98
Chapter 3
was more active (2 turnovers/h). Similar but
water soluble tungsten and molybdenum complexes are known [223-226]
which would allow the use of water as solvent for such reactions. It is
noteworthy, though, that ionic hydrogenation of ketones by dihydrogen
complexes has so far been observed only in non-aqueous solutions
[223,227]; perhaps the coordination of a ketone is disfavoured in water due
to competition by
3.4 HYDROGENATION OF MISCELLANEOUS
ORGANIC SUBSTRATES
3.4.1
Hydrogenation of nitro compounds and imines
Amines are extremely important intermediates and end products of the
chemical industry and are often obtained by hydrogenation of the
corresponding nitro compounds or imines. A search of the literature reveals,
that hydrogenation of nitro compounds catalyzed by well-defined molecular
complexes in aqueous solutions is rare. One reason may line in the fact, that
reduction of the function proceeds in one-electron steps, while many
soluble hydrogenation catalysts act in the “oxidative addition of
elimination of the product” cycles in which the central metal ion
(formally) looses or gains two electrons at a time. It is not surprising
therefore, that the catalysts of nitro-hidrogenations are either metal centered
radicals themselves or are capable of delocalizing the
temporary surplus of electron(s) on their large conjugated system
or on a cluster framework Catalysts, operating
through formation of intermediate monohydrides, which does not require the
change of the oxidation state of the metal, are good candidates of nitro-
reduction (see also 3.8.2) on reductions with ). Unfortunately,
other functional groups in a molecule are usually even more reactive
towards hydrogenation than the nitro functionality - therefore selective

derived catalysts [232] (see Ch .10).
Asymmetric hydrogenation of imines was studied in aqueous/organic
biphasic systems and presented a puzzle which is still not solved completely.
It was first discovered by Bakos et al. [233], that acetophenone benzylimines
were hydrogenated to the corresponding amines with unprecedented
enantioselectivity up to 96% under very mild conditions with the Rh-
complexes of sulfonated BDPP,
36
, provided the degree of sulfonation of
BDPP was close to 1 (in fact it was 1.41-1.65) (Scheme 3.26). With
increasing number of sulfonate substituents in 36 the enantioselectivity
decreased sharply.
100
Chapter 3
This “monosulfonation effect” was investigated in detail by de Vries et
al. [234,235], who isolated the sulfonated BDPP with one, two, three and
four sulfonate groups (each phenyl ring carries only one). In hydrogenation
of acetophenone benzylimine it was confirmed, that indeed, the highest
enantioselectivity (94 %) could be achieved by using monosulfonated BDPP
as ligand in the in situ prepared Rh-catalyst, whereas with the bissulfonated
ligand a practically racemic product (2 % e.e.) was obtained Note, that
monosulfonated BDPP is chiral at one of the phosphorus atoms, and it was
determined by HPLC that it contained a 1:1 ratio of the two epimers. Now
the puzzle is in that how can a ligand, which is a 1:1 mixture of two
diastereomers, induce such outstandingly high enantioselectivity what was
found with the Rh-complex of monosulfonated BDPP in the hydrogenation
of imines. It is also important to add, that under comparable conditions, the
enantiomeric excess of the hydrogenation of acid and
its methyl ester decreased monotonously with increasing degree of
sulfonation (from 87 % to 65 % and from 74 % to 45 %, respectively).

always keep this warning in mind.
3.5 TRANSFER HYDROGENATION AND
HYDROGENOLYSIS
Transfer hydrogenation is a reaction in which hydrogen is catalytically
transferred from a suitable hydrogen donor to a reducible substrate
(S) yielding the hydrogenated product and the oxidized form of the
donor molecule (D) [236-238].
Several of the most common hydrogen donors, such as formic acid and
formates, ascorbic acid, EDTA or 2-propanol are well or at least sufficiently
soluble in water. In addition, itself can serve as a source of hydrogen.
Frequently, hydrogenation of unsaturated substrates is achieved by using
mixtures; such reactions are discussed in 3.8. As written in eq.
(3.11) the hydrogen transfer reaction is often reversible, an obvious example
being the reduction of ketones using 2-propanol as donor.
Reductions with hydrogen transfer are attractive for at least two reasons.
First, the concentration of in the reaction mixture can be much higher
than that of under high pressure (cf. for example and
in water at 1 bar pressure). This may be beneficial for a
faster reaction. Second, the use of a soluble or liquid hydrogen donor also
eliminates the safety hazard of handling high pressure hydrogen.
Formic acid and formates were among the most effective donors used for
the reduction of olefins with or catalysts in
non-aqueous systems [239-241]. No wonder, the water soluble analogues of
these catalysts became widely used in aqueous solutions. In a series of
investigations [242-245] with Ru/TPPMS and Rh/TPPMS catalysts olefins
(such as 1-heptene) were hydrogenated in mixtures of HCOOH/HCOONa.
Crotonaldehyde was selectively reduced to butyraldehyde by the
catalyst [245]. It was also established that (unfiltered)
ultraviolet irradiation accelerated the reactions [245].
Dimethyl itaconate was reduced by hydrogen transfer from aqueous

The same reaction was investigated in a reverse experimental setup, i.e.
having the water-soluble catalyst excess TPPMS,
and the hydrogen donor HCOONa in the aqueous phase and the substrate
aldehyde together with the products in the organic (chlorobenzene) phase
[249,250]. Unsaturated aldehydes, such as cinnamaldehyde (Scheme 3.18)
104
Chapter 3
and citral (Scheme 3.31) were reduced to the corresponding unsaturated
alcohols with high selectivity. No cis-trans isomerization was observed
around the double bond.
It is important to note, that in this arrangement of an aqueous/organic
biphasic reaction the substrate inhibition discussed above was hardly
observable. Although the aldehydes are sufficiently soluble in water to allow
a fast reaction, still most of the substrate is found in the organic phase at all
times. Therefore the concentration of the aldehydes in the catalyst-
containing aqueous phase is not high enough to cause efficient inhibition of
catalysis [250]. Under comparable conditions, Ru(II) and Rh(I) complexes
of PTA behaved very similar to their TPPMS-containing analogues in that
led to exclusive formation of unsaturated alcohols [27,204]
while catalysis by selectively produced saturated aldehydes in
reduction of unsaturated aldehydes with [27,28,204].
In contrast to the case of the water soluble complexes (P =
PTA, TPPMS or TPPTS) which did not promote the reduction of
function in aldehydes or ketones in biphasic systems, was
found an active catalyst for reduction of ketones with aqueous HCOONa
(Scheme 3.32). The reaction was aided by phase transfer catalysis using
Aliquat-336 and required a large excess of to prevent reduction of
rhodium into inactive metal. Substrates like acetophenone, butyrophenone,
cyclohexanone and dibenzyl-ketone were reduced to the corresponding
secondary carbinols with turnover frequencies of [251].

The reaction rate of the reduction of these carbonyl compounds showed a
sharp maximum at pH 3.2, which coincides with the value of HCOOH
in the studied concentration, and there was no reaction above pH 5. The lack
of reactivity at higher pH can be attributed to the formation of the
catalytically inactive hydroxide-bridged trimer, which,
however, is in equilibrium with the starting catalyst precursor at the
optimum pH of the reaction. The active form of the catalyst is most probably
the dimeric which happens to form to the
hand, is known to be a good catalyst for ketone hydrogenation
in the presence of amines [253].
It is instructive to see, that in biphasic aqueous organometallic catalysis a
seemingly minor change (dissolving the catalyst in the aqueous or, contrary,
in the organic phase) may lead to major changes in the rate and/or the
selectivity of the catalyzed reaction under otherwise identical conditions.
Hydrogenation
107
highest extent at pH 3.2; the compound was characterized in solution and in
isolated from, as well. It is supposed that reduction of the carbonyl
compounds takes place on this dimer (Scheme 3.35).
108
Chapter 3
In the presence of aqueous NaOH, palladium(II) chloride was effective
for the transfer hydrogenation of unsaturated acids, azlactones and
phenylpyruvic acid (Scheme 3.36) at 65 °C although in quite long reaction
times (typically 16 h) [255]. For these water-soluble substrates organic
solvents were not required. No attempt was made to clarify the nature of the
active catalytic species, which -under these conditions- may well be a fine
colloid of Pd metal.
Hydrogen transfer to ketones from 2-propanol was developed into an
extremely efficient method of obtaining secondary alcohols [256,257] and

being less favoured.
Chloroarenes were efficiently hydrodechlorinated with a
( or ) catalyst in biphasic systems under mild conditions [267].
The catalyst tolerates the presence of a variety of functional groups (R, OR,
COAr, COOH, ). Some chloro heterocycles (e.g. 5-chloro-l-ethyl-
2-methylimidazole) can be readily dehalogenated, but 2-chlorotiophene does
not react at all.
Several aliphatic and benzylic halides were dehalogenated by hydrogen
transfer from sodium or ammonium formate with
or catalysts [268]. As it is seen from Table
3.10, carbon tetrachloride and benzyl chloride were particularly reactive. As
expected, the order of reactivity was and chlorobenzene
remained unchanged. Interestingly, of the two ruthenium complexes
was a less effective catalyst for the reactions of carbon
tetrachloride and chloroform, however, it showed appreciably higher
catalytic activity in the dehalogenation of hexyl halides. The turnover
numbers (TON-s) in Table 3.10 were obtained in 3 h reactions and there is a
remarkable difference to an analogous system, with as
catalyst, where benzyl chloride was reduced by HCOOLi in refluxing
dioxane with only 26 turnovers in 6 h [269].
110
Chapter 3
Under the conditions of Table 3.10, HCOOH is decomposed and the
yield of reaches only 3.5% (in contrast to HCOONa, where the 478
TON corresponds to 60 % conversion). The use of 5 bar instead of
formate as hydrogen donor leads to 1.8 % conversion. The reason for this
low conversion can be the
low
concentration of dissolved hydrogen (<0.004
M under the reaction conditions) as opposed to that of formate (1.67 M);

Under “ligandless” conditions catalyzed the hydrogenolysis of
several 4-substituted aryl chlorides in alkaline aqueous solutions using
as reductant (Scheme 3.40) [275]. In case of certain ortho-
substituted substrates, such as 2-chlorophenolate and 2-chloroaniline,
strong chelation in the intermediate palladacycle completely inhibited the
reaction. On the other hand, in case of 2-chlorobenzoic acid addition of
iodide led to 86 % yield of benzoic acid.
could also be used for the hydrogenolytic removal of phenolic
hydroxy groups. However, in this case phenols, e.g. 4-methoxyphenol, had
to be transformed first into monoaryl sulfates which, in turn, could be
reduced by into the corresponding arenes (Scheme 3.40)
[276]. Again, no phosphines or other organic ligands were required for an
efficient reaction.
Homogeneous catalytic asymmetric hydrogenolysis of epoxysuccinates
offers a route to the preparation of chiral malic acid derivatives [277] which
are useful building blocks in natural product synthesis. The reaction was
studied in aqueous solution using a catalyst prepared from
and sulfonated BDPP (
36
) with varying degree of sulfonation [278] (Scheme
3.41). In contrast to the hydrogenation of prochiral imines (Table 3.9) the
enantioselectivity of the hydrogenolysis of sodium cis-epoxysuccinate
decreased monotonously with the increasing number of sulfonate groups, i.e.
no “monosulfonation effect” was observed (see 3.4.1). The reason probably
is in that the sodium salt of cis-epoxysuccinic acid dissolves well in water
where the catalysis takes place, in contrast to imines and esters of
dehydroamino acids. Therefore the exceptional solubility of the Rh(I)-
complex of monosulfonated BDPP in organic solvents does not play a role
here.
In a related reaction, racemic sodium trans-phenylglycidate was

”
by metal complexes
has been reviewed quite frequently [279-285] therefore this short chapter
covers only the partially or fully aqueous systems.
Since carbon dioxide is a thermodynamically stable, highly oxidized
compound, its synthetic utilization requires some kind of a reduction -
reaction with molecular hydrogen is a distinct possibility. Stepwise
reduction of with may yield formic acid, formaldehyde, methanol
and finally methane, together with CO or Fischer-Tropsch-type derivatives
as shown on Scheme 3.42. In aqueous organometallic catalysis the most
common product of such a reduction is formic acid. Formation of carbon
monoxide, formaldehyde, and methane has already been reported, however,
methanol and Fischer-Tropsch type products were not observed.
114
Chapter 3
Reaction of two gaseous compounds resulting in a liquid product are
biased by a decrease in enthropy which –depending on the temperature–
may make the whole process thermodynamically unfavourable. This is also
the case for the hydrogenation of to HCOOH (eq. 3.12) with
However, in aqueous solution hydration of the solutes makes the overall
enthropy difference smaller, and reaction (3.13) becomes slightly exergonic
with
Thermodynamics tells, therefore, that in water this reaction is likely to
proceed; we must not forget, though, that these data refer to standard
conditions, and in order to eliminate the kinetic activation barrier at 25 °C
highly active catalysts are needed. Unfortunately, the same catalysts can also
be active in the reverse process, i.e. in the decomposition of formic acid to
and at low pressures; decomposition to CO and (i.e. the reverse
water gas shift) is rarely observed.
Equation (3.13) can also be shifted to the right by further reactions of

From the reaction mixture could be isolated,
which was also supposed to be the key catalytic intermediate in the reaction.
Although the role of water was not specifically addressed it is worth noting
that exclusively formic acid i.e. no ethyl formate was produced despite the
presence of 80 % ethanol in the solvent.
Tsai and Nicholas used as a
catalyst precursor for hydrogenation in THF and also observed
acceleration of the reaction in the presence of water [290]. With careful
spectroscopic measurements they could detect the formation of the
dihydrides, and and also that of the
bidentate formato complex, It was therefore suggested
that the mechanism of the reaction involved the insertion of into the
Rh-H bond of the dihydride yielding a hydridorhodium-formato
intermediate, followed by reductive elimination of formic acid then
oxidative addition of to regenerate the dihydride (Scheme 3.43).
116
Chapter 3
It was also suggested [290] that the rate accelerating effect of water was
due to formation of an intermolecular hydrogen bond between the
ligand in and the incoming within the insertion
transition state, such as depicted (A) on Scheme 3.44.
Hydrogenation
117
The existence of such or closely related intermediates received support
from studies on the water effect in the hydrogenation of carbon dioxide
catalyzed by Tp = hydridotris(pyrazolyl)borate
[291]. This complex catalyzes the reduction of to formic acid with an
average (100 °C, 50 bar total pressure,
15 mL/5 mL, 2 mL). Since addition of inhibited the
reaction, it was concluded, that the catalytic cycle probably does not involve

additive. It is also interesting, that the primary products of the reaction were
formic acid and formaldehyde
,
which later decomposed to give CO and
(and ). Although without any spectroscopic or other evidence, the
catalytic cycle was suggested to involve formation of a metallocarboxylic
118
Chapter 3
acid (via “abnormal insertion” [279] of into the Ru-H bond), as shown
on Scheme 3.45
Water-soluble rhodium complexes, such as or the ones
prepared in situ from and TPPTS and from
and TPPTS were succesfully used by
Leitner et al. [282,295] for the hydrogenation of in aqueous solutions in
the presence of amines or aminoalkanols. In this system no other products of
carbon dioxide reduction, such as formaldehyde or methanol could be
detected. There was no formation of HCOOH in the absence of an amine,
however, a formic acid concentration of 3.63 M was obtained in an aqueous
solution containing 3.97 M (well soluble in water as compared to
) and 5.4 mM Initial turnover frequencies were
substantially higher than any other before, e.g. at 81 °C and 40 bar total
pressure a was observed. For this reaction
an overall activation barrier was determined. Interestingly,
under the same conditions the ruthenium complex, proved
much inferior to the Rh-TPPTS catalysts with a TOF of only In the
supposed catalytic cycle key role was assigned to a monohydrido-rhodium
complex (Scheme 3.46) which at that time could
not be supported by spectroscopic methods but which later became
characterized by and NMR spectroscopy [86].
Hydrogenation

solutions with concentrations up to 0.93 M [296].
Hydrogenation
121
It should be mentioned here, that the heterogeneous catalytic
hydrogenation of bicarbonate in aqueous solution is a well-known process
[300,301]. Coupled to a catalytic decomposition of formate back to and
this reaction was even suggested as a method for storage and
transport of hydrogen [272]. The best catalysts of bicarbonate hydrogenation
consist of metallic palladium, either supported, such as Pd/C, or colloidal,
stabilized by The latter [273] actively catalyzes the
photochemically assisted reduction of
A very interesting finding was published by Pruchnik et al. who studied
the hydrogenation of with catalysts prepared from
or other water soluble small phosphines, such as and
[302]. When a mixture of and was passed through a
flow reactor ( each) containing an aqueous solution of the Rh-
PTA catalyst, the major product was CO accompanied by a few % of
methane (Scheme 3.48). At 70 °C the activity of the catalyst for CO
production reached a The peculiarity of this system is in
the production of methane which had not been observed before with
homogeneous catalysts. Unfortunately, no further results have been
published and no suggestion concerning the catalytic cycle have been made
yet.
122
Chapter 3
The iridium cluster with = pyridylphosphines 67
( 1, or 2) and the rhodium complex with
and acac = acetylacetonate, have been claimed recently as
catalysts for the removal of C oxides from mixtures of or CO and
by passing the gas mixture through a homogeneous aqueous acid solution of