university of groningen oxidative dehydrogenation of

University of Groningen

Oxidative dehydrogenation of ethylbenzene under industrially relevant conditionsZarubina, Valeriya

IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite fromit. Please check the document version below.

Document VersionPublisher's PDF, also known as Version of record

Publication date:2015

Link to publication in University of Groningen/UMCG research database

Citation for published version (APA):Zarubina, V. (2015). Oxidative dehydrogenation of ethylbenzene under industrially relevant conditions: onthe role of catalyst structure and texture on selectivity and stability. [S.n.].

CopyrightOther than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of theauthor(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).

The publication may also be distributed here under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license.More information can be found on the University of Groningen website: https://www.rug.nl/library/open-access/self-archiving-pure/taverne-amendment.

Take-down policyIf you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediatelyand investigate your claim.

Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons thenumber of authors shown on this cover page is limited to 10 maximum.

Download date: 14-12-2021

https://research.rug.nl/en/publications/oxidative-dehydrogenation-of-ethylbenzene-under-industrially-relevant-conditions(c5f5bc05-8fba-4968-ad9c-69c6b0cea3cf).html

Making Coke a more Efficient Catalyst in the Oxidative Dehydrogenation of Ethylbenzene using wide-pore Transitional Aluminas

The thermal activation of a silica-stabilized -alumina impacts positively on the oxidative

dehydrogenation of ethylbenzene (EB) to styrene (ST). A systematic thermal study reveals that the

transition from -alumina into transitional phases at 1050 oC leads to an optimal enhancement of

both conversion and selectivity under pseudo-steady state conditions; where active and selective

coke has been deposited. The effect is observed in the reaction temperature range of 450-475 oC

at given operation conditions resulting in the highest ST yield, while at 425 oC this effect is lost due

to incomplete O2 conversion. The conversion increase is ascribed to the ST selectivity improvement

that makes more O2 available for the main ODH reaction. The fresh aluminas and catalytically

active carbon deposits on the spent catalysts were characterized by gas adsorption (N2 and Argon),

acidity evaluation by NH3-TPD and pyridine adsorption monitored by FTIR, thermal and elemental

analyses, solubility in CH2Cl2 and MALDI-TOF to correlate the properties of both phases with the ST

selectivity enhancement. Such an increase in selectivity was interpreted by the lower reactivity of

the carbon deposits that diminished the COx formation. The site requirements of the optimal

catalyst to create the more selective coke is related to the higher density of Lewis sites per surface

area, no mixed Si-Al Brønsted sites are formed, while the acid strength of the formed Lewis sites is

relatively weaker than those of the bare alumina.

Chapter 2

32

2.1 Introduction

Styrene (ST) is industrially produced by direct dehydrogenation of ethylbenzene (EB) using steam at 580-630 °C [1]. The process suffers from high energy consumption due to low conversion per pass because of equilibrium limitations, and the high

temperatures required for the endothermic reaction. The operation cost mainly depends on oil prices and utilities; decreasing the reaction temperature can have a remarkable market impact. Oxidative dehydrogenation (ODH) can address these

drawbacks [2], but it is not commercialized among others mainly due to the limited catalyst stability and the lower selectivity to styrene.

A distinctive catalyst support for this reaction is -alumina due to its mild

acidity as compared to e.g. zeolites. The coke deposits generated on the alumina

under these oxidative reaction conditions do not produce deactivation, but promote the activity and selectivity [3-6]. There is a general consensus that these carbon

deposits species are partially oxygenated and act as active and selective catalytic sites [3-17]. This has been confirmed by employing pre-coked catalysts [18] and by measuring the intrinsic activity of the carbon deposits separated from the alumina.

Those carbon deposits were superior in respect to the alumina [6,15]. Nederlof et al. [17] demonstrated that this also holds for CO2-asisisted EB ODH; during the first 15 h time on stream the alumina shows an increase in ethylbenzene conversion from 15

to 60% and styrene selectivity from 60 to 92%. Characterizations of the Al2O3 samples after different times on stream show a clear correlation between the formed coke and the activity and selectivity. This is also shown in this work for the O2-based

EB ODH, where the activation period can be observed at reaction temperatures lower than 475 oC; at temperatures > 475 oC the coke build-up is fast.

The alumina by itself is not an inactive support but it provides Lewis centres

and those generate „good coke‟ [7,14,15]; the Brønsted acidity does not play a key role in forming selective coke for this reaction. These findings on active coke established a trend setting by using synthetic carbons that mimic the in-situ

produced deposits during EB ODH [18-41]. Typically, their conversion levels are acceptable, 50-70%, while the styrene selectivity is relatively low for commercialization, up to 70%, compared to the commercial process of steam

dehydrogenation (ST selectivity > 95%). In the framework of the EB ODH catalyst development, less attention has

been given on the use of inorganic supports, in particular about strategies to improve

either conversion, selectivity or stability of their carbonaceous sites or combinations thereof. Few examples indicate that the acidity enhancement of -Al2O3 by H3PO4

[11] or HBO3 [42,43] has a positive impact on the styrene yield. The effect of the

texture of the alumina phases at high temperature, to the best of our knowledge, has not been investigated systematically. In this study, we use thermally treated SiO2 promoted -Al2O3 at high temperature; the pore size is expanded by coarsening

the nanoparticles to investigate the effect of pore size on the reaction performance.

We have observed that this treatment also modifies the nature of the carbon deposit formed; making it less reactive and consequently more selective to ST. The relationships between EB ODH performance and the alumina and coke properties are

discussed in this paper and rationalize previous observations [44] where the EB conversion and ST selectivity were optimal for a specific phase composition at high temperature.


33

2.2 Experimental

2.2.1 Materials

In this study a low SiO2 stabilized γ-Al2O3 extrudates (Albemarle Catalysts BV) were employed. The extrudates were crushed and sieved into a 212-425 μm fraction used

for the catalytic tests and characterization. The material was thermally activated at temperatures ranging from 500 to 1200 oC in ceramic crucibles using a box furnace in static air at a rate of 4 oC/min and held for 8 h. The samples were labeled as

AluX00, where X00 corresponds to the treatment temperature in degree Celsius. Other inorganic commercial supports and catalysts were tested as reference: TiO2 (Degussa, P25), SiO2 (Fuji Silysia, G-6 5 microns), CeO2 nanopowder (Sigma-Aldrich,

700290), mesoporous Al-MSU-F type (Sigma Aldrich, 643629) and Merck ultrapure alumina (1.01095.1000).

2.2.2 Characterization of bare and spent catalysts

The coke content and stability of the spent catalysts (after 60 h time on stream)

were determined by thermogravimetric analysis in a Mettler-Toledo analyzer (TGA/SDTA851e) using a flow of synthetic air of 100 ml/min (STP). The temperature was increased from 30 to 900 °C at 10 °C/min. Blank curve subtraction using an

empty crucible was employed. CHN elemental analyses were carried out in a EuroVector 3000 CHNS

analyzer. Approximately 2 mg of sample was accurately weighed in a 6-digit analytic

balance (Mettler Toledo). The samples were burnt at 1800 oC in the presence of an oxidation catalyst and decomposed into CO2, H2O, and N2. These gases are then separated in a Porapak QS column at 80 oC and quantified with a TCD detector.

Acetonitrile (99.9%) was used as an external standard. Textural analysis of the fresh and spent catalysts was carried out by N2 and Ar

physisorption at -196 oC and -186 oC, respectively, in a Micromeritics ASAP 2420

analyzer. Prior to the measurements all samples were outgassed under vacuum at 350 oC for 10 h for the fresh materials, while for the spent catalysts a mild degassing was applied, 130 oC for 24 h. The surface area was calculated by BET method [45],

SBET. The single point pore volume (VT) was estimated from the amount adsorbed at a relative pressure of 0.98 in the desorption branch. The pore size distribution was derived from the BJH model [45]. For the fresh materials, the adsorption pore sizes

are used. In case of spent catalysts, the desorption branch is favored to evidence the presence of closed pores.

NH3-TPD experiments were carried out in a Micromeritics AutoChem II system

equipped with a thermal conductivity detector (TCD). The sample (ca. 30 mg) was pre-treated by heating it up to 500 °C in He at 10 °C/min. The sample was cooled to 120 °C at a similar cooling rate, and then exposed to 1 vol% NH3/He (25 ml/min) for

30 min. Subsequently, a flow of He (25 ml/min) was passed through the reactor for 60 min to remove weakly adsorbed NH3 from the samples‟ surface. After this

baseline stabilization, the desorption of NH3 was monitored in the range of 120-1000 °C using a heating rate of 10 °C/min.

The acidity was also evaluated by pyridine adsorption monitored by IR. FTIR

spectra were recorded on a Thermo Scientific Nicolet 6700 FTIR spectrometer equipped with a MCT-B detector and a quartz home-made IR cell having CaF2 windows connected to a vacuum system that allows the controlled dosing of pyridine.

Around 50 mg of sample were pressed in a 1.767 cm2 self-supporting wafer and mounted into the quartz cell. Before dosing the pyridine, the sample was preheated under vacuum (5.10-6 mm Hg) as follows: 120 oC for 2 h at 1 oC/min, then at 400 oC

for 2h, 1 oC/min followed by cooling to 150 oC. Pyridine gas was dosed into the

Chapter 2

34

sample wafer until full saturation at 150 oC (i.e. the pyridine bands become stable).

Afterwards, the sample wafer was evacuated and cooled down to room temperature. Spectra were measured by accumulating 256 scans at a resolution 4 cm−1; the spectrum of the sample after evacuation was subtracted to each spectrum. The

strength of the acid sites was evaluated by recording the spectra after evacuating the sample at increasing temperatures, with 50 oC increments. The spectra were taken once the pressure remains stable. The density of acid sites was calculated

according to Tamura et al. [46]: W

SACW

where CW is the acid site density (mol/g); is the molar extinction coefficient (1.71

cm.mol-1 for Lewis sites on alumina [46]); S is the surface of the sample disc area

(cm2), W the sample weight (mg), and A is the peak area (cm-1). The skeletal density was obtained by He pycnometry at room temperature

after evacuating the sample chamber (1 ml) 5 times and measuring in 10 cycles; the standard deviation for each analysis is given.

XRD analysis was carried out in a Philips PW1840 using a CuK radiation. The

phase identification was done using the JCPDS database.

2.2.3 Characterization of the active coke

The coke nature and composition were analyzed after separating it from the inorganic support. The spent catalyst was treated in dichloromethane (CH2Cl2) under

reflux for 6 h, which removes the soluble coke compounds retained on the sample [47]. The percentage of insoluble coke (IC) in total was determined by the difference in the TGA weight loss between the spent catalyst and the catalyst after refluxing in

CH2Cl2. The carbonaceous insoluble compounds were obtained by treating the alumina matrix in a hydrofluoric acid solution (40 wt.% HF) at room temperature overnight, washing of the residue by centrifugation and subsequently drying at 90 °C

for 10 h. This produces brown-to-black particles that were subjected to CHN analysis to quantify the empirical CxHyOz formula. For the high-temperature spent samples, the alumina matrix could not be fully dissolved due to its inertness; the inorganic

residue was taken into account for the calculations of the corrected C and H concentrations in the coke while O was calculated by difference. The presence of O2, a temperature of 1800 oC and an oxidation catalyst in the CHN analysis can ensure

that the formation of Al-oxycarbide is avoided. The insoluble carbon material was mildly analyzed by MALDI time of flight

(TOF) spectrometry [48-51]. These IC particles were suspended in tetrahydrofuran

(THF) by sonication and mixed with dithranol as a matrix. MALDI spectra were recorded on an Applied Biosystems, Inc. Voyager-DE PRO spectrometer by pulsed ion extraction using a solid state laser with a wavelength of 355 nm. Laser desorption-

ionization (LDI)-TOF was also performed in the same instrument. Samples were suspended in THF, 1 L was spotted directly on a stainless steel target plate. In both

modes, the measurements were performed in linear mode using positive ionization.

Spectra were accumulated between 100 or 300 and 3000 or 10000 Da (depending on spectrum) and calibrated externally.

2.3 Catalytic tests

The catalytic tests were carried out in a parallel flow micro reactor using a fixed volume of catalyst (0.8 ml) using down-flow 4 mm quartz reactors. The reactors are

typically loaded as follows: quartz wool, 10 cm glass pearls (0.5 mm diameter), the catalyst, 10 cm glass pearls (0.5 mm diameter), and a second quartz wool plug. In this way, it is assured that the catalyst bed is located in the isothermal operational

zone of the furnace. In all cases the same volume of catalyst bed was used,


35

corresponding to 65 mm of bed length; so the sample mass increases with increasing packing density of the catalyst. The inertness of the glass beads was verified as they showed less than 3% EB conversion under all applied conditions. The reactor gas

feed is a mixture that can consist of CO2, N2, and air that counts for a gas-flow rate of 36 ml (NTP)/min; the liquid EB-feed flow rate is 1 g/h (3.54 ml (NTP)/min vapour) that is evaporated upstream each reactor in a -Al2O3 column, resulting in a 1:9

molar ratio of ethylbenzene to gas (10 vol. % EB). This corresponds with operation at a GHSV of 3000 l/l/h. The total pressure was 1.2-1.3 105 Pa. The pseudo steady-state conditions refer at the reaction time where both the conversion and selectivity

achieve the highest values, which are typically achieved after 2 to 6 h time on stream; the reported conversion data are given at 6-10 h time on stream or longer. The conversion and selectivity were averaged in the isothermal period after reaching

pseudo steady-state conditions. We state „pseudo steady state conditions‟ because after the conversion and selectivity achieve a maximum, they start both to decline slowly; most probably due to excessive coke build-up. The standard deviations were

below 2% (see example Table S-1 in appendix) for all cases. The reactor exhaust gas was analyzed by gas chromatography using a combination of columns (0.3m Hayesep Q 80-100 mesh with back-flush, 25m × 0.53mm Porabond Q, 15 m ×

0.53mm molsieve 5A and RTX-1 with 30m × 0.53mm), and TCD and FID detectors. This configuration allows the quantification of permanent gases such as CO2, H2, N2, O2, CO as well as hydrocarbons (typically, methane, ethane, ethene, benzene,

toluene, ethylbenzene, styrene, and heavy aromatics). The catalytic tests were carried out under practical conditions of 20 % excess

O2 with respect to the ODH reaction, a concentrated EB feed of 10% and constant

bed volume. Unless stated otherwise, for all EB conversion data the oxygen conversion is complete. The oxygen is used to generate coke, to convert ethylbenzene mainly into styrene, and to convert deposited coke into COx. Since the

selectivities to side products are small and almost constant as a function of the reaction conditions, the selectivity to styrene (ST) is in principle directly coupled to COx. These two combined selectivities (COx and ST) are on average responsible for

96% of the converted ethylbenzene. All physical characterizations for the spent catalysts and derived coke were done after the complete testing cycle of 60 h. Due to the O2 gradient in the reactor, the coke will never be a steady-state material.

Because of that, we took the complete sample, mix it and used it for characterization as an average sample of the reactions conditions applied, location in the bed, and

time on stream.

2.4 Results and discussion

2.4.1 Pseudo steady-state ODH performance

Under pseudo steady-state conditions, typically after 5 h on-stream, the bare -Al2O3

outperforms other commercial supports such as CeO2, TiO2, and SiO2 (data not

shown). The bare alumina (SBET = 272 m2/g and 0.389 ml/g pore volume) is an optimized industrial catalyst support that is stabilized with SiO2; it has an extraordinarily high surface area and pore volume. The thermal activation of the

bare -alumina was carried out to expand the pore size by coarsening the

nanoparticles and investigate its effect on the reaction performance. Remarkably, both conversion of EB and selectivity to ST steadily increase with pretreatment temperature (Figure 1), showing an optimal temperature at 1050 oC (Alu1050)

above which both quantities dropped significantly.

Chapter 2

36

Figure 1. Ethylbenzene conversion and styrene selectivity at 475oC, 1g EB/h and O2:EB=0.6 with a total flow of 36 mL STP/min measured at 105 Pa with 0.8 ml catalyst after 10 h time on stream; GHSV of 3000 h-1.

In all cases the oxygen was completely converted. The maximum ST yield of

35% was found for the Alu1050, which corresponds to ca. 22% relative increase to

the Alu500 at the given GHSV of 3000 l/l/h. Figure 2-a represents the ST yield versus time on stream at various temperatures and EB:O2 ratios for Alu500 and Alu1050. The positive effect of the calcination was observed at 450 and 475 oC

reaction temperature at which full O2 conversion was observed. At 425 oC the O2 conversion was not complete for the Alu1050 and the catalyst performance trend was reversed (Alu500>Alu1050). It should be noted that the O2 conversion was not

always complete for some catalysts at 425 oC; this means that the O2 concentration profiles decrease variably over the reactor bed for the Alu series (i.e. smoother or sharper). Hence, the „active catalyst reactor volume‟ where oxygen is present is

variable as a function of reactor length. This makes the comparison of the intrinsic activity complex at 425 oC. Therefore, the discussion in this paper refers to reaction temperatures ≥450 oC, where full oxygen conversion was achieved. At these

temperatures, the complete catalyst bed was taken as reference for the interpretation.

Figure 2-a also evidences that the catalysts are deactivating as a function of

time on stream, most probably due to ongoing coke deposition on the coked catalysts. Figure 2-b represents the time dependency at 450 oC for Alu500 (O2:EB=0.6); it shows the activity and selectivity increase during the initial coke

formation before achieving a pseudo stationary state. This is in agreement with Nederlof et al. [17] for the CO2-assisted EB ODH, it is attributed to the formation and deposition of active coke on the alumina surface. From a practical standpoint it is

beneficial starting the catalyst performance tests at the highest temperature, i.e. 475 oC, because the coke build-up is faster than that at 425-450 oC; in this way the pseudo-steady state is achieved at shorter time. This can be seen by comparing the

initial performance (<5 h) for Alu500 in Figure 2-a (475 oC) and Figure 2-b (450 oC): at 450 oC the pseudo stationary conditions are achieved after 6 h, while this is

reduced to 3 h at 475 oC. From Figure 2-a, increasing the O2:EB ratio may look favorable in terms of the

overall ST yield but this will have a penalty in the selectivity to ST; a high selectivity

is one of the major targets in the O2 EB ODH to compete with the existing steam-based process. It must be emphasized that the catalytic performance in Figure 1

400 600 800 1000 1200

10

20

30

40

50

60

EB

con

vers

ion

/ %

Pretreatment temperature / oC

50

60

70

80

90

100

ST

sele

ctivity / %

SELECTIVITY CONVERSION


37

corresponds to pseudo-steady state conditions between 5-10 h time on stream with significant coke build up. Assuming the density of graphene (0.76 mg/m2) the number of coke layers is nearly 2.4 for Alu500 and it increases up to 3.4 for

Alu1050. Therefore, we assume this increased performance corresponds to the coke deposits; no interaction between the reactants with the alumina surface or the interface alumina-coke is expected.

Figure 2. a) Time on stream performance at various temperatures of the optimal Alu1050 and

Alu500 catalysts at various temperatures (475, 450, 425, and 450 oC) and O2:EB= 0.6, 0.4 and 0.2 (vol). GHSV of 3000 l/l/h; b) time on stream dependency at 450 oC and O2:EB= 0.6 (Alu500).

We first attempt to correlate this promoting effect to the total amount of carbon deposited, as it is considered the dominating active phase once the deposits are formed under pseudo-steady state conditions. The coke contents are displayed in

Figure 3 as gravimetric quantity (i.e. g coke/g fresh alumina) as well as the volumetric coke (g coke/ml reactor bed); the latter can be more appropriately related to the product yield as it represents the actual amount of coke per reactor

volume, taking into account that the performance is measured at fixed catalyst volume. From this representation, it can be unambiguously seen that the amount of volumetric coke progressively decreases with calcination temperature with a

pronounced reduction above 1050 oC. Considering the increased ST product yield, this suggests that the coke becomes apparently more active (different species) or more selective for the ODH reaction. Calculation of the coke per surface area

indicates that the coke on the low-temperature aluminas is more dispersed (160 mol C/m2 for Alu500) than the optimal Alu1050 (205 mol C/m2).

0 10 20 30 40 50 60 700

5

10

15

20

25

30

35

40

45

50 Alu1050

Alu500

Yie

ld / %

Time / h

425 450 450

0.6 0.4 0.6 0.6 0.4 0.6 0.4

475 oC

O2/EB

0 25 50 75 1000

10

20

30

40

ST

se

lectivity

/ %

EB

co

nve

rsio

n

/ %

Time / h

70

80

90

100

a)

b)

Chapter 2

38

Figure 3. Coke on catalyst (gravimetric and volumetric) determined by TGA as a function of the pre-treatment temperature of the bare alumina after 60 hours, time on stream.

Regarding the selectivity effect, less of the available oxygen is used for the

combustion-gasification of the deposited carbon; for Alu500 around 35% of the

converted oxygen is used for ODH reaction, whereas this increases to 50% for the Alu1050 catalyst. The conversion increase cannot be justified by an additional pathway. Because we run at full O2 conversion, part of the bed will be O2 free. We

did not observe changes in the heavy by-products composition indicating the stability of the by-products in the absence of O2 (2% benzene/toluene and 2% oxygenated and heavy aromatics were observed in all the cases). The reaction temperatures

(425-475 oC) are relatively low to justify the direct EB dehydrogenation; the conventional steam dehydrogenation is operated above 600 oC. If direct dehydrogenation would occur in the absence of O2, a decrease in the EB conversion

is expected due to the low equilibrium conversion at these temperatures (e.g. 20% at 450 oC, EB partial pressure of 0.1 bar [52]). The key aspect to explain the conversion increase is otherwise the improved selectivity to ST; a lower COx

formation will make more O2 available for the main ODH reaction, and this is responsible for the conversion increase. 2.4.2 Textural properties of the alumina upon thermal treatment

The alumina surface has limited intrinsic activity with a low EB selectivity in the order of approximately 75% (Fig. 2-b), also observed in [15]. It is the coke deposited on

the alumina surface that makes the catalyst more selective and active. However, the initial surface area and pore volume of the fresh alumina (and acidity) are important as these will determine how much and well dispersed the coke is. In addition,

accessibility is highly important in heterogeneous catalysis, as diffusion limitations may completely deteriorate selectivity and reduce activity, especially for the much larger ethylbenzene with respect to oxygen. The texture and accessibility of the fresh

aluminas were evaluated by textural properties as derived from N2 physisorption. The textural parameters are plotted in Figure 4-a as a function of the calcination temperature. The marked reduction of surface area and pore volume is associated to

the thermal coarsening of the alumina nanoparticles because the pore size increases. The BJH pore size distributions (PSD) shift towards larger pores (Figure 4-b); this is consistent with the reduction of the textural parameters. Figure 4-b additionally


Co

ke o

n c

atal

yst

g/g

or

g/m

L)


39

shows a broadening of the pore size distribution that is more evident for the Alu1000 upwards while full collapse occurs at 1150 oC. At 1050 oC, this corresponds to the optimal catalyst performance, the pore size distribution is broad, but still well-

defined with a surface area of 101 m2/g. Analysis of an ultrapure gamma-alumina (without silica stabilization) calcined at 1050 oC shows 11 m2/g (Table S-3) and no PSD (Figure 4-c) due to structural collapse. Therefore, the preservation of the

texture, or lack of pore collapse for the Alu1050 is associated to the promotion effect of silica [53]. The changes in texture were accompanied by a densification of the material with an increase of the skeletal density from 3.035 to 4.067 g/ml (Table S-2

in supplementary information). The changes in texture of two relevant spent catalysts containing the coke

were investigated as well and compared to the fresh counterparts. The isotherm of

the fresh Alu500 is of the type IV with hysteresis H1 [54]; representing solids with cylindrical pore geometry with relatively high pore size uniformity and facile pore connectivity. However, the hysteresis changes into type H2 for the spent catalyst

with a closure point at 0.45 relative pressure. Hysteresis H2 is believed to occur in solids where the pores have narrow necks and wide bodies, also called ink-bottle type pores, or when the porous material has interconnected pores. As the original

material has no interconnectivity around 0.45 relative pressure, the spent Alu500 possesses pore neck restrictions attributed to the excessive coke build-up. This is

consistent with the observed TSE effect at p/po = 0.45 in the desorption branch, supposedly showing a uniform pore at 3.9 nm (inset in Figure 5-top). This is not due to a real pore but a physical phenomenon associated to the nature of the adsorbate;

it is known that pore restrictions below 3.9 nm cannot be detected by N2 because of this effect [55]. Therefore, we can only conclude that the TSE points out to pore restrictions smaller than 3.9 nm; it is uncertain whether these restrictions are

relevant for this reaction in terms of mass-transfer limitations. The isotherm for the spent Alu1050 (Figure 5-bottom) maintains the H1 hysteresis (as for the fresh counterpart) with no detection of the TSE effect in the desorption PSD. This indicates

that the pore expansion by high temperature calcination and limited coke build-up keep the pores accessible during the reaction.

Chapter 2

40

Figure 4. a) Specific surface area and total pore volume as a function of the calcination temperature of the alumina; b) BJH pore size distribution using the adsorption branch of the isotherm; both derived from N2 physisorption at -196 oC; c) Comparison of the PSD between Alu1050 and an ultrapure alumina (Merck) calcined at 1050 oC (C1050).

20 100 10000.0

0.5

1.0

Alu1050

C1050

BJH

Ad

so

rptio

n d

V/d

log

(w)

Po

re V

olu

me

pore size /

100 10000.0

0.5

1.0

1.5

2.0

Alu500

Alu600

Alu700

Alu800

Alu900

Alu1000

Alu1050

Alu1100

Alu1150

Alu1200

BJH

Ad

so

rptio

n d

V/d

log

(w)

Po

re V

olu

me

pore size / Å

a)

b) 400 600 800 1000 12000

50

100

150

200

250

300


SBET (m

2/g

)

0.0

0.2

0.4

0.6

0.8

1.0

VT (c

m3/g

)

c)

Å


41

Figure 5. N2 physisorption isotherms and corresponding desorption BJH PSD (inset) for Alu500 and Alu1050: a) fresh and b) spent catalysts. The indicated pore sizes correspond to the desorption branch.

2.4.3 Acidic properties of the alumina upon thermal treatment

The acidity properties were evaluated by temperature programmed desorption of NH3 in order to understand the reasons why the wide-pore SiO2 promoted -alumina

phase becomes a better selective coke-forming support with the calcination

temperature. The TPD profiles (Figure S-1) are formed by two contributions centered at 250 and 600 oC that do not change in position in the Alu series, representing weak and stronger sites. The quantification of the total acidity per catalyst volume displays

a threshold value of 385 to 420 mol/ml-cat, which drops significantly after

calcination of the alumina at 1100 oC to 245 mol/ml-cat. A commercial ultra-pure -

Al2O3 was thermally treated at 1000 oC and its acidity was quantified by NH3-TPD (Figure S-1 and Table S-2). It clearly shows a less intense desorption profile with a

very low acidity (90 mol/g), while the corresponding Alu1000 has 436 mol/g;

hence the high acidity is attributed to the silica promotion. The acidity of relevant catalysts of the series (Alu500 and Alu1050) was

further assessed by pyridine adsorption, monitored by FTIR; this technique is unique in providing information about the nature and strength of the acid sites. The spectrum of the bare Alu500 (Figure 6-left and a) shows a distinctive adsorption of

pyridine on Lewis Al sites at 1449 cm-1; no Brønsted sites were detected that

Chapter 2

42

typically appears at 1545 cm-1 [46]. This consequently implies that there are no

aluminosilicate domains; silica and alumina are present as separate phases. This is in agreement with results of Daniell et al. [56] that observed independent phases up to 5 wt.% silica in alumina. Therefore, the 1 wt. % silica present in bare alumina is a

textural promoter that avoids thermal coarsening of the alumina crystallites and prevents pore collapse, as discussed in 2.4.2. At room temperature the Lewis sites density of Alu500 is 126 mol/g (Table 1) that is in agreement with reported values

[46,57]. It is remarkable that the density obtained from NH3 TPD is a factor 5 higher than pyridine. This is likely due to the fact that NH3 is less selective and accounts for the adsorption on crystallographic defects. The density of acid sites per surface area

was further quantified using Argon physisorption (Table S-2) instead of N2. The quadrupole moment of the nitrogen molecule on highly hydroxylated surfaces causes an orientation effect of the adsorbed nitrogen molecules [58]; this can be solved

using Argon physisoprtion at 87 K. The acid site surface density results in 0.33 sites/nm2 for the Alu500 (Table 1). The strength of the Lewis sites was evaluated after desorbing pyridine at increasingly higher temperatures up to 400 oC (Figure 6-

left, spectra b-g). Quantification of the acid sites density at each evacuation temperature indicates a drop from 126 to 33.9 mol/g at 400 oC (Table 1).

Alu1050 shows Lewis sites at 1449 cm-1 (Figure 6-right and a) with a density

of 138.9 mol/g (Table 1). Brønsted sites were not detected, thus the thermal

treatment at 1050 oC does not produce the mixing between silica and alumina and remain as separate phases. The acid site surface density using Ar physisorption is of 1.39 sites/nm2, that is relatively higher than that of the Alu500 (0.33 sites/nm2). The

acid strength of the Lewis sites was distinctively inferior to the Alu500. At 250 oC Alu1050 has 5 times less sites than that of Alu500, and pyridine was fully desorbed

at 350 oC, while Alu500 still has 49.7 mol/g at this temperature of 350 °C.

It is concluded that Alu1050 has a high acid site density per surface area, higher surface area than non-silica promoted alumina calcined at 1050 oC; both effects are ascribed to the silica promotion, while the Lewis sites are relatively

weaker than that of Alu500.


43

Figure 6. IR spectra of adsorbed pyridine after evacuation at various temperatures for Alu500 (left) and Alu1050 (right): a) room temperature; b) 150; c) 200; d) 250; e) 300; f) 350 and g) 400oC.

Table 1. Lewis acid sites densities determined by FTIR pyridine adsorption.

Alu500 Alu1050

Temperature (oC) a mol/g sites/nm2 b mol/g sites/nm2 b

25 126.2 0.33 138.9 1.39

150 140.7 0.37 57.5 0.58

200 119.2 0.31 31.6 0.32

250 85.3 0.23 16.4 0.17

300 66.6 0.18 4.7 0.05

350 49.7 0.13 0 0

400 33.9 0.09 0 0

a. Evacuation temperature; b. Using SBET values derived from Ar physisorption at -186 oC.

2.4.4 Coke nature and reactivity

In this section the nature of the coke will be discussed based on solubility in CH2Cl2,

thermal analysis (TPO), chemical analysis (CHN), and MALDI information; the latter two techniques were applied on carbonaceous residues after a process of dissolution of the alumina.

The importance of the insoluble coke comes from the fact that the fraction of coke that is less active in styrene production, burns more easily under oxidative

1650 1600 1550 1500 1450 1400

Absorb

ance

Wavenumbers / cm-1

1650 1600 1550 1500 1450 1400

Absorb

ance

Wavenumbers / cm-1

a

b

c

d

e

f

g

a

b

c

d

e

f

pyH+ py on Al3+ 0.2 1545 cm-1 1456 cm-1

0.2 pyH+ py on Al3+ 1545 cm-1 1456 cm-1

Chapter 2

44

conditions, and its nature corresponds to soluble coke; therefore insoluble coke (IC)

determines the product yield (conversion times selectivity) [9,12,15]. The proportion of insoluble coke after refluxing the spent catalysts in CH2Cl2 was quantified (Table 2). The differences among the aluminas are undetectable, always IC > 97%, and

cannot explain the differences in selectivity discussed in Figure 1. The chemical composition of the insoluble carbon deposits was further characterized after dissolving in HF the alumina from the spent catalysts. It was found that the high

temperature aluminas were substantially inert and hard to dissolve completely, showing an ash residue of ca. 8-9 wt. %. Because of that, the elemental composition was tentatively corrected using the TGA residue. The chemical composition of

representative samples of the low- and high-temperature treatments is given in Table 2 and plotted in a Van Krevelen-type plot (Figure 7-a) together with several

reference catalysts. The C:H molar ratio stays in the typical range (1.8-2.5) reported for other untreated -Al2O3 and modified counterparts [7,15], that is typical of

polycondensed/aromatic coke. The high temperature aluminas produced a priori a substantial amount of oxygen-rich coke with O:C molar ratios of 0.20 (0.19 duplo)

and 0.29 (0.27 duplo). In particular, the Alu1050 has a higher concentration of oxygenates and less hydrogen, C:H=2.21; indicating less condensation apparently.

Table 2. Chemical characteristics of the coke for representative low and high temperature spent catalysts after HF treatment.

Sample IC (%)

a

Residue

(wt. %) b

Compositional

formula (at.) C:H (at.) O:C (at.) PD (Å)c

Alu500 99 1.0 C5.52H2.02O 2.73 0.18 86

Alu600 98 1.4 C6.72H2.39O 2.81 0.15 93 Alu1000 97 8.6 C5.20H1.81O 2.76 0.20 141/165d Alu1050 >99 8.4 C3.47H1.57O 2.21 0.29 230

a) Insoluble coke after refluxing in CH2Cl2.

b) Alumina residue due to incomplete dissolution in 30 wt. % HF. c) Pore diameter (PD) for the corresponding alumina at the pore size distribution maxima

(Figure 4). d) Cylindrical geometric value calculated as 4(VT)/SBET as having a substantially more

asymmetric pore size distribution.


45

Figure 7. a) Modified Van Krevelen plots representing the composition of the carbon deposits of various Alu samples (Table 2) as well as data adapted from prior work [,7] and [,15]; b)

oxidation rate in air (TPO) using the TGA derivative patterns of the spent Alu catalysts (>60 h time on stream).

It is remarked that the variability in the ash content within a sample for the

high temperature alumina introduces uncertainty in the representativeness of the samples used in the TGA and CHN analyses. As the oxygen content is very sensitive to the ash residue, we believe that the relatively high oxygen content for Alu1050

derived coke can be due to an overestimation. This is supported by the coke oxidation patterns (discussed later in Figure 7-b); no promotion to lower temperature was observed that would be expected for such an oxygen-rich coke.

The insoluble coke was analyzed by MALDI and LDI time-of-flight mass spectrometry. These techniques use a mild ionization and reflect directly the molecular distribution of the carbon deposits as the energy input of the laser is low.

LDI is particularly selective to oxidized products and was chosen for these ODH carbon deposits, as the matrix addition made the spectra interpretation more complex. The spectrum of the Alu500 derived insoluble coke shows a wide molecular

distribution of masses centered on ca. 1500 Da (Figure 8-a). The spectrum can be assigned to large platelets of polyaromatic hydrocarbons up to 4500 Da, as it shows long series with 24 Da of mass increase revealed upon high resolution magnification

(insets in Figure 8). This is due to the addition of –CH=CH– entity to build a new aromatic ring, with the abstraction of two hydrogen atoms from the main aromatic-polycondensed structure. The interpretation of this spectrum, is in agreement with

the Iwasawa-Ogasawara quinone-based active site model postulated for the EB ODH reaction: condensed aromatic rings with conjugated double bonds containing quinone/hydroxyquinone C=O functional groups [5,59-61]. However, the O groups

must be embedded in the lighter units (<500 m/z) as no mass increase involving O was observed. The spectrum of the Alu1050 (Figure 8-c) shows also a 24 Da C2H2 increase, in agreement with the enlargement of the condensed rings. The emission

intensity of a LDI spectrum gives a proof of the coke nature; the intensity is a measure of the coke reactivity (i.e. oxidation ability) as the emission is directly coupled to ionization. The comparison of the emission intensities in Figure 8

evidences the low reactivity of the Alu1000 and Alu1050 derived cokes as compared to the Alu500, in a factor of ca. 2.5 to 10.

a) b)

Chapter 2

46

Figure 8. LDI time-of-flight spectra of the thermally treated alumina: a) 500 (300 shoots, top); b) 1000 oC (400 shoots, middle) and c) 1050 oC (400 shoots, bottom).

The thermal stability of the carbon deposits is related to the mechanism of

total oxidation in EB ODH and can explain the selectivity trends; COx in EB ODH results from the combustion of the carbon deposits rather than from the direct combustion of ethylbenzene [15]. Looking at the oxidation rate patterns of the spent

catalysts (Figure 7-b), a trend exists with a shift of ca. 20 K towards higher temperatures in the maxima with increasing the calcination temperature of the alumina. This enhanced oxidation stability can explain the lower formation of COx

and, consequently the increase of the ST selectivity. The TPO shift in the carbon oxidation rate is consistent with the lower ionization intensity of the MALDI pattern for the optimal Alu1050. Hence, the increased oxidation stability indicates the lower

reactivity of the Alu1050-derived coke and can explain the selectivity increase to ST. The increased oxidation stability of the coke can be ultimately attributed to the acidity and textural changes upon thermal treatment. Alu1050 has higher density of

Lewis acid sites per surface area, it does not collapse texturally keeping 101 m2/g (compared to a non-silica promoted alumina calcined at 1050 oC with 11 m2/g); both effects are ascribed to the silica promotion. Si-Al Brønsted acid sites are not

involved, while the acid strength of the Lewis sites is relatively weaker (the “good” Lewis acidity) than the low-temperature Alu500. Other phenomenon that changes the surface chemistry of the alumina upon thermal treatment cannot be ruled out to

play a role too; for instance, the surface segregation of impurities upon thermal treatment [62].

2.5 Conclusions

Modification of a SiO2-stabilized -alumina by thermal treatment into a transitional

phase having wider pores and a high density of Lewis sites per surface area enhance the conversion and selectivity for the oxidative dehydrogenation of ethylbenzene to

styrene; the effect is observed when the O2 conversion is complete. The EB conversion increase comes from the ST selectivity enhancement that makes more O2 available for the main ODH reaction. The ST selectivity improvement is attributed to

the coke nature that is less reactive and consequently less prone to COx formation. The site requirements of the optimal catalyst to create the more selective coke seem to be related to the higher density of Lewis sites per surface area, no mixed Si-Al

1000 2000 3000 4000 50000

150

300

450

600

750

900

1050

1200

1350

1500

Absolu

te e

mis

sio

n in

tensity

mass (m/z)

cb

c

a

a


47

Brønsted sites are involved, while the acid strength of the Lewis sites is relatively weaker.

2.6 References

[1] Meyers RA. Handbook of petrochemicals production processes. New York NY:

McGraw-Hill. 2005: 11.3-11.34.

[2] Cavani F, Trifiro F. Alternative processes for the production of styrene. Appl

Catal A-Gen 1995;133(2):219-39.

[3] Lisovskii AE, Alkhazov TG, Dadasheva AM, Feizullaeva SA. Oxidative

dehydrogenation of alkylaromatic hydrocarbons on alumina catalysts. 4. Effect

of alkalis on acid properties of alumina and its catalytic activity in oxidative

dehydrogenation of ethyl benzene, Kinet Catal 1975;16(2):385-9.

[4] Kozharov AI, Makhlis LA, Lisovskii AE, Alkhazov TG, Vasserberg BE.

Mechanism of oxidational dehydrogenation of ethylbenzene on an aluminum

oxide catalyst. Rus Chem Bul 1977;26(3):477-80.

[5] Iwasawa Y, Nobe H, Ogasawara S. Reaction mechanism for styrene synthesis

over polynaphthoquinone. J Catal 1973;31(3):444-9.

[6] Alkhazov TG, Lisovskii AE. Role of condensation products in oxidative

dehydrogenation process of ethylbenzene on aluminium-oxide catalyst. Kinet

Catal 1976;17(2):375-9.

[7] Fiedorow R, Przystajko W, Sopa M, Dalla Lana IG. The nature and catalytic

influence of coke formed on alumina: oxydative dehydrogenation of

ethylbenzene. J Catal 1981;68(1):33-41.

[8] Schraut A, Emig G, Sockel HG. Composition and structure of active coke in the

oxydehydrogenation of ethylbenzene. Appl Catal 1987;29(2):311-26.

[9] Cadus LE, Arrua LA, Gorriz OF, Rivarola JB. Action of activated coke as a

catalyst – oxidative dehydrogenation of ethylbenzene to styrene. Eng Chem

Res 1988;27(12):2241-6.

[10] Schraut A, Emig G, Hofmann H. Kinetic investigations of the

oxydehydrogenation of ethylbenzene. J Catal 1988;112(1):221-8.

[11] Vrieland GE. Oxydehydrogenation of ethylbenzene to styrene over metal

pyrophosphates. 1. Catalyst composition and reaction variables. J Catal

1988;111(1):1-13.

[12] Vrieland GE. Oxydehydrogenation of ethylbenzene to styrene over metal

pyrophosphates.2.Microbalance studies of carbon deposition and burnoff. J

Catal 1988;111(1)14-22.

[13] Cadus LE, Gorriz OF, Rivarola JB. Nature of active coke in the

oxydehydrogenation of ethylbenzene to styrene. Ind Eng Chem Res

1990;29(7):1143-6.

[14] Vrieland GE, Menon PG. Nature of the catalytically active carbonaceous sites

for the oxydehydrogenation of ethylbenzene to styrene – A brief review. Appl

Catal 1991;77(1):1-8.

[15] Lisovskii AE, Aharoni C. Carbonaceous deposits as catalysts for

oxydehydrogenation of alkylbenzenes. Catal. Rev. Sci. Eng. 1994;36(1):25-

74, and references therein.

[16] Grunewald GC, Drago RS. Oxidative dehydrogenation of ethylbenzene to

styrene over carbon-based catalysts. J Mol Catal 1990;58(2):227-33.

Chapter 2

48

[17] Nederlof C, Kapteijn F, Makkee M. Catalysed ethylbenzene dehydrogenation in

CO2 or N2 - Carbon deposits as the active phase. Appl Catal A-Gen

2012;417:163-73.

[18] Allen RH, Alfrey T, Yats LD. Oxidative dehydrogenation of alkylbenzenes. US

Patent 3497564, 1967.

[19] Drago RS, Jurczyk K. Oxidative dehydrogenation of ethylbenzene to styrene

over carbonaceous catalysts. Appl Catal A-Gen 1994;112(2):117-24.

[20] Klimkiewicz R, Morawski AW, Mista W. Bi-K-graphite intercalation compounds

as a new catalyst for styrene synthesis. J Catal 1993;144(2):627-31.

[21] Guerrero-Ruiz A, Rodriguez-Reinoso F. Oxydehydrogenation of ethylbenzene

to styrene catalyzed by graphites and activated carbons. Carbon

1994;32(1):23-9.

[22] Sugino M, Shimada H, Turuda T, Miura H, Ikenaga N, Suzuki T. Oxidative

dehydrogenation of ethylbenzene with carbon-dioxide. Appl Catal A-Gen

1995;121(1):125-137.

[23] Badstube T, Papp H, Kustrowski P, Dziembaj R. Oxidative dehydrogenation of

ethylbenzene with carbon dioxide on alkali-promoted Fe active carbon

catalysts. Catal Lett 1998;55(3-4):169-72.

[24] Pereira MFR, Orfão JJM, Figueiredo JL. Oxidative dehydrogenation of

ethylbenzene on activated carbon catalysts I. Influence of surface chemical

groups. Appl Catal A-Gen 1999;184(1):153-60.

[25] Sakurai Y, Suzaki T, Ikenaga N, Suzuki T. Dehydrogenation of ethylbenzene

with an activated carbon-supported vanadium catalyst. Appl Catal A-Gen

2000;192(2):281-8.

[26] Ikenaga N, Tsuruda T, Senma K, Yamaguchi T, Sakurai Y, Suzuki T.

Dehydrogenation of ethylbenzene with carbon dioxide using activated carbon-

supported catalysts. Ind Eng Chem Res 2000;39(5):1228-34.

[27] Mestl G, Maksimova NI, Keller N, Roddatis VV, Schlögl R. Carbon

nanofilaments in heterogeneous catalysis: An industrial application for new

carbon materials? Angew Chem Int Ed 2001;40(11):2066-8.

[28] Keller N, Maksimova NI, Roddatis VV, Schur M, Mestl G, Butenko YV,

Kuznetsov VL, Schlögl R. The catalytic use of onion-like carbon materials for

styrene synthesis by oxidative dehydrogenation of ethylbenzene. Angew Chem

Int Ed 2002;41(11):1885-8.

[29] Pereira MFR, Orfão JJM, Figueiredo JL. Oxidative dehydrogenation of

ethylbenzene on activated carbon fibers. Carbon 2002;40(13):2393-401.

[30] Pereira MFR, Figueiredo JL, Orfão JJM, Serp P, Kalck P, Kihn Y. Catalytic

activity of carbon nanotubes in the oxidative dehydrogenation of

ethylbenzene. Carbon 2004;42(14):2807-13.

[31] Su DS, Maksimova N, Delgado JJ, Keller N, Mestl G, Ledoux MJ, et al.

Nanocarbons in selective oxidative dehydrogenation reaction. Catal Today

2005;102-103:110-4.

[32] Delgado JJ, Vieira R, Rebmann G, Su DS, Keller N, Ledoux MJ, et al.

Supported carbon nanofibers for the fixed-bed synthesis of styrene. Carbon

2006;44(4):809-12.


49

[33] Delgado JJ, Su DS, Rebmann G, Keller N, Gajović A, Schlögl R. Immobilized

carbon nanofibers as industrial catalyst for ODH reactions. J Catal

2006;244(1):126-9.

[34] Zhao TJ, Sun WZ, Gu XY, Rønning M, Chen D, Dai YC, et al. Rational design of

the carbon nanofiber catalysts for oxidative dehydrogenation of ethylbenzene.

Appl Catal A-Gen 2007;323:135-46.

[35] Zhang J, Su D, Zhang A, Wang D, Schlögl R, Hebert C. Nanocarbon as robust

catalyst: Mechanistic insight into carbon-mediated catalysis. Angew Chem Int

Ed 2007;46:7319-23.

[36] Su D, Maksimova NI, Mestl G, Kuznetsov VL, Keller V, Schlögl R, et al.

Oxidative dehydrogenation of ethylbenzene to styrene over ultra-dispersed

diamond and onion-like carbon. Carbon 2007;45(11):2145-51.

[37] Rinaldi A, Zhang J, Mizera J, Girgsdies F, Wang N, Hamid SBA, et al. Facile

synthesis of carbon nanotube/natural bentonite composites as a stable

catalyst for styrene synthesis. Chem Commun 2008;48:6528-30.

[38] Su DS, Delgado JJ, Liu Xi, Wang D, Schlögl R, Wang L, Zhang Z, Shan Z, Xiao

FS. Highly ordered mesoporous carbon as catalyst for oxidative

dehydrogenation of ethylbenzene to styrene. Chem Asian J 2009;4(7):1108-

13.

[39] Delgado JJ, Chen X, Tessonnier JP, Schuster ME, Del Rio E, Schlögl R, et al.

Influence of the microstructure of carbon nanotubes on the oxidative

dehydrogenation of ethylbenzene to styrene. Catal Today 2010;150(1-2):49-

54.

[40] Rinaldi A, Zhang J, Frank B, Su DS, Hamid SBA, Schlögl R. Oxidative

purification of carbon nanotubes and its impact on catalytic performance in

oxidative dehydrogenation reactions. ChemSusChem 2010;3(2):254-60.

[41] Zhang J, Su DS, Blume R, Schlögl R, Wang R, Yang X, et al. Surface chemistry

and catalytic reactivity of a nanodiamond in the steam-free dehydrogenation

of ethylbenzene. Angew Chem Int Ed 2010;49(46):8640-4.

[42] Gasanova NI, Lisovskii AE, Alkhazov TG. Oxidative dehydrogenation of

ethylbenzene on aluminoboron catalysts. Kinet Catal 1979;20(4): 748-52.

[43] Izumi I, Shiba T. Characterization of the alumina-boria catalyst. Bull Chem

Soc Japan 1964;37(12):1797-809.

[44] Zarubina V, Nederlof C, Heeres HJ, Kapteijn F, Makkee M, Melián-Cabrera I.

Proceedings of the 15th International Congress on Catalysis. Munich:2012.

[45] Leofantia G, Padovanb M, Tozzolac G, Venturellic B. Surface area and pore

texture of catalysts. Catal Today 1998;41(1-3):207-19.

[46] Tamura M, Shimizu KI, Satsuma A. Comprehensive IR study on acid/base

properties of metal oxides. Appl Catal A-Gen 2012;433-434:135-45.

[47] Guisnet M, Magnoux P. Organic chemistry of coke formation. Appl Catal A-Gen

2001;212(1-2):83–96.

[48] Nielen MWF. MALDI time-of-flight mass spectrometry of synthetic polymers.

Mass Spectrom Rev 1999;18:309-44.

[49] Hanton SD. Mass spectrometry of polymers and polymer surfacesChem Rev.

2001;101:527-69.

Chapter 2

50

[50] Johnson MP, Donnet JB, Wang TK, Wang CC, Locke RW, Brinson BE, Marriott

T. A dynamic continuum of nanostructured carbons in the combustion furnace.

Carbon 2002;40(2):189-94.

[51] Hillenkamp F, Peter-Katalinic J. MALDI MS: A Practical Guide to

Instrumentation, Methods and Applications. Weinheim: Wiley VCH. 2007.

[52] Moulijn JA, Makkee M, van Diepen AE. Chemical Process Technology, 2nd ed.

Chichester: Wiley. 2013: 336-8.

[53] Béguin B, Garbowski E, Primet M. Stabilization of alumina toward thermal

sintering by silicon addition. J Catal 1991;127(2):595.

[54] Kruk M, Jaroniec M. Gas adsorption characterization of ordered organic-

inorganic nanocomposite materials. Chem Mater 2001;13(10):3169-83.

[55] Groen JC, Pérez-Ram rez J. Critical appraisal of mesopore characterization by

adsorption analysis. Appl Catal A-Gen 2004;268(1-2):121-5.

[56] Daniell W, Schubert U, Glöckler R, Meyer A, Noweckb K, Knözinger H.

Enhanced surface acidity in mixed alumina-silicas: a low-temparature FTIR

study. Appl Catal A-Gen 2000;196(2):247-60.

[57] Carre S, Tapin B, Gnep NS, Revel R, Magnoux P. Model reactions as probe of

the acid-base properties of aluminas: nature and strength of active sites.

Correlation with physicochemical characterization. Appl Catal A-Gen

2010;372(1):26-33.

[58] Thommes M. Physical adsorption characterization of nanoporous materials.

Chem Ing Tech 2010;82(7):1059-73.

[59] Iwasawa Y, Ogasawara S. Control of selectivity and increase of catalytic

activity of polynaphthoquinone by various Lewis acids. J Catal

1975;37(1):148-57.

[60] Iwasawa Y, Fujitsu H, Onishi T, Tamaru K. Various reactions catalyzed by

electron donor-acceptor complex of polynaphthoquinone with potassium. J.

Chem. Soc. Faraday T 1 1974;70(2):202-7.

[61] Iwasawa Y, Mori H, Ogasawara S. Catalytic hydrogen transfer reactions

between hydroaromatics and nitrobenzene over polynaphthoquinone. J. Catal

1980;61(2):366-73.

[62] Yang BL, Lee SB, Cheng DS, Chang WS. Surface enrichment of impurities and

its effect on catalytic properties of iron oxide. J Catal 1993;141:161-70.


51

Chapter 2

52

Figure S-1. NH3 temperature programmed desorption (TPD) profiles: a) Alu series and b) comparison with a commercial Alumina (Merck 1.01095.1000) calcined at 1000 oC.

Figure S-2. Nitrogen sorption isotherms at -196 oC K for the fresh Alu series.

200 300 400 500 600 700 800 900 1000

0.000

0.004

0.008

0.012

Alu bare

Alu900

Alu1000

Alu1050

Alu1100

Alu1200

TC

D S

ign

al (a

.u.)

Temperature, °C

200 300 400 500 600 700 800 900 1000

0.000

0.004

0.008

0.012 Alu1000

gamma alumina Merk (1000)

TC

D S

ign

al (a

.u.)

Temperature, °C

a)

b)

0,0 0,2 0,4 0,6 0,8 1,0

0

100

200

300

400

500 Alu500

Alu600

Alu700

Alu800

Alu900

Alu1000

Alu1050

Alu1100

Alu1150

Alu1200

Qu

an

tity

Adsorb

ed (

cm

³/g

ST

P)

Relative Pressure (p/p°)


53

Table S-1. Example of the accuracy of the conversion and selectivity quantities

Alu500

Data points Conversion (%) Selectivity (%)

1 35.71 81.36

2 36.18 82.71

3 35.62 84.13

4 36.08 82.64

5 34.99 82.94

6 34.98 82.26

7 34.54 82.79

Average 35.44 82.69

(%) 1.6 0.9

Chapter 2

54

Tab

le S

-2.

Str

uctu

ral, t

extu

ral and a

cid

ic p

ropert

ies o

f th

e fre

sh t

herm

ally t

reate

d a

lum

inas.a

Sam

ple

P

hase

VT

(cm

3/

g)

VT’

(cm

3/

ml

bed

)b

SB

ET

(m

2/g

cat.

)

S’ B

ET

b

(m

2/

ml

bed

)

D B

JH

ad

s

(Å

)

Den

sit

yc

(g

/cm

3)

Acid

ity

(m

ol/

g)

Acid

ity

b

(m

ol/

ml

bed

)

Bare

e

0.6

39

(0.6

28)

0.3

89

(0.3

82)

272

(227)

165

(138)

84

3.0

35

637

388

Alu

500 e

0.6

49

(0.6

61)

0.4

31

(0.4

39)

271

(228)

180

(152)

86

3.1

08

Alu

600

0.6

44

0.4

46

255

177

93

3.0

84

Alu

700

0.6

35

0.4

02

239

151

101

3.0

03

Alu

800

0.6

36

0.4

36

214

147

118

3.0

99

Alu

900 e

0.6

08

(0.5

93)

0.4

73

(0.4

61)

179

(138)

139

(108)

138

3.2

95

540

420

Alu

1000

0.4

92

0.4

35

119

105

141

3.3

16

436

385

Alu

1050 e

0.4

58

(0.3

82)

0.3

27

(0.2

73)

101 (

60)

72 (

43)

230

3.3

76

398

284

Alu

1100

0.3

54

0.3

56

54

54

294

3.6

58

244

245

Alu

1150

0.1

65

0.1

57

20

19

330 d

4.0

09

Alu

1200

0.1

17

0.1

43

16

20

293 d

4.0

67

20

25

C1050 f

0.0

51

11

185 d

a)

N2 (

-196 o

C)

isoth

erm

s a

re g

iven in F

igure

S-2

and N

H3-T

PD

in F

ig.

S-1

. b)

Quantity

per

reacto

r volu

me.

c)

Skele

tal density.

d)

Geom

etr

ical pore

siz

e a

s t

here

is n

o m

axim

um

in t

he B

JH p

ore

siz

e d

istr

ibution.

e)

Valu

es b

etw

een b

rackets

are

deri

ved fro

m A

rgon p

hysis

orp

tion a

t -1

86 o

C.

f)

Ultra

pure

alu

min

a (

com

merc

ial:

Merc

k 1

.01095.1

000)

therm

ally t

reate

d a

t 1050 o

C.

university of groningen oxidative dehydrogenation of

Documents