Evidence for the role of colloidal palladium in the catalytic formation of H2O2 from H2 and O2

Article abstract of DOI:10.1006/jcat.2001.3501

The direct production of H₂O₂ from H₂ and O₂ (O₂/H₂ = 2) at 25°C and 760 Torr occurs in an aqueous phase over colloidal palladium that may be introduced either via PdCl₂ or via Pd supported on silica gel (Pd/SiO₂). In the latter case, aqueous HCl facilitates the dissolution of the supported Pd. The presence of colloidal palladium was confirmed by electron microscopy. When the solution was either 0.1 M or 1.0 M in HCl, removal of the silica, along with any remaining supported Pd, did not affect the rate of H₂O₂ formation because the amount of active colloidal Pd remained unchanged. The specific activity of the supported Pd is only 3% of that for colloidal Pd, probably because of transport limitations within the pores of the silica.

Full text of DOI:10.1006/jcat.2001.3501

Journal of Catalysis 206, 173–176 (2002)
doi:10.1006/jcat.2001.3501, available online at http://www.idealibrary.com on
PRIORITY COMMUNICATION
Evidence for the Role of Colloidal Palladium in the Catalytic
Formation of H2O2 from H2 and O2
Dhammike P. Dissanayake1 and Jack H. Lunsford2
Department of Chemistry, Texas A&M University, College Station, Texas 77842-3012
Received November 12, 2001; revised December 7, 2001; accepted December 17, 2001
co-workers (7). More typical supports include silica, silicic
acid, carbon and alumina. The reaction is usually carried
out at elevated pressures in an aqueous slurry that contains
HCl or HBr. The presence of protons is known to inhibit the
decomposition of H2O2 (8); however, the role of the halide
The direct production of H2O2 from H2 and O2 (O2/H2 = 2) at
25◦C and 760 Torr occurs in an aqueous phase over colloidal palla-
dium that may be introduced either via PdCl2 or via Pd supported
on silica gel (Pd/SiO2). In the latter case, aqueous HCl facilitates
the dissolution of the supported Pd. The presence of colloidal palla-
dium was confirmed by electron microscopy. When the solution was
has not previously been established. The results reported
here demonstrate the importance of colloidal palladium in
either 0.1 M or 1.0 M in HCl, removal of the silica, along with any
the catalytic conversion of H2 and O2 to H2O2, even when
remaining supported Pd, did not affect the rate of H2O2 formation
onebeginswithpalladiumsupportedonsilica. Thepresence
because the amount of active colloidal Pd remained unchanged. The
specific activity of the supported Pd is only 3% of that for colloidal
Pd, probably because of transport limitations within the pores of
of HCl facilitates the dissolution of metallic Pd as PdCl42−
which is an intermediate in the colloid formation.
,
c
the silica.
ꢀ 2002 Elsevier Science (USA)
Key Words: palladium colloid; hydrogen peroxide; palladium/
silica; supported palladium catalyst.
EXPERIMENTAL
The catalytic reactions were carried out at 760 Torr and
25◦C in a 22-mm-i.d. Pyrex reactor so that changes in the
systemcouldbeobservedvisually. Thegases, oxygen(20mL
min−1) and hydrogen (10 mL min−1), were allowed to flow
through a fine glass frit at the bottom of the reactor. The
reactor was open at the top. Note that this is a potentially
explosive mixture and appropriate precautions should be
taken. For example, if a supported palladium catalyst is
used, there should be no dry catalyst on the walls of the
reactor.
Palladium was introduced to 10 mL of an aqueous phase
either as PdCl2 or via a Pd/SiO2 material. Before addition
of the Pd, the water was acidified with HCl and saturated
with the H2/O2 mixture. The Pd/SiO2 material consisted of
5 wt% Pd on Davison grade 57 silica gel that had been re-
duced in H2 at 300◦C (9). After the Pd was added to the
aqueous phase, the solution or slurry was rapidly stirred
with an overhead stirrer (∼500 rpm). Small aliquots of the
liquid phase were removed from the reactor, and the H2O2
formed was analyzed by colorimetry after complexation
with a TiOSO4/H2SO4 reagent (10).
INTRODUCTION
For environmental reasons, hydrogen peroxide has be-
come a substitute for chlorine as a bleach in the pulp and pa-
per industry and as an oxidant in wastewater treatment (1).
Moreover, it has recently been proposed as a reagent for the
removal of residual aromatic sulfur compounds in fuels (2).
ThecurrentmethodforthecommercialproductionofH2O2
is a circuitous process that involves the catalytic hydro-
genation of an alkylanthroquinone to the corresponding
hydroquinone, followedbyitstreatmentwithO2 toproduce
H2O2 and the original anthroquinone (1). In the quest for a
less expensive route, attention has been given to the direct
production of H2O2 from the reaction of H2 and O2 over
a supported palladium catalyst, and a number of patents
have resulted from the work at duPont and other compa-
nies (3–6). A fundamental study of a catalyst containing
hafnium phosphate and viologen phosphonate with sup-
ported Pd has recently been reported by Thompson and
Theamountofcolloidpresentinsolutionwasdetermined
by first analyzing for the amount of Pd2+, using the method
of Onishi (11). Then the colloid was oxidized by expos-
ing the solution to air for 12 h, and the solution was again
1 On leave from the Department of Chemistry, University of Colombo,
Colombo 3, Sri Lanka.
2 To whom correspondence should be addressed. Fax: (979) 845-4719.
E-mail: lunsford@mail.chem.tamu.edu.
173
0021-9517/02 $35.00
c
ꢀ 2002 Elsevier Science (USA)
All rights reserved.

174
DISSANAYAKE AND LUNSFORD
analyzed for Pd2+. The difference between the amounts of
Pd2+ is taken to be the amount of colloidal Pd.
Samples for electron microscopy were prepared by re-
moving aliquots of liquid after 1 h of reaction from the so-
lution that originally contained 1.0 × 10−3 M PdCl42−. This
was neutralized with NaOH, deposited on a carbon-coated
copper grid, dried, and analyzed with a Zeiss 10C transmis-
sion electron microscope.
RESULTS AND DISCUSSION
The formation of H2O2 in aqueous solutions that initially
contained 1.0 × 10−3 or 1.0 × 10−4 M PdCl24− in 1.0 M HCl is
depicted in Fig. 1. Although it is not evident from this figure,
there was generally a brief induction period during which
almost no H2O2 was formed. After this period the rate of
H2O2 formation was essentially constant for several hours
and then began to decrease. During the course of the reac-
tion the yellow color that results from PdCl24− began to de-
crease in intensity and the solution eventually became col-
orless. The electron micrograph of Fig. 2 shows the presence
of colloidal Pd with a broad distribution of particle sizes
but with an average diameter of about 6 nm. Henglein (12)
has previously reported the formation of Pd colloids by re-
duction of PdCl24− by hydrogen; however, sodium citrate
was used as a stabilizer, which resulted in a narrow size dis-
tribution and an average particle size of about 4 nm. From
the data of Fig. 1 it was established that after 1 h, for the
nominally 1 × 10−4 M PdCl42− solution, the rate of H2O2
formation was 1 × 10−4 M min−1 and the colloid concen-
FIG. 2. Electron micrograph of a Pd colloid obtained after exposure
of a 1.0 × 10−3 M PdCl42− solution to H2/O2 for 1 h at 25◦C.
tration was 2 × 10−5 M. For the nominally 1 × 10−3
M
PdCl42− solution, the rate was 3 × 10−4 M min−1 and the
colloidconcentrationwas5×10−5 M. Thus, theratesareap-
proximately proportional to the colloid concentration. The
specific rate was ca. 50 mmole H2O2 g−1 colloidal Pd min−1
.
As shown in Fig. 1, the rates of H2O2 formation are nearly
constant for 3–5 h, but thereafter the rates, as well as the
colloid concentrations, continuously decrease. After 24 h,
for example, the rate of H2O2 formation and the Pd colloid
concentration were 7 × 10−5 M min−1 and 7 × 10−6 M, re-
spectively, for the nominally 1 × 10−3 M solution. The lower
colloid concentration is approaching the limit of detection;
nevertheless, it is apparent that the colloid is disappearing
from the aqueous phase. Meanwhile the frit became dark as
a result of the deposition of palladium. The PdCl24− is not a
factor since it has decreased below a detectable level even
after 1–2 h on stream. Neither is the decomposition of H2O2
a factor in the decreasing rates under the conditions of these
experiments;although, atsmallerH+ concentrationsand/or
at larger wt% H2O2, the decomposition reaction becomes
significant, particularly in the presence of Pd/SiO2.
One may speculate as to why the palladium prefers to
reside on the frit. The answer may be found in the local
concentration of H2 that exists at the frit where the gases
are introduced. At 25◦C the concentration of H2 in the gas
phase is approximately 50 times that found in the aqueous
phase, and the Pd2+ ions may preferentially be reduced at
the gas–liquid film on the frit where the H2 concentration is
the largest. Attempts were made to stabilize the colloid with
sodium citrate or poly (vinyl alcohol), but without success.
Thepalladiummetalonthefritexhibitedverylittlecatalytic
activity, probably as a result of its small surface area. It
FIG. 1. Catalytic formation of H2O2 at 25◦C in an aqueous solution
that was 1 M in HCl and originally (ꢀ) 1.0 × 10−4 M or (ꢁ) 1.0 × 10−3
M
in PdCl24−
.

COLLOIDAL PALLADIUM IN H2O2 FORMATION
175
may be largely digested back into solution as PdCl24− by
TABLE 1
passing only O2 through the frit. Then upon addition of the
H2/O2 mixture, the aqueous phase again becomes active. In
general, the formation of the colloid appears to be autocata-
lytic, as has been noted previously for Pt colloids (13), and
to achieve reproducible results it was necessary that a small
amount of Pd remain on the frit from one experiment to the
next. When the frit was rigorously cleaned with nitric acid,
long induction periods were observed, and the palladium
in the form of PdCl24− was extensively reduced on the frit,
without the formation of a significant amount of colloid.
In the range of conditions described here, colloidal palla-
dium also plays a major role as a catalyst even when Pd/SiO2
is present in the system. The formation of H2O2 over 2.2 mg
of Pd/SiO2 in the slurry was followed as a function of time
at three different HCl concentrations. The amount of Pd
on the silica is equivalent to that present in 10 mL of a 1 ×
10−4 M solution. As shown in Fig. 3, the reaction was char-
acterized by a brief induction period, during which time
the aqueous phase became yellow and then clear again
for the systems containing 0.1 M and 1 M HCl. Analysis
of the yellow solution by UV/vis spectroscopy indicated
that PdCl24− was formed. After 60 min on stream the solid
phase was rapidly removed by centrifuging, and the reac-
tion was continued only in the liquid phase. Surprisingly,
the rate of H2O2 formation (ca. 5 × 10−5 M min−1) was es-
sentially the same after the solid phase had been removed
Rates of H2O2 Formation and Pd Colloid Concentrations
in Several Catalytic Systems
Rate
Colloid
concentration (M)
System
(M min−1
)
Nominally 1 × 10−4M PdCl42−
in 1 M HCl solution
After removal of 2.2 mg Pd/SiO2
from 1 M or 0.1 M HCl solution
Prior to removal of Pd/SiO2
from 0.01 M HCl solution
1.5 × 10−4
5.1 × 10−5
1.5 × 10−5
2.9 × 10−5a
1.8 × 10−5
n.a.b
a The estimated error in the rates is 10% and the estimated error in
the colloid concentrations is 3 × 10−6 M.
b The colloid concentration after removal of the Pd/SiO2 was <3 ×
10−6 M.
from the solutions that contained the larger acid concentra-
tions. These results confirm that the reaction was catalyzed
by a form of palladium in the liquid phase, i.e., colloidal
palladium. Obviously, the palladium remaining on the sil-
ica gel, as well as that on the frit, contributed very little
to the activity. The smaller rate that was observed after the
solid phase was removed from the 0.01 M HCl solution may
result, in part, from the partial loss of colloidal palladium.
Although this loss of colloidal palladium may take place at
all three acid concentrations, it is most apparent when the
colloid concentration is small, which occurs when the acid
strength is low. For comparison with the results obtained
with the Pd/SiO2 slurry, the catalytic properties of 1.0 ×
10−4 M PdCl24− in 1.0 M HCl were determined over the
same time on stream, and the results are shown in Fig. 3.
The rates of H2O2 formation and the respective Pd colloid
concentrations are summarized in Table 1. Within the limits
of uncertainty, the rates are again found to be proportional
to the colloid concentration. When the solutions were 1.0
or 0.1 M HCl, 40% of the palladium remained on the sil-
ica or was deposited on the frit. In the 0.01 M solution, the
palladium remained largely on the silica. The rate obtained
for the Pd/SiO2 in the 0.01 M solution was 1.5 × 10−5 M−1
,
which is only 10% of the value found for the nominally
1.0 × 10−4 M aqueous phase. If one makes the comparison
based on the amount of colloid present in the latter system,
the activity of the colloid is about 50 mmol H2O2 g−1 Pd
min−1, whereas the activity of the Pd/SiO2 is only 1.4 mmol
H2O2 g−1 Pd min−1. At H2O2 concentrations <0.01 wt%,
the decomposition of H2O2, even at 0.01 M HCl, is not a
factor.
The question arises as to why the Pd/SiO2 itself is so
inactive for H2O2 production. As pointed out above, the
concentration of H2 in the gas phase is much greater than
in the aqueous phase, and, as a consequence, H2O2 may be
formed mainly at the interface between a gas bubble and
a colloidal Pd particle (i.e., in a region where the reagent
concentration is the largest). The absence of this interface
FIG. 3. Catalytic formation of H2O2 in an aqueous phase before and
after the removal of 2.2 mg of 5 wt% Pd/SiO2: (ꢁ) before and (ꢂ) after
the removal of the solid phase from a 1 M HCl solution; (ꢃ) before and
(ꢄ) after the removal of the solid phase from an 0.1 M HCl solution; (ꢅ)
before and (ꢆ) after the removal of the solid phase from an 0.01 M HCl
solution. Peroxide formation in a nominally 1.0 × 10−4 M PdCl42− solution
containing 1 M HCl is depicted by (ꢀ).

176
DISSANAYAKE AND LUNSFORD
in the pores of the silica gel would result in the small activity
of the Pd/SiO2 phase, even though the Pd particle size in
this phase, at least initially, was about 6 nm (9). In this three-
phase system, transport of H2 across the liquid film between
a bubble and a colloidal particle may be the rate-limiting
factor in the formation of H2O2, similar to that found in
the partial hydrogenation of benzene to cyclohexene over
a supported ruthenium catalyst (14). Thompson and co-
workers(7) have addressed the issue of mass transfer in
the context of H2O2 formation; however, their focus was
on the transport of hydrogen, the limiting reagent, from
the gas bubble into the liquid phase. This aspect of mass
transfer was not a limiting factor in the results reported
here since the rates were proportional to the amount of Pd
present as the colloid, which would not be expected if the
global transfer of the gases into the liquid phase were rate
limiting.
ACKNOWLEDGMENT
Electron microscopy was carried out at the Electron Microscopy Center,
Texas A&M University. This research was supported by the Office of Basic
Energy Sciences, U.S. Department of Energy.
REFERENCES
1. Hess, W. T., in “Kirk–Othmer Encyclopedia of Chemical Technology”
(J. I. Kroschwitz and M. Howe-Grant, Eds.), 4th ed., Vol. 13, p. 961.
Wiley: New York, 1995.
2. Hulea, V., Fajula, F., and Bousquet, J., J. Catal. 189, 179 (2001).
3. Gosser, L. W. (du Pont), U.S. Patent 5,135,731 (1992).
4. Gosser, L. W., and Schwartz, J.-A.T. (du Pont), U.S. Patent 4,832,938
(1989).
5. Izumi, Y., Miyazaki, H., and Kawahara, S. (Tokuyama Soda Kabushiki
Kaisha), U.S. Patent 4,009,252 (1977).
6. Brill, W. F. (Halcon), U.S. Patent 4,661,337 (1987).
7. Krishnan, V. V., Dokoutchaev, A. G., and Thompson, M. E., J. Catal.
196, 366 (2000).
8. Pospelova, T. A., and Kobozev, N. I., Russ. J. Phys. Chem. (Engl.
Transl.). 35, 584 (1961).
9. Fajula, F., Anthony, R. G., and Lunsford, J. H., J. Catal. 73, 237 (1982).
10. Cohen, I. R., Purcell, T. C., and Altshuller, A. P., Environ. Sci. Technol.
1, 247 (1967).
In conclusion, the catalytic conversion of H2 and O2 to
H2O2 occurs over colloidal Pd in the aqueous phase. This
colloidal Pd may be derived from the hydrogen reduction
of PdCl24− that is introduced either directly or indirectly
via the oxidation of metallic Pd in contact with an aqueous 11. Onishi, H., “Photometric Determination of Traces of Metals.” Wiley,
New York, 1989.
12. Henglein, A., J. Phys. Chem. 104, 6683 (2000).
13. Henglein, A., Ershov, B. G., and Malow, M., J. Phys. Chem. 99, 14129
HCl solution. A solid consisting of reduced Pd supported
on silica gel may serve as one source of palladium. Indirect
evidence suggests that the transfer of reagents from small
bubbles within the liquid phase to the colloid may be the
rate limiting step in the reaction.
(1995).
14. Struijk, J., d’Angremond, M., Lucas-de Regt, W. J. M., and Scholten,
J. J. F., Appl. Catal. A. 83, 263 (1992).