Induced acceleration of phosphine exchange in metal carbonyls by pendant groups of coordinated polyphosphines. Two dangling phosphine arms are much better than one

Article abstract of DOI:10.1021/om980289h

The kinetics and thermodynamics of isomerization of the linkage isomers (OC)₅W[η¹-PPh₂-CH₂CH(PPh ₂)₂] (1) and (OC)₅W[η¹-PPh₂CH(PPh₂)CH ₂PPh₂] (2) in chloroform have been studied by ³¹P{^'H} NMR over the temperature range of 10-55 °C. Conversion of 1 to 2 is exothermic (△H = -12.25 ± 0.1 kJ mol^-1) and is accompanied by a large decrease in entropy (△S = -28.2 ± 0.3 J mol^-1 K^-1). The reaction proceeds much faster than expected [k = (1.18 ± 0.01) × 10^-5 s^-1] at 25 °C, and its small activation enthalpy (△H^? = 92.6 ± 1.9 kJ mol^-1) and large negative activation entropy (△S^? = -28.2 ± 6.2 J mol^-1 K^-1) suggest an associative mechanism. The reaction is 4 orders of magnitude faster than the isomerization of (OC)₅W[η¹-PPh₂CH₂CH ₂P(p-tol)₂] (3) to (OC)₅W[η¹-P(p-tol)₂CH₂CH ₂PPh₂] (4). Exchange of coordinated and dangling phosphines is faster than chelation for each of the four complexes, but chelation of 1 or 2 is much faster than chelation of 3 or 4. It appears that the second dangling phosphine arm present in 1 and 2 accelerates the exchange of all metal-attached ligands. Also observed is long-range phosphorus-carbon coupling, possibly enhanced by a through-space interaction, between the short-armed dangling phosphine and the equatorial carbonyl groups (⁴J_PC = 3.89 Hz).

Full text of DOI:10.1021/om980289h

Organometallics 1998, 17, 4291-4297
4291
In d u ced Acceler a tion of P h osp h in e Exch a n ge in Meta l
Ca r bon yls by P en d a n t Gr ou p s of Coor d in a ted
P olyp h osp h in es. Tw o Da n glin g P h osp h in e Ar m s Ar e
Mu ch Better Th a n On e1
Richard L. Keiter,* J ohn William Benson, Ellen A. Keiter,* Weiying Lin,
Zhongjiang J ia, Donna M. Olson, Douglas E. Brandt, and J eremy L. Wheeler
Department of Chemistry, Eastern Illinois University, Charleston, Illinois 61920
Received April 15, 1998
The kinetics and thermodynamics of isomerization of the linkage isomers (OC)5W[η1-PPh2-
CH2CH(PPh2)2] (1) and (OC)5W[η1-PPh2CH(PPh2)CH2PPh2] (2) in chloroform have been
studied by 31P{1H} NMR over the temperature range of 10-55 °C. Conversion of 1 to 2 is
exothermic (∆H ) -12.25 ( 0.1 kJ mol-1) and is accompanied by a large decrease in entropy
(∆S ) -28.2 ( 0.3 J mol-1 K-1). The reaction proceeds much faster than expected [k )
(1.18 ( 0.01) × 10-5 s-1] at 25 °C, and its small activation enthalpy (∆Hq ) 92.6 ( 1.9 kJ
mol-1) and large negative activation entropy (∆Sq ) -28.2 ( 6.2 J mol-1 K-1) suggest an
associative mechanism. The reaction is 4 orders of magnitude faster than the isomerization
of (OC)5W[η1-PPh2CH2CH2P(p-tol)2] (3) to (OC)5W[η1-P(p-tol)2CH2CH2PPh2] (4). Exchange
of coordinated and dangling phosphines is faster than chelation for each of the four complexes,
but chelation of 1 or 2 is much faster than chelation of 3 or 4. It appears that the second
dangling phosphine arm present in 1 and 2 accelerates the exchange of all metal-attached
ligands. Also observed is long-range phosphorus-carbon coupling, possibly enhanced by a
“through-space” interaction, between the short-armed dangling phosphine and the equatorial
carbonyl groups (4J PC ) 3.89 Hz).
In tr od u ction
Robust complexes such as (OC)5M(η1-dppm) and
(OC)5M(η1-dppe) (M ) Cr, W; dppm ) Ph2PCH2PPh2;
dppe ) Ph2PCH2CH2PPh2) undergo chelation extremely
slowly at room temperature with half-lives of many
months.6b,7 Early 31P{1H} NMR studies further con-
firmed their inert nature by showing that phosphorus
exchange of coordinated and dangling phosphines was
slow relative to the NMR time scale.6b Thus, all
evidence indicated that the uncoordinated phosphine of
these dangling ligand complexes is inactive until sub-
jected to thermal or photolytic stimulation that leads
to CO loss followed by chelation.
Transition metal complexes in which polydentate
phosphorus ligands are incompletely coordinated, some-
times known as “dangling” ligand complexes, have
become important precursors for the synthesis of het-
erobimetallic compounds.2 They were first observed
spectroscopically3 soon after the first chelated complexes
of these ligands were isolated,4 and within a few years
random examples were obtained in pure form.5 Not
surprisingly, synthetic success was possible only when
chelation was slow compared to the time required for
isolation. Subsequently, selective syntheses based on
controlling the number of metal sites available to the
polydentate phosphine were developed for a number of
these complexes.6
Recent qualitative results from our laboratory, how-
ever, demonstrated that for at least one mononuclear
monoligated complex, (OC)5W[η1-PPh2CH2CH(PPh2)2],
exchange between coordinated and uncoordinated phos-
phine occurs much more rapidly than chelation and is
fast enough to measure conveniently even below room
temperature.8
(1) Presented in part in the Symposium on Reactivity at Transition
Metal-Nitrogen, Sulfur, and Phosphorus Bonds at the 211th ACS
National Meeting, New Orleans, 1996.
(2) (a) Orth, S. D.; Terry, M. R.; Abboud, K. A.; Dodson, B.; McElwee-
White. Inorg. Chem. 1996, 35, 916-922. (b) Carr, S. W.; Fontaine, X.
L. R.; Shaw, B. L. J . Chem. Soc., Dalton Trans. 1991, 1025-1027. (c)
Carr, S. W.; Fowles, E. H.; Fontaine, X. L. R.; Shaw, B. L. J . Chem.
Soc., Dalton Trans. 1990, 573-579. (d) Fontaine, X. L. R.; J acobsen,
G. B.; Shaw, B. L.; Thronton-Pett, M. J . Chem. Soc., Dalton Trans.
1988, 741-750. (e) Chaudret, B.; Delavaux, B.; Poilblanc, R. Coord.
Chem. Rev. 1988, 86, 191-243.
(3) Mawby, R. J .; Morris. D.; Thorsteinson, E. M.; Basolo, F. Inorg.
Chem. 1966, 5, 27-33.
(4) Chatt, J .; Hart, F. A. J . Chem. Soc. 1960, 1378.
(5) (a) King, R. B.; Efraty, A. Inorg. Chem. 1969, 8, 2374-2379. (b)
Rigo, P. Longato, B.; Favero, G. Inorg. Chem. 1972, 11, 300-303. (c)
Brown, M. L.; Cramer, J . L.; Ferguson, J . A.; Meyer, T. J .; Winterton,
N. J . Am. Chem. Soc. 1972, 94, 8707-8710.
(6) (a) Keiter, R. L.; Shah, D. P. Inorg. Chem. 1972, 11, 191-193.
(b) Connor, J . A.; Day, J . P.; J ones, E. M.; McEwen, G. K. J . Chem.
Soc., Dalton Trans. 1973, 347-354. (c) Grim, S. O.; Del Gaudio, J .;
Molenda, R. P.; Tolman, C. A.; J esson, J . P. J . Am. Chem. Soc. 1974,
96, 3416-3422. (d) Keiter, R. L.; Sun, Y. Y.; Brodack, J . W.; Cary, L.
W. J . Am. Chem. Soc. 1979, 101, 2638-2641. (e) Keiter, R. L.; Brodack,
J . W.; Borger, R. D, Cary, L. W. Inorg. Chem. 1982, 21, 1256-1258.
(f) Keiter, R. L.; Rheingold, A. L.; Hammerski, J . J .; Castle, C. K.
Organometallic 1983, 2, 1635-1639. (g) Gan, K.-S.; Lee, H. K.; Hor,
T. S. A. J . Organomet. Chem. 1993, 460, 197-201.
(7) Benson, J . W.; Keiter, E. A.; Keiter, R. L. J . Organomet. Chem.
1995, 495, 77-81.
S0276-7333(98)00289-1 CCC: $15.00 © 1998 American Chemical Society
Publication on Web 08/21/1998

4292 Organometallics, Vol. 17, No. 19, 1998
Keiter et al.
techniques. The compounds, (Ph2P)2CdCH2,10 (OC)5WPPh2H,6d
Cl2Pt(PPh2)2CdCH2,11 P(p-tol)2H,12 (Ph2P)2CHCH2PPh2,13 (OC)5-
WNH2Ph,6a and (OC)5WPPh2CHdCH2 were prepared accord-
6d
ing to published procedures. Phosphorus-31 NMR spectra
(referenced to 85% phosphoric acid), carbon-13 spectra, and
infrared spectra were recorded with GE QE-300 NMR and
Nicolet 20 DXB FT-IR spectrometers, respectively. Elemental
analyses were performed at the University of Illinois Mi-
croanalytical Laboratory, Urbana, Illinois.
We now report the quantitative kinetic and thermo-
dynamic details of this reaction as determined by 31P-
{1H} NMR spectroscopy.
For purposes of comparison we have also synthesized
both (OC)5W[η1-PPh2CH2CH2P(p-tol)2] (3) and (OC)5W-
[η1-P(p-tol)2CH2CH2PPh2] (4) in pure form and moni-
tored their rates of isomerization.
Syn th esis of Cl2P t(P P h 2)2CHCH2P P h 2W(CO)5. A THF
solution (25 mL) containing Cl2Pt(PPh2)2CdCH2 (0.700 g, 1.06
mmol), (OC)5WPPh2H (0.539 g, 1.06 mmol), and KOBut (0.0590
g, 0.528 mmol) was refluxed for 2 h. The residue remaining
after the solvent was removed by vacuum was recrystallized
from a 1:1 solution of CH2Cl2/CH3OH to give 0.964 g (77.8%)
of white needlelike crystals (dec 205-210 °C). IR (CHCl3): E
1
a
(OC)5W[η -PPh2CH2CH2P(p-tol)2]
3
(OC)5W[η1-P(p-tol)2CH2CH2PPh2]
(2)
+ A1 , 1940(s) cm-1; A1 , 2073(m) cm-1
(CDCl3): δ 17.7 ppm (t, 1J WP ) 246 Hz, 3J PP ) 5.01 Hz), δ -37.4
.
31P{1H} NMR
1
2
4
ppm (d, J PtP ) 3077 Hz, J PP ) 5.01 Hz14). Anal. Calcd for
43H33Cl2O5P3PtW: C, 44.05; H, 2.84; P, 7.92. Found: C,
1
3
Remarkably, reaction 1 is 4 orders of magnitude faster
than reaction 2. In this paper we propose mechanistic
models to account for the very different exchange rates
and present evidence suggesting that reaction 1 pro-
ceeds primarily by an associative mechanism in which
the transition state requires active participation of all
three phosphino groups, one dangling phosphine arm
interacting with the cis carbonyl groups and a second
dangling phosphine arm displacing the coordinated
phosphine. Reaction 2, on the other hand, lacking an
accelerating short phosphine arm, proceeds at a normal
rate, consistent with a mechanism in which ligand
dissociation is of primary importance.
If interaction of an R group (in this case a dangling
phosphine) of coordinated PR3 with the equatorial
carbonyl ligands of a complex such as (OC)5WPR3 leads
to accelerated phosphine exchange, it would be expected
that the same interaction would lead to accelerated
carbonyl replacement. Thus, if the rate of isomerization
in reaction 1 is accelerated by a phosphine-carbonyl
interaction, the same interaction should lead to an
accelerated rate of chelation. We have evaluated this
hypothesis by measuring rates of chelation for both
reaction 1 and 2 and comparing them to rates of
isomerization.
These studies are of fundamental interest because
they suggest that complex reactivity may be profoundly
influenced by the nature of pendant groups attached to
a coordinated ligand. Although it is widely recognized
that steric repulsion between pendant groups and other
ligands of a complex may significantly influence reaction
rates, the importance of attractive interactions between
these groups and coordinated ligands has received much
less attention.9 It would appear that for organometallic
reactions in general it may be possible to influence
substitution rates by employing ligands with pendant
groups which have an affinity for other coordinated
ligands of the complex.
C
44.14; H, 2.88; P, 7.77.
Syn th esis of (OC)5W[η1-P P h 2CH2CH(P P h 2)2], 1, a n d
(OC)5W[η1-P P h 2CH(P P h 2)CH2P P h 2], 2. Meth od A.
A
mixture of Cl2Pt(PPh2)2CHCH2PPh2W(CO)5 (0.457 g, 0.390
mmol) and KCN (0.102 g, 1.56 mmol) in ethanol (25 mL) was
stirred for 72 h at room temperature. During this time the
color of the suspension changed from colorless to yellow. The
insoluble material was collected by filtration and extracted
with CH2Cl2 (10 mL). The solvent was removed, and the
residue was shown to consist of a 1:5 mixture of isomers 1
1
and 2.13a 31P{1H} NMR (CDCl3) for 1: δ 13.1 ppm (t, J WP
)
242 Hz, J PP ) 7.5 Hz), δ -3.1 ppm (d, J PP ) 7.5 Hz). 13C-
3
3
{1H} NMR: δ(CO)ax ) 199.8 ppm (d, J PC ) 23.1 Hz); δ(CO)eq
2
197.2 ppm (d, 2J PC ) 7.27 Hz). 31P{1H} NMR (CDCl3) for 2: δ
1
2
3
34.1 ppm (dd, J WP ) 252 Hz, J PP ) 207 Hz, J PP ) 22.2 Hz),
2
3
δ -8.3 ppm (d, J PP ) 207 Hz), δ -16.6 ppm (d, J PP ) 22.2
Hz). 13C{1H} NMR for 2: δ(CO)ax 199.6 ppm (d, J PC ) 23.7
2
1
2
Hz, J WC ) 142.1 Hz), δ(CO)eq 197.1 ppm (dd, J PC ) 6.52 Hz,
4J PC ) 3.89 Hz, J WC ) 126.5 Hz). The solvent was removed,
1
and the residue was recrystallized from a mixture of 10 mL of
CH2Cl2 and 15 mL of CH3OH to give white crystals of the less
soluble 2 (0.196 g, 55.4%, mp ) 164.0-164.5 °C). IR (CHCl3):
A1 + E ) 1937(s) cm-1, A1 ) 2070(m) cm-1
. It was not
1
2
possible to obtain the more soluble 1 in pure form by either
recrystallization or column chromatography; however, when
the cyanide reaction was run for 4 h under the above
conditions, the ratio of 1 to 2 was 3.6:1 as shown by 31P{1H}
NMR, indicating that 1 formed initially and then partially
1
2
isomerized to 2. IR (CHCl3): A1 + E ) 1936(s) cm-1, A1
)
2071 cm-1. When pure 2 was placed in CHCl3, isomerization
took place to give a mixture of 1 and 2.
Meth od B. To (OC)5WNH2Ph (1.00 g, 2.33 mmol) and
(Ph2P)2CHCH2PPh2 (1.36 g, 2.33 mmol) was added 20 mL of
toluene. The solution was stirred for 12 h at room tempera-
ture, toluene was removed, and the residue was chromato-
graphed on Al2O3 with CH2Cl2/hexane (1:2). The first fraction
contained isomers 1 and 2 to give a combined yield of 1.1 g
(52%).
F or m a tion of (OC)4W[η2-P P h 2CH(P P h 2)CH2P P h 2], 5,
a n d (OC)4W[η2-P P h 2CH2CH(P P h 2)2], 6. These complexes
slowly formed at 55 °C from mixtures of 1 and 2 in CDCl3
Exp er im en ta l Section
(10) Colquhoun, I. J .; McFarlane, W. J . Chem. Soc., Dalton Trans.
1982, 1915-1921.
Gen er a l Con sid er a tion s. All reactions were carried out
under a dry nitrogen atmosphere using standard Schlenk
(11) Higgins, S. J .; Shaw, B. L. J . Chem. Soc., Dalton Trans. 1989,
1527-1530.
(8) (a) Keiter, R. L.; Keiter, E. A.; Olson, D. M.; Bush, J . R.; Lin,
W.; Benson, J . W. Organometallics 1994, 13, 3752-3754. (b) Evidence
for phosphine exchange in (OC)5MoPPh2CHdC(Me)PPh2 has been
described recently. Maitra, K.; Catalano, V. J .; Nelson, J . H. J .
Organomet. Chem. 1997, 529, 409.
(9) (a) Brown, T. L. Inorg. Chem. 1992, 31, 1286-1294. (b) Woo, T.
K.; Ziegler, T. Inorg. Chem. 1994, 33, 1857-1863.
(12) (a) Grim, S. O.; Yankowsky, A. W. J . Org. Chem. 1977, 42,
1236-1239. (b) Cotton, F. A.; Kitagawa, S. Inorg. Chem. 1987, 26,
3463-3468.
(13) (a) Bookham, J . L.; McFarlane, W., Colquhoun, I. J . J . Chem.
Soc., Dalton Trans. 1988, 503-507. (b) Schmidbaur, H.; Paschalidis,
C.; Reber, G.; Muller, G. Chem. Ber. 1988, 121, 1241-1245.
(14) 3J PP was reported incorrectly in ref 8a.

Phosphine Exchange in Metal Carbonyls
Organometallics, Vol. 17, No. 19, 1998 4293
solution in sealed NMR tubes (see kinetics results). 31P{1H}
1
3
NMR (CDCl3) for 5:13a δ 3.9 ppm (t, J WP ) 206 Hz, J PP ) 8.5
Hz), δ -18.1 ppm (d, J PP ) 8.5 Hz). 31P{1H} NMR (CDCl3)
3
for 6:13a δ 50.6 ppm (dd, 1J WP ) 230.0 Hz, 3J PP ) 1.0 Hz,2J PP
27.1 Hz), δ 33.2 ppm (dd, 1J WP ) 235 Hz, 3J PP ) 1.0 Hz,3J PP
)
)
3
2
2.0 Hz), δ -14.4 ppm (dd, J PP ) 2.0 Hz, J PP ) 27.1 Hz).
Syn th esis of (OC)5W[P (p-tol)2H] a n d cis-W(CO)4[P (p-
tol)2H]2. A solution containing W(CO)6 (1.84 g, 5.23 mmol),
Me3NO‚2H2O (0.581 g, 5.23 mmol), P(p-tol)2H (1.12 g, 5.23
mmol), and CH2Cl2 (50 mL) was stirred at room temperature
for 72 h. The solvent was removed, and 80 mL of CH2Cl2/
CH3OH (1:3) was added to the residue followed by filtration
to remove unreacted W(CO)6. The filtrate was cooled to -5
°C, which gave a precipitate of 1.21 g of gray solid. Separation
of products was achieved by eluting the sample on an alumina
column (Brockman I deactivated with 5% water) with CH2-
Cl2/hexane (1:5). The first band was collected to yield 0.367 g
(13.0%) of (OC)5WPPh2(p-tol)2H (mp ) 89-90 °C). IR (CH2Cl2),
F igu r e 1. 31P{1H} NMR spectrum of a CDCl3 solution of
1 and 2 at 10 °C. Signals A and B belong to isomer 1, while
signals D, C, and E belong to 2.
Ta ble 1. Equ ilibr iu m Con sta n ts (K) for
Isom er iza tion of Lin k a ge Isom er s 1 / 2 a n d 3 / 4
in CDCl3
A1 + E ) 1940(s) cm-1, A1 ) 2074(m) cm-1
.
31P NMR
1
2
1
(CDCl3): δ -15.5 ppm (1J wp ) 228 Hz, J PH ) 332 Hz). Anal.
Calcd for C19H15O5PW: C, 42.41; H, 2.81. Found: C, 42.57;
H, 2.88. From the second fraction cis-W(CO)4[P(p-tol)2H]2 was
obtained (0.048 g, 2.5%; mp ) 138-139 °C). IR (CH2Cl2): A1
) 2021(m) cm-1, A1 ) 1916(sh) cm-1, B1 ) 1903(s) cm-1, B2 )
temp (K)
K
1 / 2
283
298
313
328
6.14 ( 0.04
4.74 ( 0.01
3.76 ( 0.05
3.01 ( 0.02
1886(sh) cm-1
.
31P NMR (CDCl3): δ -4.4 ppm (1J WP ) 224
1
Hz, J PH ) 337 Hz). Anal. Calcd for C32H30O4P2W: C, 53.06;
H, 4.17. Found: C, 52.90; H, 4.17. A trace of a third fraction,
presumably fac-W(CO)3[P(p-tol)2H]3, was detected. 31P{1H}
NMR (CDCl3): δ 4.6 ppm (1J WP ) 216 Hz).
3 / 4
328
1.62 ( 0.08
an integral that included 99% of the intensity. Each spectrum
was integrated four times, and the average value was used in
subsequent calculations. Equilibrium was assumed to have
been reached when the integral ratio of isomers remained
unchanged after three consecutive runs spaced by several days.
Equilibrium and kinetics experiments with 3 and 4 were
carried out as described for 1 and 2, except that pure 3 (20.0
mg, 0.0263 mmol) and pure 4 (20.0 mg, 0.0263 mmol) were
placed in separate NMR tubes; thus, the equilibrium between
3 and 4 was approached from both directions starting with
pure reactants.
Syn th esis of (OC)5W[η1-P P h 2CH2CH2P (p-tol)2], 3.
A
mixture of (OC)5WPPh2CHdCH2 (3.0 g, 5.6 mmol), P(p-tol)2H
(1.1 g, 5.1 mmol), and 2,2′-azobisisobutyronitrile (AIBN) (0.13
g) was heated without solvent at 60° C for 24 h. Recrystalli-
zation of the crude product gave 3 (1.71 g, 44%) (mp ) 112-
o
1
2
115 C). IR (CH2Cl2): A1 + E ) 1937(s) cm-1, A1 ) 2071(m)
cm-1, B2 ) 1981(w) cm-1
.
31P{1H} NMR (CDCl3): δ 13.6 ppm
1
3
3
(d, J WP ) 239.3 Hz, J PP ) 37.6 Hz), δ -13.7 ppm (d, J PP
)
37.6 Hz). 13C{1H} NMR (CDCl3): δ(CO)ax 199.3 ppm (d, J PC
2
1
2
) 21.4 Hz, J WC ) 143.7 Hz), δ(CO)eq 196.9 ppm (d, J PC
)
6.46 Hz, 1J WC ) 125.6 Hz). Anal. Calcd for C33H28O5P2W: C,
52.82; H, 3.76. Found: C, 52.64; H, 3.75.
Syn th esis of (OC)5W[η1-P (p-tol)2CH2CH2P P h 2], 4.
Resu lts a n d Discu ssion
A
mixture of (OC)5W[P(p-tol)2H] (0.31 g, 0.57 mmol), PPh2CHd
CH2 (0.15 mL, 0.66 mmol), and AIBN (0.12 g) without solvent
was heated to 80 °C for 24 h. The crude reaction mixture was
recrystallized from CH2Cl2/CH3OH (3:1) to give 0.33 g of 6,
which was further purified chromatographically (Brockman I
deactivated with 5% water) by eluting with a hexane/CH2Cl2
Th er m od yn a m ics. Equilibrium constants for reac-
tion 1 were obtained from 31P{1H} NMR spectra of
CDCl3 solutions at 10, 25, 40, and 55 °C (Table 1).
Phosphorus signals of the first-order spectra of 1 and 2
are well-separated, as shown in Figure 1, allowing
straightforward integration. From a van’t Hoff plot (R2
) 0.999 86), calculations of ∆H, ∆S, and ∆G (298 K) for
1
solution (1:3) (0.19 g, 44%). IR (CH2Cl2): A1 + E ) 1936(s)
2
cm-1, A1 ) 2071(m) cm-1, B2 ) 1981(w) cm-1
.
31P{1H} NMR
3
1
the isomerization gave values of -12.25 ( 0.1 kJ mol-1
-28.2 ( 0.3 J mol-1 K-1, and -3.86 ( 0.14 kJ mol-1
respectively.
,
,
(CDCl3): δ 11.5 ppm (d, J PP ) 38.0 Hz, J WP ) 238.4 Hz), δ
-12.0 ppm (d, 3J PP ) 38.0 Hz). 13C{1H} NMR (CDCl3): δ(CO)ax
199.5 ppm (d, 2J PC ) 20.9 Hz, 1J WC ) 143.7 Hz); δ(CO)eq 197.2
(d, J PC ) 7.01 Hz, J WC ) 125.2 Hz). Anal. Calcd for
33H28O5P2W: C, 52.82; H, 3.76. Found: C, 52.65; H, 3.64.
2
1
Electronic considerations would lead us to believe that
the basicities of the nonequivalent phosphorus groups
of (PPh2)2CHCH2PPh2 are quite similar, as each phos-
phorus atom has two phenyl substituents and is sepa-
rated from other phosphorus atoms by either one or two
methylene groups. For example, the pKa values of
Ph2PCH2CH2PPh2 and Ph2CH2PPh2 are 3.86 and 3.81,16
respectively, and photoelectron spectroscopy studies
show that Me2PCH2CH2PMe2 and Me2PCH2PMe2 are
electronically nearly identical in LMo(CO)5 complexes.17
Interestingly, the more sterically congested end of the
phosphine ligand is coordinated in the more stable
C
Equ ilibr iu m a n d Kin etics Mea su r em en ts. A mixture
of 1 and 2 (65.0 mg, 0.0717 mmol) and CDCl3 (0.50 mL) was
flame sealed under vacuum at liquid nitrogen temperature in
an NMR tube, thawed, and thermostated in a constant-
temperature bath at the appropriate temperature. The NMR
probe was brought to the same temperature, and the 31P{1H}
NMR spectrum was recorded. Meaningful spectral integra-
tions were obtained by using a delay time of 15 s and a pulse
1
sequence that gave H decoupling but eliminated NOE effects
on 31P signal intensities. Integrations were carried out fol-
lowing standard procedures15 in which the width at half-height
was obtained for each signal and multiplied by (31.8 to give
(16) Sowa, J . R., J r.; Angelici, R. J . Inorg. Chem. 1991, 30, 3534-
(15) Rabenstein, D. L.; Keire, D. A. In Modern NMR Techniques and
Their Application in Chemistry; Popov, A. I., Hellenga, K., Eds.; Marcel
Dekker: New York, 1991.
3537.
(17) Lichtenberger, D. L.; J atcko, M. E. Inorg. Chem. 1992, 31, 451-
455.

4294 Organometallics, Vol. 17, No. 19, 1998
Keiter et al.
complex, 2. Furthermore, the decrease in entropy is far
greater than can be rationalized on the basis of simple
conformational changes. These intriguing thermody-
namic results can be accounted for if one assumes that
the short dangling phosphine which is separated from
the coordinated phosphine by one carbon atom experi-
ences some attractive interaction for the four equatorial
carbonyl groups. Thus 2 may be stabilized by a van der
Waals interaction between the short phosphine arm and
the carbonyl groups. The latter interaction, which
diminishes the mobility of the dangling arm, would also
explain the large entropy decrease for the reaction (eq
1). As the temperature of the reaction is increased,
isomer 1 gains in stability relative to isomer 2 (the
equilibrium constant decreases), consistent with the
breakdown of the weak phosphine-carbonyl interaction
present in isomer 2 (but not in 1) and with the favorable
entropy increase for the reverse reaction. The origin of
the weak interaction that exists between the dangling
phosphine and the carbonyl groups is uncertain; it may
be a forced consequence of the steric demands of a
pentacarbonyltungsten and four phenyl groups in close
proximity.
Although a single crystal of 2 has not been obtained,
the molecular structure18 of (OC)5W[η1-PPh2CH2PPh2],
a complex similar to 2, shows that the unshared pair of
electrons of the dangling phosphine is tilted toward two
of the equatorial carbonyl groups, bisecting the C-W-C
angle and giving a phosphorus-carbon separation (3.55
Å) equal to the approximate sum of the van der Waals
radii for C and P (3.50-3.55 Å).19 Long-range phos-
phorus-carbon coupling is also observed for this mol-
ecule (4J PC ) 3.0 Hz, CDCl3) and for 2, suggesting that
they have similar structures in solution. The 13C{1H}
NMR spectrum of the equatorial carbonyl region of each
consists of a doublet of doublets (4J PC ) 3.89 Hz for 2),
a pattern without precedent in (OC)5W(phosphine)
complexes.20 The possibility of endo and exo isomers
was considered but ruled out because neither the
remainder of the 13C{1H} NMR or the 31P{1H} NMR
spectra provided evidence for their existence. The
possibility of two sets of nonequivalent CO groups
resulting from restricted rotation about the W-P bond
was also considered as an explanation for the four-line
equatorial carbonyl C-13 patterns (two doublets rather
than a doublet of doublets) but dismissed because
13C{31P} established the presence of long-range phos-
phorus-carbon coupling.21 The close approach of the
phosphine to the equatorial carbonyl groups, as shown
by the structure of (OC)5W[η1-PPh2CH2PPh2], would not
allow space for an intervening solvent molecule. Five-
bond phosphorus-carbon coupling involving the longer
F igu r e 2. Plot of ln([1] - [1∞]) vs time for the reaction, 1
a 2, at 10 °C in CDCl3.
phosphine arm present in 1 or 2 is not observed.
Similarly no long-range phosphorus-carbon coupling is
observed in complexes 3 and 4, neither of which contains
a short phosphine arm.
The possibility of “through-space” coupling between
the short dangling phosphine arm in 2 was considered
because it seemed unlikely that coupling would be
observed through four bonds, a 90° angle, and a transi-
tion metal atom.22 Also, long-range phosphorus cou-
pling to the axial carbonyl ligand was not observed even
though it would be expected to be larger than coupling
to equatorial carbonyl groups. The infrared spectrum
of 2 in the carbonyl region, however, is consistent with
C4v symmetry, implying that any interaction that exists
between the dangling phosphine of 2 and its equatorial
carbonyl groups is not localized. Furthermore, the
carbonyl spectra of 1 and 2 are identical within experi-
mental error, suggesting that any interaction that exists
in 2 must be rather weak. It has been noted, however,
that vibrational spectroscopy is not very sensitive to
small electronic differences in LW(CO)5 (L ) PMen-
Ph3-n).20b The observed long-range coupling may arise
from a particularly favorable conformational arrange-
ment of the dppm ligand or from a through-space
interaction or a combination of both.
Linkage isomers 3 and 4 are easily distinguished from
one another with phosphorus-31 NMR spectroscopy, the
former having a coordinated phosphine chemical shift
lying over 2 ppm downfield and a dangling phosphine
chemical shift lying nearly 2 ppm upfield from the
corresponding signals in the latter. Because isomers 3
and 4 move to equilibrium extremely slowly, even at
55 °C, time requirements made it impractical to obtain
K at lower temperatures. Isomer 4 is thermodynami-
cally more stable than 3, in agreement with our expec-
tation that coordination of the more basic ditolylphos-
phino end of Ph2PCH2CH2P(p-tol)2 will lead to a more
stable complex in solution than coordination of the
isosteric diphenylphosphino group.23
(18) Benson, J . W.; Keiter, R. L.; Keiter, E. A.; Rheingold, A. L.;
Yap, G. P. A.; Mainz, V. V. Organometallics, in press.
(19) Huheey, J . E.; Keiter, R. L.; Keiter, E. A. Inorganic Chemistry,
4th ed.; HarperCollins: New York, 1993.
Kin etics. Rate constants and half-lives to equilib-
rium for reactions 1 (10, 25, and 40 °C) and 2 (55 °C)
are shown in Table 2. A plot of ln{[1] - [18]} vs t at 10
(20) (a) Nelson, J . H. Coord. Chem. Rev. 1995, 139, 245-280. (b)
Bancroft, G. M.; Dignard-Bailey, L.; Puddephatt, R. J . Inorg. Chem.
1986, 25, 3675-3680. (c) Buchner, W.; Schenk, W. A. Inorg. Chem.
1984, 23, 132-137. (d) Bodner, G. M.; May, M. P.; McKinney, L. E.
Inorg. Chem. 1980, 19, 1951-1958. (e) Inubushi, Y.; Huy, N. H. T.;
Mathey, F. Chem. Commun. 1996, 1903-1904. (f) Streubel, R.;
Kusenberg, A.; J eske, J .; J ones, P. G. Angew. Chem. 1994, 33, 2427-
2428. (g) Huang, J .-T.; Chand, P.; Fronczek, F. R.; Watkins, S.F.;
Lammertsma, K. Organometallics 1993, 12, 1401-1405.
(22) (a) Holmes, R. R.; Prakasha, T. K.; Pastor, S. D. In Phosphorus-
31 NMR Spectral Properties in Compound Characterization and
Structural Analysis; Quin, L. D., Verkade, J . G., Eds.; VCH: New York,
1994. (b) Mathey, F.; Mercier, F.; Nief, F.; Fischer, J .; Mitschler, A. J .
Am. Chem. Soc. 1982, 104, 2077. (c) Pastor, S. D.; Shum, S. P.; DeBellis,
A. D.; Burke, L. P.; Rodebaugh, R. K.; Clarke, F. H.; Rihs, G. Inorg.
Chem. 1996, 35, 949. (d) Szalontai, G.; Bakos, J .; Toth, I.; Heil, B.
Magn. Reson. Chem. 1987, 25, 761. Gu¨nther, H. NMR Spectroscopy,
2nd ed.; J ohn Wiley: New York, 1995; p 129.
(21) The 13C{31P} NMR spectrum was provided by Vera Mainz of
the Varian Oxford Instrument Center for Excellence in NMR Labora-
tory at the University of Illinois at Urbana.
(23) Li, C.; Nolan, S. P. Organometallics 1995, 14, 1327-1332.

Phosphine Exchange in Metal Carbonyls
Organometallics, Vol. 17, No. 19, 1998 4295
Ta ble 2. Ra te Con sta n ts for Isom er iza tion of
Lin k a ge Isom er s, 1, 2, 3, a n d 4 a n d for th e
Ch ela tion of 1 a n d 2 in (CDCl3)
Sch em e 1
ln 2/
(k1 + k-1)
temp (K)
k1 (s-1
)
k-1 (s-1
)
1 / 2
283
298
313
(1.60 ( 0.04) × 10-6 (2.61 ( 0.06) × 10-7 4.31 days
(1.18 ( 0.01) × 10-5 (2.50 ( 0.01) × 10-6 13.5 h
(7.95 ( 0.3) × 10-5
(2.11 ( 0.1) × 10-5
2.4 h
3 / 4
sociation.29 Notable exceptions include associative mech-
anisms involvingnitrosyl(linear tobent)or cyclopentadienyl
(η5 to η3) complexes in which the ligands change
hapticity during the course of the reaction. As in these
cases, a dissociative (D) mechanism for reaction 1 may
be ruled out for several reasons: (1) the rate of the
reaction is much faster than the rate of phosphine
exchange in similar LM(CO)5 complexes, as indicated
in the previous section,27 and enthalpies of activation
for both forward and reverse reactions are much smaller
than expected for a D mechanism.27a Values for the
328
328
328
(2.03 ( 0.06) × 10-8 (1.25 ( 0.06) × 10-8 245 days
1 f 5 + CO
(1.29 ( 0.03) × 10-6
1 + 2 f 6 + CO
(1.35 ( 0.1) × 10-6
Ta ble 3. Activa tion P a r a m eter s for Isom er iza tion
of Lin k a ge Isom er s 1 / 2
∆Hq(forward) ) 92.6 ( 1.9 kJ mol-1
∆Sq(forward) ) -28.2 ( 6.2 J mol-1 K-1
∆Gq298(forward) ) 101.0 ( 2.6 kJ mol-1
Ea(forward) ) 95.3 ( 1.8 kJ mol-1
∆Hq(reverse) ) 104.5 ( 1.8 kJ mol-1
∆Sq (reverse) ) -1.0 ( 6 J mol-1 K-1
∆Gq298(reverse) ) 104.8 ( 2.5 kJ mol-1
Ea(reverse) ) 107.0 ( 1.6 kJ mol-1
forward and reverse reactions are 93 and 105 kJ mol-1
,
respectively. Hoff and Nolan have shown that for a
series of phosphines in L3M(CO)3 complexes the W-P
bond energy is 1.48 times the Cr-P bond energy.30
Assuming that a similar ratio exists in LM(CO)5 com-
plexes and using the Cr-P bond energy reported for
(OC)5CrPPh3 (151 kJ mol-1),27a a W-P bond energy of
about 223 kJ mol-1 is predicted, much larger than our
observed activation enthalpies. (2) When the reaction
is carried out in the presence of a 5-fold excess of PPh3,
no (OC)5WPPh3 is formed, indicating that a vacant
coordination site does not become available during
phosphine exchange. (3) The isomerization takes place
without the formation of phosphine-bridged complexes
such as (OC)5W[µ-PPh2CH2CH(PPh2)PPh2]W(CO)5, also
arguing against the availability of a coordination site
to an external ligand.
Abel and co-workers have examined a number of
fluxional dithioether complexes, (OC)5W(η1-RSCH2SR)
(R ) alkyl), with variable-temperature proton NMR
spectroscopy and have proposed a pseudo-seven-coor-
dinate transition state for the observed 1,3-metal sulfur
shift.31 It would appear that to a first approximation
reaction 1 requires a similar transition state, one
involving a 1,4-metal phosphorus shift (Scheme 1). In
our view, however, a simple associative (A) mechanism
does not satisfactorily explain why reaction 1 is 4 orders
of magnitude faster than reaction 2. In both reactions
the exchange that takes place involves diphenylphos-
phino groups separated by two carbon atoms and the
exchanging phosphines would be expected to be elec-
tronically very similar. The large difference in rates of
isomerization for reactions 1 and 2 must in some
manner depend on the presence of the short CH2PPh2
arm in reaction 1. A modification of the transition state
in Scheme 1, as shown in Scheme 2, nicely accounts for
the observed results. The weakening of the existing
W-P bond and the formation of the new W-P bond are
accompanied by the onset of an interaction between the
°C is shown in Figure 2. Activation parameters, ∆Hq,
∆Sq, and ∆Gq298, for both forward and reverse directions
of eq 1 as obtained from Eyring plots, are given in Table
3. It was not possible to determine k for reaction 1 at
55 °C with phosphorus-31 NMR because the reaction
proceeds too quickly, relative to the time required for
signal accumulation, to allow direct determination, but
not fast enough to be determined by 31P 2D-EXSY
experiments.24 However, the calculated kfor (3.7 × 10-4
s-1) at that temperature is 4 orders of magnitude faster
than kfor (3.3 × 10-8 s-1) for reaction 2.
Many studies have confirmed the expectation that d6
low-spin complexes will be substitutionally inert.25 For
example, the dissociative rate constant (k1) for CO
substitution by phosphines in W(CO)6 at 30 °C is 1 ×
10-14 s-1 26
The reverse reactions in which phosphines
.
are replaced by CO have received much less attention.27
The rate constant for PPh3 dissociation from (OC)5-
CrPPh is 3 × 10-11 s-1 at 30 °C,27a about 6 orders of
3
magnitude slower than the rate of isomerization in
reaction 1. Phosphine dissociation for tungsten carbonyl
complexes would be expected to be even slower than for
chromium complexes if generally observed trends are
followed.28
Rea ction Mech a n ism s. Most transition metal car-
bonyl complexes undergo substitution by ligand dis-
(24) Orrell, K. G.; Sik, V.; Stephenson, D. Prog. Nucl. Magn. Reson.
Spectrosc. 1990, 22, 141.
(25) (a) Wilkins, R. G. Kinetics and Mechanism of Reactions of
Transition Metal Complexes, 2nd ed.; VCH: New York, 1991. (b)
Atwood, J . D. Inorganic and Organometallic Reaction Mechanisms;
Brooks/Cole: Monterey, 1985. (c) Basolo, F.; Pearson, R. G. Mecha-
nisms of Inorganic Reactions, 2nd ed.; J ohn Wiley: New York, 1967.
(26) (a) Atwood, J . D.; Brown, T. L. J . Am. Chem. Soc. 1976, 98,
3160-3166. (b) Lichtenberger, D. L.; Brown, T. L. J . Am. Chem. Soc.
1978, 100, 366-373.
(27) Darensbourg, D. J . Graves, A. H. Inorg. Chem. 1979, 18, 1257-
1261. Yang, G. K.; Peters, K. S.; Vaida, V. Chem. Phys. Lett. 1986,
125, 566.
(28) Kirtley, S. W. In Comprehensive Organometallic Chemistry;
Wilkinson, G., Stone, F. G. A., Abel, E. W., Eds.; Pergamon: Oxford,
U.K., 1982.
(29) (a) Basolo, F. Polyhedron 1990, 13, 1503-1535. (b) Darens-
bourg, D. J . Adv. Organomet. Chem. 1982, 21, 113-150.
(30) Mukerjee, S. L.; Lang, R. F.; J u, T.; Kiss, G.; Hoff, C. D.; Nolan,
S. P. Inorg. Chem. 1992, 31, 4885-4889.
(31) Abel, E. W.; Blackwall, E. S.; Orrell, K. G.; Sik, V. J . Organomet.
Chem. 1994, 464, 163-170.

4296 Organometallics, Vol. 17, No. 19, 1998
Keiter et al.
Sch em e 2
Sch em e 3
expected that rate acceleration would also be observed
for CO substitution. Loss of CO by either isomer would
result in the formation of a chelated complex. Only one
product, 5, is possible when isomer 1 undergoes chela-
tion, but two products, 5 and 6, result when 2 loses CO
(Scheme 3). Within the temperature range used for the
phosphine exchange experiments (10-40 °C), chelation
was too slow to monitor in practical time periods. At
55 °C, however, it was possible to obtain kobs for the
irreversible formation of both 5 and 6 (Table 2). Com-
parison with chelation rates from previous studies
substantiates that 1 and 2 undergo chelation much
faster than expected. Kinetic studies of chelation of
(OC)5W[η1-PPh2CH2CH2PPh2] at 123, 134, and 150
°C6b,33 allow calculation of k at 55 °C to give a direct
comparison with chelation rates for 1 and 2. A value
of 3 × 10-10 s-1 is obtained, which is about 6 orders of
magnitude slower than chelation of 1 and 2. Isomers 3
and 4 undergo chelation so slowly that chelated products
were seen only in trace concentrations even after several
hundred days. It was not possible to extract meaningful
rate constants from these baseline concentrations.
However, as we were able to measure phosphine ex-
change rates of 3 and 4 as described earlier, we can say
with certainty that chelation is much slower than
exchange and that kobs for chelation of 3 and 4 is at least
an order of magnitude smaller than 10-8 s-1. Thus,
chelation rates as well as phosphine exchange rates are
much faster in 1 and 2 than in 3 and 4, leading us to
conclude that the presence of the short phosphine arm
in reaction 1 accelerates both phosphine and carbonyl
replacement.
equatorial carbonyl groups and the dangling CH2PPh2.
The large decrease in entropy for the forward activation
step (∆Sq ) -28.2 J mol-1 K-1) is consistent with an
for
interaction involving all three associated arms in the
transition state. When the reverse reaction takes place,
however, as the new W-P bond forms, the weak
dangling phosphine-carbonyl interaction diminishes in
strength, as does the existing W-P bond. The combina-
tion of two interactions decreasing while a third is
increasing leads to opposing entropy changes which
nearly nullify each other (∆Sq ) -1.0 J mol-1 K-1).
rev
There are a number of examples in the literature in
which the interaction of an approaching nucleophile
with coordinated CO accelerates the rate of ligand
substitution. For example, Basolo has shown that CO
substitution by AsPh3 in Fe(CO)2(NO)2 is catalyzed by
a variety of bases [R4NX (X ) Cl, Br, I, N3), NaOMe,
pyrrolidine].29a Similarly, Darensbourg’s studies have
shown that ligand substitution in M(CO)6 (M ) Cr, Mo,
W) is catalyzed by OH-.29b In each example acceleration
is believed to occur because of a transient interaction
between the base and coordinated CO, which in turn
leads to labilization of carbonyl groups. Furthermore,
it is well-known that substitution rates in metal car-
bonyls are greatly influenced by the nature of ligands
cis to the ligand that is replaced.26 The transition state
in these substitution reactions is stabilized by the
presence of cis ligands that are π donors.32 We are
suggesting a neighboring group cis labilization effect in
which the transition state is stabilized by an interaction
between the short dangling phosphine and the cis
carbonyl groups.
Con clu sion s
In this work we have found that the rate of exchange
of coordinated and terminal phosphines in (OC)5W[η1-
PPh2CH2CH(PPh2)2] is 4 orders of magnitude faster
than the rate of exchange in (OC)5W[η1-PPh2CH2CH2P-
(p-tol)2], a striking result when one considers the
electronic similarities of the two phosphorus ligands.
Chelation rates for the two complexes also differ by at
least 4 orders of magnitude but are significantly slower
in each than the rate of phosphine exchange. We
believe that phosphine exchange in (OC)5W[η1-PPh2CH2-
CH(PPh2)2] proceeds primarily by an associative mech-
anism in which the transition state exhibits three
phosphine interactions, one in which the reactant W-P
bond is weakening, one in which the incoming dangling
Rate constants for reaction 2 at 55 °C in CDCl3 are
2.03 × 10-8 and 1.25 × 10-8 s-1 for the forward and
reverse reactions, respectively, about 4 orders of mag-
nitude slower than for reaction 1. The mechanistic
model presented in Scheme 2 requires the existence of
an exchange-accelerating short phosphine arm not
found in isomers 3 and 4. Isomerization rates for
reaction 2, although an order of magnitude faster than
PPh3 dissociation from (OC)5CrPPh3 (kcalc ) 2.9 × 10-9
s-1), do not clearly distinguish between an A, IA, or ID
mechanism, although the slowness of the reaction
suggests a significant dissociative component.
If phosphine exchange in 1 and 2 is accelerated by
the presence of a dangling phosphine, it would be
(32) Darensbourg, D. J .; J oyce, J . A.; Bischoff, C. J .; Reibenspies, J .
H. Inorg. Chem. 1991, 30, 1137.
(33) Connor, J . A.; Riley, P. I. J . Organomet. Chem. 1975, 94, 55-
60.

Phosphine Exchange in Metal Carbonyls
Organometallics, Vol. 17, No. 19, 1998 4297
phosphine is interacting with tungsten to form a new
bond, and one in which the short arm experiences an
interaction with the equatorial carbonyl groups. The
latter interaction may help to stabilize the product
isomer, (OC)5W[η1-PPh2CH(PPh2)PPh2], and help ac-
count for the large decrease in entropy associated with
the isomerization, as well as the long-range coupling
between the short arm and the equatorial carbonyl
groups. We believe that the acceleration of the reaction
occurs because the interaction of the dangling phosphine
arm with a carbonyl ligand in the transition state leads
to a weakening (cis labilization) of the metal-phospho-
rus bond, allowing the other dangling phosphine arm
to displace the coordinated phosphine. The isomeriza-
tion of (OC)5W[η1-PPh2CH2CH2P(p-tol)2], lacking an
accelerating arm, proceeds at the expected slower pace.
These results suggest that complex reactivity may be
greatly influenced by the extent to which R groups of a
bound ligand can interact with other ligands of the
complex. The principles involved should be applicable
to a wide range of organometallic reactions and have
important catalytic implications.
Ackn owledgm en t. We thank the Camille and Henry
Dreyfus Foundation for a Scholar/Fellow grant, and the
National Science Foundation (CHE-9708342 and CHE
8851619) for support of this work and for the purchase
of a General Electric QE300 NMR spectrometer. D.M.O.
is the recipient of a Council on Undergraduate Research
AIURP Fellowship sponsored by Eli Lilly and Co.
OM980289H