Selective oxidation of glycerol to dihydroxyacetone over a Pd–Ag catalyst

Abstract

Glycerol oxidation with molecular oxygen was conducted using Pd-based catalysts under neutral conditions. Pd–Ag/C showed higher selectivity to dihydroxyacetone (DHA) and higher activity than Pd/C. The DHA yield reached 44% at 52% glycerol conversion over Pd–Ag/C (Ag/Pd = 1). The Pd–Ag alloy phase is responsible for the high activity and DHA selectivity.

---

Much attention has been given to the transformation of biomass and biomass-derived chemicals to fuel and valuable chemicals for the sustainable development, and the facility that integrates the processes is called biorefinery.^1–4 In the biorefinery, glycerol can be regarded as a renewable feedstock because it is the coproduct of triglyceride saponification and biodiesel fuel manufacture.⁵ The present demand for glycerol cannot compensate its production, and new efficient procedures for the transformation of glycerol into valuable chemicals are highly desired.^6,7 Oxidation with molecular oxygen is an attractive way of conducting the transformation since the reactant is very cheap and abundant.

The oxidation of glycerol leads to a complex reaction pathway in which a large number of products such as dihydroxyacetone (DHA), glyceric acid (GLYAC), hydroxypyruvic acid (HYPAC), dicarboxylic acids, etc., could be obtained (Scheme 1). Oxidation of the terminal OH group of glycerol gives GLYALD and GLYAC. GLYAC is currently produced using inorganic oxidizing agents. The production of GLYAC by catalytic aerobic oxidation of glycerol has been intensively investigated using monometallic or bimetallic catalysts such as Au, Pt and Pd.⁸ Over 95% GLYAC yield has been achieved at full conversion.⁹

Scheme 1 Reaction routes in the oxidation of glycerol.

Oxidation of the secondary OH group of glycerol gives DHA. DHA is currently produced in biocatalytic processes using the acetobacterium Gluconobacter suboxydans with only minor space time yields.¹⁰ Besides the use of DHA as a tanning agent in the cosmetics industry, DHA at a lower market price could be more widely used as building blocks for new degradable polymers.¹¹ However, direct aerobic oxidation of glycerol to DHA has been less studied than that to GLYAC. Bimetallic Pt–Bi catalysts have been mentioned as selective catalysts for DHA formation. Since the discovery of Pt–Bi catalysts by Kimura and Tsuto in 1993,¹² several research groups have developed the oxidation systems using catalysts with various Bi/Pt ratios and supports.^7,13–16 In the recent publications the initial selectivity has been reported to be about 80% and the DHA yield about 50% using a batch reactor. Other than Pt–Bi systems, only Au-based ones have been mentioned for glycerol oxidation to DHA. Claus et al. succeeded in the synthesis of DHA using Au-based catalysts in 2007, while DHA yield was less than that on Pt–Bi catalysts.¹⁷ Dimitratos et al. and Rodrigues et al. also reported an Au catalyst with high selectivity to DHA and the yield was about 20%.^18,19

In this communication, for the first time we developed Pd-based catalytic systems for glycerol oxidation to DHA. Pd catalysts are known to be active for alcohol oxidation.²⁰ We explored Pd-based bimetallic catalysts and found that a Pd–Ag alloy catalyst is very effective for selective oxidation of glycerol to DHA.

Pd–Ag/C catalysts were prepared by a co-impregnation method using XC-72 carbon and mixed aqueous solution of Pd(NO₃)₂ and AgNO₃. The active components were reduced to the metallic states under H₂ flow at 473 K. The catalytic activity of Pd–Ag/C for aerobic glycerol oxidation is shown in Table 1. In comparison, the catalytic activities of Pd/C and Ag/C prepared by similar procedures are also shown. The reaction did not proceed without catalyst (<0.1% conversion). The formations of dicarboxylic acids were negligible in all cases. Pd/C showed some catalytic activity and the selectivity to DHA was around 65% (entry 1). The high DHA selectivity is contrasted with high selectivity to GLYAC reported for Pd/C under basic conditions, although the activity in this case was lower than that under basic conditions.^13,21 Ag/C showed higher selectivity than Pd/C, while the activity was much lower (entry 2). In contrast, Pd–Ag/C (Ag/Pd = 1) showed high activity for the oxidation of the secondary OH group of glycerol (entry 3). The physical mixture of Pd/C and Ag/C showed almost no enhancing effect on the DHA selectivity and was even less active than Pd/C (entry 4), showing the synergetic effects in Pd–Ag/C catalysts. Pd–Ag catalysts on oxide supports such as TiO₂ were also selective, but less active.

Table 1 Glycerol oxidation over Pd and Ag based catalysts

Entry	Catalyst	Ag/Pd molar ratio	Reaction time/h	Conversion/%	Selectivity/%
					DHA	GLYALD	GLYAC	HYPAC	Glycolic acid
Reduction conditions: 473 K, H2 flow 30 cc min^−1, 4 h. Pretreatment conditions: 5 wt% glycerol aqueous solution 20 g, O2 pressure 0.1 MPa, 353 K, 4 h. Reaction conditions: 5 wt% glycerol aqueous solution 20 g, catalyst 50 mg (1 mg Pd), initial O2 pressure 0.3 MPa, 353 K.a Ag (2 wt%)/C, 50 mg.b Pd/C 50 mg, Ag/C 50 mg.c After one use, catalyst 45 mg, 5 wt% glycerol aqueous solution 18 g.d After two uses, catalyst 41 mg, 5 wt% glycerol aqueous solution 16 g.e After three uses, catalyst 36 mg, 5 wt% glycerol aqueous solution 15 g. DHA, dihydroxyacetone; GLYALD, glyceraldehyde; GLYAC, glyceric acid; HYPAC, hydroxypyruvic acid.
1	Pd/C	—	4	2.8	66.1	22.5	10.9	<0.1	0.5
2	Ag/C^a	—	24	0.3	84.0	5.0	11.0	<0.1	<0.1
3	Pd–Ag/C	1	4	6.7	74.6	14.6	7.2	2.1	1.4
4	Pd/C+Ag/C^b	1	4	1.8	70.7	18.8	7.1	1.2	2.2
5	Pd–Ag/C	1	12	18.3	75.1	5.4	14.8	0.3	4.5
6	Pd–Ag/C	1	24	24.5	79.1	3.6	11.2	0.4	5.8
7	Pretreated Pd–Ag/C	1	4	9.5	81.9	8.9	5.6	2.0	1.6
8	Pretreated Pd–Ag/C	0.5	24	10.9	77.7	8.6	9.6	1.3	2.7
9	Pretreated Pd–Ag/C	1	24	20.0	82.2	4.6	8.0	1.7	3.4
10	Pretreated Pd–Ag/C	2	24	16.4	85.0	4.6	5.4	2.8	2.2
11	Pretreated Pd–Ag/C^c	1	24	14.8	80.8	4.9	9.2	2.2	2.9
12	Pretreated Pd–Ag/C^d	1	24	14.5	82.4	5.0	8.2	1.9	2.6
13	Pretreated Pd–Ag/C^e	1	24	12.5	83.0	5.0	8.0	2.0	2.0

---

The reaction time dependence of the aerobic glycerol oxidation over Pd–Ag/C was investigated (entries 3, 5, and 6). The conversion increased with an increase in the reaction time. Interestingly, the DHA selectivity also increased. These data showed that the Pd–Ag/C catalyst at longer reaction time was more selective to DHA than at the initial stage. We treated Pd–Ag/C with glycerol aqueous solution and the pretreated catalyst was applied to the aerobic glycerol oxidation (entry 7). The selectivity to DHA (81.9%) was higher than that of the catalyst without pretreatment. At the same time, the initial rate also increased with the pretreatment.

The effect of Ag loading on the catalytic activity was examined. The amount of Ag was varied from 0.5 to 2 by the molar ratio of Ag to Pd (entries 8–10). The catalysts with more Ag tend to be more selective for DHA formation. The activity is maximum at the molar ratio of 1. Based on the balance of activity and selectivity, we used the catalyst with Ag/Pd = 1 in the following studies.

The results of reuse experiments for Pd–Ag/C (Ag/Pd = 1) are shown in entries 9 and 11–13. The activity of the catalyst used once was about 3/4 of that of the fresh one. The decrease in activity in further reuses was lower than that in the first use. The selectivity was almost unchanged in the reuses. ICP analysis of reaction solution after the filtration of the catalyst shows very slight leaching of Pd and Ag (0.1% and 0.5%, respectively), indicating that leaching is not the main factor in the deactivation.

Comparison between entries 7 and 9 shows that the reaction rate greatly decreased in the longer reaction time. The high selectivity to DHA was maintained, which is in contrast to the case of the Pt–Bi system where DHA selectivity decreased at longer reaction time. With larger amounts of the catalyst, 44% DHA yield was achieved at 52% glycerol conversion (eqn (1)). The value of the yield was comparable to those reported for Pt–Bi systems (48% DHA yield at 80% glycerol conversion).¹⁵ Considering the recyclability of Pd–Ag/C, the major cause of the deactivation of the catalyst over a longer reaction time should not be the irreversible change in the catalyst such as leaching or overoxidation of the active metal. Since several oxidation products of glycerol (GLYAC, HYPAC and dicarboxylic acids) are known chelating agents,²² glycolic acid or propionic acid (0.2 mmol, 0.02 equiv. to glycerol) was added to the initial reaction mixture to test the influence of the product acid (details in ESI†). The glycerol conversion after 4 h in the presence of glycolic acid and propionic acid was 3.4% and 5.6%, respectively, while 9.5% conversion was observed without acid. The DHA selectivity in the presence of acid was 78%. Therefore, poisoning by produced acid was one factor in the deactivation of Pd–Ag/C over a longer reaction time. The decrease in activity of Pd/C after the addition of glycolic acid was much lower (2.8% conv. → 2.0%).

Pd–Ag/C (Ag/Pd = 1) catalysts were characterized by TPR, TEM, XRD, and adsorption studies. The TPR profile confirmed that the reduction of Pd and Ag was complete below 473 K.

The particle size distribution of Pd–Ag/C after the reduction determined by TEM is shown in Fig. 1(a). Particles with sizes of 3–4 nm were dominant, while significant amounts of larger particles with the size of >8 nm were also present. The size distributions of the catalysts after pretreatment with glycerol + O₂ and after use are shown in Fig. 1(b) and (c), respectively. The distributions were similar except the slight shift to larger size.

Fig. 1 Particle size distribution of Pd–Ag/C determined by TEM. (a) After the reduction at 473 K, (b) after the reduction at 473 K and the pretreatment with glycerol + O₂, (c) after the reduction at 473 K, the pretreatment with glycerol + O₂ and the catalytic use for 24 h.

In the XRD patterns (Fig. 2), no peaks assigned to pure Ag, Pd, and metal oxides were observed for reduced, pretreated and used catalysts. The Pd–Ag/C after reduction showed weak, broad peaks around 2θ = 39° (Fig. 2b). The peaks were located between the Bragg lines of pure Ag and Pd. When Pd–Ag/C was treated with glycerol + O₂, sharper and more intense peaks were observed at the same position (Fig. 2c). When Pd–Ag/C was used for the 24 h reaction, the peaks became still more intense (Fig. 2d). The pattern of the Pd–Ag/C used four times was almost the same as that of the Pd–Ag/C used once (Fig. 2e). Considering that the particle size was not so different from TEM analysis, the much larger width of the peaks of Pd–Ag/C after reduction indicates the less crystalline nature of the particles. The bimetallic particles attained uniform Pd–Ag alloy structure after pretreatment with glycerol + O₂. Considering the induction period and the reusability, the crystalline Pd–Ag alloy phase may be catalytically active.

Fig. 2 XRD patterns of C (a) and Pd–Ag/C (b–e). (a) C, (b) after the reduction at 473 K, (c) after the reduction at 473 K and the pretreatment with glycerol + O₂, (d) after the reduction at 473 K, the pretreatment with glycerol + O₂ and the catalytic use for 24 h, (e) after the reduction at 473 K, the pretreatment with glycerol + O₂ and the catalytic use four times.

Table 2 shows the results of CO and O₂ adsorption on Pd/C, Ag/C and Pd–Ag/C. Pd/C adsorbs both CO and O₂, and Ag/C adsorbs only O₂. The CO and O₂ uptakes of Pd–Ag/C after reduction were similar to those of Pd/C. On the other hand, Pd–Ag/C after the pretreatment or the reaction adsorbed lower amounts of CO and O₂, and the decrease was more vivid for O₂ than CO. These results suggest that the nature of the Pd site in Pd–Ag/C after reduction is rather similar to pure Pd and changed during the pretreatment.

Table 2 Adsorption of CO and O₂ on Pd/C, Ag/C and Pd–Ag/C^a

Entry	Catalyst	Pretreatment	CO uptake amount/μmol g^−1	O2 uptake amount/μmol g^−1
a Pd 188 μmol g^−1, Ag 185 μmol g^−1. Reduction conditions: 473 K, H2 flow 30 cc min^−1, 4 h. Pretreatment conditions: 5 wt% glycerol aqueous solution 20 g, O2 pressure 0.1 MPa, 353 K, 4 h. Reaction conditions: 5 wt% glycerol aqueous solution 20 g, catalyst 50 mg, initial O2 pressure 0.3 MPa, 353 K, 24 h.
1	Pd/C	Reduction	50.1	35.0
2	Ag/C	Reduction	0.3	8.0
3	Pd–Ag/C	Reduction	42.9	18.1
4	Pd–Ag/C	Glycerol + O2	16.2	5.0
5	Pd–Ag/C	Reaction	11.1	3.3

---

In summary, the Pd–Ag/C catalyst is active and highly selective to DHA for aerobic glycerol oxidation. The activity and DHA selectivity increased upon treating the catalyst with glycerol aqueous solution. A maximum of 44% DHA yield was achieved at 52% glycerol conversion. Overoxidation of DHA hardly proceeded and higher DHA yield may be achieved by separating product acid from the reaction mixture. Characterization results indicated that the uniform crystalline phase of a Pd–Ag alloy was formed by reduction and pretreatment with glycerol + O₂. This phase can be responsible for the selective DHA formation.

This research is supported by the JSPS KAKENHI (23760737) and the Cabinet Office, Government of Japan, through its “Funding Program for Next Generation World-Leading Researchers”.

Notes and references

1. A. Corma, S. Iborra and A. Velty, Chem. Rev., 2007, 107, 2411.
2. P. Gallezot, Green Chem., 2007, 9, 295.
3. M. Schlaf, Dalton Trans., 2006, 4645.
4. R. M. West, E. L. Kunkes, D. A. Simonetti and J. A. Dumesic, Catal. Today, 2009, 147, 115.
5. C. H. Zhou, J. N. Beltramini, Y. X. Fan and G. Q. Lu, Chem. Soc. Rev., 2008, 37, 527.
6. Y. Nakagawa and K. Tomishige, Catal. Sci. Technol., 2011, 1, 179.
7. A. Brandner, K. Lehnert, A. Bienholz, M. Lucas and P. Claus, Top. Catal., 2009, 52, 278.
8. G. L. Brett, Q. He, C. Hammond, P. J. Miedziak, N. Dimitratos, M. Sankar, A. A. Herzing, M. Conte, J. A. Lopez-Sanchez, C. J. Kiely, D. W. Knight, S. H. Taylor and G. J. Hutchings, Angew. Chem., Int. Ed., 2011, 50, 10136.
9. B. Katryniok, H. Kimura, E. Skrzyńska, J. Girardon, P. Fongarland, M. Capron, R. Ducoulombier, N. Mimura, S. Paul and F. Dumeignil, Green Chem., 2011, 13, 1960.
10. S. Varga, R. Bauer, N. Katsikis and D. Hekmat, Chem. Eng. Technol., 2004, 76, 1306.
11. H. Kimura and K. Tsuto, J. Am. Oil Chem. Soc., 1993, 70, 1027.
12. H. Kimura and K. Tsuto, Appl. Catal., A, 1993, 96, 217.
13. R. Garcia, M. Besson and P. Gallezot, Appl. Catal., A, 1995, 127, 165.
14. L. Xie and J. Hainan, Normal Univ. (Nat. Sci.), 2007, 20, 251.
15. W. Hu, D. Knight, B. Lowry and A. Varma, Ind. Eng. Chem. Res., 2010, 49, 10876.
16. D. Liang, S. Cui, J. Gao, J. Wang, P. Chen and Z. Hou, Chin. J. Catal., 2011, 32, 1831.
17. S. Demirel, K. Lehnert, M. Lucas and P. Claus, Appl. Catal., B, 2007, 70, 637.
18. N. Dimitratos, A. Wagland, G. J. Hutchings and E. H. Stitt, Catal. Today, 2009, 145, 169.
19. E. G. Rodrigues, M. F. R. Pereira, X. Chen, J. J. Delgado and J. J. M. Orfao, J. Catal., 2011, 281, 119.
20. J. W. Medlin, ACS Catal., 2011, 1, 1284.
21. N. Dimitratos, F. Porta and L. Prati, Appl. Catal., A, 2005, 291, 210.
22. N. Wörz, A. Brandner and P. Claus, J. Phys. Chem. C, 2010, 114, 1164.

---

Footnote

† Electronic supplementary information (ESI) available: Experimental details and TEM images. See DOI: 10.1039/c2cy20062g

---