- Email: [email protected]
Gas phase hydroxylation of benzene with air–ammonia mixture over copper-based phosphate catalysts
Accepted Manuscript Title: Gas phase hydroxylation of benzene with air-ammonia mixture over copper-based phosphate catalysts ˇ ˇ Author: Blaˇzej Horv´ath Martin Sustek Jaroslava Skriniarov´ a M´aria Omastov´a Edmund Dobroˇcka Milan Hronec PII: DOI: Reference:
S0926-860X(14)00320-2 http://dx.doi.org/doi:10.1016/j.apcata.2014.05.006 APCATA 14826
To appear in:
Applied Catalysis A: General
Received date: Revised date: Accepted date:
10-2-2014 16-4-2014 6-5-2014
ˇ ˇ Please cite this article as: B. Horv´ath, M. Sustek, J. Skriniarov´ a, M. Omastov´a, E. Dobroˇcka, M. Hronec, Gas phase hydroxylation of benzene with air-ammonia mixture over copper-based phosphate catalysts, Applied Catalysis A, General (2014), http://dx.doi.org/10.1016/j.apcata.2014.05.006 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Gas phase hydroxylation of benzene with air-ammonia mixture over copper-based phosphate catalysts Blažej Horvátha,*, Martin Šusteka, Jaroslava Škriniarováb, Mária Omastovác, Edmund Dobročkad
ip t
and Milan Hroneca,* a
Department of Organic Technology, bInstitute of Electronics and Fotonics, Slovak University of
cr
Technology, Radlinského 9, 81237Bratislava, Slovakia c
Polymer Institute, dInstitute of Electrical Engineering, Slovak Academy of Sciences,
us
Dúbravská cesta 9, 845 41 Bratislava, Slovakia
an
Abstract
The direct gas phase hydroxylation of benzene to phenol over copper-based phosphate
M
catalysts has been studied. Nitrous oxide and air-ammonia mixtures were used as oxidizing agents. It has been found that nitrous oxide can be replaced by air-ammonia mixture without
d
significant deterioration of the phenol yield. Nanostructured Ca-Cu phosphate catalysts with
te
a dominant mixed pyrophosphate structure, prepared in the presence of suitable surfactants were found highly active and selective. Using surfactant-modified phosphate catalysts with sponge-like
Ac ce p
structure above 95% selectivity and about 6.5% and 4.2% phenol yields were achieved at the benzene WHSV 1 h-1 and under not optimized reaction conditions using N2O and air-ammonia mixture as oxidizing agents, respectively. These catalysts, compared with unmodified ones, were found to be much more resistant towards deactivation caused by the chemical reduction of catalytically active copper phase during the catalytic test. The activity of the spongy Ca-Cu phosphate catalyst after 37 h on stream decreased only by about 25%. In contrast, over similar unmodified Ca-Cu phosphate catalyst the phenol yield already after 8h on stream drops to about one third of the initial value. Keywords: benzene hydroxylation, copper phosphates, nitrous oxide, ammonia, catalyst lifetime *corresponding authors, e-mail address: [email protected] (B.Horváth) [email protected] (M.Hronec)
Page 1 of 24
1. Introduction Benzene as a raw material is widely used in many industrial processes, including phenol production. More than 90% of the world’s phenol production is based on the cumene route that
ip t
implies alkylation of benzene, oxidation of cumene by a free-radical mechanism and cleavage of the formed hydroperoxide catalyzed by a mineral acid [1]. Despite its great success, the cumene process has some disadvantages such as high environmental impact, the use of corrosive catalyst,
cr
the production of an explosive intermediate and a co-production of acetone. Therefore, the success of this process significantly depends on the marketability of the by-product acetone and
us
on the ratio between the demands for phenol vs. acetone [2]. Other processes of industrial manufacture of phenol, as the Dow process [3, 4] involving toluene oxidation to benzoic acid and
an
its decarboxylation using a copper salt catalyst to phenol, represent only a minor contribution to the commercial phenol production.
M
Several studies have been published during the past two decades on the synthesis of phenol via one-step oxidation of benzene. Direct hydroxylation of the aromatic ring is a challenging reaction to synthesize phenol, due to the formation of over-oxygenated by-products.
d
The synthesis of phenol by direct hydroxylation of benzene was extensively studied using
te
different oxidants, such as N2O, O2, H2O2 as well as the H2-O2 mixture [5-8]. Using hydrogen peroxide to oxidize benzene, the self-decomposition and the price of
Ac ce p
hydrogen peroxide is the most common limiting factor. However, using vanadium oxide supported on mesoporous SBA-16 [9] or metal doped heteropolyacids [10], hydrogen peroxide – based systems reach high selectivities. An attractive catalytic system for the hydroxylation of benzene should use oxygen, preferably air as a cheap and available oxidant. Several catalytic systems, active in the hydroxylation of benzene with O2 [11], often use a co-reducing agent apart from the oxygen [12]. It is well known, that copper-based phosphate catalysts, depending on its nanostructure can alter their catalytic properties [13]. For the direct hydroxylation of benzene the role of the oxidation state of copper and finding of the phase responsible for the catalytic activity [14-16] are other important keys for understanding the mechanism of the catalytic hydroxylation of benzene.
One of the most widely studied catalytic systems for the gas phase hydroxylation of aromatics with N2O uses Fe-zeolite catalysts [17, 18]. Fe/ZSM-5 is reported to be highly active
Page 2 of 24
and stable in DeNOx systems designed to decompose N2O [19]. Over similar catalysts phenol is formed with a selectivity exceeding 90% [20]. One of the major problems of these systems was their rapid deactivation (either published or observed in practice). Similarly, a very rapid deactivation was reported also over AlPO catalysts [21]. Another drawback of catalytic systems
ip t
using N2O as oxidant is their dependence on the source of waste nitrous oxide, usually obtained from adipic acid plants [22].
cr
In the present work we focused attention on the preparation copper-based phosphate catalysts for gas-phase oxidation of benzene, using nitrous oxide and an air-ammonia mixture as
an
influence on catalytic properties are mainly discussed.
us
oxidizing agents. The oxidation state and the structure of the copper-based domains and their
M
2. Experimental
d
2.1 Catalysts preparation and characterization
As a basis of comparison, copper-calcium-phosphate with 43 % of copper (relative to the
te
overall amount of cations) was prepared by precipitation of the calcium and copper nitrate solutions (1.00 M) with a 0.60 M ammonium monohydrate orthophosphate solution at ambient
Ac ce p
temperature. The hydrogen phosphate solution was added drop wise to the metal nitrate solution while stirring vigorously. The precipitate was stirred for 3h, then filtered and excessively washed with deionized water to extract the relics of the precursors, until pH 7. The catalysts with 5% copper content were prepared using a similar procedure. Iron-calcium-phosphate was prepared accordingly, replacing copper nitrate with an equimolar amount of ferric nitrate monohydrate. Mixed Cu-Fe-Ca-phosphate catalysts were prepared as well, by replacing 50 mol% of copper nitrate by ferric nitrate. A Pd-Cu-Ca-phosphate catalyst was prepared by introducing palladium by wet impregnation from a 1.00M Pd(acac)2 solution in toluene onto a pre-prepared Cu-Ca-phosphate sample (43% Cu). The overall palladium content of the resulting material, calculated as metallic palladium, was 1.5 wt%. To prepare phosphate catalysts with sponge-like structure, various structure-directing agents were added to the solution of metal nitrates in an amount of 10 wt% relative to the
Page 3 of 24
precipitate (calculated as metal pyrophosphates) prior to the precipitation by ammonium hydrogen phosphate. Stearyl amine polyethylene glycol ether with 7 ethoxy units (SAPEG7), Pluronic P-123, dodecylbenzene sulphonate-Na (DBS), as well as active carbon FW285 (Evonik) were tested as structure-directing agents.
ip t
Silica-modified catalysts were prepared by two methods: (i) by precipitating the metal ion-exchanged silicas with an ammonium hydrogen phosphate solution (denoted as Cu-IE), and
cr
(ii) by adding pre-hydrolyzed tetraethyl orthosilicate (TEOS) to the metal nitrate solutions in a ratio 1:1. The former set was prepared by stirring overnight a 1 wt% suspension of silica Aerosil®
us
200 (Degussa, declared BET-surface area 200 m2.g-1) with an equimolar amount of copper nitrate in aqueous solution. The silica was then filtered, washed with deionized water until neutrality to sodium sulphide and re-suspended in 10% aqueous solution of ammonium monohydrate
an
orthophosphate. The suspension was stirred overnight, filtered, washed excessively, dried at 180°C and calcined. The latter set was prepared by pre-hydrolyzing 5 wt% of TEOS in a 0.05 N
M
aqueous solution of hydrochloric acid, at the temperature of 50°C for 3h, while stirring vigorously. The resulting solution was used to dissolve the metal nitrates, and was precipitated by
d
ammonium hydrogen phosphate as described above. The precipitate was stirred at 80°C for 24h, filtered, washed until neutrality, and dried. All the catalysts were calcined prior to catalytic tests
Ac ce p
summarized in Table 1.
te
at 600°C in air flow during 5 h, with a heating rate of 100 K.h-1. The prepared catalysts are
Table 1: Summary of the prepared mixed Ca-Cu phosphate catalysts. Metal contents are given as per cents relative to overall cation content. denotation
Cu content / %
modifier
parent
43
no
5
no
Cu-Fe-p
21.5
21.5% Fe
Pd-Cu-p
43
2% Pd
SAPEG7-p
43
SAPEG7
P123-p
43
Pluronic P123
DBS-p
43
DBS
FW285-p
43
Active carbon FW285
5Cu-p
Page 4 of 24
43
TEOS
Cu-IE
<1
Silica ion-exchanged by Cu
cr
ip t
TEOS-p
The catalysts were characterized by powder X-ray diffraction using a Stadi Stoe apparatus
us
in Bragg-Brentano geometry and a Bruker D8 DISCOVER diffractometer in a parallel beam geometry with parabolic Goebel mirror in the primary beam; the diffraction patterns were recorded in grazing incidence set-up with the angle of incidence α = 6°. Temperature
an
programmed reduction and BET-measurements were carried out using a Micromeritics Pulse Chemisorb 2700 apparatus. SEM measurements were carried out using a JEOL JSM-7500F
M
microscope; XPS were measured using a Thermo Scientific K-Alpha XPS system with an Al Kα
Ac ce p
2.2 Catalytic tests
te
d
X-ray source at 6 mA, 12 kV.
The experiments were carried out in a fixed-bed glass reactor at the temperature of 450°C, using 1 g of catalyst. The catalysts prior to testing were pelletized and crushed to the grain size of 0.2 – 0.3 mm. The temperature was controlled by a thermocouple in the axis of the catalytic bed. The gaseous oxidizing agent (air or N2O) with the flow rate of 70 ml.min-1, was saturated by benzene (1.0g.h-1) before entering the reactor. In experiments where air-ammonia mixture was used as an oxidizing agent, aqueous ammonia solution (with the ammonia concentration of 25%) was added by a syringe pump (2.0g.h-1) to the reactant stream. In some experiments, only water was injected instead of ammonia solution. The reaction products were collected in an ice-cooled flask, homogenized with methanol and analyzed by GC, equipped with FID for the detection of organic products, and TCD for the analysis of CO2 in the gaseous products. To verify the composition of reaction products a Shimadzu QP-500 GC-MS was used.
Page 5 of 24
ip t
3. Results
cr
3.1 Catalysts characterization
us
XRD – characterization
The parent Ca-Cu phosphate represents a mixed phase system (Fig. 1). As being a coprecipitate, one can expect various separated phosphates of copper and calcium, and bimetallic
an
phosphates. The calcination temperature of the precipitate, 600°C, presages the dominance of pyrophosphates over orthophosphates or hydrogen phosphates in the resulting material [23-25].
M
Hydroxyapatites, if formed, can retain their structure above the calcination temperature employed. However, the pH of the reaction mixture and the Ca/P ratio is unfavorable for the formation of hydroxyapatites [24,26]. Hence, as expected, no hydroxyapatite or other separated
d
calcium orthophosphate or hydrogen phosphate was found in detectable concentrations in the Ca-
te
Cu phosphate materials. Similarly, the presence of copper hydrogen phosphate was not detected. Copper, if formed as a separated phase, is known to transform to copper pyrophosphate at the
Ac ce p
temperature of 600°C [27,28]. Copper and calcium phosphate precipitates can undergo solid-state reactions to form binary phosphates at 600°C [2, 29]. Accordingly, in the parent Ca-Cu phosphate sample (with 43% Cu), the copper and calcium pyrophosphates were found as dominant phases in the diffraction region bellow 40°. However, the shift of the crystallographic parameters (about 1°) indicates that a partial mutual substitution of Cu and Ca occured in these phases, leading to a deformation of the crystalline structure. This cation exchange takes place most probably via solid-phase reactions during the calcination. The occurence of minor binary phosphates, as Ca19Cu2(PO4)14, with a deformed whitlockite (Ca3(PO4)2) structure cannot be excluded [30]. Whitlockit, being one of the most abundant phosphate structures, can form a plethora of bimetallic phases with copper [31]. In the diffraction region 40 – 45° the presence of extra-framework CuO and Cu2O is observable. The
Page 6 of 24
occurence of paracrystalline or crystalline extra-framework copper oxides in calcined phosphates was repeatedly reported [32,33]. The phase composition of the Cu-Ca phosphate catalyst is obviously changing during the catalytic tests, using nitrous oxide or air-ammonia mixture as oxidizing agents. The initial
ip t
deformed Ca-Cu pyrophosphate structure, with minor amounts of CuO and whitlockit-type binary orthophosphates, exhibited after the catalytic tests several alternations. These changes
cr
indicate that during the catalytic tests several changes took place: (i) the partial reduction of phosphorus and an emergence of a phosphite phase, (ii) the reduction of extra-framework CuO to
Ac ce p
te
d
M
an
us
Cu2O and (iii) the appearance of metallic copper.
Fig. 1 XRD-patterns of the fresh parent Cu-Ca phosphate catalyst, and after benzene oxidation at 450°C using N2O and an air-ammonia mixture as oxidizing agents, respectively. * Ca19Cu2(PO4)14, ▼Ca2P2O7, ▲Cu2P2O7, ♦Cu(PO3)2, ◊Cu, ●CuO, ○Cu2O
The Ca-Cu phosphate prepared on a silica-based matrix, modified by the pre-hydrolyzed TEOS, has shown a similar initial phase composition, as well as similar changes upon the catalytic tests. The rate of its chemical reduction, however, was milder when using N2O as an oxidizing agent, in comparison with the unmodified parent phosphate (Fig. 2).
Page 7 of 24
Fig. 2 XRD-patterns of the fresh TEOS-p catalyst, and the catalyst after benzene oxidation at 450°C using N2O and
ip t
an air-ammonia mixture as oxidizing agents, respectively.
On the contrary to the former catalysts, the Ca-Cu phosphate prepared with a SAPEG7
cr
surfactant exhibited no significant alternation of the phase structure upon the reactor tests. No major chemical reduction occurred neither using nitrous oxide, nor an air-ammonia mixture as
us
oxidizing agents (Fig. 3). Interestingly, even in the fresh sample there was a CuO/Cu2O oxide pair already present. The amount of Cu2O phase in the used catalysts did not increase, neither a
Ac ce p
te
d
M
an
phosphite phase and metallic copper did not appear (Fig. 3).
Fig. 3 XRD-patterns of the fresh SAPEG7-p catalyst, and after benzene oxidation at 450°C using N2O and an airammonia mixture as oxidizing agents, respectively.
As the unstabilized (untemplated) phosphate catalyst is transforming into different phases, detected by XRD, the change in the texture is undoubted. The templation by SAPEG7 reduces these changes.
SEM – characterization The morphology of the copper-calcium phosphate catalysts can be significantly altered by changing the method of their preparation. A simple precipitation of metal nitrates by ammonium hydrogen phosphate resulted in a rather crystalline structure with relatively large crystals, with
Page 8 of 24
the sizes above 1μm (Fig. 4). On the other hand, a TEOS-doped phosphate exhibits much finer structure. In this case the phosphate precipitated in an aqueous solution, where the micelles of pre-hydrolyzed TEOS were already present. Thus, the role of TEOS was not to form zeolitic structures, as it is often applied in the synthesis of zeolites in combination with surfactants. The
ip t
silica, formed upon the hydrolysis of TEOS served as crystallization nuclei for the phosphate precipitates. In this way, the sizes of the phosphate particles dropped dramatically bellow 500 nm,
cr
and in the resulting material we can observe the presence of isolated phosphate and silica particles (Fig. 4 – larger phosphate particles indicated by circles, finer silica network indicated by
us
rectangles).
The precipitation of the metal nitrate solution in the presence of SAPEG7 resulted in a phosphate precipitate with a spongiest structure. For this structure there is observed a set of larger
an
cavities (about 500 nm) and finer channels with diameters ranging from several 100 nm to about 10 nm (Fig. 4). The change in the texture did affect also the specific surface area of the catalysts
M
(Tab.1). However, on the contrary to the unmodified catalyst, the apparent volume of the spongy-
Ac ce p
te
d
like Ca-Cu phosphate catalyst increased from 3.17 ml.g-1 to 6.06 ml.g-1.
Page 9 of 24
ip t cr
us
Fig. 4: SEM images of the parent (top), TEOS-p (middle) and SAPEG7-p (bottom) catalysts.
an
BET – characterization
The specific surface areas of the phosphate catalysts correspond with their morphology
M
observed by SEM. The unmodified Ca-Cu phosphate having large crystals had the lowest surface area SBET = 8 m2.g-1. The silica formed by decomposition of TEOS squashed the large crystals of the phosphate matrix, resulting in a remarkable increase of the specific surface area to 52 8 m2.g. However, taking into account a large number of silica particles (measuring about 100 nm)
d
1
te
formed, the increase in the surface area should in major part be assigned to the surface of the silica. The highest specific surface area, 63 m2.g-1, was observed for the phosphate catalyst
Ac ce p
prepared in the presence of the SAPEG7 surfactant. The untemplated phospahes, regardless the nature of cations, had rather low surface areas, < 10 m2.g-1.
TPR – characterization It was observed that during catalytic tests the unmodified Ca-Cu phosphate significantly looses the reducible components. The overall reducibility (area bellow the TPR-curve) for the
Page 10 of 24
used catalyst dropped to 32% of its initial value. Moreover, we can see a shift to higher reduction temperatures (Fig. 5). Interestingly, for the SAPEG7-modified catalyst the total reducibility decreased only slightly (to 85% of the initial value), mainly the most easily reducible moieties are
us
cr
ip t
lost at reduction temperatures bellow 323°C.
an
Fig. 5: TPR patterns of the parent (a), and SAPEG7-p (b) catalysts. Heating rate 10 K.min-1,100 mg of sample,
M
hydrogen flow 20 ml.min-1.
d
XPS –characterization
te
The XPS-patterns of the parent, fresh Ca-Cu phosphate catalyst shows phosphorus being most probably in a rather uniform pyrophosphate state. Minor amounts of other phosphates
Ac ce p
cannot be excluded; nevertheless the dominant form is P2O74- (Fig. 6, P 2p3 peak at 133.5 eV). For the copper, the oxidation state corresponds to 2+, however, two different chemical surroundings appear to be present. The first one, ca.70% of the amount of copper, being most probably a pyrophosphate (peak at 931.0 eV), and the second one, with lower binding energy. This state can represent CuO, or copper in extra-framework positions in the phosphates, what is typical e.g. for Cu in ion-exchange positions. The XPS data for the Ca-Cu-Fe phosphate catalysts have shown similar patterns for phosphorus and copper, the iron being in a single pyrophosphate form, in the oxidation state 3+.
Page 11 of 24
ip t
Fig. 6: XPS patterns of phosphorus and copper in the SAPEG7-p catalyst.
cr
3.2 Hydroxylation of benzene by nitrous oxide
us
In accordance with our previous results [16], copper-calcium phosphate catalysts have shown significant selectivity in benzene hydroxylation. Practically only two reaction products were detected, phenol and CO2, with the selectivity to phenol exceeding 95% (Fig.7). The
an
reaction products formed in trace amounts were CO, benzofuran, aniline and polyaromatics. For the Ca-Cu phosphate catalysts prepared by co-precipitation, the phenol yields were correlating
M
with the copper content, reaching a maximum at about 43% Cu content (Fig.7). The phosphate catalysts prepared by precipitation on ion-exchanged silica surface produced only trace amounts of phenol. Obviously, water injection to the reactant stream, in the aim to strip the phenol formed
d
on the catalyst, was beneficial in terms of phenol yield (Fig.7). Unfortunately, a gradual
Ac ce p
te
deterioration of the catalytic activity can be observed in the case of all catalysts from this group.
Page 12 of 24
Fig. 7.: Catalytic performance of the Ca-Cu phosphate catalysts: parent – open symbols, and 5Cu-p – full symbols without (left) and with water injection (right) to the reactant stream, using N2O as an oxidizing agent. ♦ phenol, ●
ip t
CO2.
3.3 Hydroxylation of benzene by air-ammonia mixture
cr
Interestingly, the performance of air-ammonia mixture as an oxidizing agent is similar to that of N2O – water mixture (Fig. 8). Using molecular oxygen together with a reducing agent is a
us
known oxidizing system used for the selective oxidations, e.g. the mixture of H2+O2, or the catalytic systems with co-reductants [17]. Hence, the addition of ammonia to the reactant stream
an
can have at least two effects: it can either act as a co-reductant, or it can act as a source of nitrous oxide generated in-situ by the oxidation of ammonia [16]. To distinguish between these effects we tested other amines and aldehydes instead of ammonia, as ethylenediamine, formaldehyde and
M
acetaldehyde in a form of a 25% aqueous solution. As it is seen in Fig.8, the admixture resulting in the highest phenol yield was ammonia, but similarly to other amines, the catalyst underwent
d
deactivation in the course of several hours on stream (TOS). The maximal phenol yields, obtained
Ac ce p
te
by the addition of formaldehyde or acetaldehyde, were equally low, 0.4%.
Fig. 8: Catalytic performance of the parent Ca-Cu phosphate catalyst, using air-ammonia (open symbols) and airethylenediamine (full symbols) mixtures as oxidizing agents. ♦ phenol, ● CO2.
3.4 The effect of modification of phosphates
Page 13 of 24
The effect of other metals One has to distinguish between several routes how metals can be incorporated into a
ip t
phosphate matrix prepared in aqueous media. One possible route is co-precipitation of the phosphates (or binary phosphates), the second route is using the ion-exchange capacity of
cr
phosphates (e.g. hydroxyapatites), hence introducing the doping metal into the ion-exchange positions in the previously formed phosphate matrix. Another possible route is the use of
us
organometallic precursors in non-aqueous media. Hence, the introduced metal is located neither in the ion-exchange positions, nor in the phosphate matrix. By using the ion exchange method to introduce transition metals into calcium phosphates, copper had been reported as a most
an
catalytically active metal and palladium as the second one [16]. On the other hand, using ionexchanged calcium phosphates iron was found as a rather inactive metal, in spite of iron being
M
already known as a catalytically promising metal when using in zeolitic structures [12]. Investigating the above mentioned most promising metals, copper incorporated in a phosphate matrix, we found that albeit palladium produces some phenol, but an over oxidation of the
d
substrate with a detrimental effect on the selectivity takes place. Iron itself in the phosphate
te
matrix was not active for benzene hydroxylation; however, in combination with copper it has
Ac ce p
shown synergistic effect (Fig. 9).
Fig. 9: Phenol yields over the parent (♦), Fe-p(■) and Cu-Fe-p(▲)catalysts using N2O as an oxidizing agent.
Page 14 of 24
The effect of a silica-based substructure Attempting to increase the available surface of the copper phosphate catalyst,
ip t
precipitation of the phosphate on the pre-hydrolyzed TEOS precursor was carried out. This procedure resulted in two achievements: it elongated the lifetime of the catalyst, and enhanced
cr
the phenol yields (Fig. 10). Very similar results were attained using active carbon FW285 as a
M
an
us
template (not shown).
te
The effect of surfactants
d
Fig.10: Catalytic performance of the TEOS-p sample, using N2O as an oxidizing agent. ♦ phenol, ● CO2.
Ac ce p
Addition of surfactants during the precipitation of the phosphate catalysts was found to be a promising tool to increase the catalysts performance. The effect of surfactants is well known for the preparation of zeolites. However, in the case of phosphate catalysts it obviously brings about a similar structure-directing effect (Fig. 4), and an enhancement of the phenol yield (Fig. 11, Tab. 2).The SAPEG7 surfactant, used during the precipitation of a Ca-Cu phosphate, led to the preparation of Ca-Cu phosphate with sponge-like structure, which was catalytically highly active and selective towards phenol production. Under not optimized reaction conditions and using N2O and an air-ammonia mixture as oxidants the yield of phenol at the benzene space velocity (WHSV) 1gbenzene.gcatalyst-1.h-1 reached 6.5mol% and 4.2mol%, respectively. Moreover, the high advantage of the Ca-Cu phosphate catalysts, prepared in the presence of proper surfactants, is their significantly elongated lifetime; their catalytic activity not being rapidly lost with long time on stream. For example three repeated experiments, each one with freshly prepared catalyst
Page 15 of 24
shown that the activity of the sponge-like Ca-Cu phosphate catalyst after 37 h on stream decreased only by about 25%. In contrast, over unmodified Ca-Cu phosphate catalyst the phenol
M
an
us
cr
ip t
yield already after 8h on stream drops to about one third of the initial value.
Fig. 11: Catalytic performance of the Ca-Cu phosphate catalyst prepared in the presence of the surfactant SAPEG7,
d
using N2O (open symbols) and air-ammonia mixture (full symbols) as an oxidizing agent. ♦ phenol, ● CO2.
te
Tab.2 The effect of surfactants on the maximal yield of phenol (Ymax). Catalyst: templated Ca-Cu mixed phosphate, oxidizing agent: 70 ml.min-1 N2O or air-ammonia mixture; reaction temperature 450°C.
SAPEG7
Ac ce p
surfactant
Pluronic
DBS
no
Ymax / % (N2O)
6.5
2.2
2.0
2.0
Ymax / % (air-NH3)
4.2
1.1
1.0
0.7
4. Discussion
Copper in phosphate and zeolitic matrixes have been extensively studied for direct hydroxylation of benzene [32]. An often observed phenomenon in these catalytic systems is the reduction of copper during catalytic tests. Hence, a question emerges: what oxidation state of copper is active for oxygen transfer to benzene molecule, and what structure of the copper catalyst is able to stabilize the copper in an appropriate form. Our observation, that in the deactivated catalyst a massive reduction of copper to metallic state took place, is consistent with
Page 16 of 24
previous reports [16]. Ammonia, as well as benzene are known reducing agents at the reaction temperature 450 °C. The reduction of the oxidation state of copper upon elongated contact with the reaction mixture is evidnet from the XRD, XPS and TPR patterns. Copper in different matrixes can be principally reduced by two routes: a one-step route, Cu2+ →Cu0
ip t
where Cu2+ is directly reduced to Cu0: (1)
cr
and a two-step route, where Cu2+ is first reduced to Cu1+, and consecutively to metallic copper: Cu1+ →Cu0
(2)
us
Cu2+ →Cu1+
(3)
an
Over non-interacting supports, as well as in bulk state, CuO is reduced according to reaction (1), while on strongly interacting supports the preferred route of copper reduction
M
follows reactions (2) and (3) [34]. Since Cu1+ is more resistant against reduction than Cu2+, reaction (3) is likely to occur at temperatures higher than reaction (2); the temperature difference being typically more than 100°C. Since in the TPR patterns are more than two peaks (Fig. 5),
d
there must be at least two different types of copper in the initial sample. In addition, when the
te
areas of low-temperature and high-temperature peaks do not correspond to the 1:1 stoichiometry of the reactions (2) and (3), there should have been originally at least two different forms of Cu2+
Ac ce p
in the sample [35]. For example in zeolites the distinctly different types of Cu localization, and thus the different forms of copper can be present, according to the structure of a particular zeolite. In the mixed phosphates the different possible forms of copper cannot be exactly assigned to a fixed reduction temperature; copper can be present as isolated Cu cations in ion–exchange positions, as O-bridged Cu dimers or chains, Cu0 nanoclusters, crystalline Cu, Cu2O and CuO of different sizes.
When comparing the XRD and TPR patterns, it is not surprising that the overall reducibility of the unmodified Ca-Cu phosphate after the catalytic test decreased dramatically, since a peak of metallic copper in the XRD patterns appeared. In the fresh samples, for both the SAPEG7-modified and unmodified phosphates, four well-distinguished peaks can be observed, with a shoulder at around 250°C. This shoulder can be assigned to the most easily reducible form of copper, to non-interacting CuO particles [36]. This
Page 17 of 24
shoulder, together with the first distinct peaks, at 308 or 323°C respectively, disappears in the both used samples. The remaining three peaks are preserved, though their intensity is deteriorated in the case of the unmodified sample. The most striking difference between SAPEG7-modified and unmodified phosphates
ip t
appears in the temperature range of 345-380°C, where the second distinct peak in the TPR patterns is present. For the untemplated catalyst the intensity of this peak is decreasing upon the
cr
catalytic test, what appears to be a result of the chemical reduction of copper, hence the deterioration of the amounts of reducible copper. However, for the SAPEG7-modified phosphate,
us
this peak is becoming more pronounced in the used sample. This increase in the intensity of the peak is possible only by creation of a reducible form of copper. The only form of reducible copper moieties, created in otherwise reductive reaction conditions during the catalytic test, is
an
Cu1+. According to the reactions (2) and (3), the intensity of the Cu1+ peak is possible to increase only for the price of a decrease in the intensity of Cu2+ (the first peak). That means, in the highly
M
active catalyst (SAPEG7-modified), the weakly interacting Cu2+ cations (peaks at 308 and 302°C) are transformed to Cu1+ cations. Thus, the overall amount of Cu1+ is increasing during the catalytic test, but is not reduced further to Cu0 [20]. On the other hand, for the less active catalyst
d
(the parent untemplated phosphate), instead of the increase of the amount of Cu1+ in the TPR
te
pattern, one can see a formation of Cu0 in the XRD data (Fig. 1). The high-temperature peaks above 400°C represent the bulky, crystalline phosphates with
Ac ce p
strongly bound copper. Since they are reduced above 400°C in pure hydrogen, and the catalytic reaction is carried out at 450°C in the presence of less reducing benzene, the redox function of these copper species during the catalytic tests is unreachable. These peaks, slightly shifted to higher reduction temperatures, can be assigned to the creation of more hardly reducible Cu1+ species.
One cannot correlate directly the XRD and TPR results, mainly due to the XRD being less accurate in the estimation of the exact amount of the certain compound, and due to the fact that some weakly crystalline phases remain hidden in the XRD patterns. Nevertheless, these data indicate several facts: (i) the peak at 382°C in the TPR patterns appears to be most resistant against extinction in the used catalysts, (ii) for the very active catalyst this peak is even more pronounced in the used catalyst, (iii) the crystalline structure of the very active catalysts appears (according to the XRD-patterns) to be rather unaltered during the catalytic tests. Taking into
Page 18 of 24
account the decrease of hydrogen uptake for the SAPEG7-modified catalyst (by 15 %) this should represent the portion of copper which is catalytically active for benzene hydroxylation. This portion of copper should be apparently poorly crystalline, and it did not contribute to any significant phase in the XRD-patterns.
ip t
The both co-precipitated Ca-Cu phosphates prepared without additional adjustment of pH, and with the addition of pre-hydrolyzed TEOS, were obtained under acidic conditions (the pH
cr
was ~6 or ~4, respectively). On the other hand, the addition of SAPEG7 shifted the pH of the mixture to a slightly alkaline region (pH~8). The phase distribution of the phosphates formed,
us
especially in the region of diffractions 40-45°, representing the catalytically accessible Cu2+-Cu1+ pair seems to correspond mostly to the pH value during the preparation of the catalyst. In fact, all the poorer catalysts were obtained in conditions where the pH value was less than 7.
an
The main drawback of many copper-type catalysts, although being highly selective for benzene hydroxylation, is generally their fast degradation due to the reduction of copper to the
M
metallic state. Copper-containing catalysts, where the TOF of the catalyst decreases in a few hours practically to zero, are routinely reported in the literature. This pinpoints the bottleneck of the problem of the copper-catalyzed benzene hydroxylation: the reoxidation of the copper.
d
Preparing copper-calcium phosphate by a simple procedure in the presence of surfactants,
te
resulted in an unexpected elongation of the catalysts lifetime and the increase in phenol yield. Moreover, such types of catalysts are also active when using air-ammonia mixture as an oxidizing
Ac ce p
agent. The deactivation of the catalysts is apparently caused by the reduction of copper, rather than a progressive buildup of coke deposits. After 37 h TOS the overall carbon content of the catalyst was found to be <1%. That implies that the coking does not proceed significantly after reaching a steady state level, whereas the gradual reduction of the copper is the process limiting the catalysts lifetime.
5. Conclusions A series of copper-based phosphates were prepared and tested as catalysts for gas-phase hydroxylation of benzene to phenol using nitrous oxide and an air-ammonia mixture as oxidizing agents. Among the investigated phosphate-type catalysts, the highest activity and stability was
Page 19 of 24
exhibited using catalysts prepared by precipitation in a slightly basic medium in the presence of nitrogen-containing surfactants. It was shown that using a proper surfactant, the copper-calcium phosphate precipitates in the form of a fine sponge-like structure with a set of channels ranging down to ~10 nm. A stearyl
ip t
amine polyethylene glycol ether surfactant resulted in the catalyst, where the resistance against the reduction of copper to metallic state during the catalytic tests was achieved by stabilization
cr
the Cu1+ species in the phosphate matrix. Over these sponge-like catalysts, highly selective (>95%) hydroxylation of benzene to phenol proceeds not only with nitrous oxide but also with an
us
air-ammonia mixture as oxidizing agents. Under not optimized reaction conditions and using these oxidants the yield of phenol at the WHSV 1h-1 reached 6.5% and 4.2%, respectively. The substitution of ammonia with other reducing agents, mainly those not containing a nitrogen atom,
an
led to a deterioration of the phenol yield, typically bellow 2%.
Another great advantage of copper phosphate type catalyst, prepared by a simple method
M
in the presence of suitable surfactants is that they attain high catalytic activity. The activity of the sponge-like Ca-Cu phosphate catalyst decreased only by about 25%. In contrast, over unmodified
Ac ce p
te
the initial activity.
d
Ca-Cu phosphate catalyst the phenol yield already after 8h on stream drops to about one third of
6. References
[1] R.J. Schmidt, Appl. Catal. A: Gen. 280 (2005) 89–103. [2] M.H. Sayyar, R.J. Wakeman, Chem. Eng. Res. Des. 86 (2008) 517–526. [3] C. Thurman (Ed.), Kirk-Othmer Encyclopedia of Chemical Technology: Phenol 17 (1982) 373-384. [4] H. Nakajima, M. Koya, H. Ishida, H. Minoura, Y. Takamatsu, Microporous Mater. 2 (1994) 237-243.
Page 20 of 24
[5] A.J.J. Koekkoek, H.C. Xin, Q.H. Yang, C. Li, E.J.M. Hensen, Micropor. Mesopor. Mat. 145 (2011) 172-181. [6] H. Yang, J.Q. Chen, J. Li, Y. Lv, S. Gao, Appl. Catal. A: Gen. 415 (2012) 22-28. [8] R. Dittmeyer, L. Bortolotto, Appl. Catal. A: Gen. 391 (2011) 311-318.
ip t
[7] P.P. Zhao, L. Yan, J. Wang, Chem. Eng. J. 204 (2012) 72-78. [9] Y. Zhu, Y. Dong, L. Zhao, F. Yuan, J. Mol. Catal. A: Chem. 315 (2010) 205-212.
cr
[10] L. Hu, B. Yue, X. Chen, H. He,Catal. Commun. 43 (2014) 179-183.
[11] L. Wang, S. Yamamoto, S. Malwadkar, S. Nagamatsu, T. Sasaki, K. Hayashizaki, M. Tada, Y.
us
Iwasawa, ChemCatChem 5 (2013) 2203–2206.
[12] J. Okamura, S. Nishiyama, S. Tsuruya, M. Masai, J. Mol. Catal. A: Chem. 135 (1998) 133– 142.
an
[13] J.A. van Bokhoven, ChemCatChem. 1 (2009) 363-364.
[14] B. Ene, T. Archipov, E. Roduner, J. Phys. Chem. C 115 (2011) 3688-3694.
M
[15] A. Kubacka, Z. Wang, B. Sulikowski, V. Cortés Corberán, J. Catal. 250 (2007) 184–189. [16] J. Čejka, G. Centi, J. Perez-Pariente, W. J. Roth, Catal. Today 179 (2012) 2– 15. [17] G. Centi, S. Perathoner, F. Pino, R. Arrigo, G. Giordano, A. Katovic, V. Pedulà, Catal. Today
d
110 (2005) 211-220. 128-136.
te
[18] I. Yuranov, D. A. Bulushev, A. Renken, L. Kiwi-Minsker, Appl. Catal. A: Gen. 319 (2007)
Ac ce p
[19] S. Sklenak, P. C. Andrikopoulos, B. Boekfa, B. Jansang, J. Nováková, L. Benco, T. Bucko, J. Hafner, J. Dědeček, Z. Sobalík, J. Catal. 272 (2010) 262-274. [20] A.A. Ivanov, V.S. Chernyavsky, M.J. Gross, A.S. Kharitonov, A.K. Uriarte, G.I. Panov, Appl. Catal. A: Gen. 249 (2003) 327-343.
[21] R. Navarro, S. Lopez-Pedrajas, D. Luna, J.M. Marinas, F.M. Bautista., Appl. Catal. A: Gen. (2013), doi. 10.1016/j.apcata.2013.08.043 [22] H. Tounsi, S. Djemal, C. Petitto, G. Delahay: Appl. Catal. B: Environmental 107 (2011) 158– 163 [23] P.Blazy, E.-A. Jdid, C.R. Acad. Sci. Paris 325 (1997) 761-764 [24] P.Blazy, J.-C. Samama, C. R. Acad. Sci. Paris 333 (2001) 271–276 [25] A. K. Özer , M. S. Gülaboglu, S. Bayrakceken, W. Weisweiler, Adv. Powder Technol. 17 (2006) 481–494
Page 21 of 24
[26] S. Raynaud, E. Champion, D. Bernache-Assollant, Biomaterials 23 (2002) 1073–1080 [27] P. Khemthong, P. Daorattanachai, N. Laosiripojana, K. Faungnawakij, Catal. Commun. 29 (2012) 96–100 [28] S. Induja, P.S. Raghavan, Catal. Commun. 33 (2013) 7–10.
ip t
[29] C. Díaz, M.L. Valenzuela, V. Lavayen, K. Mendoza, D.O. Peña, C. O’Dwyer, Inorg. Chim. Acta 377 (2011) 5–13.
cr
[30] N. Khan, V. A. Morozov, S. S. Khasanov, B. I. Lazoryak, Mater. Res. Bull. 32 (1997) 12111220.
us
[31] A. Benarafa, M. Kacimi, G. Coudurier, M. Ziyad, Appl. Catal. A: Gen. 196 (2000) 25-35. [32] M. Bahidsky, M. Hronec, Catal. Today 91–92 (2004) 13–16.
[33] G. Centi, S. Perathoner, Appl. Catal. A: Gen. 132 (1995) 179-259. J. Solid State Chem. 184 (2011) 1915–1923.
an
[34] D. L. Hoang, T. T. H. Dang, J. Engeldinger, M. Schneider, J. Radnik, M. Richter, A. Martin,
M
[35] G. C. Bond, S. N. Namijo, J. S. Wakeman, J. Mol. Catal. 64 (1991) 305-319. [36] C. Tom-Abreu, M.F. Ribeiro, C. Henriques, G. Delahay, Appl. Catal. B: Environmental 14
Ac ce p
te
d
(1997) 261-272.
Page 22 of 24
Highlights: - Sponge-like Ca-Cu phosphate were prepared - They catalyze selective oxidation of benzene to phenol - >95% selectivity and 4.2mol% yield was achieved
Ac ce p
te
d
M
an
us
cr
- During 37h on stream the catalyst was not notably deactivated
ip t
- The air-ammonia mixture is the oxidant
Page 23 of 24
Ac
ce
pt
ed
M
an
us
cr
i
*Graphical Abstract (for review)
Page 24 of 24
S0926-860X(14)00320-2 http://dx.doi.org/doi:10.1016/j.apcata.2014.05.006 APCATA 14826
To appear in:
Applied Catalysis A: General
Received date: Revised date: Accepted date:
10-2-2014 16-4-2014 6-5-2014
ˇ ˇ Please cite this article as: B. Horv´ath, M. Sustek, J. Skriniarov´ a, M. Omastov´a, E. Dobroˇcka, M. Hronec, Gas phase hydroxylation of benzene with air-ammonia mixture over copper-based phosphate catalysts, Applied Catalysis A, General (2014), http://dx.doi.org/10.1016/j.apcata.2014.05.006 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Gas phase hydroxylation of benzene with air-ammonia mixture over copper-based phosphate catalysts Blažej Horvátha,*, Martin Šusteka, Jaroslava Škriniarováb, Mária Omastovác, Edmund Dobročkad
ip t
and Milan Hroneca,* a
Department of Organic Technology, bInstitute of Electronics and Fotonics, Slovak University of
cr
Technology, Radlinského 9, 81237Bratislava, Slovakia c
Polymer Institute, dInstitute of Electrical Engineering, Slovak Academy of Sciences,
us
Dúbravská cesta 9, 845 41 Bratislava, Slovakia
an
Abstract
The direct gas phase hydroxylation of benzene to phenol over copper-based phosphate
M
catalysts has been studied. Nitrous oxide and air-ammonia mixtures were used as oxidizing agents. It has been found that nitrous oxide can be replaced by air-ammonia mixture without
d
significant deterioration of the phenol yield. Nanostructured Ca-Cu phosphate catalysts with
te
a dominant mixed pyrophosphate structure, prepared in the presence of suitable surfactants were found highly active and selective. Using surfactant-modified phosphate catalysts with sponge-like
Ac ce p
structure above 95% selectivity and about 6.5% and 4.2% phenol yields were achieved at the benzene WHSV 1 h-1 and under not optimized reaction conditions using N2O and air-ammonia mixture as oxidizing agents, respectively. These catalysts, compared with unmodified ones, were found to be much more resistant towards deactivation caused by the chemical reduction of catalytically active copper phase during the catalytic test. The activity of the spongy Ca-Cu phosphate catalyst after 37 h on stream decreased only by about 25%. In contrast, over similar unmodified Ca-Cu phosphate catalyst the phenol yield already after 8h on stream drops to about one third of the initial value. Keywords: benzene hydroxylation, copper phosphates, nitrous oxide, ammonia, catalyst lifetime *corresponding authors, e-mail address: [email protected] (B.Horváth) [email protected] (M.Hronec)
Page 1 of 24
1. Introduction Benzene as a raw material is widely used in many industrial processes, including phenol production. More than 90% of the world’s phenol production is based on the cumene route that
ip t
implies alkylation of benzene, oxidation of cumene by a free-radical mechanism and cleavage of the formed hydroperoxide catalyzed by a mineral acid [1]. Despite its great success, the cumene process has some disadvantages such as high environmental impact, the use of corrosive catalyst,
cr
the production of an explosive intermediate and a co-production of acetone. Therefore, the success of this process significantly depends on the marketability of the by-product acetone and
us
on the ratio between the demands for phenol vs. acetone [2]. Other processes of industrial manufacture of phenol, as the Dow process [3, 4] involving toluene oxidation to benzoic acid and
an
its decarboxylation using a copper salt catalyst to phenol, represent only a minor contribution to the commercial phenol production.
M
Several studies have been published during the past two decades on the synthesis of phenol via one-step oxidation of benzene. Direct hydroxylation of the aromatic ring is a challenging reaction to synthesize phenol, due to the formation of over-oxygenated by-products.
d
The synthesis of phenol by direct hydroxylation of benzene was extensively studied using
te
different oxidants, such as N2O, O2, H2O2 as well as the H2-O2 mixture [5-8]. Using hydrogen peroxide to oxidize benzene, the self-decomposition and the price of
Ac ce p
hydrogen peroxide is the most common limiting factor. However, using vanadium oxide supported on mesoporous SBA-16 [9] or metal doped heteropolyacids [10], hydrogen peroxide – based systems reach high selectivities. An attractive catalytic system for the hydroxylation of benzene should use oxygen, preferably air as a cheap and available oxidant. Several catalytic systems, active in the hydroxylation of benzene with O2 [11], often use a co-reducing agent apart from the oxygen [12]. It is well known, that copper-based phosphate catalysts, depending on its nanostructure can alter their catalytic properties [13]. For the direct hydroxylation of benzene the role of the oxidation state of copper and finding of the phase responsible for the catalytic activity [14-16] are other important keys for understanding the mechanism of the catalytic hydroxylation of benzene.
One of the most widely studied catalytic systems for the gas phase hydroxylation of aromatics with N2O uses Fe-zeolite catalysts [17, 18]. Fe/ZSM-5 is reported to be highly active
Page 2 of 24
and stable in DeNOx systems designed to decompose N2O [19]. Over similar catalysts phenol is formed with a selectivity exceeding 90% [20]. One of the major problems of these systems was their rapid deactivation (either published or observed in practice). Similarly, a very rapid deactivation was reported also over AlPO catalysts [21]. Another drawback of catalytic systems
ip t
using N2O as oxidant is their dependence on the source of waste nitrous oxide, usually obtained from adipic acid plants [22].
cr
In the present work we focused attention on the preparation copper-based phosphate catalysts for gas-phase oxidation of benzene, using nitrous oxide and an air-ammonia mixture as
an
influence on catalytic properties are mainly discussed.
us
oxidizing agents. The oxidation state and the structure of the copper-based domains and their
M
2. Experimental
d
2.1 Catalysts preparation and characterization
As a basis of comparison, copper-calcium-phosphate with 43 % of copper (relative to the
te
overall amount of cations) was prepared by precipitation of the calcium and copper nitrate solutions (1.00 M) with a 0.60 M ammonium monohydrate orthophosphate solution at ambient
Ac ce p
temperature. The hydrogen phosphate solution was added drop wise to the metal nitrate solution while stirring vigorously. The precipitate was stirred for 3h, then filtered and excessively washed with deionized water to extract the relics of the precursors, until pH 7. The catalysts with 5% copper content were prepared using a similar procedure. Iron-calcium-phosphate was prepared accordingly, replacing copper nitrate with an equimolar amount of ferric nitrate monohydrate. Mixed Cu-Fe-Ca-phosphate catalysts were prepared as well, by replacing 50 mol% of copper nitrate by ferric nitrate. A Pd-Cu-Ca-phosphate catalyst was prepared by introducing palladium by wet impregnation from a 1.00M Pd(acac)2 solution in toluene onto a pre-prepared Cu-Ca-phosphate sample (43% Cu). The overall palladium content of the resulting material, calculated as metallic palladium, was 1.5 wt%. To prepare phosphate catalysts with sponge-like structure, various structure-directing agents were added to the solution of metal nitrates in an amount of 10 wt% relative to the
Page 3 of 24
precipitate (calculated as metal pyrophosphates) prior to the precipitation by ammonium hydrogen phosphate. Stearyl amine polyethylene glycol ether with 7 ethoxy units (SAPEG7), Pluronic P-123, dodecylbenzene sulphonate-Na (DBS), as well as active carbon FW285 (Evonik) were tested as structure-directing agents.
ip t
Silica-modified catalysts were prepared by two methods: (i) by precipitating the metal ion-exchanged silicas with an ammonium hydrogen phosphate solution (denoted as Cu-IE), and
cr
(ii) by adding pre-hydrolyzed tetraethyl orthosilicate (TEOS) to the metal nitrate solutions in a ratio 1:1. The former set was prepared by stirring overnight a 1 wt% suspension of silica Aerosil®
us
200 (Degussa, declared BET-surface area 200 m2.g-1) with an equimolar amount of copper nitrate in aqueous solution. The silica was then filtered, washed with deionized water until neutrality to sodium sulphide and re-suspended in 10% aqueous solution of ammonium monohydrate
an
orthophosphate. The suspension was stirred overnight, filtered, washed excessively, dried at 180°C and calcined. The latter set was prepared by pre-hydrolyzing 5 wt% of TEOS in a 0.05 N
M
aqueous solution of hydrochloric acid, at the temperature of 50°C for 3h, while stirring vigorously. The resulting solution was used to dissolve the metal nitrates, and was precipitated by
d
ammonium hydrogen phosphate as described above. The precipitate was stirred at 80°C for 24h, filtered, washed until neutrality, and dried. All the catalysts were calcined prior to catalytic tests
Ac ce p
summarized in Table 1.
te
at 600°C in air flow during 5 h, with a heating rate of 100 K.h-1. The prepared catalysts are
Table 1: Summary of the prepared mixed Ca-Cu phosphate catalysts. Metal contents are given as per cents relative to overall cation content. denotation
Cu content / %
modifier
parent
43
no
5
no
Cu-Fe-p
21.5
21.5% Fe
Pd-Cu-p
43
2% Pd
SAPEG7-p
43
SAPEG7
P123-p
43
Pluronic P123
DBS-p
43
DBS
FW285-p
43
Active carbon FW285
5Cu-p
Page 4 of 24
43
TEOS
Cu-IE
<1
Silica ion-exchanged by Cu
cr
ip t
TEOS-p
The catalysts were characterized by powder X-ray diffraction using a Stadi Stoe apparatus
us
in Bragg-Brentano geometry and a Bruker D8 DISCOVER diffractometer in a parallel beam geometry with parabolic Goebel mirror in the primary beam; the diffraction patterns were recorded in grazing incidence set-up with the angle of incidence α = 6°. Temperature
an
programmed reduction and BET-measurements were carried out using a Micromeritics Pulse Chemisorb 2700 apparatus. SEM measurements were carried out using a JEOL JSM-7500F
M
microscope; XPS were measured using a Thermo Scientific K-Alpha XPS system with an Al Kα
Ac ce p
2.2 Catalytic tests
te
d
X-ray source at 6 mA, 12 kV.
The experiments were carried out in a fixed-bed glass reactor at the temperature of 450°C, using 1 g of catalyst. The catalysts prior to testing were pelletized and crushed to the grain size of 0.2 – 0.3 mm. The temperature was controlled by a thermocouple in the axis of the catalytic bed. The gaseous oxidizing agent (air or N2O) with the flow rate of 70 ml.min-1, was saturated by benzene (1.0g.h-1) before entering the reactor. In experiments where air-ammonia mixture was used as an oxidizing agent, aqueous ammonia solution (with the ammonia concentration of 25%) was added by a syringe pump (2.0g.h-1) to the reactant stream. In some experiments, only water was injected instead of ammonia solution. The reaction products were collected in an ice-cooled flask, homogenized with methanol and analyzed by GC, equipped with FID for the detection of organic products, and TCD for the analysis of CO2 in the gaseous products. To verify the composition of reaction products a Shimadzu QP-500 GC-MS was used.
Page 5 of 24
ip t
3. Results
cr
3.1 Catalysts characterization
us
XRD – characterization
The parent Ca-Cu phosphate represents a mixed phase system (Fig. 1). As being a coprecipitate, one can expect various separated phosphates of copper and calcium, and bimetallic
an
phosphates. The calcination temperature of the precipitate, 600°C, presages the dominance of pyrophosphates over orthophosphates or hydrogen phosphates in the resulting material [23-25].
M
Hydroxyapatites, if formed, can retain their structure above the calcination temperature employed. However, the pH of the reaction mixture and the Ca/P ratio is unfavorable for the formation of hydroxyapatites [24,26]. Hence, as expected, no hydroxyapatite or other separated
d
calcium orthophosphate or hydrogen phosphate was found in detectable concentrations in the Ca-
te
Cu phosphate materials. Similarly, the presence of copper hydrogen phosphate was not detected. Copper, if formed as a separated phase, is known to transform to copper pyrophosphate at the
Ac ce p
temperature of 600°C [27,28]. Copper and calcium phosphate precipitates can undergo solid-state reactions to form binary phosphates at 600°C [2, 29]. Accordingly, in the parent Ca-Cu phosphate sample (with 43% Cu), the copper and calcium pyrophosphates were found as dominant phases in the diffraction region bellow 40°. However, the shift of the crystallographic parameters (about 1°) indicates that a partial mutual substitution of Cu and Ca occured in these phases, leading to a deformation of the crystalline structure. This cation exchange takes place most probably via solid-phase reactions during the calcination. The occurence of minor binary phosphates, as Ca19Cu2(PO4)14, with a deformed whitlockite (Ca3(PO4)2) structure cannot be excluded [30]. Whitlockit, being one of the most abundant phosphate structures, can form a plethora of bimetallic phases with copper [31]. In the diffraction region 40 – 45° the presence of extra-framework CuO and Cu2O is observable. The
Page 6 of 24
occurence of paracrystalline or crystalline extra-framework copper oxides in calcined phosphates was repeatedly reported [32,33]. The phase composition of the Cu-Ca phosphate catalyst is obviously changing during the catalytic tests, using nitrous oxide or air-ammonia mixture as oxidizing agents. The initial
ip t
deformed Ca-Cu pyrophosphate structure, with minor amounts of CuO and whitlockit-type binary orthophosphates, exhibited after the catalytic tests several alternations. These changes
cr
indicate that during the catalytic tests several changes took place: (i) the partial reduction of phosphorus and an emergence of a phosphite phase, (ii) the reduction of extra-framework CuO to
Ac ce p
te
d
M
an
us
Cu2O and (iii) the appearance of metallic copper.
Fig. 1 XRD-patterns of the fresh parent Cu-Ca phosphate catalyst, and after benzene oxidation at 450°C using N2O and an air-ammonia mixture as oxidizing agents, respectively. * Ca19Cu2(PO4)14, ▼Ca2P2O7, ▲Cu2P2O7, ♦Cu(PO3)2, ◊Cu, ●CuO, ○Cu2O
The Ca-Cu phosphate prepared on a silica-based matrix, modified by the pre-hydrolyzed TEOS, has shown a similar initial phase composition, as well as similar changes upon the catalytic tests. The rate of its chemical reduction, however, was milder when using N2O as an oxidizing agent, in comparison with the unmodified parent phosphate (Fig. 2).
Page 7 of 24
Fig. 2 XRD-patterns of the fresh TEOS-p catalyst, and the catalyst after benzene oxidation at 450°C using N2O and
ip t
an air-ammonia mixture as oxidizing agents, respectively.
On the contrary to the former catalysts, the Ca-Cu phosphate prepared with a SAPEG7
cr
surfactant exhibited no significant alternation of the phase structure upon the reactor tests. No major chemical reduction occurred neither using nitrous oxide, nor an air-ammonia mixture as
us
oxidizing agents (Fig. 3). Interestingly, even in the fresh sample there was a CuO/Cu2O oxide pair already present. The amount of Cu2O phase in the used catalysts did not increase, neither a
Ac ce p
te
d
M
an
phosphite phase and metallic copper did not appear (Fig. 3).
Fig. 3 XRD-patterns of the fresh SAPEG7-p catalyst, and after benzene oxidation at 450°C using N2O and an airammonia mixture as oxidizing agents, respectively.
As the unstabilized (untemplated) phosphate catalyst is transforming into different phases, detected by XRD, the change in the texture is undoubted. The templation by SAPEG7 reduces these changes.
SEM – characterization The morphology of the copper-calcium phosphate catalysts can be significantly altered by changing the method of their preparation. A simple precipitation of metal nitrates by ammonium hydrogen phosphate resulted in a rather crystalline structure with relatively large crystals, with
Page 8 of 24
the sizes above 1μm (Fig. 4). On the other hand, a TEOS-doped phosphate exhibits much finer structure. In this case the phosphate precipitated in an aqueous solution, where the micelles of pre-hydrolyzed TEOS were already present. Thus, the role of TEOS was not to form zeolitic structures, as it is often applied in the synthesis of zeolites in combination with surfactants. The
ip t
silica, formed upon the hydrolysis of TEOS served as crystallization nuclei for the phosphate precipitates. In this way, the sizes of the phosphate particles dropped dramatically bellow 500 nm,
cr
and in the resulting material we can observe the presence of isolated phosphate and silica particles (Fig. 4 – larger phosphate particles indicated by circles, finer silica network indicated by
us
rectangles).
The precipitation of the metal nitrate solution in the presence of SAPEG7 resulted in a phosphate precipitate with a spongiest structure. For this structure there is observed a set of larger
an
cavities (about 500 nm) and finer channels with diameters ranging from several 100 nm to about 10 nm (Fig. 4). The change in the texture did affect also the specific surface area of the catalysts
M
(Tab.1). However, on the contrary to the unmodified catalyst, the apparent volume of the spongy-
Ac ce p
te
d
like Ca-Cu phosphate catalyst increased from 3.17 ml.g-1 to 6.06 ml.g-1.
Page 9 of 24
ip t cr
us
Fig. 4: SEM images of the parent (top), TEOS-p (middle) and SAPEG7-p (bottom) catalysts.
an
BET – characterization
The specific surface areas of the phosphate catalysts correspond with their morphology
M
observed by SEM. The unmodified Ca-Cu phosphate having large crystals had the lowest surface area SBET = 8 m2.g-1. The silica formed by decomposition of TEOS squashed the large crystals of the phosphate matrix, resulting in a remarkable increase of the specific surface area to 52 8 m2.g. However, taking into account a large number of silica particles (measuring about 100 nm)
d
1
te
formed, the increase in the surface area should in major part be assigned to the surface of the silica. The highest specific surface area, 63 m2.g-1, was observed for the phosphate catalyst
Ac ce p
prepared in the presence of the SAPEG7 surfactant. The untemplated phospahes, regardless the nature of cations, had rather low surface areas, < 10 m2.g-1.
TPR – characterization It was observed that during catalytic tests the unmodified Ca-Cu phosphate significantly looses the reducible components. The overall reducibility (area bellow the TPR-curve) for the
Page 10 of 24
used catalyst dropped to 32% of its initial value. Moreover, we can see a shift to higher reduction temperatures (Fig. 5). Interestingly, for the SAPEG7-modified catalyst the total reducibility decreased only slightly (to 85% of the initial value), mainly the most easily reducible moieties are
us
cr
ip t
lost at reduction temperatures bellow 323°C.
an
Fig. 5: TPR patterns of the parent (a), and SAPEG7-p (b) catalysts. Heating rate 10 K.min-1,100 mg of sample,
M
hydrogen flow 20 ml.min-1.
d
XPS –characterization
te
The XPS-patterns of the parent, fresh Ca-Cu phosphate catalyst shows phosphorus being most probably in a rather uniform pyrophosphate state. Minor amounts of other phosphates
Ac ce p
cannot be excluded; nevertheless the dominant form is P2O74- (Fig. 6, P 2p3 peak at 133.5 eV). For the copper, the oxidation state corresponds to 2+, however, two different chemical surroundings appear to be present. The first one, ca.70% of the amount of copper, being most probably a pyrophosphate (peak at 931.0 eV), and the second one, with lower binding energy. This state can represent CuO, or copper in extra-framework positions in the phosphates, what is typical e.g. for Cu in ion-exchange positions. The XPS data for the Ca-Cu-Fe phosphate catalysts have shown similar patterns for phosphorus and copper, the iron being in a single pyrophosphate form, in the oxidation state 3+.
Page 11 of 24
ip t
Fig. 6: XPS patterns of phosphorus and copper in the SAPEG7-p catalyst.
cr
3.2 Hydroxylation of benzene by nitrous oxide
us
In accordance with our previous results [16], copper-calcium phosphate catalysts have shown significant selectivity in benzene hydroxylation. Practically only two reaction products were detected, phenol and CO2, with the selectivity to phenol exceeding 95% (Fig.7). The
an
reaction products formed in trace amounts were CO, benzofuran, aniline and polyaromatics. For the Ca-Cu phosphate catalysts prepared by co-precipitation, the phenol yields were correlating
M
with the copper content, reaching a maximum at about 43% Cu content (Fig.7). The phosphate catalysts prepared by precipitation on ion-exchanged silica surface produced only trace amounts of phenol. Obviously, water injection to the reactant stream, in the aim to strip the phenol formed
d
on the catalyst, was beneficial in terms of phenol yield (Fig.7). Unfortunately, a gradual
Ac ce p
te
deterioration of the catalytic activity can be observed in the case of all catalysts from this group.
Page 12 of 24
Fig. 7.: Catalytic performance of the Ca-Cu phosphate catalysts: parent – open symbols, and 5Cu-p – full symbols without (left) and with water injection (right) to the reactant stream, using N2O as an oxidizing agent. ♦ phenol, ●
ip t
CO2.
3.3 Hydroxylation of benzene by air-ammonia mixture
cr
Interestingly, the performance of air-ammonia mixture as an oxidizing agent is similar to that of N2O – water mixture (Fig. 8). Using molecular oxygen together with a reducing agent is a
us
known oxidizing system used for the selective oxidations, e.g. the mixture of H2+O2, or the catalytic systems with co-reductants [17]. Hence, the addition of ammonia to the reactant stream
an
can have at least two effects: it can either act as a co-reductant, or it can act as a source of nitrous oxide generated in-situ by the oxidation of ammonia [16]. To distinguish between these effects we tested other amines and aldehydes instead of ammonia, as ethylenediamine, formaldehyde and
M
acetaldehyde in a form of a 25% aqueous solution. As it is seen in Fig.8, the admixture resulting in the highest phenol yield was ammonia, but similarly to other amines, the catalyst underwent
d
deactivation in the course of several hours on stream (TOS). The maximal phenol yields, obtained
Ac ce p
te
by the addition of formaldehyde or acetaldehyde, were equally low, 0.4%.
Fig. 8: Catalytic performance of the parent Ca-Cu phosphate catalyst, using air-ammonia (open symbols) and airethylenediamine (full symbols) mixtures as oxidizing agents. ♦ phenol, ● CO2.
3.4 The effect of modification of phosphates
Page 13 of 24
The effect of other metals One has to distinguish between several routes how metals can be incorporated into a
ip t
phosphate matrix prepared in aqueous media. One possible route is co-precipitation of the phosphates (or binary phosphates), the second route is using the ion-exchange capacity of
cr
phosphates (e.g. hydroxyapatites), hence introducing the doping metal into the ion-exchange positions in the previously formed phosphate matrix. Another possible route is the use of
us
organometallic precursors in non-aqueous media. Hence, the introduced metal is located neither in the ion-exchange positions, nor in the phosphate matrix. By using the ion exchange method to introduce transition metals into calcium phosphates, copper had been reported as a most
an
catalytically active metal and palladium as the second one [16]. On the other hand, using ionexchanged calcium phosphates iron was found as a rather inactive metal, in spite of iron being
M
already known as a catalytically promising metal when using in zeolitic structures [12]. Investigating the above mentioned most promising metals, copper incorporated in a phosphate matrix, we found that albeit palladium produces some phenol, but an over oxidation of the
d
substrate with a detrimental effect on the selectivity takes place. Iron itself in the phosphate
te
matrix was not active for benzene hydroxylation; however, in combination with copper it has
Ac ce p
shown synergistic effect (Fig. 9).
Fig. 9: Phenol yields over the parent (♦), Fe-p(■) and Cu-Fe-p(▲)catalysts using N2O as an oxidizing agent.
Page 14 of 24
The effect of a silica-based substructure Attempting to increase the available surface of the copper phosphate catalyst,
ip t
precipitation of the phosphate on the pre-hydrolyzed TEOS precursor was carried out. This procedure resulted in two achievements: it elongated the lifetime of the catalyst, and enhanced
cr
the phenol yields (Fig. 10). Very similar results were attained using active carbon FW285 as a
M
an
us
template (not shown).
te
The effect of surfactants
d
Fig.10: Catalytic performance of the TEOS-p sample, using N2O as an oxidizing agent. ♦ phenol, ● CO2.
Ac ce p
Addition of surfactants during the precipitation of the phosphate catalysts was found to be a promising tool to increase the catalysts performance. The effect of surfactants is well known for the preparation of zeolites. However, in the case of phosphate catalysts it obviously brings about a similar structure-directing effect (Fig. 4), and an enhancement of the phenol yield (Fig. 11, Tab. 2).The SAPEG7 surfactant, used during the precipitation of a Ca-Cu phosphate, led to the preparation of Ca-Cu phosphate with sponge-like structure, which was catalytically highly active and selective towards phenol production. Under not optimized reaction conditions and using N2O and an air-ammonia mixture as oxidants the yield of phenol at the benzene space velocity (WHSV) 1gbenzene.gcatalyst-1.h-1 reached 6.5mol% and 4.2mol%, respectively. Moreover, the high advantage of the Ca-Cu phosphate catalysts, prepared in the presence of proper surfactants, is their significantly elongated lifetime; their catalytic activity not being rapidly lost with long time on stream. For example three repeated experiments, each one with freshly prepared catalyst
Page 15 of 24
shown that the activity of the sponge-like Ca-Cu phosphate catalyst after 37 h on stream decreased only by about 25%. In contrast, over unmodified Ca-Cu phosphate catalyst the phenol
M
an
us
cr
ip t
yield already after 8h on stream drops to about one third of the initial value.
Fig. 11: Catalytic performance of the Ca-Cu phosphate catalyst prepared in the presence of the surfactant SAPEG7,
d
using N2O (open symbols) and air-ammonia mixture (full symbols) as an oxidizing agent. ♦ phenol, ● CO2.
te
Tab.2 The effect of surfactants on the maximal yield of phenol (Ymax). Catalyst: templated Ca-Cu mixed phosphate, oxidizing agent: 70 ml.min-1 N2O or air-ammonia mixture; reaction temperature 450°C.
SAPEG7
Ac ce p
surfactant
Pluronic
DBS
no
Ymax / % (N2O)
6.5
2.2
2.0
2.0
Ymax / % (air-NH3)
4.2
1.1
1.0
0.7
4. Discussion
Copper in phosphate and zeolitic matrixes have been extensively studied for direct hydroxylation of benzene [32]. An often observed phenomenon in these catalytic systems is the reduction of copper during catalytic tests. Hence, a question emerges: what oxidation state of copper is active for oxygen transfer to benzene molecule, and what structure of the copper catalyst is able to stabilize the copper in an appropriate form. Our observation, that in the deactivated catalyst a massive reduction of copper to metallic state took place, is consistent with
Page 16 of 24
previous reports [16]. Ammonia, as well as benzene are known reducing agents at the reaction temperature 450 °C. The reduction of the oxidation state of copper upon elongated contact with the reaction mixture is evidnet from the XRD, XPS and TPR patterns. Copper in different matrixes can be principally reduced by two routes: a one-step route, Cu2+ →Cu0
ip t
where Cu2+ is directly reduced to Cu0: (1)
cr
and a two-step route, where Cu2+ is first reduced to Cu1+, and consecutively to metallic copper: Cu1+ →Cu0
(2)
us
Cu2+ →Cu1+
(3)
an
Over non-interacting supports, as well as in bulk state, CuO is reduced according to reaction (1), while on strongly interacting supports the preferred route of copper reduction
M
follows reactions (2) and (3) [34]. Since Cu1+ is more resistant against reduction than Cu2+, reaction (3) is likely to occur at temperatures higher than reaction (2); the temperature difference being typically more than 100°C. Since in the TPR patterns are more than two peaks (Fig. 5),
d
there must be at least two different types of copper in the initial sample. In addition, when the
te
areas of low-temperature and high-temperature peaks do not correspond to the 1:1 stoichiometry of the reactions (2) and (3), there should have been originally at least two different forms of Cu2+
Ac ce p
in the sample [35]. For example in zeolites the distinctly different types of Cu localization, and thus the different forms of copper can be present, according to the structure of a particular zeolite. In the mixed phosphates the different possible forms of copper cannot be exactly assigned to a fixed reduction temperature; copper can be present as isolated Cu cations in ion–exchange positions, as O-bridged Cu dimers or chains, Cu0 nanoclusters, crystalline Cu, Cu2O and CuO of different sizes.
When comparing the XRD and TPR patterns, it is not surprising that the overall reducibility of the unmodified Ca-Cu phosphate after the catalytic test decreased dramatically, since a peak of metallic copper in the XRD patterns appeared. In the fresh samples, for both the SAPEG7-modified and unmodified phosphates, four well-distinguished peaks can be observed, with a shoulder at around 250°C. This shoulder can be assigned to the most easily reducible form of copper, to non-interacting CuO particles [36]. This
Page 17 of 24
shoulder, together with the first distinct peaks, at 308 or 323°C respectively, disappears in the both used samples. The remaining three peaks are preserved, though their intensity is deteriorated in the case of the unmodified sample. The most striking difference between SAPEG7-modified and unmodified phosphates
ip t
appears in the temperature range of 345-380°C, where the second distinct peak in the TPR patterns is present. For the untemplated catalyst the intensity of this peak is decreasing upon the
cr
catalytic test, what appears to be a result of the chemical reduction of copper, hence the deterioration of the amounts of reducible copper. However, for the SAPEG7-modified phosphate,
us
this peak is becoming more pronounced in the used sample. This increase in the intensity of the peak is possible only by creation of a reducible form of copper. The only form of reducible copper moieties, created in otherwise reductive reaction conditions during the catalytic test, is
an
Cu1+. According to the reactions (2) and (3), the intensity of the Cu1+ peak is possible to increase only for the price of a decrease in the intensity of Cu2+ (the first peak). That means, in the highly
M
active catalyst (SAPEG7-modified), the weakly interacting Cu2+ cations (peaks at 308 and 302°C) are transformed to Cu1+ cations. Thus, the overall amount of Cu1+ is increasing during the catalytic test, but is not reduced further to Cu0 [20]. On the other hand, for the less active catalyst
d
(the parent untemplated phosphate), instead of the increase of the amount of Cu1+ in the TPR
te
pattern, one can see a formation of Cu0 in the XRD data (Fig. 1). The high-temperature peaks above 400°C represent the bulky, crystalline phosphates with
Ac ce p
strongly bound copper. Since they are reduced above 400°C in pure hydrogen, and the catalytic reaction is carried out at 450°C in the presence of less reducing benzene, the redox function of these copper species during the catalytic tests is unreachable. These peaks, slightly shifted to higher reduction temperatures, can be assigned to the creation of more hardly reducible Cu1+ species.
One cannot correlate directly the XRD and TPR results, mainly due to the XRD being less accurate in the estimation of the exact amount of the certain compound, and due to the fact that some weakly crystalline phases remain hidden in the XRD patterns. Nevertheless, these data indicate several facts: (i) the peak at 382°C in the TPR patterns appears to be most resistant against extinction in the used catalysts, (ii) for the very active catalyst this peak is even more pronounced in the used catalyst, (iii) the crystalline structure of the very active catalysts appears (according to the XRD-patterns) to be rather unaltered during the catalytic tests. Taking into
Page 18 of 24
account the decrease of hydrogen uptake for the SAPEG7-modified catalyst (by 15 %) this should represent the portion of copper which is catalytically active for benzene hydroxylation. This portion of copper should be apparently poorly crystalline, and it did not contribute to any significant phase in the XRD-patterns.
ip t
The both co-precipitated Ca-Cu phosphates prepared without additional adjustment of pH, and with the addition of pre-hydrolyzed TEOS, were obtained under acidic conditions (the pH
cr
was ~6 or ~4, respectively). On the other hand, the addition of SAPEG7 shifted the pH of the mixture to a slightly alkaline region (pH~8). The phase distribution of the phosphates formed,
us
especially in the region of diffractions 40-45°, representing the catalytically accessible Cu2+-Cu1+ pair seems to correspond mostly to the pH value during the preparation of the catalyst. In fact, all the poorer catalysts were obtained in conditions where the pH value was less than 7.
an
The main drawback of many copper-type catalysts, although being highly selective for benzene hydroxylation, is generally their fast degradation due to the reduction of copper to the
M
metallic state. Copper-containing catalysts, where the TOF of the catalyst decreases in a few hours practically to zero, are routinely reported in the literature. This pinpoints the bottleneck of the problem of the copper-catalyzed benzene hydroxylation: the reoxidation of the copper.
d
Preparing copper-calcium phosphate by a simple procedure in the presence of surfactants,
te
resulted in an unexpected elongation of the catalysts lifetime and the increase in phenol yield. Moreover, such types of catalysts are also active when using air-ammonia mixture as an oxidizing
Ac ce p
agent. The deactivation of the catalysts is apparently caused by the reduction of copper, rather than a progressive buildup of coke deposits. After 37 h TOS the overall carbon content of the catalyst was found to be <1%. That implies that the coking does not proceed significantly after reaching a steady state level, whereas the gradual reduction of the copper is the process limiting the catalysts lifetime.
5. Conclusions A series of copper-based phosphates were prepared and tested as catalysts for gas-phase hydroxylation of benzene to phenol using nitrous oxide and an air-ammonia mixture as oxidizing agents. Among the investigated phosphate-type catalysts, the highest activity and stability was
Page 19 of 24
exhibited using catalysts prepared by precipitation in a slightly basic medium in the presence of nitrogen-containing surfactants. It was shown that using a proper surfactant, the copper-calcium phosphate precipitates in the form of a fine sponge-like structure with a set of channels ranging down to ~10 nm. A stearyl
ip t
amine polyethylene glycol ether surfactant resulted in the catalyst, where the resistance against the reduction of copper to metallic state during the catalytic tests was achieved by stabilization
cr
the Cu1+ species in the phosphate matrix. Over these sponge-like catalysts, highly selective (>95%) hydroxylation of benzene to phenol proceeds not only with nitrous oxide but also with an
us
air-ammonia mixture as oxidizing agents. Under not optimized reaction conditions and using these oxidants the yield of phenol at the WHSV 1h-1 reached 6.5% and 4.2%, respectively. The substitution of ammonia with other reducing agents, mainly those not containing a nitrogen atom,
an
led to a deterioration of the phenol yield, typically bellow 2%.
Another great advantage of copper phosphate type catalyst, prepared by a simple method
M
in the presence of suitable surfactants is that they attain high catalytic activity. The activity of the sponge-like Ca-Cu phosphate catalyst decreased only by about 25%. In contrast, over unmodified
Ac ce p
te
the initial activity.
d
Ca-Cu phosphate catalyst the phenol yield already after 8h on stream drops to about one third of
6. References
[1] R.J. Schmidt, Appl. Catal. A: Gen. 280 (2005) 89–103. [2] M.H. Sayyar, R.J. Wakeman, Chem. Eng. Res. Des. 86 (2008) 517–526. [3] C. Thurman (Ed.), Kirk-Othmer Encyclopedia of Chemical Technology: Phenol 17 (1982) 373-384. [4] H. Nakajima, M. Koya, H. Ishida, H. Minoura, Y. Takamatsu, Microporous Mater. 2 (1994) 237-243.
Page 20 of 24
[5] A.J.J. Koekkoek, H.C. Xin, Q.H. Yang, C. Li, E.J.M. Hensen, Micropor. Mesopor. Mat. 145 (2011) 172-181. [6] H. Yang, J.Q. Chen, J. Li, Y. Lv, S. Gao, Appl. Catal. A: Gen. 415 (2012) 22-28. [8] R. Dittmeyer, L. Bortolotto, Appl. Catal. A: Gen. 391 (2011) 311-318.
ip t
[7] P.P. Zhao, L. Yan, J. Wang, Chem. Eng. J. 204 (2012) 72-78. [9] Y. Zhu, Y. Dong, L. Zhao, F. Yuan, J. Mol. Catal. A: Chem. 315 (2010) 205-212.
cr
[10] L. Hu, B. Yue, X. Chen, H. He,Catal. Commun. 43 (2014) 179-183.
[11] L. Wang, S. Yamamoto, S. Malwadkar, S. Nagamatsu, T. Sasaki, K. Hayashizaki, M. Tada, Y.
us
Iwasawa, ChemCatChem 5 (2013) 2203–2206.
[12] J. Okamura, S. Nishiyama, S. Tsuruya, M. Masai, J. Mol. Catal. A: Chem. 135 (1998) 133– 142.
an
[13] J.A. van Bokhoven, ChemCatChem. 1 (2009) 363-364.
[14] B. Ene, T. Archipov, E. Roduner, J. Phys. Chem. C 115 (2011) 3688-3694.
M
[15] A. Kubacka, Z. Wang, B. Sulikowski, V. Cortés Corberán, J. Catal. 250 (2007) 184–189. [16] J. Čejka, G. Centi, J. Perez-Pariente, W. J. Roth, Catal. Today 179 (2012) 2– 15. [17] G. Centi, S. Perathoner, F. Pino, R. Arrigo, G. Giordano, A. Katovic, V. Pedulà, Catal. Today
d
110 (2005) 211-220. 128-136.
te
[18] I. Yuranov, D. A. Bulushev, A. Renken, L. Kiwi-Minsker, Appl. Catal. A: Gen. 319 (2007)
Ac ce p
[19] S. Sklenak, P. C. Andrikopoulos, B. Boekfa, B. Jansang, J. Nováková, L. Benco, T. Bucko, J. Hafner, J. Dědeček, Z. Sobalík, J. Catal. 272 (2010) 262-274. [20] A.A. Ivanov, V.S. Chernyavsky, M.J. Gross, A.S. Kharitonov, A.K. Uriarte, G.I. Panov, Appl. Catal. A: Gen. 249 (2003) 327-343.
[21] R. Navarro, S. Lopez-Pedrajas, D. Luna, J.M. Marinas, F.M. Bautista., Appl. Catal. A: Gen. (2013), doi. 10.1016/j.apcata.2013.08.043 [22] H. Tounsi, S. Djemal, C. Petitto, G. Delahay: Appl. Catal. B: Environmental 107 (2011) 158– 163 [23] P.Blazy, E.-A. Jdid, C.R. Acad. Sci. Paris 325 (1997) 761-764 [24] P.Blazy, J.-C. Samama, C. R. Acad. Sci. Paris 333 (2001) 271–276 [25] A. K. Özer , M. S. Gülaboglu, S. Bayrakceken, W. Weisweiler, Adv. Powder Technol. 17 (2006) 481–494
Page 21 of 24
[26] S. Raynaud, E. Champion, D. Bernache-Assollant, Biomaterials 23 (2002) 1073–1080 [27] P. Khemthong, P. Daorattanachai, N. Laosiripojana, K. Faungnawakij, Catal. Commun. 29 (2012) 96–100 [28] S. Induja, P.S. Raghavan, Catal. Commun. 33 (2013) 7–10.
ip t
[29] C. Díaz, M.L. Valenzuela, V. Lavayen, K. Mendoza, D.O. Peña, C. O’Dwyer, Inorg. Chim. Acta 377 (2011) 5–13.
cr
[30] N. Khan, V. A. Morozov, S. S. Khasanov, B. I. Lazoryak, Mater. Res. Bull. 32 (1997) 12111220.
us
[31] A. Benarafa, M. Kacimi, G. Coudurier, M. Ziyad, Appl. Catal. A: Gen. 196 (2000) 25-35. [32] M. Bahidsky, M. Hronec, Catal. Today 91–92 (2004) 13–16.
[33] G. Centi, S. Perathoner, Appl. Catal. A: Gen. 132 (1995) 179-259. J. Solid State Chem. 184 (2011) 1915–1923.
an
[34] D. L. Hoang, T. T. H. Dang, J. Engeldinger, M. Schneider, J. Radnik, M. Richter, A. Martin,
M
[35] G. C. Bond, S. N. Namijo, J. S. Wakeman, J. Mol. Catal. 64 (1991) 305-319. [36] C. Tom-Abreu, M.F. Ribeiro, C. Henriques, G. Delahay, Appl. Catal. B: Environmental 14
Ac ce p
te
d
(1997) 261-272.
Page 22 of 24
Highlights: - Sponge-like Ca-Cu phosphate were prepared - They catalyze selective oxidation of benzene to phenol - >95% selectivity and 4.2mol% yield was achieved
Ac ce p
te
d
M
an
us
cr
- During 37h on stream the catalyst was not notably deactivated
ip t
- The air-ammonia mixture is the oxidant
Page 23 of 24
Ac
ce
pt
ed
M
an
us
cr
i
*Graphical Abstract (for review)
Page 24 of 24