Mild temperature palladium-catalyzed ammoxidation of ethanol to acetonitrile

Applied Catalysis A: General 506 (2015) 261–267

Contents lists available at ScienceDirect

Applied Catalysis A: General journal homepage: www.elsevier.com/locate/apcata

Mild temperature palladium-catalyzed ammoxidation of ethanol to acetonitrile Conor Hamill a , Hafedh Driss b , Alex Goguet a , Robbie Burch a , Lachezar Petrov b , Muhammad Daous b,∗ , David Rooney a,∗∗ a b

School of Chemistry and Chemical Engineering, Queen’s University Belfast, Belfast BT8 5AG, UK Chemical and Materials Engineering Department, King Abdulaziz University, Jeddah, P.O. Box 80035 Saudi Arabia

a r t i c l e

i n f o

Article history: Received 21 August 2015 Received in revised form 18 September 2015 Accepted 21 September 2015 Available online 25 September 2015 Keywords: Palladium Bio-based chemicals Renewable solvents Product security Sustainable chemistry

a b s t r a c t The ammoxidation of ethanol is investigated as a renewable process for the production of acetonitrile from a bio-feedstock. Palladium catalysts are shown to be active and very selective (>99%) to this reaction at moderate to low temperatures (150–240 ◦ C), with acetonitrile yields considered a function of Pd morphology. Further investigations reveal that the stability of these catalysts is influenced by an unselective product, and that any deactivation observed is reversible. Interpretation of this deactivation allows operating conditions to be defined for the stable, high yielding production of acetonitrile from ethanol. © 2015 Published by Elsevier B.V.

1. Introduction The development of sustainable industry requires a fundamental change in societal dependence on petroleum resources. The production of chemicals from ‘carbon neutral’ biomass offers improved supply securities and could significantly reduce industry’s environmental impact [1]. Currently 95% of all organic chemicals are produced from fossil resources [2]. U.S. legislation is targeting a 25% replacement with biomass derived chemicals by 2025 [3], thus reinforcing the need for the development of sustainable processes to aid the realization of a bio-based chemical industry. Acetonitrile is an important fine chemical which is widely used as a solvent in the agrochemical, pharmaceutical and petrochemical sectors. The pharmaceutical industry in particular accounts for over 70% of the total global demand [4]. Presently, acetonitrile is a co-product in the manufacture of acrylonitrile (SOHIO Process), whereby 40–60 kg of acetonitrile are produced for every tonne of acrylonitrile [5]. Consequently, output is heavily reliant on

∗ Corresponding author. Fax: +966 26952257. ∗∗ Corresponding author. Fax: +44 2890974687. E-mail addresses: [email protected] (M. Daous), [email protected] (D. Rooney). http://dx.doi.org/10.1016/j.apcata.2015.09.030 0926-860X/© 2015 Published by Elsevier B.V.

propylene availability and acrylonitrile demand. Recently, Asahi Kasei announced the closure of a 150,000-ton/year acrylonitrile plant [6], citing increased propylene prices associated with shale gas production [7] and reduced acrylonitrile demands due to the slowdown in the Chinese economy. These markets trends are ominous for future acetonitrile reserves, and could prompt a situation similar to the observed crisis of 2008–2009 [8]. It is therefore imperative to develop a more secure supply of acetonitrile. The ammoxidation of bio-ethanol affords a possible sustainable route for the production of acetonitrile. Bio-ethanol has the potential to become a renewable platform chemical due to its existing high volume production and the continued advancements in feedstock selection [9]. Ethanol ammoxidation was previously reported [10,11], however relatively high temperatures (>350 ◦ C) were needed and the modest activity required high catalyst loading. More recently, Oishi et al. [12] have reported the ammoxidation of aromatic alcohols to nitriles. However, they note in their work, an inability to activate aliphatic alcohols. Nitrile synthesis from such alcohols has been reported previously [13]; however, the array of reagents required could be difficult and expensive to scaleup. The ammoxidation of ethane [8,14] and ethylene [15] have also been explored as alternative acetonitrile production routes. However, these suffer from high temperature demands and low product selectivity, (maximum yields of 26% reported at temperatures greater than 450 ◦ C), and ultimately do not offer a solution to

262

C. Hamill et al. / Applied Catalysis A: General 506 (2015) 261–267

the overarching issue of product security, as they remain heavily dependent on petrochemical resources. The production of acetonitrile from ethanol has also been explored via a reductive amination route [16–18], however, these again requires high temperatures (≈340 ◦ C) and display poor catalyst activity. The oxidative dehydrogenation of ‘bio’ ethylamine was proposed as a possible production route, reporting a selectivity of 80% under continuous conditions [19]. The economics of this proposed pathway are, however, doubtful, since mono-ethylamine is a more expensive chemical than acetonitrile [20]. The amination of glycerol to acetonitrile has also been examined [21], but the selectivity was less than 50%. The application of noble metals for ammoxidation is limited [22,23], with the most widely applied ammoxidation catalysts comprising of supported oxides of vanadium and molybdenum [24]. However, the use of these sites for ethanol ammoxidation [10,25] have had limited success requiring high temperatures (>400 ◦ C) to display modest yields (60–80%). In the ammoxidation reaction partial oxidation species, in particular carbonyls, are widely acknowledged as crucial intermediates in facilitating nitrogen insertion [24,26]. In the present work supported palladium catalysts have been selected to investigate the ammoxidation of ethanol given the metal’s ability to selectively oxidize alcohols to the corresponding aldehyde at low temperatures [27–29]. Herein we report on the moderate to low temperature production of acetonitrile via the ammoxidation of a renewable resource; ethanol, using highly selective palladium based catalysts under conditions yielding stable operation. 2. Experimental 2.1. Catalyst preparation The supported Pd catalysts were typically prepared using a colloidal technique called sol immobilization [30,31]. In a standard preparation, an aqueous solution of PdCl2 (Sigma–Aldrich) was prepared. Polyvinyl alcohol (PVA) solution (1 wt%) was added to obtain a resulting ratio of PVA/Pd (wt/wt) = 1.2. A freshly prepared solution of NaBH4 (0.1 M, NaBH4 /Pd (mol/mol) = 5) was then added. After 30 min of sol generation, the colloid was immobilized by the support under vigorous stirring conditions. Note TiO2 -P25 (Degussa) SiO2 (BDH) and ZrO2 (Alfa-Aesar) supports were acidified to a pH of 1 using sulfuric acid prior to immobilization, ␥-Al2 O3 (Grace) was untreated. After 2 h the slurry was filtered, the solid washed thoroughly with distilled water and dried at 120 ◦ C overnight. The same method was employed when preparing catalysts with different metal loadings. Pd/TiO2 was also synthesized via wet impregnation. A predetermined volume of PdCl2 solution was added to a specific quantity of TiO2 -P25. The mixture was stirred at 50 ◦ C until a paste like material was generated. This was then dried at 120 ◦ C for 4 h and calcined at 500 ◦ C for an additional 4 h. For details concerning characterization of these catalysts please see the supporting information.

detector (FID) and fitted with a 30 m Restek Stabilwax column. A Hiden HPR 20 mass spectrometer was operated in parallel to the GC to monitor the evolution of additional products. All GC data were recorded on a temperature ramp down. Each temperature interval was held for 40 min and an average analysis was recorded. 2.3. Catalyst characterization The XPS measurements were carried out in an ultra-high vacuum multi-technique surface analysis system (SPECS GmbH, Germany) operating at a base pressure of 10−10 m bar. A standard dual anode X-ray source SPECS XR-50 with Al-K␣, 1486.6 eV was used to irradiate the sample surface with 13.5 kV, 100 W X-ray power and a take-off-angle for electrons at 90◦ relative to sample surface plane. The high energy resolution or narrow scan spectra were recorded at room temperature with a 180◦ hemispherical energy analyzer model PHOIBOS-150 and a set of nine channel electron multipliers MCD-9. The analyzer was operated in Fixed Analyzer Transmission (FAT) and medium area lens modes at pass energy of 20 eV, step size of 0.1 eV and dwell time of 2.0 s. As is the standard practice in XPS studies, the adventitious hydrocarbon C1s line (285 eV) corresponding to CC bond has been used as the binding energy reference for charge correction. The TEM analysis of the Pd catalysts was performed using a Tecnai G2 F20 Super Twin at 200 kV with a LaB6 emitter. The microscope was fully equipped for analytical work with an energydispersive X-ray (EDX) detector with a S-UTW window and a high angle annular dark-field (HAADF) detector for STEM imaging. Unless stated otherwise, the scanning transmission electron microscopy (STEM) imaging and all analytical work were performed with a probe size of 1 nm resulting in a beam current of about 0.5 nA. TEM images and selected area diffraction (SAD) patterns were collected on an Eagle 2K HR 200 kV CCD camera. The HAADF-STEM EDX and CCD line traces were collected fully automatically using the Tecnai G2 User Interface and processed with the Tecnai Imaging and Analysis (TIA) software Version 1.9.162. The DRIFT spectra were recorded using a Bruker Equinox 55 spectrometer using an average of 128 scans and a resolution of 4 cm−1 . The DRIFTS setup consisted of an in-situ high temperature diffuse reflectance IR cell (Spectra-Tech) fitted with ZnSe windows which was modified in house to behave as a plug flow reactor, the details of which have been previously reported [32,33]. The samples were pre-reduced for 1 h at 100 ◦ C in a pure H2 stream (40 ml min−1 ). The cell was then cooled to 35 ◦ C. A background spectra was collected after purging with Ar (40 ml min−1 ) for 30 min. A 1%CO/Ar stream (40 ml min−1 ) was subsequently passed through the cell for 30 min, with the resulting spectra recorded every minute. The gas flow was then switched back to Ar (40 ml min−1 ). The cell was evacuated for a further 30 min with spectra recorded every minute. During this step any weakly adsorbed CO species are removed, and only the features corresponding to strongly adsorbed CO molecules remain. 3. Results and discussion

2.2. Catalyst testing Catalyst testing was performed in an isothermal fixed bed reactor (I.D 6 mm) placed in a ceramic tubular furnace controlled by a Eurotherm2604 PID controller. Typically a 50 s cm3 min−1 gas stream comprising of 525 ppm ethanol, 4200 ppm NH3 , 6825 ppm O2 , 1% Kr, balanced with Ar (unless otherwise stated), flowed through a 138 mg catalytic packed bed (particle size: 212–425 ␮m). The ethanol was fed via a calibrated temperature controlled saturator. The concentrations of ethanol and acetonitrile were analyzed using a PerkinElmer Clarus 500 GC, equipped with a flame ionized

Fig.1a reports the ethanol ammoxidation activity obtained with the 0.3–10 wt% Pd/TiO2 (30–1000Pd) catalysts. Complete conversion of the alcohol was achieved from circa 215 ◦ C for the highest Pd loaded catalysts. It was observed that increasing the nominal loading of palladium from 0.3 wt% to 2 wt% (30–200Pd) has a minimal impact on this conversion. Further increases to 5 wt% (500Pd) and 10 wt% (1000Pd) yielded significant improvements in the rate of ethanol conversion. Note however, that there was no difference in the conversion profiles of the 500Pd and 1000Pd catalysts. Fig. 1b reports the acetonitrile selectivity of these materials as a

Ethanol Conversion (%)

C. Hamill et al. / Applied Catalysis A: General 506 (2015) 261–267

100 90 80 70 60 50 40 30 20 10 0

Table 1 XPS and H2 chemisorption analysis of various palladium loaded catalysts supported on TiO2.

a)

100

150

200

250

Sample

Pd0 (%)

Pd2+ (%)

D (%)

dp (nm)

1000Pd 500Pd 200Pd 100Pd 50Pd 30Pd

– – 38.3 32.0 43.5 33.8

– – 61.7 68.0 56.5 66.2

14.7 18.3 20.4 22.6 26.1 28.9

7.63 6.13 5.50 4.97 4.30 3.88

300

Ethanol Conversion (%)

b)

100 90 80 70 60 50 40 30 20 10 0

a)

100

150

200

250

300

250

300

Temperature (oC) 100

150

200

250

300

Temperature (oC) Fig. 1. The effect of palladium loading on ethanol conversion (a) and acetonitrile selectivity (b) as a function of temperature. 1000Pd (×), 500Pd (䊏), 200Pd (䊉), 100(), 50Pd () and 30Pd (+), all supported on TiO2 .

Acetonitrile Selecvity (%)

Acetonitrile Selecvity (%)

Temperature (oC) 100 90 80 70 60 50 40 30 20 10 0

263

100 90 80 70 60 50 40 30 20 10 0

b)

100

150

200

Temperature (oC) Fig. 2. The effect of support type on ethanol conversion (a) and acetonitrile selectivity (b) as a function of temperature. For 1 wt% Pd (nominal) supported on ␥-Al2 O3 (), ZrO2 (䊏), TiO2 (䊉), and SiO2 (). Scheme 1. Proposed reaction pathway for the ammoxidation of ethanol over Pd/TiO2 catalysts.

function of temperature. Below about 190 ◦ C, acetonitrile selectivity increased with temperature for all catalysts; this is attributable to the suppression of ethyl acetate production. This is consistent with the results of Gaspar et al. [34] who reported ethyl acetate production over Pd catalysts via a condensation reaction between ethanol and acetaldehyde (see Scheme 1). At higher temperatures a sharp decline in acetonitrile production was observed due to the dominance of competing combustion reactions, evidenced by CO2 production. Note that the presence of HCN or acetamide was not observed in the gas phase. Fig. 1b reveals that varying the palladium content had a significant influence on the product distribution. In the low temperature region, acetonitrile selectivity increased with palladium loading, with the lower loaded catalysts forming more ethyl acetate. The catalysts with lower Pd concentration conversely displayed improved acetonitrile selectivity at higher temperatures (>240 ◦ C), due to a reduced affinity for combustion reactions evidenced by a decline in CO2 production (see Fig. S.1). From Table 1 it can be observed that the least dispersed catalyst (1000Pd) was the most selective to acetonitrile, and the least selective (30Pd) was the most disperse. The possible dependence on initial oxidation state of the catalysts was postulated to elucidate the observed relationship between selectivity and metal concentration in Fig. 1. Lee et al. [27] report that a decreased metal loading yields an increase in the palla-

dium oxide fraction [35,36]. However, XPS analysis of the palladium catalysts (30–200Pd), revealed that metal loading had a negligible impact on their initial oxidation state (see Table 1) indicating that product distribution is most likely independent of the initial oxidation state of the palladium. To further probe this hypothesis, the least selective catalyst; 30Pd was pre-reduced at 100 ◦ C under H2 for 1 h before its activity was assessed. This pre-treatment had no observable impact on performance. It has been widely reported that the interactions between a metal and its support can govern the effectiveness of a catalyst [37]. To explore the influence of the supporting material on the catalytic performance, palladium was deposited (nominal 1 wt%) onto various oxide supports; TiO2 , ZrO2 , ␥-Al2 O3 and SiO2 . The ensuing conversion of ethanol is reported in Fig. 2 a. It can be observed that ammoxidation activity follows the trend; TiO2 > ␥Al2 O3 > ZrO2 > SiO2 . The selectivity of these catalysts as a function of temperature is reported in Fig. 2b. Note the characteristic selectivity profile described previously was observed for all catalysts. It can be seen that ZrO2 was the most selective support, displaying an acetonitrile selectivity of approximately 94% at 170 ◦ C. Note also that the results demonstrate that Pd/TiO2 was the most productive catalyst as, on average, it gave the highest yield of acetonitrile. The diverse acid/base properties of the supports were investigated to elucidate the observed variations in acetonitrile selectivity. Acid sites are considered to be important for the insertion of nitrogen into partially oxidized intermediates during the ammoxidation

C. Hamill et al. / Applied Catalysis A: General 506 (2015) 261–267

Table 2 ICP, BET and H2 chemisorption analyses of palladium loaded on the various oxide supports. Sample

ICP (%)

BET (m2 /g)

D (%)

dp (nm)

1%Pd/ZrO2 1%Pd/TiO2 1%Pd/␥-Al2 O3 1%Pd/SiO2

0.50 0.68 0.61 0.50

19.0 54.9 195.6 295.3

18.1 22.6 28.9 34.3

6.19 4.97 3.88 3.47

200Pd 100Pd 50Pd 30Pd Support

0.05

Acetonitrile Selecvity (%)

264

1980

20

0

1925

40

60

Bridged / Linear CO Fig. 4. Acetonitrile selectivity at 167 ◦ C as a function of the proportion bridged bonded to linear adsorbed CO determine via DRIFTS. 30Pd/TiO2 ( ), 50Pd/TiO2 ( ), 100Pd/TiO2 ( ), 200Pd/TiO2 ( ), 500Pd/TiO2 ( ), 1000Pd/TiO2 (䊉), 100Pd/Al2 O3 ( ), 100Pd/ZrO2 ( ) 100Pd/SiO2 ( ).(For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

2000

1900

-1

1800

1700

Wavenumber (cm )

Ethanol Conversion (%)

2078

2100

100 90 80 70 60 50 40 30 20 10 0

Fig. 3. DRIFTS spectra of adsorbed CO on various palladium loaded catalysts supported on TiO2 (500Pd & 1000Pd not shown).

100 90 80 70 60 50 40 30 20 10 0

a)

100

150

200

250

300

250

300

reaction [24,38]. Gaspar et al. [39] also cite the importance of acid/base properties in the production of the ethyl acetate. To evaluate the relationship between catalyst acidity and acetonitrile selectivity NH3 -TPDs were conducted to quantify the acid sites. The results of Fig. S.2 indicate that the acid/base characteristics of the support do not play a dominant role in the production of acetonitrile. The least selective catalyst Pd/SiO2 , was less acidic than Pd/TiO2 and more acidic than Pd/ZrO2 , both of which displayed superior selectivities. The metal dispersion of these catalysts was also assessed to probe the possible relationship between nitrile selectivity and morphology. From Table 2 it can be observed that the palladium dispersion increased with support surface area. The least dispersed catalyst, Pd/ZrO2 , was the most selective, and conversely the most disperse catalyst, Pd/SiO2 , was the least selective. These results were correlated with those found for the titania-supported catalysts with different Pd loadings (see Fig. S.3), The result display that acetonitrile selectivity increases with decreasing metal dispersion, which suggests that the product distribution may be a function of Pd morphology. To further evaluate the relationship between dispersion and selectivity, a CO adsorption Diffuse Reflectance Infrared Fourier Transform Spectroscopy study was conducted. In general two types of adsorbed CO are identifiable on palladium. Fig. 3 displays the IR spectra of CO adsorbed onto various loaded Pd/TiO2 catalysts (0.3–2 wt%). Two bands appeared in the region of 1800–2100 cm−1 ; that close to 2078 cm−1 was ascribed to mono-coordinated, linearly bonded CO. The broad band appearing under 2000 cm−1 was assigned to multi-coordinated bridged–bonded CO [40]. Linearly, bonded CO species are typically associated with adsorption onto low coordinated (corner and edges) Pd atoms, whereas bridged species are synonymous with CO adsorption on facetted planes. Smaller particles expose a larger fraction of the former type of sites than larger crystallites [40]. From Fig. 3 it can be observed that increasing the palladium loading led to an increase in the intensity of bands assigned to linear and bridge adsorbed CO. However,

ACN Selecvity (%)

Temperature (oC) 100 90 80 70 60 50 40 30 20 10 0

b)

100

150

200

Temperature (oC) Fig. 5. The effect of preparation method on ethanol conversion (a) and acetonitrile selectivity (b) as a function of temperature. 100Pd via sol immobilisation (䊏) and 100Pd via impregnation (), both supported on TiO2 .

the ratio of bridged to linear (B/L) species increased with palladium loading. This infers that the average size of the palladium crystallites are also increasing, which complements the chemisorption data reported in Table 1. The B/L ratio of the other oxide supported palladium catalysts were also evaluated and compared with the Pd/TiO2 samples. It should be noted that the band positions for CO adsorption onto palladium were independent of supporting material [41]. The acetonitrile selectivity of all the Pd catalysts at 167 ◦ C is plotted as function of their B/L ratio in Fig. 4. From this it would appear that the selectivity increases with an increase in the B/L ratio; the exact reason for this remains unknown. However, these results do suggest that the terraced palladium planes are more selective for nitrile production. The possible structure sensitivity of ethanol ammoxidation was further examined by investigating the catalyst preparation method. The sol immobilization (S.I.) method, used thus far, affords control over the morphology and size distribution of supported nanoparticles [42]. 1 wt% Pd/TiO2 was prepared via the wet impregnation

C. Hamill et al. / Applied Catalysis A: General 506 (2015) 261–267

265

Fig. 6. HRTEM images of (a, b) fresh 100Pd catalyst, (c) selected area diffraction (SAD) of a particle within the 100Pd sample and (d–f) the impregnated sample100Pd-IMP. Both supported on TiO2 -P25.

(100Pd IMP) method and its performance was compared to the S.I. sample (100Pd). Fig. 5 reveals that the 100Pd catalyst is significantly more active than its impregnated counterpart, with considerable differences observed in product distribution. The impregnated sample yields considerable quantities of acetaldehyde, ethylene, diethyl ether, ethyl acetate and COx . The findings of Grunwaldt et al. [43] and Lee et al. [27] suggest that this contrasting performance could be due to the electronic properties of the supported Pd. However, XPS analysis (see Table S.1) of the samples reveals that this result was not due to variations in the initial oxidation states of palladium. TEM investigations (see Fig. 6) revealed that the impregnated catalyst possessed a broad particle size distribution; ranging from small nanoparticles (<5 nm) to large cluster ensembles (>100 nm).

Note that large PdO particles were also observed via XRD (Fig. S.5). Conversely, the S.I. method yielded more uniformly dispersed Pd nanoparticles in the range of 3–5 nm, which agrees with the chemisorption analysis in Table S.1. These results support the hypothesis that acetonitrile production over palladium is structure sensitive. They suggest that selectivity can be optimized by successfully tuning the morphology and size of the supported particles. Long term activity tests revealed that these Pd catalysts were susceptible to deactivation. To further evaluate the nature of this deactivation, a Pd/TiO2 catalyst (200Pd) was deactivated at 150 ◦ C, treated at 300 ◦ C under feed conditions and subsequently retested. Following this treatment, the original catalytic activity was fully recovered, (Fig. 7) and the catalyst proceeded to deactivate at a similar rate. This regeneration indicated that the small restruc-

C. Hamill et al. / Applied Catalysis A: General 506 (2015) 261–267

50

0.0035 150oC

100 90 80 70 60 50 40 30 20 10 0

0.003

40

0.0025

35

kd (min-1)

Ethanol Conversion (%)

300oC

150oC

45

30 25 20

0.002 0.0015 0.001

15

0.0005

10 0

5

ZrO2

TiO2

Al2O3

ACN Selecvity %

266

SiO2

0 0

200

400

600

800

1000

1200

Fig. 9. The relationship between catalyst deactivation and selectivity at 180 ◦ C, for the various supported palladium catalysts.

Time on Stream (min)

500Pd

200Pd

30Pd

Acetonitrile Yield (%)

100 90 80 70 60 50 40 30 20 10 0

0.005 0.0045 0.004 0.0035 0.003 0.0025 0.002 0.0015 0.001 0.0005 0

ACN Selecvity (%)

kd (min-1)

Fig. 7. Deactivation profile of the 200Pd catalyst (GHSV: 7126 cm−3 g−1 h−1 ) at 150 ◦ C including regeneration under feed conditions.

0

10

20

30

40

50

60

70

Time on Stream (h) Fig. 10. Acetonitrile yield over the 100Pd catalyst at 230 ◦ C.

Ethanol Conversion (%)

Fig. 8. The relationship between catalyst deactivation and selectivity at 150 ◦ C, for 500Pd, 200Pd* and 30Pd* catalysts. * Note GHSV = 7126 ml g−1 h−1 .

100 90 80 70 60 50 40 30 20 10 0 100

150

200

250

300

250

300

Temperature (oC)

ACN Selecvity (%)

turing/sintering observed in Fig. S.6 is not the major source of deactivation for Pd/TiO2 catalysts. The result suggests that the activity loss may be due to the strong adsorption of a reaction product or the deposition of carbon on the catalyst surface. A temperature programmed desorption study (see Fig. S.7) excluded acetonitrile self-poisoning as a deactivation mechanism. To evaluate carbon deposition as a deactivation mechanism, a temperature programmed oxidation (TPO) was conducted with a deactivated catalyst (100Pd). The resulting profile is displayed in Fig. S.8. and showed the release of CO2 from about 225 ◦ C. It is possible that the carbon dioxide observed via temperature programmed oxidation could be due to adsorbed ethyl acetate and/or a similar or associated product. The performance of the 500Pd, 200Pd and 30Pd catalysts was examined at 150 ◦ C to investigate the influence of product distribution on catalyst stability. Fig. 8 reports the deactivation rate constant and acetonitrile selectivity of each material. It can be observed that the most stable catalyst, 500Pd, is also the most selective to acetonitrile, and the least selective, 30Pd, is conversely the least stable. Similar analysis was conducted on the various oxide supported palladium catalysts at 180 ◦ C (see Fig. 9) and a comparable relationship between selectivity and stability was noted. These results re-enforce the suggestion that the production/interaction of an unselective species is involved in the deactivation of the catalyst. Recalling that ethyl acetate is the major by-product in the reaction would imply that it has an important role in the deactivation mechanism. The stability of the catalysts was reassessed at 230 ◦ C, where no ethyl acetate production was observed. Fig. 10 reports a typical result obtained with the 100Pd sample and stable activity was observed, with a yield to acetonitrile of approximately 93%. Finally, the performance of the catalyst was tested at increasing initial concentrations of ethanol to assess its behavior when the conditions are moved toward more realistic ones. Note that further increases in the reactant concentration were constraint by

100 90 80 70 60 50 40 30 20 10 0

100 90 80 70 60 50 40 30 20 10 0 100

150

200

Temperature (oC) Fig. 11. The effect of ethanol concentration on its conversion (a) and selectivity to acetonitrile (b) as a function of temperature over the 200Pd catalyst. 525 ppm (䊏), 1050 ppm () and 1575 ppm (). Note in all tests the concentrations of ammonia and oxygen remain constant. Ratio of NH3 :EtOH; 8, 4 and 2.7 and O2 :EtOH; 13:6.5:4.3 respectively.

the exothermic nature of the reaction (Hr 0 = −371 kJ mol−1 ) and the reactor design employed. The results obtained (Fig. 11) first, demonstrate that nitrile yield is independent of the ethanol partial pressure, and second, the NH3 :EtOH ratio can be significantly reduced (down to 2.7 in this case as NH3 partial pressure remained

C. Hamill et al. / Applied Catalysis A: General 506 (2015) 261–267

constant) without any impact on the selectivity. This result would suggest that the concentration of ethanol in the feed could be increased without significantly altering product distribution. 4. Conclusion The moderate to low temperature production of acetonitrile from ethanol has been reported over a Pd/TiO2 catalyst via ammoxidation. An apparent structure sensitive relationship was observed which suggests that acetonitrile selectivity increases with particle size. Analysis of the results indicates that particles possessing a higher density of terraced sites are more selective. The specific reasons for why this would occur were not investigated here. However the results demonstrate that the preparation method of the palladium catalysts significantly influenced performance. Reversible deactivation was observed over the Pd catalysts, which was assigned to the adsorption of an unselective product, most likely derived from ethyl acetate. On this basis, operating conditions could be devised allowing stable performance at a high yield. Acknowledgement This work was supported by the Deanship of Scientific Research, King Abdulaziz University, Jeddah, Saudi Arabia under grant No. (D-005/431). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.apcata.2015.09. 030. References [1] R.D. Perlack, L.L. Wright, A.F. Turhollow, R.L. Graham, B.J. Stokes, D.C. Erbach, DOE Technical Report. (2005) ORNL/TM-2005/66. [2] J. Rass-Hansen, H. Falsig, B. Jørgensen, C.H. Christensen, J. Chem. Technol. Biotechnol. 82 (2007) 329–333, http://dx.doi.org/10.1002/jctb.1665. [3] M.I. Hoffert, K. Caldeira, G. Benford, D.R. Criswell, C. Green, H. Herzog, A.K. Jain, H.S. Kheshgi, K.S. Lackner, J.S. Lewis, H.D. Lightfoot, W. Manheimer, J.C. Mankins, M.E. Mauel, L.J. Perkins, M.E. Schlesinger, T. Volk, T.M.L. Wigley, Science 298 (2002) 981–987. [4] I.F. McConvey, D. Woods, M. Lewis, Q. Gan, P. Nancarrow, Org. Process Res. Dev. 16 (2012) 612–624, http://dx.doi.org/10.1021/op2003503. [5] P.W. Langvardt, Acrylonitrile, Ullmann’s Encyclopedia of Industrial Chemistry, Wiley-VCH Verlag GmbH & Co., KGaA, Weinheim, 2000. [6] IHS Chemical Week, http://www.chemweek.com/lab/Asahi-Kasei-to-closeseveral-petrochemical-plants-in-Japan 59203.html. 2014 (accessed 08.04.14). [7] D. Lippe, Oil Gas J. 112.9 (2014) 100–108. ˜ [8] E. Rojas, M.O. Guerrero-Pérez, M.A. Banares, Catal. Commun. 10 (2009) 1555–1557, http://dx.doi.org/10.1016/j.catcom.2009.04.016. [9] C. Angelici, B.M. Weckhuysen, P.C.A. Bruijnincx, ChemSusChem 6 (2013) 1595–1614, http://dx.doi.org/10.1002/cssc.201300214. [10] B.M. Reddy, B. Manohar, J. Chem. Soc. Chem. Commun. (1993) 234–235, http://dx.doi.org/10.1039/C39930000234. [11] S.J. Kulkarni, R.R. Rao, M. Subrahmanyam, A.V.R. Rao, J. Chem. Soc. Chem. Commun. (1994) 273, http://dx.doi.org/10.1039/C39940000273. [12] T. Oishi, K. Yamaguchi, N. Mizuno, Angew. Chem. Int. Ed. 48 (2009) 6286–6288, http://dx.doi.org/10.1002/anie.200900418.

267

[13] L.M. Dornan, Q. Cao, J.C.A. Flanagan, J.J. Crawford, M.J. Cook, M.J. Muldoon, Chem. Commun. 49 (2013) 6030–6032, http://dx.doi.org/10.1039/ C3CC42231C. [14] Y. Li, J. N. Armor, Chem. Commun. (1997) 2013–2014, http://dx.doi.org/10. 1039/A703788K. [15] F. Ayari, M. Mhamdi, J. Álvarez-Rodríguez, A.R. Guerrero Ruiz, G. Delahay, A. Ghorbel, Appl. Catal. A: Gen. 415-416 (2012) 132–140, http://dx.doi.org/10. 1016/j.apcata.2011.12.021. [16] Y. Zhang, Y. Zhang, C. Feng, Y. Qiu, J. Hao, Catal. Commun. 10 (2009) 1454–1458, http://dx.doi.org/10.1016/j.catcom.2009.03.019. [17] H.C. Chitwood, B.T. Freure, US 2795600 (1957). [18] R.J. Card, J.L. Schmitt, US 4731464 A (1988). [19] E.C. Corker, U.V. Mentzel, J. Mielby, A. Riisager, R. Fehrmann, Green Chem. 15 (2013) 928–933, http://dx.doi.org/10.1039/C3GC36513A. [20] ICIS, http://www.icis.com/chemicals/channel-info-chemicals-a-z/. 2014 (accessed 14.04.14). [21] Y. Zhang, T. Ma, J. Zhao, J. Catal. 313 (2014) 92–103, http://dx.doi.org/10.1016/ j.jcat.2014.02.014. [22] Z. Shu, Y. Ye, Y. Deng, Y. Zhang, J. Wang, Angew. Chem. 125 (2013) 10767–10770, http://dx.doi.org/10.1002/ange.201305731. [23] T. Ishida, H. Watanabe, T. Takei, A. Hamasaki, M. Tokunaga, M. Haruta, Appl. Catal. A: Gen. 425–426 (2012) 85–90, http://dx.doi.org/10.1016/j.apcata.2012. 03.006. [24] A. Martin, V.N. Kalevaru, ChemCatChem. 2 (2010) 1504–1522, http://dx.doi. org/10.1002/cctc.201000173. [25] P. Rajaram, M.V. Joshi, EP 206632 A1 (1986). [26] G. Centi, F. Cavani, F. Trifirò, Selective Oxidation by Heterogeneous Catalysis, Springer, US, New York, 2000. [27] A.F. Lee, S.F.J. Hackett, J.S.J. Hargreaves, K. Wilson, Green Chem. 8 (2006) 549–555, http://dx.doi.org/10.1039/B601984F. [28] D. Ferri, C. Mondelli, F. Krumeich, A. Baiker, J. Phys. Chem. B 110 (2006) 22982–22986, http://dx.doi.org/10.1021/jp065779z. [29] C.M.A. Parlett, D.W. Bruce, N.S. Hondow, M.A. Newton, A.F. Lee, K. Wilson, ChemCatChem 5 (2013) 939–950, http://dx.doi.org/10.1002/cctc.201200301. [30] M. Sankar, N. Dimitratos, P.J. Miedziak, P.P. Wells, C.J. Kiely, G.J. Hutchings, Chem. Soc. Rev. 41 (2012) 8099–8139, http://dx.doi.org/10.1039/C2CS35296F. [31] C. Jia, F. Schuth, Phys. Chem. Chem. Phys. 13 (2011) 2457–2487, http://dx.doi. org/10.1039/C0CP02680H. [32] F.C. Meunier, D. Reid, A. Goguet, S. Shekhtman, R. Hardacre, W. Deng, M. Flytzani-Stephanopoulos, J. Catal. 247 (2007) 277–287, http://dx.doi.org/10. 1016/j.jcat.2007.02.013. [33] F.C. Meunier, A. Goguet, S. Shekhtman, D. Rooney, H. Daly, Appl. Catal. A: Gen. 340 (2008) 196–202, http://dx.doi.org/10.1016/j.apcata.2008.02.034. [34] A.B. Gaspar, A.M.L. Esteves, F.M.T. Mendes, F.G. Barbosa, L.G. Appel, Appl. Catal. A: Gen. 363 (2009) 9–114, http://dx.doi.org/10.1016/j.apcata.2009.05.001. [35] S.F.J. Hackett, R.M. Brydson, M.H. Gass, I. Harvey, A.D. Newman, K. Wilson, A.F. Lee, Angew. Chem. Int. Ed. 46 (2007) 8593–8596, http://dx.doi.org/10.1002/ anie.200702534. [36] C.M.A. Parlett, L.J. Durndell, K. Wilson, D.W. Bruce, N.S. Hondow, A.F. Lee, Catal. Commun. 44 (2014) 40–45, http://dx.doi.org/10.1016/j.catcom.2013.07.005. [37] G.C. Bond, R. Burch, M.D. Birkett, A.T. Kuhn, G.C. Bond, in: G.C. Bond, G. Webb (Eds.), The Royal Society of Chemistry, The Royal Society of Chemistry, London, 1983, pp. 27–60. [38] G. Cavani, Metal Oxide Catalysis, Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim, 2008. [39] A.B. Gaspar, F.G. Barbosa, S. Letichevsky, L.G. Appel, Appl. Catal. A: Gen. 380 (2010) 113–117, http://dx.doi.org/10.1016/j.apcata.2010.03.034. [40] G. Cabilla, A. Bonivardi, M. Baltanás, Catal. Lett. 55 (1998) 147–156, http://dx. doi.org/10.1023/A:1019095231484. [41] A. Guerrero-Ruiz, S. Yang, Q. Xin, A. Maroto-Valiente, M. Benito-Gonzalez, I. Rodriguez-Ramos, Langmuir 16 (2000) 8100–8106, http://dx.doi.org/10.1021/ la000387j. [42] L. Kesavan, R. Tiruvalam, M.H.A. Rahim, M.I. bin Saiman, D.I. Enache, R.L. Jenkins, N. Dimitratos, J.A. Lopez-Sanchez, S.H. Taylor, D.W. Knight, C.J. Kiely, G.J. Hutchings, Science 331 (2011) 195–199, http://dx.doi.org/10.1126/ science.1198458. [43] J. Grunwaldt, M. Caravati, A. Baiker, J. Phys. Chem. B 110 (2006) 25586–25589, http://dx.doi.org/10.1021/jp066949a.