Nanosized hydrogenation catalyst based on palladium bisacetylacetonate and phosphine: Formation, the origin of activity, and properties

Article abstract of DOI:10.1134/S0023158407050072

The nature and catalytic properties of a hydrogenation catalyst based on Pd(acac)₂ and PH₃ are considered. As demonstrated by a variety of physicochemical methods (IR and UV spectroscopy, ³¹P and ¹H NMR, electron microscopy, and X-ray powder diffraction), nanoparticles consisting of various palladium phosphides (Pd₆P, Pd_4.8P, and Pd₅P₂) and Pd(0) clusters form under the action of dihydrogen during catalyst preparation. The promoting effect of phosphine at low PH₃: Pd(acac)₂ ratios is mainly due to the ability of phosphine to increase the extent of dispersion of the catalyst.

Full text of DOI:10.1134/S0023158407050072

ISSN 0023-1584, Kinetics and Catalysis, 2007, Vol. 48, No. 5, pp. 640–648. © MAIK “Nauka /Interperiodica” (Russia), 2007.
Original Russian Text © F.K. Shmidt, L.B. Belykh, N.I. Skripov, L.N. Belonogova, V.A. Umanets, A.V. Rokhin, 2007, published in Kinetika i Kataliz, 2007, Vol. 48, No. 5,
pp. 685−694.
Nanosized Hydrogenation Catalyst Based on Palladium
Bisacetylacetonate and Phosphine: Formation, the Origin
of Activity, and Properties
F. K. Shmidt, L. B. Belykh, N. I. Skripov, L. N. Belonogova, V. A. Umanets, and A. V. Rokhin
Irkutsk State University, Irkutsk, 664033 Russia
e-mail: belykh@chem.isu.ru
Received February 22, 2006
Abstract—The nature and catalytic properties of a hydrogenation catalyst based on Pd(acac)2 and PH3 are con-
sidered. As demonstrated by a variety of physicochemical methods (IR and UV spectroscopy, 31P and 1H NMR,
electron microscopy, and X-ray powder diffraction), nanoparticles consisting of various palladium phosphides
(Pd6P, Pd4.8P, and Pd5P2) and Pd(0) clusters form under the action of dihydrogen during catalyst preparation.
The promoting effect of phosphine at low PH3 : Pd(acac)2 ratios is mainly due to the ability of phosphine to
increase the extent of dispersion of the catalyst.
DOI: 10.1134/S0023158407050072
In recent years, researchers have expressed steady ence of other reducing agents, such as sodium hypo-
interest in the synthesis and characterization of nano- phosphite [4]. Depending on the nature of the reducing
sized particles, in particular, metal nanoparticles, agent and on the extent of destruction of the organo-
whose application areas include catalysis [1–4]. In the phosphorus ligand, the resulting nanosized particles
absence of a stabilizer, metal nanoclusters are typical can be metal cores with a ligand shell, polynuclear pal-
liophobic colloids with a low aggregation resistance. ladium complexes with phosphide and phosphinidene
For this reason, nanosized particles of metals are usu- ligands, or even palladium phosphides bearing immobi-
ally synthesized in the presence of stabilizing ligands, lized Pd(0) clusters [4, 9, 18, 20, 21]. There have been
among which are low-molecular-weight compounds reports on hydrogenation catalysts based on palladium
(ortho-phenanthroline (phen) [5, 6], quarternary complexes with various organic phosphines, including
ammonium salts [7], and phosphines [4, 8, 9]), polyoxo- tertiary [4, 9, 16–19], secondary [20], and primary [21].
metallate ions (Mo(IV), W(VI), V(V), and Nb(V) [10, Unsubstituted phosphine, PH3, has not been used in the
11]), and macromolecular compounds [12, 13]. For synthesis of microheterogeneous hydrogenation cata-
example, palladium carboxylate reduction with hydro- lysts.
gen in the presence of ortho-phenanthroline followed
In order to gain deeper insight into the modifying
by the oxidation of the intermediate product with oxy-
effect of phosphorus-containing compounds and to
gen yields catalytic palladium nanoclusters with the
reduce the number of steps in the synthesis of efficient
idealized formulas [Pd561(phen)60](OAc)180 [1, 5, 6],
hydrogenation catalysts, we studied the formation and
catalytic properties of the Pd(acac)2 – PH3 system.
Pd147(phen)32O60(OCOR)30 (R = Me, Et, CHMe, CMe3)
[14], and Pd1215(phen)60O1100 [15].
The quarternary ammonium stabilizers turned out to
EXPERIMENTAL
be ineffective. The palladium nanoclusters obtained in the
presence of these compounds aggregate within 5–6 min
in a hydrogen atmosphere [7]. At the same time, by
depositing these clusters onto a solid substrate, it is pos-
sible to prepare stable heterogeneous hydrogenation
catalysts superior to the commercial Pt/C catalysts [7].
Chemicals
Solvents and substrates were purified by standard
procedures used for organometallic compounds [22].
Benzene was additionally dried by distillation from
LiAlH4 in a distillation column. It was stored over
molecular sieve 4 Å in an argon atmosphere in sealed
tubes.
Catalytic systems active in hydrogenation can also
be produced by reducing palladium(II) phosphine com-
plexes [4, 9, 16–21]. As distinct from the other above-
listed stabilizers, phosphine ligands undergo destruc-
N,N-Dimethyl formamide was held over anhydrous
tion via the stepped breaking of the P–R (R= Ar, Alk) copper sulfate until the formation of a green solution
bond during the formation and operation of hydrogena- and was twice distilled in vacuo (3 Torr) at a tempera-
tion catalysts in a hydrogen atmosphere or in the pres- ture not exceeding 42°C for drying and amine removal.
640

NANOSIZED HYDROGENATION CATALYST
641
Palladium bisacetylacetonate was obtained as transferred in an inert atmosphere to an appendix-type
1
described in [23]. H NMR data for this complex: reactor. Next, 2/3 of the solvent was removed in vacuo
δ(ëç) = 5.04 ppm (s, 1H); δ(ëç3) = 1.76 ppm (s, 6H). and diethyl ether was added until the formation of a
precipitate. The precipitate was washed with diethyl
ether in an argon atmosphere and was dried at 50°C in
vacuo (1 Torr). The product yield was 0.235 g. Elemen-
tal analysis data (%): Pd, 87.88; P, 3.59; C, 0.99; H,
0.25. The IR spectrum of the product exhibited no
bands between 4000 and 400 cm–1.
Phosphine was synthesized by dropping an alkali
solution to a toluene dispersion of red phosphorus
heated to 80°ë [24]. The resulting gas was purified of
diphosphine by passing it through a Tishchenko bottle
filled with a 50% alkali solution in DMSO; for remov-
ing the residual water, it was then passed through col-
umns packed with NaOH and P2O5. Phosphine was
condensed into benzene in an appendix-type reactor.
The phosphine concentration in benzene was deter-
mined by 31P NMR using H3PO4 as the external stan-
Catalytic Hydrogenation
Hydrogenation was carried out in a temperature-
controlled glass shaker at 30°C and an initial hydrogen
pressure of 1 atm in the presence of the catalytic system
Pd(acac)2–PH3 prepared in situ. A solution of phosphine
(0.3 × 10–5 mol) in benzene (1 ml) was added in drops
to a solution of Pd(acac)2 (0.00304 g, 1 × 10–5 mol) in
DMF (9 ml) contained in a temperature controlled
shaker under flowing hydrogen. The reaction mixture
was agitated for 5 min at room temperature. Next, the
temperature was raised to 80°C and the catalyst was
synthesized by agitating the mixture in a hydrogen
atmosphere for another 15 min. The resulting blackish
brown “solution” was cooled to 30°C, and the substrate
was introduced. Hydrogenation was carried out under
vigorous agitation to rule out the diffusion control of
the reaction. The progress of the reaction was moni-
tored volumetrically and by GLC. Hydrogenation at
other P/Pd ratios was carried out in the same way.
dard. NMR data for phosphine: δ = –238 ppm (q, JPH
=
188 Hz).
Reaction between Palladium Bisacetylacetonate
and Phosphine
The reaction was carried out in dry, deoxygenized
argon in a temperature-controlled vessel at various ini-
tial ratios of the reactants. The vessel was designed so
that it could be pumped and filled with an inert gas
before the reaction.
Phosphine (1.5 mmol) as a benzene solution (52 ml)
was added to a stirred solution of Pd(acac)2 (0.4566 g,
1.5 mmol) in benzene (55 ml) under flowing argon.
Two to three minutes after the beginning of the reac-
tion, the solution became turbid and turned into a
brown dispersion. Thereafter, the reaction mixture was
stirred for 24 h at room temperature. The filtrate was ana-
lyzed by UV spectroscopy at regular intervals. The pal-
ladium bisacetylacetonate concentration was derived
from the intensity of the absorption band at 330 nm
(ε330 = 10630 l mol–1 cm–1); the acetylacetone concentra-
tion, from the intensity of the absorption band at 290 nm
(ε290 = 5000 l mol–1 cm–1 for Hacac and 3090 l mol–1 cm–1
for Pd(acac)2). After the reaction, the precipitate was
collected on a Schott filter in an argon atmosphere,
washed with benzene and diethyl ether, and dried at
30°C in vacuo (1 Torr). The product yield was 0.22 g.
Elemental analysis data (%): Pd, 68.90; P, 24.66; C,
4.18; H, 2.25. The product was X-ray-amorphous. Its
IR spectrum showed absorption bands at 2296, 1067,
977, 725, and 672 cm–1.
Analytical Methods
IR spectra were recorded in the 4000–400 cm–1
range on an IFS25 spectrophotometer using KBr tech-
niques. The samples to be characterized, which were
either mulls in mineral oil thermally pretreated to
remove the traces of water or KBr pellets, were pre-
pared in a dry inert atmosphere in a glovebox.
NMR spectra were recorded on a VXR-500S Varian
1
pulsed spectrometer. H chemical shifts were deter-
mined relative to tetramethylsilane. 31P shifts were
determined relative to 85% H3PO4. Downfield shifts
were assigned positive values. The solutions to be char-
acterized by 31P NMR were sealed in coaxial inserts
prepumped and filled with argon.
The reaction was carried in this way for all initial
reactant ratios.
X-ray powder diffraction patterns from catalysts
were recorded on a DRON-3M diffractometer using
CuKα radiation.
Preparation of the Catalyst in a Hydrogen Atmosphere
The catalyst was prepared in a thermostated glass
shaker. A benzene solution (25 ml) of phosphine
(0.75 × 10–3 mol) was added in drops to a solution of
Pd(acac)2 (0.761 g, 2.5 × 10–3 mol) in DMF (90 ml) in a
hydrogen atmosphere (1 atm). Thus, the PH3 : Pd(acac)2
ratio (P/Pd) was 0.3. The reaction mixture was agitated
at 80°C for 15 min. After 2–5 min, it turned blackish
brown. After the reaction was complete, the resulting
TEM images were obtained with a BS-300 micro-
scope (Czech Republic). A drop of a solution was
placed onto a carbon-coated specimen grid and was
dried in an argon atmosphere. The imaging conditions
ruled out the melting or decomposition of the specimen
under the action of the electron beam.
The Pd(0) determination procedure was described in
dispersion was cooled to room temperature and was an earlier publication [25].
KINETICS AND CATALYSIS Vol. 48 No. 5 2007

642
SHMIDT et al.
Table 1. Styrene hydrogenation in the presence of the cata-
lytic system Pd(acac)2–0.3PH3
destruction under the reducing conditions, a process
converting part of the palladium into the oxidized state.
The data presented in Table 1 illustrate how the cat-
alytic properties of the Pd(acac)2–PH3 system in styrene
hydrogenation depend on the catalyst formation condi-
tions. The necessity of applying rather high tempera-
tures in this reaction stems from the fact that, due to its
chelate structure, Pd(acac)2 requires more severe condi-
tions to react with ç2 than does any other reported pal-
ladium(II) compound with oxygen-containing acido
ligands [27].
The specific styrene hydrogenation activity of the
Pd(acac)2–PH3 system as a function of P/Pd passes
through an extremum (Fig. 1). The highest hydrogena-
tion activity is observed at P/Pd = 0.3. As this ratio is
further increased, the reaction rate falls dramatically. At
P/Pd ≥ 1, phosphine acts as an inhibitor, as is clear from
a comparison between the activity of this system and
the activity of Pd black produced under the same con-
ditions. A similar trend was observed earlier for a pri-
mary phosphine as the modifier [21]. With triphe-
nylphosphine as the modifier, the alkene hydrogenation
activity of the catalyst peaks at P/Pd = 0.75. In the fam-
ily of palladium catalysts containing phosphine
ligands, hydrogenation activity increases in the follow-
ing order:
Catalyst formation conditions
Specific activity,
(mol H2)
(mol Pd)–1 min–1
time, min
temperature, °C
1
2
3
4
5
6
7
15
15
15
15
5
50
70
80
90
80
80
80
0
166
190
56
156
163
116
10
30
Note: Palladium content, C = 1 mmol/l; styrene : Pd = 870; sol-
Pd
vent, DMF; reaction temperature, T = 30°C; PH = 1 atm.
2
RESULTS AND DISCUSSION
Catalytic Properties of the Pd(acac)2–PH3 System
It was demonstrated earlier that palladium(II)
β-diketonate phosphine complexes show their catalytic
hydrogenation activity only after they are treated with a
reducing agent [4, 9, 21, 26]. The activity and selectiv-
ity of these catalysts depend strongly not only on the
reducing agent, the organophosphorus compound, and
P/Pd, but also on the order in which the reactants are
combined. Variations in the properties of the resulting
microheterogeneous palladium catalysts arise from
variations in the palladium nanocluster size, in the
nature of the ligand shell, and in the extent of phosphine
PPh3 (40) [9] < PHPh2 (140) [20]
< PH3 (190) < PH2Ph (280) [21].
Here, the parenthesized numbers are the specific activ-
ities (in (mol styrene) (mol Pd)–1 min–1) of the catalysts
with optimum P/Pd values produced in a hydrogen
atmosphere under identical conditions.
The promoting effect of PH3 on the palladium cata-
lysts is 4.75 times stronger than the effect of triphe-
nylphosphine and is 1.5 times weaker than the effect of
phenylphosphine.
Specific activity, (mol H2) (mol Pd)–1 (min)–1
200
The specific hydrogenation activity of the Pd(acac)2–
PH3 system can be increased by reducing the palladium
concentration. The highest catalytic activity in the sty-
rene hydrogenation reaction (330 mol (mol Pd)–1 min–1)
is reached at a palladium concentration of 5 × 10–4 mol/l.
The fact that the reaction rate as a function of the
Pd(acac)2 concentration passes through a maximum
(Fig. 2) indicates that active palladium particles com-
bine into less active aggregates and that the catalyst is
microheterogeneous. Apparently, extending the dura-
tion of catalyst formation in a hydrogen atmosphere
also impairs the aggregation resistance and, accord-
ingly, reduces the specific catalytic activity (Table 1).
180
160
140
120
100
80
60
40
20
Some data characterizing the behavior of the
Pd(acac)2–0.3PH3 system in the hydrogenation of a
number of substrates are presented in Table 2. This sys-
tem shows a rather high selectivity in the hydrogenation
of the triple bond and the nitro and carbonyl groups. For
example, phenylacetylene is successively hydroge-
nated to styrene and ethylbenzene and these reactions
are not accompanied by dimerization or oligomeriza-
0
0.5
1.0
1.5
2.0
P/Pd
Fig. 1. Catalytic activity of the Pd(acac) –PH system in
styrene hydrogenation as a function of the P/Pd ratio (C
1 mmol/l; substrate : Pd = 870; solvent, DMF; T = 30°ë;
PH = 1 atm).
2
3
=
Pd
2
KINETICS AND CATALYSIS Vol. 48 No. 5 2007

NANOSIZED HYDROGENATION CATALYST
643
Reaction rate, mmol l–1 min–1
tion. At a phenylacetylene conversion of 96%, the sty-
rene selectivity of the reaction is 94.4%. Nitrobenzene
is selectively reduced to aniline.
300
250
200
150
100
50
Reaction between Pd(acac)2 and Phosphine
in an Inert Atmosphere
Investigation of the interaction between the initial
components of the Pd(acac)2–PH3 system in hydrogen
and in an inert atmosphere provided information con-
cerning the nature and the formation of the catalytically
active particles.
0
1
2
3
4
5
6
7
8
CPd, mmol/l
Fig. 2. Styrene hydrogenation rate in the presence of the cat-
alytic system Pd(‡cac) –0.3PH as a function of palladium
The interaction between a palladium β-diketonate
complex and an organic phosphine in an inert atmo-
sphere can proceed via the following three routes,
depending on the nature of the phosphine and on reac-
tion conditions: (1) phosphine coordination, which
causes the acetylacetonate ligands to switch from the
O,O-chelating mode to another binding mode [28, 29];
(2) redox reaction [20, 30, 31]; and (3) replacement of
the β-diketonate ligands by phosphorus-containing
groups, yielding polynuclear palladium complexes
containing bridging phosphide or phosphinidene
ligands [20, 21]. Since phosphine has the properties of
both a Lewis base and a Brønsted acid and is a reducer,
it is pertinent to consider the interaction between the
initial components of the catalytic system in an inert
atmosphere.
According to UV spectroscopic and 1H NMR data,
the reaction between Pd(acac)2 and phosphine in ben-
zene at P/Pd = 0.5–2.0 yields acetylacetone and a
brown precipitate. At P/Pd = 0.5, the reaction yields
49% acetylacetone and 52% of the Pd(acac)2 remains
unreacted (Table 3). At P/Pd = 1 or 2, the resulting
acetylacetone concentration is close to the theoretical
value (89 and 100%, respectively). Therefore,
Pd(acac)2 and phosphine react in nearly equimolar pro-
portions. According to 31P NMR data, the solution
resulting from the reaction at P/Pd = 2 contains free
phosphine (δ = –238 ppm, JPH = 188 Hz) and some of
the phosphine hydrogen atoms are not involved in the
exchange. If all of these atoms were replaced, then, at
P/Pd = 0.5, the acetylacetone yield would be 75% on
the Pd(‡cac)2 basis.
2
3
concentration (solvent, DMF; T = 30°C; PH = 1 atm;
2
–3
amount of styrene, 8.7 × 10 mol).
The brown precipitate that forms at P/Pd = 1 is
unstable in air, and its composition can be described by
the formula Pd1P1.2C0.5H3.3 (sample I). The IR spectrum
of sample I shows an absorption band at 2296 cm–1,
which is characteristic of the stretching vibrations of
the P–H bond (νê–ç) [32], and absorption bands at
1067, 978, 725, and 673 cm–1, which are assignable to
the bending vibrations of the same bond (Fig. 3). The
absence of absorption bands characteristic of the C=O
and C=C stretchings in the O,O-chelating acac ligand at
the palladium atom (1560 and 1516 cm–1) proves that
the acetylacetonate ligands are replaced by phospho-
rus-containing groups. The presence of the P–H bend-
ing bands indicates that coordinated phosphine and/or
phosphide (PH2) fragments are present. The low solu-
bility of sample I in both polar and nonpolar solvents
did not allow us to record the 31P NMR spectrum, so we
cannot determine what particular phosphorus-contain-
ing groups (phosphide (PH2) or phosphinidene (PH)
fragments or palladium phosphides of various compo-
sitions (PdxP)) are present in sample I. The scarce infor-
mation available on the outcomes of the reactions
between phosphine and palladium(II) complexes sug-
gests that these reactions can yield various products.
For example, the reaction between PH3 and the com-
plex (PhCN)2PdCl2 in dry dioxane or chloroform yields
Table 2. Hydrogenation of various substrates in the presence of the catalytic system Pd(acac)2–0.3PH3
Reaction rate, (mol H2)
Substrate
Styrene
CPd, mmol/l
Products
Yield, %
(mol Pd)–1 min–1
1
1
5
5
190
59
21
1
Ethylbenzene
Ethylbenzene
Aniline
100
100
100
95
Phenylacetylene
Nitrobenzene
Benzaldehyde
Benzyl alcohol,
toluene
5
Note: Substrate : Pd = 870; solvent, DMF; T = 30°C; catalyst preparation conditions: hydrogen atmosphere, 80°C, 15 min.
KINETICS AND CATALYSIS Vol. 48 No. 5 2007

644
SHMIDT et al.
Table 3. Reaction between palladium bisacetylacetonate and phosphine in an inert atmosphere
Concentration, %
P/Pd
0.5
Reaction time, min
in benzene
in DMF
Pd(acac)2
Hacac
Pd(acac)2
Hacac
5
30
78
16
35*
46**
49
85
75
71
62
66
49
40
16
54
38
27
0
15
26
30
42
34
54
65
91
47
64
75
98
59*
51**
52
60
1 day
5
1.0
2.0
65
34
30
49
56
60
38
66
1 day
5
16
89
38
54
30
13*
5**
2
86*
98**
104
60
1 day
Note: C = 4.2 mmol/l; T = 20°C; solvent, benzene or DMF; concentration determination error, 10%.
Pd
* The reaction time is 40 min.
** The reaction time is 80 min.
[Pd(PH2)2]n, an insoluble amorphous compound [33]. plexes containing phosphinidene bridges (µ-êç) and a
At the same time, palladium(II) chloride dissolved in small proportion of coordinated phosphine
dioxane reacts with phosphine to yield the palla- ([Pd(µ-êç)]n(PH3)0.2n). For P/Pd ≤ 1, unreacted
dium(II) phosphide derivative (PdCl)3P as a brown pre- Pd(acac)2 remains in the reaction system.
cipitate [34].
The reaction between Pd(acac)2 and phosphine in
According to X-ray diffraction data, sample I is DMF, an inert medium, also involves the substitution
X-ray-amorphous. The amorphous halo observed at of phosphinidene groups for acetylacetonate ligands
small reflection angles (diffuse peak at d/n = 13.556 Å (Table 3). The addition of phosphine to the carbonyl
(2θ = 6.52°)) is evidence that sample I has a low-order group, a process typical of aldehydes, was not
structure typical of polymer compounds [35]. Note observed [32].
that, in the diffraction patterns from palladium phos-
A comparison between the above data and earlier
phides (Pd3P, Pd6P, Pd4.8P, Pd5P2, etc.), the reflection
data for the reactions of Pd(acac)2 with phenylphosphine
111 occurs in the 2θ = 35°–45° range [36]. The diffrac-
[21] and diphenylphosphine [20] demonstrates that a
tion pattern from sample I shows neither reflections
common feature of all of the reactions examined is the
from a crystalline phase nor an amorphous halo in this
replacement of the acido ligands at the palladium atom
region. Therefore, no palladium phosphides result from
by phosphorus-containing groups. However, these reac-
tions yield different products. The reaction between pal-
the reaction between Pd(acac)2 and PH3. Since Pd(acac)2
and phosphine react in nearly equimolar proportions to
ladium acetylacetonate and an equimolar amount of
yield ~2 mol of acetylacetone per mole of phosphine,
diphenylphosphine yields the dinuclear complex
the resulting oligomers are most likely polynuclear pal-
[Pd(acac)(µ-PPh2)]2; at P/Pd = 2, a redox reaction takes
ladium complexes with bridging phosphinidene ligands
place, yielding tetraphenyldiphosphine and the trinuclear
(µ-êç). The formation of oligomer structures is possi-
palladium cluster [Pd3(µ-PPh2)3PPh2(PHPh2)2]. Even in
ble owing to the ability of phosphinidene ligands to
an inert atmosphere, the reaction between Pd(acac)2 and
coordinate not only to two metal atoms, but even to
phenylphosphine (P/Pd = 0.3–2) yields nanosized par-
three or four atoms [37–39]. According to IR spectro-
ticles consisting of polynuclear palladium complexes
scopic and elemental analysis data, the polynuclear pal-
with bridging phenylphosphinidene ligands and coordi-
ladium complexes contain coordinated phosphine mol-
nated PH2Ph molecules. In this reaction, as distinct
ecules along with µ-êç ligands.
from the reaction involving phosphine, the acetylaceto-
Thus the reaction between Pd(acac)2 and PH3 in an nate ligands are completely replaced only at P/Pd ≥ 2
argon atmosphere involves replacement of the acido and the P/Pd ratio in the polynuclear palladium com-
ligands at the palladium atom and yields acetylacetone plexes is close to 2 (equals 1.7). As a consequence, at a
and randomly structured polynuclear palladium com- given P/Pd molar ratio, the unreacted Pd(acac)2 percent-
KINETICS AND CATALYSIS Vol. 48 No. 5 2007

NANOSIZED HYDROGENATION CATALYST
645
Transmittance, %
80
70
60
50
40
30
20
10
0
–10
4000
3500
3000
2500
2000
1750
1500
1250
1000
750
500
ν, cm–1
Fig. 3. IR spectrum of the product of the reaction between Pd(‡cac) and phosphine (sample I as mineral-oil mulls).
2
age is higher for the reaction involving phenylphos- reaction system Pd(acac)2–0.3PH3–ç2 shows weak res-
phine than for the reaction involving phosphine. We onances at δ = 1.6 ppm (d, Jêç = 635 Hz) and δ =
believe that the difference between the behaviors of 0.6 ppm (s), with their intensity ratio equal to 1 : 0.3.
phenylphosphine and phosphine arises from the lower These resonances are due to phosphorous and phospho-
basicity of the latter.
ric acids, respectively. No other signals were detected
between +500 and –500 ppm. A plausible explanation
of the absence of any strong signals from reaction prod-
ucts is that the products form as a colloidal dispersion
and can hardly be detected by NMR spectroscopy
because of the very short relaxation time of the colloid
particles [40].
Formation and Nature of the Hydrogenation Catalyst
Based on Pd(acac)2 and PH3
The formation of the hydrogenation catalyst based
on Pd(acac)2 and PH3 can arbitrarily be divided into
two steps, namely, the reaction between these compo-
nents in a hydrogen atmosphere at room temperature
(5 min) and the reduction of the resulting system by
hydrogen at 80°C (15 min).
Five minutes after addition of phosphine to a
Pd(acac)2 solution at room temperature in a hydrogen
atmosphere, the 1H NMR spectrum of the reaction sys-
tem Pd(acac)2–0.3PH3 shows weak resonances at δ =
2.02 and 2.08 ppm along with the resonances due to the
acac ligands coordinated to palladium (δ(ëç) =
5.46 ppm, s; δ(ëç3) = 1.99 ppm, s). As will be demon-
strated below, these weak resonances are due to the
methyl protons of the enol and keto forms of acetylac-
etone. Since Pd(acac)2 is irreducible by hydrogen under
the above-specified mild conditions, the minor amounts
of acetylacetone result from the side reaction between
Pd(acac)2 and phosphine.
Transmission electron microscopic examination of
the reaction system demonstrated that the reaction of
the components of the catalytic system with hydrogen
yields a microheterogeneous system dominated by
high-contrast particles 7.3–7.8 nm in diameter (Fig. 4).
Extending the duration of catalyst formation in a hydro-
gen atmosphere from 15 to 30 min results in a better
structured system (Fig. 5). The observed loose,
branched structure is typical of fractal clusters [12, 41].
These clusters result from the self-organization (aggre-
gation) of nanoparticles of similar sizes in the absence
of stabilizing factors. The formation of these fractal
clusters in the preparation of microheterogeneous pal-
ladium-based hydrogenation catalysts in the presence
of phosphine ligands was observed for the first time.
With a tertiary, secondary, or primary phosphine, the
reduction of palladium(II) complexes yields nanoparti-
cles with a higher aggregation resistance.
Treating the catalytic system Pd(acac)2–0.3PH3 with
hydrogen at 80°ë yields a blackish brown solution. The
The composition of the black precipitate isolated
1H NMR spectrum of this reaction system shows no from the reaction mixture (sample II) can be described
signals from the acetylacetonate ligands of Pd(acac)2 by the formula Pd7.5P1.0C0.7H2.1. Its diffraction pattern
and exhibits signals from the keto and enol forms of shows peaks assignable to the phosphides Pd6P (d/n
acetylacetone: for the enol form, δ(ëç) = 5.57 ppm (s), 2.696, 2.583, 2.368, 2.291, 2.256, 2.232, 2.093, and
δ(ëç3) = 1.93 ppm (s), and δ(éç) = 16.24 ppm; for the 1.998 Å), Pd4.8P (d/n 2.439, 2.327, 2.256, 2.232, 2.122,
keto form, δ(ëç2) = 3.65 ppm (s) and δ(ëç3) = and 1.998 Å), and Pd5P2 (d/n 2.734, 2.696, 2.439,
2.08 ppm (s). The stoichiometric formation of acetylac- 2.291, 2.256, 2.232, 2.122, 2.093, 1.858, and 1.432 Å)
etone is confirmed by UV spectroscopic data. After the and crystalline palladium (d/n 2.256, 1.950, and 1.375 Å)
reaction is complete, the 31P NMR spectrum of the [36]. The dominant phases are Pd6P and Pd4.8P. As
KINETICS AND CATALYSIS Vol. 48 No. 5 2007

646
SHMIDT et al.
nPd(acac)2 + nPH3 = [Pd(µ-PH)]n + 2nHacac;
(2) reduction of part of the Pd(acac)2 with hydrogen,
yielding Pd(0) clusters,
Pd(acac)2 + H2 = Pd(0) + 2Hacac,
nPd(0)
Pd(0)n;
(3) reaction between Pd(acac)2 and palladium com-
plexes with phosphorus-containing ligands ([Pd(µ-PH)]n)
under reducing conditions, yielding various palladium
phosphides,
50 nm
[Pd(µ-PH)]n + 2nPd(acac)2 + 1.5nH2
= nPd3P + 4nHacac,
(‡)
Percentage of particles, %
16
Pd3P + 3Pd = Pd6P.
Note also that phosphine itself can reduce transition
metal compounds [24]. The reducing action of phos-
phine in catalyst formation is confirmed by the forma-
tion of small amounts of phosphorous acid:
(b)
14
12
10
8
3Pd(acac)2 + PH3 + 3H2O = 3Pd(0) + H3PO3 + 6Hacac.
6
Water, which is necessary for this redox process,
may be present as an impurity in the solvent, even
though the solvent was dried. The resulting phospho-
rous acid is also a reducing agent, and it undergoes oxi-
dation into phosphoric acid. Obviously, we cannot rule
out the possibility that phosphorous and phosphoric
acids result from the partial oxidation of phosphine into
phosphorus(III) and phosphorus(V) oxides by traces of
oxygen in the reaction system followed by the hydra-
tion of these oxides. However, in a special-purpose
experiment, phosphine was heat-treated in a hydrogen
atmosphere at 80°ë and the 31P NMR spectrum of the
product showed only a signal from free phosphine (δ =
–238 ppm, q, Jêç = 188 Hz).
Transition metal phosphides are comparatively unre-
active. Active hydrogenation catalysts are Pd(0) clusters
or palladium hydride clusters resulting from the former
during the hydrogenation process [9, 20, 21]. The fol-
lowing question arises here: If most of the palladium
turns into palladium phosphides during the formation
of the Pd(acac)2–0.3PH3 system, then what is the origin
of the strong catalytic effect of this system in the hydro-
genation of various substrates?
4
2
0
3.0 4.3 5.5 6.7 7.9 9.1 10.3 11.6 12.8 14.0
d, nm
Fig. 4. (a) Electron micrograph of the Pd(‡cac) –0.3êç
2
3
system treated with hydrogen at 80°ë for 15 min. (b) Parti-
cle size distribution.
determined by chemical analysis, the Pd(0) content of
sample II is no higher than 31%. Knowing the total palla-
dium and phosphorus contents, the proportion of reduced
palladium, and the compositions of the palladium phos-
phides, one can express the composition of sample II
as [{Pd(0)}2.33{Pd6P}0.40{Pd4.8P}0.55{Pd5P2}0.026].
The source of carbon and hydrogen found in this
sample is most likely dimethyl formamide, the sol-
vent in which the synthesis was carried out. In view
of this, the refined formula of sample II appears as
[{Pd(0)}2.33{Pd6P}0.40{Pd4.8P}0.55{Pd5P2}0.026{DMF}0.3].
Thus, the reduction of palladium bisacetylacetonate
with hydrogen in the presence of phosphine yields pal-
ladium phosphides of various compositions, Pd(0), and
phosphorous and phosphoric acids. These product
composition data for the conversion of the Pd(acac)2–
0.3PH3 catalytic system in hydrogen suggest the fol-
lowing product formation scheme.
According to electron microscopic data, the particle
size of the palladium black resulting from the reduction
of palladium bisacetylacetonate with hydrogen in the
absence of phosphine is 25–30 nm, almost four time as
large as the particle size of the palladium catalyst exam-
ined. Therefore, one of the origins of the promoting
effect of phosphine is that this compound favors the for-
mation of a more finely dispersed system. The fraction
of surface atoms is 20–15% for metal particles 8–10 nm
The reaction between the components of this system
can proceed at least via three routes:
(1) formation of a complex between palladium in diameter and has a much smaller value of 3–2% for
bisacetylacetonate and phosphine followed by the a particle diameter of 25–30 nm [42]. Thus, the percent-
replacement of the bisacetylacetonate ligands by phos- age of surface atoms can increase by a factor of about
phorus-containing groups, yielding polynuclear com- 7, as does the reaction rate under the action of phos-
plexes of the [Pd(µ-PH)]n type,
phine introduced into the catalytic system. The promot-
KINETICS AND CATALYSIS Vol. 48 No. 5 2007

NANOSIZED HYDROGENATION CATALYST
647
However, the high extent of dispersion alone cannot
be fully responsible for the high hydrogenation effi-
ciency of the nanosized palladium catalysts prepared in
the presence of phosphine. The conversion of the palla-
dium phosphine complexes under the action of dihy-
drogen is a complicated multistep process. Common
features of the formation, in a hydrogen atmosphere, of
the Pd(acac)2-based hydrogenation catalysts involving
different phosphines are the formation of a microheter-
ogeneous system and the oxidation of part of the palla-
dium as a result of the decomposition of the phospho-
rus-containing ligands. The proportions of the oxidized
and reduced palladium particles depend on the rate
ratios between the reduction of the Pd(II) complexes,
the aggregation of the resulting reduced palladium par-
ticles, and the decomposition of the phosphorus-con-
taining groups in the coordination sphere of Pd(0). All
these processes exert an effect on the nature and cata-
lytic properties of the resulting nanoparticles.
100 nm
(‡)
Percentage of particles, %
18
16
(b)
14
12
10
8
ACKNOWLEDGMENTS
This study was supported by the Zamaraev Interna-
tional Fund and by grant no. III-02-000/107 from
Irkutsk State University.
6
4
2
REFERENCES
0
1. Moiseev, I.I. and Vargaftik, M.N., Zh. Obshch. Khim.,
5.0 6.7 8.3 10.0 11.7 13.3 15.0 16.7
2002, vol. 72, no. 4, p. 550.
d, nm
2. Aiken, J.D. and Finke, R.G., J. Mol. Catal., 1999,
vol. 145, nos. 1–2, p. 1.
Fig. 5. (a) Electron micrograph of the Pd(‡cac) –0.3êç
2
3
system treated with hydrogen at 80°ë for 30 min. (b) Parti-
3. Bonnemann, H., Braun, G., Brijoux, W., Brinkmann, R.,
Tilling, A.S., Seevogel, K., and Siepen, K., J. Orga-
nomet. Chem., 1996, vol. 520, p. 143.
cle size distribution.
4. Shmidt, F.K., Belykh, L.B., Umanets, V.A., Cheren-
kova, T.V., and Belonogova, L.N., Kinet. Katal., 1999,
vol. 40, no. 3, p. 466 [Kinet. Catal. (Engl. Transl.),
vol. 40, no. 3, p. 418].
5. Vargaftik, M.N., Zagorodnikov, V.P., Stolyarov, I.P.,
Likholobov, V.A., Chuvilin, A.L., Zaikovskii, V.I.,
Kochubei, D.I., Timofeeva, G.I., Zamaraev, K.I., and
Moiseev, I.I., Dokl. Akad. Nauk SSSR, 1985, vol. 284,
no. 4, p. 896.
ing effect of phosphine that is associated with an
increase in the extent of dispersion of the catalysts pro-
vides a satisfactory explanation for the observed exper-
imental data under the assumption that the nanoparticle
surface contains Pd(0) atoms. Therefore, the most
likely model of a nanoparticle of the microheteroge-
neous hydrogenation catalyst is a palladium phosphide
core on which Pd(0) clusters are immobilized.
6. Vargaftik, M.N., Kozitsyna, N.Yu., Cherkashin, N.V.,
Rudyi, R.I., Kochubei, D.I., Novgorodov, B.N., and
Moiseev, I.I., Kinet. Katal., 1998, vol. 39, no. 6, p. 806
[Kinet. Catal. (Engl. Transl.), vol. 39, no. 6, p. 740].
7. Bonnemann, H. and Brijoux, W., in Catalytically Active
Metal Powders and Colloids, Furstner, A., Ed., New
York: VCH, 1996, p. 339.
Further studies are necessary to elucidate the role of
the electron factor in the catalytic activity of the system.
Thus, the above experimental data suggest that the
promoting effect of phosphine on palladium catalysts is
not weaker and, in some cases, is even stronger than the
effect of organic phosphines. The promoting effect of
phosphine is in part due to the fact that phosphine
ensures the formation of a microheterogeneous system
that is dispersed more finely than Pd black obtained
from Pd(acac)2 under the same conditions. The result-
ing nanoparticles have a core consisting of various pal-
ladium phosphides, such as Pd6P, Pd4.8P, and Pd5P2, on
which Pd(0) clusters are immobilized.
8. Schmid, G., Polyhedron, 1988, vol. 7, p. 2321.
9. Shmidt, F.K., Belykh, L.B., and Goremyka, T.V., Kinet.
Katal., 2003, vol. 44, no. 5, p. 683 [Kinet. Catal. (Engl.
Transl.), vol. 44, no. 5, p. 623].
10. Aiken, J.D. and Finke, R.G., J. Am. Chem. Soc., 1998,
vol. 120, p. 9545.
11. Maksimov, G.M., Zaikovskii, V.I., Matveev, K.I., and
Likholobov, V.A., Kinet. Katal., 2000, vol. 41, no. 6,
KINETICS AND CATALYSIS Vol. 48 No. 5 2007

648
SHMIDT et al.
p. 925 [Kinet. Catal. (Engl. Transl.), vol. 41, no. 6, 27. Mironova, L.V., Levkovskii, Yu.S., Belykh, L.B.,
p. 844].
Lobza, G.V., Ivanova, N.A., Ratovskii, G.V., Dubin-
skaya, E.I., and Shmidt, F.K., Koord. Khim., 1985,
vol. 11, no. 12, p. 1689.
12. Pomogailo, A.D., Rozenberg, A.S., and Uflyand, I.E.,
Nanochastitsy metallov v polimerakh (Metal Nanoparti-
cles in Polymers), Moscow: Khimiya, 2000.
28. Kawaguchi, S., Coord. Chem. Rev., 1986, vol. 70, p. 51.
29. Ratovskii, G.V., Tyukalova, O.V., Tkach, V.S., and
Shmidt, F.K., Zh. Obshch. Khim., 1996, vol. 66, no. 11,
p. 1791.
30. Ratovskii, G.V., Belykh, L.B., Burlakova, O.V., and
Shmidt, F.K., Zh. Obshch. Khim., 1989, vol. 59, no. 12,
p. 2784.
31. Amotore, C., Carre, E., Jutand, A., M’Barki, M.A., and
Meyer, G., Organometallics, 1995, vol. 14, no. 12,
p. 5605.
32. Purdela, D. and Vilceanu, R., Chimia compusilor
organici ai fosforului si ai acizilor luj, Timishoara,
Romania, 1965.
33. Makarova, G.A., Zudin, V.N., Pashis, A.V., Kochubei, D.I.,
and Likholobov, V.I., XVI Vses. Chugaevskoe soveshch.
po khimii kompleksnykh soedinenii (XVI All-Union
Chugaev Conf. on the Chemistry of Coordination Com-
pounds), Krasnoyarsk, 1983, p. 608.
34. Dorfman,Ya.A., Polibetova, G.S., Aleshkova, M.M., and
Borangazieva, A.K., Koord. Khim., 1999, vol. 25, no. 7,
p. 520.
13. Toshima, N. and Yonezawa, T., New J. Chem., 1998,
p. 1179.
14. Stolyarov, I.P., Gaugash, Yu.V., Kryukova, G.N.,
Kochubei, D.I., Vargaftik, M.N., and Moiseev, I.I., Izv.
Akad. Nauk, Ser. Khim., 2004, no. 6, p. 1147.
15. Schmid, G., Emde, S., Maihack, V., Zaika, M.W., and
Peschel, St., J. Mol. Catal., 1996, vol. 107, p. 950.
16. Berenblyum, A.S., Zh. Vses. Khim. O–va im. D. I. Men-
deleeva, 1984, vol. 32, no. 1, p. 82.
17. Sisak, A., Ungvary, F., and Kiss, G., J. Mol. Catal., 1983,
vol. 18, no. 2, p. 223.
18. Bakunina, T.I., Zinchenko, S.V., Khutoryanskii, V.A.,
Kolomeichuk, V.N., and Shmidt, F.K., Metallokompleks-
nyi kataliz (Catalysis by Metal Complexes), Irkutsk:
Irkutsk. Gos. Univ., 1989.
19. Cairns, G.R., Cross, R.J., and Stirling, D., J. Mol. Catal.,
2001, vol. 172, p. 207.
20. Belykh, L.B., Goremyka, T.V., Rokhin, A.V.,
Belonogova, L.N., and Shmidt, F.K., Kinet. Katal., 2004,
vol. 45, no. 3, p. 411 [Kinet. Catal. (Engl. Transl.),
vol. 45, no. 3, p. 385].
21. Belykh, L.B., Goremyka, T.V., Belonogova, L.N., and
Schmidt, F.K., J. Mol. Catal., 2005, vol. 231, nos. 1–2,
p. 53.
35. Kitaigorodskii, A.I., Rentgenostrukturnyi analiz (X-ray
Structure Determination), Moscow: Tekhteorizdat,
1950.
36. Hanawalt Search Manual: Inorganic Phases, Int. Center
for Diffraction Data, 1992, p. 1.
22. Gordon, A.J. and Ford, R.A., The Chemist’s Companion:
A Handbook of Practical Data, Techniques, and Refer-
ences, New York: Wiley, 1972.
37. Mathey, F., Angew. Chem., 1987, vol. 99, no. 4, p. 285.
38. Belyakova, O.A. and Slovokhotov, Yu.L., Izv. Akad.
Nauk, Ser. Khim., 2003, no. 11, p. 2175.
23. US Patent 3474464, 1969.
39. Fenske, D. and Simon, F., Angew. Chem., 1997, vol. 109,
no. 3, p. 240.
24. Osadchenko, I.M. and Tomilov, A.P., Usp. Khim., 1969,
vol. 38, no. 6, p. 1089.
40. Günther, H., NMR Spectroscopy: An Introduction,
Chichester: Wiley, 1980.
25. Shmidt, A.F. and Mametova, L.V., Kinet. Katal., 1996,
vol. 37, no. 3, p. 431 [Kinet. Catal. (Engl. Transl.),
vol. 37, no. 3, p. 406].
41. Shmidt, F.K., Fraktal’nyi analiz v fiziko-khimii getero-
gennykh sistem i polimerov (Fractal Analysis in the
Physical Chemistry of Heterogeneous Systems and
Polymers), Irkutsk: Irkutsk. Gos. Univ., 2001.
26. Belykh, L.B., Goremyka, T.V., Skripov, N.I., Umanets,V.A.,
and Shmidt, F.K., Kinet. Katal., 2006, vol. 47, no. 3,
p. 373 [Kinet. Catal. (Engl. Transl.), vol. 47, no. 3, 42. Gubin, S.P., Khimiya klasterov (Chemistry of Clusters),
p. 367].
Moscow: Nauka, 1987.
KINETICS AND CATALYSIS Vol. 48 No. 5 2007