You are on page 1of 81

This article was downloaded by:

On: 20 January 2011


Access details: Access Details: Free Access
Publisher Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-
41 Mortimer Street, London W1T 3JH, UK

Catalysis Reviews
Publication details, including instructions for authors and subscription information:
http://www.informaworld.com/smpp/title~content=t713597232

Reactions of Unsaturated Ligands in Pd(II) Complexes


Eric W. Sterna
a
Division of Pullman, Inc., The M. W. Kellogg Company, Piscataway, New Jersey

To cite this Article Stern, Eric W.(1968) 'Reactions of Unsaturated Ligands in Pd(II) Complexes', Catalysis Reviews, 1: 1, 73
— 152
To link to this Article: DOI: 10.1080/01614946808064701
URL: http://dx.doi.org/10.1080/01614946808064701

PLEASE SCROLL DOWN FOR ARTICLE

Full terms and conditions of use: http://www.informaworld.com/terms-and-conditions-of-access.pdf

This article may be used for research, teaching and private study purposes. Any substantial or
systematic reproduction, re-distribution, re-selling, loan or sub-licensing, systematic supply or
distribution in any form to anyone is expressly forbidden.

The publisher does not give any warranty express or implied or make any representation that the contents
will be complete or accurate or up to date. The accuracy of any instructions, formulae and drug doses
should be independently verified with primary sources. The publisher shall not be liable for any loss,
actions, claims, proceedings, demand or costs or damages whatsoever or howsoever caused arising directly
or indirectly in connection with or arising out of the use of this material.
Reactions of Unsaturated
Ligands in Pd(ll] Complexes
ERIC W . STERN
The M . W. Kellogg Company
Division of Pullman. Inc .
Piscataway. New J e r s e y

. ...........................
Downloaded At: 12:43 20 January 2011

I INTRODUCTION 74
I1. REACTIONS O F OLEFINS ..................... 75
A . Nucleophilic Reactions: Aqueous Systems . . . . . . . . 75
8. Nucleophilic Reactions: Nonaqueous Systems ...... 89
C . Double-Bond Isomerization . . . . . . . . . . . . . . . . . . 108
.....................
D . Condensation Reactions 113
. ...............................
E T-AllylS 114
. ...........................
F Hydrogenation 117
I11. REACTIONS O F ACETYLENES . . . . . . . . . . . . . . . . . . 117
IV . REACTIONS O F AROMATICS . . . . . . . . . . . . . . . . . . . 119
.
V REACTIONS O F OLEFINICALLY UNSATURATED COM-
POUNDS CONTAINING OTHER FUNCTIONAL GROUPS . . 121
.
A Reactions of a.@-Unsaturated Compounds . . . . . . . . . 122
.
B Reactions of P. Y-Unsaturated Compounds ......... 126
.
VI REACTIONS O F CARBON MONOXIDE ............. 128
.
A Nucleophilic Reactions: Aqueous Systems ........ 128
.
B Nucleophilic Reactions: Nonaqueous Systems . . . . . . 131
.
C Reactions With Unsaturated Compounds . . . . . . . . . . 133
.
D Reactions With Aliphatics . . . . . . . . . . . . . . . . . . . 141
.
VII REACTIONS O F UNSATURATED NITROGEN
COMPOUNDS ............................. 142
.
A Nitric Oxide ............................ 142
.
B Azo Compounds .......................... 143
VIII . CONCLUSION ............................. 143
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144

73
74 ERIC W. STERN

I. INTRODUCTION

During recent years there has been growing awareness of a direct


correspondence between the catalytic chemistry of certain transition
metals and the reactions of coordination complexes of the same met-
als. For this reason a better understanding of the reactions of coor-
dinatively bound molecules has become a source of interest to cata-
lytic chemists as w e l l as to workers in inorganic and organometallic
chemistry. Of reactions discovered to date, those involving palladium
and unsaturated compounds a r e probably the most numerous and var-
ied. A description of these w i l l , therefore, s e r v e to illustrate what
has been accomplished and what remains to be accomplished in the
investigation of complex reactions.
The rapid absorption of ethylene and carbon monoxide by aqueous
solutions of PdCl,, which results in the precipitation of metallic pal-
Downloaded At: 12:43 20 January 2011

ladium and the formation of acetaldehyde and CO,, respectively, has


been known since 1894 [l]. However, the reactions involved were not
examined in detail. Credit for initiation of the current interest in
this a r e a must go to J. Smidt and his co-workers at the Consortium
for Electrochemical Industry, who in 1959 disclosed a commercially
feasible process for acetaldehyde production [ 21 by combining the
formation of acetaldehyde from ethylene, which leads to stoichiomet-
r i c reduction of Pd(I1) [Eq. (l)],with a reoxidation of the metal [Eq.
(2)] resulting in an overall process in which PdC1, functions as a cat-
alyst [Eqs. (1)-(4)].

C,H, + PdC1, + H,O - CH,CHO + Pd + 2HC1 (1)

Pd + 2CuC1, - PdCl, + 2CuC1 (2)

2CuC1 + 90, + 2HC1- 2CuC1, + H,O (3)

-
C,H, + +02 CH,CHO 14)

Commercialization of this acetaldehyde synthesis a s the Wacker


P r o c e s s has provided the impetus for an ever increasing number of
investigations of reactions of unsaturated compounds with Pd(I1) salts,
and published literature in this a r e a has grown extensively. It is the
purpose of this review not only to summarize current information
but also, wherever possible, to present a detailed analysis of reported
results.
UNSATURATED LIGANDS IN Pd(I1) COMPLEXES 75

11. REACTIONS OF OLEFINS

A. Nucleophilic Reactions : Aqueous Systems

Olefins react with water and salts of Pd(1I) to form carbonyl com-
pounds having the same carbon skeleton a s the original olefin [Eq.
( 5 ) ] . Thus the reaction which results in the conversion of ethylene to

acetaldehyde yields mixtures of ketones and aldehydes from propyl-


ene and higher terminal olefins and ketones from internal olefins.
Results obtained with various monoolefins a r e summarized in Table
1. The reactions a r e generally rapid at ordinary temperatures and
pressures and, as indicated, result in the reduction of Pd(I1) to Pd(0)
unless c a r r i e d out in the presence of a regenerating agent. Rates
and products (ketone/aldehyde ratios and by-products) depend on the
Downloaded At: 12:43 20 January 2011

structure of the olefins undergoing reaction, the temperature, the


acidity of the medium, the nature of the Pd(I1) salt employed, the na-
ture and concentration of anions, and the presence of regenerating
agents. The effect of individual variables is discussed below. It must
however, be recognized that these a r e not always readily separable,
and “facts” must therefore be interpreted with caution, particularly
since many of the effects are easily masked by side reactions.

1. Variables Affecting Product Distributions

The reaction in Eq. (5) is essentially limited to olefins having at


least one hydrogen on each of the carbon atoms in the double bond.
In a homologous series of 1-olefins, the aldehyde/ketone ratio in
products shows some inverse variation with chain length [2,3]. How-
ever, the yield of methyl ketone obtained from l-dodecene [ 51 w a s
found to be much greater than that from 1-nonene or 1-decene [Z]
under comparable conditions. Branching of an aliphatic side chain on
ethylene also appears to have little effect, a s evidenced by the forma-
tion of 3,3-dirnethyl-Z-butanone from 3,3-dimethyl- l-butene in good
yield [ 6 ] . Therefore, the conclusion that ketone formation is steri-
cally hindered by alkyl chains of increasing length and complexity [3]
does not appear to be substantiated. Aromatic substituents on the
double bond appear to favor aldehyde formation [3], but, here again,
the available data a r e contradictory and insufficient to permit firm
conclusions to be drawn.
When both hydrogens on one of the carbon atoms in ethylene a r e
substituted, the nucleophilic attack on the double bond leading to a
Downloaded At: 12:43 20 January 2011

TABLE 1
Reaction of Monoolefins with Pd(I1) in Aqueous Systems

Conditions -
Olefin Catalyst Temp.,"C Time, min Product Yield, % Ref.

Ethylene PdCly'CuC12 20 5 Acetaldehyde 85 I21


Propylene PdC12/CuC12 20 5 Acetone 90 PI
KZPdC14 70 Propionaldehyde 15.3 [31
PdCly'CuC12 70 Acetone 92-94 [41
Propionaldehyde 0.5-1.5
1-Butene PdC12/CuC12 20 10 Methyl ethyl ketone 80 [21
Pd Cl& UC12 hIethyl ethyl ketone 85-88 [41
S u t y r aldehyde 2 -4
KZPdC1, 70 Butyiddehyde 8.9 [31
1-Pantene PdClz/Cuc12 20 2 I> 2-Pentanone 81 [21
KZPdC1, 70 1-Pentanal 20 [31
1-Hexene PdC12/CUC12 30 33 2 -Hexanone 75 [ZI
KzPdC14 70 1-Hexamid 3.8 [31
3,3 -Dimethyl -1-butene 20 15 3.3 -DimethJ-l-d -butanone 66 161
1-Heptene PdC12/CuC12 50 30 2-Heptanone 65 PI
K2PdC1, 70 1-Heptanal 5 [31
1-0ctene PdC12/CuC12 50 30 2 -0ctanone 42 [21
1-Nonene PdCl,/CuC12 70 45 2-Nonanone 35 121
1-Decene PdCl,/CuCl, 70 60 2 -Decanone 34 [21
1-Dodecene PdC12/CuC12 60-70 2-Dodecanone 78-85 [51
Cyclopentene PdC12/CuC12 30 30 Cyclopentanone 61 121
Cyclohexene PdC12/CuC12 30 30 Cyc 1oh ex an o ne 65 [21
Styrene PdC12/CuC12 50 180 Acetophenone 57 121
K2PdC1, 70 Phenylacetaldehyde 75 [31
Ally1 benzene PdCl,/C uC 1 40 30 Benzyl methyl ketone 76 [a1
n-Propenylbenzene PdC12 100 60 Benzyl methyl ketone 61 [GI
2 -Phenyl-2 -butene PdC1, 100 60 2 -Phenyl-3 -butanone 14.5 161
Stilbene PdC12 Benzyl phenyl ketone 30 [61
UNSATURATED LIGANDS IN Pd(I1) COMPLEXES 77

carbonyl compound with the same skeleton is almost totally sup-


pressed. With isobutylene some of the expected 2-methylpropanal is
formed [2]. However, products of side reactions, such as hydration
to t-butyl alcohol, oxidation to acetone and 2-methylacrolein, and re-
arrangement to methyl ethyl ketone predominate. With 1,l-diphenyl-
ethylene [ 21, 1-methylcyclohexene, 2-methylstyrene, and methyl stil-
bene [S], no carbonyl products based on the parent olefin a r e formed.
However , the conversion of 2-phenyl-2-butene to 2-phenyl-3-butanone
[ 61 is normal.
In general then, increasing substitution of an olefin leads to de-
creasing product formation in the reaction pictured in Eq. (5) and
leads to increasing by-product formation. To date a systematic study
of the effect of aliphatic and aromatic substituents on product forma-
tion is lacking, and, therefore, the relative importance of steric and
inductive effects is not known.
No systematic study of temperature effects on product distribu-
tions has been carried out, and it is to be expected that such a study
Downloaded At: 12:43 20 January 2011

would be severely complicated by increasing r a t e s of by-product for-


mation. In the one case examined, the formation of propionaldehyde
relative to acetone from propylene increases with increasing temper-
ature until reaching a maximum at 70” [3]. With propylene, as with
higher olefins, higher temperatures favor the formation of n-ally1
complexes 161, which will be discussed separately.
The form in which Pd(I1) is added to reaction media has a consid-
erable effect on ketone/aldehyde ratios (Table 2) [S]. Based on the
assumption that the reaction involves an olefin complex of Pd(I1) (see
discussion below), the effect has been rationalized a s being primarily
due to ligands in the complex trans to the olefin which compete for
the same d-orbitals a s the olefins. According to this interpretation
[3], ligands which are capable of “double bonding” with the metal de-
TABLE 2
Reaction of Propylene in Aqueous Solutions
of Different Pd(II) Salts

?(, Propionaldehyde in
Solution composition, moles/liter ketone/aldehyde mixture

0.05 PdSO, 9.5


0.05 PdC1, 9.8
0.05 PdCI,, 0.05 LiCl 13.5
0.05 K2PdC1, 15.3
0.05 Li,PdBr4, 0.1 LiBr 8.4
0.05 PdF,, 0.9 H F 2.0
78 ERIC W. STERN

crease the olefin-metal double bonding and increase the overall polar-
ization of the olefin, leading to loss of selectivity with respect to ke-
tone. Anions which a r e poor ligands should have little or no effect,
as is shown in the case of fluoride. The apparent lack of difference
between sulfate, a poor ligand, and chloride might be explained on
the basis that ligand exchange with solvent water occurs to some ex-
tent during the reaction and that product distributions reflect the for-
mation of some common reactive species containing aquo or hydroxo
ligands. This exchange is depressed when excess ligand is present,
but to a greater extent with chloride than with bromide.
The effect of added acids is twofold. In experiments with HClO, it
was shown [3] that the aldehyde/ketone product ratio from propylene
varied inversely with the pH of the solution. In the case of addition
of increasing amounts of HC1, the effect was somewhat more pro-
nounced, apparently due both to decreasing pH and increasing chloride
concentration (see below). The effect of other acids is summarized
in Table 3 . A s can be seen, aldehyde production is a function of acid
Downloaded At: 12:43 20 January 2011

strength (pH effect) and the ability of the anion to complex with Pd(I1).
TABLE 3
Reaction of Propylene with Aqueous 0.05 M K,PdCl, at 70"
in the P r e s e n c e of Various Acids

Acid, m o l e d l i t e r lH,PO, lHNO, 1HzS04 1HC10, lHCl 2.5H,B03


% Propionaldehyde 5.2 12.0 13.3 15.3 20.0 3.3
in ketone/aldehyde
mixture
~

The effect of proton concentration has also been rationalized by


assuming that protonation of ligands other than olefin in the reacting
Pd(I1) complex leads to electron withdrawal from the olefin-metal
bond and therefore enhanced polarization.
Unfortunately perhaps, most of the studies of the reaction have
been carried out in the presence of regenerating agents such as CuC1,
or p-benzoquinone. Whereas the latter appears to have no effect in
t e r m s of by-product formation 171, the presence of CuC1, leads to the
formation of chlorinated products in addition to the aldehydes and
ketones normally obtained. For example, ethylene formed substantial
amounts of chlorohydrin in a solution containing large amounts of
CuCl, and KC1 181, whereas in technical installations for the produc-
tion of acetaldehyde from ethylene using CuC1, regeneration, methyl
chloride, ethyl chloride and chloroacetaldehyde a r e formed [ 91. Under
similar conditions propylene forms monochloro- and dichloroacetone ,
UNSATURATED LIGANDS IN Pd(I1) COMPLEXES 79

whereas butene yields 3,3 -dichIorobutanone and 3 -chlorobutanone


[41.
In the presence of large excesses of CuCl,, precipitation of Pd(0)
is not noted [9,10]. If, as has been suggested [ l o ] , electron transfer
between Cu(1I) and Pd(0) occurs in a complex, it may be that such
complexes, or other complexes between copper species and PdC1,
1111, are active directly in carbonyl product formation, in which case
product distributions may be affected by regenerating agents in the
same manner a s by other ligands in the system.

2. Variables Affecting Rates

Variables influencing product distribution also affect the reaction


rate. However, the overall reaction is complex, consisting of a num-
b e r of sequential steps, each of which, when they can be separated,
is affected to varying degrees. In all of the rate studies to be consid-
ered, r a t e s were measured by decreases in olefin volume o r pres-
Downloaded At: 12:43 20 January 2011

s u r e . Conveniently, this occurs in two distinct steps: a rapid initial


uptake of olefin, greater than that necessary to saturate the solution
and presumed to be indicative of complex formation; followed by a
slower absorption, presumed to be due to reaction [3,12]. Overall
rates a r e first order with respect to both olefin and palladium salt
~7,12,14,20,21,26,29].
The influence of olefin structure on rate was among the first of
the effects noted [2,13]. In their original description of the reaction
[2] Smidt and co-workers found that the rate was an inverse function
of chain length in a homologous series of 1-olefins and that terminal
olefins reacted faster than internal cis olefins, which, in turn, re-
acted faster than trans olefins. A s expected, both the fast complex
forming equilibrium and the slower complex decomposition reactions
are affected by olefin structure and these effects have now been sep-
arated qualitatively [3] as well as quantitatively j12,14] (Table 4).
The trend in the effect of olefin structure on the olefin n-complex-
forming equilibrium [Eq. (6)] w a s found to be somewhat less pro-

PdC12,- + CnH2n [CnH2nPdC1,]- f C1- (6)

nounced under conditions where reaction is inhibited [ 151 but is in


good agreement with values reported for n-complex formation with
Ag(1) [IS]. The values for k (Table 4 ) a s well as the apparent lack of
effect of olefin structure on aquation of the olefin complex [17] [Eq.
(7)] appear to substantiate the conclusion that the observed variation
80 ERIC W. STERN

TABLE 4
Effect of Olefin Structure on Rate of Reaction with Aqueous PdC1;' at 25"
~ ~~ ~~

k (M'/sec x lo5),
K. olefin complex complex Olefin
Olefin formation decomposition solubility, mM Ref.

C2H4 17.4 20.3 2.67 I121


13.1 i 0.6 ~ 5 1
c3H6 14.5 f 1.5 6.5 i 0.09 2.62 1141
1-CqHs *
11.2 1.1 3.5 i 0.4 2.59 ~ 4 1
12.4 i 0.1 I151
CiS-2-C,Hg 8.7 i 0.5 3.5 i 0.4 3.92 1141
Tr ZUl S -2 -C, Hg 4.5 f 0.5 7.5 * 0.9 3.74 1141

in reaction rate with structure is not due entirely to differences in


complex-forming ability [ 171. The finding, that the effect of olefin
Downloaded At: 12:43 20 January 2011

structure on reaction rate is small, is unexpected, particularly since


olefin oxidation by Tl(II1) in aqueous systems, which has been dis-
cussedas a related reaction, shows a large rate increase with increas-
ing substitution on the vinyl system [18]. The nature of the effect,
whether steric o r electronic, however, has not a s yet been explained.
The effect of temperature on rate has been examined over a lim-
ited range of temperatures. Results a r e summarized in Table 5. A s
TABLE 5
Effect of Temperature on Rate of Reaction
of Olefins with Aqueous PdC1,

Activation a a r a m c t e r s
AH *, Temp. range,
Olefin kcal/mole AS*,eu "C Ref.

C2H4 16.8 -6.4 10-40 1191


CZH4 19.8 -8.7 15-35 [=I
C3H6 10.4 POI
Cyclohexene 13 -3.7 7-30.3 [71

expected, temperature affects both the complex-forming equilibrium


and the rate at which the olefin complex reacts, the latter being con-
siderably more sensitive in the range studied 1121 than the former
112,151, which appears to be prone to mass transfer limitations [15].
A rate equation for reaction of ethylene under nondiffusion controlled
conditions, which includes a temperature correction on the rate con-
stant, has been developed 1211.
UNSATURATED LIGANDS IN Pd(I1) COMPLEXES 81

Most studies of the reaction of olefins with water have been c a r -


ried out in aqueous solution. However, since water is a reagent, its
concentration in nonaqueous media should affect the reaction r a t e ,
and this has been shown to be the c a s e , at least qualitatively. How-
e v e r , the reaction rate appears not to be a simple function of the wa-
ter concentration, the nature of the solvent having considerable influ-
ence on the amount of water necessary for the reaction to proceed a t
detectable rates. For example, in acetic acid the r a t e decreases with
increasing acid concentraction until, at concentrations greater than
80%, the reaction ceases almost entirely [21. A similar observation
has been made in 1,2-dimethoxyethane in which very large excesses
of water (ca. 1000 t i m e s stoichiometric on the ethylene-PdC1, com-
plex) were necessary before reaction occurred 1221. These observa-
tions a r e consistent with the relatively poor nucleophilicity of water
in displacement reactions on palladium(I1) complexes 1231 [Eq. (7)].
On the other hand, it has been observed [24] that in a hydrocarbon
medium containing acetic acid the intrinsic r a t e of the water reaction
Downloaded At: 12:43 20 January 2011

appears to be larger than that of acetate o r acetic acid, which is in-


consistent with the relative nucleophilicities of these materials but
may reflect the relative ease of solvation of the olefin-PdC1, complex
[25] or be a function of water activity which is expected to be greater
in nonpolar than in polar solvents. The effect may also depend on the
fate of the nucleophile in some step subsequent to complexing [12].
Inhibition of the reaction rate by acid has been noted in the case of
ethylene [9,12,26], propylene [14,20], and the butenes [14,17,28] but is
absent in the reaction of cyclohexene [7J. In general the inhibition
w a s found to be first o r d e r with respect to proton [ 12,14,20,26] and
affects complex decomposition only. That is, the p r e r a t e determin-
ing complex formation (initial olefin absorbtion) [Eq. (6)lis indepen-
dent of acid concentration [12]. Recently, it has been found that at
very low acid concentrations the r a t e s of reaction of ethylene, propyl-
ene, and 1-butene increase in direct proportion to acid concentra-
tion [27], a plot of the r a t e , a s measured by olefin absorption p e r
unit time against acid concentration, passing through a maximum.
The influence of ionic strength on rate is difficult to separate from
other salt e f f e d s . However, results obtained with added LiC10, [26]
and NaC10, [ 121 have been interpreted as being due purely to increased
ionic strength, since ClO; does not complex Pd(I1). These indicate a
generally linearly decreasing rate with increasing ionic strength at
p > 0.7 [26]. At lower values, data [12] indicate a r a t e increase,
reaching a maximum in the a r e a of p = 0.4. The entire effect, how-
e v e r , appears to be small [39] so that, in general, it can be stated
that ionic strength has little effect on rate.
The rate of reaction has been found to be strongly affected by
82 ERIC W. STERN

halides, inhibition decreasing in the order, I- > Br- > C1- [9]. This i s
in accord with the nucleophilic activity of these ions in displacement
reactions in planar Pd(I1) complexes [231. With chloride the inhibiting
effect on rate has been shown to be generally of second o r d e r f o r the
reaction of ethylene [12,21,29], propylene [14,20], the butenes [14],
and cyclohexene [7] in the concentration ranges studied. At low chlo-
ride ion concentrations, however, the r a t e s of reaction of ethylene,
propylene, and 1-butene are first o r d e r with respect to chloride 1271
passing through a maximum and then decreasing a s reported pre-
viously. A change in the r a t e law to second-order dependence on
PdC1, when no additional chloride is present has also been reported
[261.
Other salts which a r e not good complexing agents, such a s phos-
phates, fluorides, nitrates, and sulfates, show less effect but, in gen-
e r a l , do inhibit the reaction with increasing concentration. In some
cases, such a s with phosphates and iodides, some of the inhibiting
effect is due to formation of insoluble Pd(I1) species [S].
Downloaded At: 12:43 20 January 2011

It is to be expected that anions which are capable of complexing


with Pd(I1) will interfere both with olefin complex formation and with
the complexing of water in prerate-determining equilibria. Indeed,
the initial slow absorbtion of olefin has been shown to be subject to
first-order inhibition by chloride ion [ 121 in agreement with Eq (6).
It is also reasonable to assume that the electronic nature of ligands
will affect reaction rates, and it is precisely this type of argument
which has been advanced, to explain the accelerating effect of chloride
ions a t low concentration [26]. The greatest effect i s to be expected
from anions which are the best ligands.
Observed salt effects may also, to some extent, reflect the degree
of aggregation in reacting Pd(I1) species. That is, in the absence of
added chloride, binuclear species become important, whereas a t very
low anion concentrations the depolymerization of PdC1, may become
rate determining.
The reaction of regenerating agents such as CuC1, [9] o r p-benzo-
quinone 1261 with Pd(0) is fast relative to the reaction of olefins with
water and Pd(I1). Therefore, the presence of large excesses of such
materials is not expected to cause difficulties in rate studies if c a r e
is taken to allow for changing acid and halide concentrations [9]. How-
e v e r , the presence of Cu(I1) has been reported to increase observed
rates [ll],an effect promoted by the presence of compounds which
can form bridging groups in electron transfer reaction [lo]. This
rate enhancement may be due either to providing larger steady-state
concentrations of Pd(I1) and/or the formation of binuclear Pd(I1)-Cu(I1)
complexes which react with g r e a t e r intrinsic r a t e s than binuclear o r
mononuclear Pd(1I) species. The effect of competing olefin complex
formation by Cu(1) is expected to be small [9].
UNSATURATED LIGANDS IN Pd(I1) COMPLEXES 83

A number of isotope effects on the reaction have been reported.


Thus ethylene reacts more slowly in D,O than in H,O [9,29] with the
ratio of kH,o/kDzO= 4.05 f. 0.15 1291. The acetaldehyde produced in
this case contains no deuterium 191, indicating that only the oxygen of
the water is incorporated in the final product of reaction. A compari-
son of reaction rates of C,H, versus C,D, indicates no isotope effect,
with kC,H4/ky4= 1.07 [121.
Rate laws derived by various workers are in substantial agree-
ment [12,19,20,26,29,30]. The reaction r a t e , which i s f i r s t o r d e r in
olefin and the palladium salt (except at low chloride concentration
where it is second o r d e r in PdC1,[26]) and is inhibited by anions and
acid, is given in the majority of cases by

- d(olefin)/dt = k(PdX,)(olefin)/(X-)'(H")

The second-order rate constant is the product of the equilibrium


constant (K, Table 4 ) f o r complex formation and the rate constant
Downloaded At: 12:43 20 January 2011

for decomposition of the complex (k, Table 4) [12].


Acceleration of the reaction at low acid and chloride concentra-
tions is accommodated by the following expression a t constant olefin
partial pressure and Pd(I1) concentration (system regenerated with
p-benzoquinone) [2?]:

r = a(H+)(Cl-)/b + (H')2(Cl-)3

3 . Mechanism

A number of mechanistic schemes have evolved which seek to ex-


plain the experimental data enumerated above. Whereas these have,
understandably, been expanded to accommodate reaction of olefins
with nucleophiles in nonaqueous systems, only the aqueous case w i l l
be discussed in detail in this section.
Before proceeding to a discussion of possible mechanisms, it
would be w e l l to consider several possibilities incompatible with data.
Alcohols react with PdC1, in aqueous medium to form aldehydes and
ketones. Therefore, a possible reaction path f o r the conversion of
olefins to aldehydes and ketones could involve the hydration of olefins
to alcohols followed by oxidation of the alcohols to products accord-
ing to Eq. (8).

C,H, f H,O __
PdC12+CzH,0H - CH,CHO + Pd + 2HCl (8)

However, in the oxidation of ethylene no ethanol is found among reac-


84 ERIC W. STERN

tion products, even when the reaction is carried out on a large scale
[3]. From this it must be concluded that alcohol, if it is formed a s an
intermediate, must react with extreme rapidity. However, it has been
shown that the oxidation of olefins to aldehydes and ketones is much
faster than the corresponding reaction of alcohols [ 3,311. Therefore,
the reaction of olefins and water does not proceed via alcoholic inter-
mediates. Similar results were obtained previously in the reaction
of the ethylene-PtC1, complex [ 321 for which this mechanistic route
had been suggested [ 331.
Another possibility considered by early workers in the field [ 341
was the formation of vinyl alcohol as the actual reaction product. The
reaction was pictured as proceeding a s in Eq. (9).
Downloaded At: 12:43 20 January 2011

The conclusion that this mechanistic scheme was valid was bolstered
by the isolation of vinyl substituted products when other nucleophiles
were substituted for water in nonaqueous media [34,35]. However, if
vinyl alcohol were indeed produced in the reaction, the acetaldehyde
formed when the reaction is carried out in D,O should contain one
deuterium. That it does not 191 indicates that acetaldehyde is formed
directly and that all four hydrogens in the product come from ethyl-
ene. This result also eliminates Pd-H elimination from the inter-
mediate.
Broadly speaking, the mechanism of reaction consists of two steps:
formation of an olefin-Pd(I1) complex and decomposition of the com-
plex leading t o aldehydes and/or ketones. Separation of these steps
has already been mentioned. Efforts have also been made to obtain
direct evidence for olefin complex formation and to determine the
nature of the complexes formed. These have centered around attempts
to isolate complexes under conditions which inhibit their decomposi-
tion. For example, when ethylene was reacted with PdC1, in D,O,
complex formation was observed visually, but isolation was not
achieved [9]. Likewise, inhibition of reaction at high acid concentra-
tion also led to some complex precipitation, but, again, no isolation
PI.
On the other hand, unreactive olefins such as isobutylene, methyl-
cyclohexene, and substituted styrenes yielded isolatable complexes
from 50% acetic acid solution [S]. These materials were identical to
UNSATURATED LIGANDS IN Pd(I1) COMPLEXES 85

the dimeric olefin 8-complexes, (CnH,nPdClZ),, prepared by reaction


of olefins with benzonitrile-PdC1, [36]. The dimeric a-complexes of
unhindered olefins react rapidly with water, yielding the s a m e prod-
ucts as a r e obtained in the reaction of the olefins with water and Pd-
(11) [2,37] LEq. (10).

[RCH=CHR'PdCl,], + 2H,O -.2R-CH,-C-R' II + 4HC1 + 2Pd (10)


0
However, complexes of the n-ally1 type, (CnH,n-lPdCl), (see below),
which were formed during attempts to react hindered olefins at ele-
vated temperatures, give rise to other products [6J and are, there-
f o r e , not considered to be intermediates.
Whereas it s e e m s reasonably certain that the reacting complex is
a a-complex, its exact nature-monomer, dimer, o r diolefin PdC1,
complex-is undoubtedly dependent on reaction conditions. Mention
has already been made of the change in r a t e law at low chloride con-
Downloaded At: 12:43 20 January 2011

centration to second-order dependence on PdC1, [ 261. Conversely,


when Pd(I1) salts of noncomplexing anions, such as the sulfate, nitrate,
o r phosphate, are employed, initial olefin absorptions of g r e a t e r than
1 mole of olefin per mole of Pd(1I) were observed, the excess being
desorbed again on reaction [9]. Thus, depending on conditions, mono-
olefin-dipalladium or diolefin-monopalladium complexes may form
as well as the monoolefin-monopalladium d i m e r s or monomers.
Ligand exchange with anions o r water a s well a s bridge cleavage in
dimeric species are processes which undoubtedly occur during r e -
action and which may, under suitable circumstances, affect observed
r a t e s . The rate of solution of PdC1, itself (a chloride-bridged poly-
m e r ) is a function of the concentration of ligands capable of cleaving
Pd-C1-Pd bridges [ 221.
Based primarily on rate data, there is general agreement [12,14,
19,20,27] that olefin and hydroxide ion a r e incorporated in a palladium
complex in a series of prerate-determining equilibria and that the
rate-determining (redox) step r e s i d e s in breakdown of this complex
and includes a rearrangement from a a-complex to a o-complex. The
prerate-determining steps are summarized for reaction of ethylene
and PdCl, as follows:

[C,H,PdCl,]' + H,O G= [C,H,PdCl,(H,O)] + C1- (12)


(2 )

[C,H,PdCl,(H,O)] + H,O * [C,H,PdCl,(OH]- + H,O+


(3
86 ERIC W. STERN

Equilibrium (ll),which, a s discussed, corresponds to the ob-


served initial olefin absorption, was found to be inhibited by C1' [ 121.
Equilibrium (12) explains the additional r a t e inhibition by C1- and
equilibrium (13) explains the r a t e inhibition by proton a s well a s the
isotope effect observed when the reaction is carried out in D,O.
Since intermediate complexes have not been isolated, the stereo-
chemistry of complex (2) and, therefore, complex (3), is not known.
The question of stereochemistry has been largely ignored, perhaps
for the reason that Pd(I1) complexes, in comparison with their Pt(I1)
analogs, are labile and substitute very rapidly and can therefore be
expected to yield complexes with the stereochemistry required in
further steps. In the case where an author has specifically addressed
himself to this question [27], it is assumed that the t r a n s effect is
operative in the c a s e of substitution reactions of Pd(I1) complexes as
well a s in Pt(I1) complexes and that steps (12) and (13) lead to forma-
tionof a t r a n s complex. Two further steps [Eqs. (14) and (15)] arepostu-
lated f o r isomerization of trans-hydroxoolefin complex, (3), to cis-
Downloaded At: 12:43 20 January 2011

hydroxoolefin complex, (5), and these steps account for the additional
rate enhancement by C1' and H a s well a s the consequent additional
inhibition by these ions. The assumption of a cis-hydroxoolefin com-

trans-[ C,H,PdCl,(OH)]- + H,O [C,H,PdCl(OH),]- + C1- (14)


+ H+ (4 )

[ C,H,PdCl(OH),]- + H' + C1- =+ cis-[ C,H,PdCl,(OH)]- (15)


+ H,O (5)

plex also appears to be implicit in the reaction schemes of several


other workers [14,41]. This rearrangement has been suggested a s
being necessary for placing the olefin and OH ligands into a more
favorable position for subsequent interaction [27].
Rearrangement of the 7r-olefin complex to a u-complex, in which
the olefin has inserted into the hydroxopalladium bond, is the gener-
ally accepted next step [12,14,19,27] [Eq. (1611.

cis-[C,H,PdCl,(OH)]- - u-[ (HOCH,CH,) -PdCl,]-


(6)
(16)

The small magnitude of the effect of olefin structure on the r a t e of


reaction, as distinct from olefin complex formation [Eq. ( l l ) ] has
been interpreted as indicating that the transition state for this r e a r -
rangement has little carbonium ion character and the reaction is pic-
tured as proceeding as a concerted, nonpolar, four-center addition
[I41 [Eq. (IT)].
UNSATURATED LIGANDS IN Pd(I1) COMPLEXES 87

Cl\pd~,,CH>cH, H,O C1, ,CH,-CH,-OH


+ 2% (17)
c1’ “OH” c1 0%
Formulation of (6) as the product of a 1,2-insertion reaction in which
palladium and the hydroxo ligand have, in effect, added a c r o s s the
double bond, appears to r e s t primarily on the isolation of P-alkoxy-
alkyl palladium(I1) products from reaction of palladium chloride with
dienes in alcoholic media [42]. However, attempts to prepare ethyl
or propyl palladium compounds, either with [43] or without 1381 sta-
bilizing ligands, have been unsuccessful, as have attempts to prepare
(6) itself [38]. Formation of acetaldehyde and metallic palladium
when the attempt was made to achieve preparation by reaction of 2-
chloromercuriethanol with PdCl, in anhydrous ether has been inter-
preted a s indicating that (6) formed and decomposed [38,39]. Analo-
gies have also been drawn to and comparisons made with similar r e -
actions of Tl(III), Hg(II), and Pb(IV) in which there is good evidence
Downloaded At: 12:43 20 January 2011

for oxymetallation intermediates or in which such materials can be


isolated [12,14,39]. In these c a s e s , however, glycols a r e formed in
addition to carbonyl products and, moreover, olefin structure has a
large effect on the rate of reaction [ 181.
Product distributions with higher olefins indicate preferential at-
tachment of hydroxyl to the most substituted carbon o r the carbon
from which hydride is most readily lost [9], i.e., Markovnikoff addi-
tion. Whereas this might also indicate at least initial attachment of
palladium to the least hindered carbon and therefore constitute ad-
duced evidence for intermediate (6), the enhancement of anti-Mar-
kovnikoff addition under conditions which r e t a r d rate indicates that
product distributions a r e most likely controlled by polarization of the
olefin moiety in complexes which are p r e c u r s o r s of (6).
A number of suggestions have been made a s to the manner in
which (6) is converted to products. All these take into account the
necessity for shifting a hydride and losing the hydroxyl hydrogen a s
a proton in the final formation of acetaldehyde and Pd(0). The sim-
plest suggestion which h a s been made involves heterolytic cleavage
of the Pd-C bond in (6) [19] [Eq. (18)l.

+ CH,CHO

However, it would appear reasonable that if the oxypalladation prod-


uct, (6), formed and were prone to heterolysis, at least some hydrol-
88 ERIC W. STERN

ysis of the Pd--C bond should occur with formation of glycol a s in


Tl(II1) and Hg(I1) reactions. However, no glycol products a r e formed
in reactions of olefins involving Pd(I1) s a l t s .
This difficulty has been overcome by the suggestion that (6) re-
a r r a n g e s to a-hydroxyethyl palladium chloride prior to hydrolysis
[27,40]. A detailed view [27] of this rearrangement i s presented in
Eqs. (19)-(22). In this case, the hydride shift is accomplished by pal-

.-[
(HO-~CH,CH,)PdCl,]- * T-((HO-CH=CH,)P~HC~,]- (19)

T-[ (HO--CH=CH,)PdHCl,]- + D-[ (CH,-CH-OH)PdCl,]- (20)

p H 2
o-[(CH,-CHOH)PdCl,]- + H,O CH,-CH
‘OH
+ (PdCl,),- (21)
Downloaded At: 12:43 20 January 2011

CH,-CH
/OH,

‘OH
- CH,CHO + H,O+

ladium hydride elimination and readdition, an attractive possibility


despite the failure t o date to achieve preparation of stable palladium
hydrides. This route is consistent with the known ability of Pt(I1)
hydrides to add reversibly to olefins [44] and the apparent 6-hydrogen
elimination in palladium alkyls 138,431. However, a palladium hydride,
if formed a s an intermediate, might be expected to be prone to iso-
tope exchange, particularly in the presence of DC1[45]. The observed
lack of exchange in D,O [9] would therefore have to be explained on
the basis of an extremely short-lived hydride, a view compatible with
preparative failures.
Another suggestion for the decomposition of (6) via palladium hy-
dride formation involves heterolytic cleavage of the Pd-C bond, pro-
ton addition from the hydroxyl group to metal and hydride shift not
involving metal participation [41] [Eqs. (23) and (24)].

[HPdC1,]2- - Pd(0) + HC1 + 2C1- (24)


UNSATURATED LIGANDS IN Pd(I1) COMPLEXES 89

This route avoids the difficulty of the observed lack of isotope ex-
change since the latter is consistent with direct 1,2-hydride shifts in
organic rearrangements [46]. However, although it is known that
zerovalent platinum is capable of forming hydrides with acids, this
reaction has not been demonstrated f o r palladium [ 471. Moreover,
this route has the disadvantage of not being directly applicable to the
reactions of olefin-Pd(I1) complexes in nonaqueous media.
A view which s k i r t s the various difficulties associated with the
above mechanisms [ 121 suggests that outright palladium hydride for-
mation does not occur, but that palladium assists in the transfer of
hydride from the hydroxyl carbon to the carbon which originally was
a-bonded to palladium [Eq. (25)].

H H
H I
‘C,-C-0---H
I - CH,CHO + HC1 + Pd
II ’\ Ii
Downloaded At: 12:43 20 January 2011

-\,

Pd--- H
c1’

This suggestion is coupled with the proposal that the formation of the
oxypalladation intermediate [Eq. (16)] is the rate-determining step in
the sequence. The reasoning behind this choice is that, since no kin-
etic isotope effect was observed for the reactions of C,D,and C,H,,
C-H bond rupture, i.e., the hydride shift, cannot occur in the rate-
determining step. Step (16), as the only nonequilibrium step prior to
the hydride shift, i s therefore chosen a s being rate determining. How-
e v e r , lack of isotope effect in this instance can also be explained via
the sequence of Eqs. (19)-(22) in which the hydride shift is accom-
plished prior to the rate-determining hydrolysis in a series of r e a r -
rangement equilibria requiring little reorganization energy.

B. Nucleophilic Reactions: Nonaqueous Systems

When nucleophiles other than water react with olefins and Pd(I1)
salts, unsaturated products are formed which a r e , in many cases,
vinyl compounds. The reaction has been established for carboxylic
acids and their salts[34,35], alcohols [34,35], amines and amides [35],
cyanides [49], and carbanions [50]. Palladium(I1) is reduced to pal-
ladium(0) in these c a s e s , but, again, can be regenerated in situ by a
variety of oxidizing agents [34,51,52]. Therefore, in effect, it is pos-
sible to consider these reactions a s palladium catalyzed.
In t e r m s of product distributions the reaction of olefins with nu-
cleophiles in nonaqueous media is f a r more complex than the c o r r e -
Downloaded At: 12:43 20 January 2011

TABLE 6

E s t e r Formations by Reactions of Monoolefins with Pd(I1) in Nonaqueous Media

Olefin System Temp., "C Products, % of monoacetates Ref.

Propylenc Pd(OAC)z, HOAC 25 Isopropenyl acetate


Trans-n-propenyl acetate
1
Cis-n-propenyl acetate
Allyl acetate 90
PdCl,, NaOAc, HOAc 25 Isopropenyl acetate 44 [5 51
Trans -n-propenyl acetate 17
Cis-n-propenyl acetate 36
Allyl acetate 3
PdCl,, N%HP04, HOAc, isoctane 25 Isopropenyl acetate 64.0 ~ 4 1
n-Propenyl acetates 36.0
Pd(0Ac)z HOAC Isopropenyl acetate 98.6 [581
n-Propenyl acetates 0.5
Allyl acetate 0.9
PdClz, CUCI,, HOAC 1sop;-openyl acetate 44 [511
n -Propenyl acetates 56
1-Butene Pd(OAc)z, HOAC 25 1-Buten-2-yl acetate 80 [581
1-Buten-1 -yl acetates 9
Trans-2 -buten-1 -yl acetate 7
Cis -2-buten-1-yl acetate 2
3-Buten-1-yl acetate 2
PdCl,, NaOAc, HOAc 1-Buten-2-yl acetate 10 [561
1-Buten-1 -yl acetate 0
T r a n s -2 -buten-1-yl acetate 17.5
Cis-2-buten -1-yl acetate 1.5
Downloaded At: 12:43 20 January 2011

TABLE 6 (continued)

Olefin System Temp., "C Products, of % of monoacetates Ref.

1-Butene Trans-2 -buten-2 -yl acetate 55


Cis-2-buten-2-yl acetate 13
1-Buten-3-yl acetate 13
PdCI,, LiC1, LiOAC, Cu(OAc), HOAc, 1-Buten-2-yl acetate 2 1521
N,N -dimethylbenz amide 1-Buten-1-yl acetate 0
Trans-2-buten-1-yl acetate 40
Cis -2-buten-1-yl acetate 8
2-Buten-2-yl acetate 8
1-Buten-3-yl acetate 27
v
1-Buten-4-yl acetate 15
Pd(OAc), HOAc Trans-2-buten-1-yl acetate [571
major
3-Buten-1-yl acetate
2 -Butene Pd(OAC)z, HOAc Trans-2-buten-1-yl acetate major [571
1-Buten-3-yl acetate
Pd(OAc)z, HOAC 1-Buten-3-yl acetate 96 1581
PdCl,, LiC1, LiOAC, Cu(OAc),, HOAc, 1-Buten-2-yl acetate tl 1521
N,N-dimethylbenz amide 1-Buten-3-yl acetate 36
1-Buten-4-yl acetate 0.5
Trans-2-buten-1-yl acetate 49
Cis-8-buten-1-yl acetate 8
2-Buten-2-yl acetate 6
92 ERIC W, STERN

sponding aqueous reaction. Thus, whereas the latter is essentially


limited to aldehyde and ketone formation, the former yields monosub-
stituted vinyl and allyl compounds (Table 6) a s well a s saturated di-
substituted products. Since several stereoisomers of unsaturated
products a r e possible, the number of products increases rapidly with
increasing molecular weight of olefin, even when precautions against
such side reactions as double-bond isomerization and oligomerization
of the olefin have been taken; and, in many c a s e s , exact product anal-
yses a r e not yet available. A s in the case of the water reaction, both
the nature of products a s well as their ratios a r e sensitive to reaction
conditions such a s the temperature, olefin structure? solvent, the
nature of the Pd(I1) salt, the anion concentration in the medium, and
the presence of regenerating agents.

1. E s t e r Formation

Because of the great commercial interest in the production of


Downloaded At: 12:43 20 January 2011

vinyl acetate [Eq. ( 2 6 ) ] , ester formation has been studied more ex-
tensively than other nonaqueous nucleophilic reactions.

CH,=CH, + HOAC + PdCl, - CH,=CH-OAc + Pd + 2HC1 (26)


a. Vu/nriuhlesAffecling Product Distributions. Since few studies
have been made on a s e r i e s of olefins under the same conditions, the
effect of olefin structure on products is difficult to separate from
other effects and is therefore difficult to assess. It does appear, how-
e v e r ? in the case of l-olefins, that, as f a r a s vinyl ester formation is
concerned, secondary ester formation i s favored [24,53,56-581.
The tendency toward allyl e s t e r formation increases with increas-
ing chain length 152-54, 56-59] and with shift of the double bond away
from the terminal position [ 52,57,58,60]. In addition, 1-olefins form
primary allyl e s t e r s , whereas 2-olefins form secondary allyl ester
[58]. Cyclic monoolefins yield allylic products [61,62] or other e s t e r
products 162,631 but apparently not vinyl esters. Since no systematic
studies have included branched o r aromatically substituted olefins,
the effect of these variables is not known. On the basis of isolated
examples it might appear that such substituents increase the tenden-
cy to primary ester formation in 1-olefins. [53].
The composition of nonaqueous solutions in which the reactions
are c a r r i e d out exerts a profound influence on product distribution.
For example, the generally observed formation of 1,l-diacetates in
glacial acetic acid 134,55,64] has been shown to be a function of ace-
tic acid concentration in solvents such a s N,N-dimethylformamide,
sulfonamides , and sulfoxides , the ethylidene diacetate yield decreas-
UNSATURATED LIGANDS IN Pd(I1) COMPLEXES 93

ing from 40yG in glacial acetic acid to 1%at 25% acetic acid in di-
methyl acetamide [ 641. Solvents less effective in reducing diacetate
yields were sulfones, nitriles, ketones, and e s t e r s [64J. No diacetate
formation was observed with low concentrations of acetic acid in hy-
drocarbon [24,35] and ether 1221 solvents, in vapor phase at low ace-
tic acid concentrations relative to olefin [22], o r when t h e reaction
was c a r r i e d out with lithium acetate but without added acetic acid
(641.
Solvent effects have also been noted with respect to product dis-
tributions among unsaturated e s t e r s . For example, the amount of
cis- and trans-n-propenyl acetates produced relative to isopropenyl
acetate in the reaction of propylene with acetic acid is increased by
a factor of 2 in going from isooctane to 1,2-dimethoxyethane solvent,
under otherwise identical conditions [221. A similar effect has been
noted on adding benzonitrile or dimethyl sulfoxide to an acetic acid
medium, the amount of primary e s t e r increasing by a factor of 2 in
the former and 4 in the latter for the reaction of 1-octene [53]. Di-
Downloaded At: 12:43 20 January 2011

methyl sulfoxide has a l s o been reported to increase the amount of


ally1 acetate formed. Thus, in t h e reaction of 1-butene with palladium
acetate, the allylic acetate yield was increased from 9% in 100% gla-
cial acetic acid to 73% in 30% glacial acetic acid in dimethyl sulfoxide
581.
A complete interpretation of these effects is not possible a t this
time. However, limited rationalization is possible. For example, the
formation of 1,l-diacetates, a s shown above, is dependent on the p r e s -
ence of acetic acid, the amount formed being a function of the acetic
acid concentration. This is consistent either with formation of these
materials by the known addition of acetic acid to vinyl acetate [Eq.
(27)), or by reaction of acetic acid with an intermediate in the reaction.

CH,=CH,-OAc + HOAC -..


CH3-CH,(OAc), (27)

Evidence has been adduced supporting the latter course. When the reac-
tion of ethylenewas carriedout in CH,COOD, the ethylidene diacetate pro-
duced contained no deuterium [65], a result not possible if diacetate
formation proceeded as in Eq (27). The observation that high-boiling
products (diesters) a r e greatly increased a t short reaction times in
the reaction of 1-hexene [60] supports this view. The enhanced ability
of acid amides, sulfonamides, and sulfoxides to suppress diacetate
formation has been explained on the basis of decreased acetic acid
activity due to hydrogen bonding to solvent in these cases [64].
The influence of solvent on the position of esterification is more
difficult to explain. This effect is generally noted for donor solvents
94 ERIC W. STERN

capable of complex formation with Pd(II), i.e., functioning as ligands.


It is possible that such complexing, if it occurs, affects the manner
in which the olefin is polarized. This explanation is similar to that
advanced for the effect of acid concentration on product distribution
in the aqueous systems [3].
Another possibility is that the various unsaturated esters a r e
formed at independent rates and that changes in solvent affect these
rates differently. In the case of propylene reacting in 1,2-dimethoxy-
ethane, the rates of n-propenyl acetate and isopropenyl acetate for-
mation a r e independent, the former apparently increasing with time,
whereas the latter is reasonably constant over the same interval [22].
This observation supports the view that some product distributions
may be time dependent and indicates that apparently conflicting re-
sults reported by different workers should be compared at equal re-
action times.
A third rationalization is that, since product distributions a r e af-
fected by certain ligands (see below), those solvents which affect prod-
Downloaded At: 12:43 20 January 2011

uct distribution do so by virtue of increasing the rate of ligand ex-


change. Finally, nucleophile solvation, with a consequent decrease
in nucleophile activity, may effect differences in product distribu-
tions where possible competition between kinetic and thermodynamic
control exists. The possibility of solvent participation in the mecha-
nism w i l l be discussed later.
To date only one study, involving reaction of 1-hexene, has been
concerned with the effect of various anions and their concentrations
on product distribution [ 601. Unfortunately, unsaturated products
were not identified but were hydrogenated so that, while the position
of nucleophilic attack was found, no distinction between vinyl and
allyl products was possible. However, since double-bond isomeriza-
tion was shown to be slow relative to ester formation, the product
distributions may be accepted a s being representative of 1-hexene
reaction. The slowness of product equilibration, in the case of allyl
products, has also been demonstrated 1581. It was found that in gla-
cial acetic acid solutions containing no chloride an increase in ace-
tate concentration increased 1-substitution. Addition of chloride
(either as PdCl,, Na,PdCl,, (NH3),PdC1,, NaC1, CuCl,, or n-ally1 pal-
ladium chloride) also increased l-substitution. This effect was even
more pronounced at higher temperatures, unlike that in chloride-free
systems in which product distributions were found to be temperature
independent. Increases of both acetate and chloride relative to pal-
ladium resulted in further increases in l-substitution. No clear-cut
trend with respect t o 3-substitution was found, the latter remaining
minor in all cases in agreement with the influence of olefin structure
on allyl substitution 1581. Addition of bromide and iodide inhibited re-
UNSATURATED LIGANDS IN Pd(I1) COMPLEXES 95

action, whereas in the presence of nitrate, sulfate, or acetylaceton-


ate reaction products containing these moieties were found but not
identified.
The effect of anions has been rationalized on the basis of rapid
ligand interchange, products then being a function of the particular
palladium(I1) salt undergoing reaction. However, this explanation is
not entirely satisfactory since it merely restates the problem along
different lines.
The presence of regenerating agents has been shown to lead to
changes in product distribution among unsaturated esters a s well as
to increases in high-boiling products. With respect to the distribution
of unsaturated esters, the presence of p-benzoquinone is reported to
increase the amount of allyl product formation [ 561. In the case of
CuCl, [SO] the distribution of unsaturated products is affected only
insofar a s acetate/chloride ratios a r e changed. An investigation of
the previously discussed salt effects in copper-containin solutions
7
indicated a maximum 1-substitution of 74% at an acetate chloride
Downloaded At: 12:43 20 January 2011

ratio of 2.45, a decrease to 38% at acetate/chloride = 1.0, and an in-


crease to 78% at acetate/chloride = 0.3. The product distribution
between allyl e s t e r s from cyclohexene was not influenced by Cu(I1)
[621.
Addition of CuCl, to reaction media results in increased formation
of higher-boiling products. In the case of 1-hexene these were found
to form in increasing amounts at acetate/chloride ratios < 1.25 and
consisted of mixtures of hexene- 1 and 2-diol-di- and monoacetates.
No 1,3-disubstituted products were found in t hi s study. However, 1,3-
as w e l l a s 1,2- disubstituted products consisting of diacetates and
chloroacetates have been reported to form in reactions for olefins
larger than ethylene [39,52,63]. Comparison of disubstituted products
from 1- and 2-butene [ 631 shows that 1,2-disubstitution predominates
in 1-butene, whereas 1,2- and 1,3-disubstitutions occur to about equal
extents in the case of 2-butene. In addition, some 1,4-disubstituted
products were found. No 1,2- or 1,3-disubstituted products have been
found in the absence of copper. On the other hand, CuC1, alone does
not lead to the products in question, the presence of both PdCl, and
CuC1, being required [39]. In view of this it does not appear reason-
able to explain results with mechanistic schemes which do not include
copper 139,631 or to relate these results to the mechanism of reac-
tions in the absence of copper.
Although it is possible to view these reactions a s involving a com-
plex containing both copper and palladium, they may also be envi-
sioned a s sequential reactions involving formation of chlorinated
hydrocarbons followed by substitution, a path that appears reason-
able in view of the known ability of CuC1, to chlorinate olefins [66,67].
96 ERIC W. STERN

Few of the studies cited have dealt with the stereochemistry of


products and many of the analyses are incomplete in this respect.
Where a distinction between cis and trans isomers has been made
for vinyl products [ 551, these were formed in the thermodynamic equi-
librium mixtures [68].

b. Vaviables Affecting Rates. In general, much less is known


about the r a t e s of reaction of olefins with nucleophiles and Pd(I1)
salts in nonaqueous media than about the aqueous reactions. Again,
reaction apparently occurs in two steps, a rapid absorption of olefin
followed by much slower absorption [64]. However, equilibrium con-
stants have not been determined. Qualitative rate comparisons indi-
cate that internal olefins react more slowly than terminal olefins [60].
Originally it was reported that the ethylene-PdC1, complex,
[(C,H,)PdCl,],, would not react with glacial acetic acid in the absence
of added acetate 1341. However, it has been shown that a slow reac-
Downloaded At: 12:43 20 January 2011

tion between ethylene and acetic acid will occur without added ace-
tate in N,N-dimethylacetamide [64] o r in glacial acetic acid a t ele-
vated temperatures and pressures [ 691. Conversely, the reaction can
also be carried out with acetate alone [64]. Addition of acetate to
systems containing acetic acid increases the reaction rate [64]. A
quantitative studyofthis effect, in the reaction of PdC1, with C,H,
and HOAc 170], shows an increasing rate a s a function of acetate con-
centration at low acetate concentrations and a decreasing rate at
higher concentrations, the position of the maximum being somewhat
dependent on the PdCl, concentration in the medium but independent
of the acetic acid concentration o r the ethylene partial pressure. The
decrease in rate at higher concentrations is too large to be explained
on the basis of decreasing ethylene solubility resulting from increas-
ing acetate concentration [70].
A s expected, the rate of reaction increases with temperature (Ta-
ble 7). However, it is clear that the magnitude of the effect depends

TABLE 7
Effect of Temperature on Rate of Ester Formation

Activation parameters
AH*, Temp. range,
Olefin System kcal/mole AS*, eu "C Ref.
C2H4 PdCl,, NaOAc, HOAc, 35 < 50" [701
p-benzoquinone 8.9 > 50"
C2H4 Pd(OAC)2, HOAC, 16.6 -10.7 19-48 "711
p-benzoquinone
-- - - - - ~-- I
UNSATURATED LIGANDS IN Pd(I1) COMPLEXES 97

on the composition of the system and that the rate-determining step


or mechanism appears to change at higher temperatures. On the other
hand, since products were not analyzed, it is possible that a different
reaction takes place at higher temperatures.
Results of isotope labeling studies indicate that the reaction of
olefins with nucleophiles involves a hydride shift in nonaqueous sys-
tems a s well as in aqueous systems. When propene-2-d was reacted
with acetic acid and palladium chloride in isooctane-containing di-
sodium hydrogen phosphate, 75% of the unsaturated acetates formed
contained one deuterium [24] (Table 8). Based on the assumption that

TABLE 8
Deuterium Isotope Effects in Ester Formation from Propylene

Propene Propene-2-d
~- ~
Downloaded At: 12:43 20 January 2011

Total acetate yield, mole % based on PdC1, 15 10


Acetate, %, as isopropenyl acetate 64.0 63.6
Acetate, y,, as n-propenyl acetate 36.0 36.4
Apparent reaction velocity, acetate yield/hour 0.31 0.11
Acetate, %, containing deuterium - 75

a hydride is shifted from the carbon to which acetate ultimately be-


comes attached, both for n- and isopropenyl acetates and that hydro-
gen is subsequently lost from an adjacent carbon without appreciable
isotope effect in this step, a retention of 71% deuterium is calculated
in good agreement with the value found. Other assumptions regarding
shifts and losses give values which do not agree as well with the ob-
served retention. The finding that no deuterium was incorporated in-
toethylidene diacetate formed from ethylene and CH,COOD [65] is in
agreement with the conclusion that a hydride shift must be considered
a part of the reaction mechanism.
None of the rate studies carried out in nonaqueous media have in-
cluded a study of isotope effects. A comparison of apparent reaction
velocities (enol acetate yield per hour) indicates that the reaction of
propene is faster than that of propene-2-d by a factor of 2.8. If this
difference is real and not due to some artifact, it would appear that
the hydride shift in this case is part of the rate-determining step.
However, the ratio of n-propenyl acetates to isopropenyl acetate
formed from propene-2-d w a s only slightly greater than for unlabeled
propene. It would therefore follow that product distributions a r e de-
termined in an independent step o r steps, Such conclusions, however,
hardly seem warranted at this time and a decision as to whether the
hydride shift is, o r is not, part of the rate-determining step must
await further work.
98 ERIC W. STERN

Two studies of the r a t e of reaction of ethylene with acetic acid


and a palladium(I1) salt have been reported. However, since the sys-
tems employed differed somewhat, only limited comparison is possi-
ble. In the simpler of the two [71] the reaction of ethylene with pal-
ladium acetate was studied in glacial acetic acid containing excess
sodium acetate and p-benzoquinone. With respect t o the reagents
whose concentrations were varied, the reaction was first order in
ethylene and first order with respect t o Pd(OAc),. No effect of p-ben-
zoquinone concentration was found.
A more complete study [70] has been carried out on one of the sys-
tems originally employed to demonstrate the reaction of ethylene with
acetic acid [34]. In this case reaction order with respect to all com-
ponents in the system was investigated and the following rate expres-
sion obtained:

r = k(PdC1,)'.2(HOAc )2.2 (NaOac )O. 6-''*1'(p-benzoquinone)"P&&


Downloaded At: 12:43 20 January 2011

The second-order dependence of r a t e on acetic acid concentration


found in this case is somewhat surprising. It may well be that the re-
action actually under study was the formation of ethylidine diacetate
(the products were not analyzed and the activation energies in both
studies differed considerably; see above). It is also possible that,
since this portion of the study was carried out in p-xylene, the order
with respect to acetic acid concentration may reflect the tendency
of acetic acid to exist as a hydrogen-bonded dimer in this medium.
The variable order with respect to acetate has been mentioned
under the discussion of salt effects. It appears reasonable to conclude
that the rate inhibition at higher concentrations is analogous to the
first-order chloride inhibition of olefin complex formation in aqueous
systems, The rate increase at lower concentrations may reflect
either the activity of acetate a s a buffer o r a nucleophile, o r both.
It has also been suggested [70] that at least a portion of the observed
rate increase is due to the conversion of PdC1, to Pd(OAC), because
PdCl, is insoluble in acetic acid. This scheme, which makes Pd(OAc),
the sole reacting Pd(I1) species, is unreasonable since it would ac-
count neither for the differences in product distributions between chlo-
ride- and nonchloride-containing systems (see above) nor the differ-
ence in activation energy between two systems in which the only ob-
vious difference is the presence of PdC1,. Moreover, the rate of re-
action of the soluble PdC1,-C,H, complex is also increased by the pres-
ence of acetate.
A s can be seen from the foregoing discussion, meaningful rate
data on which mechanistic conclusions can be based a r e difficult to
obtain in view of the multiplicity of products and the sensitivity of
UNSATURATED LIGANDS IN Pd(I1) COMPLEXES 99

product distributions to conditions. In particular, caution must be


exercised in interpreting data obtained without regard to the products
formed.

c . Mechanism. Due to the difficulties in obtaining reliable rate


data and in interpreting such data as a r e available, most of the mech-
anistic schemes proposed a r e , essentially, attempts to rationalize
the formation of the various products formed. For the most p a r t , as
stated e a r l i e r , this has involved only slight modifications of the mech-
anisms proposed for reaction in aqueous systems. Again, in its broad
aspects, the reaction is thought to proceed via n-olefin complex for-
mation, incorporation of the nucleophile into the complex, rearrange-
ment from a n- to a u-complex, and decomposition of the latter to
products.
Adduced evidence indicates that, a s in the c a s e of aqueous reac-
tions, the reactive complex is a n-olefin complex. Not only can for-
mation of such materials be observed when reactions are retarded
Downloaded At: 12:43 20 January 2011

(absence of added acetate) [64], but n-olefin complexes yield the same
products observed when olefins and Pd(I1) s a l t s a r e added separately
to reaction systems [34,35,64]. On the other hand, it has been shown
that n-ally1 palladium(I1) complexes, which can form from olefins
higher than ethylene, do not react at comparable r a t e s o r yield the
s a m e products a s those formed in reactions of the parent olefins
158,621. Moreover, the observed difference in products from 1- and
2-olefins could not be explained on the basis of n-ally1 complex in-
termediacy since the same n-ally1 complex would be formed from
both isomers of the s a m e olefin [39,58]. Although n-ally1 complex
formation has been observed in s e v e r a l instances [55,60], particularly
when reactions were c a r r i e d out at elevated temperatures [SO], these
materials were shown to be by-products.
In systems containing PdC1, solubilized with excess chloride, for-
mation of the n-complex of ethylene is expected to occur by the same
equilibrium reaction a s in aqueous system [Eq. (ll)],followed by
nucleophilic attack on the complex either by acetate [Eq. (28)] o r by
acetic acid [Eq. (29)]. In the latter c a s e , a s with water, proton re-
moval must follow, and this can be assisted by either acetate or a
basic solvent [64] [Eq. (30)].

LC,H,PdCl,]- + OAc- * [C,H,PdCl,(OAc)]- + C1-


(7)

[C,H,PdCl,]- + HOAc * [C,H,PdCl,(HOAc)] + C1- (29)

[ C,H,PdCl,(HOAc)] + OAc-(B) * [ C,H,PdC12(0Ac)]-


+ HOAc(B) (30)
100 ERIC W. STERN

In the absence of evidence to the contraty, it i s not necessary to


assume that ligand exchange occurs only after olefin complexing. In
systems containing excess acetate, ligand exchange may occur p r i o r
to olefin complex formation [Eq.(31)]. Similarly, acetic acid may

[ P ~ C ~ , ( O A C ) ,+
] ~C,H,
- == [C,H,PdCl,(OAc)]- + OAc-

ionize either before or after nucleophilic attack on the complex. How-


e v e r , it is expected that the complexing of acetic acid should increase
its acidity. Analogous schemes can be written for chloride-free sys-
t e m s . On the basis of differences between chloride-containing and
chloride-free systems as well as the relative nucleophilicities of C1-
and OAc- [23], it can be assumed that chloride is present in complexes
reacting in chloride-containing systems.
Rearrangement of the s-olefin complex t o a complex in which the
palladium-nucleophile pair has added a c r o s s the double bond (oxypal-
ladation) Eq. (32) is the next step favored by most authors [38,58,62].
Downloaded At: 12:43 20 January 2011

[ C,H,PdCl,(OAc)]- =+ [AcO-CH,CH,-P~C~,]-
(8)
Again, it has not been possible to prepare such intermediates from
simple olefins. However, reaction of 2-chloromercuriethyl acetate
with palladium chloride in ether has been found to lead, at least par-
tially, to reduction of palladium and formation of vinyl acetate 1381.
The intermediacy of (8) formed by metal exchange therefore has been
assumed. A similar assumption has been made in the reaction of 2-
chloromercuriethyl acetate with PdC1, and NaOAc in glacial acetatic
acid which yielded vinyl acetate and ethylidine diacetate. However,
under the conditions employed, deoxymercuration followed by olefin-
PdC1, reaction was not precluded.
Also, as in the case of aqueous reaction, analogies have been drawn
to similar reactions of Tl(II1) and Hg(I1) [38,39], which are known to
proceed via oxymetallation. In the case of Tl(III), reaction with cyclo-
hexene in glacial acetic acid [61,72] leads to extensive ring contrac-
tion and 1,2-diacetate formation. Mercuric acetate, on the other hand,
yielded allylic acetate [61]. However, in contrast to reactions involv-
ing palladium, allylic products from Hg(OAc), reaction with linear
olefins were the same for both 1- and 2-olefins [58,60]. In moist
acetic acid, mercuric acetate reacts with cyclohexene to give 1,2-
glycols and ring-contracted products [ 611.
Stable oxypalladation products are formed by reaction of certain
cyclic diolefins, Pd(I1) s a l t s , and alcohols in the presence of a weak
base [42]. These materials have been investigated in some detail
UNSATURATED LIGANDS IN Pd(I1) COMPLEXES 101

([73] and references therein). It has been shown that the alkoxyl and
palladium moieties are located trans to each other with the palladium
endo and the alkoxyl exo in systems such as dicyclopentadiene, (9))
and that their formation is readily reversed by acid. Because of this,
it has been suggested that they are formed by nucleophilic attack of
alcohol (from the bulk solution) on the n-diolefin complex, the base
aiding in removal of the hydroxyl hydrogen simultaneously with, o r
subsequent to, addition. The same stereochemistry (addition of nu-
cleophile exo to metal) is found in reaction of (n-cyclopentadieny1)-
(n-tetraphenylcyclobutadiene) palladium bromide [ 741.
If oxypalladation does indeed occur) the stereochemistry of the
insertion might be expected to influence product distribution, partic-
ularly in reactions of cyclic olefins. The suggestion has been made
[62] that the products from reaction of cyclohexene with PdC1, and
NaOAc in glacial acetic acid, which are cyclohexen-3-yl and cyclo-
hexene-4-yl acetates but not the expected enol ester, can be explained
by assuming equatorial-axial (cis) oxypalladation followed by palla-
Downloaded At: 12:43 20 January 2011

dium-assisted hydride shift and proton loss. The nonoccurrence of


vinyl ester is explained on the basis that the necessary hydrogen (on
the carbon to which acetate has become attached) is not in position
to interact with palladium, whereas hydrogens on C-3 and C-4 are
(10). This interesting speculation requires complete stereospecificity

(9)

and nonreversibility of the adduct formation. In the case of the stable


oxypalladation adducts, the specific endo configuration for Pd is sta-
bilized by a-bonding to the remaining double bond. The cis-exo con-
figuration found for the stable norbornene oxythallation adduct [ 751
is imposed on the system sterically. Such considerations would not
appear to apply in the case of cyclohexene. A s in the oxythallation of
cyclohexene [ 721) the thermodynamically favored adduct can be ex-
pected to be diequatorial or diaxial (trans). It may, however, be
argued that the relative configuration of olefin and nucleophile in the
palladium complex prior to insertion imposes the required stereo-
chemistry on the adduct.
102 ERIC W. STERN

The question of Markovnikoff addition in oxypalladation has also


been raised. Product distributions from linear monoolefins have been
rationalized on the basis of preferred Markovnikoff addition, in a re-
versible oxypalladation step, followed by palladium hydride elimina-
tion to give either an enol o r ally1 acetate, the latter being preferred
[58]. However, this mechanism does not account for the observed
deuterium isotope effects [24,65]. Preferred Markovnikoff addition
has also been used to rationalize 1,2- and 2,3-disubstitution product
distributions in reactions of 1- and 2-butene [39]. A s pointed out,
however, such materials a r e only formed in the presence of copper,
and it is therefore questionable whether any predictions concerning
copper-free systems should be based on these observations. Finally,
as indicated earlier, the product distributions observed for any ole-
fin depend considerably on the solvent and salt composition of the
medium, Therefore, although predictions of preferred Markovnikoff
addition may be successful in some cases, such success is by no
means assured.
Downloaded At: 12:43 20 January 2011

Decomposition of (8) to products can occur in a number of ways.


Simple palladium hydride elimination, a s stated, can be neglected on
the grounds of noncompatibility with isotope effects. Heterolytic
cleavage of the palladium-carbon bond followed by rearrangement
of the carbonium ion [Eq. (33)] has again been suggested [38].

x\ Pd--CH,-CH-OAc
X

q CH\
- CH,&H-OAc
(11)
+ PdX- + X- (33)

Vinyl and 1,l-disubstituted products are then formed by proton loss


from, o r acetic acid attack on, (11). This accounts for the isotope ef-
fects and for the products normally observed, at least in the case of
ethylene. Ally1 products from higher olefins are accounted for by 1,
3-insertion followed by heterolytic cleavage and proton loss [ 55).
However, the absence of 1,2- or 1,3-disubstituted products in this
case is not explained and does not appear reasonable.
Similar objections can be made to simultaneous palladium-assisted
hydride transfer and heterolysis [ 601 and to carbonium ion rearrange-
ments with or without palladium assistance which move the positive
center down the chain prior to elimination, unless it is assumed that
such mechanisms only play a part when copper is present. Postula-
tion of a “free” carbonium ion species, (ll),has one other serious
drawback, namely, that skeletal rearrangements are not observed.
Thus, although formation of n-propenyl acetate could be rationalized
by a 1,a-methyl shift in the carbonium ion initially formed in a pref-
erential Markovnikoff addition [Eq. (34)], no rearranged products
UNSATURATED LIGANDS IN Pd(II) COMPLEXES 103

were found in the c a s e of butenes and higher olefins [24,58,60].

H&-CH-
,\ II
OAC - CH,CH,-~H-OA~
(34)
CH3
Decomposition of (8) by attack of acetate on the palladium-carbon
bond of (8) o r , in the case of higher olefins, on isomerized oxypalla-
dation products in which palladium has migrated down the chain [39]
is successful in explaining 1,2- and 1,3-products obtained from the
butenes in the presence of copper, but again fails in explaining the
absence of these products in copper-free systems.
Again, these difficulties inherent in two-carbon insertion mecha-
nisms a r e overcome by assuming a one-carbon insertion. Suggestions
concerning rearrangement of initially formed 1,Z-insertion to 1,l-in-
sertion products have already been discussed in connection with the
reaction in aqueous medium. A mechanism involving direct conver-
Downloaded At: 12:43 20 January 2011

sion of the a-olefin palladium(I1) nucleophilic complex to a 1 , l - i n s e r -


tion product in which palladium and nucleophile become bonded to the
same carbon has also been suggested for nonaqueous reactions [60].
According to this suggestion, nucleophilic attack on olefin occurs si-
multaneously with palladium-assisted hydride shift [Eq. (35)]. De-

n-[ C,H,PdCl, (0Ac)J- -.


L CH, J
U- LCH3- CH (0Ac)PdC1J-
(121
composition of (12)is then envisioned to occur by proton l o s s to form
vinyl product, o r acetate attack to form 1,l-diacetates. However,
other products must be accounted for by r e v e r s e hydride shifts which
again leads to the difficulties already discussed. Moreover, the dif-
ference in product distributions between 1- and 8-butenes a r e diffi-
cult to explain on the basis of a 1-carbon insertion, since the same
intermediate should be formed from both. Thus, as can be seen, vinyl
and 1,l-diacetates a r e best accounted for by assuming a 1,l-insertion,
whereas allyl, 1,Z- and 1,3-diacetates are more readily visualized as
being formed by 1,2-insertion. Since deuterium isotope effects have
not yet been investigated for allyl and glycol e s t e r formation, it i s
not known whether the mechanism of formation of these products
must also include a hydride shift.
104 ERIC W. STERN

Few of the suggested mechanisms have taken into account the var-
ious factors affecting product distribution. For example, since diace-
tate formation has been shown to be a function of the acetic acid con-
centration, diacetate formation must be due to acetic acid attack on a
carbonium ion or Pd-C bond in the final step. Acetate, on the other
hand, is a strong base in acetic acid and may well assist in the final
proton removal. Since it appears that the initial nucleophilic attack
can be either by acetate o r by acetic acid, the question is raised a s
to why acetate acts as a nucleophile only once in the reaction. P e r -
haps these observations constitute evidence that acetate incorpora-
tion into the complex before attack is essential.
Solvent and salt influences on product distributions a r e also largely
not a part of mechanistic suggestions. A possibility not yet considered
is that the addition of donor molecules (solvents or salts) aids in the
rearrangement of T - to a-complexes by stabilizing the latter. Involve-
ment of donor molecules in n-ally1 to u-ally1 palladium complexes
has been described [76]. Stabilization of a u-complex, if formed, may
Downloaded At: 12:43 20 January 2011

lead to additional half-life necessary for rearrangement to other spe-


cies. Stabilization of hydridic species is another possibility.
Some of the solvent and salt effects can be explained on steric
grounds if it is assumed that a one-carbon insertion mechanism is
applicable. Thus the enhancement of terminal substitution by donor
solvents, or increased acetate and chloride concentration, can be ra-
tio!ialized on the basis of increased bulk on palladium and therefore
increased tendency to form the less hindered 1,l-a-complex rather
than the 2,Z-complex.
All the foregoing reaction schemes, whether in aqueous o r non-
aqueous media, have been based on the tacit assumption that all pal-
ladium complexes in the sequence from initially formed olefin com-
plex to products are four-coordinate. The formation of five-coordi-
nate intermediates appears not to have been considered despite the
fact that axial interactions a r e known to be quite common in Pd(I1)
complexes [48], particularly in solution.
The assumption of pentacoordination in steps leading to both for-
mation of the o-bonded intermediate and its decomposition offers cer-
tain advantages. Thus in the ethylene-PdC1, complex the axis of the
ethylene molecule is perpendicular to, and centered on, the plane of
the complex 1771. Analogous structures for other olefin-Pd(I1) com-
plexes can be assumed. If all intermediates in the ensuing sequence
a r e square planar, the nucleophile cannot occupy the position trans
to the olefin prior to insertion, although t h i s is the position expected
to be preferentially substituted, since, apparently, some trans effect
is operable in t h e case of Pd(I1) complexes [78]. Therefore, cis sub-
UNSATURATED LIGANDS IN Pd(I1) COMPLEXES 105

stitution must be stipulated [27], with rotation of ethylene toward the


cis position. Whereas evidence has been obtained that ethylene ro-
tates around the coordinate bond in r-C,H,Rh(C,H,),, the rotation bar-
rier being 6 kcal [79], this has not been demonstrated for Pd(I1) com-
plexes. However, rotation is not necessary if it is assumed that the
nucleophile is complexed axially, a s attack on olefin from this posi-
tion is sterically feasible without rotation [Eq. (36)l. The tendency
for pentacoordination is increased by donor ligands [80], and the axial
positions are thought to be occupied by donor solvents in solutions of
Pd(I1) salts [25]. [The sequence pictured in Eq. (36) is for a square
pyramidal configuration. In zi trigonal bipyramidal configuration, (13),
the positions originally trans and axial become coplanar with and
equidistant from the double bond, again making olefin rotation unnec-
essary.]
Downloaded At: 12:43 20 January 2011

H L
F
-< c1
/ \
H H
As indicated in Eq. (36), nucleophilic attack can -e launchec ?om
either above o r below the plane and, therefore, can occur on either
end of the double bond, leading, in the case of propylene, to I- o r 2-
substitution. The hydride shift can be pictured a s occurring simulta-
neously to give a 1-carbon insertion product, or an oxypalladation prod-
uct can be formed. An advantage of the former view, in addition to
more readily explaining the lack of 1,2-disubstitution, is that, if donor
molecules occupy all available coordination sites, formation of the
a-bond with the terminal carbon would be less hindered, thus account-
ing for the enhancement of attack on that position in the presence of
donor solvents and excess acetate or chloride. Moreover, the in-
creasing rate of terminal substitution with time in a nonregenerated
106 ERIC W. STERN

system could be due to an increase of donor molecules and ions rel-


ative to unreacted Pd(II), again resulting in enhancement of terminal
substitution.
Breakdown of the o-bonded intermediate can also involve axial in-
teraction. It has been shown that the hydrogen on the p-carbon of the
benzene ring in trans-diodobis (dimethylphenylphosphine) palladium-
(11) interacts with an axial coordination position on palladium [ 811.
Interaction of N-hydrogen with palladium in palladous ammines [ 821
has also been noted. Thus proton removal from an incipient carbo-
ilium ion formed from the a-bonded intermediate, to give either vinyl
[Eq. (37)] or ally1 [Eq. (38)] product, can be assisted by axial coordi-
nation. The hydrogen on the allylic carbon would appear to be partic-
Downloaded At: 12:43 20 January 2011

ularly well disposed for interaction with the axial position, and this
may be the preferred decomposition route leading to such products.
Increase of the electron density on palladium by the presence of do-
nor solvents would increase both the tendency f o r pentacoordination
and proton affinity and, therefore, could explain the enhancement of
allylic substitution in the presence of such materials.
Admittedly, the above mechanism has the same drawbacks as the
previously discussed 1-carbon insertion. Further work in this a r e a
is necessary to determine which of the suggestions enumerated above
have validity and to what extent current mechanistic thinking may
need further modification.

2 . Ether Forination

Ethylene reacts with alcohols and PdC1, to yield vinyl ethers and
acetals [Eq. (39)]. Recovery of acetal from alcoholic media [34]and
of acetal and trace amounts of ether from alcoholic media containing
sodium alcoholate [38]hasbeen reported. On the other hand, either
UNSATURATED LIGANDS IN Pd(I1) COMPLEXES 107

C,H, + PdCl, + ROH -.CH,=CH-OR + CH,-CH,(OR),


+ Pd + HC1 (39)

(C,H,PdCl,), or ethylene and PdC1, react with alcohol in isooctane con-


taining a phosphate buffer to give significant amounts of vinyl ether
in addition to acetal [35]. Although it is possible to interpret these
r e s u l t s on the basis of solvent polarity [38], it is likely that the
amount of acetal formed relative to ether depends to a greater ex-
tent on the concentration of alcohol in the medium. Deuterium is not
incorporated into the acetal formed from ethylene and CH,OD [65]
and a hydride shift a s well a s alcohol attack on a reaction interme-
diate can again be assumed. Reaction of the ethyl ether of 2-bromo-
o r 2-chloromercuriethanol with PdCl, in ethyl ether yields ethyl vinyl
ether in addition to other, a s yet unidentified, products [38] and this
result has been interpreted as evidence for oxypalladation. No other
mechanistic investigations have been c a r r i e d out.
Downloaded At: 12:43 20 January 2011

3 . Vinvl Animes and Amides

A reaction of propylene with n-butylamine o r acetamide and PdC1,


in tetrahydrofuran containing disodium hydrogen phosphate is indi-
cated by palladium reduction. However, product isolation and analysis
was complicated by formation of stable complexes with unreacted
PdC1, and by the inherent instabilities of the enamines and enamides
[83] when liberated from these complexes. Hydrogenation of reaction
mixtures yielded the saturated derivatives, butylisopropylamine and
N-isopropyl acetamide, respectively [35]. Similarly, N-ethylaceta-
mide was obtained from reaction of (C,H,PdCl,), with acetamide. Iso-
lation of unsaturated products was possible when ethylene was re-
acted with carbazole or 2-pyrrolidone and PdCl, in 1,2-dimethoxy-
ethane, the expected N-vinyl products being formed in 41.5 and 23%
yields, respectively [22]. Yields obtained from the reaction of iso-
butene with diethylamine were s m a l l e r , but 1-diethylamino-2-methyl-
1-propene was isolated and identified.

4. Nitriles

Ethylene r e a c t s with Pd(CN), at elevated p r e s s u r e s and tempera-


t u r e s to give acrylonitrile 1491. Extensive by-product formation, con-
sisting of ethylene oligomers and propionitrile, was observed. Both
polarity and the nature of solvents influence the product distribution.
Polymers were formed exclusively in cyclohexane and benzene, where-
108 ERIC W. STERN

a s acrylonitrile formation predominated in benzonitrile. No reaction


was obtained with sodium tetracyanopalladite, and it appears likely
that olefin complex formation is not possible in this case. Allylic
products were found in the reaction of propylene, and the ratio of
allylic to vinylic product was somewhat solvent dependent. Again, a s
in ester formation, cyclohexene gave allylic product only.

5. Carbanions

Reaction of 1,5-cyclooctadiene palladium chloride complex with


ethyl acetoacetate or ethyl malonate, in the presence of sodium car-
bonate, leads to formation of a stable dimeric complex [50] which
appears to be related to the stable oxypalladation products obtained
f r o m cyclic dienes and alcohols. Decomposition of the product com-
plex with base leads either to cyclopropyl o r ally1 derivatives.

C. Double-Bond Isomerization
Downloaded At: 12:43 20 January 2011

Double-bond isomerization is frequently observed during nucleo-


philic reactions of higher olefins, particularly at elevated tempera-
t u r e s and extended reaction times. Under suitable conditions the re-
action is highly specific in that it i s not accompanied by skeletal iso-
merization [ 84,891, Palladium precipitation is occasionally observed
[85,86], but the extent of isomerization is not proportional to the
amount of metal deposited [86]. Deposition of metal can be avoided
if nucleophilic substances a r e absent o r if conditions a r e chosen s o
that nucleophilic reactions a r e suppressed. The thorough drying of
reagents and apparatus is also essential [84]. The reaction can be
c a r r i e d out in a neat olefin o r in a variety of nonpolar and polar sol-
vents [84-901. Olefin structure does not appear to influence product
distribution inasmuch as equilibrium mixtures of all possible double-
bond isomers a r e formed starting with any olefin in a series. An ob-
servation that chain branching adjacent to the double bond prevents
isomerization past the point of branching [87] has not been substan-
tiated [84,88,89].
The palladium(I1) necessary for catalysis can be introduced a s
PdC1, [ 84,88,90], Li,PdCl, [861, Na,PdCl, [ 87,891, (C,H,CN)2PdC1, [ 84,
85,901, (CH,CN),PdCl, [ 901, palladium salicylate [86], and the dimeric
7r-olefin-PdCl, complexes of ethylene o r cyclohexene [ 841. Simple
n-allyl-palladium(I1) complexes a r e not effective catalysts, and co-
catalysts have been found to be unnecessary [85].
No significant differences in activity have been found among cata-
lysts, except for PdC1, which is e r r a t i c in i t s behavior [84]. The be-
havior of PdCl, appears to be a matter of solubility, a s reisolation
UNSATURATED LIGANDS IN Pd(I1) COMPLEXES 109

of PdCl, f r o m HC1 solution results in a soluble material showing


high, uniform activity [88]. Diminished activity of other catalysts
under certain conditions [84] can also be related to poor solubility,
and activity enhancement, in the case of Li,PdCl,-CFSCOOH [86],
may be due to increased solubility of Pd(I1). On the other hand, the
increased activity of Li,PdCl, as a function of HC10, concentration in
aqueous medium 1861 undoubtedly s t e m s from the suppression, by acid,
of the nucleophilic reaction of olefin with water which leads to palla-
dium reduction and, therefore, decreased palladium(I1) concentration.
The apparent nondependence of catalytic activity on the particular
palladium(I1) species employed has led to the general conclusion that
the catalytically active material is formed in situ from the palladium-
(11) species and olefin. This conclusion is strengthened by the obser-
vation that the reaction i s inhibited by HC1, LiC1, and stannous chlo-
ride [86], all of which would be expected to interfere with olefin com-
plex formation, and by color changes in the course of reaction sug-
gestive of complex formation 184,891.
Downloaded At: 12:43 20 January 2011

Isomerization r a t e s for 1-olefins a r e greater than those of their


internal isomers [84,86,90]. In addition, the isomerization of 1-ole-
fins is stereoselective a s in the early stages of reaction the ratio of
cis/trans 2-olefin i s greater than equilibrium, but approaches equi-
librium values a s the reaction progresses 184,881. In the c a s e of n-
pentene isomerization in the presence of (C, H,CN)2PdC12,the 1-pen-
tene formed from cis-2-pentene reaches its equilibrium concentra-
tion much e a r l i e r than the cis-trans-2-pentene equilibrium is attained,
the maximum r a t e for cis-trans-2-pentene interconversion being
reached only after 1-pentene equilibrium is established [go]. This re-
sult appears to indicate not only that cis-2-pentene isomerizes to the
t r a n s isomer via 1-pentene but that the catalytically active species
is formed more readily from 1-pentene. The latter view is strength-
ened by the observation that cis-2-pentene isomerization has an in-
duction period, the length of which v a r i e s inversely with catalyst con-
centration and which can be eliminated by the addition of ca. 5% 1-
pentene. That the isomerization of cis to t r a n s 2-isomer is slower
than the conversion of c i s to other isomers is indicated also in the
case of methylpentene isomerization catalyzed by (C,H5CN),PdC1, in
which it was found that the cis/trans-4-methyl-2-pentene ratio drop-
ped to values well below equilibrium during 4-methyl-2-pentene con-
version L841. In contrast, in the isomerization of butenes catalyzed
by Li,PdCl,-CFsCOOH it was found that, whereas the relative rate of
conversion of cis-2-butene to 1-butene was faster than that of trans-
2-butene, the interconversion of 2-butenes w a s much faster than the
isomerization of either to 1-butene [SS]. Although it may be that this
result is due to an acid catalyzed cis-trans interconversion not re-
110 ERIC W. STERN

lated to the palladium catalyzed reaction, the question deserves fur-


ther study before conclusions can be reached,
Another aspect of the problem of relative isomerization r a t e s on
which agreement is lacking concerns the r a t e s at which the various
internal isomers a r e formed from 1-olefins. A stepwise shift to the
double bond has been observed, at least in initial stages of the reac-
tion, in the case of 4-methyl-1-pentene isomerization [84]. That is,
4-methyl-2-pentene is the chief initial product and its concentration
r i s e s until the 4-methyl- 1-pentene concentration is almost at equilib-
rium. The concentration of 4-methyl-2-pentene then falls while that
of 2-methyl-2-pentene rises at its maximum rate. However, 2-methyl-
1-pentene, which would not be expected to form at maximum rate un-
til after 2-methyl-2-pentene equilibration, rises to near equilibrium
concentration long before 2-methyl-2-pentene equilibrium is estab-
lished. This unexpected result has been rationalized by the sugges-
tion that a portion of the 2-methyl-1-pentene is formed in a direct
rearrangement of 4-methyl- 1-pentene involving cleavage at the
Downloaded At: 12:43 20 January 2011

C,-C, bond and recombination of the fragments [SS]. However, it


has not been possible to demonstrate this rearrangement. The obser-
vation, made for other olefins, that all possible internal isomers a r e
formed at comparable rates [85,92] implies that stepwise reaction
does not occur. It would appear that a thorough study in which reac-
tion conditions a r e related to rates is required.
Two mechanistic schemes have been considered in attempts t o ex-
plain palladium(I1) catalyzed olefin isomerization. These, in their
extreme forms, involve either an intramolecular hydride shift by for-
mation of a n-ally1 complex [Eq. (4011 from an initially formed n-ole-
fin complex,

\ ' (40)
,y,- ' 'PdH
I /Pf\
o r an intermolecular hydride transfer in an addition-elimination of a
palladium hydride species [Eq. (4 1)).

R-CH,CH
= CH, * RCH,-CH-CH,
Pd-H
/ I
I
'I
I
Pd ,
=+ RCH

'I
Pd-H
+ CH-CH,
(41)

Attempts to distinquish between these possibilities have chiefly in-


volved tracer experiments.
Lf the n-ally1 mechanism is applicable, an allylically deuterated
terminal olefin would be expected to isomerize to terminally deuter-
ated product. However, when l-octene-3d2 was isomerized, no ter-
UNSATURATED LIGANDS IN Pd(I1) COMPLEXES 111

minal deuteration was detected [87], products being consistent with


the shift of a single deuterium to C,. Similar results were obtained
with 1-butene-3d2 [86]. These results would appear to eliminate in-
ternal C,-C, hydride transfer via n-ally1 formation as a reasonable
mechanistic path.
The alternative scheme in which isomerization is caused by sue-
cessive additions and eliminations of palladium hydride r a i s e s the
question of hydride source. That solvents or cocatalysts a r e not, at
least, a major source of hydridic species is indicated not only by the
lack of necessity for cocatalysis but by a number of observation that
little o r no exchange between olefins and such materials a s deutero-
acetic and trifluorodeuteroacetic acids takes place [ 86-88]. This
leaves olefin as the source of hydride, and evidence that the allylic
hydrogen is involved can be adduced. Thus, when terminal olefins
which a r e deuterated on the double bond, e.g., R-CH,-CD= CHD,
a r e isomerized, no movement of deuterium i s detected [91,92], in
contrast with the movement of deuterium in allylically deuterated
Downloaded At: 12:43 20 January 2011

olefins. Moreover, vinylically deuterated terminal olefins isomerize


faster than allylically deuterated olefins [92], the latter isomerizing
much more slowly than nondeuterated olefins 186,931.
The intermolecular nature of hydride transfer was demonstrated
by an experiment in which tritiated 1-octene was isomerized in the
presence of 1-hexene, resulting in complete distribution of activity
to all olefins at equilibrium [88]. Similarly, isomerization of 1-pen-
tene in the presence of 1-heptene-3dz led to considerable deuteration
of pentenes [92]. In view of this evidence, a plausible mechanism
could involve formation of the hydridic species by Eq. (40), followed
by addition elimination via Eq. (41). However, any hydride addition-
elimination mechanism, if completely reversible, should ultimately
still lead to transfer of deuterium from allylic to terminal carbons.
Indeed, in the experiment involving 1-pentene and 1-heptene-3dz 1921,
little vinyl deuterium was found in pentenes, whereas some, but not
a l l , of the deuterium in heptenes was vinyl. Moreover, in a separate
experiment with 1-heptene-3d2 some movement of deuterium to t e r -
minal carbons was indicated.
The observed lack of terminal deuteration has been explained on
the basis of the demonstrated isotope effect on C-H bond cleavage
[86,92,94]. Since the allylically deuterated olefins which were iso-
merized were not isotopically pure, it is plausible to assume that the
initially active catalyst is a hydride. Some hydride formation involv-
ing vinylic hydrogens is also possible [85], and when cis-2-butene
was isomerized in the presence of C,D, some transfer of deuterium
to butenes was noted [86]. Thus, if it is assumed that a hydride spe-
c i e s can form, the initial isomerization of an allylically deuterated
112 ERIC W. STERN

olefin can be pictured as shown in Eq. (42)[91].


\
+
R-CD,-CH=CH,
/
PdH R-CD,-CH
f
-Pd-H
CH,

* RCD
+= RCD,-CH-CH,
I
-Pd-
T
-Pd-D
CH-CH,

I
I
I
41
* RCD-CHD-CH,
R-CD
T
-Pd-H
CD-CH,
-Pd-
I
I
Since palladium deuteride species also a r i s e in the course of this s e -
quence, the isotope effect is also assumed to control further scatter-
ing of deuterium to terminal carbon 186,911 [Eq. (43)j.
Downloaded At: 12:43 20 January 2011

+ R-CH2-CH=CH,,
-CD+ CD-CH, . R-cD,-
.CH+ CH,
-pP-D
I

I
-R-CD=CD-CH,
1I
* R-CD,-CD
RCD,-CD-CH,
-Pd-
I ?=
-Pd-H
CH, e=R-CD,-CHD-CH,
I
-Pd-
I I '(43)

Although a mechanism involving hydride formation will explainpal-


ladium(I1) catalyzed olefin isomerization and can be modified to account
for the observedisotope effects, i t i s n o t entirely satisfactory. The chief
difficulty i s that such a mechanism requires that the hydride have
considerable stability. However, palladium hydrides have not yet
been isolated, and ligands such as phosphines which stabilize other
transition metal hydrides are apparently ineffective in stabilizing pal-
ladium hydrides. Despite this, it could be argued that the particular hy-
dridic species f o r m e d h a s the required stability. It is, therefore, of in-
t e r e s t that the stable Group VIII metal hydride, trans-(Et,P),PtHCl,
catalyzed the isomerization of 1-hexene at an extremely slow rate
and then only under forcing conditions [86]. Finally, there is even
doubt that hydride addition-elimination is valid in the c a s e of catalysis
by Group VIII metal hydrides since it has been found that HCo(C0,)
and DCo(C0,) isomerized ally1 benzene at approximately the same
rates and that use of the latter led to very little product deuteration
UNSATURATED LIGANDS IN Pd(I1) COMPLEXES 113

1941. Further work is therefore needed to establish the mechanism


not only of palladium(I1) catalyzed isomerization but of the apparently
related isomerizations catalized by other Group VIII metal salts.

D. Condensation Reactions

Frequently, products of the reaction of nucleophiles with olefins


and palladium(I1) salts include oligomers and polymers of the olefin
and, under certain conditions, these materials predominate. In ad-
dition to conditions, the type of product obtained in this reaction ap-
p e a r s to be a function largely of olefin structure. Thus it was found
that decomposition of (C,H,PdCl,), at temperatures greater than 50"
primarily yielded a mixture of butenes, 10% 1-butene and 65% trans-
and 25% cis-2-butene [95]. Chlorinated products were minor in sharp
contrast to the thermal decomposition of the corresponding platinum-
(11) conplex which yielded chiefly chlorinated C,'s [96].
Solvent influenced both the r a t e and the selectivity of the dimeriza-
Downloaded At: 12:43 20 January 2011

tion, the reaction being both faster and m o r e selective in benzene than
in dioxane. Further investigation showed that the C,H,-PdC1, complex
catalyzed the formation of butene from ethylene a t 25" and atmospher-
ic p r e s s u r e . This reaction can also be c a r r i e d out in the absence of
preformed complex, but only under more severe conditions [97].
Somewhat different olefin dimerizations which have been reported in
the presence of PdC1, a r e 1,l-diphenylethylene to 1,1,4,4-tetraphenyl-
butadiene and 2-methylstyrene to 2,5-diphenyl-2,4-hexadiene[ 61.
Cyclic olefins and diolefins yield polymers. Thus norbornene and
various substituted norbornenes are polymerized in the presence of
PdC1, to low molecular weight polymers, the polymer structure de-
pending on the presence and nature of functional groups on the nor-
bornene nucleus [ 981. Norbornadiene exclusively yields 1,2-polymer
[98], whereas butadiene yields a mixture of 1,2 and trans-1,4 poly-
m e r in which the former predominates [99].
A result of particular interest has been obtained in the polymeri-
zation of butadiene in which it was shown that the polymer structure
was extremely sensitive to the nature of the palladium(I1) catalyst
employed [loo]. It was found that with PdCl, 83% of the polymer had
the 1,2-structure, the remainder being of the trans- 1,4 configuration.
The use of salts of PdC1:- and PdC12,- increased the proportion of 1,
2-polymer to 98%. Similarly, PdBr, yielded 91% 1,2-polymer. On the
other hand, PdI, yielded 85% of the t r a n s 1,4-materials and Pd(CN),
yielded 78%, whereas PdSO, and Pd(NO,), yielded noncrystalline low
molecular weight materials. The addition of various neutral ligands
also influenced the reaction. Addition of dimethyl sulfoxide and t r i -
phenyl- and tributylphosphines slowed the reaction and triethyl phos-
114 ERIC W. STERN

phite and thiocyanates completely deactivated the catalyst. Methyl


sulfide had no effect, similar to the observation that thiophene did
not effect the dimerization of ethylene to butenes [95].
Although several mechanistic suggestions have been made con-
cerning the above results, the amount of information available to date
is not sufficient to warrant making a choice among them or to offer
alternative suggestions, It does appear reasonable t o assume that the
reactions in question a r e not catalyzed by metallic palladium. How-
ever, whether the active catalyst is an olefin-palladium(I1) a-com-
plex or a material derived from the r-complex cannot be decided a t
this time. The temptation to rationalize the reaction on the basis of
the mechanism proposed for the ethylene dimerization catalyzed by
RhC1, [loll should also be resisted until more specific evidence is
accumulated for the palladium(I1) case.
Downloaded At: 12:43 20 January 2011

When olefins having at least one allylic hydrogen a r e reacted with


PdCl, a t elevated temperatures i n the absence of nucleophilic mater-
ials, o r when the corresponding olefin n-complexes are heated be-
tween 50-loo”, HC1 is evolved, and new, stable, crystalline materials
known a s n-ally1 complexes a r e formed [S] [(Eq. (44)l.

(CnH,nPdCl,), + (CnH,n-, PdCl), + 2HC1 (44)


The reversibility of this reaction is indicated by inhibition in solu-
tions containing HC1. Complex yields are a function of the degree of
branching in the vicinity of the double bond [ 1021, the stability of the
materials increasing with branching as well a s the molecular weight
of the original olefin. In general, 1- and 2-olefins with the same c a r -
bon skeleton yield the same n-ally1 complex under conditions in which
double bond isomerization prior to complex formation does not occur
[ 1031. However, steric influences over which three carbons of the n-
electron pair will be delocalized have been noted. Thus, although the
same a-ally1 complex involving carbons 1, 2 , and 3 is formed from
2-methyl- 1-butene and 2-methyl-2-butene, the complex from 2,4,4-
trimethyl- 1-pentene involves carbons 1 and 2 and the 2-methyl group,
whereas the complex from 2,4,4-trimethyl-2-penteneinvolves car-
bons 1,2, and 3 in the C, chain. Similarly, 1-methyl-1-cyclododecene
forms an “exocyclic” n-ally1 complex [ 1041.
When n-ally1 complexes a r e formed from conjugated dienes, no
HCl is evolved, but a chlorine is incorporated a to the a-ally1 system.
11051 [Eq. W I .
UNSATURATED LIGANDS IN Pd(1I) COMPLEXES 115

2CHZ=CH-CH=CH, + B(C,H,CN),PdCl, -.
CH,C1
I

CH,C1
(14)

This chlorine is readily replaced by a methoxyl group if the complex


( 14) i s treated with methanol, whereas, if the complex is prepared
in methanol solution the methoxyl group is incorporated directly [ 1061.
Allenes behave in a similar manner, chlorine becoming attached to
the central carbon atom of the n-ally1 systems [107,108]. Further re-
action, with the chloro-n-ally1 complex adding to another mole of
allene, is then possible [ 1081.
Downloaded At: 12:43 20 January 2011

As stated, the a-ally1 complexes of Pd(I1) a r e considerably more


stable than n-olefin complexes and, therefore, a r e frequently observed
by-products of reactions involving olefins and Pd(I1). When they do
react, their products differ from those normally obtained from cor-
responding n-complexes. Thus reaction of n-ally1 complexes with
water leads to formation of qp.-unsaturated carbonyl compounds and
olefins in addition t o palladium [6,102] [Eq. (46)].

(C3H,PdC1), + H,O -.CH,=CH-CHO + CH,=CH--CH,


+ 2Pd + 2HC1 (46)

In some c a s e s the carbonyl group i s found a! to the original n-ally1


system as in the c a s e of the n-ally1 complex formed from 2-methyl-l-
pentene or 2-methyl-2-pentene which yields mesityl oxide on reaction
with water rather than the expected 2-methyl- 1-pentene-3-one [ 1021.
Acetate and alcoholate anions react with the a-ally1 complex
(C3H,PdC1), in ethanol/DMSO to yield allyl acetate and allyl ether,
respectively, in low yields [log]. However, in the case of l-ethyl-2-
methyl- n-allylpalladium(1I) chloride only 2-methyl- 1-pentene and 2-
methyl-2-pentene are formed on reaction with methanolic KOH [ 1101.
Good yields of allylic products are reported from reaction of n-ally1
palladium(I1) chloride with the carbanions derived from ethyl malonate
and ethyl acetoacetate [ 1091.
The bridged dimeric a-allylic complexes a r e cleaved to mono-
meric a-allylic species by reagents such a s pyridine, thallous acetyl-
acetonate, and cyclopentadienyl sodium [ 106,111], ethylene diamine
and a,@-dipyridyl [ 1121, and quaternary ammonium and phosphonium
116 ERIC W. STERN

halides [ 1131. Reagents such a s triphenylphosphine, triphenylarsine,


and dimethyl sulfoxide convert binuclear a-allylic complexes t o mono-
nuclear 0-allylic compounds via mononuclear 7i-allylic species [ 76,
114,1151.
Thermal decomposition of 7i-allyl-Pd(II) complexes, in which the
organic moiety has four or more carbons, in the absence of solvent
under vacuum, leads to formation of conjugated dienes [116]. With
lower olefins, or when diene formation is impossible a s from a-
methallyl palladium(I1) chloride, allylic chlorides are formed [ 1171.
Since workers in the area of a-ally1 palladium(I1) complexes have
been concerned chiefly with the preparation and structure of these
materials rather than with detailed investigations of their reactions,
little can be said concerning mechanism a t this time. In the c a s e
of thermal decomposition to dienes, hydride removal from a carbon (Y
to the n-ally1 system [Eq. (47)] has been suggested [116].
Downloaded At: 12:43 20 January 2011

R’
,
CHR‘
CH/’
H I
-9 \fl
Pd-bcH
CH
--+
(47)
I
R R
R’-CH=CH-CH-CHR + Pd(0)

A hydride displacement mechanism involving hydride removal by


a second mole of PdC1, [Eq. (48)] has been used to rationalize aqueous
nucleophilic reactions [ 1021.

Rx, 80
C
.CH, I
+ C-CH, + Pd
I’
CH
‘R’ I
R’

However, in view of the known a,o-allyl interconversions with nucleo-


philic materials, such a rearrangement followed by heterolysis of
the carbon-palladium bond would also explain the products formed.
UNSATURATED LIGANDS IN Pd(I1) COMPLEXES 117

F. Hydrogenation

Very few examples of homogeneous hydrogenations of unsaturated


organic compounds catalyzed by Pd(I1) species exist. This is some-
what surprising in view of the fact that palladium hydride species are
frequently invoked intermediates in mechanisms proposed to explain
reactions involving a variety of Pd(I1) complexes and in view of the
demonstrated ability of various other Groups VIII metal hydrides to
function as hydrogenation catalysts [ 118- 1221. In the s e v e r a l examples
available, the palladium catalysts are active under very mild conditions.
For example, palladium(I1) chloride has been used to catalyze the hy-
drogenation of dicyclopentadiene in dimethylformamide solution at
low temperature and atmospheric p r e s s u r e [ 1231. The homogeneity
of the catalyst was demonstrated by the failure of thiophene t o poison
the reaction, addition of this material actually accelerating the r e -
duction rate. Similarly, catalysis by PdC1, of the reduction of benzo-
nitrile, nitrosobenzene, and nitrobenzene by KBH, has been reported
Downloaded At: 12:43 20 January 2011

[124]. A more interesting case is the homogeneous room tem-


perature, atmospheric p r e s s u r e hydrogenation of 1-hexene in hy-
drocarbon solution catalyzed by [ (n-C,H,),P],PdCl, in this instance,
a triisobutylaluminum cocatalyst was employed L 1251. The func-
tion of the latter is presumed to be alkylation of the palladium
complex, the catalytically active hydride then being formed by hydro-
genolysis of or hydride elimination from the alkyl. In view of this
r e s u l t , it may well be that the full potential of homogeneous hydrogen-
ation catalysts f r o m palladium(I1) has not yet been realized.

111. REACTIONS O F ACETYLENES

Most acetylenic compounds condense rapidly in the presence of


palladium(I1) compounds, but the exact nature of products has not
been fully elucidated. It appears, however, that, as with olefin con-
densation, the acetylene structure e x e r t s some influence on the type
of product obtained.
The preferred reaction in the case of disubstituted acetylenes ap-
p e a r s to be cyclooligomerization. Thus tolane is converted to hexa-
phenylbenzene either by alcoholic PdC1, [ 126,1271 or a benzene solu-
tion of (C,H,CN),PdCl, [ 1287. Similarly, phenylmethyl and phenylethyl
acetylenes are condensed to 1,2,4-trimethyl-3,5,6-triphenylbenzene
and the corresponding triethyl compound, respectively, by PdC1, in
methanol [129]. In addition to the cyclotrimer, other oligomers a r e
formed which complex with varying amounts of PdC1,. Thus a tetra-
phenyl cyclobutadiene complex (15) is formed by reaction of tolane
with (C,H,CN),PdCl, in 3 :1 chloroform/ethanol[130]. The number of
PdC1, units separating the cyclobutadiene groups is variable [l28].
118 ERIC W. STERN

In alcoholic medium an alkoxyl radical is incorporated into the cy-


clodiiner [ 126,1271 and the formation of this material is postulated
as occuring via a linear dimeric o-ally1 complex [ 1271 (16). Other

? ? rp
Downloaded At: 12:43 20 January 2011

as yet incompletely identified complexes of acetylene d i m e r s and


t e t r a m e r s are also found [126,129].
Cocatalysis by metal hydrides has been reported for the polymer-
ization of monosubstituted acetylenes [ 1311. Thus 1-heptyne was con-
densed to dimer, t r i m e r , t e t r a m e r , and polymeric material by the
action of [(n-C,H,),P],PdCl, or PdC1, and NaBH, in ethanol solution.
The overall yield of oligomer was better when the phosphine complex
was employed.
Acetylene is condensed to mono- o r dicarbonyl-containing dimers
by aqueous PdC1, [132]. The products in this case f o r m complexes
with PdC1,.
Two types of palladium acetylene complexes have been reported.
One of these is an acetylide, [Pd(CN),(C-CH),]2- which can be iso-
lated as a barium salt from liquid ammonia [133]. The other, formed
by reaction of hexafluoro-2-butyne o r methyl acetylynedicarboxylate
with tetrakis-triphenyl phosphine palladium(0) in methylene chloride
has the structure (17) [134].

On the basis of infrared data, the multiple bond in this material is


intermediate between a double and triple bond and the formal oxida-
UNSATURATED LIGANDS IN Pd(I1) COMPLEXES 119

tion state of the metal is therefore intermediate between zero and +2.
Both complexes a r e unstable. In the absence of air the barium salt
of the acetylide shows somewhat greater stability. However, it is
readily decomposed by water, dilute acids, and most organic solvents
with the formation of unidentified colored organic materials. Reduc-
tion of the Pd(I1) acetylide to a Pd(0) acetylide by potassium in liquid
ammonia has been c a r r i e d out [133]. The resulting compounds are
l e s s stable and decompose explosively.
It is likely that complexes of either type represent a primary stage
in the formation of oligomers and polymers from acetylenes. How-
e v e r , data available at this time are insufficient to warrant detailed
speculation. In the case of cocatalysis by metal hydrides, t r a c e r ex-
periments indicate product deuteration and palladium hydride addition
to acetylene as a primary step has therefore been proposed [131].

IV. REACTIONS O F AROMATICS


Downloaded At: 12:43 20 January 2011

Although the published literature contains only two reports of re-


actions of aromatic compounds with Pd(I1) salts, the information pre-
sented permits a rather detailed description. Products and r a t e s a r e
functions both of aromatic structure and the composition of reaction
media.
The primary reaction is coupling. Thus biphenyl is formed from
benzene in acetic acid solutions containing PdC1, and sodium acetate
at 90” [ 1351 [Eq. (49)].
2C,H, + PdC1, + 2NaOAc -C,H,-C,H, + Pd + 2NaC1
+ 2HOAc (49)
As indicated, palladium(I1) is reduced to metal. Similar results are
reported for the reaction of Pd(OAc), in benzene solution [136].
With monosubstituted benzenes mixtures of biphenyls were ob-
tained in which p a r a and meta isomers predominated. The relatively
low amounts of ortho isomers formed [135,136] is presumably due to
s t e r i c hindrance. Influence on isomer distribution by the nature of
substituent groups was also found [ 1351. Thus the p,p’-isomer was
the chief product in the coupling of diphenyl ether, whereas methyl
benzoate yielded only meta substitution products. As expected, the
effect of s t e r i c hindrance on product distribution becomes more pro-
nounced in di- and trisubstitued benzenes [ 1351. Thus p-di-t-butyl-
benzene, p-diisopropylbenzene and mesitylene failed to react while
o-xylene yielded 3,4,3’ ,4‘-tetramethylbiphenyl and 3,4,2’,3’-tetra-
methylbiphenyl in a ratio of 2.7:l. With m-diisopropylbenzene the
chief product was 3,5,3’ ,5’-tetraisopropylbiphenyl.
120 ERIC W. STERN

Whereas biphenyls a r e the only products reported when PdC1,-


NaOAc is used [ 1351, phenyl acetate was formed in addition to biphenyl
when benzene reacted with Pd(OAc), in acetic acid [136]. The ratio of
phenyl acetate to biphenyl was decreased by the presence of perchlor-
ic acid and increased by lithium acetate. Similarly, the reaction of
toluene in acetic acid with Pd(OAc), mainly leads to benzyl acetate.
However, formation of the acetate was completely suppressed by the
addition of 12% perchloric acid.
The reason for this difference in results is not readily apparent.
Acidity, acetate concentration, and the composition of the Pd(1I) spe-
c i e s involved may all be factors. Biphenyl formation i s not obtained
in the presence of PdBr, or PdI, [ 1351, whereas PdSO, caused the
dimerization of benzene in 20% H,SO, at 110" [ 1361. In the same sys-
tem toluene was oxidized to benzoic acid.
A number of factors have been found which affect the r a t e of ben-
zene condensation to biphenyl. In acetic acid containing PdCl,, no re-
action is reported to take place in the absence of sodium acetate [135].
Downloaded At: 12:43 20 January 2011

In the reaction of Pd(OAc), in acetic acid, the r a t e is accelerated both


by LiOAc and HC10, [136]. With monosubstituted benzenes, the r a t e
is somewhat increased by electron-donating substituent groups and
decreased by electron-withdrawing groups [ 1351.
The formation of biphenyl from benzene in the PdC1,-NaOAc-ace-
tic acid system was found to be first order with respect to benzene
and PdC1, concentrations and independent of the sodium acetate con-
centration [ 1351 giving the rate expression:

d(biphenyl)/dt = k,(PdCl,)(C,H,)

The above results indicate that benzene and PdC1, combine in a


rate-determining step having the characteristics of an aromatic elec-
trophilic substitution reaction [ 1351. That is, product distribution as
w e l l a s rate a r e determined in the initial complex formation [Eq. (50)l.

Irreversibility of this step is indicated by the lack of exchange of un-


reacted benzene with D,O during dimerization with PdSO, [ 1361. Sub-
sequent steps a r e fast. By analogy to reactions of Tl(OAc), and
Hg(OAc), with benzene, it has been suggested that a o-phenyl palla-
dium(I1) compound is formed [Eq. (51)] which decomposes to biphenyl
and palladium [ 1361. This view is supported by the formation of bi-
phenyl and palladium by reaction of B(OH),(C,H,) with Pd(OAc),. De-
composition by a radical mechanism is eliminated by the absence of
UNSATURATED LIGANDS IN Pd(I1) COMPLEXES 12 1

o+
'\ NaOAc
PdC1,
- @ P d C l +NaCl + HOAc (51)

phenol formation when the reaction was c a r r i e d out in the presence


of oxygen [ 1351.
Conversion of the o-phenyl species can be pictured a s occuring
either by heterolysis of the Pd-C bond with subsequent attack of the
phenonium ion on a second benzene o r by interaction with acetate.
Cleavage of the polarized Pd-C bond could a l s o occur by attack of
either benzene or acetate and palladium displacement in a concerted
step. A somewhat different suggestion [135] for decomposition of the
initial intermediate involves attack by acetate to form a n-cyclohex-
adienyl- PdCl complex which dimerizes and then decomposes to prod-
ucts [Eq. (52)].
Downloaded At: 12:43 20 January 2011

+ OAC--*
PdC1,
H

V. REACTIONS OF OLEFINICALLY UNSATURATED


COMPOUNDS CONTAINING OTHER FUNCTIONAL GROUPS

The behavior of olefinic compounds having functional groups in


addition to the double bond in systems containing Pd(I1) s a l t s is
largely dependent on the nature of the functional group and its posi-
tion relative to the double bond, That is, in some c a s e s , reactions
primarily involve the functional group, whereas in others the com-
pounds react as olefins. The latter appears always to be t r u e when
the functional group is farther removed than the ally1 position. In
most instances reactions have not yet been investigated in detail and
suggestions concerning reaction paths a r e therefare of a speculative
nature.
122 ERIC W. STERN

A. Reactions of a,p-Unsaturated Compounds

1. Aqueous Systems

As in the case of olefins, products a r e aldehydes and ketones.


However, whereas the reaction of olefins with water and Pd(I1) salts
leads to reduction of Pd(II), the reaction of some vinyl compounds
does not.
a. Reactions Inuolving P d ( I I ) Reduction

In this category a r e reactions of dienes, unsaturated aldehydes,


acids, amides, nitriles, and nitro compounds. The primary
reaction is nucleophilic attack on the double bond. Products obtained
indicate that the functional group substituent directs the attack to the
p -carbon, that is, the carbon from which hydride is most readily
displaced. Illustrative examples a r e given in Table 9.
Downloaded At: 12:43 20 January 2011

TABLE 9
Reactions of a,P-Unsaturated Compounds With Aqueous Pd(I1)

Reagent Product

Butadiene Crotonaldehyde
Crotonaldehyde Triacetylbenzene
Acrylic acid Acetaldehyde
Crotonic acid Acetone
a-methyl acrylamide Propionaldehyde
Crotonamide Acetone
Acrylonitrile P -Cyanoacetaldehyde
Nitroethylene Nitroacetaldehyde
1-Nitro-1-propene Nitroacetone

Observed products a r e frequently the result of secondary reac-


tions. Thus, in the case of the a,P-unsaturated acids and amides, the
primarily formed p-keto compounds a r e unstable under reaction con-
ditions and decarboxylate . The product obtained from crotonaldehyde
is the result of acid catalyzed cyclization of P-ketobutyraldehyde [2].
The formation of crotonaldehyde from butadiene can be rationalized
by assuming a similar reaction path, i.e., formation of 3-butenal fol-
lowed by either acid o r PdC1, catalyzed isomerization. This path
would require formation of a n-olefin complex from butadiene involv-
ing one of the double bonds. Although such a material has been pre-
pared by ligand exchange of n-(n-pentene-PdCl,), with butadiene at
-40" [ 1381, it is extremely unstable a t temperatures above -20" and
UNSATURATED LIGANDS IN Pd(I1) COMPLEXES 123

at room temperature changes almost quantitatively to the n-ally1


compound, (PdC1-C,HeC1),, normally obtained directly from butadiene
under ambient conditions [105]. Under the conditions of the reaction
(80"), it therefore appears improbable that the n-complex would be
present in significant amounts. However, since hydrolysis of the
m o r e stable n-ally1 complex would not be expected to yield croton-
aldehyde, it must be assumed the n-ally1 complex is in equilibrium
with n-complex, and that, while conditions strongly favor the f o r m e r ,
reaction of the n-complex continually shifts the equilibrium toward
the latter.

b . Reactions Not Involving Pd(I1) Reduction


(1). Saponification of vinyl esters. Vinyl e s t e r s react in aqueous
solutions of Pd(I1) salts to yield the corresponding aldehydes or ke-
tones. For example, acetaldehyde is formed from vinyl acetate and
acetone from isopropenyl acetate 1121. This reaction is catalyzed by
Downloaded At: 12:43 20 January 2011

palladium chloride? and, although vinyl e s t e r hydrolysis is readily


catalyzed by mineral acids, the reaction in the presence of PdC1, is
faster than it is in a solution having the same acidity but without
PdCl, [91.
(2). Hydrolysis of vinyl halides. Vinyl halides hydrolyze readily
in the presence of PdC1,. Vinyl chloride and bromide yield acetal-
dehyde, whereas chloro- and bromo-1-olefins as w e l l a s 2-halo- 1-
olefins mainly yield methylketones [2,139]. For example, the chief
product from 1-chloro-1-propene [27], as well a s from 2-chloro-1-
propene [Z], was acetone with only minor amounts of propionaldehyde
formed from the f o r m e r . Alkyl substitution on C-2 in 1-halo-1-ole-
fins, e.g., 1-chloro-2-methyl-l-propene, leads to an increase in the
amount of aldehyde formed. However, the chief product in this c a s e
was a-methylacrolein, which results from hydrolysis of the isobutenyl
n- ally1 complex and palladium reduction. Phenyl substitution does
not have the same effect; the chief product from P-bromostyrene i s
acetophenone. In addition, phenyl loss occurs to some extent to pro-
duce acetaldehyde [2,139].

2 . Nonaqueous Systems

a. Nucleophilic Displacement of Vinyl Halides. Displacement of


vinylic halides by nucleophiles in nonaqueous, nonpolar, and polar
media is catalyzed by salts of Pd(I1). The reaction has been demon-
strated with respect to acids [140] [Eq. (53)] and their salts [141], al-
cohols, and amines [140J. That Pd(I1) is a catalyst is indicated by the
fact that reduction of Pd(I1) does not occur and that product yields f a r
124 ERIC W. STERN

R-CH=CH,X + HY CH,-Y
PdC1,+~-~~= + HX (53)
in excess of the amount of Pd(I1) employed can be obtained in the ab-
sence of oxygen and regenerating agents [22]. An interesting facet of
the reaction in nonaqueous media is that 1-halo- 1-olefins yield l-sub-
stitution products only. Thus the only product of reaction of l-bromo-
1-propene with acetic acid in di(2-methoxyethyl) ether was l-acetoxy-
1-propene [22]. This result contrasts sharply with the PdCl, catalyzed
hydrolysis of the same compound which leads to acetone formation
with greater than 90% selectivity [ 1391. As in aqueous systems, 2-
halo-1-olefins form 2-substituted products [22].
The high reactivity of the monohaloolefins in the presence of Pd-
(11) salts is somewhat surprising in view of the demonstrated inert-
ness of such materials in displacement reactions [142]. A priori as-
sumption of 7i-complex formation between the haloolefin and Pd(I1)
appears justified in view of the similarity in kinetic pattern to olefin
Downloaded At: 12:43 20 January 2011

reactions observed in aqueous solution. However, since vinyl halide-


PdC1, complexes have not, as yet, been prepared, the nature of bond-
ing in such materials is not known. Therefore, whether halide activa-
tion occurs in the complex itself or is the result of subsequent r e a r -
rangements of the complex remains a matter for speculation. For
example, it is possible to envision halide activation as being due to
an increase of electron density in the vinyl halide by back donation
of electrons from metal to antibonding orbitals, o r by interaction of
the halide with an axial position in the complex, o r a combination of
these, that is, stabilization of a structure, (18),which could be repre-
sented as an incipient vinyl carbonium ion leading either to an Sn, or
concerted displacement mechanism.

On the other hand, product formation via a o-bonded oxypalladation


adduct is also readily envisioned [Eq. (54)].
CH
<
UNSATURATED LIGANDS IN Pd(I1) COMPLEXES 125

Here the double bond is reestablished by elimination of the Pd-C1


moiety in a r e v e r s a l of the oxypalladation step.
An examination of the sterochemistry of the reaction should per-
mit some choice among mechanistic possibilities. Thus a purely
elimination-addition type of mechanism would be expected to lead
to thermodynamic control of products, whereas an addition-elimin-
ation mechanism may lead to stereospecificity [ 1431. Preliminary
results of an investigation of the reaction of trans- 1-chloro- l-pro-
pene indicate substantial retention of configuration in the l-acetoxy-
1-propene product [22]. This finding can be rationalized on the
basis of a mechanism involving either cis oxypalladation followed
by t r a n s elimination o r trans addition followed by c i s elimination.
The latter path can be pictured as an attack by nucleophile from
solution on complexed vinyl chloride followed by r e v e r s e oxypalla-
dation, possibly with axial interaction. [Eq.(5511.
Downloaded At: 12:43 20 January 2011

H\h/
" II O n

/c, CH, "H


CH, H
(55)
H

-c-/o'I
I1
,l

6. Tyansestekficntion. A closely related reaction of vinyl com-


pounds catalyzed by Pd(I1) i s the transesterification of vinyl e s t e r s
[9] [Eq. (56)]. This reaction has been demonstrated with a wide va-

RCOOCH=CH, + R'COOH P d o _ R C O O H + R'COOCH= CH,


(56)
riety of mono and dibasic carboxyllic acids. It can also be catalyzed
by rhodium and platinum compounds, as w e l l as by Hg(I1)-BF, [ 1441,
and would, therefore, appear to be a general reaction of vinyl e s t e r s
catalyzed by Lewis acids.

c. Vinyl Estev Decomfiosition. Palladium chloride catalyzes the


cleavage of vinyl acetate to acetaldehyde and acetic anhydride in
acetic acid-acetate media [64,145] [Eq. (57)]. No reaction w a s ob-
served when PdC1, was replaced by HgCl,, ZnCl,, or HClL1451, al-

CH,-CHOAc + HOAc OAc-


P d C 1 2 c ~ ~ , +~ (A~o),o
~ ~ (57)
126 ERIC W. STERN

though a similar reaction of isopropenyl acetate catalyzed by anhy-


drous HC1 o r HBr in ether o r by sulfuric acid has been reported in
which acetone and acetic anhydride were formed [ 1441.
The palladium chloride catalyzed reaction requires the presence
of both acetic acid and the acetate buffer. It has been shown that in
acetic acid negligible conversions a r e obtained in the absence of
acetate [ 1451, whereas in dimethyl acetamide conversion to aldehyde
is a function of acetic acid concentration at constant acetate con-
centration [64].
The following reaction scheme [ Eqs. (58)- (SO)] has been suggested
[145]. This is in substantial agreement with the proposed intermed-
iacy of acetyl chloride in the HCl catalyzed decomposition of isopro-
penyl acetate [ 1461.

CH,COOCH= CH, + PdC1, * CH,COOCH


6+ t CH, (58)
Downloaded At: 12:43 20 January 2011

PdClF

CH, + OAC- - (CH,CO),O 7


+ -0-CH CH,
PdC1,
(59)

-0-CH CH, + HOAc - HOCH=CH, + PdCl, + OAc- (60)


t
PdC1, \CH,CHO

B. Reactions of /3' , y- Unsaturated Compounds

1. Aqueous Systems

Generally, products obtained from the hydrolysis of allyl com-


pounds in Pd(I1)-containing solutions a r e those obtained by hydroly-
sis of the corresponding n-allyl-palladium(I1) complexes. For ex-
ample, hydrolysis of allyl alcohol yields acrolein, and the product
from allyl chloride o r bromide is methyl glyoxal, obtained by fur-
ther reaction of acrolein [2]. Similarly, l-chloro-2-methyl-2-pro-
pene i s converted to 2-methacrolein. Dihaloolefins, such as 1,3-
and 2,3-dibromo- 1-propene, yield methyl glyoxal, apparently by a
combination of vinyl- ai,d allyl-type reactions.
Ally1 amines a r e deaminated. In contrast to halides and alcohols,
however, products are chiefly saturated mono and dicarbonyl com-
pounds. For example, in addition to small amounts of methyl gly-
oxal, allyl amine yields either propionaldehyde o r acetaldehyde a s
the major product, depending on the ratio of amine to PdCl, em-
UNSATURATED LIGANDS IN Pd(I1) COMPLEXES 127

ployed, whereas 3-amino-2-methyl- 1-propene f o r m s methyl ethyl


ketone and diacetyl. The origin of these products remains unex-
p lained.

2. Nonaqueous Systems

With allylic halides and allylic alcohols, elimination of water


from reaction media results in the formation of n-ally1 complexes
[117,147]. Olefins are also evolved and, with allyl alcohol, an un-
saturated compound C,H,,O,, produced.
The disproportionation of allyl alcohol to olefin and the dimeric
product [Eq. (Sl)] is catalyzed by PdC1, [148].

3C3H,0H PdC1z~C,H,+ C,H,oO, + H,O (61)

Formation of n-ally1 complex inactivates the catalyst. Yields a r e


Downloaded At: 12:43 20 January 2011

decreased by the addition of solvents o r water and improved by in-


creasing the temperature or the amount of allyl alcohol relative to
PdC1, and also by the presence of HC1. The C,H,,O, product has
been identified as a mixture consisting of 60- 85% 4-methylenetetra-
hydrofurfuryl alcohol, (19),and 40- 15% 4-methyl-2,5-dihydrofurfuryl
alcohol. Since the higher yields of the methylene isomer were ob-
tained under milder conditions, it is thought to be the primary r e -
action product.
The following mechanism has been suggested to explain the prod-
ucts obtained [Eq. (62)]:
\ /
128 ERIC W. STERN

This route consists of two c i s insertions followed by hydride r e -


moval. The metal hydride r e a c t s with allyl alcohol forming water
and an allyl carbanion which can either add a proton to form pro-
pylene o r react with PdC1, to form n-ally1 complex.

VI. REACTIONS OF CARBON MONOXIDE

A. Nucleophilic Reactions : Aqueous Systems

Oxidation of CO to CO, in aqueous solutions of PdC1, [Eq. (63)]


has been known for a considerable time [ 11.

PdCl, + CO + H,O - Pd + CO, + 2HCl (63)

However, despite early recognition that a palladium carbonyl com-


plex was involved and that both CO and water were bound to metal
Downloaded At: 12:43 20 January 2011

prior to reacting [ 1491, the reaction was not investigated in detail


until recently.
It has been shown that the r a t e of reaction is a function of the
particular palladium salt employed [150]. In a s e r i e s of complexes
of the type (PdL,),-, the r a t e decreased a s a function of L in the
following manner:

Br > C1> NO, > I > SCN NH, > Thio > CN-
This r a t e dependence is in r e v e r s e order of the stability of the com-
plex s a l t s except for bromide and chloride. The greater activity of
the bromide was interpreted a s being due to its greater rate of aqua-
tion.
Reaction of Pd(I1) salts with CO is inhibited by increased common
anion concentrations 1150- 1531, presumably, a s in olefin reactions,
by interference with complex forming equilibria. The reaction is
a l s o inhibited by large concentrations of acid, although, in systems
regenerated by p-benzoquinone, small amounts of acid have an ac-
celerating effect [ 1541. Addition of bases such as acetate and par-
ticularly KOH greatly accelerate the rate [ 1511.
The effectiveness of a number of inorganic and organic oxidizing
agents in maintaining palladium in the +2 oxidation state during CO
oxidation has been demonstrated. These a r e salts of Cu(I1) and
Fe(III), CrCl,, K.$r207, H202 11521, and quinones [154,155]. Com-
parison of reaction r a t e s with and without p-benzoquinone showed
enhancement by a factor g r e a t e r than two in the presence of quinone
11561. Metal precipitation was not observed when the quinone was
UNSATURATED LIGANDS IN Pd(I1) COMPLEXES 129

used [ 155,1561, and this has been interpreted as indicating that the
palladium atom acts as a bridge for electron transfer between CO
and the oxidizing agent [156]. This assumption is supported, in part,
by a calculated entropy of activation, S* = -22.4 eu, which i s in sub-
stantial agreement with reactions in which electron transfer between
metal ions is facilitated by ligand bridging. In the presence of
quinones whose redox potential is below that of the Pd(I1)-Pd(0)
system, palladium precipitates or remains in solution as complexed
Pd(0) [155]. In these systems electron transfer is incomplete, and
the quinone aids the transport of electrons to metal only [ 1561.
Formation of carbonyl complexes a s intermediates in the reac-
tion of aqueous Pd(I1) salt solutions with CO is indicated by the re-
action of stable carbonyls with water. Thus the monocarbonyl com-
plex, [Pd(CO)Cl,], [ 1571, reacts with water to f o r m CO, and Pd 11491
[Eq. W ) ] .

[Pd(CO)Cl,], + 2H,O - Pd + CO, + 2HC1 (64)


Downloaded At: 12:43 20 January 2011

In addition, evidence has been obtained from manometric measure-


ments for the formation of mono and dicarbonyl palladium(II) chlorides
as primary stages in the reaction of aqueous solutions of H,PdCl,
with CO, and for formation of unstable Pd(0) carbonyls a s end prod-
ucts [153]. A s with olefin reactions, the r a t e s of complex formation
and complex decomposition can be separated [158].
Reaction of the monocarbonyl, [ Pd(CO)Cl,], , with low concentra-
tions of water in alcoholic or ether media leads to evolution of CO,
and HC1 122,1591. Here, however, metallic palladium is not precip-
itated. Instead, an insoluble, violet palladium carbonyl chloride is
formed. This material has been tentatively identified as a polymeric
Pd(1) complex, (PdCOC1)n [ 1591. Although further reaction of water
with this complex again leads to palladium(O), it i s not likely that
the Pd(1) carbonyl represents an intermediate stage in the reaction
in aqueous media; rather, it appears more probable that its forma-
tion is the result of reaction of Pd(0) with unreacted Pd(I1) car-
bony1 LEq. (65) and (66)l.

nPd(CO),Cl, + nPd - (PdCOCl), (66)

Reaction of (20) with water would result information of two moles of


Pd(0) p e r mole of CO, formed [Eq. (67)] and reestablishment of the
overall stoichiometry expressed in Eq. (64).
130 ERIC W. STERN

2 PdCOCl + H,O - 2Pd + CO, +- 2HC1 (67)

No detailed mechanism for the oxidation of CO in aqueous Pd(I1)


solutions has been presented. Based on a kinetic investigation of the
reaction of (PdBrJ2- with CO in aqueous dioxane containing naphtho-
quinone, it has been suggested that a complex is formed by reaction
of CO with an aquated palladium bromide, [PdBr,(H,O)]-, followed
by further complexing with naphthoquinone. The latter complex de-
composes with regeneration of the aquated Pd(I1) salt [158].
A m o r e detailed picture is possible based on available informa-
tion on the water-CO-Pd(I1) reaction and analogy with reactions of
other metal carbonyls. For reaction of (PdBrJ2- without the p r e s -
ence of a regenerating agent, the following sequence is proposed
[Eqs. (68)-(72)]:

(PdBr,),- + H,O * [PdBr,(H,O)]- + Br-


Downloaded At: 12:43 20 January 2011

(21)

[PdBr,(OH)]2- + CO [Pd(CO)Br,(OH)]- + Br- (70)

Br
[Pd(CO)Br,(OH)]- + CO -
Br
+ H,O - PdCO + CO, + 2Br- + H,O+ (72)

The conversion of the aquo species, (21),to a hydroxo species, (22),


prior to reaction with CO, as w e l l a s the subsequent insertion, is
written by analogy to mechanisms suggested for reaction of aqueous
&(I) [l60]. Carbonyl insertion into an aquo-palladium bond, as
suggested for reaction of Hg(I1) [ 1611, would appear to be incompat-
ible with acid inhibition. Carbonyl formation prior to aquation, and,
a s in the case of olefin reactions, axial interactions are also pos-
sible. By analogy to carbonyl insertion in CH,Mn(CO), 11621, the
insertion step is assumed to involve hydroxo migration to the c a r -
bony1 group.
UNSATURATED LIGANDS IN Pd(1I) COMPLEXES 131

B. Nucleophilic Reactions: Nonaqueous Systems

Reactions of CO with nucleophilic reagents in nonaqueous media


have not a s yet been studied extensively. This appears due, in p a r t ,
to the fact that primary reaction products a r e frequently masked by
secondary reactions or by side reactions. For example, while
ethanol reacts with carbon monoxide in the presence of a PdC1,-
CuC1, catalyst to form ethyl carbonate, the expected product of nu-
cleophilic attack, yields a r e low and ether and ethyl chloride a r e
formed in amounts equal to o r slightly greater than carbonate [ 1631.
A better illustration of nonaqueous nucleophilic reaction is found
in the formation of isocyanates by reaction of primary aliphatic o r
aromatic amines with CO and PdC1, [164] [Eq. (73)l.

RNH, + CO + PdCl, - RNCO + Pd + 2HC1 (73)


This reaction proceeds readily in ether solvents a t low tempera-
Downloaded At: 12:43 20 January 2011

t u r e s and atmospheric p r e s s u r e and is accompanied by the precipi-


tation of metallic palladium. At high temperatures and pressures,
u r e a s , oxamides, and hydrogen a r e formed in a reaction apparently
catalyzed by palladium metal [ 1651.
With respect to the formation of isocyanates, a number of obser-
vations have been made which provide information concerning pos-
sible mechanistic pathways [164]. Reaction is observed when CO is
added to Pd(RNH,),Cl,, or amine is added to LPd(CO)Cl,],, indicating
complexing of both reacting species prior to reaction. The possibil-
ity that the observed products a r e the result of in situ phosgene for-
mation by reaction of PdC1, and CO followed by reaction of phosgene
and amine appears highly unlikely in view of the failure to observe
Pd(I1) reduction with CO in the absence of amine at temperatures
ranging from 40-200". The observed formation of phosgene during
the conversion of (C6H,CN)PdC1, to Pd,(CO),Cl in chloroform under
high CO p r e s s u r e [ 1661 i s likely due to a reaction involving chloro-
form. Phosgene is formed from CO and carbon tetrachloride in the
presence of ZnC1, [22].
Another possibility involves the intermediate formation of for-
mamides, as in the reaction of amines and CO catalyzed by a num-
ber of transition metal carbonyls and s a l t s [167,168] [Eq. (74)].

RNH, + CO - RNHCHO (741


RNHCHO + PdCl, - RNCO + P d + 2HC1 (75)

Dehydrogenation of the formamide by PdC1, [Eq. (75)] would lead to


the observed products. However, the presence of isocyanates, as
132 ERIC W. STERN

well as PdC1, reduction, shortly after combination of reagents, and


the failure to detect formamides a t any time, in conjuction with the
fact that contact of formanilide and CO for an extended period (un-
d e r conditions leading to rapid phenyl isocyanate formation from
aniline and CO) resulted in neither PdCl, reduction nor isocyanate
formation, leads to the conclusion that formamides a r e not isocy-
anate p r e c u r s o r s in this reaction.
A likely reaction path involves formation of a palladium carbonyl
ammine followed by carbonyl insertion. This has been shown to be
the primary step in the reaction of amines with Fe(CO), [169,170].
The sequence pictured here [Eq. (76)] is schematic and does not
imply actual knowledge of steric relationships in reacting complexes.
Axial interactions a r e a possibility.

O H

-
-H+ \ /
I1
C-N-R
I
Downloaded At: 12:43 20 January 2011

/pd\
co

Breakdown of the insertion product can be envisioned as occurring


either by proton loss, (23), perhaps with palladium assistance, o r
by hydride transfer, (24), to palladium [Eq. (77)].
R
R I
I
I H-N,
H-N
\C,,=
\I
-.RNCO + HPd(CO), !, f,C=O

2
Pd
'co
0
c
Pd + H'
' 'coPd (77)

(24)
(23)
Support for at least transitory formation of a palladium hydride,
possibly stabilized by carbonyl, has been obtained by the appear-
ance of a band a t 2130 cm-l in the infrared spectra of reaction mix-
t u r e s , which grew in intensity a t about the same rate a s the isocy-
anate band at 2270 cm'l [22]. Although this band could be due to a
carbonyl complex, the observed frequency is well above that of the
known Pd carbonyl chlorides and that of a complex containing amine
and CO [ 1571.
An interesting application of the reaction of CO with nucleophiles
and Pd(I1) salts is the preparation of n-ally1 complexes. It was found
UNSATURATED LIGANDS IN Pd(I1) COMPLEXES 133

that improved yields of n-ally1 palladium chloride could be obtained


by reaction of Na,PdCl, and allyl chloride with CO in aqueous meth-
anol [171]. Based on the formation of a mole of CO, for each mole
of n-ally1 complex, the following reaction scheme [Eq. (78)] has been
suggested [ 1721 :

CH,-

Cl, ,,CH
pd ~IYCH+ CO, + HCI
TH,

Similarly, reaction of amines with CO, Na,PdCl,, and allyl chloride


Downloaded At: 12:43 20 January 2011

in benzene leads to formation of isocyanates and n-ally1 complex,


whereas the same reaction in methanol yields methyl carbamates
and the complex [ 1731.

C. Reactions with Unsaturated ComDounds

1. Olefins

Olefin-PdC1, complexes react with carbon monoxide under mod-


erately high (40- 100 atm) p r e s s u r e s to yield P-chloro acylchlorides
[174,175] [Eq.(79)].

(C,H,PdCl,), + 2CO - 2CH,CH,COCl + 2Pd


I (79)

The reaction can also be c a r r i e d out with the olefins and PdC1, added
separately to the reaction mixtures. It is rapid a t room tempera-
t u r e in benzene and, when starting with the preformed complexes,
is accompanied by extensive decomposition. Product yields, which
a r e generally not high, decrease with increasing chain length of n-
1-olefin, a r e lower for internal olefins than those from the c o r r e s -
ponding terminal olefins, and a r e further reduced when the olefin
is t r a n s . Branching on the double bond of a terminal olefin has
about the same effect on yield as moving the double bond to the 2-
position, The yield from isobutene is about the same as from cis-
2-butene. Cyclic olefins yield mixtures of c i s and trans-p-chloro-
acyl chlorides in which the latter predominates. However, large
134 ERIC W. STERN

amounts of hydrochlorinated products also f o r m . Double-bond iso-


merization in the olefin prior to reaction of CO with 1-hexene and
1-heptene results in mixtures of chloroacyl chlorides. Dichloro-
olefins r e a c t as olefins a t low temperatures. Thus, P,p-dichloro-
propionyl chloride i s formed from vinyl chloride and P,cw-dichloro-
butanoyl chloride from ally1 chloride. Yields, however, a r e much
lower than those from the corresponding olefins.
In all cases involving terminal olefins, carbonylation occurs ex-
clusively a t the terminal carbon. The reaction therefore differs
from nucleophilic reactions of olefins in which terminal olefins
f o r m mixtures of 1- and 2-substituted products. Because of this,
the reaction has been viewed a s involving attack by chloride on ole-
fin followed by carbonylation rather than vice v e r s a L175,176]. A
reaction scheme [Eq. (80)] that involves formation of a complex con-
taining both coordinated CO and olefin, (25), (a material of this type
has been prepared in the case of platinum [176]), followed by dis-
placement of chloride with formation of a o-2-chloroalkyl palladium
Downloaded At: 12:43 20 January 2011

complex, (26), has been proposed [175J. Acylformationfollowsbyin-


sertion of carbon monoxide, and the acyl complex then decomposes
to products.

R-CH=CH, + PdC1, + CO -.

H
I
Pd(C0)X + R-C-CH,COCl
I
c1
Steric factors in the transition state a r e presumed to direct the
chloride to attack at C-2 in all c a s e s . Since acyl halides a r e con-
UNSATURATED LIGANDS IN Pd(I1) COMPLEXES 135

vertible to olefins, CO; and HC1 in a reaction catalyzed by palladium


[ 1771, the reaction sequence can be viewed a s being completely re-
versible. conversion of olefins to esters by reaction with CO in al-
coholic HCl has also been reported [ 1781, but the catalyst in this
case is presumed to be palladium metal.

2. Allylic Compounds

The carbonylation of allylic compounds at low temperatures has


already been mentioned. At higher temperatures (generally > 100")
and at CO p r e s s u r e s of the order of 100-200 atm, allyl compounds
r e a c t to yield products in which CO has inserted between the allyl
radical and functional group. For example allyl chloride r e a c t s
with CO, without solvent, in the presence of a PdC1, catalyst to yield
3-butenoyl chloride 11791 and minor amounts of crotonyl chloride
1Eq. (W3.
Downloaded At: 12:43 20 January 2011

CH,=CH-CH,Cl + CO PdC1z* CH,= CH-CH,COCl


+ CH,-CH=CH-COCl (81)

In ethanol the major product is the corresponding ethyl e s t e r . Sub-


stantial amounts of ethyl crotonate and ethyl isobutyrate a r e formed
as by-products. However, these can be eliminated by controlling the
HC1 concentration of the reaction medium [ 1801. n-Ally1 palladium
chloride also yields 3-butenoyl chloride [ 1811, and the reaction has
been demonstrated with other allylic chlorides and n-ally1 complexes
derived from these materials [179,182]. Reaction of alkyl substi-
tuted n-ally1 complexes indicates preferential carbonylation a t the
least hindered carbon [ 1821,
The intermediacy of a n-ally1 complex is demonstrated by the
fact that the same product, 3-pentenoyl chloride, is formed from
both 1-chloro-3-butene and 3-chloro-1-butene [ 1791. Basically, any
compound capable of n-ally1 complex formation will participate in
the reaction [180]. Allyl alcohol r e a c t s in ethanol to yield ethyl-3-
butenoate, whereas the products in the absence of solvent a r e allyl-
3-butenoate and diallyl ether. Again, as expected, both 2-buten- 1-01
and 1-buten-3-01 give the s a m e product in ethanol, ethyl-3-penteno-
a t e . Allyl e s t e r s are converted to mixed anhydrides [Eq. (82)].

CH,=CH-CH,OAc + CO PdC1z+ CH,= CH-CH,C-OC-CH,


Il II
0 0
(82)
136 ERIC W. STERN

Diallyl ether f o r m s 3-butenoic anhydride, presumably via allyl-3-


butenoate, the primary carbonylation product. The n-ally1 com-
plexes derived f r o m esters of o,p- or p,y-unsaturated carboxylic
acids yield unsaturated e s t e r s . Thus diethylglutaconate is produced
from the n-ally1 complex formed from either ethyl-2 or 3-butenoate
[183].
The reaction does not take place at low CO p r e s s u r e s [182J. It
is also specific for materials capable of forming n-ally1 complexes.
Thus isopropyl, t-butyl, benzyl, and n-propenyl chlorides, a s well
as a product of low temperature carbonylation of n-ally1 complexes,
P-chloropropionylchloride,failed to react [ 1791.
Catalysis by salts of palladium(I1) other than chloride as well a s
by palladium metal has been reported, but best results a r e obtained
with either PdC1, or n-ally1 palladium chloride [179,180].
No specific mechanism has been suggested other than that n-ally1
complexes a r e the reactive species in the reaction. The suggestion
of probable mechanistic similarity to the carboxylation of allyl
Downloaded At: 12:43 20 January 2011

halides by nickel carbonyl [180], for which formation of an acyl di-


carbonyl chloride has been proposed [ 1841, appears reasonable, but,
again, no details are given. In view of the known ability of n-ally1
complexes to rearrange to o-ally1 species in the presence of nucleo-
philic materials, it s e e m s reasonable to expect that such a r e a r -
rangement may precede insertion. The reaction can therefore pic-
tured as follows [Eq.(83)]:

CH,= CH-CH,-Cl + PdC1, -.CHA<- H , P,d, p,,Pd, ,GH?,


--,I,CH
CH, c1 CH,
I

CH,=CH-CH,-COCl + Pd(CO),

Since the reaction is catalytic, the metal o r Pd(0) carbonyl must re-
act with more allyl compound to regenerate the active species as
has been demonstrated for Pd and allyl bromide [ 1851.
UNSATURATED LIGANDS IN Pd(I1) COMPLEXES 137

3. Diolefins

a. Conjugated Diolefins. As described earlier, butadiene forms


a chloromethyl- n-ally1 complex with palladium chloride in which
the chlorine in the chloromethyl group is very labile. It might there-
fore be expected that the complex would react with CO to yield 5-
chloro-3-pentenoyl chloride, and, possibly, the diacyl chloride a s
well. In fact, however, the product distribution is considerably more
complicated. At room temperature, in benzene, reaction with CO
results in formation of 1,4-dichloro-2-butene and the expected 5-
chloro-3-pentenoyl chloride in 3:2 ratio [186]. In ethanol the major
product is ethyl 3-pentenoate. The dichloro product is also formed,
however, in decreased yield.
Products from the ethyoxy-substitued isoprene complex are even
more numerous. Whereas at room temperature, in benzene, the
main product, after conversion of acid chlorides to ethyl esters,
was the normal carbonylation product of the n-ally1 complex, ethyl-
Downloaded At: 12:43 20 January 2011

5-ethoxy- 3- methyl- 3-pentenoate; ethyl-4-methyl- 3-pentenoate ,


ethyl- 3-methyl- 3- hexenedioate , and ethyl-4-ethoxy- 4-methyl valer-
ate were also formed. At 100" the main product was 3-methyl-3-
hexenedioate with Y,y-dimethylbutyrolactone and ethyl-4-ethoxy-4-
methyl valerate a s by-products. In ethanol, at room temperature,
products consisted of ethyl- 5- ethoxy-3 -methyl- 3-pentenoate plus
minor amounts of the lactone, whereas a t 100" the latter was the
main product formed together with l e s s e r amounts of the products
formed in benzene at room temperature plus ethyl-4-methyl-4-pen-
tenoate. Addition of HC1 favored lactone and unsubstituted unsatur-
ated ester formation.
Some of these products have been rationalized by assuming either
normal carbonylation of the n-ally1 complex or a t the chloromethyl
group, or both, with possible subsequent reactions of initially formed
products also playing a part. For example, 3-methyl-3-hexenedioate
can be formed as a product of carbonylation of ethyl-5-ethoxy-3-
pentenoate which is an ally1 ether. It is also possible to form the
same product by first carbonylating the ethoxymethyl group in the
complex before attacking the s-ally1 system.
The formation of unsubstituted p ,y-unsaturated esters from con-
jugated diolefins in alcoholic solution appears to be a general reac-
tion catalyzed by a number of palladium salts 11871. Yields of these
materials depend on the anion associated with the metal, Na,PdI,
giving considerably better results than either the chloride or bro-
mide. The addition of tri-n-butyl phosphine to the system stabilizes
the catalyst and makes possible the use of higher reaction tempera-
tures. Best results have been obtained with (n-Bu,P),PdI,. Pres-
s u r e s of 1000 atm a r e required, but on addition of p-toluenesulfonic
138 ERIC W. STERN

acid, p r e s s u r e s as low as 250 atm give reasonable rates 11881. It


has been assumed that the catalytically active species in this case
is a hydride which f o r m s a n-ally1 complex with the diene and is
then carbonylated [Eq. (84)].

CH,=CH-CH=CH, + HPd(L),X ,m., Pd,/L


- CH/- co
‘CHi’ X
CH,CH= CH-CH,COOR, f HPd(L),X (L = n-Bu,P)
(84)

b . Unconjugated Diolefins Under comparable conditions noncon-


,

jugated dienes react faster than the corresponding conjugated ma-


terials with CO in the reaction catalyzed by (n-Bu,)PPdI, 11881.
Products in methanol a r e mixtures of unsaturated monoesters and
Downloaded At: 12:43 20 January 2011

saturated diesters, the latter being formed consecutively from the


former. Thus the products from cycloocta-1, 5-diene a r e cycloocta-
4-ene- 1-carboxylate plus a cyclooctadicarboxylate. Nonconjugated
linear dienes, in which both double bonds are terminal, a r e con-
verted directly to cyclic ketoester [Eq. (85)l.

),PdIq
CH,=CH-CH,-CH,-CH=CH, + CO (BusP
MeOH

CH,COOMe
(85)

Rationalization of this product is again based on formation of a pal-


ladium hydride as shown below [Eq. (86)l.

CO, MeOH

CH,COOMe C H,Pd(L, )X
UNSATURATED LIGANDS IN Pd(I1) COMPLEXES 139

Formation of bicycl0-[3,3 ,l]-non-Z-en- 9-one by reaction of cyclo-


octa- l,5-diene with CO in tetrahydrofuran [189] has been rational-
ized similarly.

4. Allene

The r-ally1 complexes derived from allene a r e also carbonylated


[19OJ. In ethanol, 2-chloro-n-ally1 palladium chloride r e a c t s to form
ethyl-3-chloro-3-butenoate [Eq. (87)l.

CH,= C- CH,C OOEt


I
c1
Downloaded At: 12:43 20 January 2011

Direct reaction of allene without prior formation of the complex re-


sults in formation of ethyl itaconate, (27), [Eq. (88)].

CH,- C= CH, + PdC1,


co
/!% /CH, / CH
-2,\
CO, EtOH
ClCO-C\
Y-
Pd
/ \cl/ Pd
\ --7
’ c-cocl
CH, CH,
CH,= C-CH,COOEt
I
COOEt
(27)

Reaction of the complex in which 2-chloro- a-ally1 palladium chloride


has added to allene results in mono-, di-, and tricarbonylation. For-
mation of methyl methacrylate in the reaction of allene and CO cat-
alyzed by PdC1,-GeC1, and PdCl,-SnCl, has also been reported [ 1911.

5. Acetylenes
Reaction of acetylenes with CO results in products that reflect
the ability of the palladium catalysts to promote both mono- and
multicarbonylation of unsaturated compounds and coupling of ace-
tylenes. The particular products obtained are frequently a function
of reaction conditions. In most cases it is unclear whether palla-
dium metal or palladium(I1) i s the catalyst. However, results a r e
usually somewhat better when palladium is added a s Pd(II), and the
140 ERIC W. STERN

intermediacy of complexes in which the oxidation state of the metal


is probably +2 can be inferred.
Acetylene itself r e a c t s with CO in ethanol containing palladium
to yield ethyl acrylate, maleate, and fumarate [192]. In addition to
these, the presence of propionate and succinate has also been ob-
served when the solution contained iodide [ 1931. Reaction of ace-
tylene with CO and (C,H,CN),PdCl, in benzene at loo”, at a p r e s s u r e
of ca. 100 atm, leads to formation of mucanoyl chloride, (281, to-
gether with considerable amounts of maleyl and fumaryl chlorides
11941 [Eq. (89)l.

CH= CHCOCl H \c/cocl


H C r C H + (@CN),PdCl,+ I + I1
CH= CHCOCl C
128) H
’ ‘COCl (89)
Downloaded At: 12:43 20 January 2011

H\C/coC1
+ II
/c,
COCl H

In these products cis carbonylation as well as c i s coupling appear


to predominate. From these results it appears that acetylene coup-
ling is favored in benzene solution. This is confirmed by the obser-
vation that products from the reaction of diphenyl acetylene in 10%
ethanolic HC1 containing a catalytic amount of PdC1, a r e 6,y-diphenyl-
y-crotonolactone, (29), and ethyl diphenylmaleate [Eq. (go)], where-
as in benzene saturated with HCl no carbonylation takes place, the
only product being hexaphenylbenzene [ 1921.

The presence of HC1 also affects product distribution. Thus in


ethanol containing little o r no HC1 the yield of lactone is decreased
considerably, the diester is absent, and, instead, minor amounts of
ethyl-2,3-diphenylacrylate a r e found. As noted previously for re-
actions of olefins, higher HC1 concentrations favor the production
of saturated e s t e r s [195], perhaps in reactions consecutive to c a r -
bonylation of acetylenics. Products from the reaction of ethyl ace-
UNSATURATED LIGANDS IN Pd(I1) COMPLEXES 141

tylenemonocarboxylate in ethanol at room temperature a r e , in de-


creasing yields, ethyl fumarate , ethyl ethylenetricarboxylate, ethyl-
1,1,2-ethanetricarboxylate,and ethyl-l,3-butadiene-1,1,4,4-tetra-
carboxylate, In 10% ethanolic HC1 the saturated tricarboxylate be-
comes the chief product. Similar observations were made in the
c a s e of ethyl acetylenedicarboxylate.
The reactions of propargyl alcohol and chloride [ 1961 are of
particular interest, since products formed a r e identical with those
obtained from the reaction of allene [190]. Thus the major product
from reaction of either propargyl alcohol o r propargyl chloride,
in methanol, was methyl itaconate with small amounts of methyl-2-
(methoxymethyl) acrylate formed in the former case and methyl-3-
chloro-3-butenoate in the latter. A third product, methyl itaconitate,
(30),reflecting both dicarbonylation of the acetylenic and carbonyl
insertion into the reactive alcohol group, was found in the reaction
of propargyl alcohol [Eq. (91)].
Downloaded At: 12:43 20 January 2011

HC=C-CH,OH + CO + MeOH PdC1zb CH,=C-COOMe


I
CH,COOMe
CH-COOMe
I1
+ C-COOMe + CH,=C-COOMe
I I
CH,C OOMe CH,-OMe
(30)
These results suggest that under the conditions employed the allene
n-ally1 complex is either formed by direct reaction of the propargyl
compound with PdC1, (as with n-ally1 complex formation from ally1
compounds) o r that allene was formed prior to complex formation.
The formation of methyl methacrylate from allene [191] implies
that this process i s reversible.

D. Reactions with AliDhatics

1. Palladium Alkyk

Since the formation of acyl palladium complexes a s intermediate


stages in carbonylation reactions has been inferred frequently, it is
of interest that the insertion of CO into an alkyl palladium bond has
also been demonstrated. Reaction of the methyl palladium complexes,
trans-PdXMe(Et,P), (X = C1, Br , I) in hydrocarbon solution at room
temperature and atmospheric p r e s s u r e , yielded the corresponding
142 ERIC W. STERN

trans-acyls, PdX(C0Me) (Et,P), [197]. Stable acyl was not obtained


from the dimethyl complex [198]. The halogen atom in the acetyl
complexes i s labile and is readily replaced by other anions [198].
Of further interest to the mechanism of carbonylations catalyzed
by Pd(I1) is the findings that carbonyl insertion probably involves a
5- o r 6-coordinate intermediate [197] and that the reaction is re-
versible [198].

2 . Cyclopropane

Carbonylation of cyclopropane proceeds readily in benzene solu-


tion at 90" and 100-atm p r e s s u r e in the presence of PdC1, [ 199).
Products a r e a 5: 1:2 mixture of a,p, and y chlorobutyryl chlorides and
propyl benzene. Formation of a-chlorobutyryl chloride a s the ma-
jor product is somewhat surprising. Whereas no cyclopropane-
PdCl, complex appears to have been prepared, the structure of the
cyclopropane-PtC1, complex is thought to have a ring opened struc-
Downloaded At: 12:43 20 January 2011

t u r e , (31), [ZOO], andreactionof such material should yield y-iso-

m e r . The p-isomer and propyl benzene appear to a r i s e f r o m reac-


tions involving propylene, which w a s found to form from cyclopro-
pane and PdC1, in the absence of CO. Some alkylation of benzene
has also been found when (C,H,PdC12), was heated in benzene solu-
tion at reflux for extended periods [22].

VII. REACTION O F UNSATURATED NITROGEN COMPOUNDS

A. Nitric Oxide

A s might be expected from the ('oxidation" of other unsaturated


compounds in aqueous solutions of Pd(I1) salts, NO reacts with
aqueous PdC1, to form a nitro palladium salt [ Z O l ] [Eq. (92)].

2PdC1, + 2NO + H,O - (PdC13N0,)2- + 2H' + PdNOCl (92)

Here palladium reduction to metal i s not observed, and, instead,


the brown, insoluble PdNOCl i s formed. The same material has
UNSATURATED LIGANDS IN Pd(I1) COMPLEXES 143

also been obtained in the reactions of olefins with aqueous palladium


nitrite solution.
Similar results are obtained when the dinitroso complex,
Pd(NO),Cl, [202], is reacted with water. The mononitroso complex,
PdNOC1, r e a c t s with ammonia o r KCN in aqueous solution to yield
the tetrammine or tetracyanide complexes plus N,O [Eq. (93)],
whereas the dinitroso complex, under the same conditions, liber-

2PdNOC1 + NH, + H,O - 2[Pd(NHJ4I2*+ 2C1- + 20H-+ N,O


(93)

ates NO. Thermal decomposition of either complex leads to NO


evolution.
The formal oxidation state of the metal in these complexes is in
doubt.

B. Azo Compounds
Downloaded At: 12:43 20 January 2011

Azobenzene and substituted azobenzenes react with PdCl, in


dioxane-water with formation of complexes, (32), in which a a-bond
i s formed between metal and one of the benzene rings [203] [Eq.
(94)I .

@-N=N-@ + PdC1, -

This compound reacts with LiAlH, to yield hydrazobenzene.

VIII. CONCLUSION

It is evident from the foregoing discussion that the major em-


phasis to date has been on the discovery of new reactions, particu-
larly those with synthetic application. As a result many more ques-
tions have been raised than answered. Even in those reactions which
144 ERIC W. STERN

have been studied most extensively, a detailed interpretation of ob-


servations is not possible and obvious gaps remain to be filled. How-
e v e r , it is to be expected that this deficiency w i l l be corrected be-
fore long and that the resulting expansion of understanding of the
chemistry typified by reactions of unsaturated compounds in the
presence of palladium(I1) w i l l contribute greatly to increased under-
standing of catalytic phenomena a s a whole.

Acknowledgement

The author i s grateful to D r . Harry P. Leftin for his careful


reading of the manuscipt and for numerous suggestions made in its
revision. The cooperation of D r . Patrick M. Henry and D r . Robert
G. Schultz in making available the texts of talks presented a t the
152nd National A.C .S.Meeting is a l s o gratefully acknowledged.
Downloaded At: 12:43 20 January 2011

REFERENCES

111 F. C. Phillips,Am. Chem.J., 16, 255 (1894).


[2] J. Smidt, W. Hafner, R. Jira, J. Sedlmeier, R. Sieber, R.
Ruttinger, and H. Kojer,Angew. Chem., 71, 176 (1959).
131 W.Hafner, R. J i r a , J. Sedlmeier, and J. Smidt, Ber.,95,
1575 (1962).
141 J. Smidt and H. Krekeler, Erdoel Kohle, 16, 560 (1963).
[5] W.H. Clement and C. M. Selwitz, J. Org. Chem.,29, 241
(1964).
[6] R. Huttel, J. Kratzer, and M. Bechter, Ber.,94, 766 (1961).
[7] M. N. Vargaftik, I. I. Moiseev, and Ya. K. Syrkin, Dokl.
Akad. Nauk S S S R , 139, 1396 (1961).
[8] Farbwerke Hoechst, A . - G . , Belg. Pat. 626,669 (1963).
[9] J. Smidt, W. Hafner, R. J i r a , R. Sieber, J. Sedlmeier, and
A. Sabe1,Angew. Chem.,74, 93 (1962).
[lo] K. I. Matveev, A. M. Osipov, V. F. Odyakov, Yu. V.
Suzdal’nitskaya, I. F. Bukhtoyarov, and 0. A . Emel’yanova,
Kinelika i Kalaliz, 3, 661 (1962).
[ll] K. I. Matveev, I. F. Bukhtoyarov, N. N. Shul’ts, and 0. A.
Emel’yanova, Kinelika i Kataliz, 5, 649 (1964).
[12] P. M. Henry, J . Am. Chem. SOC., 86, 3246 (1964).
[13] I. I. Moiseev, M. N. Vargaftik, and Ya. K. Syrkin, Dokl.
Akad. Nauk SSSR, 130, 820 (1960).
[14] P. M. Henry, J . Am. Chem.Suc., 88, 1595 (1966).
[15] S. V. Pestrikov, I. I. Moiseev, and T. M. Romanova, Russ.
J . Znorg. Chem., 10, 1199 (1965).
UNSATURATED LIGANDS IN Pd(I1) COMPLEXES 145

[16] M. A. Muhs and F. T. W e i s s , J . A m . Chem. SOC.,84, 4697


(1962).
[17] S. V. P e s t r i k o v and I. I. Moiseev, Izv. Akad. Nuuk S S S R Ser.
Khim., 1965, 349.
1181 P. M. Henry, J . A m . Chem. SOC.,8 7 , 4423 (1965).
[19] I. I. Moiseev, M. N. Vargaftik, and Ya. K. Syrkin, Dokl.
Akad. Nauk S S S R , 153, 140 (1963).
[20] T. Dozono and T. Shiba, Bull. Jupun Petrol. Inst., 5, 8 (1963).
[21] K. T e r a m o t o , T. Oga, S. Kikuchi, and M. Ito, Yuki Gosei
Kugaku Kyokai Shi, 21, 298 (1963).
[22] E . W. S t e r n , unpublished results.
[23] R. G. P e a r s o n a n d D. A. Johnson, J . A m . Chem. SOC.,86,
3983 (1964).
[24] E. W. S t e r n , Proc. Chem. SOC., 1963, 111.
[25] C. M. H a r r i s , S. E. Livingston, and I. H. Reece, J . Chem.
Soc., 1959, 1505.
[26] M. N. Vargaftik, I. I. Moiseev, and Ya. K. Syrkin, Dokl.
Downloaded At: 12:43 20 January 2011

Akad. Nuuk SSSR, 147, 399 (1962).


[27] R. Jira, J. S e d l m e i e r , and J. Smidt, Ann., 693, 99 (1966).
[28] B. L. Kozik, S. V. P e s t r i k o v , and A. P. Savel’yev, Khim. i
Tekhnol. Topliw i Musel, 8, 11 (1963).
[29] I. I. Moiseev, M. N. Vargaftik, and Ya. K. Syrkin, Izv. Akud.
Nauk SSSR Otd. Khim. Nuuk, 1963, 1144.
[30] I. I. Moiseev, M. N. Vargaftik, a n d Ya. K. Syrkin, k z v . Akud.
Nuuk SSSR Otd. Khim. Nuuk, 1963, 1147.
[31] A. V. Nikiforova, I. I. Moiseev, and Ya. K. Syrkin, Zh.
Obshch. Khim., 33, 3239 (1963).
[32] R. J o y and M. Orchin, 2. Anorg. Allgem. Chem., 305, 236
(1960).
[33] J. Chatt and 1. Leden, J . Chem. Soc., 2036 (1955).
[34] I. I. Moiseev, M. N. Vargaftik, and Ya. K. Syrkin, Dokl.
Akad. Nuuk S S S R , 133, 377 (1960).
[35] E . W. S t e r n a n d M. L. Spector, Proc. Chem. SOC.,1961, 370.
1361 M. S. Kharasch, R. C. S e y l e r , a n d F. R. Mayo, J . A m . Chem.
SOC., 60, 882 (1938).
[37] M. Nakamura and K. Gunji, J . Jupan Petrol. Inst., 6, 191
(1963).
1381 I. I. Moiseev a n d M. N. Vargaftik, Dokl. Akud. Nuuk SSSR,
166, 370 (1966).
[39] P. M. Henry, 152nd Meeting of the American Chemical Soci-
ety, New York, Sept. 1966.
[40] S. V. P e s t r i k o v , Russ. J. Phys. Chem., 39, 218 (1965).
[41] J. Halpern, Chem. Eng. News, 44, NO. 45, 68, (1966).
[42] J. Chatt, K. M. Vallarino, and L. M. Venanzi, J . Chem. Soc.,
1957, 3413.
146 ERIC W. STERN

1431 G. E . Coates and G. Calvin, J . Chem. SOC., 1960, 2008.


[44] J. Chatt and B. L. Shaw, J . Chem. SOC.,1962, 5075.
[45] C. D. Falk and J. Halpern, J . Am. Chem. Soc., 87, 3523
(1965).
[46] L. G. Cannel and R. W. Taft, Jr., J . Am. Chem. Soc.,78,
5812 (1956).
[47] F. Cariati, R. Ugo, and F. Bonati, Inorg. Chem., 5, 1128
(1966).
[48] C. H. Langford and H. B. Gray, Ligand Substitution Pro-
c e s s e s , Benjamin, New York, 1966, p. 49.
[49] Y. Odaira, T . Oishi, T. Yukawa, and S. Tsutsumi, J . Am.
Chem. Soc., 88, 4105 (1966).
1501 J. Tsuji and H. Takahashi, J . A m. Chem. Soc., 87, 3275
(1965).
[51] Consortium f . Electrochem. Ind., Fr. Pat., 1,370,867 (1965).
[52] Imperial Chem. Ind., Ltd., Belg. Pat. 635,426 (1964).
[53] D. Clark, P. Hayden, W. D. Walsh, and W. E . Jones, Brit.
Downloaded At: 12:43 20 January 2011

Pat. 964,001 (1964).


[54] W. D. Shaeffer, U.S. Pat. 3,260,739 (1966).
1551 A. P. Belov, G. Yu. Pek, and I. I. Moiseev, Izv. Akad. Nauk
S S S R Ser . Khim., 1965, 2204.
1561 A. P. Belov and I. I. Moiseev, Izv. Akad. Nauk S S S R Ser.
Khim., 1966, 139.
1571 D. R. Bryant, J. E. McKeon, and P. S. Starcher, Abstract,
2nd International Symbosium on Organometallic Ch e m i sl ry ,
Madison, W is c. , 1965, p.94.
1581 W. Kitching, Z . Rappoport, S. Winstein, and W. G. Young,
J . Am. Chem. SOC., 88, 2054 (1966).
1591 M. N. Vargaftik, I. I. Moiseev, Ya. K. Syrkin, a n d V . V.
Yakshin, Izv. Akad. Nauk SSSR Otd. Khim. Nauk, 1962, 930.
[60] R. G. Schultz, 1.52n.d Meeting of the Am e ri c a n Chemical
Society, New Y o r k , Sept. 1966,
[61] C. B. Anderson and S. Winstein, J . Org. Chem., 28, 605
(1963).
1621 M. Green, R. N. Haszeldine, and J. Lindley, J . Organometal.
Chem. ( A m s t e r d a m ) , 6, 107 (1966).
1631 W. C. Baird, Jr., J . Org. Chem., 31, 2411 (1966).
1641 D. Clark and P. Hayden, 152nd Meeting of the American
Chemical Society, New Y o r k , Sefit. 1966.
[65] I. I. Moiseev and M. N. Vargaftik, I z u . Akad. Nauk SSSR Ser.
Khim., 1965, 893.
[66J L. Friend, L. Wender, and J. C. Yarze, 152nd Meeting of the
A m eri can Chemical Society, New Y o r k , Sept. 1966.
[67] M. L. Spector, H. Heinemann, and K. D. Miller, 152nd Meet-
ing of the A mer ican Chemical Society, New Y o r k , Sept. 1966.
UNSATURATED LIGANDS IN Pd(I1) COMPLEXES 147

D. J. Foster a n d E . Tobler, J. Org. Chem., 27, 834 (1962).


Nippon Gosei, Nagaku Kogyo Kabushiki Kaisha, Fr. Pat.
1,324,029 (1963).
R. Ninomiya, M. Sato, a n d T. Shiba, Bull. Japan Petrol. Inst.,
7 , 31 (1965).
A. P. Belov, I. I. Moiseev, and N. G. Uvarova, Izv. Akad.
Nauk S S S R Ser. Khim., 1965 2224.
J. B. L e e and M. J. Price, Tetrahedron, 20, 1017 (1964).
J. K. Stille and R. A. Morgan, J . A m . Chem. SOC.88, 5135
(1966).
P. M. M a i t l i s , A. E f r a t y , a n d M. L. G a m e s , J . A m . Chem.
Soc., 87, 719 (1965).
K. C. P a n d e a n d S. Winstein, Tetrahedron Letters, 1964,
3393.
K. C. R a m e y and G. L. Statton, J . A m . Chem. S O C . , 88, 4387
(1966).
J. N. D e m p s e y a n d N. C. B a e n z i g e r , J . A m . Chem. SOC., 77,
Downloaded At: 12:43 20 January 2011

4984 (1955).
J. R. Durig, P. Layton, D. W. Sink, a n d B. R . Mitchell,
Spectrochim. Acta, 21, 1367 (1965).
R. C r a m e r , J . A m . Chem. SOC,,86, 217 (1964).
A. D. Westland, J . Chem. SOC.,1965, 3060.
B. D. Fockhart, C. Naccache, J. A. N. Scott, a n d R . C . P i n k ,
Proc. Chem. SOC.,1965, 238.
J. Chatt, L. A. Duncanson, a n d L . M. Venanzi, J . Chem. SOC.,
1958, 3203.
G. Opitz, H. Hellmann, and H. W. Schubert, Ann., 623, 112
(1959).
M. B. Sparke, L. T u r n e r , and A. J. M. Wenham, J . Catalysis
4, 332 (1965).
J. F. H a r r o d and A. J. Chalk, J . A m . Chem. SOC., 86, 1776
(1964).
R. C r a m e r and R. V. Lindsey, Jr., J. A m . Chem. SOC.,88,
3534 (1966).
N. R. Davies, Nature, 201, 490 (1964).
F. A s i n g e r , B. Fell, a n d P. K r i n g s , Tetrahedron Letters,
1966, 633.
B. Cruikshank and N. R. Davies, Australian J . Chem., 19,
815 (1966).
G. C. Bond and M. H e l l i e r , J. Catalysis, 4 , 1 (1965).
J. F. H a r r o d a n d A. J. Chalk, Nature, 205, 280 (1965).
J. F. H a r r o d and A. J. Chalk, J. A m . Chem. S O C . , 88, 3491
(1966).
N. R. D a v i e s , Nuture, 205, 281 (1964).
ERIC W. STERN

L. Roos and M. Orchin, J . A m . Chem. S O C . , 87, 5502 (1965).


J. T. van Gemert and P. R. Wilkinson, J . Phys. Chem.,68,
645, (1964).
A. S. Gow and H. Heinemann, J . Phys. Chem., 64, 1574 (1960).
Consortium f . Electrochem. Ind., Brit. Pat. 887,362 (1962).
R. G. Schultz, J . Polymer Sci., 4B, 541 (1966).
A. J. Canale, W. A. Hewett, T. M. Shryne, and E. A. Young-
man, Chern. Ind. (London), 1962, 1054.
A. J. Canale and W. A. Hewett, J . Polymer Sci.,B2, 1041
(1964).
R. C r a m e r , J . A m . Chem. Soc., 87, 4717 (1965).
R. Huttel and H. Christ, Ber., 97, 1439 (1964).
R. Huttel and H. Christ, Bey., 96, 3101 (1963).
R. Huttel and H. Dietl, Be??., 98, 1753 (1965).
B. L. Shaw, Chem. Ind. (London), 1962, 1190.
S . D. Robinson and B. L. Shaw, J . Clzem. SOC., 1964, 5002.
M. S. Lupin and B. L , Shaw, Tetmhedron Letters, 1964, 15.
Downloaded At: 12:43 20 January 2011

R. G. Schultz, Tetrahedron, 20, 2809 (1964).


J. Tsuji, H. Takahashi, and M. Morikawa, Tetrahedron
Letters, 1965, 4387.
H. Christ and R. Huttel, Angew Chem., 75, 921 (1963).
S. D. Robinson and B. L. Shaw,J. Chem. Soc., 1963, 4807.
G. P a i a r o and A. MUSCO,Tclmhedron Letters, 1965, 1583.
R. J. Goodfellow and L. M. Venanzi, J . Chem. Soc., A1966,
784.
J. C. W. Chien and H. C. Dehm. Chem. Ind. (London), 1961,
745.
J. Powell, S. D. Robinson, and B. L. Shaw, Chem. Cornmiin.,
1965, 78.
M. Donati and F. Conti, Tetrahedron Letters, 1968, 4953.
R. Huttel and J. Kratzer, Angew. Chem., 71, 465 (1959).
J. C. Bailar, Jr., and H. Itatani, Inorg. Chem., 4, 1618 (1965).
F. H. Jardine, J. A. Osborn, G. Wilkinson, and J. F. Young,
Chem. Ind. (London), 1965, 560.
R. D. Gillard, J. A. Osborn, P. B. Stockwell, and G. Wilkin-
son, Proc. Chem. Soc., 1964, 284.
L. Vaska and R. E . Roades, J . Am. Chem. SOC., 87, 4970
(1965).
I. Jardine and F. J. McQuillin, Tetrahedron Letters, 1966,
4871.
R. N. Rylander, N. Himelstein, D. R. Steele, and J. Kreidl,
EngZehard Ind. Tech. Bull., 3, 61 (1962).
M. Pesez and J. F. Burtin, Bidl. SOC.Chim. France, 1959,
1996.
UNSATURATED LIGANDS IN Pd(I1) COMPLEXES 149

M. F. Sloan, A. S. Matlack, and D. S. Breslow, J . A m . Chem.


SOC., 85, 4014 (1963).
L. Malatesta, G. Santarella, L. Vallarino, and F. Zingales,
Angew. Chem., 72, 34 (1960).
L. M. Vallarino and G. Santarella, Gazz. Chim. Ital., 94, 252
(1964).
P. M. Maitlis, D. Pollock, M. L. Games, and W. J. Pryde,
Can. J . Chem., 43, 470 (1965).
F. Zingales, Ann. Chim. (Rome),52, 1174 (1962).
R. Huttel and H. J. Neugebauer, Tetrahedron Letters, 1964,
3541.
L. B. Luttinger and E. C. Colthup, J . Org. Chem., 27, 3752
(1962).
0. N. Temkin, S. M. Brailovskii, R. M. Flid, M. P. Strukova,
V. B. Belyanin, and M. G. Zaitseva, Kinetika i Kataliz, 5,
192, (1964).
R. Nast and W. Horl, B e y . , 95, 1470 (1962).
Downloaded At: 12:43 20 January 2011

E. 0. Greaves and P. M. Maitlis, J . Organometal, Chem.


(Amsterdam), 6, 104 (1966).
R. Van Helden and G. Verberg, Rec. ’I’rau. Chim., 84, 1263
(1965).
J. M. Davidson and C. Triggs, Chem. Ind. (London),457
(1966).
J. Smidt and R. Sieber, Angew Chem., 71, 626 (1959).
M. Donati and F. Conti, Tetrahedron Letters, 1966, 1219.
J. Smidt, R. Sieber, W. Hafner, and R. Jira, Ger. Pat.
1,176,141 (1964).
E. W. Stern, M. L. Spector, and H. P. Leftin, J . Catalysis,
6, 152, (1966).
Shell International Research Maatschappig N. V., Brit. Pat.
1,010,024 (1965).
H. A . Smith and W. H. King, J . A m . Chem. SOC., 72, 95 (1950).
S. Patai and Z. Rappoport, The Chemistry of the Alkenes
(S. Patai, ed.), Wiley (Intersciene), New York, 1965, pp. 525ff.
H. J. Hagemeyer, Jr., and D. C. Hull, Ind. Eng. Chem., 41,
2920 (1949).
W. H. Clemment andC. M. Selwitz, Tetrahedron Letters, 1962,
1081.
E. A. Jeffrey and D. P. N. Satchell, Chem. Ind. (London),
1960, 1444.
J. Smidt and W. Hafner,Angew. Chem., 71, 284 (1959).
W. Hafner, H. Prigge, and J . Smidt,Ann., 693, 109 (1966).
W. Manchot and J. Konig, B e y . , 59, 883 (1926).
A. B. Fasman, G . G. Kutyukov, and D. V. Sokol’skii, Russ.
J . Inorg. Chem., 10, 727 (1965).
150 ERIC W. STERN

G. D. Zakumbaeva, N. F. Noskova, E. N. Konaev, and D. V.


Sokol’skii, Dokl. Akad. Nauk SSSR, 156, 1386 (1964).
G. D. Zakumbaeva, N. F. Noskova, E. N. Konaev, and D. V.
Sokol’skii, Dokl. Akad. Nauk SSSR, 159, 1323 (1964).
V. A. Golodov. G. G. Kutyukov, A. B. Fasman, and D. V.
Sokol’skii, Zh. Neorg. Khim., 9, 2319 (1964).
V. A. Golodov, A. B. F a s m a n , and D. V. Sokol’skii, Dokl.
Akad. Nauk S S S R , 151, 98 (1963).
A. B. Fasman, V. A. Golodov, and D. V. Sokol’skii, Dokl.
Akad. Nauk S S S R , 155, 434 (1964).
A. B. F a s m a n and V. A. Golodov, Kinetika i Kataliz, 6, 956
(1965).
R. J. Irving and E . A. Magnusson, J . Chem. SOC.,1958, 2283.
G. G. Kutyukov, A. B. F a s m a n , A. E . Lyuts, Yu. A. Kushuri-
kov, V. F. Vozdvizhenski, and V. A. Golodov, Zh. Fiz. Khim.,
40, 1468 (1966).
A. T r e i b e r , Tetrahedron L e t t e r s , 1966, 2831.
Downloaded At: 12:43 20 January 2011

S. Nakamura and J. Halpern, J . Am. Chem. SOC.,83, 4102


(1961).
A. C. Harkness and J. Halpern, J . Am. Chem. SOC.,83, 1258
(1961).
R. J. Mawby, F. Basolo, and R. G. P e a r s o n , J . A m . Chew.
SOC., 86, 5043 (1964).
I. L. Mador and A. U. Blackham, U.S. Pat. 3,114,762 (1963).
E. W. Stern and M. L. Spector, J . Org. Chem., 31, 596 (1966).
J. Tsuji and N. Iwamoto, Chem. Commun., 380 (1966).
E. 0. F i s c h e r and A. Vogler, J . Organometal. Chem. (Amster-
dam), 3 , 1 6 1 (1965).
F. Calderazzo,Inorg. Chem., 4, 293 (1965).
T. Saegusa, S. Kobayashi, and K. Hirota, Tetrahedron
L e t t e r s , 1966, 6125.
W. F. Edgell, M. T. Yang, B. J. Bulkin, R. Bayer, and N.
Koizumi, J . Am. Chem. Soc., 87, 3080 (1965).
W. F. Edgell and B. J. Bulkin, J . Am. Chem. SOC., 88, 4839
(1966).
W. T. Dent, R. Long, and A. J. Wilkinson, J . Chem. SOC.,
1964, 1585.
J. K. Nicholson, J. Powell, and B. L . Shaw, Chem. Commun.,
1964, 174.
J. Tsuji and N. Iwamoto, Chem. Commun., 1966, 828.
J. Tsuji, M. Morikawa, and J. Kiji, Tctrahedvon L e t t e r s ,
1963, 1061.
J. Tsuji, M. Morikawa, and J. Kiji, J . Am. Chem. SOC., 86,
4851 (1964).
UNSATURATED LIGANDS IN Pd(I1) COMPLEXES 151

A. J. Chalk, Tetrahedron Letters, 1964,2627.


J. Tsuji, K. Ohno, and T. Kajimoto, Tetrahedron Letters,
1965,4565.
J. Tsuji, M. Morikawa, and J. Kiji, Tetrahedron Letters,
1963,1437.
W. Dent, R. Long, and G. H. Whitfield, J . Chem. SOC.,1964,
1588.
J. Tsuji, J. Kiji, S. Imamura, and M. Morikawa, J. Am.
Chem. SOC., 86,4350 (1964).
J. Tsuji, J. Kiji, S. Imamura, and M. Morikawa, Tetrahedron
Letters, 1963,1811.
R. Long and G. H. Whitfield, J . Chem. SOC., 1964,1852.
J. Tsuji, S. Imamura, and J. Kiji, J . Am. Chem. Soc., 86,
4491 (1964).
R. F.Heck, J. Am. Chem. SOC., 85,2013 (1963).
E . 0. Fischer and G. Burger, Z . Naturforsch., 16b,702 (1961).
J. Tsuji and S. Hosaka, J . Am. Chem. SOC., 87,4075 (1965).
Downloaded At: 12:43 20 January 2011

S.Brewis and P. R. Hughes, Chem. Commun., 1965,157.


S . Brewis and P. R. Hughes, Chem. Commun., 1965,489.
S . Brewis and P. R. Hughes, Ghem. Commzm., 1966,6.
J. Tsuji and T. Susuki, TetruFiedron Letters, 1965,3027.
E. L. Jenner and R. V. Lindsey, Jr., U.S. Pat. 2,876,254(1959).
J. Tsuji and T. Nogi, J . Am. Chem. S O C . , 88,1289 (1966).
G.Jacobsen and H. Spathe, Ger. Pat. 1,138,760(1962).
J. Tsuji, M.Morikawa, and N. Iwamoto, J . Am. Chem. Soc.,
86,2095 (1964).
J. Tsuji and T. Nogi, J . Org. Chem., 31,2641 (1966).
J. Tsuji and T . Nogi, Tetrahedron Letters, 1966,1801.
G . Booth and J. Chatt,Proc. Chem. SOC., 1961,67.
G. Booth and J. Chatt, J . Chem. SOC.,A1966,634.
J. Tsuji, M. Morikawa, and J. Kiji, Tetrahedron Letters,
1965,817.
D.M. Adams, J. Chatt, and R. G. Guy, Proc. Chem. SOC.,
1860,179.
J. Smidt and R. Jira, Ber., 93,162 (1960).
W. Manchot and A. Waldmuller, Bey., 59,2363 (1926).
R. F. Nutt, B. Arison, F. W. Holly, and E. Walton, J . Am.
Chem. S O C . , 3272
~ ~ , (1965).
NOTE ADDED IN PROOF
The following pertinent publications have appeared since this re-
view was submitted:
The dependence of formation constants for (CnHm)PdCl, and
(CnH,)PdC12(OH2) on ionic strength [204]; kinetics of vinyl acetate
formation from C,H+ Pd(OAc),, NaOAc, and p-benzoquinone in
HOAc 1205, 2061,and of l-butene isomerization by aq. PdCl, [207];
152 __ .. ERIC W. STERN

dimerization of ethylene [208,209],benzene [210],vinyl acetate [211],


and /3-substituted a-olefins [212];hydrogenation of polyolefinic mate-
r i a l s to monoolefins [213,214];preparation and characterization of a
stable palladium hydride [215];decomposition of vinyl acetate [216];
N-vinylation of cyclic amides [217];reversible n-ally1 complex for-
mation from allene [218];carbonylation of piperilene [219], a,w-dienes
[220];and n-ally1 complexes derived from a,p- and p , y-unsaturated
esters [2211;formation of complexes containing carbon-metal 0 bonds
from tertiary allylic amines and PdC1, /222]; and formation of glyoxal
from ethylene, NOz, or HNO,, and Pd salts in aqueous media [223].
Oxypalladation of styrene leads, as expected, to 1,2-disubstitution
only [2241.

[204] S. V. Pestrikov, I. I. Moiseev, and L. M. Sverzh, Zh. Neorg-.


Khim., 11, 2081 (1966).
[205] A. P. Belov, I. I. Moiseev, and N. G. Uvarova, Izv. Akad. Nauk
SSSR. Ser. Khim., 1966,1642.
Downloaded At: 12:43 20 January 2011

[206] I. I. Moiseev, A. P. Belov, V. A. Igoshin, and Ya. K. Syrkin,


Dokl. Akad. Nauk SSSR, 173,863 (1967).
[207] I. I. Moiseev, S. V. Pestrikov, and L. M. Sverzh, Izv. Akad.
Nauk SSSR, Ser. Khirvz, 1966,1866.
[208] Y. Kusunoki, R. Katsuno, N. Hasegawa, S. Kurematsu, Y. Nagao,
K. Ishii, and S. Tsutsumi, Bull. Chem. SOC.Japan, 39, 2021 (1966).
[209] A. D. Ketley, L. P. Fisher, A. J. Berlin, C. R. Morgan,
E. H. Gorman, and T. R. Steadman, Inorg-. Chem., 6,657 (1967).
[210] J. M. Davidson and C. Triggs, Chem. Ind. (London), 1967,1361.
[211] C. F. Kohl and R. van Helden, Rec. Trau. Chim., 86,193 (1967).
[212] H. C. Volger, Rec. Trav. Chim., 86,677 (1967).
[213] J. C. Bailar and H. Itatani, J . A m . Chem. Soc., 89,1592 (1967).
[214] H. A. Tayim and J. C. Bailar, J . A m . Chem. Soc.,89,4330 (1967).
[215] E. H. Brooks and F. Glocking, J . Chem. SOC.(A), 1967,1030.
[216] R. G. Schultz and P. R. Rony, 153rd Meeting of the American
Chemical Society, Miami, April 1967.
[217] J. E. McKeon and P. S. Starcher, U.S. Pat. 3,318,906(1967).
[218] M. S. Lupin, J. Powell, and B. L. Shaw, J . Chem. SOC.(A),
1966,1687.
[219] C. Bordencaand W.E. Marsico, Tetrahedron Letters, 1967,1541.
[220] S. Brewis and P. R. Hughes, Chem. Commun., 1967,71.
[221] J. Tsuji and S. Imamura, Bull. Chem. SOC. Japan, 40,197 (1967).
[222] A. C. Cope, J. M. Kliegman, and E. C. Friedrich, J . A m .
Chem. SOC.,89,287 (1967).
[223] R. Platz and W. Fuchs, Brit. Pat. 1,041,376(1966).
[224] S. Uemura and K. Ichikawa, Bull. Chem. SOC.Japan, 40,1016
(1967).

You might also like