Chelate bite effects for [Pd(triphosphine)(solvent)] (BF4)2 complexes in electrochemical CO2 reduction and the heterolytic cleavage of molecular hydrogen

Article abstract of DOI:10.1021/om960072s

A series of [Pd(triphosphine)(CH₃CN)](BF₄)₂ complexes has been prepared with different chelate bites. Stoichiometric reactions of these complexes with triethylphosphine, NaBH₄, and H₂ have been studied, as well as the catalytic electrochemical reduction of CO_2. All of these reactions show significant chelate effects. [Pd(ttpE)(CH₃CN)](BF₄)₂ (where ttpE is bis(3-(diethylphosphino)propyl)phenylphosphine) catalyzes the electrochemical reduction of CO₂ to CO in acidic dimethylformamide solutions and reacts with NaBH₄ or H₂ to form [Pd(ttpE)(H)](BF₄). The latter complex is the decomposition product formed under catalytic conditions. X-ray diffraction studies of [Pd(ttpE)(CH₃CN)](BF₄)₂ and [Pd(ttpE)(H)](BF₄) provide insight into possible steric origins of reactivity differences between the last two complexes and analogous complexes with smaller chelate bites. [Pd(ttpE)(CH₃CN)](BF₄)₂ has a square-planar structure with one methylene group of the ethyl substituents making a close contact with the nitrogen atom of acetonitrile. This steric interaction likely contributes to some of the reactivity differences observed. [Pd(ttpE)(H)](BF₄) also has a square-planar structure with the two terminal phosphorus atoms of the triphosphine ligand distorted slightly toward the hydride ligand. Extended Hückel molecular orbital calculations suggest that small chelate bites shift electron density onto the hydride ligand, making it more hydridic, while larger chelate bites shift electron density away from the hydride ligand, making it more acidic.

Full text of DOI:10.1021/om960072s

3360
Organometallics 1996, 15, 3360-3373
Ch ela te Bite Effects for [P d (tr ip h osp h in e)(solven t)](BF 4)2
Com p lexes in Electr och em ica l CO2 Red u ction a n d th e
Heter olytic Clea va ge of Molecu la r Hyd r ogen
Sheryl A. Wander,† Alex Miedaner,† Bruce C. Noll,‡ Robert M. Barkley,‡ and
Daniel L. DuBois*,†
National Renewable Energy Laboratory, Golden, Colorado 80401, and Department of
Chemistry and Biochemistry, University of Colorado, Boulder, Colorado 80309
Received February 5, 1996X
A series of [Pd(triphosphine)(CH3CN)](BF4)2 complexes has been prepared with different
chelate bites. Stoichiometric reactions of these complexes with triethylphosphine, NaBH4,
and H2 have been studied, as well as the catalytic electrochemical reduction of CO2. All of
these reactions show significant chelate effects. [Pd(ttpE)(CH3CN)](BF4)2 (where ttpE is
bis(3-(diethylphosphino)propyl)phenylphosphine) catalyzes the electrochemical reduction of
CO2 to CO in acidic dimethylformamide solutions and reacts with NaBH4 or H2 to form
[Pd(ttpE)(H)](BF4). The latter complex is the decomposition product formed under catalytic
conditions. X-ray diffraction studies of [Pd(ttpE)(CH3CN)](BF4)2 and [Pd(ttpE)(H)](BF4)
provide insight into possible steric origins of reactivity differences between the last two
complexes and analogous complexes with smaller chelate bites. [Pd(ttpE)(CH3CN)](BF4)2
has a square-planar structure with one methylene group of the ethyl substituents making
a close contact with the nitrogen atom of acetonitrile. This steric interaction likely contributes
to some of the reactivity differences observed. [Pd(ttpE)(H)](BF4) also has a square-planar
structure with the two terminal phosphorus atoms of the triphosphine ligand distorted
slightly toward the hydride ligand. Extended Hu¨ckel molecular orbital calculations suggest
that small chelate bites shift electron density onto the hydride ligand, making it more
hydridic, while larger chelate bites shift electron density away from the hydride ligand,
making it more acidic.
In tr od u ction
But P(CH2)nPBut , n ) 2, 3) has been shown to favor
2
2
an agostic ethyl complex for n ) 3 and the hydrido
ethylene complex for n ) 2.4 This result is consistent
with molecular orbital calculations by Thorn and Hoff-
mann, suggesting that alkyl formation would be favored
by P-Pt-P angles larger than 90°.5 Similar conclusions
have been reached for the insertion of carbon monoxide
into M-C bonds of [M(diphosphine)(CH3)(CO)]+ inter-
mediates (where M ) Pt, Pd) based on both experimen-
tal data and theoretical analyses.6,7 The insertion of
olefins into Pd-aryl bonds is another reaction that
exhibits significant chelate effects.8 A final and poten-
tially practical example is the use of bidentate ligands
with large bite sizes to control the regioselectivity of
hydroformylation reactions.9
For bidentate phosphine ligands there are numerous
examples in which the chelate bite has a major effect
on the structure and reactivity of transition-metal
complexes. For example, many square-planar d8
[M(diphosphine)2]2+ and [MCl2(diphosphine)] complexes
exhibit tetrahedral distortions.1,2 This distortion is
larger for complexes with large chelate bites and has
been attributed to both steric and electronic effects.
Regardless of its origin, this distortion results in a lower
energy for the LUMO, which is reflected in more positive
reduction potentials for complexes with large chelate
bites.3 Differences in reactivity as a function of the
chelate bite size have been observed for a number of
stoichiometric and catalytic reactions. The insertion of
ethylene into the Pt-H bond in [Pt(diphosphine)-
For metal complexes with tridentate and tetradentate
phosphine ligands, much less is known about the
relationship between the chelate bite size of the ligand
and the structure and reactivity of the metal complexes.
(ethylene)(H)]+ complexes (where diphosphine
)
† National Renewable Energy Laboratory.
‡ University of Colorado.
X Abstract published in Advance ACS Abstracts, J uly 1, 1996.
(1) (a) Hall, M. C.; Kilbourn, B. T.; Taylor, K. A. J . Chem. Soc. A
1970, 2539. (b) Anderson, M. P.; Pignolet, L. H. Inorg. Chem. 1981,
20, 4101. (c) J ones, R. A.; Real, F. M.; Wilkinson, G.; Galas, A. M. R.;
Hursthouse, M. B.; Malik, K. M. A. J . Chem. Soc., Dalton Trans. 1980,
511. (d) Marder, T. B.; Williams, I. D. J . Chem. Soc., Chem. Commun.
1987, 1478. (e) Clark, G. R.; Skelton, B. W.; Waters, T. N. Acta
Crystallogr., Sect. C 1987, C43, 1708. (f) Mercier, F.; Mathey, F.;
Fischer, J .; Nelson, J . H. J . Am. Chem. Soc. 1984, 106, 425.
(2) (a) van Hecke, G. R.; Horrocks, W. D., J r. Inorg. Chem. 1966, 5,
1968. (b) McPhail, A. T.; Komson, R. C.; Engel, J . F.; Quin, L. D. J .
Chem. Soc., Dalton Trans. 1972, 874. (c) Palenik, G. J .; Mathew, M.;
Steffen, W. L.; Bevan, G. J . Am. Chem. Soc. 1975, 97, 1059. (d) Steffen,
W. L.; Palenik, G. J . Inorg. Chem. 1976, 15, 2432.
(4) Mole, L.; Spencer, J . L.; Carr, N.; Orpen, A. G. Organometallics
1991, 10, 49.
(5) Thorn, D. L.; Hoffmann, R. J . Am. Chem. Soc. 1978, 100, 2079.
(6) (a) Dekker, G. P. C. M.; Elsevier, C. J .; Vrieze, K.; van Leeuwen,
P. W. N. M. Organometallics 1992, 11, 1598. (b) To´th, I.; Elsevier, C.
J . J . Am. Chem. Soc. 1993, 115, 10388. (c) To´th, I.; Ke´gl, T.; Elsevier,
C. J .; Kolla´r, L. Inorg. Chem. 1994, 33, 5708. (d) Scrivanti, A.; Botteghi,
C.; Toniolo, L.; Berton, A. J . Organomet. Chem. 1988, 344, 261.
(7) Sakaki, S.; Kitaura, K.; Morokuma, K.; Ohkubo, K. J . Am. Chem.
Soc. 1983, 105, 2280.
(8) (a) Portnoy, M.; Ben-David, Y.; Rousso, I.; Milstein, D. Organo-
metallics 1994, 13, 3465. (b) Portnoy, M.; Ben-David, Y.; Milstein, D.
Organometallics 1993, 12, 4734. (c) Portnoy, M.; Milstein, D. Orga-
nometallics 1993, 12, 1655.
(3) Miedaner, A.; Haltiwanger, R. C.; DuBois, D. L. Inorg. Chem.
1991, 30, 417.
S0276-7333(96)00072-6 CCC: $12.00 © 1996 American Chemical Society
+
+

Bite Effects for [Pd(triphosphine)(solvent)](BF4)2
Organometallics, Vol. 15, No. 15, 1996 3361
In this paper we describe some observations for pal-
ladium triphosphine complexes that have arisen during
our studies of electrochemical CO2 reduction. It has
been shown that [Pd(triphosphine)(CH3CN)](BF4)2 com-
plexes will catalyze the electrochemical reduction of CO2
to CO in acidic acetonitrile or dimethylformamide
solutions.10,11 Two decomposition products were char-
acterized by spectroscopic and X-ray diffraction studies
and shown to be Pd(I) dimers with bridging triphosphine
ligands, i.e., structures 1a and 1b. Presumably, part
on the chelate bite size. In addition, the ligand occupy-
ing the fourth coordination site of [Pd(triphosphine)L]-
(BF4)2 complexes is more labile for complexes containing
three bridging methylene groups compared to those
containing two methylene groups. Possible reasons for
some of these chelate effects are explored, based on
structural data and qualitative molecular orbital cal-
culations.
Resu lts
Syn th esis a n d Ch a r a cter iza tion of Liga n d s. To
compare the effect of chelate bite on the chemistry and
structure of palladium triphosphine complexes, ligands
were needed with different numbers of methylene
groups connecting the three phosphorus atoms, i.e.,
R2P(CH2)nP(R′)(CH2)mPR′′2 (where R, R′, and R′′ are
alkyl or aryl groups and n and m are 1-3). The free-
radical synthesis of etpE (where R ) R′′ ) Et, R′ ) Ph,
and n ) m ) 2) from divinylphenylphosphine and
diethylphosphine has been described elsewhere.10a This
synthetic method provided a convenient route to a
variety of symmetric triphosphine ligands containing
two-carbon backbones. Efforts to extend this approach
to the synthesis of trimethylene linkages were unsuc-
cessful, as a similar reaction between diallylphenylphos-
phine and diethylphosphine resulted in the formation
of less than 5% of the desired ligand ttpE (2), on the
basis of 31P NMR spectra of the crude product. Appar-
ently, the presence of two allyl groups leads to undesir-
able side reactions. However, 2 is readily prepared in
moderate yield from phenylphosphine and allyldieth-
ylphosphine using the free-radical-catalyzed route shown
in eq 1.12 Purification of 2 can be accomplished by
of the driving force for formation of these dimers is the
release of ring strain in reduced catalytic intermediates.
On the basis of mechanistic studies, the rate-determin-
ing step under normal catalytic conditions is the reac-
tion of a Pd(I) monomeric species with CO2. Because
monomeric Pd(I) complexes would be expected to un-
dergo tetrahedral distortions, a significant amount of
strain would result in these complexes, which contain
two fused five-membered rings. This strain could be
relieved by forming dimers such as 1. This suggested
that dimer formation might be retarded by using
triphosphine ligands with a trimethylene linkage be-
tween the central and terminal phosphorus atoms. The
Pd(I) intermediates formed should be more stable
because the six-membered rings could more readily
accommodate tetrahedral distortions. If this line of
reasoning is correct, then using triphosphine ligands
with trimethylene bridges could prevent or retard dimer
formation and enhance catalyst lifetime.
In this paper, the synthesis and characterization of
[Pd(triphosphine)(CH3CN)](BF4)2 complexes are de-
scribed in which the number of methylene groups
bridging the phosphorus atoms of the triphosphine
ligand is varied in a systematic fashion. This has led
to the identification of a new degradation pathway for
some of these complexes in which stable palladium
hydrides are formed and the discovery of a subtle change
in the reaction mechanism during catalytic electro-
chemical reduction of CO2. Interesting new reactions
in which hydrogen is activated by heterolytic cleavage
have also been observed and found to depend markedly
vacuum distillation. The product is a viscous, yellow,
air-sensitive oil. The ligand ttpC (3), in which the
terminal ethyl groups of 2 are replaced with cyclohexyl
substituents, was prepared in a strictly analogous
fashion. These reactions provide access to ligands with
three-carbon backbones.
Reaction of a 4-fold molar excess of PhPH2 with
vinyldiethylphosphine forms Et2PCH2CH2P(H)Ph (4),
which can be converted to the asymmetric triphosphine
eptpE (5) by a second free-radical-catalyzed reaction
with allyldiethylphosphine (eq 2). The synthesis of 5
was designed in this order so that 4 could also be used
as an intermediate in the preparation of metpE (6);
however, the order of P-H addition may be reversed to
go via the propylene intermediate Et2PCH2CH2CH2P-
(H)Ph. The ligand metpE (6) was synthesized by
reacting the lithium salt of 4, Et2PCH2CH2PPh-Li+,
with (chloromethyl)diphenylphosphine (eq 3). Ph2PCH2-
Cl was prepared by a method described by Stelzer and
(9) (a) Billig, E.; Abatjoglou, A. G.; Bryant, D. R. (Union Carbide)
U.S. Patent 4,769,498, 1988. (b) Hughes, O. R.; Unruh, J . D. J . Mol.
Catal. 1981, 12, 71. (c) Casey, C. P.; Petrovich, L. M. J . Am. Chem.
Soc. 1995, 117, 6007. (d) Casey, C. P.; Whiteker, G. T.; Melville, M.
G.; Petrovich, L. M.; Gavney, J . A., J r.; Powell, D. R. J . Am. Chem.
Soc. 1992, 114, 5535. (e) Moasser, B.; Gladfelter, W. L.; Roe, D. C.
Organometallics 1995, 14, 3832. (f) Kranenburg, M.; van der Burgt,
Y. E. M.; Kamer, P. C. J .; van Leeuwen, P. W. N. M.; Goubitz, K.;
Fraanje, J . Organometallics 1995, 14, 3081.
(10) (a) DuBois, D. L.; Miedaner, A.; Haltiwanger, R. C. J . Am.
Chem. Soc. 1991, 113, 8753. (b) DuBois, D. L.; Miedaner, A. J . Am.
Chem. Soc. 1987, 109, 113.
(11) (a) Bernatis, P. R.; Miedaner, A.; Haltiwanger, R. C.; DuBois,
D. L. Organometallics 1994, 13, 4835. (b) Herring, A. M.; Steffey, B.
D.; Miedaner, A.; Wander, S. A.; DuBois, D. L. Inorg. Chem. 1995, 34,
1100. (c) Steffey, B. D.; Miedaner, A.; Maciejewski-Farmer, M.;
Bernatis, P. R.; Herring, A. M.; Allured, V.; Caperos, V.; DuBois, D. L.
Organometallics 1994, 13, 4844.
(12) The closely related ligand PhP(CH2CH2CH2PMe2)2 has been
reported: Arpac, E.; Dahlenburg, L. Angew. Chem., Int. Ed. Engl. 1982,
21, 931.
+
+

3362 Organometallics, Vol. 15, No. 15, 1996
Ta ble 1. 31P NMR Sp ectr a l Da ta for Liga n d s a n d Com p lexesa
δ (ppm)
Wander et al.
J (Hz)
compd
PA
PB
PX
PY or H
J AB J AX J AY J BX J BY J XY
ttpE (2)
ttpC (3)
eptpE (5)
metpE (6)
-24.4 (s) [-26]
-6.6 (s) [-10]
-26.4 (s) [-28]
-26.8 (s) [-28]
0
0
-21.5 (d) [-20] -24.4 (s) [-26] -18.8 (d) [-22]
-18.6 (d) [-20] -21.4 (d) [-24] -25.5 (dd) [-28]
23
20
25
32
43
22
40
0
0
21
7
12
4
0
118
Pd(ttpE)(NCCH3)2+ (7a )
Pd(ttpE)(PEt3)2+ (7b)
Pd(ttpE)(H)+ (7c)
-0.6 (d)
-5.8 (dd)
11.4 (d)
10.4 (d)
27.4 (d)
76.0 (d)
68 (dd)
65.3 (dd)
61.9 (d)
55.9 (dd)
74.5 (d)
77.1 (m)
71.2 (d)
87 (d)
8.7 (t)
-13.3 (dt)
-11.2 (t)
12.6 (t)
-8.5 (t)
61.8 (d)
8.2 (dt)
36
29
315
-5.0 (d)b
200c
Pd(ttpC)(CH3CN)2+ (8a )
Pd(ttpC)(H)+ (8c)
-5.1 (d)b
200c
Pd(eptpE)(NCCH3)2+ (9a )
Pd(eptpE)(PEt3)2+ (9b)
Pd(eptpE)(H)+ (9c)
3.6 (dd)
-5 (ddd)
17.3 (dd)
321
307
327
35
47
43
50 (dd)
12 (ddd)
25
17
307
40.3 (dd)
114.1 (t)
115.5 (dt)
115.6 (t)
120.4 (m)
93.1 (t)
-3.7 (d)b
206c
Pd(etpE)(NCCH3)2+ (10a )d
Pd(etpE)(PEt3)2+ (10b)d
Pd(etpC)(NCCH3)2+ (11a )d
Pd(etpC)(PEt3)2+ (11b)
Pd(etpC)(H)+ (11c)
10.0 (dt)
30
30
303
7.7 (dt)
<10
16
300
-2.9 (d)b
213c
Pd(metpE)(NCCH3)2+ (12a )e
Pd(metpE)(PEt3)2+ (12b)e
-44.3 (d)
-49.4 (ddd)
49.7 (m)
29.5 (dd)
230
73 (dd)
18.4 (ddd) 281
0
26
47
320
a
All ligand spectra were recorded in toluene-d8. All spectra for the metal complexes were recorded in CD3CN, with the exception of
the triethylphosphine complexes, which were recorded in CD2Cl2 or CD3NO2. Hydride resonance observed in 1H NMR. c P-H coupling
b
constant between the hydride and the central phosphorus atom of the triphosphine ligand. Data taken from ref 10a. e A dimeric species
d
is also present in solution. See text for discussion.
phorus atoms. For eptpE, no coupling is observed
through the trimethylene bridge, consistent with the
results for ttpE and ttpC. A 23 Hz coupling is observed
between nuclei A and X of eptpE, which is consistent
with the three-bond couplings observed for a variety of
other triphosphine ligands with ethylene linkages.10,11
A similar coupling is observed between nuclei A and X
for metpE, but the two-bond coupling between B and X
is much larger (118 Hz), as expected on the basis of
related compounds.13b Compound 4 exhibits a doublet
in the proton-coupled 31P NMR spectrum due to P-H
coupling (1J PH ) 204 Hz).
Syn th esis a n d Ch a r a cter iza tion of Meta l Com -
p lexes. Reaction of the tridentate ligands ttpE, ttpC,
and eptpE with [Pd(NCCH3)4](BF4)2 provides a conve-
nient route to complexes of the type [Pd(triphosphine)-
(CH3CN)](BF4)2, as shown in eq 4. The same method
co-workers that involves the phase-transfer-catalyzed
alkylation of Ph2PH with CH2Cl2 in dichloromethane/
toluene/water.13 Attempts to prepare Et2PCH2Cl in an
analogous manner were unsuccessful because Et2PH is
unreactive under these conditions.
31P NMR data listed in Table 1 support the structures
assigned to the ligands 2-6. The calculated values for
the chemical shifts (shown in brackets in Table 1) are
in reasonable agreement with the experimentally ob-
served values.14 For ttpE (2) and ttpC (3), two singlets
are observed with integration ratios of 2:1 for the
resonances of the terminal phosphorus atoms compared
to the resonances of the central phosphorus atoms. No
P-P coupling is observed for these two ligands. Other
triphosphine ligands containing trimethylene linkages
exhibit either small or no P-P coupling.12,15 For the
has been used previously to prepare [Pd(etpE)(CH3CN)]-
unsymmetrical ligands eptpE (5) and metpE (6), three
(BF4)2 and [Pd(etpC)(CH3CN)](BF4)2.10 The new com-
resonances are observed for the three different phos-
plexes are pale yellow solids, stable to oxygen, and are
somewhat hygroscopic. They have been characterized
(13) (a) Langhans, K.-P.; Stelzer, O.; Weferling, N. Chem. Ber. 1990,
123, 995. (b) Langhans, K.-P.; Stelzer, O. Chem. Ber. 1987, 120, 1707.
(14) Fluck, E.; Heckmann, G. In Phosphorus-31 NMR Spectroscopy
in Stereochemical Analysis; Verkade, J . G., Quin, L. D., Eds.; Methods
in Stereochemical Analysis 8; VCH: Deerfield Beech, FL, 1987; pp 88-
93.
(15) (a) Green, L. M.; Meek, D. W. Polyhedron 1990, 9, 35. (b)
Baacke, M.; Hietkamp, S.; Morton, S.; Stelzer, O. Chem. Ber. 1981,
114, 2568.
by 1H and 31P NMR spectroscopy, IR spectroscopy, cyclic
voltammetry, and elemental analyses. [Pd(ttpE)(CH3-
CN)](BF4)2 has also been characterized by X-ray crystal-
lography, as discussed below.
The 31P NMR spectra of these complexes are sum-
marized in Table 1. The spectra of [Pd(ttpE)(CH3CN)]-
+
+

Bite Effects for [Pd(triphosphine)(solvent)](BF4)2
Organometallics, Vol. 15, No. 15, 1996 3363
(BF4)2 and [Pd(ttpC)(CH3CN)](BF4)2 each consist of a
doublet and triplet for the terminal and central phos-
phorus atoms, respectively. The chemical shifts and
coupling constants are typical of phosphorus atoms
bound to palladium in six-membered rings.15,16 The
spectrum of [Pd(eptpE)(NCCH3)]2+ consists of two dou-
blets and a doublet of doublets. The doublet of doublets
is assigned to the terminal phosphorus atom, PB, which
is in the six-membered ring. The strong trans coupling
constant of 321 Hz is consistent with either PA or PB,
but its chemical shift (3.6 ppm) is similar to that of the
terminal phosphorus atom of [Pd(ttpE)(CH3CN)](BF4)2.
PA is shifted much further downfield (76 ppm) due to
the five-membered-ring effect16 and has a large trans
coupling. The remaining resonance is assigned to the
central phosphorus atom, PX, with the doublet structure
arising from the cis coupling to PB.
The products of the reaction of [Pd(CH3CN)4](BF4)2
with metpE are solvent- and temperature-dependent.
At room temperature, two isomers are formed in aceto-
nitrile, which we propose are a monomer, [Pd(metpE)-
4+
(NCCH3)]2+ (12a ), and a dimer, [Pd(metpE)(NCCH3)]2
(13a ), in equilibrium (eq 5). When the solution is
F igu r e 1. Experimental (upward) and calculated (down-
ward) 31P NMR spectra for isomer 13a . Parameters used
in the simulation are listed in the Experimental Section.
Each minor division represents 20 Hz.
warmed to 60 °C, only the resonances assigned to the
monomer are observed. When it is cooled to -35 °C,
only resonances assigned to the dimer are observed. The
31P NMR spectrum of monomer 12a has a doublet at
-44.3 ppm (PB). This chemical shift is diagnostic of a
coordinated phosphorus atom in a four-membered ring.16
The large splitting of 230 Hz, while smaller than the
trans couplings described above, requires a trans ar-
rangement with respect to PA. The latter resonance has
a large downfield shift (87 ppm) due to its incorporation
in a five-membered ring. The remaining resonance at
49.7 ppm is an unresolved multiplet, and it is assigned
to PX. Similarly, three resonances are also assigned to
dimer 13a . In this case, the resonance for the diphe-
nylphosphino group, PB, occurs at 9.3 ppm, consistent
with incorporation into a larger eight-membered ring.
In addition to smaller coupling to PX and PX′, PB exhibits
a strong doublet splitting of 347 Hz (NAB) due to the
trans phosphine ligand, PA. The chemical shift of the
doublet for PA occurs at 98.6 ppm, consistent with
incorporation in a five-membered ring. The experimen-
tal and simulated spectra of 13a are shown in Figure
1. Dimer 13a is one of three possible isomers and has
a cis arrangement of the two PCH2P bridges. Dimers
13b and 13c have trans PCH2P bridges, with the solvent
molecule binding on either the same side or opposite
sides of the plane formed by the two diphosphine
bridges. Dimer 13a is the only dimer observed in
acetonitrile. In acetone solutions containing approxi-
mately 5% acetonitrile, three isomers are observed, as
indicated by three separate resonances for the dieth-
ylphosphino group, PA (98.6, 86.4, and 85.3 ppm). As
might be expected, there is significant spectral overlap
for resonances attributed to PM and PX and a complete
spectral analysis of the three overlapping AA′MM′XX′
spectra was not attempted. Redissolution of material
isolated from acetone solutions with no excess aceto-
nitrile contained only one dimer, 13b or 13c, which have
trans bridging groups and acetone as the solvent ligand.
Simulated and experimental spectra for this isomer are
given in Figure 3s of the Supporting Information, and
the parameters used in the simulation are listed in the
Experimental Section. Although satisfactory analytical
data were obtained for the material isolated from the
reaction of [Pd(CH3CN)4](BF4)2 with metpE, crystals
suitable for X-ray diffraction studies were not obtained,
perhaps due to the number of isomers and equilibria
involving solvent molecules. However, the 31P NMR
spectral data under different conditions are unambigu-
ous in demonstrating the existence of various isomers
of 13 and a monomer-dimer equilibrium in acetonitrile.
Tr ieth ylp h osp h in e Com p lexes. The acetonitrile
ligands in complexes 7a -12a and 13 are labile. This
(16) Garrou, P. E. Chem. Rev. 1981, 81, 229.
+
+

3364 Organometallics, Vol. 15, No. 15, 1996
Wander et al.
lability is demonstrated by the observation that the
addition of small amounts of acetonitrile to solutions of
[Pd(triphosphine)(CH3CN)](BF4)2 in acetone-d6 or CD2-
Cl2 results in a shift of the resonance of coordinated
acetonitrile toward the position of free acetonitrile.
Advantage can be taken of the lability of the acetonitrile
ligand to replace it with monodentate phosphine ligands
or hydrides. Reaction of 7a and 9a -11a with trieth-
ylphosphine readily forms the corresponding [Pd-
(triphosphine)(PEt3)](BF4)2 complexes. The 31P NMR
spectra for these complexes are summarized in Table
1. In comparison to the corresponding acetonitrile
complexes, the triethylphosphine complexes exhibit new
resonances assigned to the triethylphosphine ligands
with chemical shifts ranging from 8 to 18 ppm. These
resonances display large trans couplings between 305
and 320 Hz. The resonances assigned to the tridentate
ligand in these complexes show an additional splitting
compared to their corresponding acetonitrile complexes
due to the presence of coordinated triethylphosphine.
Because complex 12a exists in equilibrium with 13a (eq
5), it is not surprising that two products are formed in
this reaction. The major product is monomeric [Pd-
(metpE)(PEt3)]2+ (12b; ∼75%), while minor components
appear to be the triethylphosphine adducts of 13a , 13b,
and 13c (∼25%). In the case of [Pd(ttpC)(CH3CN)]-
(BF4)2, which contains bulky cyclohexyl terminal sub-
stituents, no reaction is observed with triethylphosphine
even in noncoordinating solvents such as acetone. In
contrast, the two-carbon-chain analogue [Pd(etpC)(CH3-
CN)](BF4)2 reacts readily to form [Pd(etpC)(PEt3)](BF4)2.
The three-carbon linkage in the ttpC ligand decreases
the stability of the triethylphosphine complex compared
to its two-carbon analogue. The origin of this difference
is discussed in more detail below.
from a dichloromethane/ethanol mixture produces pre-
dominantly one isomer, but the separation is not
complete. The analytical data, however, are satisfactory
as expected for isomeric impurities. The 31P NMR
spectrum of [Pd(metpE)]2(BF4)2 is consistent with an
AA′MM′XX′ spin system. The resonance for PX is
observed at -8.8 ppm, compared to a chemical shift of
-5.9 ppm for the bridging dppm ligand in [Pd(dppm)-
(CH3CN)]2(BF4)2 (dppm is bis(diphenylphosphino)-
methane).17 This resonance has a large splitting (NMX
) 420 Hz) that is largely due to the trans coupling to
PM (35.0 ppm). The resonance for PA is a complex
multiplet at 52.5 ppm. A FAB mass spectrum of the
tetrafluoroborate salt of 14 exhibits strong ion peaks
-
at m/z 1149 and 1061 consistent with loss of BF4 and
-
H(BF4)2 anions, respectively. These peaks have ap-
propriate isotopic ratios for two palladium atoms.
The hydrides [Pd(ttpE)(H)]+ (7c) and [Pd(ttpC)(H)]+
(8c) can also be formed by direct reaction of 7a and 8a
with 1 atm of hydrogen at room temperature in di-
methylformamide or acetone solution, as shown in eq
6. This reaction is reversible for 7a . Addition of HBF4
Meta l Hyd r id e Com p lexes. The hydride complexes
[Pd(ttpE)(H)]+, [Pd(ttpC)(H)]+, [Pd(eptpE)(H)]+, and
[Pd(etpC)(H)] (7c, 8c, 9c, and 11c, respectively) can be
prepared from their corresponding acetonitrile com-
plexes by reaction with NaBH4 on alumina. The 31P
NMR spectra (Table 1) of these complexes are similar
1
to those of their acetonitrile analogues. The H NMR
spectra exhibit broad doublets between -2 and -6 ppm
to dimethylformamide solutions of 7c results in the
formation of a mixture of 7a , 7c, and hydrogen (detected
by 1H NMR spectroscopy and gas chromatography). The
forward and reverse reactions shown in eq 6 have half-
lives of approximately 1.5 h. For the cyclohexyl ana-
logue 8c, the equilibrium appears to lie farther to the
right. Reaction 6 is observed only for complexes con-
taining triphosphine ligands with two trimethylene
linkages. No reaction is observed for [Pd(eptpE)(CH3-
CN)]2+ (9a ) or [Pd(etpE)(CH3CN)]2+ (10a ), which con-
tain one and two ethylene linkages, respectively.
2
with J PH values of approximately 200 Hz (see Table
1). These values are similar to those observed previ-
ously for [Pd(ttp)(H)](BF4)10 (where ttp is bis(3-(diphen-
ylphosphino)propyl)phenylphosphine). In addition, weak
Pd-H stretches are observed for Nujol mulls of each
complex (1830, 1864, and 1896 cm-1 for 7c, 8c, and 9c,
respectively).
For [Pd(etpE)(CH3CN)](BF4)2, a complex mixture of
products is formed upon reaction with NaBH4 on
alumina. Reaction of [Pd(metpE)(CH3CN)]2+ with
2+
NaBH4 produces a Pd(I) dimer, [Pd(metpE)]2 (14),
Str u ctu r a l Stu d ies. X-ray diffraction studies were
carried out on [Pd(ttpE)(CH3CN)](BF4)2 and [Pd(ttpE)-
(H)](BF4) to obtain insight into structural features that
play a role in the reactivity and stability of these
complexes. Crystals of [Pd(ttpE)(CH3CN)](BF4)2 were
grown from a mixture of dichloromethane and ethanol
-
and consist of [Pd(ttpE)(CH3CN)]2+ cations, BF4 an-
ions, and molecules of dichloromethane. A drawing
illustrating the structure and atom-numbering scheme
for the cation is shown in Figure 2. Selected bond
quantitatively by 31P NMR. No evidence of a hydride
is observed. Two isomers of 14 are expected due to the
chiral phosphorus atoms PM. Selective precipitation
(17) Miedaner, A.; DuBois, D. L. Inorg. Chem. 1988, 27, 2479.
+
+

Bite Effects for [Pd(triphosphine)(solvent)](BF4)2
Organometallics, Vol. 15, No. 15, 1996 3365
F igu r e 2. Thermal ellipsoid drawing of the cation [Pd-
(ttpE)(CH3CN)]2+ (7a ) at the 50% probability level showing
the atom-numbering scheme. Hydrogen atoms have been
omitted for clarity.
F igu r e 3. Thermal ellipsoid drawing of the cation [Pd-
(ttpE)(H)]+ (7c) at the 50% probability level showing the
atom-numbering scheme. All hydrogen atoms except for the
hydride ligand have been omitted for clarity.
Ta ble 2. Selected Bon d Dista n ces (Å) a n d Bon d
An gles (d eg) for [P d (ttp E)(CH3CN)]2+ (7a ) a n d
[P d (ttp E)(H)]+ (7c)
7a
7c
Bond Distances
2.071(6) Pd-H(1)
Pd-N
1.64(3)
Pd-P(1)
Pd-P(2)
Pd-P(3)
N-C(1)
C(1)-C(2)
2.393(2)
2.264(2)
2.383(2)
1.140(8)
1.471(10)
Pd-P(1)
Pd-P(2)
Pd-P(3)
2.2815(8)
2.2923(7)
2.2751(8)
Bond Angles
N-Pd-P(2)
178.5(2)
88.8(2)
90.5(2)
89.75(6)
90.96(6)
174.23(7)
174.0(6)
178.3(8)
118.3(2)
118.4(5)
H(1)-Pd-P(2)
H(1)-Pd-P(1)
H(1)-Pd-P(3)
P(2)-Pd-P(1)
P(2)-Pd-P(3)
P(1)-Pd-P(3)
177.2(10)
N-Pd-P(1)
86.1(10)
87.0(10)
92.83(4)
94.05(3)
173.10(2)
N-Pd-P(3)
P(2)-Pd-P(1)
P(2)-Pd-P(3)
P(1)-Pd-P(3)
C(1)-N-Pd
N-C(1)-C(2)
C(16)-P(2)-Pd
C(17)-C(16)-P(2)
C(18)-C(17)-C(16) 114.1(6)
C(17)-C(18)-P(3)
C(18)-P(3)-Pd
C(14)-P(2)-Pd
C(15)-C(14)-P(2)
C(16)-C(15)-C(14) 113.1(2)
C(15)-C(16)-P(3)
C(16)-P(3)-Pd
117.87(8)
114.3(2)
terminal and central phosphorus atoms of 15 can be
attributed to steric constraints imposed by the five-
membered rings. The bending of the terminal phos-
phorus atoms of 7c toward the hydride ligand is a
common feature of hydrides and can be attributed to
both steric and electronic effects.18
114.7(5)
117.1(2)
116.9(2)
116.23(8)
distances and bond angles are given in Table 2. Crys-
tals of [Pd(ttpE)(H)](BF4) were grown from a mixture
of acetone and ethanol at -20 °C. The crystals consist
-
of [Pd(ttpE)(H)]+ cations and BF4 anions. A drawing
The different ring sizes for 15 compared to 7a and 7c
also have another effect. Figure 4 gives edge-on views
of 15 and 7a , looking down what are nominally the
P(1)-Pd-P(3) axes. From this perspective, it can be
seen that the P(2)-Pd-N axis of 15 approximately
bisects the angle formed by the two ethyl groups of the
terminal phosphorus atoms. For 7a , one of the ethyl
groups is almost parallel to the P(2)-Pd-N axis, while
the second is slightly less than perpendicular. As a
result, one of the ethyl substituents in 7a is much closer
to the acetonitrile ligand than is the case for 15. As a
measure of this interaction, the C(3)-N and C(5)-N
distances in 7a are 4.16 and 3.23 Å, respectively, while
the analogous distances in 15 are 3.90 and 3.72 Å,
respectively. The 3.23 Å distance between C(3) and N
of the [Pd(ttpE)(H)]+ cation is shown in Figure 3 with
the atom-numbering scheme. Selected bond distances
and bond angles for this cation are also given in Table
2.
It is useful to compare some structural features of [Pd-
(ttpE)(CH3CN)]2+ (7a ) and [Pd(ttpE)(H)]+ (7c) with
those of the previously studied [Pd(MesetpE)(CH3CN)]2+
(15).11 The bond angles between the terminal and
central phosphorus atoms are very close to the ideal
value of 90° for 7a but deviate significantly for 15
(approximately 85°) and 7c (approximately 93°). Simi-
larly, the P-Pd-N bond angles between the terminal
phosphorus atoms and the nitrogen atom of the aceto-
nitrile ligand are about 5° smaller for 7a compared to
15, and the P-Pd-H angle of 7c is approximately 3°
smaller than the corresponding P-Pd-N angle in 7a .
The less than 90° P-Pd-P bond angles between the
(18) (a) Elian, M.; Hoffmann, R. Inorg. Chem. 1975, 14, 1058. (b)
Albright, T. A.; Hoffmann, R.; Thibeault, J . C.; Thorn, D. L. J . Am.
Chem. Soc. 1979, 101, 3810.
+
+

3366 Organometallics, Vol. 15, No. 15, 1996
Wander et al.
F igu r e 4. Drawing illustrating the effect of ring size on
the positioning of the substituents of the terminal phos-
phorus atom with respect to the coordinated acetonitrile
ligand for [Pd(MesetpE)(CH3CN)]2+ (15) and [Pd(ttpE)(CH3-
CN)]2+ (7a ).
for 7a is less than the sum of the van der Waals radii
for a methyl group (2.0 Å) and nitrogen (1.5 Å), and this
suggests a steric interaction between the substituents
on the terminal phosphorus atoms and the fourth ligand
in palladium complexes containing triphosphine ligands
with trimethylene linkages. The Pd-P bond distances
of terminal phosphorus atoms are 0.1 Å longer for 7a
compared to 7c. The long bond distances of 7a likely
reflect the repulsion between C(3) and N. The repulsion
appears to be manifested in the terminal Pd-P bond
distances of 7a and not in the Pd-N bond distance,
which is similar to that observed for 15 and other
second-row transition-metal complexes of acetonitrile.19
The lengthening of the Pd-P bonds and not the Pd-N
bond may arise from the fact that two methylene groups
interacting with the acetonitrile ligand oppose each
other. The resulting steric force is perpendicular to the
acetonitrile ligand and parallel to the Pd-P bonds. As
a result, it is the Pd-P bonds that are lengthened.
The Pd-H bond distances range from 1.46 to 1.79 Å
for the relatively few palladium hydrides in which the
hydrogen atom has been located by X-ray diffraction
studies.20 The Pd-H distance of 1.64 Å observed for
7c falls within this range, and it appears normal for
other second-row transition metals such as Ru and Rh.21
E lect r och em ica l St u d ies. Tr iet h ylp h osp h in e
Com p lexes. Figure 5 shows the cyclic voltammogram
F igu r e 5. Cyclic voltammogram of 1.0 × 10-3 M [Pd-
(eptpE)(PEt3)](BF4)2 in a dimethylformamide solution con-
taining 0.3 N NBu4BF4 at 20 mV/s.
Ta ble 3. Electr och em ica l Da ta for
[P d (tr ip h osp h in e)L](BF 4)2 Com p lexes in
Dim eth ylfor m a m id e
current
efficiencyc
nb CO H2
turnover
complex
E1/2
no.d
a
[Pd(ttpE)(CH3CN)](BF4)2
[Pd(ttpC)(CH3CN)](BF4)2
-1.27 (q) 1.0 95
-1.15 (q) 1.1 70
7
31
39
37
7
120
3
83
10
130
[Pd(eptpE)(CH3CN)](BF4)2 -1.45 (q) 1.4 64
e
[Pd(etpE)(CH3CN)](BF4)2 -1.25 (i) 1.1 65
e
[Pd(etpC)(CH3CN)](BF4)2 -1.28 (i) 1.0 97
[Pd(ttpE)(PEt3)](BF4)2
[Pd(eptpE)(PEt3)](BF4)2
[Pd(etpE)(PEt3)](BF4)2e
[Pd(etpC)(PEt3)](BF4)2
-1.18 (r) 1.2
-1.30 (r) 1.0
-1.35 (r) 1.9
-1.33 (r) 1.0
a
Potentials are given versus the ferrocene/ferrocenium couple
and correspond to the half-wave potential under catalytic condi-
tions for the acetonitrile complexes. The letters in parentheses
indicate whether the wave is reversible (r), quasi-reversible (q),
or irreversible (i). b Number of faradays passed per mole of complex
during bulk electrolysis at a potential approximately 0.1 V more
negative than the listed E1/2 value. Reductions of acetonitrile
complexes were carried out under an atmosphere of CO2. Reduc-
tions of triethylphosphine complexes were carried out under
nitrogen. c Current efficiencies were calculated on the assumption
that two electrons are required for CO and H2 production.
of [Pd(eptpE)(PEt3)]2+ (9) in dimethylformamide.
A
single diffusion-controlled, two-electron reduction is
observed at -1.30 V vs the ferrocene/ferrocenium couple.
A plot of the peak current vs the square root of the scan
rate is linear between 0.05 and 1.0 V/s, confirming
diffusion control over this range of scan rates. The
observed wave has a peak-to-peak separation of 33 mV
and an ip,a/ip,c ratio of 0.96 at a scan rate of 20 mV/s.
These data are consistent with a reversible two-electron
transfer, for which a peak-to-peak separation of 30 mV
and an ip,a/ip,c ratio of 1.0 are expected.22 However,
controlled-potential electrolysis of this complex at -1.4
V resulted in the passage of 1.0 faraday/mol of complex,
indicating that, on the longer time scale required for
bulk electrolysis, a one-electron reduction is observed.
d
Number of moles of CO formed per mole of catalyst after current
had decayed to 5-10% of initial value. e Data from ref 10a.
Similar experiments were carried out for the complexes
[Pd(ttpE)(PEt3)](BF4)2 and [Pd(etpE)(PEt3)](BF4)2, and
the results are summarized in Table 3. In order to
obtain a fully reversible wave for [Pd(ttpE)(PEt3)](BF4)2,
it was necessary to add a 5-10-fold excess of trieth-
ylphosphine to retard ligand dissociation. Under these
conditions, a reversible two-electron reduction is ob-
served using the criterion of peak-to-peak separation
and the ratio of anodic to cathodic currents. Data are
not reported for [Pd(metpE)(PEt3)](BF4)2, since this
compound could not be isolated in pure form as dis-
cussed above. It can be seen from the data in Table 3
that increasing the bite size of the tridentate ligand
leads to a less negative redox potential for the Pd(II)/
Pd(0) couple. This trend is expected since larger chelate
bites should produce more stable tetrahedral Pd(0)
complexes, which would ideally have P-Pd-P bond
angles of 109°.
(19) Storkhoff, B. N.; Lewis, H. C., J r. Coord. Chem. Rev. 1977, 23,
1.
(20) (a) Bugno, C. D.; Pasquali, M.; Leoni, P.; Sabatino, P.; Braga,
D. Inorg. Chem. 1989, 28, 1390. (b) Leoni, P.; Sommovigo, M.; Pasquali,
M.; Midollini, S.; Braga, D.; Sabatino, P. Organometallics 1991, 10,
1038. (c) Portnoy, M.; Frolow, F.; Milstein, D. Organometallics 1991,
10, 3960. (d) Fryzuk, M. D.; Lloyd, B. R.; Clentsmith, G. K. B.; Rettig,
S. J . J . Am. Chem. Soc. 1991, 113, 4332.
(21) (a) Fryzuk, M. D.; J ones, T.; Einstein, F. W. B. Organometallics
1984, 3, 185. (b) J ia, G.; Meek, D. W.; Gallucci, J . Inorg. Chem. 1991,
30, 403.
(22) Bard, A. J .; Faulkner, L. R. Electrochemical Methods; Wiley:
New York, 1980; p 218.
Aceton itr ile Com p lexes. The cyclic voltammogram
+
+

Bite Effects for [Pd(triphosphine)(solvent)](BF4)2
Organometallics, Vol. 15, No. 15, 1996 3367
faraday/mol for 7b. This result supports a one-electron
reduction of 7a in the presence of CO2. Bulk electrolysis
of 7a under CO2 is again consistent with the chrono-
coulometric results, as 1.0 faraday/mol of palladium
complex was consumed. The 31P NMR spectra of the
electrolysis products under nitrogen and CO2 were
identical with that observed for [Pd(ttpE)(H)](BF4). The
results show that the hydride 7c can be formed by either
one- or two-electron-reduction pathways. Under an
atmosphere of CO2, only the one-electron pathway is
operative, indicating that the Pd(I) complex acts as a
radical abstracting a hydrogen atom from the solvent.
Under a nitrogen atmosphere, transient Pd(I) and
Pd(0) species react to abstract a hydrogen atom or a
proton, respectively. In the latter case, the proton is
likely obtained from trace amounts of water remaining
in the solvent or from the supporting electrolyte.
Similarly, [Pd(triphosphine)(H)]+ complexes 8b, 9b, and
11b are formed as the major products during electro-
chemical reduction of the corresponding [Pd-
(triphosphine)(CH3CN)](BF4)2 complexes in dimethyl-
formamide solutions saturated with CO2. For
[Pd(etpE)(CH3CN)]2+ (10a ) electrochemical reduction
produces dimer 1a in nearly quantitative yield as
reported previously.10a
In the presence of CO2, a second reduction wave is
observed for [Pd(ttpE)(CH3CN)](BF4)2 as shown in
Figure 6 (trace b). This wave may be due to either a
PdI(CO2) complex or to [Pd(ttpE)(H)](BF4), which has a
reduction wave at the same potential (within 0.02 V),
as verified by cyclic voltammograms of the hydride in
the presence of CO2. Similar results were obtained for
the other complexes that form hydrides, which suggests
that hydride formation is responsible for the second
wave and that the lifetime of the PdI(CO2) complexes is
short (less than 0.1 s). In the presence of acid, the
PdI(CO2) intermediates undergo further reduction to
form CO in a catalytic process, as discussed below.
F igu r e 6. Cyclic voltammograms of dimethylformamide
solutions of 2.90 × 10-3 M [Pd(ttpE)(CH3CN)](BF4)2 under
(a) nitrogen (s) and (b) CO2 (- - -) and cyclic voltammo-
grams of dimethylformamide solutions containing 2.70 ×
10-2 M HBF4 and 2.30 × 10-3 M [Pd(ttpE)(CH3CN)](BF4)2
under (c) nitrogen (s) and (d) CO2 (- - -).
of [Pd(ttpE)(CH3CN)]2+ (7a ) shown by the solid line (a)
in Figure 6 is typical of the acetonitrile complexes 7a -
9a under nitrogen atmospheres. All of the complexes
exhibit an irreversible or quasi-reversible cathodic wave
between -1.15 and -1.45 V vs the ferrocene/ferroce-
nium couple. These waves are diffusion-controlled
between 0.05 and 1.0 V/s, as indicated by linear plots
of peak currents vs the square root of the scan rate. To
determine the number of electrons involved in the
reduction wave for 7a , the slopes of chronocoulometric
plots of charge vs t1/2 were compared for 7a and 7b. For
such plots, the slope is proportional to the number of
electrons involved in the charge-transfer step.23 In the
presence of excess triethylphosphine, a two-electron-
reduction process was established for 7b as discussed
above. The observed ratio of the slope of 7a to 7b was
found to be 0.85, indicating approximately 1.7 faraday/
mol of complex is passed during the reduction of 7a . This
value is consistent with the 1.6 faraday/mol of complex
observed for bulk electrolysis experiments carried out
under nitrogen at -1.4 V. These data suggest that both
one- and two-electron processes are occurring during the
reduction of 7a . In the presence of CO2, the voltam-
mogram shown by the dashed line (b) in Figure 6 is
observed. The cathodic wave is decreased in height and
shifted in a positive direction, which is consistent with
a fast chemical reaction of the reduction product of 7a
with CO2. The ratio of the slopes of the chronocoulo-
metric plots of the first reduction process for 7a under
CO2 and 7b under nitrogen is 0.57. This value corre-
sponds to 1.1 faraday/mol of 7a assuming a value of 2.0
Ca ta lytic Stu d ies. Cyclic voltammograms of [Pd-
(ttpE)(CH3CN)](BF4)2 in acidic dimethylformamide so-
lutions (0.027 M HBF4) saturated with nitrogen and CO2
at 620 mmHg are shown by traces c and d, respectively,
of Figure 6. In the presence of CO2, a current enhance-
ment is observed. To determine the nature of the
products formed during the reduction process, a con-
trolled-potential electrolysis of 7a in the presence of CO2
and acid was carried out in a sealed cell at -1.4 V. Gas
chromatographic analysis of the gases formed during
the electrolysis indicated that, initially, approximately
95% of the charge passed resulted in the production of
CO, while 7% produced hydrogen. The selectivity for
CO gradually decreases with time. Similar experiments
were carried out for 8a and 9a , and the resulting data
are listed in Table 3. A turnover number of 120 was
calculated on the basis of the moles of CO produced per
mole of catalyst. The catalyst was considered deacti-
vated when the current had decayed to approximately
10% of its initial value. Similar experiments were
carried out for other acetonitrile complexes, and the
results are presented in Table 3.
The dependence of the catalytic current on substrate
and catalyst concentrations is shown in Figures 7 and
8. Plots of the catalytic currents for 7a and 8a vs the
acid concentration are biphasic (Figure 7, open circles).
The catalytic current initially increases with acid con-
(23) Anson, F. C. Anal. Chem. 1966, 38, 54.
+
+

3368 Organometallics, Vol. 15, No. 15, 1996
Wander et al.
F igu r e 8. (Top) Dependence of the catalytic current on
CO2 concentration for a dimethylformamide solution con-
taining 0.04 M HBF4 and 2.30 × 10-3 M [Pd(ttpE)(CH3-
CN)](BF4)2. The scan rate was 0.05 V/s. (Bottom) Depen-
dence of the catalytic current on [Pd(ttpE)(CH3CN)](BF4)2
concentrations for dimethylformamide solutions containing
0.04 M HBF4 and 0.18 M CO2. The scan rate was 0.1 V/s.
F igu r e 7. (Top) Dependence of the peak current on acid
concentration for 2.30 × 10-3 M solutions of [Pd(ttpE)(CH3-
CN)](BF4)2 saturated with nitrogen (solid circles) and
carbon dioxide (open circles). All data were recorded at a
scan rate of 0.05 V/s. The solid line shows the best-fit line
using eq 7 with k1 ) 22 M-1 s-1 and k2 ) 790 M-1 s-1. The
dashed line represents the best-fit line assuming a second-
order dependence on acid with k1 ) 13 M-1 s-1 and k2 )
3.7 × 104 M-2 s-1. Because of the scan rate dependence of
the peak current, k1 and k2 cannot be regarded as true rate
constants, as discussed in the text. (Bottom) Dependence
of the peak current on acid concentration for 1.90 × 10-3
M solutions of [Pd(eptpE)(CH3CN)](BF4)2 saturated with
nitrogen (solid circles) and carbon dioxide (open circles).
All data were recorded at 0.05 V/s. The solid line shows
the best-fit line using eq 7 with k1 ) 43 M-1 s-1 and k2 )
400 M-1 s-1. The dashed line represents the best-fit line
assuming a second-order dependence on acid with k1 ) 30
M-1 s-1 and k2 ) 6.3 × 104 M-2 s-1. Because of the scan
rate dependence of the peak current, k1 and k2 cannot be
regarded as true rate constants, as discussed in the text.
The catalytic current for 7a was decreased 25-30%
when the CO2 pressure was held constant at 1 atm and
the partial pressure of CO was increased to 0.5 atm and
allowed to equilibrate for approximately 15 min. This
indicates that the product, CO, inhibits its production.
However, no evidence was observed by 31P NMR spec-
troscopy for the reaction of 7a with CO when an acidic
dimethylformamide solution was saturated with CO at
1 atm. These results suggest that CO inhibition occurs
by binding to reduced forms of the catalyst. Similar
experiments were carried out with 0.25 atm of H2. After
equilibration for 15 min, no significant inhibition was
observed.
Qu a lita tive Molecu la r Or bita l Ca lcu la tion s. Ex-
tended Hu¨ckel calculations were carried out for [Pd-
(PH3)3]2+ and [Pd(PH3)3H]+ to obtain a greater under-
standing of the influence of the chelate bite on the
hydride complexes. Two parameter sets were used in
these calculations and are listed in the Supporting
Information. One was a relatively hard parameter set
used by Thorn and Hoffmann for modeling olefin inser-
tion reactions in Pt-H complexes.5 The second param-
eter set was taken from the default values of a standard
molecular modeling package.25 Both parameter sets
give the same trends. As the bond angle between the
central phosphorus atom and the two terminal phos-
phorus atoms decreases from 100 to 80°, the calculated
charge on the hydride ligand of [Pd(PH3)3H]+ becomes
more negative (-0.27 and -0.51, respectively, for the
hard parameter set and -0.61 and -0.73, respectively,
centration and then becomes independent of it. This
dependence is discussed in more detail below. The solid
circles in Figure 7 show the currents observed for
solutions of different acid concentrations in the absence
of CO2. At all acid concentrations, a significant current
enhancement is observed for solutions purged with CO2
compared to those purged with N2. Plots of the catalytic
current vs the square root of the CO2 concentration and
vs the catalyst concentration are shown in Figure 8 for
compound 7a . The square root dependence of the
catalytic current on CO2 concentration is consistent with
a rate-determining step that is first order in CO2.24 The
linear dependence of the current on catalyst concentra-
tion implies a first-order dependence on the catalyst
concentration as well. The catalytic current of 8a also
shows a square root dependence on the CO2 concentra-
tion and a linear dependence on catalyst concentration.
(24) Save´ant, J . M.; Vianello, E. Electrochim. Acta 1963, 8, 905.
(25) CAChe Scientific, Inc., 13475 SW Karl Braun, Beaverton, OR.
+
+

Bite Effects for [Pd(triphosphine)(solvent)](BF4)2
Organometallics, Vol. 15, No. 15, 1996 3369
for the softer parameter set). This result implies that
the hydridic nature of metal hydride complexes can be
influenced by controlling the P-M-P bond angle. Hy-
dride complexes with ligands that have small chelate
bites should be more hydridic and less acidic than
analogous complexes with ligands that have larger
chelate bite sizes. The origin of this effect can be traced
NaBH4 on alumina produce the corresponding hydrides.
These results demonstrate that small chelate bites favor
formation of Pd(I) dimers in the presence of NaBH4
while large chelate bites promote hydride formation.
The differences in the products formed in the reactions
of the various triphosphine ligands with [Pd(NCCH3)4]-
(BF4)2 and the products formed during the reaction of
the [Pd(triphosphine)(CH3CN)](BF4)2 complexes with
NaBH4 can be readily understood in terms of ring strain
arguments.
to the LUMO or dx -y orbital of the [Pd(PH3)3]2+
fragment. This fragment orbital contains a significant
contribution from the px orbital, which results in the
polarization
2
2
There are other differences in the reactivity patterns
of the mononuclear [Pd(triphosphine)(CH3CN)](BF4)2
complexes containing two-carbon linkages compared to
those with three-carbon chains which are not readily
understood on the basis of differences in ring strain
present in reduced intermediates. For example, the
reaction of triethylphosphine with [Pd(etpC)(CH3CN)]-
(BF4)2 in noncoordinating solvents such as acetone
produces the corresponding triethylphosphine complex
[Pd(etpC)(PEt3)](BF4)2, but [Pd(ttpC)(CH3CN)](BF4)2 is
unreactive with triethylphosphine under the same
conditions. The rate of triethylphosphine exchange for
[Pd(ttpE)(PEt3)](BF4)2 with free triethylphosphine also
appears to be more rapid than that of [Pd(etpE)(PEt3)]-
(BF4)2. In deuterionitromethane solutions containing
1 equiv of triethylphosphine, broad resonances are
observed for the 31P NMR spectrum of [Pd(ttpE)(PEt3)]-
(BF4)2 at room temperature, while the resonances of [Pd-
(etpE)(PEt3)](BF4)2 are sharp. Excess triethylphosphine
is required to obtain a fully reversible reduction wave
for [Pd(ttpE)(PEt3)](BF4)2, while no additional trieth-
ylphosphine is required to obtain reversible waves for
[Pd(etpE)(PEt3)](BF4)2 and [Pd(eptpE)(PEt3)](BF4)2. This
is consistent with triethylphosphine dissociation at low
concentrations for the former complex but not for the
latter complexes. A comparison of the X-ray crystal
structures of [Pd(MesetpE)(CH3CN)](BF4)2 and [Pd-
(ttpE)(CH3CN)](BF4)2 is helpful in understanding these
results. As can be seen from Figure 3, one of the
methylene carbons of the two ethyl substituents on each
terminal phosphorus atom of [Pd(ttpE)(CH3CN)](BF4)2
has a close contact with the nitrogen atom of the
acetonitrile ligand. This repulsive interaction would be
expected to increase if the acetonitrile ligand is replaced
with triethylphosphine and apparently becomes pro-
hibitive for formation of the triethylphosphine complex
[Pd(ttpC)(PEt3)](BF4)2. The differences in the reactivity
of the [Pd(triphosphine)(CH3CN)](BF4)2 complexes with
triethylphosphine are attributed to the steric interaction
between triethylphosphine and the substituents on the
terminal phosphorus atoms that lie along the aceto-
nitrile axis.
As the P-M-P angle (θ) increases, this orbital de-
creases in energy, causing electron density to move from
the hydride ligand into this orbital. As a result, large
P-M-P angles will produce more acidic hydrides. The
lower energy of this orbital should also facilitate electron
transfer from molecular hydrogen to this orbital, pro-
moting heterolytic cleavage of hydrogen.
Discu ssion
A variety of new triphosphine ligands have been
prepared which contain one, two, and three bridging
methylene groups between the central and terminal
phosphorus atoms of the triphosphine ligand. These
ligands permit a systematic study of the effect of chelate
bite size for triphosphine ligands on various properties
of metal complexes. Because of our interest in CO2
reduction by [Pd(triphosphine)(CH3CN)](BF4)2 com-
plexes, we have focused on palladium complexes of the
type [Pd(triphosphine)(CH3CN)](BF4)2, [Pd(triphosphine)-
(PEt3)](BF4)2, and [Pd(triphosphine)(H)](BF4).
The reactions of [Pd(NCCH3)4](BF4)2 with triphos-
phine ligands containing two-carbon, three-carbon, and
mixed two- and three-carbon chains result in the forma-
tion of mononuclear [Pd(triphosphine)(CH3CN)](BF4)2
complexes (eq 4). The reaction of metpE, which contains
a two-carbon and a one-carbon chain, with [Pd(NCCH3)4]-
(BF4)2 in acetonitrile produces an equilibrium mixture
of a monomeric (12a ) and a dimeric species (13a ) (eq
5). In the latter complex, the strained four-membered
ring of the monomer has opened to form a bridge to a
second palladium atom. This equilibrium can be shifted
by varying the temperature and the solvent. The
formation of such a dimeric species is not surprising,
given the large number of dimeric species formed by
ligands such as dppm with a methylene bridge between
the phosphorus atoms.26
The reaction of hydrogen with [Pd(triphosphine)(S)]-
(BF4)2 complexes (where S is a weakly coordinating
solvent such as DMF or acetone) is another reaction that
exhibits a strong dependence on the chelate bite size.
Only complexes with two trimethylene backbones have
been observed to react with hydrogen to form the
corresponding hydride, as shown in eq 6. One possible
explanation for these differences is that the steric
interaction observed in the formation of the trieth-
ylphosphine complexes also plays an important role in
the formation of the hydride complex. Steric interac-
tions between the coordinated solvent molecules and the
terminal substituents in the [Pd(triphosphine)(S)](BF4)2
Reaction of the equilibrium mixture of 12a and 13a
with NaBH4 on alumina results in the clean and rapid
formation of the Pd(I) dimer 14, in which ring strain
has been reduced by the formation of the larger five-
membered rings upon Pd-Pd bond formation. The
analogous reactions of the complexes 7a -9a with
(26) (a) Puddephett, R. J . Chem. Soc. Rev. 1983, 12, 99. (b) Miedaner,
A.; DuBois, D. L. Inorg. Chem. 1988, 27, 2479.
(27) (a) Baacke, M.; Hietkamp, S.; Morton, S.; Stelzer, O. Chem. Ber.
1981, 114, 2568. (b) Miedaner, A.; Curtis, C. J .; Barkley, R. M.; DuBois,
D. L. Inorg. Chem. 1994, 33, 5482. (c) King, R. B.; Cloyd, J . C., J r. J .
Am. Chem. Soc. 1975, 97, 46.
+
+

3370 Organometallics, Vol. 15, No. 15, 1996
Wander et al.
complexes with trimethylene chains could be significant
and destabilize the reactants by weakening the Pd-S
bond. These steric interactions would be expected to
be less important for the products, [Pd(ttpE)(H)](BF4)
and [Pd(ttpC)(H)](BF4), because of the smaller steric
requirements of the hydride ligand. These consider-
ations would suggest that steric interactions should
favor formation of the hydride complexes for complexes
with three-carbon chains. The lower energy of the
current, id is the peak current due to reversible reduc-
tion of the catalyst, v is the scan rate, n is the number
of electrons involved in catalyst reduction, k1 and k2 are
rate constants for the two rate-determining steps dis-
cussed below, and σ is a factor that depends on whether
the electron transfer to the catalyst occurs at the
electrode surface or in solution. Equation 7 assumes a
first-order dependence of the catalytic rate on acid at
low acid concentrations and first-order dependence on
CO2 at high acid concentrations. For [Pd(triphosphine)-
(CH3CN)](BF4)2 complexes containing triphosphine
ligands with two ethylene bridges, two rate-determining
steps are also observed. For the ethylene-bridged
complexes, a second-order dependence of the catalytic
rate on acid concentration has been observed for several
catalysts at low acid concentrations and a first-order
dependence on CO2 at high acid concentrations.10,11
These results suggest that the rate-determining step for
the reduction of CO2 is the same at high acid concentra-
tions for [Pd(ttpE)(CH3CN)](BF4)2, [Pd(eptpE)(CH3CN)]-
(BF4)2, and the catalysts with two-carbon-chain back-
bones. This step has been proposed previously to be the
reaction of a Pd(I) intermediate with CO2 as shown in
eq 8. This is consistent with the one-electron reduction
2
2
LUMO or dx -y orbital for large P-M-P angles indi-
cates that there is also an electronic component that
favors hydride formation.
A second interesting aspect of the hydride complexes,
suggested by the molecular orbital calculations dis-
cussed above, is the possibility of using the chelate bite
size to partially control the hydridic character of the
hydride ligand. The smaller the P-M-P bond angle,
the more hydridic the hydride ligand should become,
whereas larger bond angles favor more acidic hydride
ligands. In our hands, the addition of HBF4 to dimeth-
ylformamide solutions of [Pd(etpC)(H)](BF4), which
contains five-membered chelate rings, results in the
rapid evolution of H2 and the formation of [Pd(etpC)-
(DMF)](BF4)2. For [Pd(ttpC)(H)](BF4), no reaction is
observed under these conditions. Similarly, many
square-planar hydrides are synthesized by protonation
of zerovalent metal complexes using weak to strong
acids.20,28 The complex [(PPh3)3PdH]+ is relatively
stable in aqueous solutions containing up to 80% CF3-
COOH,29 and the hydride ligand in trans-[(Cy3P)2Pd-
(H)(H2O)]+ exchanges more rapidly with D2O than the
protons bound to coordinated water.20b These results
are consistent with acidic hydride complexes. The
average O-Pd-P bond angle in the latter complex is
approximately 95°. An even larger angle would be
expected for [(PPh3)3PdH]+.
observed for [Pd(ttpE)(CH3CN)](BF4)2 in the presence
of CO2, as discussed above in the Results Section. At
low acid concentrations, there appears to be a subtle
difference in the mechanism of the rate-determining
step. For the top graph of Figure 7, it can be seen that
the data are best fit by the solid line, which corresponds
to a first-order dependence on acid at low acid concen-
trations. The dotted line, which corresponds to a
second-order dependence on acid, does not fit as well.
The first-order dependence of the catalytic rate on acid
observed for [Pd(ttpE)(CH3CN)](BF4)2 suggests loss of
hydroxide from the activated complex as shown in eq
9, whereas the second-order dependence observed for
catalysts containing ethylene linkages is consistent with
loss of water. Precedents for both pathways exist in the
literature.30-32 The greater steric interaction between
the ethyl substituents on the terminal phosphorus
atoms and the hydroxycarbonyl ligand should promote
loss of hydroxide from [Pd(ttpE)(COOH)]+. There are
steps in addition to eqs 8 and 9 that must occur during
the catalytic reduction of CO2 to CO by [Pd(triphos-
phine)(CH3CN)](BF4)2 complexes. An electron transfer
precedes reaction 8, and a second electron transfer,
protonation, and solvent loss occur between steps 8 and
9. Following reaction 9, CO must be replaced by a
solvent molecule to regenerate a [Pd(triphosphine)-
Electr oca ta lytic Red u ction of CO2. The increased
catalytic current observed in the presence of CO2 for
acidic dimethylformamide solutions of [Pd(ttpE)(CH3-
CN)]2+ (7a ) suggested that this compound catalyzes the
electrochemical reduction of CO2 (Figure 5, traces c and
d). This was confirmed by controlled-potential elec-
trolysis experiments carried out at -1.4 V and identi-
fication of CO as the major reduction product by gas
chromatography. Similar results were obtained for [Pd-
(eptpE)(CH3CN)]2+ (9a ). The dependence of the cata-
lytic current on acid concentration for 7a and 9a (Figure
7) is consistent with two different rate-determining
steps. At low acid concentrations, the catalytic current
is dependent on the acid concentration, but at higher
acid concentrations, no dependence is observed. The
best-fit line of these data using eq 7 is shown by the
ic
k1k2[CO2][H+]
1/2
1/2
σ
RT
)
(7)
+
(
) (
)
id 0.447 nFv
k1[CO2] + k2[H ]
solid curves in Figure 7. In eq 7, ic is the catalytic
(28) (a) Gerlach, D. H.; Kane, A. R.; Parshall, G. W.; J esson, J . P.;
Muetterties, E. L. J . Am. Chem. Soc. 1971, 3544. (b) Siedle, A. R.;
Newmark, R. A.; Gleason, W. B. Inorg. Chem. 1991, 30, 2005. (c)
Seligson, A. L.; Cowan, R. L.; Trogler, W. C. Inorg. Chem. 1991, 30,
3371.
(29) (a) Zudin, V. N.; Chinakov, V. D.; Nekipelov, V. M.; Rogov, V.
A.; Likholobov, V. A.; Yermakov, Y. I. J . Mol. Catal. 1989, 52, 27. (b)
Zudin, V. N.; Chinakov, V. D.; Nekipelov, V. M.; Likholobov, V. A.;
Yermakov, Y. I. J . Organomet. Chem. 1985, 289, 425.
(30) (a) Byrd, J . E.; Halpern, J . J . Am. Chem. Soc. 1971, 93, 1634.
(b) Catellani, M.; Halpern, J . Inorg. Chem. 1980, 19, 566.
(31) Bennett, M. A.; Robertson, G. B.; Rokicki, A.; Wickramasinghe,
W. A. J . Am. Chem. Soc. 1988, 110, 7098.
(32) Ford, P. C.; Rokicki, A. Adv. Organomet. Chem. 1988, 28, 139.
+
+

Bite Effects for [Pd(triphosphine)(solvent)](BF4)2
Organometallics, Vol. 15, No. 15, 1996 3371
(solvent)]2+ species, which is reduced to start the next
catalytic cycle. Evidence for these reactions has been
presented previously and is not repeated here.10 How-
ever, steps 8 and 9 are rate-determining and therefore
control the kinetic behavior of [Pd(ttpE)(CH3CN)](BF4)2.
[Pd(eptpE)(CH3CN)](BF4)2, which contains both tri-
methylene and ethylene bridges in the triphosphine
ligand, appears to exhibit behavior midway between
that expected for complexes containing two ethylene or
two trimethylene linkages. As shown by the bottom
graph of Figure 7, the catalytic current appears to fall
between the best-fit curves for first- and second-order
acid dependence. This may indicate that both pathways
are available to this catalyst. At high acid concentra-
tions, the catalytic rate is first order in CO2 and catalyst,
as previously observed for complexes with two ethylene
linkages.
Ideally, the catalytic current should not depend on
the scan rate.22 However, that is not true for complexes
7a -9a , perhaps due to product inhibition by CO or to
rapid catalyst decomposition. Because of this behavior,
it is not meaningful to compare the catalytic rates of
these complexes to those of their two-carbon-chain
analogues reported earlier.10,11 It is interesting, how-
ever, to compare the decomposition products formed
during the catalytic process. The electrochemical re-
duction of [Pd(etpE)(CH3CN)](BF4)2, [Pd(eptpE)(CH3-
CN)](BF4)2, [Pd(ttpE)(CH3CN)](BF4)2, and [Pd(ttpC)(CH3-
CN)](BF4)2 in the presence of CO2 leads to the formation
of [Pd(etpE)]2(BF4)2, [Pd(eptpE)(H)](BF4), [Pd(ttpE)(H)]-
(BF4), and [Pd(ttpC)(H)](BF4), respectively, as the major
products. These are the same products that are ob-
served by 31P NMR spectroscopy of the crude reaction
mixtures at the end of catalytic experiments. Increasing
the chain length of the triphosphine ligand from two
carbons to three carbons apparently prevents dimer
formation, but a new decomposition pathway that
involves abstraction of a hydrogen atom from the
reaction medium becomes dominant. Due to the stabil-
ity of the hydride complexes containing three-carbon
chains in acidic media, hydride formation constitutes
an effective decomposition pathway. The stability of the
hydrides containing trimethylene linkages under rela-
tively acidic solutions and their formation by direct
heterolytic cleavage of hydrogen suggest that these
complexes may have other interesting catalytic proper-
ties. This area will be the subject of future studies.
chelate bites. The ligand containing one ethylene and
one methylene backbone, metpE, forms bridged, dimeric
complexes for Pd(II) and Pd(I) oxidation states. For
ligands with two ethylene backbones, bridged complexes
are not observed for the Pd(II) oxidation state, but
Pd(I) dimers are formed readily during electrochemical
reductions. For complexes containing two trimethylene
backbones, hydride species are formed exclusively under
similar conditions.
Crystal structures of [Pd(ttpE)(CH3CN)](BF4)2 and
[Pd(ttpE)(H)](BF4) indicate that a significant steric
interaction exists between the substituents on the
terminal phosphorus atoms of the triphosphine ligand
and the fourth ligand in square-planar complexes. This
steric interaction undoubtedly plays a significant role
in the failure of [Pd(ttpC)(CH3CN)](BF4)2 to coordinate
triethylphosphine. Formation of the hydrides [Pd(ttpE)-
(H)](BF4) and [Pd(ttpC)(H)](BF4) from the analogous
solvent complexes is also favored by weakening the
Pd-S bond. Finally, this steric interaction promotes the
loss of hydroxide from the hydroxycarbonyl complex
formed during electrochemical reduction of CO2 to CO.
In the analogous ethylene-bridged complexes, water is
lost rather than hydroxide.
The larger bite size of the complexes with trimethyl-
ene linkages also produces an electronic stabilization
of the hydride complexes compared to their two-carbon-
chain analogues. This suggests the possibility of using
chelating ligands with large bite sizes as a strategy for
promoting heterolytic cleavage of hydrogen by com-
plexes containing weakly coordinating solvent mol-
ecules. This aspect will be pursued in future work. In
addition, the qualitative molecular orbital calculations
used here indicate that the chelate bite can be used to
modulate the charge associated with hydride ligands.
Exp er im en ta l Section
Ma ter ia ls a n d P h ysica l Meth od s. Details of methods
used for drying solvents, spectral measurements, and gas
chromatography experiments have been presented else-
where.10,11 Elemental analyses were performed by Schwarz-
kopf Microanalytical Laboratory, Inc. The following reagents
were purchased from commercial suppliers and used as
obtained: allylmagnesium chloride, vinylmagnesium bromide,
phenylphosphine, chlorodiethylphosphine, NaBH4 supported
on basic alumina, triethylphosphine, butyllithium, and 2,2′-
azobis(2-methylproprionitrile) (AIBN). Bis(3-(dicyclohexyl-
phosphino)propyl)phenylphosphine (ttpC),15a [Pd(NCCH3)4]-
(BF4)2,33 diethylvinylphosphine,34 (chloromethyl)diphenyl-
Su m m a r y a n d Con clu sion s
10
phosphine,13 and [Pd(etpC)(CH3CN)](BF4)2 were prepared
according to literature methods.
One of the original objectives of this work was to
prepare [Pd(triphosphine)(CH3CN)](BF4)2 complexes in
which the triphosphine ligand contained two trimeth-
ylene chains. It was felt that such complexes would
prevent the formation of Pd(I) dimers of type 1 that were
observed as the decomposition products during the
electrochemical reduction of CO2 catalyzed by [Pd-
(triphosphine)(CH3CN)](BF4)2 complexes with two eth-
ylene bridges. Although dimer formation was not found
to be a major decomposition pathway for the complexes
containing two trimethylene bridges, catalyst lifetime
was not significantly improved, since a new decomposi-
tion pathway, the formation of [Pd(triphosphine)(H)]-
(BF4) complexes, emerged to replace it.
Coulometric measurements were carried out at 25-30 °C
using a Princeton Applied Research Model 173 potentiostat
equipped with a Model 179 digital coulometer and a Model
175 universal programmer. The working electrode was con-
structed from a reticulated vitreous carbon rod with a diameter
of 1 cm and length of 2.5 cm (60 pores per inch, The
Electrosynthesis Co., Inc.), the counter electrode was a Pt wire,
and a Ag wire immersed in a permethylferrocene/permethyl-
ferrocenium solution was used as the pseudoreference elec-
trode.35 The electrode compartments were separated by Vycor
disks (7 mm diameter, EG&G Princeton Applied Research).
(33) (a) Sen, A.; Ta-Wang, L. J . Am. Chem. Soc. 1981, 103, 4627.
(b) Hathaway, B. J .; Holah, D. G.; Underhill, A. E. J . Chem. Soc. 1962,
2444.
(34) Askham, F. R.; Stanley, G. G.; Marques, E. C. J . Am. Chem.
Soc. 1985, 107, 7423.
As expected, the tendency to form bridged complexes
is greater for the triphosphine ligands with smaller
(35) Bashkin, J . K.; Kinlen, P. J . Inorg. Chem. 1990, 29, 4507.
+
+

3372 Organometallics, Vol. 15, No. 15, 1996
Wander et al.
Measurements of current efficiencies for gas evolution were
carried out in a sealed flask from which aliquots were removed
for GC analysis. Catalytic coulometric experiments were
considered complete when the current had decayed to 5-10%
of its original value in the presence of 0.18 M CO2 and 0.1 M
HBF4, i.e., the same as the initial conditions. Cyclic voltam-
metry and chronoamperometry experiments were carried out
using a Cypress System computer-aided electrolysis system
at 21 ( 0.5 °C unless stated otherwise. The working electrode
was a glassy-carbon electrode (Cypress Systems, Inc.). Fer-
rocene was used as an internal standard, and all potentials
are reported vs the ferrocene/ferrocenium couple.36 The
solubility of CO2 in dimethylformamide was taken from ref
37. All solutions for CV and coulometric experiments were
0.3 N NEt4BF4 in dimethylformamide unless otherwise stated.
FAB mass spectra were recorded on a VG Analytical 7070 EQ-
HF tandem mass spectrometer. Samples were dissolved in
nitrobenzyl alcohol (NOBA).
Syn th eses. Allyld ieth ylp h osp h in e.38 A 500 mL three-
necked flask was fitted with an N2 inlet, mechanical stirrer,
and 250 mL addition funnel. PEt2Cl (30 g, 0.24 mol) was
transferred by cannula to the flask, dissolved in 150 mL of
THF, and cooled to -80 °C. Allylmagnesium chloride (120 mL
of a 2 M solution in THF, 0.24 mol) was transferred by cannula
to the addition funnel and added dropwise over 1 h to the
phosphine solution. After complete addition, the slurry of a
white precipitate in a pale gray solution was warmed to room
temperature and stirred for 2 h. The volatile components were
transferred under vacuum, and the product was collected by
fractional distillation (138-140 °C/615 mmHg) as a clear,
colorless liquid (24.7 g, 78%). 1H NMR (toluene-d8): δ 5.69
(m, 1H, CH2dCH); 4.93 (m, 2H, CH2dCH); 2.04 (d, 2H,
CHsCH2P); 1.20 (m, 4H, CH2CH3); 0.96 (m, 6H, CH2CH3). 31P-
{1H} NMR (toluene-d8): -23.7 ppm (s).
98%) was washed with hexane (2 × 30 mL) and dried in vacuo
for 1 h. Caution! This compound is pyrophoric. 1H NMR
(THF-d8): δ 6.2-7.5 (br m, 5H, Ph); 1.9 (br s, 2H, CH2); 1.6
(br s, 2H, CH2); 1.4 (m, 4H, CH2CH3); 1.0 (m, 6H, CH2CH3).
31P{1H} NMR (THF-d8): -22.7 ppm (d, Et2P, 3J ) 13.4 Hz),
-40.1 ppm (br, PhP).
(2-(Die t h ylp h osp h in o)e t h yl)(3-(d ie t h ylp h osp h in o)-
p r op yl)p h en ylp h osp h in e (ep tp E; 5). A mixture of (2-
(diethylphosphino)ethyl)phenylphosphine (4; 4.5 g, 0.02 mol),
allyldiethylphosphine (2.6 g, 0.02 mol), and AIBN (100 mg)
was irradiated for 1.5 days in a sealed Schlenk flask, as
described above. The crude product was distilled at 163-165
°C at reduced pressure (0.01 mmHg) to obtain a clear, pale
yellow oil (5.4 g, 76%).
(2-(Die t h ylp h osp h in o)e t h yl)((d ip h e n ylp h osp h in o)-
m eth yl)p h en ylp h osp h in e (m etp E; 6). A -78 °C THF
solution of lithium (2-(diethylphosphino)ethyl)phenylphosphide
(3.0 g, 13 mmol) was transferred slowly by cannula to a -78
°C THF solution of (chloromethyl)diphenylphosphine (3.03 g,
13 mmol). After complete addition, the solution was warmed
to room temperature and stirred for 15 h. The solvent was
removed under vacuum, and the resultant oil was dissolved
in 40 mL of an acetone/ethanol (1:1) mixture. The solution
was concentrated to 25 mL and placed in a -10 °C freezer for
3 days. A viscous oil formed on the bottom of the flask. The
supernatant was removed by cannula, and the oil was washed
with degassed H2O (2 × 20 mL) and extracted with Et2O (30
mL). The ether was stripped under vacuum, leaving a viscous,
yellow oil (4.17 g, 87.1%). The ligand was used without further
purification.
[P d (ttp E)(NCCH3)](BF4)2 (7a (BF4)2). [Pd(NCCH3)4](BF4)2
(0.77 g, 1.7 mmol) and ttpE (0.68 g, 1.8 mmol) were dissolved
in CH3CN (20 mL). After it was stirred for 30 min, the solution
was concentrated to 10 mL and layered with Et2O (30 mL).
After a yellow solid precipitated, the supernatant was removed
by cannula. The solid (1.2 g, 97%) was dried under vacuum
at 50 °C for 1 h. The crude product can be recrystallized from
a dichloromethane/ethanol mixture containing 5% acetonitrile.
Anal. Calcd for C22H40NB2F8P3Pd: C, 38.21; H, 5.83; N, 2.03.
Found: C, 37.62; H, 5.93; N, 1.82. 1H NMR (CD3CN): δ 7.65-
7.90 (m, 5H, Ph); 2.42, 2.2, 2.1, 1.75, 1.20 (complex multiplets,
Bis(3-(d ie t h ylp h osp h in o)p r op yl)p h e n ylp h osp h in e
(ttp E; 2). A mixture of allyldiethylphosphine (4.72 g, 36
mmol), phenylphosphine (2.0 g, 18 mmol), and AIBN (100 mg)
was irradiated in a sealed Schlenk flask using a Rayonet
photoreactor with 254 and 350 nm lamps for 4.5 days. Volatile
materials were removed under vacuum at 150 °C for 2 h. The
crude, yellow oil (3.2 g, 48%) was distilled under vacuum (175
°C/0.1 mmHg). A similar procedure was used to prepare bis-
(3-(dicyclohexylphosphino)propyl)phenylphosphine (3). In this
case, the ligand could not be distilled, but its spectral param-
eters are identical with those reported using a different
synthetic route.15a
C2H5 and (CH2)3). IR (Nujol mull): νCN 2321 and 2292 cm-1
.
Addition of acetonitrile (2.5 µL) to a solution of 7a (20 mg) in
acetone-d6 resulted in a shift of the resonance assigned to
acetonitrile from 2.7 to 2.2 ppm.
[P d (t t p C)(CH 3CN)](BF 4)2 (8a (BF 4)2), [P d (ep t p E )-
(CH3CN)](BF 4)2 (9a (BF 4)2), a n d [P d (m et p E )(CH3CN)]-
(BF 4)2 (12a (BF 4)2), a n d [P d (m etp E)L]2(BF 4)4 (13(BF 4)4).
These complexes were prepared in a manner analogous to that
used for 7a . All of these complexes gave satisfactory analyses,
as shown by the following example. Anal. Calcd for [Pd-
(metpE)(CH3CN)]n(BF4)2n, C27H34NB2F8P3Pd: C, 43.48; H, 4.60;
N, 1.88. Found: C, 43.73; H, 4.82; N, 2.01. In acetonitrile-
d3, both monomeric (12a ) and dimeric (13a ) species are
observed by 31P NMR spectroscopy. Spectral data for 12a are
given in Table 1. Parameters used for simulating the 31P NMR
spectrum of 13a (Figure 1) are as follows: 98.57 ppm, PA; 9.26
ppm, PB; 62.37 ppm, PX; -337 Hz, J AB; 0.0 Hz, J AB′; 2.0 Hz,
(2-(Dieth ylp h osp h in o)eth yl)p h en ylp h osp h in e (4).
A
mixture of phenylphosphine (25 g, 0.227 mol), diethylvin-
ylphosphine (6 g, 0.052 mol), and AIBN (100 mg) was irradi-
ated for 19 h in a sealed Schlenk flask using a Rayonet
photoreactor with 254 and 350 nm bulbs. The crude mixture
was fractionally distilled to separate unreacted phenylphos-
phine. The product was isolated as a clear oil (9.5 g, 81%). 1H
NMR (toluene-d8): δ 7.34, 7.04 (m, 5H, Ph); 4.14 (dt, 1H, PH,
1J PH ) 204 Hz, J HH ) 7 Hz); 1.73 (m, 2H, CH2); 1.36 (m, 2H,
CH2); 1.11 (m, 4H, CH2CH3); 0.89 (m, 6H, CH2CH3). 31P{1H}
NMR (toluene-d8): -19.9 ppm (d, Et2P, 3J PP ) 17.1 Hz), -45.5
ppm (d, PhP).
Lith iu m (2-(Dieth ylph osph in o)eth yl)ph en ylph osph ide.
(2-(Diethylphosphino)ethyl)phenylphosphine (4.0 g, 0.018 mol)
was dissolved in hexane (30 mL) and cooled to -78 °C.
Butyllithium (7 mL, 0.02 mol) was added slowly by syringe.
After complete addition, the solution was warmed to room
temperature and stirred for 2 h. The hexane supernatant was
removed by cannula and the remaining yellow solid (4.1 g,
J
AX; 0.0 Hz, J AX′; -31.28 Hz, J BX; 16.99 Hz, J BX′; 8.50 Hz, J BB′.
If this compound is isolated from acetone solutions with no
added acetonitrile, dimer 13b or 13c, where L is acetone, is
observed by 31P NMR. Parameters for simulation of this
spectrum are as follows: 75.01 ppm, PA; 9.74 ppm, PB; 64.24
ppm, PX; 25.0 Hz, J AB; 0.0 Hz, J AB′; 5.04 Hz, J AX; -0.17 Hz,
J
AX′; 397 Hz, J BX; 16.39 Hz, J BX′. The experimental and
simulated spectra are shown in Figure 3s of the Supporting
Information.
(36) (a) Gagne´, R. R.; Koval, C. A.; Lisensky, G. C. Inorg. Chem.
1980, 19, 2854. (b) Gritzner, G.; Kuta, J . Pure Appl. Chem. 1984, 56,
462. (c) Hupp, J . T. Inorg. Chem. 1990, 29, 5010.
(37) Stephen, H., Stephen, T., Eds. Solubilities of Inorganic and
Organic Compounds; Pergamon Press: New York, 1958; Vol. 1, p 1063.
(38) Hewertson, W. German Patent 2057771 710603, 1991; Chem.
Abstr. 1971, 75, 49333y.
[P d (ep tp E)(P Et3)](BF 4)2 (9b(BF 4)2). Triethylphosphine
(122 µL, 0.82 mmol) was added via syringe to a solution of
[Pd(eptpE)(NCCH3)](BF4)2 (0.508 g, 0.75 mmol) in CH3CN (30
mL). After the reaction mixture was stirred for 30 min, 30
mL of ethanol was added by cannula, and the volume was
+
+

Bite Effects for [Pd(triphosphine)(solvent)](BF4)2
Organometallics, Vol. 15, No. 15, 1996 3373
Ta ble 4. Cr ysta l Da ta , Da ta Collection Con d ition s, a n d Solu tion Refin em en t Deta ils for
[P d (ttp E)(CH3CN)](BF 4)2 (7a (BF 4)2) a n d [P d (ttp E)(H)](BF 4) (7c(BF 4))
7a (BF4)2
7c(BF4)
empirical formula
color, habit
space group
cryst syst
a, Å
C22.125H40.25B2Cl0.25F8NP3Pd
colorless block
C2/c
monoclinic
C20H38BF4P3Pd
pale yellow wedge
P1h
triclinic
27.505(6)
8.217(2)
b, Å
9.112(2)
12.330(3)
c, Å
26.749(5)
12.794(3)
R, deg
90
94.64(3)
100.01(3)
107.29(3)
1206.6(4)
â, deg
101.57(3)
90
6568(2)
γ, deg
V, Å3
Z
d
8
2
calcd, g/cm3
1.420
1.554
fw
702.1
0.790
564.6
1.004
abs coeff, mm-1
radiation
Mo KR (λ ) 0.710 73 Å)
178
R1 ) 0.0627, wR2 ) 0.1678
R1 ) 0.1035, wR2 ) 0.1891
7277/0/341
Mo KR (λ ) 0.710 73 Å)
178
R1 ) 0.0282, wR2 ) 0.0630
R1 ) 0.0348, wR2 ) 0.0657
5649/220/292
temp, K
final R indices (I > 2σ(I)]a
R indices (all data)a
no. of data/restraints/params
no. of rflns obsd
4517
4925
a
2
R1 ) ∑||Fo| - |Fc||/∑|Fo|; wR2 ) [∑[w(Fo - Fc2)]/∑[w(Fo2)2]]1/2
.
reduced under vacuum to 30 mL. The resulting yellow powder
(0.38 g, 51%) was collected by filtration and dried under
vacuum at 50 °C for 1 h. Anal. Calcd for C25H50B2F8P4Pd:
C, 39.79; H, 6.68; P, 16.42. Found: C, 39.71; H, 6.82; P, 15.35.
A similar procedure was used to prepare [Pd(etpC)(PEt3)](BF4)2
and [Pd(ttpE)(PEt3)](BF4)2, but dichloromethane was used as
the solvent. We were unable to prepare the analogous [Pd-
(ttpC)(PEt3)](BF4)2 complex.
[P d (ttp E)(H)](BF 4) (7c(BF 4)). [Pd(ttpE)(NCCH3)(BF4)2
(0.72 g, 1.04 mmol) and NaBH4 supported on alumina (∼10%
by weight, 0.421 g, 1.12 mmol) were slurried in 25 mL of CH3-
CN. After it was stirred for 30 min, the solution was filtered,
concentrated to ∼10 mL, and layered with Et2O (30 mL). The
resulting solid (0.43 g, 73%) was collected by filtration, washed
with degassed H2O (25 mL), and dried in vacuo for 4 h.
Crystals were grown from a mixture of acetone and ethanol
at -20 °C. Anal. Calcd for C20H38BF4P3Pd: C, 42.54; H, 6.78;
P, 16.46. Found: C, 43.25; H, 7.00; P, 15.27. Similar
procedures were used to prepare [Pd(ttpC)(H)](BF4) (8c(BF4)),
[Pd(eptpE)(H)](BF4) (9c(BF4)), and [Pd(etpC)(H)](BF4)
(11c(BF4)).
(NOBA): parent ion minus BF4-, m/z 1149; parent ion minus
H(BF4)2-, m/z 1061. Complex isotopic patterns match those
calculated for a dimer with two Pd atoms. Electrospray mass
spectrometry also gave ion peaks with maxima at m/z 1149
and 1061.
X-r a y Diffr a ction Stu d ies. A crystal of [Pd(ttpE)(CH3-
CN)](BF4)2‚0.125CH2Cl2 was selected under Exxon Paratone
N oil, mounted with a small amount of silicone grease, and
placed under a cold stream of N2 (153 K) for data collection.
Similarly, crystals of [Pd(ttpE)(H)](BF4) were examined at low
temperature, approximately 210 K, under Paratone N oil. A
suitable crystal was selected and mounted under a cold stream
of N2 (178 K). Data were collected on a Siemens P3/F
autodiffractometer. Mo KR radiation (monochromatized by
diffraction off a highly oriented graphite crystal) was used in
this study. Programs in the Siemens X-ray package was used
for data collection and for structure solution and refinement.
Details of the experimental conditions are given in the
Supporting Information. Table 4 summarizes the crystal data
for [Pd(ttpE)(CH3CN)](BF4)2 and [Pd(ttpE)(H)](BF4).
[P d (m etp E)]2(BF 4)2 (14(BF 4)2). NaBH4 on alumina (200
mg) was added to a solution of [Pd(metpE)(CH3CN)](BF4)2
(0.25 g, 0.34 mmol) in acetonitrile (20 mL). The solution
immediately turned orange-brown. The reaction mixture was
stirred for 15 min and the solvent removed with a vacuum.
Dichloromethane (20 mL) was added to the residue, and the
resulting slurry was filtered to remove excess NaBH4/alumina.
Ethanol (20 mL) was added to the filtrate, and the volume of
the solution was reduced to approximately 7 mL. Upon
standing overnight, a precipitate formed. The orange-brown
solid (0.104 g, 51%) was collected by filtration and dried under
vacuum at 55 °C for 3 h. Anal. Calcd for C50H62P6Pd2B2F8:
C, 48.62; H, 5.06. Found: C, 48.52; H, 4.91. 31P NMR (aceto-
nitrile-d3): second-order spectrum with complex multiplets at
53.4 (PA), 35.9 (PM), and -8.0 ppm (PX). N values: NAM ) 5
Hz; NAX ) 14 Hz; NMX ) 420 Hz. FAB mass spectrum
Ack n ow led gm en t. This work was supported by the
United States Department of Energy, Office of Basic
Energy Sciences, Chemical Sciences Division.
Su p p or tin g In for m a tion Ava ila ble: Tables 1s-16s,
containing crystal data, data collection conditions, solution and
refinement details, atomic coordinates and equivalent isotropic
displacement parameters, bond lengths, bond angles, aniso-
tropic displacement parameters, and hydrogen atom coordi-
nates and isotropic parameters for 7a and 7c, Table 17s, giving
parameters used in extended Hu¨ckel calculations, and Figure
3s, showing experimental and calculated 31P NMR spectra for
13b or 13c (22 pages). Ordering information is given on any
current masthead page.
OM960072S
+
+