Investigations into the conversion of ethanol into 1,3-butadiene

Abstract

In this paper we report the utility of a variety of silica impregnated bi- and trimetallic catalysts for the conversion of ethanol into 1,3-butadiene. The highest selectivity observed was 67%. The catalysts have been characterised via²⁹Si solid-state NMR spectroscopy, TEM, XPS, pXRD and nitrogen adsorption studies. Different silica materials have been investigated as supports and there appears to be a relationship between the pore diameter and the selectivity to 1,3-butadiene. When the same metals were impregnated on non-acidic supports the conversion dramatically reduced.

---

Introduction

With the increasing scarcity of fossil fuels, issues with supply, the ever growing demands on the world's natural resources coupled with “man-made” global warming due to CO₂ emissions now is the time to investigate sustainable and carbon neutral alternatives to fossil fuels.^1–4 One of the most abundant sustainable raw materials is ethanol, produced by the fermentation of sugars.⁵ There are numerous applications for the conversion of ethanol into commodity chemicals.⁶ In this paper we report, for the first time, a systematic study of bi- and trimetallic catalysts for the conversion of ethanol into 1,3-butadiene (1,3-BD). 1,3-BD is used in the production of rubbers, with ≈25% of world rubber production utilising 1,3-BD. There is an exigent need for the sustainable production of 1,3-BD. This process was initially attempted in the early 20th century around the time of the Second World War. The USA and the former USSR were trying to produce their rubber from non-petrochemical based methods.^7,8 The mechanism for the process is still open to debate. However, it is generally accepted that ethanol is first dehydrogenated to acetaldehyde which then undergoes an aldol condensation with ethanol forming acetaldol, Fig. 1. Acetaldol is then dehydrated to produce cis/transcrotonaldehyde, both aldehydes undergo a Meerwein–Ponndorf–Verley type reduction⁹ to either 3-hydroxybutanol or crotyl alcohol also generating acetaldehyde. Both alcohols generated are rapidly dehydrated to form 1,3-BD. Acid sites in the material can convert ethanol into ethene and diethyl ether, which are undesired side processes.¹⁰ A notable catalytic example is by Quattlebaum et al.¹¹ They reported a series of tantanlum oxides on silica were effective for the desired process.¹¹ More recently catalysts that garnered significant attention were silica/magnesia materials and sepiolite systems.^12–17 Noteworthy, is that when the MgO ∶ SiO₂ systems were washed with aqueous NaOH an increase in selectivity to 1,3-BD was observed due to the reduction in acidity of the support, with an increase in 1,3-BD yield from 44% to 87%.¹⁴ For the sepiolite work it was noted that Mn(II) doped systems were far more active for the production of 1,3-BD than the undoped material, with the highest conversion to 1,3-BD being 33% cf. 0.7% for pure sepiolite.¹⁵ One of the most recent examples is by Tsuchida et al. who have prepared a series of hydroxyapatite materials which produce 1,3-BD and 1-butanol.^18–20 The selectivities of which depended significantly on the ratio of acid ∶ base sites in the material. The highest selectivity to 1,3-BD was 14%, their target was 1-butanol with a maximum selectivty of 71%.

Fig. 1 Generally accepted mechanism for the production of 1,3-butadiene from ethanol.

The aims of this study are twofold—to investigate a range of metal centres for the conversion of ethanol into 1,3-BD with the aim of identifying good combinations with various Lewis acidities and to study the effect that the support plays.

Results and discussion

Our initial attempts focussed on the zeolites Ti/Fe–ALPO-5. However, due to their acidity only ethene and diethyl ether were detected at 350 °C. Attempts then focused on screening a range of silica impregnated bimetallic catalysts, Table 1. These metals were chosen since the majority of literature examples involve at least one of these and they possess a range of Lewis acidities.^10,15 The catalysts were initially studied for 1 h, with 1 g of silica material, an argon flow rate of 25 ml min⁻¹ and ethanol was added to the Ar flow at a rate of 0.1 ml min⁻¹. It was observed that these conditions gave reproducible results. This is a multi-step process and ethanol is required in three steps, therefore being “ethanol rich” should assist the formation of 1,3-butadiene.

Table 1 Results for bimetallic catalysts impregnated onto 60 Å pore diameter silica, all tests are for 1 h and catalysts were calcined at 500 °C

Catalyst ^a	Con/%^b	Selectivity/%
		1,3-BD	Ethene	Acet.	Ether	1-Butene
a All loadings were 1 wt% of each metal except the Zr ∶ Zn system which was 1.5 ∶ 0.5 wt%. b Temp = 375 °C. The LHSV for all runs was ≈1.5 h^−1.
Co ∶ Zn	17.0	6.5	29.8	47.4	16.1	0.4
Cu ∶ Zn	29.0	20.9	42.1	30.1	4.6	2.2
Co ∶ Zr	20.0	3.9	66.9	5.9	14.1	9.4
Cu ∶ Co	17.0	10.8	39.3	37.6	10.9	1.4
Co ∶ Mn	13.0	23.0	47.0	19.2	9.6	1.3
Ce ∶ Zr	24.0	26.6	40.0	27.8	3.5	2.1
Hf ∶ Zn	15.0	4.9	26.9	57.0	11.0	0.2
Mn ∶ Zr	10.5	28.8	46.4	9.3	15.5	0
Cu ∶ Mn	18.0	10.1	10.6	61.7	15.4	2.1
Mn ∶ Zn	17.0	19.0	28.0	35.5	16.5	1.0
Zr ∶ Zn	46.0	38.9	41.1	10.3	6.7	3.0

---

The catalysts were prepared by an incipient wetness method, see ESI† for full details. Initially, a loading of 1 wt% of each metal was used. However, the Zr ∶ Zn system was shown to produce marginally better results with a 0.5 wt% Zn to 1.5 wt% loading of Zr(IV). This loading was further confirmed by ICP-AES measurements and used in all experiments. It was observed that if the wt% of either Zr or Zn increased then more ethene was produced. This is presumably due to the increase in Lewis acidity of the catalyst, which favours the side products. The materials were analysed viapXRD, which indicated no detectable order in the oxide phase. To further characterise the materials several catalysts have been investigated vianitrogen adsorption, ²⁹Si solid-state NMR spectroscopy, TEM and XPS. Table 2 has relevant ²⁹Si NMR data, BET N₂ adsorption analysis and the XPS surface concentration, for further details see ESI†. The materials were studied viapyridine TPD and this showed that the CuZrZn on 150 Å silica had half the amount of acidic groups compared to the 60 Å catalyst.

Table 2 Selected analytical data for the catalysts

Catalyst	NMR data			BET SA^a/m^2 g^−1	Surface composition^c
	Q^2	Q^3	Q^4		Zn	Zr	Cu
a More information is given in the ESI.^1 Samples were pre-treated at 300 °C for 420 minutes under vacuum prior to N2 adsorption. b Calcined at 500 °C. c Surface atom%. d These are the spent catalysts.
40 Å SiO2 ZrZn^b	4.9	31.4	63.6	484.7	0.10	0.32	—
60 Å SiO2 ZrZn^b	3.4	26.7	69.9	475.2	0.09	0.37	—
150 Å SiO2 ZrZn^b	2.6	21.7	75.7	272.3	0.07	0.14	—
60 Å SiO2 ZrZnCu^b	—	—	—	450.3	0.39	0.39	0.39
150 Å SiO2 ZrZnCu^b	—	—	—	246.2	0.22	0.18	0.12
60 Å SiO2 ZrZnCu^bd	—	—	—	476.9	0.23	0.14	0.23
150 Å SiO2 ZrZnCu^bd	—	—	—	289.8	0.21	0.12	0.07

---

For all bimetallic catalyts significant amounts of both ethene and diethyl ether were produced, Table 1. The Brønsted acidic nature of the support is able to dehydrate ethanol. A test with pure silica afforded a 5% conversion, with ethene being the dominant product. It must also be noted that catalysts involving cobalt gave significantly lower carbon balances than other systems. It is well known that CoO can catalyse the hydration of ethanol to produce methane and carbon dioxide, neither of which we are quantifying.²¹ The results from Table 1 show that catalysts with CuO or ZnO tended to be more selective towards the production of acetaldehydevia the dehydrogenation of ethanol.^22–24 This is a positive result as the first step in the process is the dehydrogenation of ethanol to acetaldehyde, Fig. 1. The stand out catalyst from Table 1 was the ZrO₂ ∶ ZnO system and attempts were made to optimise this system, Table 3. This was initially attempted by varying the nature of the support.

Table 3 Results for bimetallic and trimetallic catalysts, all tests are for 1 h

Entry	Catalyst ^a	Support^b	Catalyst calcination temperature/°C	Reactor temp/°C	Con/%	Selectivity/%
						1,3-BD	Ethene	Acet.	Ether	1-butene
a All loadings were 1 wt% of each metal except the Zr ∶ Zn system which was 1.5 ∶ 0.5 wt%. b The number represents the pore diameter of the silica material used as the support in Å. c The catalyst was stirred with a 0.01 M solution of NaOH to neutralise the support. d ZrO2 was used as the support. e LHSV ≈ 1.5 h^−1, f LHSV ≈ 0.75 h^−1, g LHSV ≈ 8 h^−1, h LHSV ≈ 3 h^−1.
1^e	Zr ∶ Zn	60	300	300	4.5	8.1	17.4	64.9	6.9	2.7
2^e	Zr ∶ Zn	60	300	350	27.5	44.0	29.0	19.6	7.3	0.2
3^e	Zr ∶ Zn	60	300	375	44.5	42.9	32.6	15.4	6.1	3.0
4^e	Zr ∶ Zn	60	300	400	50.0	45.8	30.7	13.8	6.0	3.7
5^f	Zr ∶ Zn	60	300	375	60.9	34.8	52.2	4.4	4.5	4.2
6^e	Zr ∶ Zn	40	500	375	55.0	27.7	48.8	8.1	15.0	0.3
7^e	Zr ∶ Zn	60	500	375	46.0	38.9	41.1	10.3	6.7	3.0
8^e	Zr ∶ Zn	150	500	375	48.0	47.9	25.8	9.4	14.0	3.0
9^f	Zr ∶ Zn	150	500	375	50.0	36.8	46.9	10.6	2.6	3.4
10^e	Zr ∶ Zn	60-Na ^c	500	375	18.5	0	1.9	21.0	75.3	0
11^g	Zr ∶ Zn	ZrO2^d	500	375	17.0	13.8	29.2	51.5	5.0	0.2
12^e	Cu ∶ Zr ∶ Zn	60	500	300	21.0	29.7	4.0	45.8	10.2	10.3
13^e	Cu ∶ Zr ∶ Zn	60	500	350	35.0	25.3	23.0	33.5	14.7	3.5
14^e	Cu ∶ Zr ∶ Zn	60	500	375	38.5	50.1	24.9	15.6	4.6	4.8
15^f	Cu ∶ Zr ∶ Zn	60	500	375	44.0	42.7	31.6	18.6	2.6	4.5
16^h	Zr ∶ Zn	SG	500	375	17.8	1.6	74.0	19.9	3.1	1.4
17^h	Cu ∶ Zr ∶ Zn	SG	500	375	29.6	0.9	65.6	8.7	24.0	0.7

---

It is assumed that ZnO is active for the dehydrogenation of ethanol. The silica support and/or a combination of the Lewis acidic metal centres could catalyse the subsequent aldol condensation of ethanol and acetaldehyde. ZrO₂ will catalyse the Meerwein–Ponndorf–Verley reduction of acetaldol.^25–27 The generated alcohols (crotyl alcohol or 3-hydroxybutanol) will easily dehydrate to produce 1,3-BD. The calcination temperature of the catalysts was lowered to 300 °C (entry 3vs.7). The lower calcination temperature marginally improved the selectivity towards 1,3-BD. The reactor temperature was also varied (entries 1 to 4 and 12 to 14) as expected as the temperature increases the conversion and amount of 1,3-BD increased markedly. At lower temperatures a high proportion of acetaldehyde is produced which must be consumed at high temperatures and converted into 1,3-BD. The pore diameter of the support has been varied from 40 Å to 60 Å to 150 Å (entries 6–8) with minimal differences in the overall conversion. However, there were significant increases in the selectivity towards 1,3-BD with an increase from 27.7% for 40 Å material to 47.9% for the 150 Å material. These materials were studied via²⁹Si solid-state NMR spectroscopy to investigate the nature of the silicon environments.† It was observed that the catalyst prepared with 150 Å material has a significantly lower proportion of the acidic Q² and Q³ sites and consequently a higher proportion of Q⁴ sites (75.7% for 150 Å vs. 69.9% for 60 Å and 63.6% for 40 Å, Table 3). Essentially, the Brønsted acidity (this was also confirmed viapyridine adsorption measurements)† of the support is responsible for the reduction in ethene selectivity. If the catalyst was washed with NaOH_(aq) to neutralise the support (entry 10) the effect was to hault the production of 1,3-BD. This implies that a degree of acidity in the support is important for our process. This is in contrast to previous sepiolite work¹⁵ in which a dramatic increase in 1,3-BD was observed with the slight addition of a base. However, the data presented in this work clearly show the opposite effect when using a silica based supported system. To further investigate the nature of the support zirconia was utilised in place of silica (entry 11). It was seen that the conversion dropped significantly as did the selectivity to 1,3-BD. This result indicates that a degree of acidity from the silica itself is required for the process. In addition to the results shown in Table 3MgO and Na₂O/Al₂O₃ were used as supports. In these cases conversions obtained were below 5%. An essential conclusion is for this process to yield significant amounts of 1,3-BD it was essential to use silica impregnated materials.

In the mechanism various processes can be acid catalysed. Therefore it is possible for SiO₂ to play an important role in these processes, which merits further investigation. Sol–gels are an effective method for dispersing small metal-oxide particles onto a silica matrix.²⁸ Sol–gel catalysts were prepared by taking TEOS (tetraethyl ortho silicate), ZrO(NO₃)₂, Zn(NO₃)₂ and Cu(OAc)₂ (entries 16 and 17) with the same target loading as the silica systems. Interestingly, there was a significant reduction in the conversion and selectivity with these catalysts. Catalysts were also prepared by taking silica and reacting with organometallic species {Zn(Me)₂ or Zr(OⁱPr)₄} at 1 wt%, to form anchored species.^29,30 After calcination the catalysts were tested, low conversions ca. 10% were observed with no selectivity towards 1,3-BD. Therefore, simply impregnating the metal salts onto the silica support in a water slurry (see ESI†) and calcining to generate the oxide is the method of choice.^31,32 The active catalysts have been prepared in 20 g batches. Importantly, batch-to-batch variability is minimal as different batches of catalysts have analogous reactivity. To investigate any effect of changing the LHSV the reactions were run with half the flow (entries 5, 9, 15) of EtOH (0.05 ml min⁻¹). As expected the conversions are slightly higher but the selectivity to 1,3-BD is significantly less. This implies that the extra ethanol present in the higher LHSV processes facilitates the full conversion to 1,3-BD. It was noted from Table 1 that CuO systems produced high levels of acetaldehyde. To harness this a trimetallic CuZrZn catalyst was also tested (entries 12–15), which has the highest selectivity towards 1,3-BD. The spent catalysts were analysed viaTGA. These showed a small weight loss of approximately 4 wt%. The weight loss occurred below 150 °C so is probably due to adsorbed ethanol or water.

To test the longevity of our catalysts the most promising systems were studied for 3 h, Table 4. For evolution of gases with time see ESI† and Fig. 2. We have monitored several of the catalysts with time, to gain a deeper insight into the catalytic behaviour. Interestingly, the CuZr impregnated system on 60 Å silica appears to deactivate on-line as the selectivity to 1,3-BD decreased markedly with time. This reduction was not observed for the ZrZn system impregnated on the same support. However, all catalysts tended to show a reduced conversion over this 3 h period compared to the 1 hour run. It is possible to reuse the ZrZn catalyst on a 150 Å SiO₂ support with a small drop in overall selectivity to 1,3-BD, Fig. 2 and Table 4. However, for this system selectivity to diethyl ether has slightly increased, indicating a small increase in acidity of the catalyst. We also studied the CuZrZn system impregnated onto 60 Å and 150 Å silica. Interestingly and reproducible, when we monitored the 150 Å catalyst with time, bottom Fig. 2, we observed an initial lag in the product formation, with very low conversions after 30 min compared to 60 min. This implies there is an activation process present in this system, which we do not observe in other catalysts. However, after 1 h stable gas compositions are observed. This is in contrast to the 60 Å material which showed a high selectivity initially but the 1,3-butadiene selectivity tailed off with time. This must be related to deactivation of the catalyst or the pores becoming blocked. Noteworthy, is that the observed tailing off was only observed for the 60 Å materials. This is potentially related to the pores becoming blocked or obstructed, which is not observed to the same extent in the larger pore materials.

Table 4 Results for bimetallic and trimetallic catalysts, all tests are for 3 h and catalysts were calcined at 500 °C

Catalyst ^a	Support^b	Con/%	Selectivity/%
			1,3-BD	Ethene	Acet.	Ether	1-Butene
a All loadings were 1 wt% of each metal except the Zr ∶ Zn system which was 1.5 ∶ 0.5 wt%. b The number represents the pore diameter of the silica material used as the support. c Recycling experiments.
Zr ∶ Zn	60	33.0	34.7	40.6	16.3	6.3	2.1
Zr ∶ Zn—1st	150	21.0	36.1	27.1	14.9	16.3	2.7
Zr ∶ Zn—2nd	150^c	23.0	33.5	25.0	13.3	26.0	2.2
Zr ∶ Zn—3rd	150^c	22.0	34.0	26.5	12.5	22.0	4.8
Cu ∶ Zr	60	16.5	25.9	48.0	9.4	13.8	2.8
Cu ∶ Zr ∶ Zn	150	44.6	67.4	20.8	5.3	2.8	3.6
Cu ∶ Zr ∶ Zn—2nd	150^c	39.6	56.2	16.2	18.1	6.4	3.1
Cu ∶ Zr ∶ Zn	60	27.0	33.0	26.6	32.8	5.2	2.4

---

Fig. 2 Top: GC measurements of the evolution of ethene, 1,3-BD and 1-butene for the Zr ∶ Zn system supported on a 150 Å pore diameter silica fresh and recycled data. Middle GC measurements of the evolution of ethene, 1,3-BD and 1-butene for the Zr ∶ Zn system supported on a 150 Å at various feed compositions. Bottom 1,3-butadiene and ethene on-line measurement for a variety of catalysts—see caption for details.

Finally, it has been shown that significant improvements can be observed if acetaldehyde is added to the feed. For example, Kvisle observed for a MgO ∶ SiO₂ catalyst an increase in selectivty to 1,3-BD from 30% (pure ethanol) to 53% for the 9 ∶ 1 ethanol-to-acetaldehyde feed. Encouragingly, with a 8 ∶ 2 ethanol-to-acetaldehyde feed the selectivity was 68%.¹² Quattlebaum et al. also observed an enhancement with acetaldehyde and crotonaldehyde added to the feed.¹¹ Thus, our promising ZrZn system was tested. Under these conditions, to minimise pressure due to the high volatility of acetaldehyde, the argon flow was reduced to 12 ml min⁻¹ and a flow rate of 0.05 ml min⁻¹ for the feed was utilised, Table 5.

Table 5 Results for varying the ratio of EtOH ∶ acetaldehyde in the feed, the time for these runs was 3 h and catalysts were calcined at 500 °C

Catalyst ^a	Support^b	E ∶ A^c	Con/%	Selectivity/%
				1,3-BD	Ethene	Acet.	Ether	1-Butene
a Metal loading was Zr ∶ Zn 1.5 ∶ 0.5 wt%. b The number represents the pore diameter of the silica material used as the support. c The ethanol-to-acetaldehyde feed ratio. The LHSV for all runs was ≈0.75 h^−1 at a feed flow of 0.05 ml min^−1 and in all cases an Ar flow of 12 ml min^−1 was used.
Zr ∶ Zn	60	10 ∶ 0	42.0	35.7	42.2	9.6	6.7	4.7
Zr ∶ Zn	60	9 ∶ 1	53.0	44.7	33.3	11.5	4.5	6.0
Zr ∶ Zn	150	10 ∶ 0	39.0	42.1	31.1	13.9	8.6	4.4
Zr ∶ Zn	150	9 ∶ 1	41.0	61.8	18.8	15.6	3.5	0.3
Zr ∶ Zn	150	8 ∶ 2	45.0	66.0	9.8	14.7	4.8	4.8
Cu ∶ Zr ∶ Zn	150	9 ∶ 1	46.0	37.0	20.9	30.5	10.5	1.8
Hf ∶ Zn	60	9 ∶ 1	26.0	6.7	20.0	58.0	14.9	0.3

---

Compared to previous conditions (0.1 ml min⁻¹ feed and 25 ml min⁻¹argon) there is an increase in conversion and selectivity to 1,3-BD this is presumably due to the higher residence time of the ethanol on the support. The amount of catalyst remained constant and the gas flow and ethanol feed rate both halved, thus effectively doubling the residence time. As previously observed the 150 Å system gave a high selectivity to 1,3-BD compared to the 60 Å at a similar conversion. When acetaldehyde was added to the feed there was a dramatic increase in the selectivity towards 1,3-BD compared to ethene, see Fig. 2. This is implying that the ethanol is being preferentially utilised in the aldol condensation with the added acetaldehyde and therefore is dehydrated to a lesser extent. Essentially the catalyst is jumping to stage 2 of the process (aldol condensation). A poor performing Hf ∶ Zn catalyst was also investigated with added acetaldehyde in the feed in an attempt to improve its selectivity, this was unfortunately unsuccessful. Also, addition of acetaldehyde to the CuZrZn catalyst did not result in any significant enhancement. This is presumably due to the fact that this catalyst did produce significant levels of acetaldehyde with pure ethanol.

Conclusions

In conclusion catalytic data have been presented for the conversion of ethanol into 1,3-BD and the nature of the support has been investigated. We have found that a degree of acidity in the support is critical as several steps in the mechanism are acid catalysed. However, the more acidic supports form larger amounts of the by-products ethene and diethyl ether. The most promising catalyst investigated is a Zr ∶ Zn system impregnated onto silica. Both Zn(II) and Zr(IV) are Lewis acidic and this is believed to enchance the activity. Furthermore, if the feed is doped with acetaldehyde then an enhancement in selectivity is observed, since the added aldehyde can enhance the aldol condensation reaction.

Experimental

The catalysts were easily prepared by taking a slurry of a water soluble salt of each metal and silica.† For the initial experiments Davisil grade 635 silica was used with a pore diameter of 60 Å (35–65 Mesh), 150 Å Davisil 645 (35–65 Mesh) and 40 Å Merck Grade 10181 (35–70 Mesh) were used in subsequent studies). Initially, a loading of 1 wt% of each metal was targeted in our synthesis. The water was allowed to evaporate to leave the silica impregnated materials which are calcined in air at either 300 °C or 500 °C for five hours. In each test 1.0 g of the catalyst was placed in a quartz U-tube (internal diameter of 1 cm and a length of 16 cm), this was then placed in a furnace and heated to the desired temperature. Using a mass flow controller heated argon was blown over the sample at 25 ml min⁻¹, at atmospheric pressure. With the aid of an HPLC pump liquid ethanol was introduced into the stream of argon at a rate of 0.1 ml min⁻¹. Safety: as the tube was Quartz it was decided to minimise the potential for pressure build up and use a relatively low flow of argon. The exhaust gases were bubbled through two acetone bubblers (the first of which was cooled to 0 °C) to collect any unconverted ethanol and any condensables formed in the reaction. The acetone was analysed via¹H NMR spectroscopy to determine the amount of dissolved ethanol (hence conversion) and other condensables, which was used to corroborate the GC analysis. Prior to the gases being bubbled through acetone it is possible to take a sample of the exhaust stream for analysis viaGC. Thus, it is possible to obtain a carbon balance for this process, which was typically greater than 90%.† Gas samples were analysed on a Hewlett Packard 5890 GC with a CP-LowOx column (10 m by 0.53 mm) and analysed viaFID. The GC has been calibrated using a certified BOC 5 blend gas mixture. The contact time = volume of voids in the bed per total flow rate. These are estimated to be 2.3 seconds for the 25 ml min⁻¹ runs and 4.5 seconds for the 12 ml min⁻¹ runs.

Acknowledgements

We wish to thank the EPSRC and Ineos for funding MDJ, CGK, and for solid-state NMR (Durham, Dr Apperley) Dr Zhou (St. Andrews) for TEM measurements and XPS (Cardiff EP/F019823/1 Dr Morgan) and NCESS (Dr Law). We also thank the referees for useful discussions during the reviewing process.

Notes and references

1. J. Goldemberg, Science, 2007, 315, 808–810.
2. A. Corma, S. Iborra and A. Velty, Chem. Rev., 2007, 107, 2411–2502.
3. G. W. Huber, S. Iborra and A. Corma, Chem. Rev., 2006, 106, 4044–4098.
4. J. van Haveren, E. L. Scott and J. Sanders, Biofuels, Bioprod. Biorefin., 2008, 2, 41–57.
5. J. Rass-Hansen, H. Falsig, B. Jorgensen and C. H. Christensen, J. Chem. Technol. Biotechnol., 2007, 82, 329–333.
6. J. Llorca, P. R. de la Piscina, J. Sales and N. Homs, Chem. Commun., 2001, 641–642.
7. L. U. Spence, E. Park, D. J. Butterbaugh and D. G. Kundiger, United States Pat., 2438464, 1948.
8. I. L. Murray, J. L. Marsh, W. Va and S. P. Smith, US Patent, 2403742, 1946.
9. V. A. Ivanov, J. Bachelier, F. Audry and J. C. Lavalley, J. Mol. Catal., 1994, 91, 45–59.
10. S. K. Bhattacharyya and B. N. Avasthi, Ind. Eng. Chem. Process Des. Dev., 1963, 2, 45–51.
11. W. M. Quattlebaum, W. J. Toussaint and J. T. Dunn, J. Am. Chem. Soc., 1947, 69, 593–599.
12. S. Kvisle, A. Aguero and R. P. A. Sneeden, Appl. Catal., 1988, 43, 117–131.
13. H. Niiyama, E. Echigoya and S. Morii, Bull. Chem. Soc. Jpn., 1972, 45, 655–659.
14. R. Ohnishi, T. Akimoto and K. Tanabe, J. Chem. Soc., Chem. Commun., 1985, 1613–1614.
15. Y. Kitayama and A. Michishita, J. Chem. Soc., Chem. Commun., 1981, 401–402.
16. Y. Kitayama, M. Satoh and T. Kodama, Catal. Lett., 1996, 36, 95–97.
17. V. Gruver, A. Sun and J. J. Fripiat, Catal. Lett., 1995, 34, 359–364.
18. T. Tsuchida, J. Kubo, T. Yoshioka, S. Sakuma, T. Takeguchi and W. Ueda, J. Catal., 2008, 259, 183–189.
19. T. Tsuchida, S. Sakuma, T. Takeguchi and W. Ueda, Ind. Eng. Chem. Res., 2006, 45, 8634–8642.
20. T. Tsuchida, T. Yoshioka, S. Sakuma, T. Takeguchi and W. Ueda, Ind. Eng. Chem. Res., 2008, 47, 1443–1452.
21. F. Haga, T. Nakajima, H. Miya and S. Mishima, Catal. Lett., 1997, 48, 223–227.
22. F. W. Chang, H. C. Yang, L. S. Roselin and W. Y. Kuo, Appl. Catal., A, 2006, 304, 30–39.
23. J. L. Gole and M. G. White, J. Catal., 2001, 204, 249–252.
24. Y. Z. Hao, L. X. Tao and L. B. Zheng, Appl. Catal., A, 1994, 115, 219–228.
25. S. H. Liu, G. K. Chuah and S. Jaenicke, J. Mol. Catal. A: Chem., 2004, 220, 267–274.
26. Y. Z. Zhu, S. H. Liu, S. Jaenicke and G. Chuah, Catal. Today, 2004, 97, 249–255.
27. A. Ramanathan, D. Klomp, J. A. Peters and U. Hanefeld, J. Mol. Catal. A: Chem., 2006, 260, 62–69.
28. G. Ennas, A. Musinu, G. Piccaluga, D. Zedda, D. Gatteschi, C. Sangregorio, J. L. Stanger, G. Concas and G. Spano, Chem. Mater., 1998, 10, 495–502.
29. M. D. Jones, M. G. Davidson, C. G. Keir, A. J. Wooles, M. F. Mahon and D. C. Apperley, Dalton Trans., 2008, 3655–3657.
30. M. D. Jones, C. G. Keir, A. L. Johnson and M. F. Mahon, Polyhedron, 2010, 29, 312–316.
31. Z. L. Wang, Q. S. Liu, J. F. Yu, T. H. Wu and G. J. Wang, Appl. Catal., A, 2003, 239, 87–94.
32. R. Moreno-Tost, E. R. Castellon and A. Jimenez-Lopez, J. Mol. Catal. A: Chem., 2006, 248, 126–134.

---

Footnote

† Electronic supplementary information (ESI) available: Full preparation and characterisation of the catalysts, catalytic procedure and sample GC and NMRs. See DOI: 10.1039/c0cy00081g

---