; IR C6H6 : n CO
.
.
1
1
1886 cmy . Found for 1b: H NMR C6D6 : d
Xs
s
.
Ž
.
.
d 4a yd avg
Ž
.
Ž
31
1
Ž
.
Ä
4
Ž
y30.4 br. t ; P H NMR: C6D6 : d 48.0
Material balances on the total osmium and ni-
trile charged to the system yield:
y
1
Ž .
Ž
.
Ž
.
s ; IR C6H6 : n CO 1889 cm
.
Complexes 2a and 2b were prepared by ex-
posing hexane suspensions of 1a and 1b, re-
Os s 1a q 4a
T
X
X
R CN s 4a q R CN .
spectively, to pure O2 as detailed by Esteruelas
T
1
w x
Ž
.
et al. 7 . Found for 2a: H NMR C6D6 : d
It follows that the equilibrium constant KCN
may be expressed as:
2
31
1
Ž .
Ä
4
y2.5 t , JP – H s30.4 Hz; P H NMR:
Ž
.
Ž .
Ž
.
Ž
.
C6D6 : d 13.7 s ; IR nujol : n CO 1948
X
1
1
cmy . Found for 2b: H NMR C6D6 : d y2.43
KCN
s
.
Ž
.
X
2
31
R CN yXr 1qX Os
1
Ž
.
T
T
Ž .
Ä
4
Ž
.
t , JP – H s30.1 Hz; P H NMR: C6D6 : d
y
1
Ž
.
Ž
.
Ž .
Ž
.
Ž
.
d 1a and d 4a were measured at 23, 70, 110,
1328C using a solution of pure 1a and one
containing 1a plus 60 equivalents of nitrile,
respectively. Interpolation of these data pro-
23.4 s ; IR nujol : n CO 1945 cm
.
(
)(
)(
)
2.2. Preparation of OsHCl CO PhCN PR3 2
(
)
4a: RsCy; 4b: Rsi-Pr
Ž
.
Ž
.
vided the estimates of d 1a and d 4a required
to calculate KCN at each temperature.
The formation constants D H8 and DS8 for
the 1a–3a and the 1a–4a equilibria were calcu-
lated from a non-linear, least-squares regression
of
A suspension of ca. 0.3 g 1a or 1b in 5 ml
Ž
.
hexane was treated with an excess ca. 0.2 ml
of PhCN at 238C. The mixture was stirred for
10 min, resulting in a colour change from brown
to yellow. The solution was decanted and the
Ž
.
solid washed with MeOH 2=5 ml and diethyl
DS8
D H8
Ž
.
ether 1=5 ml and dried under vacuum. Found
Ks
exp y
.
1
2
ž
/
R
RT
Ž
.
Ž .
for 4a: H NMR C6D6 : d y12.3 t , JP – H
s
31
1
Ä
4
Ž
.
Ž .
19.0 Hz; P H NMR: C6D6 : d 14.5 s ; IR
Estimates of D H8 and DS8 are accompanied by
95% confidence intervals.
y
1
y1
Ž
.
Ž
.
Ž
.
C6H6 : n CO 1883 cm , n CN 2230 cm
.

(
)
288
J.S. Parent et al.rJournal of Molecular Catalysis A: Chemical 135 1998 285–293
2.4. Measurement of phosphine exchange rates
To a known mass of 1b was added a solution
of the required phosphine in toluene-d8. Once
charged to the sapphire NMR tube, the N2
atmosphere was purged with hydrogen and an
H2 pressure of 24 bar was applied. At 238C, no
phosphine exchange could be detected during
this sample preparation. Therefore, time zero
was marked by the insertion of the sample into
the preheated NMR probe. The concentrations
of exchange products were measured at 2-h
31
1
Ä
4
intervals by P H NMR using an inverse-
gated pulse sequence. Estimates of k1 and ky1
were derived from a least-squares regression of
31
Ä1
4
Fig. 1.
P
H NMR spectra illustrating the transformation of 2b
Ž .
a Complex 2b at 258C, 1 atm
to 3b; PH s24 bar, T s658C.
N2. b Complex 2b at 658C, 24 bar H2 after 1 min.
2
Ž .
w
x
d 3b rdt vs. product concentrations as per Eq.
Ž .
1 of the text.
for the PCy3 analogues. Consequently, the cat-
alytic behaviour of 2a,b is expected to be iden-
tical to that observed for the five-coordinate
complexes, 1a,b.
3. Results and discussion
3.1. Fate of 2a,b exposed to hydrogenation
conditions
Over the time frame of minutes, a colour
change from pale yellow to deep violet is ac-
companied by the degradation of the NMR sig-
nal. While detectable levels of phosphine oxide
can be observed, the structure and composition
of the resulting osmium complex are unknown.
The instability of these dioxygen adducts in
solution is consistent with the observations of
At ambient temperature, O2 will displace
Ž
.Ž .Ž
.
olefin or H2 from OsHCl CO L PR3 2 to pro-
w
x
duce the dioxygen adducts 2a,b 19,5 . The air
stability of these solids makes them convenient
catalyst precursors, and 2b has been used as
such for the hydrogenation of phenylacetylene
w
x
w
Ž
.Ž
.
xw
x
Mezzeti et al. 24 for OsCl O2 dcpe BPh4 .
2
w x
3 . In this application, an observed induction
However, the effect of dissociated O2 on indus-
trial hydrogenations is likely to be negligible,
due to the insignificant concentration of O2
created by catalytic amounts of 2a,b when
charged to a high H2 pressure environment.
period was attributed to the displacement of O2
by the alkyne. We have observed that refluxing
2a with excess PCy3 in methoxyethanol regen-
erates 1a in good yields, indicating that while it
may be strongly bound, O2 can be dissociated
to yield a five coordinate complex.
3.2. Thermodynamics of H2 addition to 1a
Complex 1a reacts with H2 to produce a
Fig. 1 illustrates the response of 2b when
exposed to hydrogenation conditions that ap-
proach those applied industrially. Under N2 at
trans-hydrido-dihydrogen complex, OsHCl-
238C, the 31P H NMR spectrum of 2b consists
of a singlet at 23.4 ppm. Exposed to 24 bar H2
at 658C, the 2b signal is quickly replaced by the
singlet at 36.1 ppm that is characteristic of 3b.
This transformation of 2b to 3b is confirmed in
CO h -H2 PCy3 2 3a . This complex is the
1
2
Ä
4
Ž
.Ž
.Ž
.
Ž
.
PCy3 analogue of the Pi-Pr3 system previously
w x
synthesized by Esteruelas et al. 7 . The interac-
tions of H2 with 1b have been studied in detail
w
x
by Bakhmutov et al. 21 in terms of the rate of
H2 exchange and the equilibrium distribution of
1
the H NMR spectrum and has been observed

(
)
J.S. Parent et al.rJournal of Molecular Catalysis A: Chemical 135 1998 285–293
289
Scheme 2.
methodology to similarly characterize the 1a–3a
equilibrium.
The relative abundance of 1a and 3a was
determined by 1H NMR from the exchange
averaged chemical shift of the hydride reso-
Ž
.
nance Fig. 2a while the distribution of free
and coordinated H2 was measured from the
position of the exchange averaged dihydrogen
Ž
.
resonance Fig. 2b . The equilibrium constants
Ž
w
x w xw
x.
KH s 3a r 1a H2 calculated from these
spectra are listed in Table 1. Least squares
2
regression of the response of KH to changes in
2
temperature yields D H8sy49.1"2.4 kJrmol
and DS8sy95.7"7.9 JrmolPK. These com-
pare with reported values of D H8sy59.0"
2.1 kJrmol and DS8sy125.0"4.2 JrmolPK
w
x
for the 1b–3b equilibrium 21 .
Fig. 2. 1H NMR spectra of the 1a–3a system under rapid ex-
2
Ž .
Ž .
Ž
.
change; a Hydride region, b Os h -H2 rfree H2 region.
3.3. ReactiÕity of 1a,b with nitriles
The deleterious influence of nitrile on hydro-
genation activity has been reported for a number
of efficient catalyst systems. Mohammadi and
Ž
.
products Scheme 1 . Having identified the cor-
responding exchange process for the PCy3 sys-
tem, we have adopted their experimental
w
x
Rempel 25 have observed the inhibition of
Table 1
1H NMR study of the 1a–3a equilibrium; KH vs. temperature
2
Temperature
davg H2
davg Os–H
KH
2
y1
Ž .
K
Ž
.
Ž
.
Ž
M
.
ppm dihydrogen region
ppm hydride region
307
340
360
374
394
411
3.04
3.33
3.63
3.86
4.02
4.12
y6.85
y8.85
y11.80
y15.35
y19.55
y23.80
2345
375
122
51.5
25.6
15.1
Os T s6.68=10y3 M.
w
x
Ž
.
dH free s4.50 ppm.
2Ž
.
dH bound sy1.52 ppm.
d1a2sy32.65 ppm.
d3a sy6.31 ppm.

(
)
290
J.S. Parent et al.rJournal of Molecular Catalysis A: Chemical 135 1998 285–293
.
ppm; 4b: y12.6 ppm suggesting that nitrile is
Ž
.
trans disposed Scheme 2 as is the case with
other addition reactions of this system. Prece-
dents for nitrile complexes of this type are
w
x
abundant. Gilbert and Wilkinson 28 have pre-
Ž
.Ž
.
pared RuCl2 PhCN PPh3 3 by heating under
Ž
.
reflux RuCl2 PPh3 3 with benzonitrile. A recent
report by Schlaf et al. 29 has defined the
reactivity of the osmium–nitrile complex
w
x
2
w
Ž
.Ž
.Ž
.
xw
x
Os h -H2 MeCN dcpe BPh4 2. In both of
2
these cases, s-donation of the nitrogen lone pair
is the accepted mode of nitrile coordination.
The relative stabilities of the nitrile and H2
31
1
Ä
4
complexes of 1 are demonstrated by P H
31
Ä1
4
NMR study of the 3b–4b
NMR spectra of 1b containing 1.25 equivalents
Fig. 3. Variable temperature
P
H
equilibrium; 1.25 equiv. PhCN.
Ž
.
of PhCN Fig. 3 . At y48C and 53 bar H2, two
singlets corresponding to 3b and 4b are re-
solved. Considering the intensity of the signals,
the coordination of nitrile is favoured at this
temperature. Reducing the system pressure by
half produces the expected shift of the equilib-
rium towards the nitrile complex. Warming the
sample to 298C produces two exchange broad-
ened signals, one at 32.5 ppm and a more
pronounced resonance at 24.5 ppm. These coa-
lesce at 508C as the system reaches the fast
exchange limit of the NMR timescale. At 708C
and beyond, the chemical shift of the exchanged
averaged 3bl4b signal is evidence of an equi-
table distribution of the two complexes, in spite
of the higher concentration of H2 relative to
Ž
.
RhCl PPh3 3 by the nitrile within NBR as have
Martin et al. 26 for the RuHCl CO PCy3 2
system. In a study of OsH2 dcpe , Farneti et
w
x
Ž
.Ž
.
Ž
.
2
w
x
al. 27 attributed the relatively low rate of
PhHC5CH–CN hydrogenation to a competitive
coordination of nitrile to the metal centre. We
have recently demonstrated a similar influence
of nitrile on the activity of 2a 12 . To establish
the mode and significance of nitrile coordina-
tion, we have isolated and characterized alkyl
and aryl nitrile complexes of 1a,b.
Solutions of 1a,b rapidly transform from
brown to yellow with the addition of one equiv-
alent of PhCN. A sharp new singlet appears in
the 31P NMR spectra at 14.5 ppm for 4a and
w
x
X
R CN. This suggests that during the hydrogena-
24.1 ppm for 4b. The hydride resonance in the
tion of a nitrile copolymer, complexes 3 and 4
are the predominant chemical species.
1
Ž
H NMR spectrum shifts downfield 4a: y12.3
Table 2
31
Ä1
4
P
H NMR study of the 1a–4a equilibrium. Hydrogenated acrylonitrile–butadiene copolymer in monochlorobenzene
w
x
Temperature
d1a
d4a
)
Temperature
dqavg
Xs 4a r
KCyN 1
Ž .
K
Ž
.
Ž
.
Ž
.
K
Ž
.
w
x
Ž
M
.
ppm
ppm
ppm
1a
296
323
343
383
405
37.17
37.37
37.55
37.78
37.95
13.86
y
14.28
14.68
15.03
351
367
388
406
18.50
21.04
25.84
29.76
4.595
2.560
1.085
0.503
3050
914
204
65.6
X
w
x
w
x
) 1a s0.0161 M; R CN s0.6450 M.
X
qw
x
w
x
1a s0.0127 M; R CN s0.0119 M.

(
)
J.S. Parent et al.rJournal of Molecular Catalysis A: Chemical 135 1998 285–293
291
Ž
.
In the fast exchange domain T)708C the
31P chemical shift of the exchanged averaged
1al4a resonance reflects the relative abun-
dance of these complexes in solution. We have
exploited this principle to derive the equilibrium
X
w
x w xw
x
constant, KCN s 4a r 1a R CN , for the addi-
tion of the alkyl nitrile found in hydrogenated
acrylonitrile–butadiene copolymer to 1a. The
chemical shifts of complexes 1a and 4a and the
exchange averaged 1al4a resonance are listed
in Table 2 along with the calculated equilibrium
constants. The influence of temperature on KCN
yields, for the formation of 4a from 1a and
hydrogenated acrylonitrile–butadiene copoly-
mer, D H8sy81.2"2.8 kJrmol and DS8s
y165"7.3 JrmolPK.
Fig. 4. Exchange product concentrations vs. time; PH s24 bar,
w
x
Ž
.
Ž 2
.
T s718C, PCy3 0 s0.32 M. v Os Pi-Pr3 2 , B Os PCy3 2 , '
Ž
.Ž .
.
Ž
.Ž
Os Pi-Pr3 PCy3 , % Pi-Pr3; OssOsHCl CO H2 .
The influence of phosphine on the hydro-
genation activity of 2a could also result from an
unfavourable shift of a required PCy3 dissocia-
tion equilibrium. However, 31P spectra of 1a,b
recorded in toluene at 1008C under N2 or H2
showed no sign of dissociated phosphine. While
it therefore appears that direct evidence to sup-
port either inhibition mechanism is unavailable,
the phosphine ligands bound to 1a,b and 3a,b
are labile. We have observed that charging a
second bulky, alkyl phosphine to a solution of 1
or 3 yields a mixture of the original complex,
the monosubstituted complex and the disubsti-
3.4. Phosphine exchange processes of 3b
The results of our catalytic research have
shown that one equivalent of PCy3 relative to
w
x
2a reduced the hydrogenation activity by an
w
x
order of magnitude 12 . While central to the
hydrogenation mechanism, this result is difficult
to rationalise. One possibility is the coordina-
tion of a third phosphine to the coordinatively
unsaturated complex, 1a. While the PPh3 ana-
logue of 1a is the tris-phosphine complex, Os-
Ž
.Ž
.
w x
HCl CO PPh3 3 30 , all attempts by Moers
w
x
Ž
.Ž
.
w
14 to prepare OsHCl CO PCy3 3 were unsuc-
cessful. Esteruelas and Werner 18 have iso-
lated products of PMe3 and P OMe addition
x
.
Ž
3
to 1b, but observed no similar coordination of a
third Pi-Pr3 ligand. Consistent with these re-
ports, we observed no change in the room tem-
perature 31P NMR spectra of 1a or 1b when
exposed to an excess of their respective phos-
phines.
w
x
Fig. 5. 3b vs. time for different phosphines; PH s24 bar,
2
w
x
w
x
T s718C. ` PtBu3 0 s0.13 M, % PMet-Bu2 o s0.13 M, B
w
x
w
x
w
x
PCy3 0 s0.11 M, l PCy3 0 s0.10 M, v PCy3 0 s0.32 M,
w x
' PCp3 0 s0.11 M.
Scheme 3.

(
)
292
J.S. Parent et al.rJournal of Molecular Catalysis A: Chemical 135 1998 285–293
Table 3
Phosphine exchange rate constants
y1
y1
w
x
Ž
.
²
:
Ž
.
² :
Lower 95% Upper
Phosphine
PR3
M
k1 M h
Lower 95% Upper
ky1 M h
PCy3
PCy3
PCy3
PMet-Bu2
Pt-Bu3
PCp3
0.10
0.11
0.32
0.13
0.13
0.11
0.814
0.873
0.842
0.773
0.000
2.72
0.729
0.802
0.762
0.677
y
2.48
0.899
0.461
0.914
0.531
17.04
y
0.148
0.358
y0.055
12.98
y
0.773
0.945
0.922
0.870
y
2.96
1.469
1.117
21.10
y
0.714
0.113
1.314
PH s24 bar.
2
Ts718C.
tuted product. Scheme 3 illustrates this process
for the 3bqPCy3 system. Both substitution
reactions are reversible, and the product distri-
butions can be established from charging PCy3
to 3b or by adding Pi-Pr3 to a solution of 3a.
Up to 1308C, the rate of exchange between
free Pi-Pr3 and that bound to 3b is insufficient
to support spin-saturation labelling as a kinetic
method. Therefore, a traditional study was un-
dertaken wherein the concentration of the ex-
change products was monitored with time by
31P NMR. All the data acquired in one of the
3b–PCy3 exchange experiments are presented
three PCy3 trials yield a single set of rate con-
Ž .
stants, as would be expected if Eq. 1 was a
suitable model for the process.
From this limited set of data, the exchange
rate appears to depend on the nature and abun-
dance of the entering phosphine, which is evi-
dence of an associative-type mechanism. That
Pt-Bu3 does not participate in such an exchange
process is not surprising, given its extraordinary
cone angle of 1808. Nevertheless, an unequivo-
cal assignment of an exchange mechanism based
on such few data is unwarranted.
w
x
in Fig. 4. Plots of 3b vs. time illustrate the key
Ž
.
result of each experiment Fig. 5 . The repro-
ducibility of the experimental technique is evi-
4. Conclusions
w
x
dent from the two trials employing PCy3 s0.1
When employed as a hydrogenation catalyst
precursor, 2 is activated by O2 displacement by
H2 to form the dihydrogen complex, 3. How-
ever, solutions of 2 will decompose over time
unless dissociated O2 is expelled from solution.
During the hydrogenation of acrylonitrile–
butadiene copolymers, the predominant com-
plexes are the nitrile and H2 adducts, 4 and 3,
respectively. Under typical hydrogenation con-
Ž
.
M Fig. 5 .
To quantify the first substitution reaction in
which 3b is transformed into OsHCl CO -
H2 Pi-Pr3 PR3 , the rate constants k1 and
ky defined by Scheme 3 have been fitted to the
Ž
.
Ž
.Ž
.Ž
.
1
second-order rate expression:
d 3b
y
sk1 3b PR3 yky1 OsHCl CO
Ž .
dt
X
ditions, the coordination of R CN to 1 can ac-
count for a reduction in catalytic activity associ-
ated with increased copolymer concentration.
The phosphine ligands of 1 and 3 are labile,
undergoing exchange with free PR3 at a rate
that is slow relative to the NMR timescale.
Preliminary kinetic data suggests this is an asso-
ciative process.
= H Pi- Pr PR
Pi- Pr3
Ž
2 .Ž
3.Ž
.
3
1
Ž .
An estimate of k1 represents the inherent rate of
the forward substitution reaction at a given
phosphine type and concentration. Regression
values of k1 and ky1 are listed in Table 3. The

(
)
J.S. Parent et al.rJournal of Molecular Catalysis A: Chemical 135 1998 285–293
293
w
w
w
x
11 G.L. Rempel, N.T. McManus, J.S. Parent, United States
Acknowledgements
Patent 5,561,197, 1996.
x
12 J.S. Parent, N.T. McManus, G.L. Rempel, Ind. Eng. Chem.
The authors thank Mrs. J. Venne and Mr. F.
Wassmer for their technical assistance. Support
from the Natural Sciences and Engineering Re-
Ž
.
Res. 1998 submitted.
x
13 N.T. McManus, G.L. Rempel, J. Macromol. Sci., Rev.
Ž
.
Macromol. Chem. Phys. C 35 1995 239.
w
w
x
x
Ž
.
14 F.G. Moers, J. Chem. Soc., Chem. Commun. 1971 79.
15 F.G. Moers, J.P. Langhout, Rec. Trav. Chim. Pays Bas 91
Ž
.
search Council NSERC and funding from
Petro-Canada to JSP is gratefully acknowl-
edged.
Ž
.
1972 591.
w
w
w
w
w
w
x
16 F.G. Moers, R.W.M. ten Hoedt, J.P. Langhout, Inorg. Chem.
Ž
.
12 1973 2196.
x
17 F.G. Moers, R.W.M. ten Hoedt, J.P. Langhout, J. Organomet.
Ž
.
Chem. 65 1974 93.
References
x
Ž
.
18 M.A. Esteruelas, H. Werner, J. Organomet. Chem. 303 1986
221.
w x
1
R.A. Sanchez-Delgado, M. Rosales, M.A. Esteruelas, L. Oro,
J. Mol. Catal. A: Chem. 96 1995 231.
x
19 F.G. Moers, R.W.M. ten Hoedt, J.P. Langhout, J. Inorg.
Ž
.
Ž
.
Nucl. Chem. 36 1974 2279.
w x
2
M.A. Esteruelas, L.A. Oro, C. Valero, Organometallics 10
x
Ž
.
20 F.G. Moers, J.P. Langhout, J. Inorg. Nucl. Chem. 39 1977
Ž
.
1991 462.
591.
w x
3
A. Andriollo, M.A. Esteruelas, U. Meyer, L.A. Oro, R.A.
Sanchez-Delgado, E. Sola, C. Valero, H. Werner, J. Am.
x
21 V.I. Bakhmutov, J. Bertran, M.A. Esteruelas, A. Lledos, F.
Maseras, J. Modrego, L.A. Oro, E. Sola, Chem. Eur. J. 2
Ž
.
Chem. Soc. 111 1989 7431.
Ž
.
1996 815.
w x
4
M.A. Esteruelas, L.A. Oro, C. Valero, Organometallics 11
w
x
Ž
.
22 P.G. Jessop, R.H. Morris, Coord. Chem. Rev. 121 1992
Ž
.
1992 3362.
155, and references therein.
w x
5
R.A. Sanchez-Delgado, A. Andriollo, N. Valencia, J. Chem.
w
w
x
x
Ž
.
23 D.C. Roe, J. Magn. Reson. 63 1985 388.
24 A. Mezzeti, E. Zangrando, A. Del Zotto, P. Rigo, J. Chem.
Ž
.
Soc. Chem Commun. 1983 444.
w x
6
R.A. Sanchez-Delgado, A. Andriollo, E. Gonzalez, N. Valen-
cia, V. Leon, J. Espidel, J. Chem. Soc., Dalton Trans. 1985
Ž
.
Soc., Chem. Commun. 1994 1597.
Ž
.
w
w
w
x
Ž
.
25 N.A. Mohammadi, G.L. Rempel, Macromolecules 20 1987
1859.
2362.
w x
7
M.A. Esteruelas, E. Sola, L.A. Oro, U. Meyer, H. Werner,
Angew. Chem., Int. Ed. Engl. 27 1988 1563.
x
26 P. Martin, N.T. McManus, G.L. Rempel, Stud. Surf. Sci.
Ž
.
Ž
.
Catal. 73 1992 161.
w x
8
M.A. Esteruelas, E. Sola, L.A. Oro, H. Werner, U. Meyer, J.
x
27 E. Farneti, M. Graziani, A. Mezzeti, A. Del Zotto, J. Mol.
Ž
.
Mol. Catal. 53 1989 43.
Ž
.
Catal. 73 1992 147.
w x
9
M.A. Esteruelas, E. Sola, L.A. Oro, H. Werner, U. Meyer, J.
w
w
x
Ž
.
28 J.D. Gilbert, G. Wilkinson, J. Chem. Soc. A 1969 1749.
Ž
.
Mol. Catal. 45 1988 1.
x
29 M. Schlaf, A.J. Lough, P.A. Maltby, R.H. Morris,
w
x
10 M.A. Esteruelas, C. Valero, L.A. Oro, U. Meyer, H. Werner,
Ž
.
Organometallics 15 1996 2270.
Ž
.
Inorg. Chem. 30 1991 1159.
w
x
Ž
.
30 L. Vaska, J. Am. Chem. Soc. 86 1964 1943.