Direct synthesis of hydrogen peroxide in water in a continuous trickle bed reactor optimized to maximize productivity

Article abstract of DOI:10.1039/c3gc40811f

Hydrogen peroxide direct synthesis was studied in continuous mode over a 5% wt Pd/C commercial catalyst in a Trickle Bed Reactor. The target of the study was to maximize the hydrogen peroxide production. The catalyst was uniformly diluted in quartz sand at different concentrations to investigate their effects on the direct synthesis. The amount of catalyst and the distribution of the catalyst along the bed were optimized to obtain the highest possible yield. The distribution of the catalyst along the bed gave the possibility to significantly improve the selectivity and production of hydrogen peroxide (up to 0.5% under selected conditions). Higher production rate and selectivity were found when the catalyst concentration was decreased along the bed from the top to the bottom as compared to the uniformly dispersed catalyst. The H₂/Pd ratio was found to be an important parameter that has to be investigated in the hydrogen peroxide direct synthesis. The effect of a pretreatment of the catalyst with a solution of sodium bromide and phosphoric acid was studied; the results showed how a catalyst pretreatment can lead to a real green hydrogen peroxide synthesis in water. Some optimization guidelines are also provided.

Full text of DOI:10.1039/c3gc40811f

Green Chemistry
View Article Online
View Journal | View Issue
PAPER
Direct synthesis of hydrogen peroxide in water in a
continuous trickle bed reactor optimized to maximize
productivity
Cite this: Green Chem., 2013, 15, 2502
Pierdomenico Biasi,*a Juan García-Serna,*b Alice Bittantea and Tapio Salmia
Hydrogen peroxide direct synthesis was studied in continuous mode over a 5% wt Pd/C commercial cata-
lyst in a Trickle Bed Reactor. The target of the study was to maximize the hydrogen peroxide production.
The catalyst was uniformly diluted in quartz sand at different concentrations to investigate their effects
on the direct synthesis. The amount of catalyst and the distribution of the catalyst along the bed were
optimized to obtain the highest possible yield. The distribution of the catalyst along the bed gave the
possibility to significantly improve the selectivity and production of hydrogen peroxide (up to 0.5%
under selected conditions). Higher production rate and selectivity were found when the catalyst concen-
tration was decreased along the bed from the top to the bottom as compared to the uniformly dispersed
catalyst. The H2/Pd ratio was found to be an important parameter that has to be investigated in the
hydrogen peroxide direct synthesis. The effect of a pretreatment of the catalyst with a solution of sodium
bromide and phosphoric acid was studied; the results showed how a catalyst pretreatment can lead to a
real green hydrogen peroxide synthesis in water. Some optimization guidelines are also provided.
Received 29th April 2013,
Accepted 1st July 2013
DOI: 10.1039/c3gc40811f
www.rsc.org/greenchem
because this is what nature does. In this paper we study the
direct synthesis as an intermediate option between using
1. Introduction
Green chemistry and Green Engineering are becoming year anthraquinone (traditional) and producing H2O2 without the
after year an important philosophy-in-practice in Chemical need of hydrogen (biomimicry and electrochemical reduction
Engineering research both in Academia and Industry. Hydro- of oxygen to superoxide).
gen peroxide and hydrogen peroxide direct synthesis are part
The market that will be of particular interest for the direct
of this philosophy. Hydrogen peroxide is an effective green synthesis includes oxidation processes like the propylene oxide
chemical, it is the simplest of the peroxides, the byproducts formation, pulp bleaching, electronics and wastewater treat-
are only oxygen and water and its oxidation potential is high. ment as examples.3 The epoxidation of propylene into propyl-
In a future vision of a sustainable world, hydrogen peroxide ene oxide is one of the most important petrochemical
will replace all those chemicals with high oxidation potential applications for H2O2 produced via direct synthesis.7 A concen-
that need a complicated and expensive waste treatment after trated H2O2 is necessary, because of the side reactions of H2O
their use in chemical synthesis.1–3 The current process to forming glycols with the propylene oxide. The use of H2O as a
produce hydrogen peroxide cannot be considered green in solvent instead of methanol for the PO process reduces the
terms of the organics used (i.e. anthraquinone); for that side products considerably over the use of methanol (that
reason the direct synthesis is becoming an interesting possi- forms methoxy propanols). On the other hand, methanol is
bility to produce hydrogen peroxide for in situ applications.1–3 convenient as it dissolves both propylene and propylene oxide,
There are other options, such as the electrochemical reduction while water creates a two-phase liquid–liquid reaction.
of oxygen via a four-step-one-electron reaction into a super-
Either for the traditional anthraquinone synthesis or the
direct synthesis, the concentrations of H2O2 are normally low
oxide, hydrogen peroxide and finally a hydroxyl radical.4–6
The last process is excellent in terms of atom economy, not (i.e. below 1% wt typically) and the need of a downstream con-
needing hydrogen, and it is probably one of the best options, centration is there in both cases. Producing H2O2 at high con-
centrations directly has the drawback of a reduced selectivity,
as the active metal (e.g. palladium) causes both hydrogenation
aDepartment of Chemical Engineering, Åbo Akademi University, Turku/Åbo,
and decomposition. For decomposition a high H2O2 concen-
FI-20500, Finland. E-mail: bpierdom@abo.fi; Fax: +358 2 215 4479;
tration is extremely inconvenient, as decomposition is a first
Tel: +358 2 215 4555
bDepartment of Chemical Engineering and Environmental Technology, University of
Valladolid, Valladolid, Spain. E-mail: jgserna@iq.uva.es
order reaction with H2O2. Furthermore, the percentage of
decomposition during the concentration operations, i.e. via
2502 | Green Chem., 2013, 15, 2502–2513
This journal is © The Royal Society of Chemistry 2013

View Article Online
Green Chemistry
Paper
vacuum distillation, is considerably lower than by reaction. the possibility of enhancing the concentration of the hydrogen
That is why in the traditional process, 30%, 50% and 70% peroxide by a subsequent distillation column.
H2O2 solutions are produced via distillation. The direct syn-
In this work we demonstrate that: (1) the H2/Pd ratio plays
thesis avoids transportation costs of concentrated hydrogen an important role during the direct synthesis15,29; (2) the
peroxide and wastewater treatment. So far, the auto-oxidation control of reactor and reactor conditions shifts the balance in
process is still the most popular process to produce hydrogen favour of productivity.
peroxide and the continuous direct synthesis process is not
sufficiently developed.2
In the last two decades, industry8–11 and academia12–27
tried hard to make progress in the direct synthesis research
2. Results and discussion
A set of 75 experiments was designed in order to study the
effect of the pressure, liquid and gas flow rates, temperature,
amount of catalyst and volume of bed. The trickle bed reactor
operated in continuous mode modifying the experimental con-
ditions. The bed was charged with the quartz sand plus cata-
lyst mixture and operated under different conditions. The
stabilization time ranged between 30 and 90 min depending
on the flow rate and the magnitude of the step. After the
system stabilizes, three measurements were taken and the
average was calculated. Table 1 contains the experimental con-
ditions and the average results.
focusing on catalyst design, reactor systems and modeling.
The results obtained are promising and the combination of
different kinds of expertise will lead to novel solutions and to
a big step further in the field.
Catalyst design and improvement of them in the direct syn-
thesis covers around 80% of the publications. Usually noble
metals were chosen as the active phase, and monometal-
lic12,17,28,29 and bimetallic18,22,30,31 catalysts based on palla-
dium were employed. The supports for the active phase cover
different inorganic and organic materials. Modifications to the
supports were also investigated. Recently the attention paid to
different kinds of continuous reactors became strong. Micro-
reactors,20,21,32 continuous membrane reactors,26 upflow reac-
tors and trickle bed reactors (TBRs)13–15,33 showed the
possibility to continuously produce hydrogen peroxide in a
promising efficient way. The issues that are still unsolved in
the direct synthesis are selectivity, concentration and catalyst
deactivation. The selectivity and concentration of hydrogen
peroxide are intimately related, as it is difficult to obtain high
concentrations of hydrogen peroxide together with high
selectivity.14 When halides and acids are used as promoters,
selectivity increases considerably but leaching of the active
metal may occur. In the same way, at high concentrations of
H2O2, the selectivity is usually low and vice versa, as indicated
before. This fact is due to the nature of the catalyst; the active
metal/metals active for the direct synthesis are also active for
the direct formation of water, hydrogenation and decompo-
sition of hydrogen peroxide.29,34 The direct formation of water
is difficult to measure; hydrogenation was well demonstrated
as a key reaction during the direct synthesis and the decompo-
sition gives a small contribution to the overall reaction rate as
described previously.29,34 The numerous investigations on
The effects of the main variables have been divided accord-
ing to the physical–chemical influence on the reaction out-
comes: primarily, mass transfer, kinetics and catalysis.
2.1. Mass transfer related influences
The reactions take place at the catalyst surface. The gas
reagents, i.e. H2 and O2, must dissolve from a gas phase into a
liquid phase and then be transported to the catalyst.
Oxygen is in excess over the stoichiometric need versus
hydrogen; therefore, the limiting reagent due to mass transfer
limitations is hydrogen.
2.1.1. Influence of H2 partial pressure. The variation of
the H2 partial pressure changes the equilibrium solubility of
H2 in the liquid phase, which modifies the mass transfer
differential concentration and ultimately the concentration of
H2 available at the surface of the catalyst.
The solubility of H2 in water between 5 and 35 °C and at
hydrogen partial pressures between 0.27 and 1.45 bara were
simulated using Aspen Plus ONE7.1 using the PRSK thermo-
dynamic package35 as shown in Table 1.
2.1.1.1. Influence of H2 inlet concentration. Due to safety
direct synthesis try to understand the phenomena that regulate reasons, it is necessary to operate away from the flammability
the hydrogenation reaction aiming to switch it off.18
region, and this implies using inlet H2 concentrations lower
Here we investigate the direct synthesis in a down-flow than 4.6 mol%. As an additional safety measure, in the set-up,
trickle bed reactor, using water as a reaction medium. The we operated with the H2 premixed with CO2 in the cylinder. To
catalyst (due to its low amount) was dispersed in quartz sand. increase the mass transfer, the maximum H2 in the cylinder
The aim of the work is to set variables in order to maximize was 5 mol% (4 mol% in the final reaction mixture, under the
the reaction over a commercial catalyst. The first part of the flammability restriction). Then, the experimental system was
work deals with the conventional way to optimize the pro- limited by the maximum reaction pressure up to 28 barg
ductivity, finding the suitable temperature, pressure and gas (because that of the gas cylinder mixture 5% H2/CO2 is up to a
and liquid flow rates. The second part is focused on unconven- maximum of 32 barg).
tional ways to optimize the reactor parameters, such as transi-
To reveal the effect of the inlet concentration, we carried
ent experiments (changing the liquid feed during the out two experiments (runs #01 and #02) using 2.5 mol%-H2/
experiments) and changing the concentration of the catalyst 97.5 mol%-CO2 and 5 mol%-H2/95 mol%-CO2 cylinders,
along the catalytic bed. The third part takes into consideration respectively.
This journal is © The Royal Society of Chemistry 2013
Green Chem., 2013, 15, 2502–2513 | 2503

Table 1 Design of the experiment table for the study of H2O2 direct synthesis in a trickle bed reactor. Reactant medium: H2O2 + 0.004 M NaBr + 0.003 M H3PO4
Total
H2 inlet
conc.
H2 inlet
press.
(bara)
Catalyst
amount
(% Pd/C) (mg cat)
press. Temp. LFR
GFR
H2 inlet flow
H2 solubility Cat.
Bed volume
(mL)
Productiv.
Yield
(%)
H2O2 conc.
(% wt/vol)
Run (barg) (°C)
(mL min−1
)
(mL min−1
)
(% mol) (mmol min−1
)
(mol mol−1
)
(mmol min−1
)
#01
#02
#03
#04
#05
#06
#07
#08
#09
#10
#11
#12
#13
#14
#15
#16
#17
#18
#19
#20
#21
#22
#23
#24
#25
#26
#27
#28
#29
#30
#31
#32
#33
#34
#35
#36
#37
#38
#39
#40
#41
#42
#43
#44
#45
10
10
10
28
28
28
28
28
28
28
28
28
28
28
28
28
28
28
28
15
15
15
15
15
15
15
5
25
25
25
25
25
25
25
25
25
25
25
25
25
5
10
15
20
35
40
5
10
15
5
15
25
35
5
15
25
35
5
15
25
35
5
1.00
4.00
4.00
2.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
2.5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
37.5
75.0
37.5
0.27
0.55
0.55
1.45
1.45
1.45
1.45
1.45
1.45
1.45
1.45
1.45
1.45
1.45
1.45
1.45
1.45
1.45
1.45
0.80
0.80
0.80
0.80
0.80
0.80
0.80
0.30
0.30
0.30
0.30
0.80
0.80
0.80
0.80
0.30
0.30
0.30
0.30
0.80
0.80
0.80
0.80
0.30
0.30
0.30
2.80E-06
2.80E-06
2.80E-06
1.00E-05
1.00E-05
1.00E-05
1.00E-05
1.00E-05
1.00E-05
1.00E-05
1.00E-05
1.00E-05
1.00E-05
9.94E-06
9.83E-06
9.82E-06
9.90E-06
1.05E-05
1.09E-05
4.19E-06
4.25E-06
4.34E-06
4.19E-06
4.34E-06
4.62E-06
5.01E-06
1.17E-06
1.29E-06
1.45E-06
1.63E-06
4.19E-06
4.34E-06
4.62E-06
5.01E-06
1.17E-06
1.29E-06
1.45E-06
1.63E-06
4.19E-06
4.34E-06
4.62E-06
5.01E-06
1.17E-06
1.29E-06
1.45E-06
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
150
150
150
150
150
150
150
150
75
75
75
75
75
75
75
75
75
75
75
30
30
80
80
80
80
80
80
80
80
40
40
40
40
40
40
40
40
40
40
40
80
80
80
80
80
80
80
80
80
80
80
80
80
80
80
80
80
80
80
80
80
80
80
80
80
80
11.6
27.8
7.5
30.9% 0.039%
37.1% 0.095%
19.9% 0.025%
27.3% 0.875%
32.0% 0.513%
34.0% 0.272%
36.1% 0.193%
38.0% 0.152%
37.6% 1.205%
48.7% 0.780%
53.2% 0.426%
62.1% 0.331%
63.4% 0.254%
26.3% 0.211%
28.4% 0.227%
30.2% 0.242%
46.9% 0.376%
48.4% 0.388%
52.6% 0.421%
26.9% 0.096%
31.0% 0.111%
19.8% 0.071%
61.6% 0.220%
64.1% 0.229%
64.1% 0.229%
61.0% 0.218%
50.7% 0.059%
58.8% 0.068%
62.3% 0.072%
63.8% 0.074%
60.3% 0.215%
63.4% 0.226%
63.2% 0.226%
60.0% 0.214%
53.5% 0.062%
61.4% 0.071%
65.0% 0.075%
64.7% 0.075%
55.7% 0.199%
61.1% 0.218%
64.1% 0.229%
64.5% 0.230%
60.7% 0.070%
59.1% 0.068%
64.1% 0.074%
1.00
1.00
0.25
0.50
1.00
1.50
2.00
0.25
0.50
1.00
1.50
2.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
235.6
235.6
235.6
235.6
235.6
235.6
235.6
235.6
235.6
235.6
235.6
235.6
235.6
235.6
235.6
235.6
105.0
105.0
105.0
105.0
105.0
105.0
105.0
34.0
34.0
34.0
34.0
105.0
105.0
105.0
105.0
34.0
34.0
34.0
64.4
75.5
80.1
85.0
89.4
88.6
114.7
125.4
146.2
149.4
62.1
66.9
71.2
110.5
114.1
123.9
28.3
32.6
20.8
64.7
67.3
67.3
64.1
17.2
20.0
21.2
21.7
63.3
66.6
66.4
63.0
18.2
20.9
22.1
22.0
58.5
64.2
67.3
67.8
20.6
20.1
21.8
30
150
150
150
150
150
150
150
150
100
100
100
100
100
100
100
100
75
5
5
5
15
15
15
15
5
5
5
15
25
35
5
15
25
35
5
5
34.0
15
15
15
15
5
5
5
105.0
105.0
105.0
105.0
34.0
34.0
34.0
75
75
75
75
75
75
15
25

Table 1 (Contd.)
Total
H2 inlet
conc.
H2 inlet
press.
(bara)
Catalyst
amount
(% Pd/C) (mg cat)
press. Temp. LFR
GFR
H2 inlet flow
H2 solubility Cat.
Bed volume
(mL)
Productiv.
Yield
(%)
H2O2 conc.
(% wt/vol)
Run (barg) (°C)
(mL min−1
)
(mL min−1
)
(% mol) (mmol min−1
)
(mol mol−1
)
(mmol min−1
)
#46
#47
#48
#49
#50
#51
#52
#53
#54
#55
#56
#57
#58
#59
#60
#61
#62
#63
#64
#65
#66
#67
#68
#69
#70
#71
#72
#73
#74
#75
#76
5
15
15
15
15
28
28
28
28
28
28
28
28
28
28
28
28
28
28
28
28
28
28
28
28
28
28
28
28
28
28
35
5
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
34.0
105.0
105.0
105.0
105.0
235.6
235.6
235.6
235.6
235.6
235.6
235.6
235.6
235.6
235.6
235.6
235.6
235.6
235.6
235.6
235.6
235.6
235.6
235.6
235.6
235.6
235.6
235.6
235.6
235.6
235.6
0.30
0.80
0.80
0.80
0.80
1.45
1.45
1.45
1.45
1.45
1.45
1.45
1.45
1.45
1.45
1.45
1.45
1.45
1.45
1.45
1.45
1.45
1.45
1.45
1.45
1.45
1.45
1.45
1.45
1.45
1.45
1.63E-06
4.19E-06
4.34E-06
4.62E-06
5.01E-06
9.83E-06
9.82E-06
9.90E-06
1.00E-05
1.03E-05
1.05E-05
9.94E-06
9.82E-06
1.00E-05
1.05E-05
9.94E-06
9.82E-06
1.00E-05
1.05E-05
9.94E-06
9.82E-06
1.00E-05
1.05E-05
9.94E-06
9.82E-06
1.00E-05
1.05E-05
9.94E-06
9.82E-06
1.00E-05
1.05E-05
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
5%
75
75
75
75
75
150
150
150
150
150
150
100
100
100
100
75
80
40
40
40
40
80
80
80
80
80
80
80
80
80
80
80
80
80
80
80
80
80
80
40/40
40/40
40/40
40/40
23.4
72.3
77.8
55.6
60.5
73.7
76.6
78.9
80.0
79.2
78.8
83.3
92.0
91.9
82.7
127.9
116.4
112.6
98.7
124.3
97.5
76.7
61.9
132.9
135.3
122.4
115.3
68.9% 0.080%
68.9% 0.246%
74.1% 0.265%
53.0% 0.189%
57.7% 0.206%
31.3% 0.251%
32.5% 0.260%
33.5% 0.268%
34.0% 0.272%
33.6% 0.269%
33.5% 0.268%
35.4% 0.283%
39.0% 0.313%
39.0% 0.312%
35.1% 0.281%
54.3% 0.435%
49.4% 0.396%
47.8% 0.385%
41.9% 0.336%
52.7% 0.423%
41.4% 0.332%
32.5% 0.261%
26.3% 0.226%
56.4% 0.452%
57.4% 0.460%
51.9% 0.416%
48.9% 0.392%
62.6% 0.502%
60.0% 0.481%
53.8% 0.431%
49.7% 0.398%
15
25
35
10
15
20
25
30
35
5
15
25
35
5
15
25
35
5
15
25
35
5
75
75
75
300
300
300
300
50/25
50/25
50/25
50/25
15
25
35
5
15
25
35
37.5/25/12.5 26.6/26.6/26.6 147.5
37.5/25/12.5 26.6/26.6/26.6 141.4
37.5/25/12.5 26.6/26.6/26.6 126.7
37.5/25/12.5 26.6/26.6/26.6 117.2

View Article Online
Paper
Green Chemistry
Fig. 1 H2O2 production vs. total pressure for 150 mg 5% Pd/C and 80 mL SiO2
Fig. 2 H2O2 production vs. liquid flow rate for a 5% Pd/C and SiO2 bed at 28
barg and 25 °C. H2O2 production for different catalyst distributions in the trickle
bed and 4% mol-H2. H2O2 production (μmol min−1): ( ) 5 °C, ( ) 15 °C and (
□
◊
△
)
◆
■
▲
●
◊
25 °C. H2O2 Yield (%): ( ) 5 °C, ( ) 15 °C, ( ) 25 °C and ( ) 35 °C.
bed reactor: (x) 300 mg of Pd/C in 80 mL bed, ( ) 150 mg of Pd/C in 80 mL
△
○
)
bed, (*) 100 mg of Pd/C in 80 mL bed, ( ) 75 mg of Pd/C in 80 mL bed, (
□
75 mg of Pd/C in 40 mL bed, ( ) 50/25 mg of Pd/C in 40/40 mL bed and (+)
The productivity increased considerably (more than double)
when the H2 was doubled, producing 27.8 μmol-H2O2 min−1
vs. 11.6 μmol-H2O2 min−1 with 2 mol%-H2, as more H2 is dis-
solved and available for the H2O2 direct synthesis. However,
the efficiency of the process decreased, as can be seen by com-
paring the yields. The yield using 4 mol%-H2 (0.55 bara of H2)
was 37.1% compared to 30.9% at 2 mol%-H2 (0.27 bara of H2).
This means that more H2O2 is hydrogenated and that H2O pro-
duction by direct synthesis is enhanced proportionally.
37.5/25/12.5 mg of Pd/C in 26.6/26.6/26.6 mL bed.
can be seen, a ratio of O2/H2 = 9 leads to yields of 31.0% and a
production of 11.6 μmol-H2O2 min−1 while O2/H2 = 4.5 leads
to yields of 62.3% and a production of 22.2 μmol-H2O2 min−1
at the same partial pressure of H2 of 0.27 barg.
2.1.2. Influence of flow rate. The variation of flow rate
influences the individual mass transfer coefficients, as these
coefficients are a function of Reynolds and Schmidt numbers.
On the other hand, the liquid flow rate has a linear effect on
the liquid residence time, which governs the time in which the
catalyst and the dissolved reagents are in contact.
2.1.2.1. Liquid flow rate. The increase of the flow rate has a
direct effect on the mass transfer, but also on the concen-
tration of the H2O2 produced. It must be considered that the
hydrogen gas stream at the inlet was constant. This means
that an increase in the production implies the same increase
in the yield, as the same H2 stream is introduced in all cases.
The higher flow rate of liquid used dilutes the H2O2 being pro-
duced, causing a decrease in concentration.
2.1.1.2. Influence of total reaction pressure. The system was
investigated at three different total pressures using 150 mg of
5% Pd/C catalyst, ensuring that the system was mass transfer
controlled (runs #06, #23–#30 and #51–#56). Fig. 1 shows that
increasing pressure increased the quantity of H2O2 formed. As
can be seen, the effect of temperature was almost negligible
compared to the pressure effect. This means that the increase
in kinetics does not increase the production, because of mass
transfer limitations. The increase in the total pressure implies
an increase in the hydrogen flow rate at the reactor entrance,
as we have decided to maintain the same gas hydrodynamics
of 4 mL min−1 of gas inside the TBR in this work. The yield in
this case has to be analysed. The yield of the reaction is con-
stant at 50% changing from 5 barg to 15 barg, but it decreases
down to 35% at 28 barg. The effect shows how between 5 and
15 barg the system is almost controlled by the mass transfer of
hydrogen, while at 28 barg all the hydrogen is not consumed,
or at least it is not consumed for producing H2O2. Again, as in
the previous section, the excess of H2 dissolves and is available
for the undesirable hydrogenation at the end of the reactor
(where the concentration of H2O2 is higher), as will be shown
later in section 2.1.2. Therefore, the increase in the total
pressure is an option to increase the mass transfer, but care
must be taken to control the kinetics, especially when higher
concentrations of H2O2 are available.
It can be seen in Fig. 2 how for 150 mg 5% Pd/C in 80 mL
◊
bed ( ) it is demonstrated that the increase of the liquid vel-
ocity (or flow rate) increases the mass transfer and so improves
the production. However the effect is not completely indepen-
dent, as the increase in flow rate causes a decrease of the H2O2
concentration and reduces the undesired hydrogenation and
decomposition. Also, due to a decrease in the residence time,
and depending on the rate of direct synthesis and direct com-
bustion reactions, the final H2O2 may change.
To clarify the effect, we made some tests using different
amounts of catalyst and bed sizes (runs #04–#13, #58, #62,
#66, #70 and #74), as it will be explained later in more detail.
For this case, it can be observed that for 75 mg 5% Pd/C in
○
40 mL bed ( ), the production was 125.4 μmol-H2O2 min−1
,
2.1.1.3. Influence of the O2/H2 ratio. Several authors stated
that the O2/H2 ratio must be high to ensure minimization of
H2O2 decomposition and an oxidizing atmosphere. However,
we carried out two experiments to show this fact (runs #01 and
#37); an excess of O2 is not as profitable as expected, probably
because of the inhibition of certain sites on the catalyst. As
◊
ca. 50% higher compared to 150 mg 5% Pd/C in 80 mL bed ( )
of 80.1 μmol-H2O2 min−1
.
This would mean that, even with half of the quartz bed, the
quantity of H2 transferred would be higher. Nevertheless, it is
clear that using a shorter bed, with the same catalyst
2506 | Green Chem., 2013, 15, 2502–2513
This journal is © The Royal Society of Chemistry 2013

View Article Online
Green Chemistry
Paper
Fig. 4 H2O2 production vs. catalyst amount in bed, 5% Pd/C and 80 mL SiO2
Fig. 3 H2O2 concentration vs. liquid flow rate for a 5% Pd/C and SiO2 bed at
28 barg and 25 °C. H2O2 production for different catalyst distributions in the
bed at 28 barg at 5 °C, 15 °C, 25 °C and 35 °C.
◊
trickle bed reactor: (x) 300 mg of Pd/C in 80 mL bed, ( ) 150 mg of Pd/C in
△
80 mL bed, (*) 100 mg of Pd/C in 80 mL bed, ( ) 75 mg of Pd/C in 80 mL bed,
○
□
surface of the catalyst, i.e. at the palladium active sites. Consid-
ering the kinetics of the direct synthesis, as has been shown in
several studies,29,37 the number of active sites is included in
the reaction rate, mainly considering the first-order influence.
For the case of the continuous TBR, the total amount of cata-
lyst is doubly important for: the total number of sites (consid-
ering the whole reactor) and also their distribution in the
support. The effect of the catalyst concentration on the
support and its distribution has not been studied previously,
and it is presented in this paper for the first time.
We conducted experiments with four catalyst amounts with
80 mL of SiO2 support bed (runs #06 and #51–#75) at four
temperatures. In general, the increase in the concentration of
catalyst in the bed reduces the production at high temp-
eratures (e.g. 25 °C and 35 °C), as shown in Fig. 4, which is
explained by the increase in the H2O2 decomposition and
hydrogenation rates.
However, at low temperatures, the effect is not clear, as the
production is higher at 75 mg and 300 mg of 5% Pd/C catalyst
at 5 °C and 15 °C, respectively. This effect can be explained
similarly, considering that at low temperatures the decompo-
sition and hydrogenation are minimized; therefore, increasing
the catalyst amount increases these undesired reactions. On
the other hand, increasing the catalyst implies more H2 con-
sumption and more H2O2 production, although probably the
overall selectivity might decrease, and the production
increases again. In this case, the highest production rates were
obtained at low temperatures.
(
) 75 mg of Pd/C in 40 mL bed, ( ) 50/25 mg of Pd/C in 40/40 mL bed and
(+) 37.5/25/12.5 mg of Pd/C in 26.6/26.6/26.6 mL bed.
concentration in the bed, the hydrogenation and decompo-
sition of H2O2 are considerably suppressed, increasing the
efficiency of the process. In this study, we could not measure
the H2 in the off-gas due to experimental limitations, but this
experiment confirms the statement. Moreover, using 75 mg of
catalyst with 80 mL of quartz, the production was in between
the cases just discussed, 112.6 μmol-H2O2 min−1, because a
larger bed implies more H2 transferred into the liquid phase
and enhanced hydrogenation, too.
Returning to the influence of the liquid flow rate related to
the decrease in the H2O2 concentration, we can observe that at
low flow rates, e.g. between 0.25 and 0.5 mL min−1 of H2O, the
effect of mass transfer is more pronounced and in Fig. 2 a step
change is observed at such a flow rate level, while at high flow
rates, the mass transfer effect is less important and the quasi-
linear increase in H2O2 production can be explained by the
H2O2 concentration decrease, as shown in Fig. 3.
2.1.2.2. Gas flow rate. The gas flow rate was optimized in
previous studies13–15 with a similar system, so the gas flow rate
was kept constant at 4 mL min−1 for all the experiments.
However, in order to check the influence we carried out two
experiments at 2 mL min−1 and 4 mL min−1 (runs #02 and
#03). At low gas flow rates, i.e. at 2 mL min−1, the yield was
19.9% and 7.45 μmol-H2O2 min−1 while at 4 mL min−1 the
yield was 37.1% and the production 27.8 μmol-H2O2 min−1. At
low gas flow rates, both the yield and production were lower,
indicating that low quantities of H2 were transferred. A further 2.3. Catalyst and catalyst bed improvement
increase in the gas flow rate may cause bad liquid distribution
Following the results obtained with different amounts of
and bad wetting of the catalyst particles in the TBR, as demon-
strated in previous studies, unless the liquid is increased pro-
portionally.36 Such a hydrodynamic study is out of the scope of
this work.
hydrogen fed inside the reactor experiments, the amount of
hydrogen was kept fixed and the catalyst amount and the
catalyst distribution inside the reactor were varied.
Experiments were performed with fixed gas flow rate as
reported above.
2.2. Kinetic study
2.3.1. Volume of the support bed. The catalyst was distrib-
2.2.1. Influence of catalyst amount. In the TBR, the inert uted with two different concentrations in the catalyst bed.
support acts to increase the surface area for gas transfer into 150 mg of catalyst were distributed in 100 g of quartz sand and
the liquid phase. Nevertheless, the reaction takes place at the fed inside the reactor and 75 mg of catalyst were diluted in
This journal is © The Royal Society of Chemistry 2013
Green Chem., 2013, 15, 2502–2513 | 2507

View Article Online
Paper
Green Chemistry
production of hydrogen peroxide, probably due to some
factors: (1) when the residence time in the catalytic bed is the
lowest, subsequent hydrogenation and decomposition of
hydrogen peroxide affect less the direct synthesis; (2) the
H2/Pd ratio plays an important role in the subsequent hydro-
genation after the H2O2 formation.
2.3.2. Concentration distribution of the catalyst in the
support bed. Following the results with low catalyst amounts
and considering that the H2/Pd ratio plays an important role
in the direct synthesis, three different catalyst bed arrange-
ments were designed and charged into the trickle bed reactor
to confirm the hypothesis reported above. Bed (A) consisted of
75 mg of catalyst uniformly distributed in 80 mL of quartz sand,
△
Fig. 5 H2O2 production and yield vs. liquid flow rate at different supports. (
)
▲
H2O2 produced and ( ) H2O2 yield at 28 barg, 25 °C, 150 mg 5% Pd/C/80 mL
□
■
quartz sand, ( ) H2O2 produced and ( ) H2O2 yield at 28 barg, 25 °C, 75 mg 5% while bed (B) was composed of 40 mL of quartz on the bottom
Pd/C/40 mL quartz sand.
with 25 mg of catalyst perfectly mixed together and above that
50 mg of catalyst dispersed in 40 mL of quartz sand and, finally,
bed (C) was divided into three parts of 26.6 mL of quartz sand
50 g of quartz sand and fed into the reactor (to fill completely
the reactor in the latter case 50 g of quartz sand were used).
Experiments were performed at 25 °C varying the liquid flow
rate and keeping constant the gas flow rate at 4 mL min−1. The
results are illustrated in Fig. 5.
and from the bottom the quantity of the catalyst was 12.5, 25
and 37.5 mg in each part of the quartz sand. Visual distribution
of the catalytic bed is reported in Fig. 6. First of all, the concen-
tration of hydrogen peroxide at the outlet and the yield of H2O2
decreased linearly with increasing the temperature.
The temperature, pressure and gas flow rate were kept con-
stant and the only parameter changed during the experiments
was the liquid flow rate. An increase in the liquid flow rate
always corresponds to an increase on the production of hydro-
gen peroxide as can be seen. The increase of H2O2 production
is more pronounced in the range between a liquid flow rate of
0.25 and 1 mL min−1, while between 1 and 2 mL min−1 the
increase is less pronounced but still remarkable. It seems that
a plateau is not reached, meaning that there is a possibility to
further enhance the production of H2O2. The fact that an
increasing liquid flow rate increases the quantity of H2O2 is
probably due to two factors: (1) the mass transfer of hydrogen
is increased with increasing the liquid flow rate; more H2
reacts to produce more water or hydrogen peroxide; (2) hydro-
genation and decomposition of H2O2 are decreased due to the
lower contact time of the liquid phase with the catalyst. The
same trend can be found in the results related to the yield of
H2O2 (solid black triangles), where the yield constantly
increases with the liquid flow rate. The simultaneous analysis
of both the results of the production and the yield of H2O2
seems to confirm the two hypotheses made above. A confir-
mation of the hypothesis made above originates from the
results obtained halving the catalyst bed in the trickle bed.
The experiments conducted with 75 mg of active catalyst and
40 mL of quartz sand showed the same trend as the previous
experiment with 150 mg of catalyst in 80 mL of quartz sand.
Interestingly, the production of H2O2 and the yield were higher
compared to the former experiments. These results can be
explained with the fact that H2O2 is less hydrogenated as the
reaction is performed in the bed with less catalyst. Usually
when the quantity of the catalyst is diminished, the amount of
the products is less. For the hydrogen peroxide direct synthesis
that counts parallel and consecutive reactions, the results are
the opposite. The lower the catalyst amount the higher is the
The idea was developed to confirm that there is an
optimum H2/Pd ratio and if it is kept constant or above some
values (without decreasing it too much), hydrogenation can be
suppressed. The results in Fig. 7 confirmed our hypothesis
that working with a gradient of the catalyst concentration
along the bed can lead to an enhancement on both production
and yield of H2O2. The results showed that bed (C) (empty tri-
angles = production of H2O2, solid black triangles = yield%)
was the one giving the best results followed by bed (B) (empty
squares = production of H2O2, solid black squares = yield%)
and bed (A) (empty diamonds = production of H2O2, solid
black diamonds = yield%). From those preliminary results
from bed (A) to bed (C), there is an average enhancement of
10% in the yield and of 20 μmol min−1 of H2O2 produced only
with a different disposition of the catalyst bed. The best result
was found at 5 °C, as the yield was improved by a value of 10%
from 50% with bed (A) to 60% with bed (C). The production
rate was 130 μmol min−1 with bed (A) and 150 μmol min−1
Fig. 6 Visual distribution of the catalytic bed.
2508 | Green Chem., 2013, 15, 2502–2513
This journal is © The Royal Society of Chemistry 2013

View Article Online
Green Chemistry
Paper
Fig. 7 H2O2 production and yield vs. temperature at different catalyst concen-
trations in the bed support with 4 mL min−1 gas flow rate and 1 mL min−1
Fig. 8 H2O2 production and yield vs. time at different unsteady Br-inlet con-
□
△
centrations: ( ) H2O2 produced and ( ) H2O2 yield at 28 barg, 25 °C, 5% Pd/C :
quartz sand = 12.5, 25 and 37.5 mg : 26.6, 26.6 and 26.6 mL.
△
▲
) H2O2 produced and yield at 28 barg 5% Pd/C : quartz
liquid flow rate: (
,
□
■
) H2O2 produced and yield at 28 barg 5% Pd/
sand = 75 mg : 80 mL, (
,
◊
◆
C : quartz sand = 25 and 50 mg : 40 and 40 mL, ( , ) H2O2 produced and yield
at 28 barg 5% Pd/C : quartz sand = 12.5, 25 and 37.5 mg : 26.6, 26.6 and
26.6 mL.
production and yield remained constant for around
400 minutes without any additives. After that (at 850 minutes)
both the H2O2 production and yield decreased faster. To
understand the activity reported above, palladium and
bromide concentrations were analyzed during all the experi-
ments and coupled with H2O2 production rate results. ICP
measurements for Pd and Br− revealed the presence of
330 ppm of bromide and 0.46 ppm of palladium at the outlet
in the presence of an acidic–bromide solution. However when
the feeding of the acidic–bromide solution was stopped, a con-
centration of 5 ppm min−1 of Br- or lower and 0.017 ppm of Pd
was coming out from the outlet. The value of 5 ppm represents
around 1.5% of the bromide fed inside the reactor when the
experiments were in the acidic–bromide feeding zone. This
means that when the acidic solution is fed into the reactor,
there is probably an adsorption of the bromide on the carbon
and on the inert particles (SiO2) surface. When the acidic–
bromide solution feeding was stopped, probably the bromide
adsorbed on the catalyst surface (and inert particle surface)
was slowly released, maintaining with its effect some of the
catalytic properties of the palladium to produce hydrogen per-
oxide. A detailed analysis of the palladium and bromide con-
centrations at the outlet of the reactor is reported in Fig. 9.
From the results analysis, it can be said that the Pd leach-
ing is promoted by the bromide and the acid contained in the
reaction liquid mixture. In the experiment zone where the
bromide–acidic solution was fed, a total amount of 0.1 mg of
with bed (C). The amount of the active catalyst was the same
and if the hydrogenation was not influenced by the local H2/
Pd ratio, the results should have been exactly the same. These
results showed that the hypothesis made was correct because
with this different disposition, the hydrogenation is kept lower
in bed (C) than in bed (A). All the results obtained with bed (B)
were showing the same trend as in the former cases: the
decreasing of the production and yield of H2O2 when the
temperature was increased.
2.3.3. Bromide concentration. Bromide plays an important
role in the direct synthesis, but its effect on the reaction is still
not completely understood. The experiments were performed
with the same catalyst bed that gave the most promising results
(e.g. three parts of 26.6 mL of quartz sand and from the bottom,
the quantity of the catalyst was 12.5, 25 and 37.5 mg in each
quartz sand portion). The experiment was conducted as follows:
the feeding was only water until 220 minutes, then it was
switched to the mixture water + acid + NaBr until 450 minutes
and then again to water alone; the results are shown in Fig. 8.
No H2O2 was detected at the outlet when the feeding was
only water. This implicates two possibilities: (1) there is no
direct synthesis if the bromide and the acid are not present in
the reacting medium; (2) the reaction happens, but the sub-
sequent formation of water is faster and the result is that all
H2O2 produced is converted into water. Studies in a different
reactant medium are needed to identify which one of the pos-
sibilities is the real one. Suddenly, when the feeding was
switched to the reactant medium mixture (e.g. water + acid +
NaBr), the production of H2O2 is remarkable and reaches the
maximum value after 120 minutes, i.e. the same value of H2O2
in terms of production and yield obtained in another experi-
ment with the same disposition of the catalyst bed (excellent
reproducibility). After we confirmed that, with the reaction
medium mixture we used for all experiments, it is possible to
obtain reproducible results, the feeding was again switched to
pure water and H2O2 production and yield were halved in
50 minutes. The real surprising result was that the H2O2
■
◆
Fig. 9 Palladium ( , red) and bromide ( , light blue); the same conditions of
the experiment reported in Fig. 8.
This journal is © The Royal Society of Chemistry 2013
Green Chem., 2013, 15, 2502–2513 | 2509

View Article Online
Paper
Green Chemistry
but the hydrogen peroxide production rate at the outlet was
still remarkable.
From the results depicted in Fig. 9, we have estimated that
running a reaction using 7.5 mg of Pd (e.g. 150 mg of 5% Pd/C
wt catalyst) using promoters, the palladium will leach in
approx. 10 days (considering 0.5 ppm and 1 mL min−1), while
if no promoters were used it would take ca. 100 days (consider-
ing 0.05 ppm and 1 mL min−1). We must recognise that the
productivity is a function of the promoter concentration, but a
convenient selection of the promoter concentration, target pro-
ductivity and amount of catalyst can provide a workable
process. The non-steady state operation of promoters can be
key to operating the process at maximum efficiency.
Fig. 10 TEM images of the fresh catalyst (a) and spent catalyst (b); focus
160 K.
The ICP and TEM results coupled together construct the
idea that the bromide acted as a palladium nanoparticle modi-
fier with two effects. The first effect was a surface restructura-
tion of the palladium nanoparticles making them more
effective for the direct synthesis. The second effect was the sin-
tering of the palladium nanoparticles (an effect noticed also in
the other experiments with only acidic–bromide feeding).
The first effect was probably the most important. The
interpretation of the data led to the supposition that the
surface restructuration was reversible: when the solution con-
tained phosphoric acid and bromide, we have one palladium
surface structure, whereas when the fed did not contain the
promoters we have another palladium surface structure. When
the surface was similar to the initial state (fresh catalyst), no
hydrogen peroxide could be detected. A possible speculation
can be that bromide interacts with some palladium sites, pro-
moting the palladium leaching and influencing the surface
structure, enhancing in this way the direct synthesis. When
the acidic–bromide solution was substituted with water the
surface went back to the initial (and thermodynamically
favoured) structure. A drawback of the acidic–bromide solution
was the quite significant leaching of the palladium (measured
using ICP-MS). Nevertheless, these discoveries can help to
identify new reaction conditions and can contribute to the
development of a new catalyst. Different experiments are
needed to identify the real effect of the bromide in the direct
synthesis. Catalytic properties need to be investigated during
all of the transient experiment performed, not only at the
beginning or at the end of the experiment. Catalyst pretreat-
ment is probably a tool that has to be investigated deeply to
understand how the direct synthesis can be improved with this
new method.
Fig. 11 Palladium nanocluster diameter before and after the direct synthesis.
Pd leached from the catalyst. The catalyst lost 2.5 wt% of Pd in
around four hours. After the bromide–acidic solution was
stopped, Pd leaching was almost zero (0.2%) in 8 hours. It is
worth mentioning that previous calculation studies showed that
a pretreatment with acids and halides weaken the interaction
between the metal surface and the reagents (i.e. changing the
metal surface of the palladium cluster with the leaching).38
The TEM analysis revealed how the palladium nanocluster
diameter changed during the reaction (Fig. 10). Fig. 10a rep-
resents the catalyst before the reaction and Fig. 10b represents
the catalyst after the reaction. A sintering effect is shown in
Fig. 10b compared to Fig. 10a.
It is interesting to notice from the TEM analyses (Fig. 10
and 11) how the distribution of the Pd nanoparticles changed:
the average diameter before the reaction was around 3 nm
with a narrow distribution and that after the reaction was
around 12 nm with a broad distribution. The results of the ICP
and TEM analyses allow for some considerations. It is clear
that the bromide–acidic solution (and probably the H2O2
formed during the reaction) modifies the palladium nanoparti-
cles. The leaching of the palladium happened mostly when
bromide and acid are in the solution fed. Under these con-
ditions, the hydrogen peroxide formed reached the maximum
production rate of all the experiments (Fig. 8 and 9). The inter-
esting part of the experiment was the second half: when the
bromide–acidic solution was substituted with water, the H2O2
production rate was halved (Fig. 8 and 9). The bromide and
palladium concentrations after only water had been fed were
very low (5 ppm of bromide and 0.0166 ppm of palladium),
2.4. Optimization guidelines
As demonstrated in this research, the production of H2O2 in
continuous mode is a feasible alternative using a TBR. The key
outcome variables are the H2O2 molar flow production, the
concentration of H2O2, the total H2 conversion and selectivity.
In this work, due to experimental limitations, we have not
measured conversion and we have inferred the selectivity
through the analysis of the yield. Although it does not give the
same information, it is enough to understand the system and
the extract conclusions on how to optimize it.
2510 | Green Chem., 2013, 15, 2502–2513
This journal is © The Royal Society of Chemistry 2013

View Article Online
Green Chemistry
Paper
First, as can be seen, the maximum productivity (run #13,
149.4 μmol-H2O2 min−1), yield (run #48, 74.1%) and H2O2 con-
centration (run #09, 1.205% wt) were obtained with 75 mg 5%
Pd/C and 40 mL bed. This implies a recommended concen-
tration to be used of ca. 0.1 mg-Pd mL−1-bed that, for the
quartz sand bed used in this case (0.2–0.8 mm particle size),
can be estimated as between 3.2 and 12.7 mg-Pd m−2-surface
area of the bed.
Second, the maximum productivity and the maximum con-
centration were obtained at the maximum pressure screening,
Fig. 12 H2O2 flow sheet for concentration of H2O2 using Aspen Plus ONE 7.1.
i.e. 28 barg (PH = 1.16 bara) and 25 °C, while the critical
2
parameter in this case was the liquid flow rate. Furthermore,
high liquid flow rates implied the maximum productivity
of 149.4 μmol-H2O2 min−1, because of a combination of
increased mass transfer and decreasing H2O2 concentration
(0.254% wt) therefore decreased the decomposition/hydrogen-
ation effect, while, on the other hand, the maximum concen-
tration of 1.205% wt was obtained at 0.25 mL min−1 of liquid,
with only 88.6 μmol-H2O2 min−1 produced. However, as it can
be seen between these two experiments, the increase in the
concentration from 2 mL min−1 down to 0.25 mL min−1 was
about 370%, while the decrease in production was only 41%.
According to our preliminary simulations, in order to compare
the energy requirements, to concentrate the H2O2 produced
from 0.254% or 1.205% wt to a value of 10% or 30% wt, the
energy flows, both cooling and heating up, are comparable for
each starting concentration (as shown in Table 2 and Fig. 12).
In fact, concentrating the H2O2 from 0.254% wt requires
10–11% more energy in cooling and heating than starting
from 1.205% wt. This means that the final optimization must
be a compromise between the costs of H2 consumed, the
molar flow of H2O2 produced and the energy requirements.
Third, in terms of yield, better yields were obtained as the
inlet flow rate of H2 was balanced with the concentration of
and 63.4% respectively; second, runs #72 and #73, where 28
barg and low catalyst densities were employed.
Fourth, considering reaction temperature, direct synthesis
can work between 10 and 60 °C as we have demonstrated in
this paper and many authors have proven before. Working
below 30 °C is not recommended from an industrial point of
view as it would be difficult to remove the heat produced in
the reaction (i.e. a low temperature utility is expensive). Never-
theless, working between 40 and 60 °C is affordable. These
temperatures are used also in traditional synthesis.
Finally, drafting an energy requirement, in both traditional
and direct synthesis vacuum distillation is needed, so similar
levels of energy consumption are expected between both pro-
cesses. However, the use of an organic solvent or the use of
methanol (for direct synthesis) can sometimes simplify the dis-
tillation (as they have lower boiling points) but they comprise a
high number of distillation columns required in series, as many
side products are formed (from the oxidation of the organics).
3. Conclusions
H2O2 obtained and most of the H2 was used in the reactor to A down-flow trickle bed reactor was set up to perform hydro-
produce H2O2, but not to hydrogenate it. Mainly, yields over gen direct synthesis continuously over a 5% Pd/C catalyst. The
60% coincide with the low and medium pressures in this reaction conditions were changed in order to maximize the
system. This confirms that it is really important to keep the hydrogen peroxide production. The catalyst was uniformly
maximum H2O2 value below a certain value of ca. 0.2–0.3% wt diluted in quartz sand in different concentrations to investi-
to minimize this undesired side reaction and maximize the gate their effects on the direct synthesis. The amount of cata-
production. There are two exceptions: first, runs #12 and #13, lyst and the distribution of the catalyst along the bed were
where 28 barg and high liquid flow rates were used (i.e. 1.5 optimized to obtain the highest possible yield. The distri-
and 2 mL min−1 respectively), where the yields reached 62.1 bution of the catalyst along the bed gave the possibility to sig-
nificantly improve the yield and the production of hydrogen
peroxide (up to 65% and 0.5% w/w respectively). This configur-
ation with a gradient of the catalyst concentration from the top
a
Table 2 Simulated heat flows for the distillation of H2O2
H2O2 inlet H2O2 final Heat flow Heat flow Heat flow Heat flow to the bottom in the reactor gave important results and infor-
conc.
(% wt )
conc.
(% wt )
condenser subcool
total cond. reboiler
mation on the H2O2 direct synthesis. Higher production rates
and yields were obtained as the catalyst concentration
decreased along the bed from the top to the bottom compared
to the uniformly dispersed catalyst. The H2/Pd ratio was found
to be an important parameter that has to be investigated in
the hydrogen peroxide direct synthesis. The effect of a pre-
treatment of the catalyst with a solution of sodium bromide
and phosphoric acid was studied; the results showed how a
catalyst pre-treatment can lead to an improvement of the
(kJ kg−1
)
(kJ kg−1
)
(kJ kg−1
)
(J kg−1
)
0.254%
0.254%
1.205%
1.205%
10.0%
30.0%
10.0%
30.0%
−4868
−4953
−4390
−4794
−540.0
−549.4
−486.7
−531.7
−5408
−5502
−4877
−5326
5435
5524
4928
5356
a The simulation has been carried out using Aspen Plus ONE7.1, the
distillation was carried out in a 12 theoretical stages RadFrac column
at 0.6 barg condenser with subcooling temperature of 35 °C and
thermodynamic property package Peng–Robinson.
This journal is © The Royal Society of Chemistry 2013
Green Chem., 2013, 15, 2502–2513 | 2511

View Article Online
Paper
Green Chemistry
direct synthesis of hydrogen peroxide in water, especially by
controlling the leaching of palladium and by the use of promo-
ters to increase selectivity.
The combination of catalyst development and chemical
reaction engineering is a key point to successfully understand/
improve the hydrogen peroxide direct synthesis. The catalyst
gradient along the reactor and the catalyst pre-treatment are
new ideas to improve hydrogen peroxide direct synthesis by
controlling the ratio of H2–H2O2 and H2–active metal along the
reactor. Hydrogen peroxide direct synthesis was improved with
these non-conventional experiments and reactor design.
Optimization guidelines were provided in order to increase the
hydrogen peroxide production.
Fig. 13 Schematic flow diagram of the trickle bed reactor (TBR) system; D-01 =
water and promoters inlet vessel, P-01 = HPLC pump, D-02 = H2 + CO2 cylinder,
D-03 = O2 cylinder, D-04 = liquid product collecting vessel, FCV-01 = gas flow
controller, FCV-02 = gas flow controller, LV = liquid sampling valve, R-01 = TBR
(RS = refrigerant supply, RR = refrigerant return).
4. Experimental section
4.1. Materials and methods
4.1.1. Materials. For this research the catalyst was micro-
particles of an activated carbon support with an average of
5 wt% Pd (2 to 5 nm size) purchased from Degussa. The cata-
lyst powder was weighed and mixed homogeneously with SiO2
particles creating a catalytic bed.
figures and tables were estimated using mixture densities
determined with the Redlich–Kwong–Soave with the Boston–
Mathias modification thermodynamic package in Aspen Pro-
perties software.35
Deionised water was used as the reaction solvent. NaBr
(Flucka) and H3PO4 (Sigma-Aldrich) were used as H2O2 promo-
ters dissolved in the water inlet stream. SiO2 microparticles of
150–200 μm (Sigma-Aldrich) were used as the trickle bed inert
support. Potassium iodide, H2SO4, Na2S2O3·5H2O 99.5%,
starch from Sigma–Aldrich and K2Cr2O7 (Riedel de Haën) were
used for titration analysis.
The TBR was made of AISI 316 stainless steel (Fig. 13),
60 cm long and 1.5 cm I.D. (ca. 105 mL) passivized using nitric
acid 30% wt during 4 hours to prevent H2O2 decomposition
due to iron catalysis. The bed was composed of 2 cm of glass
wool at the bottom and up to 110 g of quartz sand
(ca. 80–100 mL). The reactor accommodated a catalyst bed up
to 60 cm long. The reactor operates in continuous co-current
gas and liquid. The TBR operates in isothermal mode between
−20 °C and 60 °C by controlling the temperature through a
heating–cooling jacket. The temperature inside the reactor was
measured with a K thermocouple after the initial mixing zone
in the bed. The pressure inside the reactor was controlled
using a back pressure automatic controller installed at the off-
gas outlet (PCV-01). A rupture disc was installed for safety
reasons in the off-gas line. The liquid entrance was inserted
inside the bed, so that the quartz acted as a liquid distributor.
It is extremely important to start up the TBR at low liquid flow
rates ensuring proper wetting of the bed. The gas entered into
a small void volume (ca. 3 mL) acting as pressure compensator
and gas distributor. A manual globe valve was used to sample
the instantaneous liquid phase.
4.2. Methods
H2O2 concentration in H2O was determined using iodometric
titration. The yield was calculated as H2O2 produced in mmol
min−1 divided by H2 entering the reactor in mmol min−1
.
TEM specification: energy-filtered transmission electron
microscopy (EFTEM, LEO 912 Omega, acceleration voltage of
120 kV, LaB6 filament).
The determination of Pd and Br has been done using an
ICP-MS instrument, a PerkinElmer Sciex and an ICP Mass
Spectrometer 6100 DRC Plus. The analysis was conducted
using the quantitative standard mode. The timing parameters
were: Sweeps/Reading: 11, Readings/Replicate: 1, the Number
of Replicates: 7, Dwell time: 50.0 ms, Scan mode: Peak
Hopping. The calibration solutions have been prepared from a
commercial single element solution for both Pd and Br,
diluted into standard serial from 1 ppb to 100 ppb.
Acknowledgements
4.3. Experimental apparatus
Pierdomenico Biasi gratefully acknowledges the Otto A. Malm
The system consists of four sections: liquid inlet, gas inlet, Foundation for financial support. This work is part of the activi-
three-phase reaction and biphasic separation. The liquid was ties at the Åbo Akademi Process Chemistry Centre (PCC) within
pumped at room temperature using an HPLC pump (P-01). the Finnish Centre of Excellence Programmes (2000–2005 and
The gases were fed independently from gas cylinders (D-02 2006–2011) of the Academy of Finland. Juan Garcia Serna
with CO2/H2 97.5/2.5 mol%, or CO2/H2 95/5 mol% and D-03 acknowledges the Spanish Science and Innovation Ministry,
with pure O2) controlled by mass flow controllers (FCV-01 and Project Reference: CTQ2011-23293 and ENE2012-33613 and
FCV-02 respectively). Volumetric flow rates reported in the Junta de Castilla y León, Project Reference: VA254B11-2 for
2512 | Green Chem., 2013, 15, 2502–2513
This journal is © The Royal Society of Chemistry 2013

View Article Online
Green Chemistry
Paper
funding and “Programa Salvador Madariaga 2012” for mobility 20 T. Inoue, K. Ohtaki, S. Murakami and S. Matsumoto, Fuel
scholarship. Stefano Sterchele is gratefully acknowledged for his
help in analyzing TEM pictures.
Process. Technol., 2013, 108, 8–11.
21 T. Inoue, M. A. Schmidt and K. F. Jensen, Ind. Eng. Chem.
Res., 2007, 46, 1153–1160.
22 F. Menegazzo, P. Burti, M. Signoretto, M. Manzoli,
S. Vankova, F. Boccuzzi, F. Pinna and G. Strukul, J. Catal.,
2008, 257, 369–381.
References
23 F. Menegazzo, M. Signoretto, G. Frison, F. Pinna,
G. Strukul, M. Manzoli and F. Boccuzzi, J. Catal., 2012, 290,
143–150.
24 T. Moreno, J. Garcia-Serna and M. J. Cocero, Green Chem.,
2010, 12, 282–289.
1 J. M. Campos-Martin, G. Blanco-Brieva and J. L. G. Fierro,
Angew. Chem., Int. Ed., 2006, 45, 6962–6984.
2 G. Centi, S. Perathoner and S. Abate, in Modern Hetero-
geneous Oxidation Catalysis, ed. N. Mizuno, Wiley-VCH
Weinheim, Germany, 2009, pp. 253–287.
25 T. Moreno, J. Garcia-Serna, P. Plucinski, M. J. Sanchez-
Montero and M. J. Cocero, Appl. Catal., A, 2010, 386, 28–
33.
3 C. Samanta, Appl. Catal., A, 2008, 350, 133–149.
4 M. B. Grisham, Comp. Biochem. Physiol., Part A: Mol. Integr.
Physiol., 2013, 165, 429–438.
26 A. Pashkova, R. Dittmeyer, N. Kaltenborn and H. Richter,
Chem. Eng. J., 2010, 165, 924–933.
5 C. Zhang, F.-R. F. Fan and A. J. Bard, J. Am. Chem. Soc.,
2008, 131, 177–181.
27 T. M. Rueda, J. G. Serna and M. J. C. Alonso, J. Supercrit.
Fluids, 2012, 61, 119–125.
6 K. Otsuka and I. Yamanaka, Electrochim. Acta, 1990, 35,
319–322.
28 S. Abate, P. Lanzafame, S. Perathoner and G. Centi, Catal.
Today, 2011, 169, 167–174.
7 G. F. Thiele and E. Roland, J. Mol. Catal. A: Chem., 1997,
117, 351–356.
29 P. Biasi, N. Gemo, J. R. Hernandez Carucci, K. Eranen,
P. Canu and T. O. Salmi, Ind. Eng. Chem. Res., 2012, 51,
8903–8912.
30 F. Menegazzo, M. Signoretto, M. Manzoli, F. Boccuzzi,
G. Cruciani, F. Pinna and G. Strukul, J. Catal., 2009, 268,
122–130.
8 G. Paparatto, R. D’Aloisio, G. De Alberti and R. Buzzoni, US
Patent 6,630,118, 2003.
9 J. Van Weynbergh and J.-P. Schoebrechts, US Patent
5,447,706, 1995.
10 K. M. Vanden Bussche, S. F. Abdo and A. R. Oroskar, US
Patent 6,713,036, 2004.
31 M. Piccinini, N. E. Ntainjua, J. K. Edwards, A. F. Carley,
J. A. Moulijn and G. J. Hutchings, Phys. Chem. Chem. Phys.,
2012, 12, 2488–2492.
32 S. J. Freakley, M. Piccinini, J. K. Edwards, E. N. Ntainjua,
J. A. Moulijn and G. J. Hutchings, ACS Catal., 2013, 3, 487–
501.
11 B. Zhou and L. K. Lee, US Patent 6,168,775, 2001.
12 S. Abate, K. Barbera, P. Lanzafame, S. Perathoner and
G. Centi, Prepr. – Am. Chem. Soc., Div. Pet. Chem., 2008, 53,
227–229.
13 P. Biasi, P. Canu, F. Menegazzo, F. Pinna and T. O. Salmi,
Ind. Eng. Chem. Res., 2012, 51, 8883–8890.
33 J. Kim, Y.-M. Chung, S.-M. Kang, C.-H. Choi, B.-Y. Kim,
Y.-T. Kwon, T. J. Kim, S.-H. Oh and C.-S. Lee, ACS Catal.,
2012, 2, 1042–1048.
34 T. Moreno, J. Garcia-Serna and M. J. Cocero, J. Supercrit.
Fluids, 2011, 57, 227–235.
35 A. T. Aspen Properties V7.1, Inc., 200 Wheeler Road, Bur-
lington, Massachusetts 01803.
36 P. A. Ramachandran and R. V. Chaudhari, Three-Phase Cat-
alytic Reactors, Gordon and Breach, New York, 1983.
37 T. Deguchi and M. Iwamoto, Ind. Eng. Chem. Res., 2011, 50,
4351–4358.
14 P. Biasi, F. Menegazzo, F. Pinna, K. Eranen, P. Canu and
T. O. Salmi, Ind. Eng. Chem. Res., 2010, 49, 10627–10632.
15 P. Biasi, F. Menegazzo, F. Pinna, K. Eranen, T. O. Salmi and
P. Canu, Chem. Eng. J., 2011, 176–177, 172–177.
16 G. Blanco-Brieva, E. Cano-Serrano, J. M. Campos-Martin
and J. L. G. Fierro, Chem. Commun., 2004, 1184–1185.
17 C. Burato, S. Campestrini, Y.-F. Han, P. Canton, P. Centomo,
P. Canu and B. Corain, Appl. Catal., A, 2009, 358, 224–231.
18 J. K. Edwards, B. Solsona, N. E. Ntainjua, A. F. Carley,
A. A. Herzing, C. J. Kiely and G. J. Hutchings, Science, 2009,
323, 1037–1041.
38 J. Li, T. Ishihara and K. Yoshizawa, J. Phys. Chem. A, 2011,
115, 25359–25367.
19 E. Ghedini, F. Menegazzo, M. Signoretto, M. Manzoli,
F. Pinna and G. Strukul, J. Catal., 2010, 273, 266–273.
This journal is © The Royal Society of Chemistry 2013
Green Chem., 2013, 15, 2502–2513 | 2513