Nature of the modifying action of white phosphorus on the properties of nanosized hydrogenation catalysts based on bis(dibenzylideneacetone)palladium(0)

Article abstract of DOI:10.1134/S002315841105003X

The catalytic properties and nature of the nanoparticles forming in the system based on Pd(dba)₂ and white phosphorus are reported. A schematic mechanism is suggested for the formation of nanosized palladium-based hydrogenation catalysts. The m

Full text of DOI:10.1134/S002315841105003X

ISSN 0023ꢀ1584, Kinetics and Catalysis, 2011, Vol. 52, No. 5, pp. 702–710. © Pleiades Publishing, Ltd., 2011.
Original Russian Text © L.B. Belykh, N.I. Skripov, L.N. Belonogova, V.A. Umanets, T.P. Stepanova, F.K. Schmidt, 2011, published in Kinetika i Kataliz, 2011, Vol. 52, No. 5, pp. 718–726.
Nature of the Modifying Action of White Phosphorus
on the Properties of Nanosized Hydrogenation Catalysts Based
on Bis(dibenzylideneacetone)palladium(0)
L. B. Belykh*, N. I. Skripov, L. N. Belonogova, V. A. Umanets, T. P. Stepanova, and F. K. Schmidt
Irkutsk State University, Irkutsk, 664033 Russia
eꢀmail: belykh@chem.isu.ru
Received December 17, 2010
Abstract—The catalytic properties and nature of the nanoparticles forming in the system based on Pd(dba)2
and white phosphorus are reported. A schematic mechanism is suggested for the formation of nanosized palꢀ
ladiumꢀbased hydrogenation catalysts. The mechanism includes the formation of palladium nanoclusters via
the interaction of Pd(dba)2 with the solvent (N,Nꢀdimethylformamide) and substrate and the formation of
palladium phosphide nanoparticles. The inhibiting effect exerted by elemental phosphorus on the catalytic
process is due to the conversion of part of the Pd(0) into palladium phosphides, which are inactive in hydroꢀ
genation under mild conditions, and the formation of mainly segregated palladium nanoclusters and pallaꢀ
dium phosphide nanoparticles. By investigating the interaction between Pd(dba)2 and white phosphorus in
benzene, it has been established that the formation of palladium phosphides under mild conditions consists
of the following consecutive steps: Pd(0)
→
PdP2
→
Pd5P2
→
Pd3P. It is explained why white phosphorus can
produce diametrically opposite effects of on the catalytic properties of nanosized palladiumꢀbased hydrogeꢀ
nation catalysts, depending on the nature of the palladium precursor.
DOI: 10.1134/S002315841105003X
A central problem of catalytic chemistry is that of and P4 and, accordingly, the variation of the nature of
elucidating the nature, regularities of formation, and the resulting nanoparticles.
the ways of functioning and deactivation of catalytiꢀ
For chemical modeling of particular stages of the
cally active particles. Without understanding these
formation of palladiumꢀbased hydrogenation catalysts
issues, it is impossible to devise a purposeful approach
modified with elemental phosphorus, we studied the
to design of highꢀefficiency catalytic systems and to
interaction between bis(dibenzylideneacetone)pallaꢀ
preparation of new catalysts. Earlier, we demonstrated
dium(0) (Pd(dba)2) and white phosphorus and the catꢀ
that not only phosphines, but also elemental phosphoꢀ
alytic properties of the resulting system. Pd(dba)2 was
rus can serve as an effective promoter for nanosized
chosen for the following reasons: firstly, this precursor
palladiumꢀbased hydrogenation catalysts prepared
needs no reduction in the preparation of catalytically
under the action of molecular hydrogen [1–3]. The
active particles; secondly, dibenzylideneacetone is the
introduction of white phosphorus prior to the reducꢀ
most labile among the known ligands and can readily
π
tion of the palladium precursor PdX2 raises, by a factor
of 4 to 9 at P : Pd = 0.3 mol/mol, the turnover freꢀ
quency (X = acac, OAc) and/or turnover number (X =
acac, OAc, Cl) of the reaction catalyzed by the resultꢀ
ing nanoparticles. The strong promoting effect
observed at small P : Pd ratios (X = acac) is due to the
increase in the degree of dispersion of the catalyst, for
which the following model was suggested: the core of
the nanoparticle consists of palladium phosphides,
mainly Pd6P, and the surface of the nanoparticles conꢀ
sists of hydrogenationꢀactive Pd(0) clusters. However,
depending on the acido ligand in the precursor, on the
reducing agent, and on the modifier concentration,
elemental phosphorus exerts either a promoting or an
inhibiting effect on palladium catalysts [3]. We believe
be displaced from the coordination sphere of the
metal. At the same time, unlike phosphines, this
ligand does not undergo destruction in the coordinaꢀ
tion sphere of palladium.
EXPERIMENTAL
Solvents (benzene and
N,Nꢀdimethylformamide
(DMF)) were purified by standard methods used in
organometallic chemistry [4]. For dehydration and
amine removal, DMF was held over anhydrous copper
sulfate until the formation of a green solution and was
distilled twice in vacuo (8 Torr) at a temperature not
higher than 42 С. Benzene was distilled from LiAlH4
°
that this is due to the variation of the ratios between the for deeper dehydration and was stored in an argon
rate of precursor reduction to Pd(0), the Pd(0) clusterꢀ atmosphere over molecular sieve 4A in sealed tubes. As
ing rate, and the rate of the interaction between Pd(0) determined by the Fischer method [5], the water conꢀ
702

NATURE OF THE MODIFYING ACTION OF WHITE PHOSPHORUS
703
tent of benzene was 1.1
was 0.8 mol/l.
×
10–3 mol/l and that of DMF
The reaction between bis(dibenzylideneacetone)palꢀ
ladium(0) and white phosphorus was carried out at difꢀ
ferent reactant ratios in a benzene medium under dry,
deoxygenated argon in a fingerꢀshaped vessel.
Dibenzylideneacetone (dba) was synthesized by
reacting benzaldehyde with acetone [6] followed by
1
the recrystallization of the product from ethanol. H
Example 1. A benzene solution of white phosphoꢀ
, ppm): 7.76 (2НA, d, 3J = 16 Hz) and
rus (9.4 ml, 18.8
×
10–4 mol) was added to a burgundy
NMR (CDCl3,
7.11 (2НВ, d, 3J = 16 Hz). 13C NMR (CDCl3,
189.4 (CO), 143.8 (C1, d, 1JСН = 154 Hz), 125.8 (C2,
δ
solution of Pd(dba)2 (0.5405 g, 9.4
×
10–4 mol) in 210 ml
δ
, ppm):
of benzene. As the solution of elemental phosphorus
was being added, the color of the solution changed
quickly from burgundy to black. In 10–20 s, a black
solid precipitated and the solution turned yellow. After
the completion of the reaction, the solution was sepaꢀ
rated from the precipitate by decantation. The soluꢀ
tion was analyzed by 31P NMR and UV spectroscopy.
1
d, JСН =158 Hz), 135.9 (1C, Ph), 129.7 (2C, ꢀPh),
о
130.1 (2C,
mꢀPh), and 131.4 (1C,
p
HB
C5
ꢀPh).
HB
C1
O
C3
C2
Ph
C4
Ph
The precipitate (sample
(10 ml) three times and was dried in vacuo (30
Torr). Yield: 0.16 g. Sample was Xꢀrayꢀamorphous.
It was crystallized by heat treatment at 400 in an
inert atmosphere in a sealed tube for 4 h and was then
cooled slowly. The Xꢀray diffraction pattern of crystalꢀ
lized sample 1 showed the following reflections (
I0 ): 4.004(46), 3.340(15), 2.998(31), 2.930(65),
1
) was washed with benzene
HA
HA
°
С,
1
1
Bis(dibenzylideneacetone)palladium(0) (Pd(dba)2)
was synthesized by PdCl2 reduction with methanol in
the presence of sodium acetate and dibenzylideneaceꢀ
°
С
tone [7, 8]. Dibenzylideneacetone (3.4500 g, 1.472
×
×
10–2 mol), sodium acetate trihydrate (4.8525 g, 3.568
d/n, Å
10–2 mol), and methanol (113 ml) were placed in a
(I/ )
2.878(94), 2.722(100), 2.515(29), 2.448(15), 2.284(32),
2.122(31), 2,111(37), 2.080(38), 2.052(38), 2.002(35),
1.982(15), 1.900(23), 1.838(33), 1.795(31), 1.735(23),
1.689(20), 1.489(15), 1.470(25), 1.367(15), 1.349(11),
and 1.308(11) (hereafter, the numbers in parentheses
are relative reflection intensities). The reflections at
2.930, 2.878, 2.722, 2.515, 2.052, 2.002, 1.900, 1.689,
1.470, and 1.349 Å are consistent with the Xꢀray difꢀ
fraction data for the palladium phosphide PdP2 [10].
The reflections at 3.340, 2.722, 2.515, 2.448, 2.284,
2.122, 2.111, 2.080, 1.982, 1.838, 1.735, 1.470, 1.367,
1.349, and 1.308 Å are consistent with the Xꢀray difꢀ
fraction data for the palladium phosphide Pd5P2 (PDF
19ꢀ887) [11]. The reflections at 4.004, 2.998, and
1.489 Å were not assigned to any phase.
twoꢀneck roundꢀbottom flask in an argon atmosphere.
The mixture was stirred at 50
PdCl2 (0.7875 g, 4.434
resulting solution. The reaction solution was stirred in
an argon atmosphere at 40 for 4 h. The resulting
dark violet precipitate was collected on a Schott filter
and was washed with water and acetone. The solid was
dried in vacuo (30 , 2–3 Torr) for 3 h. Yield: 2.4 g
(94% of the theoretical value); mp 152 С. Pd(dba)2
°
С
for 45–60 min, and
×
10–3 mol) was added to the
°
С
°
С
°
,
,
which exists in solution as the dimer Pd2(dba)3Solv)
contains three bridging bidentate dibenzylideneaceꢀ
1
tone ligands in the sꢀcis–sꢀtrans conformation. H
NMR (CDCl3,
δ
, ppm): dba (1):
c
isꢀalkene—6.65
(НA, d, 3
J
= 13.4 Hz) and 6.77 (НВ, d, 3
J = 13.4 Hz);
transꢀalkene—4.99 (НА, d, 3J = 13.4 Hz) and 5.94
(НВ, d, 3J = 13.4 Hz); dba (2): cisꢀalkene—6.37 (НА,
Example 2. A benzene solution of white phosphoꢀ
d, 3
J
= 13.4 Hz) and 6.47 (НВ, d, 3
alkene—4.96 (НА, d, 3J = 13.4 Hz) and 5.89 (НВ, d,
3J = 13.4 Hz); dba (3): cisꢀalkene—5.13 (НА, d, 3
13.4 Hz) and 5.34 (НВ, d, 3J = 13.4 Hz); transꢀalkꢀ
ene—6.16 (НА, d, 3 = 12.2 Hz) and 6.81 (НВ, d, 3
12.2 Hz); dba (free): 7.76 (2НА, d, 3J = 16 Hz) and
7.11 (2НВ, d, 3J = 16 Hz) [9]. 13C NMR (СDCl3,
ppm): 184.9 (CO), 182.2 (СO), 181.1 (СO), 111.3
(d, 1 СН = 160 Hz), 108.2 (d, 1
СН = 158 Hz), 107.9 (d,
1JСН = 157 Hz), 94.6 (d, 1 СН = 160 Hz), 89.8 (d, 1JСН
158 Hz), and 85.6 (d, 1
СН = 159 Hz); dba (free): 189.4
(СO), 143.8 (d, 1 СН = 154 Hz), and 125.8 (d, 1JСН
J = 13.4 Hz); transꢀ
rus (4 ml, 2.4
×
10–4 mol) was added to a burgundy
solution of Pd(dba)2 (0.69 g, 1.2
×
10–3 mol) in 260 ml
J
=
of benzene. A black precipitate formed at the early
stages of the reaction, but the solution retained its burꢀ
31
J
J =
gundy color. The solution was analyzed by P NMR
and UV spectroscopy. After 6 days, the solution was
separated from the precipitate by decantation. The
δ
,
solid (sample
three times and was dried in vacuo (30
Yield: 0.15 g. Sample was a combination of crystalꢀ
line and amorphous phases. The Xꢀray diffraction patꢀ
tern of sample showed, in a scattering angle range of
= 30 –75 and against the background of an amorꢀ
phous halo at = 35 –45 , the following reflections
from a crystalline phase ( n, I0 ): 2.910(8),
2
) was washed with benzene (10 ml)
J
J
°
С, 1 Torr).
J
=
2
J
J
=
2
154 Гц). Here, cisꢀ and transꢀalkene designate the alkꢀ
ene moieties of dibenzylideneacetone that are, respecꢀ
tively, cis and trans to the carbonyl group.
2
θ
°
°
2
θ
°
°
d/
Å (I/ )
White phosphorus was mechanically purified from 2.871(24), 2.796(6), 2.739(17), 2.667(5), 2.623(15),
surface oxidation products and was washed with dehyꢀ 2.548(7), 2.505(28), 2.473(15), 2.420(18), 2.362(28),
drated benzene immediately before use. A benzene 2.320(34), 2.288(40), 2.247(53), 2.222(47), 2.188(35),
solution of white phosphorus was prepared and stored 2.168(27), 2.118(24), 2.100(22), 2.075(20), 2.068(20),
in an inert atmosphere in a fingerꢀshaped vessel whose 2.044(16), 2.022(12), 1.987(11), 1.958(11), 1.933(11),
design allowed it to be vacuumized and filled with 1.920(11), 1.902(10), 1.882(6), 1.880(4), 1.837(7),
argon. 31P NMR:
δ
= –522 ppm (s).
1.802(4), 1.779(4), 1.767(4), 1.749(4), 1.714(2), 1.680(2),
KINETICS AND CATALYSIS Vol. 52
No. 5
2011

704
BELYKH et al.
Monitoring data for the decomposition of the Pd(dba)2 comꢀ
plex in DMF in an inert atmosphere
dibenzylideneacetone concentration, from the intenꢀ
sity of the absorption band at 325 nm ( → π transiꢀ
tion, ε325 = 41 280 l mol–1 cm–1) [12], with the absorꢀ
bance of the Pd(dba)2 complex at this wavelength (
ꢀtransition, ε325 = 33 540 l mol–1 cm–1) taken into
account.
n
*
Pd(dba)2 concentration
dba concentraꢀ
n
→
Time, min
tion, mmol/l
mmol/l
%
π
*
20
60
90
120
150
180
0.711
0.570
0.313
0.164
0.095
0
71.1
57.0
31.2
16.4
09.5
00
0.59
0.90
1.36
1.74
1.93
2.06
NMR spectra were recorded on a VXRꢀ500S
31
pulsed spectrometer (Varian). P chemical shifts are
reported relative to 85% orthophosphoric acid, with
13
positive values corresponding to downfield shifts. C
and 1H chemical shifts are reported relative to tetramꢀ
ethylsilane.
Note: CPd(dba) = 1 mmol/l, 30°C.
Catalyst samples were characterized by Xꢀray difꢀ
2
fraction on a DRONꢀ3M diffractometer (Cu
K radiaꢀ
α
tion). Catalysts were also examined by transmission
electron microscopy (TEM) on a Philips EMꢀ410
microscope. A drop of a catalyst solution prepared in
situ was placed on a carbonꢀcoated specimen grid and
was dried in an argon atmosphere. The imaging condiꢀ
tions ruled out the melting or decomposition of the
specimen under the action of the electron beam.
1.595(2), 1.519(2), 1.454(1), 1.374(4), 1.366(4), and
1.318(7). The reflections at 2.910, 2.871, 2.739, 2.505,
2.044, 1.987, 1.902, 1.714, 1.454, 1.366, and 1.318 Å are
consistent with the Xꢀray diffraction data for PdP2 [10].
The reflections at 2.667, 2.548, 2.505, 2.362, 2.320, 2.288,
2.247, 2.222, 2.118, 2.100, 2.075, 1.987, 1.882, 1.880,
1.837, 1.779, 1.767, 1.749, 1.680, 1.595, 1.519, 1.454, and
1.374 Å are consistent with the Xꢀray diffraction data
for the palladium phosphide Pd12P3.2 (PDF 42ꢀ0922)
[11]. The formula of the palladium phosphide Pd12P3.2
is also written as Pd3P0.8. The reflections at 2.796,
2.623, 2.548, 2.505, 2.420, 2.362, 2.320, 2.247, 2.222,
2.168, 2.118, 2.075, 2.068, 2.022, 1.987, 1.958, 1.933,
1.902, 1.882, 1.880, and 1.802 Å are consistent with
the Xꢀray diffraction data for the palladium phosphide
Pd4.8P (PDF 19ꢀ0890) [11]. It is possible that pallaꢀ
RESULTS AND DISCUSSION
The palladium catalyst based on the Pd(dba)2 comꢀ
plex in the DMF medium is active in the hydrogenaꢀ
tion of terminal bonds of alkenes, including styrene,
under mild conditions. The maximum value of the
turnover frequency of the catalyst is 270 (mol styrene)
(mol Pd)–1 min–1, and the turnover number is
5235 (mol styrene)/(mol Pd). UV spectroscopic data
demonstrated that, after styrene hydrogenation is comꢀ
plete, the reaction mixture contains neither the initial
palladium(0) complex nor free dibenzylideneacetone;
that is, srtyrene hydrogenation is accompanied by
dibenzylideneacetone hydrogenation. The maximum
catalytic activity in styrene hydrogenation, as calculated
from the initial rate of the reaction, decreases from 270
to 100 and 60 (mol styrene) (mol Pd)–1 min–1 if the
DMF solution of Pd(dba)2 is used 2 and 4 h after prepꢀ
aration, respectively. The UV spectroscopic study of
bis(dibenzylideneacetone)palladium(0) conversion in
DMF showed the Pd(dba)2 concentration decreases
dium was present in the sample (d/n, Å (I/I0): 2.247,
1.958, and 1.374). Some reflections at small scattering
angles were not identified.
Catalytic hydrogenation was performed in a temꢀ
peratureꢀcontrolled glass shaker at 30 С and an initial
°
hydrogen pressure of 1 atm in the presence of a pallaꢀ
dium catalyst obtained in situ. The catalyst was prepared
as follows. A DMF solution of Pd(dba)2 (1.1 mmol/l,
9 ml) was placed in a fingerꢀshaped vessel in an inert
atmosphere, and 1 ml of a benzene solution of white
phosphorus was added in drops (P : Pd = 0.2–1.5).
The solution was agitated for 10 min at room temperꢀ
ature and was then transferred into a prevacuumized
and hydrogenꢀfilled shaker. The shaker was closed
with a Teflon stopper with a rubber septum, a hydroꢀ
gen pressure of 2 atm was established, and the subꢀ
strate (styrene) was syringed into the reaction mixture.
Hydrogenation was conducted under vigorous agitaꢀ
tion to rule out diffusion control of the reaction. The
reaction was monitored as the pressure drop measured
with a manometer, as well as by GLC on a Chromꢀ5
chromatograph (3.6ꢀmꢀlong column packed with
Carbowax 20M, flameꢀionization detector).
with time and, in 3 h, Pd(dba)2 at 30 С turns entirely
°
into palladium nanoclusters, leaving free dibenzylideꢀ
neacetone in the solution (table).
The decrease in the Pd(dba)2 concentration upon
the dissolution of the complex in DMF is due to the
displacement of coordinated dibenzylideneacetone by
DMF molecules, which cannot stabilize Pd(0) comꢀ
plexes. The decomposition of Pd(dba)2 in DMF at
30 С is described by a pseudoꢀfirstꢀorder rate equaꢀ
°
tion:
w = k1C
,
)
2
UV spectra were recorded on an SFꢀ2000 spectroꢀ
photometer using a sealed cuvette (light path length of
0.01 cm). The Pd(dba)2 concentration was calculated
from the intensity of the absorption band at 525 nm
(
Pd dba
where
w
is the reaction rate, k1 is the rate constant
is the current Pd(dba)2 concentration.
10–2 min–1.
(min–1), and
C
(
d
→
d*ꢀtransition, ε525 = 6400 l mol–1 cm–1); the The k1 value is 1.28
×
KINETICS AND CATALYSIS Vol. 52
No. 5
2011

NATURE OF THE MODIFYING ACTION OF WHITE PHOSPHORUS
705
50 nm
50 nm
(b)
(а)
Fig. 1. TEM images of the (a) Pd(dba) –DMF and (b) Pd(dba) –styrene–DMF systems.
С = 1 mmol/l, 30°С, Р = 1 atm.
Pd Ar
2
2
The resulting palladium nanoclusters, whose averꢀ system in styrene hydrogenation decreases nearly linearly
age size is ~3 nm, according to TEM data, are prone with an increasing modifier concentration (Fig. 2). A
to aggregation and interact to form branched chains similar trend was observed earlier for the PdCl2–P–H2
(Fig. 1a).
system [3]. By contrast, with Pd(acac)2 or Pd(OAc)2 as
the precursor, the catalytic activity increases with an
increasing modifier concentration and passes through
a maximum at P : Pd = 0.3 [2]. The hydrogenation
activity at this point is 280 (mol substrate) (mol Pd)–1
min–1, one order of magnitude higher than the activity
of the unmodified system.
In order to obtain an experimental substantiation
for the inhibiting effect of white phosphorus on the
catalytic properties of palladium nanoclusters, we
studied the products of the interaction between
Pd(dba)2 and elemental phosphorus. According to
TEM data, highꢀcontrast particles form in the
Pd(dba)2–P system in DMF. Their average size is 2.7 nm
at P : Pd = 0.3 (Fig. 3).
The fact that the highꢀcontrast particles forming in
the absence of phosphorus and those forming in the
presence of P4 are similar in size suggests that the
inhibiting effect of elemental phosphorus on the cataꢀ
lytic properties of the Pd(dba)2ꢀbased system is associꢀ
ated with the nature of the resulting highꢀcontrast
nanoparticles rather than with their size.
Thus, when the Pd(dba)2 complex in DMF is used
as the precursor, the solution before the hydrogenation
stage may contain both the initial palladium(0) comꢀ
plex and palladium nanoclusters.
It was demonstrated by additional experiments
that, upon the introduction of styrene into the DMF
solution of Pd(dba)2 in an inert atmosphere, the soluꢀ
tion changes its color from burgundy through greenish
yellow to brown within a few seconds and turns into a
colloidal solution. The selfꢀorganization of base pallaꢀ
dium nanoparticles ~3 nm in size yields spherical 60ꢀ
to 70ꢀnm supramolecular structures (Fig. 1b). A simiꢀ
lar situation was observed earlier for the nanosized
hydrogenation catalyst forming in the PdCl2–P–H2
system [3].
The formation of two types of supramolecular
structures may be due to the rateꢀlimiting steps of their
formation being different. Both in the diffusionꢀconꢀ
trolled aggregation model and in the cluster–cluster
aggregation model, the rateꢀlimiting step of the forꢀ
mation of fractal structures is the diffusion of base
nanoparticles, with the probability of sticking of priꢀ
It was demonstrated earlier that white phosphorus
mary particles being unity [13, 14]. If the rateꢀlimiting can react with Pd(II) compounds under mild condiꢀ
step is particle growth, then the weak sticking of base tions to yield palladium phosphides of different comꢀ
particles to one another will be favorable for deeper positions, which are inactive in hydrogenation under
interpenetration of primary particles and for the forꢀ mild conditions [2]. Taking this fact into considerꢀ
mation of more compact aggregates.
ation, we investigated the interaction between
Pd(dba)2 and white phosphorus at varied reactant
ratios. To rule out the conversion of bis(dibenzylideꢀ
neacetone)palladium(0) into palladium nanoclusters,
which would take place in DMF, we conducted the
interaction between Pd(dba)2 and elemental phosphoꢀ
rus in benzene, in which Pd(dba)2 at a concentration of
about 3.0 mmol/l is stable over 24 h.
According to 31P NMR and UV spectroscopic data,
the reaction between Pd(dba)2 and elemental phosꢀ
phorus in benzene in an inert atmosphere takes place
Our experimental data indicate that the catalytic
activity observed in the case of Pd(dba)2 as the precurꢀ
sor is due not to the palladium(0) complex itself, but to
the palladium nanoclusters. Note that no example of
the participation of mononuclear Pd(0) complexes in
hydrogenation catalysis has been reported to date.
Once nanoparticles have formed in the system, hydroꢀ
genation will take place on their surface [15].
The addition of white phosphorus to the DMF
solution of Pd(dba)2 changes the catalytic properties of
the system. The catalytic activity of the Pd(dba)2–Р fairly readily at room temperature. At P : Pd =
2
KINETICS AND CATALYSIS Vol. 52
No. 5
2011

706
BELYKH et al.
w
, (mol H2) (mol Pd)–1 (min)–1
300
250
200
150
100
50
0
0.2 0.4 0.6 0.8 1.0 1.2 1.4
P/Pd
Fig. 2. Catalytic activity of the Pd(dba) –P system in styrene hydrogenation as a function of the reactant ratio.
С
= 1 mmol/l,
2
Pd
[
Ph–CH=CH ]: [Pd(dba) ] = 873, DMF, 30
°
С, PH = 1 atm.
2
2
2
(а)
(b)
Percentage of particles, %
90
80
70
60
50
40
30
20
10
0
50 nm
1.3 1.8 2.2 2.6 3.1 3.5 4.0 4.4 4.8
1.1 1.5 2.0 2.4 2.9 3.3 3.7 4.2 4.6
Fig. 3. (a) TEM image of the Pd(dba) –0.3Р system in DMF;
С = 1 mmol/l, 30°С, PH = 1 atm. (b) Particle size distribution.
Pd
2
2
(
CPd(dba) = 3.280 mmol/l), the reaction is complete
According to Xꢀray diffraction data, sample 1 (P : Pd =
2
2) is an Xꢀray amorphous substance (Fig. 4, curve a).
After it is heatꢀtreated in an inert atmosphere (400 С,
4 h), its diffraction pattern shows broad reflections
from the crystalline palladium phosphides PdP2 and
within 10–20 s, yielding a black precipitate (sample 1).
°
The dibenzylideneacetone concentration at this point
is 5.864 mmol/l, as calculated from the UV spectrum
of the reaction system; therefore, nearly the entire
dibenzylideneacetone of the Pd(dba)2 complex remains
in the solution. The absence of the resonance signal
Pd5P2, the PdP2 phase being dominant (Fig. 4, curve b).
Therefore, the reaction between Pd(dba)2 and white
phosphorus in benzene at P : Pd = 2 mainly yields
PdP2, the palladium phosphide richest in phosphorus.
from white phosphorus (31P NMR:
δ
= –522 ppm) in
31
the P NMR spectrum of the Pd(dba)2–2P system
recorded 10 min after the beginning of the reaction and
the absence of resonances due to phosphorus conversion
products between +500 and –550 ppm indicate that eleꢀ
At reactant ratios of P : Pd < 2 (P : Pd = 0.4 and 1),
complete or nearly complete Pd(dba)2 conversion in
the Pd(dba)2–
nР system is also possible, but, according
mental phosphorus is incorporated in sample 1.
to UV spectroscopic data, the reaction time increases
KINETICS AND CATALYSIS Vol. 52
No. 5
2011

NATURE OF THE MODIFYING ACTION OF WHITE PHOSPHORUS
Reflection intensity
707
ꢀ1
ꢀ2
ꢀ3
b
а
30
40
50
60
70
2
θ
, deg
Fig. 4. Xꢀray diffraction patterns of sample
1
(P : Pd(dba) = 2) (
a) before and (b) after heat treatment in argon (400°С, 4 h):
2
(1) PdP , (2) Pd P , and (3) unidentified phase.
2 5 2
from 10–20 s at P : Pd = 2 to 3 h at P : Pd = 1.0 and to phides PdP2, Pd3P0.8, and Pd4.8P. It is possible that the
24 h at P : Pd = 0.4 (90% conversion). At P : Pd = 0.2, sample contains finely dispersed palladium. The pallaꢀ
the Pd(dba)2 conversion is only 70% in 6 days. Note
that the 31P NMR spectra of the Pd(dba)2–
reaction
systems with = 0.2 and 1.0, like the spectrum
recorded for P : Pd =2, show no resonance from white
phosphorus ( = –522 ppm) as soon as 10 min after
the beginning of the reaction, and no other signals
between +500 and –550 ppm are observed either at the
early stages of the reaction or in 1 day.
dium phosphide Pd3P0.8 is known to crystallize in a
cementiteꢀtype (Fe3C) structure. The difference between
this phase and the ideal phosphide Pd3P (Pd : P = 3 : 1)
is due to random phosphorus vacancies [16]. A numꢀ
nР
n
δ
ber of reflections from crystalline phases in the
–30 range were not identified.
2 =
θ
6
°
°
The presence of PdP2 in sample 2 (P : Pd = 0.2) is
consistent with the above hypothesis that, irrespective
of the reactant ratio, the reaction between Pd(dba)2
and white phosphorus proceeds via the formation of
palladium diphosphide, which then reacts slowly with
Pd(0) to yield palladiumꢀrich phosphides. Bis(dibenꢀ
zylideneacetone)palladium(0), whose empirical forꢀ
The Pd(dba)2 concentration versus time curves
indicate two distinct kinetic regions for the palladium
complex–white phosphorus interaction (Fig. 5). In
the first (initial) region, the reaction rate as a funcꢀ
tion of the white phosphorus concentration changes
from 1.1
10–4 mol l–1 min–1 for P : Pd = 0.4 and 4.54
l–1 min–1 for P : Pd = 1.0, suggesting that the reacꢀ
×
10–4 mol l–1 min–1 for P : Pd = 0.2 to 2.04
×
×
10–4 mol
×
×
CPd
4.0
3.5
3.0
2.5
2.0
1.5
1.0
0.5
,
mmol/l
tion is firstꢀorder with respect to phosphorus. In the
second kinetic region, the reaction rate is lower by a
factor of 15–40. The totality of spectroscopic and
kinetic data suggests that, at the initial stages, the reacꢀ
tion between Pd(dba)2 and white phosphorus yields
PdP2 at various reactant ratios. The further conversion
of Pd(dba)2 is due to its reaction with PdP2. Because
PdP2 is a poorly soluble compound, the Pd(dba)2 conꢀ
version takes place on the surface of nanosized pallaꢀ
dium phosphide. It is the crossover of the reaction
from homogeneous to heterogeneous kinetics that
causes the sharp decrease in the reaction rate.
1
2
3
4
This hypothesis was verified by an Xꢀray diffraction
analysis of the black precipitate forming in the
0
10 20 30 40 50 60 70 80 90 100110120130
min
Pd(dba)2–0.2Р system—sample
2 (Pd(dba)2 converꢀ
t
,
sion of 70%). According to Xꢀray diffraction data, this
sample is a mixture of crystalline and amorphous
phases (Fig. 6). Its diffraction pattern shows, against
Fig. 5. Time variation of the Pd(dba) concentration in the
2
interaction between Pd(dba) and white phosphorus in benꢀ
2
zene at different reactant molar ratios: [P] : [Pd(dba) ] =
2
2
the background of an amorphous halo at
2
θ
= 30 –
°
(
1
) 0.2, (
2
) 0.4, (
3) 1.0, and (
4
) 2.0.
С
=
Pd(dba)
75°, reflections from the crystalline palladium phosꢀ
3.3 mmol/l, 20°С.
KINETICS AND CATALYSIS Vol. 52
No. 5
2011

708
BELYKH et al.
Reflection intensity
ꢀ1
ꢀ2
ꢀ3
ꢀ4
30
40
50
60
70
2
θ
, deg
Fig. 6. Xꢀray diffraction patterns of sample 2 (P : Pd(dba) = 0.2): (1) PdP , (2) Pd P , (3) Pd P, and (4) unidentified phase.
2 2 3 0.8 4.8
mula is Pd(dba)2, is actually an M2(dba)3(dba)ꢀtype as a result of the reaction between Pd(dba)2 and phosꢀ
(M = Pd, Pt) dinuclear complex with three bridging
dibenzylideneacetone ligands in sꢀcis and sꢀtrans conꢀ
formations, with the fourth dibenzylideneacetone
molecule replacing the solvent molecule during
recrystallization [17]. The formation of palladium
phosphides can, therefore, be represented as the folꢀ
lowing consecutive reactions:
phorus in benzene can likely be extended to the reacꢀ
tion taking place in DMF. The radical difference
between the reactions occurring in these solvents is
that palladium nanocluster formation in DMF is a side
reaction. It is natural that the palladium phosphides
forming in benzene and those forming in DMF may
differ in composition at the same initial P : Pd ratio
and reaction time because, in the case of DMF, part of
the Pd(dba)2 will turn into palladium nanoclusters by
reacting with DMF and the P : Pd ratio in the solution
will change. However, the sequence of elementary
steps yielding palladium phosphides will be the same.
(I)
(II)
Pd dba dba + Solv ꢀ Pd dba Solv + dba,
(
)3 (
)
(
)3 (
)
2
2
Pd dba Solv + P = 2PdP + 3dba + Solv,
(
)
2
4
2
3
(III)
(IV)
(V)
PdP + 2Pd dba = Pd P + 6dba,
(
)
2
2
5 2
3
2Pd P + Pd dba = 4Pd P+3dba,
(
)
5
2
2
3
3
0.93Pd P + Pd dba = Pd P + 3dba,
(
)
The above data suggest that the following two proꢀ
cesses occur at a high rate in the Pd(dba)2–P–styrene
catalytic system at room temperature: chemical conꢀ
version of the Pd(dba)2 complex into the phosphides
PdP2 and Pd5P2, which are inactive in hydrogenation
under mild conditions [3], and Pd(dba)2 conversion
into palladium nanoclusters under the action of DMF
3
2
4.8
3
where Solv = solvent.
It was demonstrated by additional experiments that
PdP2 indeed reacts with Pd(dba)2 in benzene at room
temperature at an initial rate of
7
×
10–7 mol l–1 min–1
(
CPd(dba) = 3.55 mmol/l, Pd(dba)2 : PdP2 = 11). The
2
above mechanism of palladium phosphides formation and styrene:
Pd
Pd
Pd
PdP2
Pd3P
Pd5P2
Pd6P
–dba +P4
DMF,
styrene
Pd
Pd
Pd
Pd
Pd
Pd(dba)2
Pd
Pd
Pd
–dba
The fact that the catalytic activity is inversely proꢀ introduction of phosphorus, part of the palladium
portional to the P : Pd ratio indicates that, upon the turns into hydrogenationꢀinactive form. We think that
KINETICS AND CATALYSIS Vol. 52
No. 5
2011

NATURE OF THE MODIFYING ACTION OF WHITE PHOSPHORUS
709
this is due to the dominant formation, under the cataꢀ mechanism of Pd nucleation will dominate in the sysꢀ
lytic conditions, of segregated nanoparticles of pallaꢀ tem. In this mechanism, the crystallization centers are
dium phosphides, which are inactive in hydrogenation nanoparticles of the poorly soluble palladium phosꢀ
under mild conditions, and palladium nanoclusters. phides that form in the solution as a result of the redox
The segregation of the nanoparticles may be caused by reaction between Pd(II) and phosphorus. This situaꢀ
the following two factors. The first is the high rate of tion is typical of Pd(acac)2, which reacts readily with
Pd(dba)2 conversion into palladium nanoclusters via white phosphorus at room temperature to yield pallaꢀ
the replacement of dibenzylideneacetone in the coorꢀ dium phosphides and, because of its chelate structure,
dination sphere of palladium by the substrate (styꢀ is reducible with hydrogen at a noticeable rate only at
rene), which alone is incapable of stabilizing mononuꢀ 80°С [2]. In this case, the palladium phosphide nanoꢀ
clear Pd(0) complexes, even though it is in large excess particles, serving as crystallization centers for the
in the system. The second factor is the low rate of the Pd(0) resulting from the hydrogenolysis of the remainꢀ
chemical reaction between Pd(0) and PdP2 (which ing Pd(acac)2, increase the degree of dispersion and,
results from the introduction of phosphorus into the accordingly, the catalytic activity of the nanosized palꢀ
system) yielding the palladiumꢀrich phosphides Pd5P2, ladiumꢀbased hydrogenation catalyst.
Pd3P, and Pd4.8P via a number of consecutive steps. It
is also possible that the interaction of Pd(0) with the
palladium phosphides present in the solution yields
core–shell type particles in which the core is PdxP and
the shell is Pd(0). However, the proportion of these
nanoparticles is likely small.
With the order in which the components were
introduced into the solution (phosphorus and, after
10 min, styrene), the Pd(dba)2 conversion into nanoꢀ
sized palladium phosphides and palladium nanoclusꢀ
ters depends on the P : Pd ratio.
Thus, this study of the properties and nature of the
nanoparticles resulting from the interaction between
the palladium(0) complex Pd(dba)2 and white phosꢀ
phorus and earlier experimental studies [2, 3] have
provided an explanation and experimental substantiaꢀ
tion for the diametrically opposite effects of white
phosphorus on the catalytic properties of nanosized
palladiumꢀbased hydrogenation catalysts.
ACKNOWLEDGMENTS
This work was carried out in the framework of the
federal target program “Research and Pedagogical
Cadre for Innovative Russia” (state contract
no. P1344).
Our investigation of the interaction between
Pd(dba)2 and elemental phosphorus verifies some steps
of the above mechanism of the formation of the
Pd(II)ꢀbased, phosphorusꢀmodified, nanosized
hydrogenation catalysts [3]. We hypothesized in our
earlier work [3] that, depending on the ratio of the
rates of Pd(II) reduction with phosphorus and hydroꢀ
gen and the rate of subsequent nanoparticle growth,
the process can yield nanoparticles consisting of pallaꢀ
dium nanoclusters, palladium phosphide nanopartiꢀ
cles of various compositions, or core–shell nanopartiꢀ
cles. The high rate of Pd(II) reduction by hydrogen
and, as a consequence, a high degree of supersaturaꢀ
tion are favorable for Pd(0) nucleation via homogeꢀ
neous condensation. In this case, at P : Pd < 1 a
mechanical mixture of palladium phosphide nanoparꢀ
ticles and palladium nanoclusters will mainly form in
the reaction system. The conversion of part of the palꢀ
ladium into hydrogenationꢀinactive phosphides will
decrease the catalytic activity of the system at any
reactant ratio; that is, phosphorus will act as a catalyst
poison. This is the case in catalytic systems in which
the palladium precursor is in the zero oxidation state
(e.g., Pd(dba)2) and in systems in which the Pd(II)
compound is reduced at a high rate, for example,
PdCl2–Р–Н2 [3] and Pd(acac)2–Р–AlEt3 [18]. Note
that, in these catalytic systems, owing to the high
nucleation (Pd(II) reduction) rate, finely dispersed
catalysts with a particle size of 3–5 nm form in the
absence of phosphorus as well. If the rate of Pd(II)
reduction with hydrogen is low, the heterogeneous
REFERENCES
1. RF Patent 2323776, 2008.
2. Belykh, L.B., Skripov, N.I., Belonogova, L.N., Umaꢀ
nets, V.A., and Schmidt, F.K., Kinet. Catal., 2010,
vol. 51, no. 1, p. 42.
3. Skripov, N.I., Belykh, L.B., Belonogova, L.N., Umaꢀ
nets, V.A., Ryzhkovich, E.N., and Schmidt, F.K.,
Kinet. Catal., 2010, vol. 51, no. 5, p. 714.
4. Gordon, A.J. and Ford, R.A., A Handbook of Practical
Data, Techniques, and References, New York: Wiley,
1972.
5. Mitchell, J. and Smith, D., Aquametry, New York: Pleꢀ
num, 1977.
6. Vul’fson, N.S., Preparativnaya organicheskaya khimiya
(Preparative Organic Chemistry), Moscow: Khimizꢀ
dat, 1959.
7. Dzhemilev, U.M., Metallokompleksnyi kataliz v organꢀ
icheskom sinteze: Alitsiklicheskie soedineniya (Catalysis
by Metal Complexes in Organic Synthesis: Alicyclic
Compounds), Moscow: Khimiya, 1999.
8. Ukai, T., Kawazura, H., and Ishii, Y., J. Organomet.
Chem., 1974, vol. 65, p. 253.
9. Kawazura, H., Tanaka, H., Yamada, H., and Takaꢀ
hashi, T., Bull. Chem. Soc. Jpn., 1978, vol. 51, no. 12,
p. 3466.
KINETICS AND CATALYSIS Vol. 52
No. 5
2011

710
BELYKH et al.
10. Wiehage, G., Weibke, Fr., and Biltz, W., Z. Anorg. Allg.
Chem., 1936, vol. 228, p. 357.
Physical Chemistry of Heterogeneous Systems and
Polymers), Irkutsk: Irkutsk. Gos. Univ., 2001.
11. Powder Diffraction File, Hanawalt Search Manual: Inorꢀ 15. Durand, J., Teuma, E., and Gómez, M., Eur. J. Inorg.
ganic Phases, sets 1–42, Swarthmore: JCPDS, 1992. Chem., 2008, no. 23, p. 3577.
12. Harley, P.D., Adar, F., and Gray, H.B., J. Am. Chem. 16. Andersson, Y., Rundqvist, S., Tellgren, R., and Thoꢀ
Soc., 1989, vol. 111, no. 4, p. 1312. mas, J.O., J. Solid State Chem., 1980, vol. 32, p. 321.
13. Pomogailo, A.D., Rozenberg, A.S., and Uflyand, I.E., 17. Rubezhov, A.Z., Usp. Khim., 1988, vol. 57, no. 12,
Nanochastitsy metallov v polimerakh (Metal Nanopartiꢀ
p. 2078.
cles in Polymers), Moscow: Khimiya, 2000.
14. Schmidt, F.K., Fraktal’nyi analiz v fizikoꢀkhimii geteroꢀ
gennykh sistem i polimerov (Fractal Analysis in the
18. Belykh, L.B., Titova, Yu.Yu., Umanets, V.A., and
Schmidt, F.K., Russ. J. Appl. Chem., 2006, vol. 79,
no. 8, p. 1271.
KINETICS AND CATALYSIS Vol. 52
No. 5
2011