Oxidation of phenol to dihydroxybenzenes by nitrous oxide

Article abstract of DOI:10.1016/j.apcata.2011.11.029

Gas phase oxidation of phenol by nitrous oxide for preparation of dihydroxybenzenes (DHB) is of significant interest. However, due to experimental difficulties caused by the high boiling points of DHB (240.285 °C), no detailed investigation of this reaction has been conducted until now. In the present work, the reaction was studied for the first time using a catalytic setup specially designed for operation with high-boiling compounds. FeZSM-5 zeolites were shown to be efficient catalysts for the title reaction. An unusual isomeric distribution of DHB depending on reaction conditions was found. Formation of resorcinol, in addition to hydroquinone and catechol, is a particular feature of the reaction. Although the fraction of resorcinol averaged over 12 h time-on-stream is not high (6.9 mol.%), in the initial period of reaction it may comprise over 70% of the total amount of DHB. A comparison with the current liquid-phase processes of phenol oxidation by H₂O ₂ shows that the oxidation by N₂O may open a new promising way for alternative production of DHB in the gas phase.

Full text of DOI:10.1016/j.apcata.2011.11.029

Applied Catalysis A: General 415–416 (2012) 10–16
Contents lists available at SciVerse ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Oxidation of phenol to dihydroxybenzenes by nitrous oxide
D.P. Ivanov, L.V. Pirutko, G.I. Panov∗
Boreskov Institute of Catalysis SB RAS, Novosibirsk 630090, Russia
a r t i c l e i n f o
a b s t r a c t
Article history:
Gas phase oxidation of phenol by nitrous oxide for preparation of dihydroxybenzenes (DHB) is of sig-
nificant interest. However, due to experimental difficulties caused by the high boiling points of DHB
(240–285 ◦C), no detailed investigation of this reaction has been conducted until now. In the present
work, the reaction was studied for the first time using a catalytic setup specially designed for operation
with high-boiling compounds.
Received 20 October 2011
Received in revised form
24 November 2011
Accepted 25 November 2011
Available online 4 December 2011
FeZSM-5 zeolites were shown to be efficient catalysts for the title reaction. An unusual isomeric dis-
tribution of DHB depending on reaction conditions was found. Formation of resorcinol, in addition to
hydroquinone and catechol, is a particular feature of the reaction. Although the fraction of resorcinol aver-
aged over 12 h time-on-stream is not high (6–9 mol.%), in the initial period of reaction it may comprise
over 70% of the total amount of DHB.
A comparison with the current liquid-phase processes of phenol oxidation by H2O2 shows that the
oxidation by N2O may open a new promising way for alternative production of DHB in the gas phase.
© 2011 Elsevier B.V. All rights reserved.
Keywords:
Dihydroxybenzenes
Phenol oxidation
Phenol hydroxylation
N2O
FeZSM-5
Hydroquinone
Resorcinol
Catechol
1. Introduction
H2O2 oxidant in the liquid phase [2–7]. However, the progress
seems to be rather modest.
Dihydroxybenzenes (hydroquinone, resorcinol, and catechol)
are important intermediates used in chemical, agrochemical, phar-
maceutical and food industries. Present-day processes of their
production are quite imperfect. Catechol (CH) and hydroquinone
(HQ) are usually produced simultaneously by liquid phase oxi-
dation (hydroxylation) of phenol with hydrogen peroxide. The
strong acids HClO4 and H3PO4 (Rhodia process), Fenton’s reagent
of Fe++/Co++ (Brichima process) or TS-1 zeolite (Enichem process)
are used as catalysts for this reaction [1]. In all cases, intense tar
formation and H2O2 decomposition take place. The total dihydrox-
ybenzenes (DHB) selectivity based on phenol is 80–90%, and that
based on H2O2 is 50–70%.
The synthesis of resorcinol (RS) is usually conducted via an
intermediate preparation of some aromatic m-isomers, viz. 1,3-
benzenedisulfonic acid, which is then subjected to alkali fusion to
form the desired product.
Currently, the oxidation of phenol to DHB is extensively studied
in many laboratories to find more effective ways for implementa-
tion of this difficult reaction. Various metals, metal oxides, organic
and inorganic metal complexes have been tested as catalysts using
The transfer of a reaction into the gas phase can provide some
important technological advantages related, in particular, to con-
tinuous operation of the process, catalyst separation, and catalyst
regeneration. In particular, the latter operation can be conveniently
done in the same reactor by burning out carbonaceous deposits
from the catalyst surface. Therefore, it would be of significant inter-
est to develop a gas-phase process of phenol oxidation. Previous
attempts to carry out this reaction in the gas phase using dioxygen
proved unsuccessful because of the very low selectivity of DHB. One
may expect that the use of nitrous oxide can give a more favorable
prospect for this transformation:
C6H5OH + N2O → C6H4(OH)2 + N2
(1)
Indeed, the related reaction of benzene oxidation to phenol by N2O
over FeZSM-5 zeolites proceeds with a very high selectivity [8–16].
The efficiency of these catalysts is due to the presence of ␣-sites,
which are formed from admixed or specially introduced iron in
the process of high-temperature zeolite activation [10,17–19]. The
␣-sites consist of reduced FeII complexes located in the microp-
ore space of the zeolite matrix. A remarkable feature of ␣-sites is
that they are inert to dioxygen, but are readily oxidized by nitrous
oxide to generate a very reactive anion radical species of ␣-oxygen
[8,11,17,20–22]:
∗
Corresponding author.
E-mail address: panov@catalysis.ru (G.I. Panov).
–
N2O + (FeII)␣ → (FeIII
O
)␣ + N2
(2)
•
0926-860X/$ – see front matter © 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.apcata.2011.11.029

D.P. Ivanov et al. / Applied Catalysis A: General 415–416 (2012) 10–16
11
Table 1
Characteristics of FeZSM-5 samples.
Samples
C␣ (site/g)
Texture parameters
Vꢀ (cm3/g)
V˙ (cm3/g)
ABET (m2/g)
Aext (m2/g)
1. Parent FeZSM-5 (500 ◦C, dry air)
2. Parent FeZSM-5 (steamed at 650 ◦C)
3. FeZSM-5/Al2O3 (500 ◦C, dry air)
4. FeZSM-5/Al2O3 (steamed at 650 ◦C)
5. FeZSM-5/Al2O3 (steamed at 750 ◦C)
0.5 × 1017
6.5 × 1017
0.5 × 1017
2.7 × 1017
1.7 × 1017
0.165
0.160
0.100
0.085
0.080
0.275
0.265
0.390
0.400
0.350
443
430
390
345
335
75
68
155
144
138
A number of studies have shown that ␣-oxygen is involved in
the hydroxylation of benzene. Especially convincing results were
obtained by Uranov et al. [23] and Chernyavsky et al. [24], who
showed that the rate of benzene oxidation increased linearly by
about 2 orders of magnitude with increasing the concentration of
␣-sites.
MFI structure (Table 1). High crystallinity of the zeolite as well as the
absence of foreign phases are evidenced by XRD data collected both
before and after its activation (Fig. SM. 1 in supplementary mate-
rial section).
The parent FeZSM-5 zeolite (sample 1, Table 1) was used for
preparation of formed catalysts. For that, the zeolite was mixed
with boehmite peptized by dilute HNO3 (65 wt.% FeZSM-5 and
35 wt.% of the boehmite). Afterwards, the material was extruded
using a die of 3 mm diameter, dried at 383 K and calcined in dry
air at 773 K (sample 3, Table 1). A calculated composition of the
formed and calcined catalyst corresponds to 70 wt. % of the zeolite
and 30 wt.% of Al2O3 binder. Activation of the catalyst was done
by steaming in a He flow containing 50 mol.% H2O at two different
temperatures: one sample at 650 ◦C and the other one at 750 ◦C
(samples 4 and 5 in Table 1). Further in the text, the activated sam-
ples will be referred to as FeZSM-5 (650 ◦C), FeZSM-5/Al2O3 (650 ◦C)
and FeZSM-5/Al2O3 (750 ◦C), accordingly.
From texture parameters presented in Table 1 one can see
that both the forming procedure and the steam activation of the
formed samples lead to a decrease of Vꢀ. In the case of formed
catalysts, it probably resulted from a partial sealing of microp-
ores by a binder. Since ␣-sites locate precisely in the micropore
space [8,13,30–32], it resulted in a drop of the ␣-site concentration
(C␣): from 6.5 × 1017 site/g for the parent zeolite to 2.7 × 1017 and
1.7 × 1017 site/g for the formed samples (Table 1).
As for the oxidation of phenol, this reaction has not previously
been studied in detail. It was first mentioned in [25] with refer-
ence to patent [26]. The patent is mainly focused on the oxidation
of benzene, suggesting also some examples with other aromatic
compounds, including phenol. Costine et al. [27] tested phenol in
line with some other aromatic molecules to elucidate the effect
of a substituent nature on their reactivity toward N2O. In both
cases, the oxidation of phenol provided DHB products compris-
ing only HQ and CH with a total selectivity of 92–98%. The latter
values are certainly overestimated since the selectivities were cal-
culated without taking coke into account, which, as we will see
below, may have a significant effect on the results. Ivanov et al. [28]
investigated the oxidation of phenol in its mixture with benzene.
The authors showed that benzene significantly improved stabil-
ity and selectivity of the reaction. However, they could use only
minor concentrations of phenol in the feed (4.6%) to prevent plug-
ging of the setup gas lines by condensing DHB products. Therefore,
productivity of the reaction was very low.
The lack of detailed studies on phenol oxidation by N2O is due
to significant experimental difficulties caused by the high boiling
points of DHB. In conventional catalytic setups, the temperature
of the gas lines usually does not exceed 180 ◦C. Since the boiling
points of CH, RS, and HQ are 240, 281 and 285 ◦C, respectively, this
temperature is not high enough to prevent their condensation. It
leads to an increasing resistance to the feed flow with time-on-
stream, and ultimately to the failure of the run. This problem is
discussed in [28].
The value of C␣ was calculated from the N2 amount evolved
at the N2O decomposition at 230 ◦C. At this temperature the reac-
tion proceeds stoichiometrically according to Eq. (2). Additionally,
the isotopic exchange of ␣-oxygen 16O␣ with dioxygen 18O2 was
conducted at 100 ◦C, providing similar results. In more detail,
procedures for measuring the concentration of ␣-sites are given
elsewhere [31,33].
In the present work, the oxidation of phenol by N2O was stud-
ied for the first time using a catalytic setup specially designed for
operation with high-boiling compounds. The temperature of the
analytical part of the setup can be maintained at up to 330 ◦C. The
obtained results revealed high catalytic potential of FeZSM-5 zeo-
lites, which open a new promising way for preparation of DHB.
2.2. Flow setup
Catalytic experiments were performed in an automated flow
setup having a high-temperature switching valves with rotors
made of “polyamide/PTFE/carbon” composite, which operating
temperature is 150–330 ◦C (Valco Instrument Co., Inc.). Analytical
part of the setup was accommodated in a ventilated thermal box
at 290 ◦C. This temperature prevents condensation of the reaction
products and provides a reliable on-line GC analysis. A picture of the
setup and a typical on-line chromatographic pattern are presented
in Figs. SM. 2 and SM. 3 of the supplementary material section.
To conduct catalytic runs, 1 g of a catalyst (0.5–1.0 mm particles)
was placed into a quartz reactor with the inner diameter 7 mm.
Before testing, the catalyst was treated in flowing air at 550 ◦C.
Nitrous oxide and helium were fed by flow mass controllers (MKS
Instruments); phenol of C.P. grade was fed by a high-performance
syringe pump 500 D (ISCO) capable of keeping phenol in a liquid
state at 60 ◦C.
2. Experimental
2.1. Catalysts
A parent FeZSM-5 zeolite was prepared according to [29]. Its
chemical composition in H-form comprises 0.9 wt.% Al, 0.01 wt.%
Na, and 0.03 wt.% Fe (ICP data). This composition was specially
designed to provide a small Fe concentration, which would allow
one to conduct reaction at a rather high temperature thus ensur-
ing an efficient desorption of DHB from the catalyst surface. Under
this condition, an excess concentration of iron and consequently
that of ␣-sites may lead to overoxidation of the products. Texture
parameters of the zeolite (micropore volume, Vꢀ; total pore vol-
The reaction mixture was automatically sampled and analyzed
each 16 min. Analysis of N2O and N2 was performed at room tem-
perature using a packed column filled with Porapak Q and TCD.
For a more accurate COx measurement, carbon oxides were first
ume, V˙; BET surface area, ABET; and external surface area, Aext
)
obtained with the low-temperature N2 adsorption are typical of the

12
D.P. Ivanov et al. / Applied Catalysis A: General 415–416 (2012) 10–16
Table 2
Data on the oxidation of phenol by N2O averaged over 12 h runs.
Run no. Reaction conditions
DHB
Conversion(%)SelectivitiesbasedonPhOH(%) SDHB(N (%) DHB distribution (%)
Coke; Deactivation
O)
2
productivity
(mmol/(g h))
T (◦C) N2O:PhOH
tres (s)
PhOH N2O DHB COx Side
products
Coke
HQ CH RS BQ HQ + BQ mcoke
D (%)
(mol.%)
(g/gcat)
1
2
3
4
5
6
7
475
450
500
475
475
475
475
3:30
3:30
3:30
1.5:30
5:30
3:30
3:30
1.7
1.7
1.7
1.7
1.7
3.4
1.7
1.7
0.85
1.9
0.85
2.3
1.1
4.2
2.3
5.1
2.2
5.9
5.0
4.1
43
22
70
48
40
69
45
83
79
82
80
86
80
82
4.2 2.0
10.8 76
14.4 83
10.0 60
11.8 83
8.4 74
48 37
45 37
52 36
53 33
46 39
60 30
48 38
6
7
7
9
6
8
6
9
11
5
5
9
57
56
57
58
55
62
56
0.24
0.18
0.28
0.14
0.32
0.19
0.22
20
∼5
35
∼5
45
10
20
3.0 3.5
5.5 2.5
4.2 3.6
3.8 1.7
3.7 2.3
4.3 2.1
14.0 65
10.6 72
2
8
1.6
hydrogenated over a nickel catalyst and then analyzed as methane
using FID. Analysis of organic compounds was performed at a pro-
grammed temperature elevation from 180 to 280 ◦C with a capillary
column DB-1701 (J&W Scientific) using FID. In run no. 1 (Table 2),
the reaction products were accumulated and analyzed by GC–MS
method, which results well agree with the results of on-line GC
analysis.
case with FeZSM-5 catalysts. Numerous data on the oxidation of
benzene to phenol by N2O showed that in the initial period the
catalyst has the highest activity, which gradually decreases with
time-on-stream due to coking. In our case, regarding high boil-
ing points of DHB, one may think that the low initial activity is
most likely related to intense adsorption of the reaction products.
Upon completion of the adsorption, the activity achieves its maxi-
mum value and then gradually decreases due to continuing coking
process.
Chromatographic data were used to calculate main parame-
ters of the reaction presented in Table 2: DHB productivity (Pr);
conversions of phenol and N2O (XPhOH, XN O); selectivities of
Fig. 1 clearly reveals different activity of the samples, which
can be inferred from their maximum productivity. The parent
FeZSM-5 zeolite providing 2.7 mmol/(g h) is the most active sam-
ple. The formed samples have lower productivities: 1.9 mmol/(g h)
and 1.2 mmol/(g h) depending on the activation temperature. This
activity sequence correlates with the concentration of ␣-sites,
which decreases in the same order, from 6.5 × 1017 to 2.7 × 1017
and 1.7 × 1017 site/g, respectively.
2
their transformation to DHB, COx and side products; isomeric dis-
tribution of DHB; and catalyst deactivation. Equations used for
calculation of the aforementioned parameters are presented in
supplementary material section.
The carbon material balance was very close to 100%, which is
theoretically expected at low phenol conversions used in the work
(2.2–5.9%). The oxygen material balance defined as the sum of oxy-
gen containing products vs. the N2O conversion was 100 5%. In
the very initial period of reaction the balance may drop down to
85%.
Another important result seen in Fig. 1 is a beneficial effect of
Al2O3 binder on the catalytic stability. This is evidenced by a much
slower deactivation of both formed samples with the time-on-
stream. Based on a favorable combination of activity and stability,
we selected the FeZSM-5/Al2O3 (650 ◦C) catalyst for a more detailed
study. All further data were collected with this sample.
2.3. Regeneration and coke amount
In the reaction course, the catalyst deactivates due to coke for-
mation. After the reaction completion, coke was carefully burned
out to determine its amount (mcoke) and regenerate the catalyst.
For that, the sample was first blown off at 500 ◦C in flowing helium
for 30 min. Carbonic deposits, not removed from the surface under
these conditions, were considered as coke. The burning out was
performed in flowing helium with 10 mol.% air, at a gradual tem-
perature elevation from 450 to 550 ◦C. The value of mcoke expressed
in gram of coke per gram of catalyst (g/gcat) was calculated from the
total amount of COx assuming coke elementary composition to be
identical to that of phenol. Some control TGA measurements pro-
vided well-consistent results on the coke amount. The coke burning
out resulted in complete restoration of the initial catalytic activity.
3.2. Results of run no. 1
Oxidation of phenol over the FeZSM-5/Al2O3 (650 ◦C) sample
was studied at different temperatures, feed mixture composition,
and residence time. Results are collected in Table 2. For a bet-
ter understanding of the parameters under consideration, we will
comment below upon Fig. 2 presenting results of the run no. 1.
3.0
1
2.5
2.0
2
3. Results
1.5
3
3.1. Catalyst selection for detailed study
1.0
0.5
0.0
First, we performed preliminary catalytic experiments for
selecting a suitable sample for the further detailed study. For that,
all three activated samples (nos. 2, 4, 5 of Table 1) were tested
under identical conditions: 475 ◦C, residence time 1.7 s, mole ratio
N2O:PhOH = 3:30, He balance. Results are shown in Fig. 1 as a total
DHB productivity vs. time-on-stream. One can see that in all cases, a
significant increase of productivity takes place in the initial period.
Such phenomenon is known to occur sometimes as a result of cata-
lyst activation (working up) due to a tuning of the surface chemical
composition by reaction mixture [34]. But this is obviously not the
0
2
4
6
8
10
12
Time-on-stream, h
Fig. 1. DHB productivity vs. time-on-stream: (1) FeZSM-5 (650 ◦C); (2) FeZSM-
5/Al2O3 (650 ◦C); (3) FeZSM-5/Al2O3 (750 ◦C). Reaction conditions: 475 ◦C;
N2O:PhOH = 3:30 mol.%; residence time 1.7 s.

D.P. Ivanov et al. / Applied Catalysis A: General 415–416 (2012) 10–16
13
2.5
2.0
1.5
1.0
0.5
0.0
100
a
b
S*DHB (N2O)
80
PrDHB
60
40
20
0
XN2O
0
2
4
6
8
10
12
0
2
4
6
8
10
12
80
60
40
20
0
c 100
d
*
SPhOH
HQ
CH
80
60
*
XPhOH
4
3
2
1
0
BQ
RS
0
2
4
6
8
10
12
0
2
4
6
8
10
12
Time-on-stream, h
Time-on-stream, h
Fig. 2. Variation of reaction parameters with time-on-stream in run no. 1 (for reaction conditions see Table 2): (a) DHB productivity; (b) N2O conversion (XN O) and DHB
2
selectivity based on N2O, SDHB(N O); (c) phenol conversion (XP∗hOH) and DHB selectivity based on phenol, SD∗ HB(PhOH); and (d) isomeric distribution of DHB.
2
Fig. 2a shows how the catalyst productivity varies with time-
on-stream, which shortly has been outlined above. In the initial
period, Pr significantly increases starting from 1.1 mmol/(g h) and
achieving in 2 h 1.9 mmol/(g h). Then, nearly a linear decrease of Pr
is observed, which becomes 1.5 mmol/(g h) at 12 h time-on-stream.
It means that to the end of run the catalyst retains 80% of its maxi-
mum activity. The productivity averaged over the 12 h operation is
1.7 mmol/(g h).
to 42%, CH rises to 40%, while RS goes to a stationary value of ca.
6%. BQ, which may result from oxidation of HQ, passes a minimum
and then linearly increases to 13%.
Upon averaging over 12 h, run no. 1 provides the following
results (Table 2): DHB productivity is 1.7 mmol/(g h); phenol con-
version is 4.2%; and N2O conversion is 43%. DHB selectivity based on
phenol is 83%. The rest 17% of phenol transforms into COx (4.2%);
the sum of side organic products, mainly dibenzofuran and phe-
noxyphenols (2.0%); and coke (10.8%). DHB selectivity based on N2O
is 76%. Fractions of DHB isomers are as follows: 48% HQ, 37% CH, 6%
RS, and 9% BQ. The total fraction of p-isomers (HQ + BQ) comprises
57%. The mass of coke accumulated on the catalyst charge (1 g) is
0.24 g; the catalyst deactivation is 20%.
Fig. 2b shows reaction parameters related to N2O. The DHB
selectivity based on nitrous oxide (SDHB(N O)) grows substantially:
2
from 22% at the beginning of the run to 82% at its end. Meanwhile,
N2O conversion (XN O) shows an opposite trend, monotonically
decreasing from 62%2to 35%.
Fig. 2c illustrates similar parameters related to phenol. In this
case, the symbols of phenol conversion (XP∗hOH) and DHB selectivity
based on phenol (SD∗ HB(PHOH)) are marked by an asterisk to indicate
Run no. 7 was conducted under the same conditions as run
no. 1 to verify the reproducibility of results, which proved to be
satisfactory.
that these parameters are calculated without regard for coke (val-
ues of mcoke were measured upon termination of the runs). In the
initial period of reaction, the value of SD∗ HB(PHOH) rises from 64%
3.3. Effect of reaction conditions
to 94% and then remains virtually constant. The value of XP∗hOH
changes only slightly, having a weak maximum at 2 h. However,
this∗maximum is likely a seeming event resulting from the method
of XPhOH calculation, based on the amount of products recorded. As
noted above, at the initial time the products are intensely consumed
for adsorption and coking, thus giving underrated initial values of
This effect will be discussed using mainly the averaged data of
Table 2. These data were calculated taking into account (where
appropriate) the amount of coke, mcoke, formed on the catalyst
during the run.
XP∗hOH
.
3.3.1. Temperature
Unusual changes are observed with the isomeric distribution
of DHB. Fig. 2d presents fractions of HQ, CH, RS and BQ (p-
benzoquinone; the latter compound we also consider as DHB). One
can see that all the isomers behave differently. Most strong changes
occur at the very beginning of the reaction. Thus, with the reaction
time increasing from 6 min (first point) to 22 min (second point),
HQ fraction increases from 39% to 67%, RS fraction decreases from
47% to 10%, CH fraction increases from zero to 19%, and BQ frac-
tion decreases from 13% to 6%. After that, HQ gradually decreases
From Table 2 one can see that raising the reaction tempera-
ture from 450 to 500 ◦C (runs 1–3) increases the DHB productivity
from 0.85 to 1.9 mmol/(g h). A pronounced increase is observed
also for conversion of both phenol (2.3% → 5.1%) and nitrous oxide
(22 → 70%). Interestingly, DHB selectivity based on phenol shows
only a minor change (79–83%), whereas that based on N2O drops
substantially: from 83% at 450 ◦C to 60% at 500 ◦C. This drop is
accompanied by an increase in selectivity to COx (3.0 → 5.5%),
which formation may proceed mainly due to oxidation of coke.

14
D.P. Ivanov et al. / Applied Catalysis A: General 415–416 (2012) 10–16
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0.0
80
a
b
1.5% NO
2
5% N2O
HQ
60
40
20
3% N2O
CH
RS
1.5% NO
2
BQ
0
0
2
4
6
8
10
12
0
2
4
6
8
10
12
80
60
40
20
0
80
c
d
5% N O
3% N2O
2
60
40
20
0
HQ
CH
HQ
CH
BQ
RS
BQ
RS
0
2
4
6
8
10
12
0
2
4
6
8
10
12
Time-on-stream, h
Time-on-stream, h
Fig. 3. Effect of N2O concentration on DHB productivity (a) and DHB distribution (b, c, d). For reaction conditions see Table 2, runs 1, 4, 5.
Indeed, as selectivity to COx increases, the selectivity to coke
decreases (14.5% → 10%).
3.3.3. Residence time
Runs no. 1 and no. 6 (Table 2) were conducted under sim-
ilar conditions, except for the residence time (tres), which was
1.7–3.4 s, respectively. An increase of tres raises the conversion
of both phenol and nitrous oxide, but decreases the productiv-
ity (1.7 → 1.1 mmol/(g h)). Other essential changes are as follows:
a decrease of N2O selectivity to DHB (76 → 65%); an increase of
phenol selectivity to coke (10.8 → 14.0%); and an increase of HQ
fraction (48 → 60%). Changes of the total p-isomer fraction are less
significant (57 → 62%).
The DHB distribution proved to be low sensitive to the tem-
perature. Fractions of CH (36–37%) and RS (6–7%) remain virtually
constant. An increase in HQ fraction (45 → 52%) with temperature
elevation is compensated by a decreasing fraction of BQ (11 → 5%).
So, the sum of p-isomers also remains nearly constant (56–57%).
Temperature elevation strongly increases the catalyst deactiva-
tion. Thus, at 450 ◦C deactivation is weak, ∼5%, whereas at 500 ◦C
it is 35%.
3.3.2. Feed composition
4. Discussion
Fig. 3a illustrates changes in the catalyst productivity with the
time-on-stream at different concentrations of nitrous oxide (CN
)
O
4.1. Catalyst deactivation
2
in the feed mixture: 1.5% N2O, 3.0% N2O and 5.0% N2O. In all cases,
phenol concentration is 30 mol.%, and temperature is 475 ◦C. It is
In the liquid phase oxidation by hydrogen peroxide, 10–20%
of phenol is consumed for tar formation [1]. In our case, coke is
formed instead of tar, being responsible for 8.4–14.4% of the phenol
consumed (Table 2). There is a correlation between coke amount
and catalyst deactivation. Indeed, run no. 5 with mcoke = 0.32 g/gcat
shows the highest deactivation (45%), whereas runs no. 2 and no. 4
with mcoke of 0.18 and 0.14 g/gcat show the least one (∼5%). How-
ever, amount of coke does not correlate with the phenol selectivity
to coke, since mcoke depends also on the reaction rate. With increase
of the rate, the amount of coke increases, but not so strongly as the
DHB productivity. Therefore, run no. 5 with the largest value of
mcoke has the lowest phenol selectivity to coke, 8.4%.
It is generally accepted that coke formation on zeolite surfaces
is mediated by both the Bronsted and the Lewis acid sites. But
the mechanism of its deactivating effect is less clear. It may relate
to both the diffusion limitations and the poisoning of active sites.
The mechanism of FeZSM-5 deactivation in the oxidation of ben-
zene to phenol by N2O was thoroughly studied in [30,35]. A loss
in activity was shown to result from coke deposition in the zeolite
seen that the growth of CN O increases both productivity and deac-
2
tivation of the catalyst. Indeed, Table 2 (runs 1, 4, 5) indicates that
the averaged productivity at 1.5% N2O is 0.85 mmol/(g h), whereas
at 5% N2O it is 2.3 mmol/(g h). Simultaneously, the deactivation
rises from ∼5% to 45%. The increase in CN is accompanied by a
O
considerable growth of phenol conversion 2(2.2% → 5.9%) and some
decrease of N2O conversion (48 → 40%). At the same time, phe-
nol selectivity to DHB rises (80 → 86%), while N2O selectivity falls
(83 → 74%).
Fig. 3b–d shows changes in the isomeric distribution of DHB.
In all cases, HQ and CH are predominant products, their fractions
differently depending on both the N2O concentration and especially
the time-on-stream. Thus, with CN = 1.5%, the ratio HQ:CH at
O
2
1 h of time-on-stream is 3.3, whereas at 12 h it is 1.2. Variations
in the averaged distribution (Table 2) are not so significant. The
most noticeable change is for HQ, which fraction decreases from
53% to 46% as CN increases. However, the total p-isomer fraction
O
(HQ + BQ) varies 2only slightly, 55–58%.

D.P. Ivanov et al. / Applied Catalysis A: General 415–416 (2012) 10–16
15
micropores, which is exactly the place of ␣-sites location [8,13,32].
100
80
60
40
20
0
Coke decreases the number of operating sites, but the turnover
frequency calculated per a single operating site remains constant
irrespective of the coke amount and deactivation degree. This result
clearly testifies that deactivation is caused by poisoning of active
sites rather than by difficult diffusion of the reactants to and from
the sites.
Most likely, this deactivation mechanism holds also for the oxi-
dation of phenol to DHB. But taking into account a much more
intense coking in the latter reaction, one can assume that deactiva-
tion in this case may be additionally aggravated also by worsening
of the reactants diffusion or partial blocking of micropores due to
coking of the external surface of zeolite. In particular, this assump-
tion may explain an improved stability of the formed samples
compared to that of the parent zeolite. Due to Al2O3 binder, the
formed samples have a much greater external surface area (Table 1),
which may be a suitable depot for accumulation of external coke
not decreasing the catalytic activity.
RS
HQ
BQ
CH
0
1
2
3
4
5
N2O concentration, mol%
Fig. 4. Effect of N2O concentration in the feed on the initial DHB distribution (Table 2,
runs 1, 4, 5).
One can hardly think that this picture gives a real reaction pat-
tern on a “clean” (non-coked) catalyst surface. Quite probably, this
behavior is a result of an interplay of several factors like a different
involvement of isomeric molecules in the adsorption, coke forma-
tion, and other processes, which proceed most intensely in the
initial period. Identification of parameters affecting the distribution
of DHB would be of significant interest.
4.2. Isomeric distribution of DHB
As we could see above, the oxidation of phenol by N2O leads
to all three DHB isomers, which is not typical of the oxidation by
H2O2. In the aromatic ring, the OH group is known to activate the o-
and p-positions, but deactivate the m-position. This phenomenon
explains why only CH and HQ form at phenol oxidation by H2O2
[2–7]. In this case, the OH radicals (e.g. in Fenton system) or the
peroxide groups bound to isolated metal atoms (e.g. on TS-1) are
supposed to be active oxidants [36], which reactivity is not enough
to attack the m-position. In the case of N2O, the active oxidant is
␣-oxygen, which is a radical species O•− bound to iron. Because of a
very high reactivity of ␣-oxygen [8,37,38], the difference between
activated and non-activated positions of phenol may become less
important, leading to hydroxylation of m-position, too.
As noted above, the isomeric distribution of DHB strongly varies
with the time-on-stream. According to a generally accepted idea,
this variation can be explained by a transport limitation in the cat-
alyst pore system imposed by coke. This limitation may change the
composition of reaction products in favor of the molecules having
smaller cross-sections and therefore greater diffusivities. However,
this explanation, being verified in many cases, has a difficulty with
our results. According to the above reasoning, one may expect that
with increasing reaction time the fractions of CH and RS should
decrease, while the fraction of HQ (its molecules have the small-
est cross-section) should increase. But the actual changes are quite
different. In Fig. 3b and c, neglecting the very initial time, one can
see that HQ fraction monotonically decreases, whereas CH frac-
tion increases and RS fraction remains nearly constant. So, some
other ideas are needed for explaining this phenomenon. In partic-
ular, one should possibly take into consideration stability of DHB
products under the reaction conditions, as well as their possible iso-
meric redistribution, which may take place on the external surface
covered by coke. This important point is worth a further study.
Note that in the case of phenol oxidation by hydrogen peroxide
no significant variation of DHB distribution with the reaction time
was observed [3].
4.3. Comparison with the present-day processes
The formed FeZSM-5/Al2O3 (650 ◦C) catalyst studied in this
work is certainly not optimal. Its performance can be improved by
tuning the chemical composition, preparation procedure, activa-
tion conditions, etc. Nevertheless, it seems interesting to compare
the obtained results with the present-day processes of phenol
oxidation to DHB. It may allow one to estimate whether these lab-
oratory data can serve as a potential bases for developing a new
DHB process, or they are too inferior for having such a prospect.
The comparison can be made using the data of Table 3, composed
by Notari [1] for the liquid-phase phenol oxidations with H2O2. Cor-
responding parameters for the N2O oxidation are added from our
Table 2, run no. 5.
One can see that in the case of N2O, phenol conversion (5.9%)
is on a level with Rhone Poulenc process (5%), but ranks below
Brichima (10%) and particularly Enichem (25%) processes.
Phenol selectivity to DHB in the case of N2O (86%) takes an
intermediate position: it is somewhat lower than the selectivity
of Rhone Poulenc and Enichem processes (both 90%) and higher as
compared to Brichima (80%).
With both oxidants, a considerable fraction of phenol transforms
into coke or tar. With N2O, the proper phenol selectivity to coke is
8.4%. However, to provide a fair comparison with the selectivity
to tar in the case of H2O2, this value should be added with the
selectivities to COx and side products, which are virtually absent
in the case of liquid-phase processes. Being summed up, they give
14%, which is somewhat higher than phenol selectivity to tar for
Enichem (12%), but lower than that for Brichima (20%).
Nitrous oxide selectivity to DHB (74%) exceeds the selectivities
of all H2O2 based processes (50–70%).
It is interesting to consider the DHB distribution at a minimum
amount of coke on the surface, i.e. at the very initial reaction time
(6 min). The initial isomeric fractions obtained at 475 ◦C with dif-
ferent N2O concentrations (first points in Fig. 3b–d) are shown in
Fig. 4. As CN rises from 1.5% to 5%, the fraction of RS strongly
Since HQ is a more valuable product than CH, the ratio HQ:CH is
considered to be an important parameter [1]. In the case of nitrous
oxide, the HQ:CH ratio is equal to 1.2 (or 1.4 if BQ is taken into
account). With this parameter, N2O surpasses the H2O2 based pro-
cesses, including the process on TS-1, where this ratio has the
highest value equal to 1.0.
Summarizing the above comparison, we should admit that
a formal ranging of processes are quite questionable. Efficiency
of a process is determined by several parameters, sometimes
O
2
decreases (73 → 12%) and the fraction of HQ increases (18 → 72%),
nearly compensating each other. Amounts of CH and BQ are rather
insignificant and do not affect much a general picture. Remarkably,
when dependences in Fig. 4 are extrapolated to CN = 0, the frac-
O
tion of RS tends to 100%, whereas all other fractio2ns tend to zero.

16
D.P. Ivanov et al. / Applied Catalysis A: General 415–416 (2012) 10–16
Table 3
Comparison of DHB preparation via phenol oxidation by N2O and H2O2.
Process parameters
Oxidation by N2O (run no. 5, Table 2)
Oxidation by H2O2 [1]
Rhone Poulenc (HClO4, H3PO4)
Brichima (Fe++/Co++
)
Enichem (TS-1)
Phenol conversion, %
5.9
86
8.4 (14.0a)
74
1.2 (1.4b)
5
90
10
70
10
80
20
50
25
90
12
70
1.0
Phenol selectivity to DHB, %
Phenol selectivity to coke or tars, %
N2O (H2O2) selectivity to DHB, %
HQ:CH ratio
0.71
0.43
a
COx and side products are included in coke.
The ratio calculated for the sum of (HQ + BQ).
b
differently directed. Nevertheless, one can see that parameters pro-
vided by N2O are comparable with those provided by H2O2. It may
open a new way for alternative preparation of DHB. In this connec-
tion, it is appropriate to note that nitrous oxide is not considered
anymore as an exotic “laughing gas” as it was the case two decades
ago. Nowadays it is a rather conventional compound whose unique
oxidation chemistry is well studied, and appreciated not only in lab-
oratory research but also in industry [8,31,39]. Recently, two new
commercial oxidation processes have been put in operation first
using the N2O oxidant [40].
[2] J. Sun, X. Meng, Y. Shi, R. Wang, S. Feng, D. Jiang, R. Xu, F.-S. Xiao, J. Catal. 193
(2000) 199–206.
[3] S. Seelan, A.K. Sinha, Appl. Catal. A: Gen. 238 (2003) 201–209.
[4] K. Yube, M. Faruta, N. Aoki, K. Mal, Appl. Catal. A: Gen. 327 (2007) 278–286.
[5] K.M. Parida, S. Singha, P.C. Sahoo, Catal. Lett. 136 (2010) 155–163.
[6] H.S. Abbo, S.J.J. Titinchi, Top. Catal. 53 (2010) 254–264.
[7] E.A. Karakhanov, A.L. Maximov, Y.S. Kardasheva, V.A. Skorkin, S.V. Kardashev,
V.V. Predeina, M.Y. Talanova, E. Lurie-Luke, J.A. Seeley, S.L. Cron, Appl. Catal. A:
Gen. 385 (2010) 62–72.
[8] G.I. Panov, CATTECH 4 (2000) 18.
[9] J. Jia, K.S. Pillay, W.M.H. Sachtler, J. Catal. 221 (2004) 119–126.
[10] I. Yuranov, D.A. Bulushev, A. Renken, L. Kiwi-Minsker, Appl. Catal. A: Gen. 319
(2007) 128–136.
[11] L.V. Pirutko, S.V. Chernyavsky, A.K. Uriarte, G.I. Panov, Appl. Catal. A: Gen. 227
(2002) 143–157.
[12] E.J.M. Hensen, Q. Zhu, M.M.R.M. Hendrich, A.R. Overweg, P.J. Kooyman, M.V.
Sychev, R.A. van Santen, J. Catal. 221 (2004) 560–574.
[13] A. Ribera, I.W.C.E. Arends, S. de Vries, J. Perez-Ramires, R.A. Sheldon, J. Catal.
195 (2000) 287–297.
[14] G. Centi, S. Perathoner, R. Arrigo, G. Giordano, A. Katovic, V. Pedula, Appl. Catal.
A: Gen. 307 (2006) 30–41.
[15] A.J.J. Koekkoek, H. Xin, Q. Yang, C. Li, E.J.M. Hensen, Micropor. Mesopor. Mater.
145 (2011) 172–181.
[16] D. Meloni, R. Monac, V. Solianas, G. Berlier, S. Bordiga, I. Rossetti, C. Oliva, L.
Forni, J. Catal. 214 (2003) 169–178.
[17] K.A. Dubkov, N.S. Ovanesyan, A.A. Shteinman, E.V. Starokon, G.I. Panov, J. Catal.
207 (2002) 341–352.
[18] L. Kiwi-Minsker, D.A. Bulushev, A. Renken, J. Catal. 219 (2003) 273–285.
[19] A.R. Overweg, M.W.J. Craje, A.M. van der Kraan, J.W.C.E. Arends, A. Ribera, R.A.
Sheldon, J. Catal. 223 (2004) 262–270.
5. Conclusion
A formed FeZSM-5 zeolite, having small concentration of iron
and accordingly that of ␣-sites, was studied in detail at 450–500 ◦C
in the oxidation of phenol to DHB by nitrous oxide. For the first
time the reaction was carried out in an automated catalytic setup
specially designed for the gas-phase operation with high-boiling
compounds. It allowed safe catalytic runs with reliable on-line
GC analysis. The results obtained were completed with the data
on coke amount. Deposition of coke on the catalyst surface was
shown to cause a pronounced variation in the isomeric distribution
of DHB, which does not correlate with diffusivities of the product
molecules.
[20] G.D. Pirngruber, J.-D. Grunwaldt, P.K. Roy, J.A. van Bokhoven, O. Safonova, P.
Glatzel, Catal. Today 126 (2007) 127–134.
[21] E. Berrier, O. Ovsitser, E.V. Kondratenko, M. Schwidder, W. Grünert, A. Brückner,
J. Catal. 249 (2007) 67–78.
[22] J.B. Taboada, E.J.M. Hensen, I.W.C.E. Arends, G. Mul, A.R. Overweg, Catal. Today
110 (2005) 221–227.
[23] I. Yuranov, D.A. Bulushev, A. Renken, L. Kiwi-Minsker, J. Catal. 227 (2004)
138–147.
A comparison with the present-day processes based on H2O2
application shows that phenol oxidation by N2O may be considered
as a promising basis for developing an alternative technology of
DHB preparation.
[24] V.S. Chernyavsky, L.V. Pirutko, A.K. Uriarte, A.S. Kharitonov, G.I. Panov, J. Catal.
245 (2007) 466–469.
Supplementary material
[25] G.I. Panov, A.S. Kharitonov, V.I. Sobolev, Appl. Catal. A: Gen. 98 (1993) 1–20.
[26] US Patent 5110995 (1992).
The manuscript has
a supplementary material section. It
[27] A. Costine, T. O’Sillivan, B.K. Hodnett, Catal. Today 99 (2005) 199–208.
[28] D.P. Ivanov, V.I. Sobolev, L.V. Pirutko, G.I. Panov, Adv. Synth. Catal. 344 (2002)
986–995.
[29] E.G. Derouane, S. Determmerie, Z. Gabelica, N. Blom, Appl. Catal. 1 (1981)
201–224.
[30] D.P. Ivanov, M.A. Rodkin, K.A. Dubkov, A.S. Kharitonov, G.I. Panov, Kinet. Catal.
41 (2000) 771–775.
[31] G.I. Panov, A.K. Uriarte, M.A. Rodkin, V.I. Sobolev, Catal. Today 41 (1998)
365–385.
includes: XRD patterns of parent FeZSM-5 zeolite before and after
activation; a picture of catalytic setup used in the work; a typi-
cal pattern of on-line GC analysis; and a list of equations used for
calculating main reaction parameters.
Acknowledgements
[32] L.V. Pirutko, O.O. Parenago, E.V. Lunina, A.S. Kharitonov, L.G. Okkel, G.I. Panov,
React. Kinet. Catal. Lett. 52 (1994) 275–283.
[33] M.V. Parfenov, E.V. Starokon, S.V. Semikolenov, G.I. Panov, J. Catal. 263 (2009)
173–180.
[34] G.K. Boreskov, Heterogeneous Catalysis, Nova Sci. Publ. Inc., NY, 2002, 236 p.
[35] D.P. Ivanov, V.I. Sobolev, G.I. Panov, Appl. Catal. A: Gen. 241 (2003) 113–121.
[36] C. Perego, A. Carati, P. Ingallina, M.A. Mantegazza, G. Belussi, Appl. Catal. A: Gen.
221 (2001) 63–72.
This work was supported by RFBR (Grant No. 11-03-00427-a).
The authors are grateful to E.V. Starokon for the measurement of ␣-
sites concentration, and to A.S. Kharitonov for valuable comments.
Appendix A. Supplementary data
[37] K.A. Dubkov, V.I. Sobolev, E.A. Talsi, M.A. Rodkin, N.H. Watkins, A.A. Shteinman,
G.I. Panov, J. Mol. Catal. 123 (1997) 155–161.
[38] G.I. Panov, K.A. Dubkov, E.V. Starokon, Catal. Today 117 (2006) 148–155.
[39] V.N. Parmon, G.I. Panov, A. Uriarte, A.S. Noskov, Catal. Today 100 (2005)
115–131.
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.apcata.2011.11.029.
References
[40] BASF News Release, http://www.chemicalonline.com/article.mvc/BASF-Starts-
Up-A-New-Production-Facility-For-0001, November 30, 2009.
[1] B. Notari, Stud. Surf. Sci. Catal. 37 (1988) 413–425.