Selective hydrogenolysis of raw glycerol to 1,2-propanediol over Cu–ZnO catalysts in fixed-bed reactor

Abstract

The catalytic properties of Cu–ZnO catalysts for glycerol hydrogenolysis to 1,2-propanediol (1,2-PDO) were tested in a fixed-bed reactor at 250 °C and 2.0 MPa H₂. The relation between composition, surface properties, and catalytic performance of glycerol hydrogenation of Cu–ZnO catalysts was studied using nitrogen adsorption (BET methods), XRD, H₂ temperature-programmed reduction, and N₂O chemisorptions. It was found that there was a close link between the surface CuO amount of Cu–ZnO catalyst and the reactivity for glycerol hydrogenation. The Cu–ZnO catalyst (Cu/Zn = 1.86) which had the highest surface Cu amount showed the best catalytic activity for glycerol hydrogenolysis. Furthermore, Cu–ZnO catalyst presented good stability and remarkable catalytic activity for glycerol hydrogenolysis to 1,2-PDO using raw glycerol derived from the fat saponification as feedstock.

Graphical abstract

Glycerol hydrogenolysis to 1,2-PDO on Cu–ZnO catalyst in fixed-bed reactor.

Key words:

• glycerol
• hydrogenolysis
• Cu–ZnO
• 1,2-propanediol

Biodiesel is a new type of clean and environmentally friendly fuel, as it contains very little sulfur and can substitute for fossil fuel thus reducing the emission of carbon dioxide. Every 9 kg of biodiesel produced, around 1 kg of crude glycerol will be obtained as a by-product. Therefore, large amount of glycerol becomes surplus as a result of the rapid expansion of biodiesel manufacture. Also, glycerol can be produced by other industries, such as fat saponification and alcoholic beverage production units. In order to maintain the balance of glycerol in chemical industry chain, as well as to avoid environmental issues associated with its disposal, converting this very biodiesel by-product to other chemical products has become highly decisive.

By now, glycerol utilization has received extensive attention and been widely studied. Some important chemical intermediates, such as 1,3-propanediol (1,3-PDO),¹⁾ citric acid,²⁾ succinic acid³⁾ can also be obtained under milder conditions via biotechnological processes. Besides, acrolein,^4,5) an important intermediate for manufacturing acrylic resin, and olefin can be produced by catalytic conversion of glycerol. However, more efforts^6–9) have been devoted to the conversion and utilization of glycerol to economically value-added product, 1,2-propanediol (1,2-PDO), which is widely used as biodegradable functional fluids such as deicing reagents, antifreezes and coolants, and precursors in the syntheses of polyester resins.

The hydrogenolysis of glycerol to 1,2-PDO is suggested to proceed via dehydration of glycerol to acetol by acid catalysis followed by catalytic hydrogenation, which has been supported by the observation of acetol in the reaction products and its selective hydrogenation to 1,2-PDO.^6,10) Some investigations have implied that acidity of catalyst supports^11–13) or of reaction system^14,15) should be a positive factor to improve the reaction activity during the hydrogenolysis of glycerol to 1,2-PDO. In the published works, Cu–ZnO bifunctional catalyst showed good performance in glycerol hydrogenolysis.^16,17) It was found that ZnO catalyzed the reaction by dehydrating the glycerol and further hydrogenolysis took place on the CuO site.¹⁷⁾

Although there are numerous reports demonstrating a high yield of 1,2-PDO using batch reactors, just a few reports have evaluated the feasibility of raw glycerol as a substitute for 1,2-PDO production. In this study, raw glycerol derived from the fat saponification process was utilized without any purification to produce 1,2-PDO. Pure glycerol was also tested under the same conditions. The catalytic performances of Cu–ZnO catalysts in glycerol hydrogenolysis to 1,2-PDO were investigated. Fixed-bed flow reactor was chose for glycerol hydrogenolysis, as it was more economic and efficient for continuous and large-scale production compared to batch reactor. The relation between composition, surface properties, and catalytic performance of glycerol hydrogenation of Cu–ZnO was discussed. Important reference information of a suitable and industrial catalyst for catalytic hydrogenation of glycerol to 1,2-propanediol was also provided.

Materials and methods

CuO–ZnO oxide samples were prepared via co-precipitation method. A series of copper zinc hydroxycarbonate precursors were prepared by adding ammonium carbonate solution (3 M) to aqueous solutions of calculated amounts mixed solution of Cu(NO₃)₂·3H₂O and Zn(NO₃)₂·6H₂O at room temperature. Precipitation reagent ((NH₄)₂CO₃, 3 M) was controlled to adjust the desired pH 7 in the mother liquor. The resulting precipitates were aged 3 h in the mother liquor, followed by filtration and washing with distilled water and then dried overnight at 100 °C and calcined at 673 K for 3 h. CuO–ZnO samples were measured by ICP analysis to get accurate Cu/Zn molar ratio (0.39, 1.08, 1.86, 3.01, 4.03, 4.88, respectively). For comparison, pure CuO and ZnO were also prepared by precipitation method.

The catalytic activity was carried out in a fixed-bed stainless steel reactor consisting of a stainless steel tube with an internal diameter of 6 mm and a coaxially centered thermocouple with its tip located in the middle of the bed using 0.9 g Cu–ZnO oxide catalysts (20–40 mesh). Prior to activity test, oxide catalysts were reduced in H₂ flow at 300 °C for 3 h under atmospheric pressure and then dropped to reaction temperature. The following reaction conditions were employed during activity test: T = 250 °C; H₂, p = 2.0 MPa, flow rate = 90 ml/min; glycerol water solution (40 wt%) flow rate = 3 or 6 ml/h. The liquid products were condensed and collected each 2 h under the fixed-bed reactor after reaction system was steady-going. The collected liquid products mainly included acetol, 1,2-propanediol, methanol, ethanol, 1-propanol, ethylene glycol were analyzed using a gas chromatograph with a PEG capillary column (50 m × 0.32 mm) and a flame ionization detector. Glycerol conversion and selectivity of the products were calculated on a carbon basis. Raw glycerol (98 wt%) was derived from the fat saponification process.

Powder X-ray diffraction patterns were recorded on a Philips X′pert diffractometer (Cu Ka radiation, 40 kV, 40 mA, scanning range: 5°–60°). The average crystallite sizes of Cu and ZnO were calculated from XRD patterns using Scherrer equation. For H₂-TPR (Hydrogen Temperature-programmed reduction) measurement, 50 mg catalyst (20–40 mesh) was loaded in a quartz tube reactor and then treated with 5 mol% H₂ in Ar at a heating rate of 5 °C/min from room temperature to 500 °C. The amount of the consumed hydrogen was determined by a thermal conductivity detector. For comparison, the TPR experiments of pure CuO and ZnO samples were also carried out in the same way. The copper surface dispersion was determined by N₂O decomposition (N₂O + 2Cu_s → N₂ + Cu_s–O–Cu_s) at 90 °C followed by TPR.⁹⁾ Cu–ZnO oxide samples were reduced at 300 °C for 3 h and then flushed with Ar and cooled to 90 °C, afterward they were flushed with 25 vol % N₂O in N₂ for 30 min. Finally, the samples were again flushed with Ar and cooled to room temperature followed by TPR. TPR run was carried out in a 5% H₂/Ar mixture at heating ramp rate of 10 °C/min. The amount of chemisorbed O atoms was calculated from the TPR peak area of pure CuO. The specific surface area of the metallic copper was calculated assuming a reaction stoichiometry of two Cu atoms per oxygen atoms, a Cu surface density of 1.46 × 10¹⁹ Cu atoms/m².

Results and discussion

Cu–ZnO catalysts with varying Cu/Zn molar ratio were studied to optimize the catalyst composition for the glycerol hydrogenolysis in a fixed-bed reactor at the reaction conditions of 2 MPa H₂ and 250 °C. Fig. 1 shows glycerol conversion and main products selectivity of the Cu–ZnO catalyst (Cu/Zn = 4.03) are stable after 10-h reaction. Glycerol hydrogenolysis activity of Cu–ZnO catalysts with different composition in reaction time 10 h is presented in Table 1. With increasing of Cu content, the selectivity of 1,2-PDO increased, and the glycerol conversion initially increased and then decreased with a further increase in Cu composition. The optimum conversion recorded for the catalyst with Cu/Zn = 1.86. The product distributions over the Cu–ZnO catalysts suggest that the selectivity toward the formation of 1,2-PDO increased with increasing of Cu content (or decreasing of ZnO content).

Fig. 1. The catalytic properties of Cu–ZnO catalyst (Cu/Zn = 4.03).

Table 1. Effect of Cu/(Cu + Zn) ratio on the glycerol hydrogenolysis activity.

Cu/Zn (molar ratio)	Glycerol conversion (mol%)	Selectivity (mol%)
		1,2-PDO	Acetol	1-PO	EG	Et-OH	Me-OH	Condensations
ZnO	/	/	/	/	/	/	/	/
0.39	15.8	42.3	51.2	0.1	1.7	0.3	0.3	4.1
1.08	25.2	41.8	49.1	0.1	2.3	0.2	0.3	6.2
1.86	47.9	50.4	29.8	0.1	2.3	0.6	0.5	16.3
3.01	43.1	53	29.1	0.3	5.7	0.9	0.7	10.3
4.03	41.9	57.8	28	0.2	4.5	0.7	0.7	8.1
4.88	33.1	67.6	17.9	0.2	5.3	0.5	0.6	7.9
CuO	9.8	40.5	52.7	/	3.3	0.5	0.6	2.4

Notes: Reaction conditions: CuO–ZnO sample weight: 0.9 g; reaction temperature: 250 °C; H₂ pressure: 2 MPa; H₂ flow rate 90 ml/min; glycerol (40 wt% in water) flow rate: 6 ml/h. Collection time: 10–12 h.

XRD patterns of the catalysts are presented in Fig. 2. CuO–ZnO oxide sample with Cu/Zn = 1.86, as shown by XRD results, contained merely ZnO and CuO crystallites. CuO in CuO–ZnO oxide sample was reduced to CuO after H₂ 300 °C 3-h pre-reduction, as proved by the concurrent appearance of CuO crystallite diffraction peaks at 43.3°, 50.6°, and 74.2°, etc. and disappearance of CuO diffraction peaks at 35.1°, 38.7°, 48.8°, etc. Compared to the fresh pre-reduced sample, CuO crystallite size became bigger after reaction, since Cu crystallite diffraction peaks became sharp and narrow distinctly. Fig. 3 showed the XRD patterns of all Cu–ZnO catalysts after pre-reduction. CuO and ZnO crystallite diffraction peaks were clearly observed in all samples after pre-reduction. With the Cu content increase in Cu–ZnO catalysts (from curve a to curve f), CuO crystallite diffraction peaks intensity increased while ZnO crystallite diffraction peaks intensity decreased. Also, the FWHM (full width at half maximum) of CuO crystallite diffraction peaks decreased with the increasing of Cu loadings which indicated the increasing of CuO crystal particle size. It was discussed in detail in Table 2.

Fig. 2. X-ray diffraction patterns of Cu–ZnO catalysts (Cu/Zn) = 1.86.

Fig. 3. X-ray diffraction patterns of Cu–ZnO catalysts after pre-reduction.

Table 2. Physicochemical properties of Cu–ZnO catalysts.

Cu/Zn	BET surface area (m^2/g cat)		CuO crystallite size^a (nm)		Surface Cu^b (μmol/g cat)
	Before reaction	After reaction	Before reaction	After reaction	Before reaction	After reaction
0.39	35.4	20.6	10.4	40.7	768.8	269.2
1.08	26.1	19.2	29.6	52.7	870.7	349.2
1.86	23.4	14.5	42.1	59.5	560.2	390.5
3.01	13.3	9.5	44.0	62.9	550.5	249.8
4.03	9.3	8.7	46.1	67.7	509.3	220.7
4.88	9.3	8.0	51.8	70.2	451.1	208.6

^a Calculated from XRD date by Scherer formula.

^b Calculated from N₂O adsorption.

The H₂-TPR profiles of CuO–ZnO samples are reported in Fig. 4. The TPR profile of CuO shows a wide peak with maximum at 343 °C. Pure ZnO did not reduce under the same experimental conditions. Integrated hydrogen consumption demonstrated that a complete reduction of Cu(II) to Cu(0) was always achieved at the final temperature of each TPR run. With the Cu content decrease in CuO–ZnO samples, the obvious shift of the CuO reduction temperature to lower level demonstrated the promoting effect of ZnO in copper dispersion and reducibility was caused by the strong interaction between ZnO and copper components. All the reduction peaks were wide, and even showed a shoulder-peaks (curve b and c), which could be assigned to the different CuO species¹⁶⁾ formed during the catalyst preparation process.

Fig. 4. H₂-TPR profiles of CuO–ZnO catalysts.

In order to discuss regular changes of Cu–ZnO catalysts activity with different Cu/Zn ratio, those catalyst measures characterized by XRD, N₂O chemisorption, and BET were employed, and the results are listed in Table 2. CuO crystallite sizes calculated from the XRD data increased while surface amount of CuO by N₂O adsorption decreased after the reaction, indicating that CuO crystallites agglomerated during the reaction process. Thus, we will discuss the stable surface CuO amount based on the data after reaction. When Cu content was low, the glycerol conversion showed remarkable increase with Cu content increasing. When Cu content exceeded Cu/Zn = 1.86 and continued to increase, it led to glycerol conversion decrease. The stable surface CuO amount of Cu–ZnO catalyst reached the highest level while Cu/Zn = 1.86, which resulted in the optimum conversion of glycerol.

Fig. 5 illustrates glycerol reaction routes over copper catalyst surface. The mechanism to produce 1,2-PDO from glycerol via acetol intermediate was proposed and validated by Dasari et al.⁶⁾ Relative low H₂ pressure and short contacting time in our fixed-bed reactor led to incomplete conversion of the acetol, which was dehydrated from glycerol, into 1,2-PDO, therefore relatively low 1,2-PDO selectivity was detected in our work than others.^7,17,18) Some condensation products, such as 2,4-dimethyl-1,3-dioxolane-2-methanol, formed by aldol reaction of acetol with 1,2-PDO, was detected by GC-MS measurement, which was reported in glycerol hydrogenation for the first time. Prior researchers^17,18) considered that glycerol was dehydrated to acetol and glycidol intermediates on acidic ZnO surfaces, followed by their hydrogenation on Cu surfaces. In our work, that CuO afforded the hydrogenation function as well as hydrogenolysis function was proved by catalytic ability of CuO reduced from pure CuO in continuous fixed-bed reactor, though the conversion (9.8%) was relatively low. Most researchers also thought reduced copper catalyst was effective for dehydration of glycerol to produce acetol but maybe in different reaction pathway. Sato et al.¹⁹⁾ proposed that dehydration of glycerol directly occurred on reduced copper catalyst while Liu et al.^16,20) thought dehydrogenation of glycerol directly occurred first on reduced copper catalyst, as shown in Fig. 5 and Equations (1) and (2), respectively.

Fig. 5. Schematic routes of proposed reaction mechanism.

Raw glycerol from fat saponification and pure glycerol hydrogenolysis results are shown in Fig. 6. It indicated that formed Cu–ZnO catalyst pellet still maintained a good stability after 200-h reaction. Using the saponification by-product glycerol as raw material, Cu–ZnO catalyst pellet showed a good application prospects with 93 mol% glycerol conversion and 79 mol% 1,2-PDO selectivity, although slightly lower than using pure glycerol (glycerol conversion 96%, 1,2-PDO selectivity 83%). Also, the main by-product acetol selectivity was 13.6% higher than using pure glycerol (10.6%). The sum of all other by-products including condensations, 1-propanol, ethanol, methanol, and ethylene glycol was about 7%.

Fig. 6. Crude and pure glycerol hydrogenation properties on formed Cu–ZnO catalyst.

The optimum conversion and condensations selectivity were recorded for the catalyst with Cu/Zn = 1.86 in a fixed-bed reactor with reaction conditions of 2 MPa H₂ pressure, 250 °C, reaction time of 12 h. In the Cu–ZnO catalyst for glycerol hydrogenation, CuO functioned as the dehydration as well as the hydrogenation. There is a close link between the Cu amount in Cu–ZnO catalyst and the reactivity for glycerol hydrogenation. Furthermore, Cu–ZnO catalyst presented good stability and remarkable catalytic activity for glycerol hydrogenolysis to 1,2-PDO using raw glycerol derived from the fat saponification as feedstock.

Authors contribution

FY and XB conceived and designed the study. GQ, XB, and TQ performed the experiments.GQ and XB wrote the paper. GQ and XB reviewed and edited the manuscript. All authors read and approved the manuscript.

Disclosure statement

No potential conflict of interest was reported by the authors.

Funding

This work was supported by the Specialized Research Fund for the Doctoral Program of Higher Education [grant number SRFDP-2012009111001]; Natural Science Foundation of Jiangsu Province [grant number BK2010304].