фонд поддержки науки

  • Увеличить размер шрифта
  • Размер шрифта по умолчанию
  • Уменьшить размер шрифта

DESIGN OF PALLADIUM CATALYSTS SUPPORTED ON MESOPOROUS CARBON FOR HARDENING OF VEGETABLE OILS

Abstract

Nanosized palladium catalysts supported on mesoporous carbon Sibunit were developed for hydroprocessing of natural oil feedstocks to produce harmless nickel-free margarine and confectionery fat with high organoleptic quality. The effects of supports chemical pretreatment, catalyst particle size, Pd loading and the type of vegetable oil were studied. The laboratory as well as pilot-scale behavior of Pd catalysts for the oil hydrogenation in comparison with Ni counterparts was evaluated. Sunflower oil hydrogenation was studied in a laboratory batch reactor over Pd/C (Sibunit) at 413 K and hydrogen pressure of 9 bar using catalyst samples with different palladium loading and support grain size. 0.1 wt.% Pd/C showed the same activity during 21 repeated cycles of sunflower oil hydrogenation with catalyst/oil weight ratio = 1/100-1/1000 producing as much as 1797.5 kg hydrogenating oil per 1 g Pd without being deactivated. The same samples 0.1 wt.% and 1 wt.% Pd/C were tested in a pilot set up with the catalyst/oil weight ratio = 1/100-1/1000 displaying the similar activity as commercial Ni catalysts.

Keywords: hydrogenation, vegetable oil, palladium, mesoporous carbon

1. Introduction

Substantial contribution to understanding of fats hydrogenation was made along the years by catalytic scientists working in Kazakhstan [1-5]. It was demonstrated that among the metals catalyzing oil hydrogenation only palladium, nickel and copper are sufficiently selective [1-6]. Nickel and copper tend to be dissolved in the oil to form corresponding salts which then catalyze aerobic oxidation of the oil which results in off-flavors in the product and product toxicity. Ecological friendly palladium catalysts are the most active in the wide range of temperature and hydrogen pressure, being, however, more expensive [7]. A series of studies has been reported showing that supported Pd catalysts containing low amounts of the metal are proved to be right compromise for commercial industrial application in the oil hydroprocessing [8-14].

The aim of this work is to develop the Pd/C catalyst for the commercial batch hydrogenation of vegetable oil to produce confectionary fat and to evaluate Pd catalyst behavior in comparison with Ni catalysts in laboratory and pilot scale before further introduction into production-scale operation.

2. Experimental

2.1. Materials

Sunflower oil (Novosibirsk oil plant, Russia) as well as palm oil (triglycerides of fatty acids) was dried by heating up to 373 K before hydrogenation in solvent free conditions. The distribution of methyl esters of fatty acids for tested oils is presented in the Table 1.

Carbon composite materials Sibunit is produced by pyrolitic carbon deposition on a granulated carbon black. Pre-granulated carbon black is covered by a fixed amount of pyrolitic carbon through condensation or chemical vapor deposition. During a subsequent activation stage with steam at 973–1123 K a part of carbon is removed by gasification. A sponge-like system including meso- and macropores whose dimensions depend on the dispersion of the initial carbon black is formed. When the carbon black is burnt out almost completely the obtained granules lose their mechanical strength and destroy into shell-like fragments. Such procedure allows to control over a very wide range structural and textural properties (the specific surface area can be varied from 0.1 to 800 m2/g, the pore volume from 0.1 to 2.0 cm3/g). In the case of hydrogenation of fatty acids a system of wide pores and sufficient pore volume is able to provide efficient transport of oil and hydrogen molecules to palladium particles.

Table 1. Distribution of free fatty acids in sunflower and palm oils

 

Type

of oil

Methyl esters of fatty acids

composition, mass.%

Iodine value,

gI2/100 g

С 16:0

С 18:0

С 18:1

С 18:2

sunflower

6.6

4.4

17.7

71.3

138

palm

41.1

4.4

41.8

11.2

55


The samples of the carbon support of 40-70 ?m, 70-100 ?m, 100-180 ?m, 180-250 ?m were wet sieved after grinding to prepare 0.1 or 1 wt.% Pd/C (Table 2), some of them being characterized by nitrogen adsorption. The surface area was calculated using the BET equation, while the pore distribution in the support fractions 40-70 ?m, 70-100 ?m, 100-180 ?m, was calculated by Barrett-Joyner-Halenda (BJH) method. Part of Sibunit carbon support 40-70 ?m and 40-100?m samples was activated by a treatment with 38 wt.% HCl aqua solution at 321 K for 3 h to purify from possible metallic impurities followed by washing with distilled water and drying in air at 373 K. The Pd on Sibunit catalyst samples were prepared by depositing palladium hydroxide obtained by hydrolysis of palladium chloride at pH 8-10 [12]. Table 1 features Pd/C catalyst samples prepared with different size of the support size, metal loading and pretreatment by HCl (Table 1).

Some samples of Pd catalyst were analyzed by CO impulse technique using Autochem 2910 apparatus (Micromeritics) and TEM (JEM 2010 (JEOL, Japan) [15]. The benefits using mesoporous synthetic carbon as a support material in catalytic hydrogenation of fatty acids is that this material is mechanically stable and due to its mesopores large organic compounds exhibit better accessibility to the active sites compared to the microporous support material [11].

2.2. Catalytic experiments

Laboratory scale hydrogenation was performed in a three-phase 150 ml stainless steel autoclave with automatic recording of hydrogen uptake. A stirring rate of 1000 rpm was used in the experiments to avoid mass transfer limitations. The catalyst/oil weight ratio was 1/100?1/1000, while the amount of substrate was equal to 13-15 g in laboratory tests. The experiments were conducted at 413 K and hydrogen pressure of 9 bar. The product composition was analyzed chromatographically (“Tsvet -500”) using 15 m - 0.25 mm - 0.5 mm quartz capillary column, Carbowax-DVB and a flame ionization detector (FID) operating at 523 K. In the gas chromatographic analysis di-unsaturated acids, mono-unsaturated acids and saturated acids were calculated separately. In addition hydrogen consumption from a calibrated reservoir was measured.

Table 2. Prepared catalysts samples

Catalyst

Pd, %

Support fraction, ?m

Treatment by HCl

K-1

1

40-71

no

К-2

1

40-71

yes

К-3

1

71-100

no

К-4

0.1

40-71

no

К-6

1

180-250

no

К-8

1

100-180

no

К-9

1

40-100

yes

K-10

0.1

40-100

yes

K-11

0.05

40-100

yes


The pilot tests were conducted in the 5 L batch set up at 468 K and hydrogen pressure 1-2 bar. The weight ratio of catalyst/oil was 1/1000-1/2000, while the amount of substrate was equal to 3 kg. The hydrogenation products were first esterified with subsequent analysis by GC (“Agilent”) using 60 m - 0.25 mm - 0.5 mm quartz capillary column (HP-DB) and a flame ionization detector (FID) operating at 523 K. The samples were periodically taken from the reaction and the content of solid triglycerides at the certain predetermined analysis temperatures (283, 288, 293, 298, 303 K) was measured by NMR 1H (GOST Р 52179-2003, Minispec analyzer РС 120). For the determination of trans-isomer content infrared transmission spectra were recorded in the 400-6000 sm-1 region with 4 sm-1 resolution using spectrometer Shimadzu FTIR 8300 equipped with the diffusion reflection accessories DRS-8000 [12].

3. Results


3.1. Characterization

The BET specific surface areas of the support 40-70 ?m, 70-100 ?m, 100-180 ?m were determined by nitrogen adsorption. The mesoporous volume of the supports ranging from 2 to 100 nm is given in Table 3. The fraction above 100?m was the most mechanically stable after grinding which is a consequence of its lower surface area.

The metal dispersion according to CO chemisorption for 1% Pd/C was 38%. The average metal particle sizes for the fresh and spent 1% Pd/C catalysts with the support fractions 40-70 ?m and 70-100 ?m were 2.5-2.7 nm according to TEM measurements indicating that the metal dispersion remained constant during the catalytic experiments (not shown).

 

Table 3. The BET specific surface areas and relative volumes for the support materials measured by nitrogen adsorption

 

Fraction, ?м

SBET, m2/g

Vtotal, cm3/g

Vmesopor, cm3/g

(2-100 nm)

40-70

405

0.64

71.7

70-100

398

0.62

67.8

100-180

304

0.57

70.6

3.2. Effect of mass transfer

The effect of the catalyst to sunflower oil weight ratio was tested in order to evaluate the influence of gas-liquid mass transfer. When dissolution of hydrogen is not limiting the reaction rate the latter is proportional to the catalyst mass. As follows from Figure 1 the increase in the catalyst load from 0.001 to 0.005 in the reactor proportionally increased the productivity. Since hydrogen uptake at the catalyst to sunflower oil weight ratio equal to 0.01 is marginally larger than at 0.005 gas-liquid mass transfer cannot be completely ruled out at higher ratios. On the other hand as will be demonstrated below catalytic activity even at the ratio of 0.01 was dependent on the support pretreatment by HCl pointing out that gas liquid mass transfer cannot be solely responsible for observed catalytic data even in this case.

Figure 1. Effect of the catalyst (К-3) to sunflower oil weight ratio on hydrogenation rate. Reaction conditions: Т = 314 K, РH2 = 9 bar, m (Pd/C) / m (oil) = 0.001?0.01.

Catalytic data for hydrogenation of sunflower oil performed over 1 wt.% Pd/C (Sibunit) catalyst (fraction 40-70 ?m) using different stirring speed of 300, 650 и 1000 rotations per minute are presented in Figure 2. It was shown that the initial reaction rate does not depend on the stirring speed thus external mass transfer limitations could be neglected.

Further experiments were performed with the stirring speed of 1000 min-1. The impact of internal diffusion was elucidated by hydrogenating sunflower oil over 1 wt.% Pd/C (Sibunit) catalyst using four different support fractions 40-70 ?m, 70-100 ?m, 100-180 ? m and 180-250 ?m at 413 K under hydrogen pressure 9 bar. Catalyst activity and stability during the hydrogenation of fatty acids has a minor dependence on the catalyst particle size for the range 40-180 ?m, whereas decreasing for 180-250 ?m at the catalyst to oil amount ratio equal to 0.01 (Figure 3).

Figure 2. Effect of the stirring speed on hydrogenation rate over 1 wt.% Pd/C (К-1). Reaction conditions: Т = 314 K, Рн2 = 9 bar, m (Pd/C) / m (oil) = 0.01

Figure 3. Effect of catalyst particle size on initial hydrogenation rate of sunflower oil over 1 wt.% Pd/C (catalysts К-1, К-3, К-6, К-8). Reaction conditions: Т = 314 K, Рн2 = 9 bar, m (Pd/C) / m (oil) = 0.01; * m (Pd/C) / m (oil) = 0.005.

Figure 4. Effect of support treatment by HCl in

hydrogenation of sunflower oil over К-1 and К-2. Reaction conditions: Т = 314 K, Рн2 = 9 bar, stirring speed = 300 min-1, m (Pd/C) / m (oil) = 0.01.

3.3. Effect of HCl pretreatment

As it was stated above pretreatment with HCl can remove impurities from the support, which will eventually improve the catalyst behavior. The catalytic results are given on Figure 4, confirming the beneficial influence of HCl treatment on catalyst performance.

3.4. Catalyst activity and stability in catalytic runs

Table 4 contains catalytic data generated in the repeated experiments (see also section 3.3) when a series of experiments with oil were conducted in a consecutive manner, e.g. performing filtration the catalyst after hydrogenation experiments and using it in the following batches. Hydrogen uptake was very similar and there were minor fluctuations in the initial hydrogenation rates, confirming high stability of the catalyst against deactivation.

Overall during 21 catalytic runs 321.753 g of hydrogenated oil was obtained using 1.79?10-4 g Pd, which corresponds to 1797.5 kg per 1 g Pd.

3.5. Comparison with Ni catalysts

Since one of the aims was to compare palladium catalysts, developed in this study with the catalysts available commercially a comparison between them was performed.

Product melting points dependences on reaction time for K-9 and commercial samples of Ni catalysts are illustrated in Figure 5. It should be kept in mind that the metal loading of palladium in the tested samples (1.0 wt.%) is significantly lower than the nickel load (ca. 21% of Ni and NiO) pointing out on superior intrinsic activity of palladium.

Such promising results encouraged the authors to perform pilot plant tests at the industrial plant site, which are discussed in the following section.

3.6. Pilot tests

The pilot tests were performed in autoclave, which installed at the plant site and operated at Рн2= 2 bar and the stirring speed 300 min-1. The melting point of the hydrogenation product depends on the extent of saturation (presence of saturated acids) as well as on the trans/cis ratio in mono-unsaturated acids. In industry the content of solid triglycerides is evaluated at predetermined temperatures used for analysis. For instance in the reaction mixture after 1 h (Figure 6) the content of solid substances is rather low and the mixture has a melting point of ca. 301 K.

During the progress of the hydrogenation reaction the melting point is increasing, illustrated by the observation on Figure 5 that after 4 h of reaction the melting point of the product is clearly above 313 K, as at this temperature the content of solid triglycerides is 100%.


Table 4. Composition of free fatty acids during catalytic runs over 0.1 wt.% Pd/C (К-4).

Reaction conditions: Т = 314 K, Рн2 = 9 bar

Run

Н2 uptake,

mol

Н2/mol oil

Composition of free fatty acids

Iodine value,

gI2/100 g

Initial hydrogenation rate W0H2 ,

mol/(gPd•s)

С 16:0

С 18:0

С 18:1

С 18:2

sunflower oil

-

6.6

4.4

17.7

71.3

138

-

1а

2.45

6.9

18.8

71.7

2.6

66

0.35

2а

2.47

7.1

17. 8

72.6

2.5

67

0.36

3а

2.47

7.0

16.0

74.0

2.9

68

0.31

4а

2.47

7.1

18.0

73.1

1.7

66

0.49

5а

2.50

6.9

19.0

72.6

1.5

65

0.55

6а

2.48

7.0

17.9

73.3

1.8

66

0.55

7а

2.47

6.7

19.1

71.5

2.7

66

0.66

8а

2.33

7.5

14.0

76.8

1.7

68

0.37

9а

2.47

6.8

18.7

72.0

2.5

66

0.51

10а

2.45

7.4

17.4

72.4

2.8

67

0.47

11а

2.49

7.1

18.5

71.6

2.9

66

0.34

12а

2.42

6.8

19.3

71.1

2.9

66

0.57

13а

2.45

6.7

17.3

72.2

3.9

68

0.58

14а

2.47

6.5

16.0

72.8

4.6

70

0.51

15а

2.01

6.8

11.0

74.8

7.4

77

0.36

16b

2.36

6.9

14.9

72.8

5.5

72

1.05

17b

2.43

6.9

17.8

69.4

5.9

70

1.34

a m (Pd/C) / m (oil) = 0.01; b m (Pd/C) / m (oil) = 0.001.


Table. 5. Catalyst activity in pilot scale hydrogenation. Reaction conditions: Т = 570 K, Рн2 = 2 bar, stirring speed= 300 min-1


Catalyst

Catalyst/oil

Composition of free fatty acids

Iodine value,

gI2/100 g

Trans isomer content, % (by FTIR)

Reaction

time, h

С 16:0

С 18:0

С 18:1

С 18:2

K-9

3.0 g /3 kg

6.6

6.7

82.0

3.0

79

53.2

2

K-9а

1.5 g /3 kg

6.7

5.1

79.1

7.1

84

51.4

1.5

K-9b

1.0 g /2 kg

6.7

5.2

80.7

5.6

83

46.2

1.0

K-9b

1.0 g /2 kg

6.7

10.8

79.1

1,9

75

51.2

1.5

Nyss 800

6.0 g /3 kg

6.7

13.0

77.6

1.1

72

49.1

2.3

Nyss 210

6.0 g /3 kg

6.7

8.3

82.0

1.5

77

57.0

4.0

K-9c

2.0 g /3 kg

17.7

4.7

67.5

8.1

75

35.8

1

a 1st catalytic run; b 2d catalytic run; c mixture of sunflower oil (70%) and palm oil (30%).


Figure 5. Dependence of the reaction mixture melting point on hydrogenation time. Reaction conditions: Т = 570 K, РH2 = 2 bar, stirring speed = 300 min-1,

m (K-9) / m (oil) = 0.0005, m (Nissosel800) / m (oil) = 0.0072, m (K-9) = 1.5 g, m (Nyssosel800) = 2.16 g,

m (Nyssosel210) = 6.00 g, m (oil ) = 3.0 kg.


Figure 6. Distribution of solid triglyceride content

(determined by NMR) versus triglyceride heating temperature in the hydrogenation product with reaction time over K-9 catalyst. Reaction conditions: Т = 570 K, РH2 = 2 bar, stirring speed = 300 min-1, m (Pd/C) / m (oil) = 0.001, m (K-9) = 3.0 g, m(oil) = 3.0 kg.

The results for tests with palladium catalysts are given in Table 5 along with the data for nickel catalysts, demonstrating that similar iodine values (corresponding to a similar degree of unsaturation) can be achieved for palladium catalysts, as with commercial nickel counterparts. At the same time the trans isomer content after Pd hydrogenation does not exceeds that for Ni.

Conclusions

Catalytic selective hydrogenation of sunflower as well as the mixture of sunflower and palm oil was studied over Pd nanocomposite carbon (Sibunit) catalysts. The catalysts, containing palladium, showed activity similar or superior to commercial nickel containing samples in laboratory and pilot tests, being not only environmentally benign but also displaying stable behavior in the repeated tests.

Acknowledgements

The authors thank A.V. Predybailo, A.V. Shakhov and Yu.Ya. Savchenko for assistance in Pd/C and Ni catalysts pilot tests. The part of research work was supported by RFBR Grant № 08-03-91758


References

1. Zhubanov K.A., Sokolskii D.V., Gidrirovanie rastitelnyx zhirov (Hydrogenation of plant fats). Alma-Ata, Nauka. 1972. 182 p.

2. Zhubanov K.A., Shalabaev M.Sh., Slepov S.K., Krupenya N.G., Safronov V.M., Izv. Akad. Nauk Kazakh. SSR, Seriya Khim. 1976. V. 26. P. 52.

3. Buvalkina L.A., Tkacheva G.D., Zhubanov K.A., Izv. Akad. Nauk Kazakh. SSR, Seriya Khim. 1980. V. 5. P. 24.

4. Sheludyakov Y.L., Khabibullin F.K., Zhubanov K.A., Tomilov A.G., Gimadi S., Izv. Nats. Akad. Nauk Resp. Kazakhstan, Ser. Khim. 2006. V. 1. P. 55.

5. Sheludyakov Y.L., Khabibullin F.K., Zhubanov K.A., Tomilov A.G., Gorlachev I.D., Euras. Chem.-Technol. J. 2007. V. 9. P. 199.

6. Gray J.I., Russel L.F., J. Am. Oil Chem. Soc. 1979. V. 56. P.36.

7. Parry J.D., Winterbottom J.M., J. Chem. Tech. Biotechnol. 1991. V. 50. P. 67.

8. Patent RF 2105050, 1996.

9. Semikolenov V.A., Simakova I.L., Sadovnichii G.V., Chimicheskaia promyshlennost (Russian Chemical Industry). 1996. V. 3. P. 184.

10. Perez-Cadenas A.F., Zieverink M.M.P., Kapteijn F., Moulijn J.A., Carbon, 2006. V. 44. P. 173.

11. Simakova I.L., Simakova O., Romanenko A.V., Murzin D.Yu., Ind. Eng. Chem. Res. 2008. V. 47. P. 7219.

12. Simakova O.A., Simonov P.A., Romanenko A.V., Simakova I.L., React. Kinet. Catal. Lett., 2008. V. 95. P. 3.

13. Patent RF 2318868.

14. Patent RF 2323046.

15. Simakova I.L., Simakova O.A., M?ki-Arvela P., Simakov A.V., Estrada M., Murzin D.Yu., Appl. Catal. A: Gen. 2009. V. 355. P. 100.