Transesterification of castor oil with trimethylchlorosilane: simultaneous formation of fatty acid alkyl esters and α-monochlorohydrin

Abstract

The reaction between castor oil and various alcohols in the presence of trimethylchlorosilane (TMSCl) as acidic mediator provides two remarkable one pot reactions: transesterification of triglycerides, relevant to BioDiesel production, and chlorination of glycerol, with formation of α-monochlorohydrin as the predominant product, relevant to the synthesis of epoxide resins. The reaction has been thoroughly investigated to optimize the reaction conditions and separation of the products obtained in good yields. This synthesis represents a highly sustainable production of industrial commodities from available biomass.

---

1. Introduction

BioDiesel (BD) is a well known commodity, basically consisting of a mixture of fatty acid methyl (FAME) or ethyl (FAEE) esters. There is a lot of literature on BD production methods^1–4 and nowadays the most used process is transesterification of triglycerides (vegetable oil or animal fat) with a basic catalyst, such as sodium methoxide (or ethoxide), but also acid catalysis is used.

In a worldwide scale, BD production deals with economic, political, environmental and ethical issues. In fact, despite BD shows many advantages with respect to fossil fuels, mainly biodegradability and CO₂ compensation, there are some significant drawbacks in BD production, such as the so called “food versus fuel dispute”³ and the high quantity of glycerol, obtained as co-product of the reaction (about 10% w/w). The latter is mostly a market problem associated to the employment of the oversupply of glycerol coming from the BD production. In this context glycerol could be converted into additives for BD, like glycerol tert-butyl ether, or solvents like glycerol carbonates, otherwise it becomes a cheap raw material for the production of fine chemicals (chlorohydrins and epichlorohydrin, acrolein, glyceraldehyde).^4–9

By the way, the main issue deals with the ethical problem of foodstuff destination. In order to avoid BD production from edible oils (such as soybean, rapeseed and sunflower oil), many efforts have been made to convert feedstock such as non-edible oils (jatropha, karanja or mahua oil), exhausted fried oils, oils from microalgae and waste animal fats into BD.^3,4,10

Among non-edible oils, castor oil seems to be a good candidate for BD production.^11–17 It is obtained from the seeds of Ricinus communis plant. India is the world's largest exporter of castor oil,¹⁸ followed at considerable distance by China, Brazil and Thailand.¹⁹ Nevertheless castor plant is a perennial crop that tolerates variable weather conditions and soil types, so it could be cultivated also in countries where this plant is not autochthonous. Recently Portugal made a preliminary assessment of castor plant for BD production.¹⁶

Castor oil contains high percentage (up to 90%) of ricinoleic acid [(R)-12-hydroxy-cis-9-octadecenoic acid] (see Fig. 1), responsible for castor oil peculiar properties, such as higher hygroscopy, viscosity and alcohol solubility than other oils.^11,15,16 These properties render the use of pure castor BD as fuel unpractical if not in blends with reference diesel.²⁰

Fig. 1 Ricinoleic acid and ricinolein, the prevalent triglyceride in castor oil.

Castor oil is currently used itself as lubricant, functional fluid and industrial process oil. Castor oil can be also processed to obtain drying oils, surfactants, cosmetics, oil for inks and plasticize coatings. Sebacic acid can be produced from castor oil to synthesize polyamides (i.e. nylon 6,10) partially derived from renewable material. More recently a great focus has been made on polyurethanes derived from castor oil, since it is the only commercially available natural polyhydroxylated oil.^18,19

The interest in BD production from castor oil raised in the last decade. The transesterification reaction (with methanol and ethanol) has been studied both with basic and acid catalysts.^{11–17,21–23} For base-catalyzed process, studies concerning yield optimization, have been also performed with statistical approaches, like the surface method.^12,13,22 Encinar and coworkers¹⁴ accomplished the reaction in basic conditions with ultrasonic radiation. The frequency and supplied power dissipation affect the reaction rate, but not the final yield.

Unfortunately, castor oil suffers for a lower conversion with respect to common vegetable oils like cotton²¹ or soybean²³ oil when processed in the same basic conditions. On the other hand, the higher solubility of castor oil in alcoholic medium allows the reactions to occur in a monophasic system and kinetic studies revealed comparable reaction rate for acid- and base-catalyzed processes.¹¹ In general acid-catalyzed reactions are competitive with respect to the basic ones, because they can tolerate the presence of free fatty acids in the starting materials.

Recently we developed a new process for BD production by transesterification of sunflower oil, exhausted oils, and animal fats in the presence of trimethylchlorosilane (TMSCl) as acid mediator.²⁴ TMSCl in our process must be used in large amounts, not as a simple catalyst, but its hydrolysis product, the hexamethyldisiloxane (TMSOTMS), is quantitatively recovered and can be converted back into TMSCl by literature simple procedures.²⁵

In this context, the present work is focused on the application of the “trimethylchlorosilane method” to the production of BD from castor oil.

2. Experimental

2.1. Materials

All reagents were purchased from Sigma-Aldrich with the exception of castor oil purchased from Herboris Orientis Dacor. Trimethylchlorosilane was distilled before use and purity was checked by ¹H-NMR analysis. Methanol, ethanol and n-butanol were previously dried with magnesium.²⁶ All other reactants were used without any further purification.

2.2. Analytical methods

NMR spectra have been recorded with Varian Mercury plus 400 and Varian VXR 200 instruments. ¹³C-NMR spectroscopy has been used as semi-quantitative method to determine the molar ratio between chlorinated products and residual glycerol, since the CH signals of glycerol, 1-MCH and 1,3-DCH are easily recognizable in the spectra (Fig. 2).⁹

Fig. 2 A typical ¹³C-NMR spectrum (100 MHz) in DMSO-d₆ of glycerol and chlorohydrins, used for semi-quantitative analysis.

Molar fraction of estolide (χ_E) was estimated via ¹H-NMR taking into account the integrals of signals of H-12′ and H-9 by the following equation:

where E and F are the integrals of signals respectively of H-12′ and H-9 and r is the molar percentage of ricinoleic acid in castor oil, which was evaluated via ¹H-NMR by comparing the integrals of H-12 signal at 3.60 ppm with that of the terminal CH₃ at 0.87 ppm, resulting into 87%. Beside this method could be affected by the presence of oligomers and cyclic forms of estolides, that could change the relation between E and F, there is a quite good agreement between the quantity of estolides calculated by NMR and by GC analyses. So the equation can be used for a qualitative estimation.

GC analyses were performed, according to previous work,²⁴ on a Shimadzu GC-2010 gas chromatograph equipped with a programmable temperature vaporizing (PTV) injector, a flame ionization detector (FID) and a Zebron ZB-5 HT INFERNO capillary column (5% phenyl–95% polydimethylsiloxane, 15 m length, 0.32 mm internal diameter, 0.1 μm film thickness). Oven conditions: isotherm at 50 °C (2 min), ramp at 15 °C min⁻¹ to 180 °C, isotherm at 180 °C (10 min), ramp at 25 °C min⁻¹ to 380 °C, isotherm at 380 °C (1 min). Gas carrier was helium, with constant flow rate of 1.83 mL min⁻¹. PTV conditions: ramp at 200 °C min⁻¹ from 50 °C to 300 °C and isotherm at 300 °C (10 min). FID temperature set at 385 °C. Quantitative analysis, preparing samples with internal standard, were performed only for BD samples coming from Kugelrohr distillation. A sample volume of 1 μL was injected using a split mode, with a split ratio of 10 : 1. Samples have been prepared mixing 300 μL of a solution of BD (ca. 250 mg in 10 mL of ethyl acetate) with 300 μL of a methyl heptadecanoate solution (ca. 25 mg in 5 mL of ethyl acetate). The purity of BD was detected according to the equation:

where C is the content of methyl esters, expressed as a mass percentage; and A_STD are respectively the total peak area of the methyl esters and the area of methyl heptadecanoate; M_SAMPLE and M_STD are respectively the mass of sample and the mass of internal standard in the final solution. For all other reaction mixtures, coming from trials performed according to procedures A and B, only semi-quantitative analyses were done to determine BD and estolides content, preparing samples without internal standard.

Accurate mass spectra were recorded on a LTQ-Orbitrap high-resolution mass spectrometer (Thermo, San Jose, CA, USA), equipped with a conventional ESI source.

2.3. Optimization of the reaction conditions (see Table 1)

For all the reactions the molar weight of castor oil has been considered an average value of 929 g mol⁻¹ as reported in the literature.¹² All the reactions reported in Table 1 refer to a complete disappearance of starting triglycerides as determined by ¹H-NMR.

Table 1 Reactions of castor oil following procedures A and B: reaction conditions and analyses

Entry (procedure)	Step	MeOH^a	TMSCl^a	Temp	Time	χE^b	Glycerol^c	1-MCH^c	1,3-DCH^c	Esters^d	Glycerides^d	Estolides^d
a Molar ratio with respect to starting castor oil.b χE molar fraction of estolides evaluated by ^1H-NMR.c Relative percentages obtained via ^13C-NMR analyses based on the CH signals of 1-MCH, 1,3-DCH, and eventually glycerol.d Relative percentages determined by GC analyses based on the chromatographic peaks of the ‘ester fraction’. The data were used only for semi-quantitative purpose, because the determination of estolides was not in complete accordance with the gravimetric data obtained by distillation.e 10% mol mol^−1 of acetic acid has been added to the mixture.
1 (A)	1	5.0	3.0	rt	8 h	6%	25%	75%	0%	87%	4%	8%
2 (A)	1	4.9	3.0	60 °C	8 h	19%	0%	85%	15%	76%	2%	22%
3 (A)	1	10.0	3.0	60 °C	8 h	12%	19%	81%	0%	81%	5%	14%
4 (A)	1	22.4	3.0	60 °C	8 h	6%	52%	48%	0%	85%	3%	13%
5 (A)	1	45.3	3.0	60 °C	8 h	3%	83%	17%	0%	96%	>1%	3%
6 (A)	1	44.8	3.0	100 °C	8 h	3%	74%	26%	0%
7 (A)^e	1	45.0	3.0	60 °C	8 h	3%	83%	17%	0%
8 (B)	1	4.7	2.2	60 °C	8 h	19%	7%	82%	11%	76%	3%	21%
	2	45.0	2.2	60 °C	+3 h	11%	2%	88%	11%	89%	1%	10%
9 (B)	1	4.5	3.0	60 °C	5 h	15%	0%	82%	18%	77%	9%	14%
	2	47.4	3.0	60 °C	+3 h	5%	0%	81%	19%	92%	1%	7%
10 (B)	1	4.5	3.0	60 °C	8 h	22%	0%	83%	17%	80%	1%	19%
	2	45.3	3.0	60 °C	+6 h	7%	0%	75%	25%	93%	>1%	6%
11 (B)	1	4.4	5.0	60 °C	5 h	15%	0%	73%	27%	75%	7%	17%
	2	45.2	5.0	60 °C	+3 h	3%	0%	58%	42%	94%	1%	5%
12 (B)	1	4.5	5.0	100 °C	5 h	26%	0%	76%	24%	70%	3%	26%
	2	45.0	5.0	100 °C	+3 h	15%	0%	63%	37%	81%	2%	17%

---

2.3.1. Procedure A. In a typical experiment, under nitrogen atmosphere, castor oil (0.5 g, 0.54 mmol) was placed in a 15 mL screw-cap Pyrex® tube, methanol and TMSCl (purity 94% w/w by ¹H-NMR) were added in the reported molar ratios and the mixture was kept at the desired temperature for a variable time (for reaction conditions see Table 1). At the end of the reaction, the mixture was cooled to room temperature. Methanol, TMSOTMS, and excess TMSCl were removed at low pressure.† NMR analyses were performed both in CDCl₃ and DMSO-d₆ as solvents, semi-quantitative GC analyses were also carried out to determine the conversion of castor oil into methyl esters.

2.3.2. Procedure B. After the first reaction performed according to procedure A, a further amount of methanol was added to obtain the desired ratio with respect to the initial castor oil, as reported in Table 1, and the resulting mixture was allowed to react for a further time at the reported temperature. At the end of the reaction, the mixture was cooled to room temperature, methanol and volatile compounds were evaporated at low pressure, and analyses were performed as reported for procedure A.

2.4. Synthesis of esters of fatty acids from castor oil

2.4.1. Reaction of castor oil with methanol: synthesis of fatty acid methyl esters. Castor oil (2.015 g, 2.17 mmol), methanol (0.39 mL, 9.61 mmol) and TMSCl (0.87 mL, 6.44 mmol) were placed, under nitrogen atmosphere, in a 25 mL screw-cap Pyrex® glass tube. The mixture was allowed to react under continuous stirring at 60 °C. After 8 hours, the reaction was stopped and cooled to room temperature. A further amount of methanol (2.6 mL, 86.5 mmol) was added and the mixture was kept for further 6 hours at 60 °C. At the end of the reaction methanol and the other volatile compounds were removed by evaporation at low pressure obtaining a yellow liquid (2.232 g). Distillation under vacuum (2.7 Pa) in a Kugelrohr apparatus allowed to recover four fractions. The fraction collected at 100 °C (0.095 g) was a 19 : 1 mixture (¹³C-NMR analysis) of α-monochlorohydrin (3) (yield 37%) and α,γ-dichlorohydrin (4) (yield 2%). Between 170 and 240 °C two fractions were collected. The minor one (0.201 g) was a mixture of BD and 1-MCH (3), containing about 27% mol mol⁻¹ (or 49% w/w) of BD (evaluated by ¹H-NMR). The major fraction (1.649 g; yield 81%; GC purity 91.0%) consisted of methyl ricinoleate (2a)²⁷ as predominant product; ¹H-NMR (400 MHz, CDCl₃) δ (ppm) 5.54 (m, 1H; H-9), 5.39 (m, 1H; H-10), 3.65 (s, 3H; OCH₃), 3.60 (m, 1H; H-12), 2.29 (t, 2H; H-2), 2.20 (m, 2H; H-11), 2.03 (m, 2H; H-8), 1.60 (m, 2H; H-3), 1.45 (m, 2H; H-13), 1.39–1.21 (m, 16H; CH₂ chain), 0.87 (m, 3H; H-18); ¹³C-NMR (100 MHz, CDCl₃) δ (ppm) 174.3 (s, C-1), 133.3 (d, C-9), 125.2 (d, C-10), 71.5 (d, C-12), 51.4 (q, OCH₃), 36.8 (t, C-13), 35.3 (t, C-11), 34.05 (t, C-2), 31.8 (t), 29.5 (t), 29.3 (t), 29.1 (t), 29.05 (t), 27.3 (t, C-8), 25.7 (t), 24.9 (t, C-3), 22.6 (t), 14.0 (q, C-18); HRMS (ESI) m/z calcd for C₁₉H₃₆NaO₃ [MNa]⁺ 335.2557, found 335.2562.

The distillation residue (0.229 g; yield 12%) was a brownish liquid, containing 5a as the main product; ¹H-NMR (400 MHz, CDCl₃) δ (ppm) 5.51 (m, 1H; H-9), 5.47–5.33 (m, 2H; H-9′, H-10), 5.29 (m, 1H; H-10′), 4.85 (m, 1H; H-12′), 3.63 (s, 3H; OCH₃), 3.58 (m, 1H; H-12), 2.30–2.21 (m, 6H; H-2, H-2′, H-11′), 2.20 (m, 2H; H-11), 2.05–1.93 (m, 4H; H-8, H-8′), 1.63–1.53 (m, 4H; H-3, H-3′), 1.49 (m, 2H; H-13′), 1.43 (m, 2H; H-13), 1.37–1.17 (m, 32H; CH₂ chain), 0.84 (m, 6H; H-18, H-18′); ¹³C-NMR (100 MHz, CDCl₃) δ (ppm) 174.3 (s, C-1′), 173.5 (s, C-1), 133.3 (d, C-9), 132.5 (d, C-9′), 125.2 (d, C-10), 124.3 (d, C-10′), 73.7 (d, C-12′), 71.5 (d, C-12), 51.4 (q, OCH₃), 36.85 (t, C-13), 35.4 (t, C-11), 34.7 (t, C-11′), 34.1 (t, C-2′), 33.6 (t, C-13′), 32.0 (t, C-2), 31.8 (t), 31.7 (t), 29.6 (t), 29.5 (t), 29.3 (t), 29.2 (t), 29.15 (t), 29.1 (t), 27.4, 27.3 (t, C-8, C-8′), 25.7 (t), 25.3 (t), 25.1, 24.9 (t, C-3, C-3′), 22.6 (t), 22.55 (t), 14.1, 14.0 (q, C-18, C-18′).

2.4.2. Reaction of castor oil with ethanol: synthesis of fatty acid ethyl esters. Operating as above, castor oil (2.023 g, 2.18 mmol), ethanol (0.58 mL, 10.2 mmol) and TMSCl (0.88 mL, 6.52 mmol) were allowed to react under continuous stirring at 60 °C. After 8 hours, ethanol (5.25 mL, 92.1 mmol) was added and the mixture was kept for further 6 hours at 60 °C. Evaporation of volatiles compounds left a dark yellow liquid (2.388 g) tha was subjected to Kugelrohr distillation under vacuum (2.7 Pa) affording three fractions. The fraction collected until 150 °C (0.206 g) was a 85 : 15 mixture (¹³C-NMR analysis) of α-monochlorohydrin (yield 73%) and α,γ-dichlorohydrin (yield 11%). The major fraction was collected between 170 and 240 °C (1.698 g; yield 80%; GC purity 91.3%) was mainly constituted of ethyl ricinoleate (2b) ¹H-NMR (400 MHz, CDCl₃) δ (ppm) 5.53 (m, 1H; H-9), 5.39 (m, 1H; H-10), 4.10 (q, 2H; OCH₂), 3.59 (m, 1H; H-12), 2.26 (t, 2H; H-2), 2.19 (m, 2H; H-11), 2.02 (m, 2H; H-8), 1.59 (m, 2H; H-3), 1.44 (m, 2H; H-13), 1.36–1.20 (m, 18H; CH₂ chain + CH₃), 0.86 (m, 3H; H-18); ¹³C-NMR (100 MHz, CDCl₃) δ (ppm) 173.8 (s, C-1), 133.2 (d, C-9), 125.2 (d, C-10), 71.5 (d, C-12), 60.1 (t, OCH₂), 36.8 (t, C-13), 35.3 (t, C-11), 34.3 (t, C-2), 31.8 (t), 29.5 (t), 29.3 (t), 29.1 (t), 29.0 (t), 27.3 (t, C-8), 25.7 (t), 24.9 (t, C-3), 22.5 (t), 14.2 (q, CH₃), 14.0 (q, C-18); HRMS (ESI) m/z calcd for C₂₀H₃₉O₃ [MH]⁺ 327.2894, found 327.2898. The distillation residue (0.426 g; yield 19%) was a brownish liquid, mainly containing 5b ¹H-NMR (400 MHz, CDCl₃) δ (ppm) 5.55 (m, 1H; H-9), 5.50–5.36 (m, 2H; H-9′, H-10), 5.31 (m, 1H; H-10′), 4.88 (m, 1H; H-12′), 4.12 (q, 2H; OCH₂), 3.60 (m, 1H; H-12), 2.31–2.23 (m, 6H; H-2, H-2′, H-11′), 2.20 (m, 2H; H-11), 2.09–1.96 (m, 4H; H-8, H-8′), 1.65–1.56 (m, 4H; H-3, H-3′), 1.52 (m, 2H; H-13′), 1.45 (m, 2H; H-13), 1.37–1.20 (m, 35H; CH₂ chain + CH₃), 0.88 (m, 6H; H-18, H-18′); ¹³C-NMR (100 MHz, CDCl₃) δ (ppm) 173.8 (s, C-1′), 173.5 (s, C-1), 133.3 (d, C-9), 132.5 (d, C-9′), 125.2 (d, C-10), 124.3 (d, C-10′), 73.6 (d, C-12′), 71.5 (d, C-12), 60.1 (t, OCH₂), 36.8 (t, C-13), 35.4 (t, C-11), 34.6 (t, C-11′), 34.4 (t, C-2′), 33.6 (t, C-13′), 32.0 (t, C-2), 31.8 (t), 31.7 (t), 29.6 (t), 29.5 (t), 29.3 (t), 29.2 (t), 29.15 (t), 29.1 (t), 27.4, 27.3 (t, C-8, C-8′), 25.7 (t), 25.3 (t), 25.1, 24.9 (t, C-3, C-3′), 22.6 (t), 22.55 (t), 14.2 (q, CH₃), 14.1, 14.0 (q, C-18, C-18′).

2.4.3. Reaction of castor oil with n-butanol: synthesis of fatty acid n-butyl esters. Following the same procedure, castor oil (2.002 g, 2.16 mmol), n-butanol (0.88 mL, 9.60 mmol) and TMSCl (0.87 mL, 6.44 mmol) were allowed to react under continuous stirring at 60 °C. After 8 hours, n-butanol (8.0 mL, 87.2 mmol) was added and the mixture was kept for further 6 hours at 60 °C. Evaporation at low pressure of the volatile compounds gave a dark yellow liquid (2.489 g). Kugelrohr distillation under vacuum (1.3 Pa) afforded three fractions. The fraction collected until 150 °C (0.158 g) was a 19 : 1 mixture (¹³C-NMR analysis) of 1-MCH (yield 63%) and 1,3-DCH (yield 3%). The major fraction was collected between 200 and 240 °C (1.8670 g; yield 82%; GC purity 96.1%) and mainly consisted of n-butyl ricinoleate (2c). ¹H-NMR (400 MHz, CDCl₃) δ (ppm) 5.53 (m, 1H; H-9), 5.39 (m, 1H; H-10), 4.05 (t, 2H; OCH₂), 3.59 (m, 1H; H-12), 2.27 (t, 2H; H-2), 2.19 (m, 2H; H-11), 2.03 (m, 2H; H-8), 1.65–1.54 (m, 4H; H-3, OCH₂CH₂), 1.44 (m, 2H; H-13), 1.36 (m, 2H; CH₂CH₃), 1.36–1.20 (m, 16H; CH₂ chain), 0.91 (t, 3H; CH₃), 0.86 (m, 3H; H-18); ¹³C-NMR (100 MHz, CDCl₃) δ (ppm) 173.9 (s, C-1), 133.2 (d, C-9), 125.2 (d, C-10), 71.4 (d, C-12), 64.0 (t, OCH₂), 36.8 (t, C-13), 35.3 (t, C-11), 34.3 (t, C-2), 31.8 (t), 30.7 (t, OCH₂CH₂), 29.5 (t), 29.3 (t), 29.1 (t), 29.0 (t), 27.3 (t, C-8), 25.7 (t), 24.9 (t, C-3), 22.6 (t), 19.1 (t, CH₂CH₃), 14.0 (q, C-18), 13.6 (q, CH₃); HRMS (ESI) m/z calcd for C₂₂H₄₃O₃ [MH]⁺ 355.3207, found 355.3210. The distillation residue (0.400 g; yield 20%) was a brownish liquid mainly containing compound 5c ¹H-NMR (400 MHz, CDCl₃) δ (ppm) 5.54 (m, 1H; H-9), 5.50–5.35 (m, 2H; H-9′, H-10), 5.31 (m, 1H; H-10′), 4.87 (m, 1H; H-12′), 4.06 (t, 2H; OCH₂), 3.60 (m, 1H; H-12), 2.31–2.22 (m, 6H; H-2, H-2′, H-11′), 2.20 (m, 2H; H-11), 2.08–1.96 (m, 4H; H-8, H-8′), 1.65–1.55 (m, 6H; H-3, H-3′, OCH₂CH₂), 1.52 (m, 2H; H-13′), 1.46 (m, 2H; H-13), 1.42–1.20 (m, 34H; CH₂ chain + CH₂CH₃), 0.92 (t, 3H; CH₃), 0.86 (m, 6H; H-18, H-18′); ¹³C-NMR (100 MHz, CDCl₃) δ (ppm) 173.9 (s, C-1′), 173.5 (s, C-1), 133.3 (d, C-9), 132.5 (d, C-9′), 125.2 (d, C-10), 124.3 (d, C-10′), 73.6 (d, C-12′), 71.5 (d, C-12), 64.1 (t, OCH₂), 36.8 (t, C-13), 35.3 (t, C-11), 34.6 (t, C-11′), 34.4 (t, C-2′), 33.6 (t, C-13′), 32.0 (t, C-2), 31.8 (t), 31.7 (t), 30.7 (t, OCH₂CH₂), 29.6 (t), 29.5 (t), 29.3 (t), 29.15 (t), 29.1 (t), 29.05 (t), 27.4, 27.3 (t, C-8, C-8′), 25.7 (t), 25.3 (t), 25.1, 25.0 (t, C-3, C-3′), 22.6 (t), 22.5 (t), 19.1 (t, CH₂CH₃), 14.05, 14.0 ppm (q, C-18, C-18′) 13.7 (q, CH₃).

3. Results and discussion

The reaction of castor oil with MeOH and a stoichiometric amount of TMSCl at room temperature provides the complete disappearance of triglycerides after 8 hours leading to an emulsion containing glycerol, methyl ricinoleate (2a), α-monochlorohydrin (1-MCH) (3), and minor amounts of side products (Table 1, entry 1). On the other hand performing the reaction at 60 °C for 8 hours, a homogeneous solution was obtained. NMR and GC analyses showed the complete disappearance of glycerol and the formation of methyl ricinoleate (2a) and 1-MCH (3) as predominant products, together with minor amount of α,γ-dichlorohydrin (1,3-DCH) (4) and methyl ricinoleate estolide (5a) (Scheme 1, Table 1, entry 2).

Scheme 1 Reaction of castor oil with alcohols using TMSCl as acid mediator, considering ricinolein as the main triglyceride.

These results, appear different from those observed in the reaction with sunflower oil, in particular for the complete conversion of glycerol. In fact, with sunflower oil, a two-phase system consisting of BD and a 1 : 1 mixture of glycerol and α-monochlorohydrin was obtained.²⁴ The different behavior of castor oil can be associated to its peculiar properties with respect to other vegetable oils, as previously stated.^11,15,16 Now, the higher solubility of the ricinoleic moieties in alcoholic medium allows the transesterification reaction in homogeneous phase. This situation is also responsible for the complete conversion of glycerol into α-monochlorohydrin (1-MCH) and α,γ-dichlorohydrin (1,3-DCH) (Table 1, entry 2) favoring the reaction with TMSCl as chlorinating agent, as confirmed by ¹³C-NMR analysis of the reaction mixture. In fact, the ¹³C-NMR technique allowed a semi-quantitative analysis affording the relative percentages of compounds by comparison of the integrals related to the CH signals of glycerol and chlorohydrins. An example is reported in Fig. 2.

Moreover, the ¹H-NMR analysis of the reaction mixture revealed signals neither attributable to BD nor to chlorohydrins or glycerol esters (i.e. mono-, di- and triglycerides), that, on the basis of literature data, were assigned to ricinoleic estolides.^28–36

Estolides are a class of esters obtained from hydroxylated fatty acids and organic acid (generally a fatty acid). In particular ricinoleic acid gives estolides in which the OH group at C-12 position is esterified by another fatty acid. Otherwise estolides can be also synthesized from unsaturated fatty acid, like oleic acid, via oxidative process (i.e. with perchloric acid).²⁹ Estolides, like vegetable oils, are successfully used as lubricants and functional fluids in place of mineral oil, because of their better viscosity, lower toxicity and higher biodegradability. Moreover they have a lower pour point and higher resistance to thermal oxidation than vegetable oils, that could be suitable for the use as fuels.^29,30,34 Since ricinoleic acid is a non-toxic and natural hydroxylated fatty acid, estolides obtained from it are used also in the alimentary field as viscosity controllers for chocolate or as emulsifier for margarine.³⁶

However to provide the best selectivity of the process the present study was focused on the search of the best reaction conditions to convert castor oil into BD and chlorohydrins avoiding or at least limiting the estolides formation. Then, in the transesterification reaction with methanol, the first problem was to evaluate the amount of estolides with respect to FAME. This evaluation was achieved via ¹H-NMR analyses of the reaction crudes, by comparison with the data of samples obtained by distillation (see experimental). In Fig. 3 the upper ¹H-NMR spectrum refers to the fraction of FAME, with methyl ricinoleate (2a) as predominant component, while the second one is the spectrum of the distillation residue, in which the main product was the methyl ricinoleate estolide (5a). The detection of 5a in a reaction mixture is mainly ascertained by the presence of signals that are shifted with respect to those of 2a. In particular H-12′ shifts from 3.60 to 4.85 ppm, H-9′ from 5.54 to 5.43 ppm, H-10′ from 5.39 to 5.29 ppm, H-11′ from 2.20 to 2.27 ppm and H-13′ from 1.45 to 1.49 ppm.

Fig. 3 ¹H-NMR spectra (400 MHz) in CDCl₃ of methyl ricinoleate (top) and of the ricinoleate estolide of methyl ricinoleate (bottom) with signal attribution, according to literature.²⁸

Thereby, with some approximation, it is possible to estimate the molar fraction of estolides (χ_E) by ¹H-NMR analysis (see experimental). These data were also confirmed by semi-quantitative GC analyses (see Table 1).

With the aim to optimize the BD and chlorohydrins production, the castor oil transesterification was repeated under different reaction conditions. Initially a direct procedure (see procedure A in the experimental section) was applied simply by mixing all the reagents and heating at the desired temperature for 8 hours. By increasing the alcohol to castor oil ratio the formation of estolides was reduced (Table 1, entries 2–5). Nevertheless the high dilution of the oil dramatically affected the chlorination process, leading to minor amount of 1-MCH together with unreacted glycerol. Even rising the reaction temperature or adding a small amount of acetic acid, that was found to be an excellent catalyst for the glycerol chlorination with TMSCl,⁹ the yields of chlorohydrins were not improved (Table 1, entries 6 and 7). For these reasons, a different procedure (see procedure B in experimental section) was tested, performing the reaction in two steps: a first one, achieved with a minor amount of alcohol (methanol to oil ratio about 4.5) to favour glycerol chlorination and a second one, involving a significant methanol addition (final methanol to oil ratio about 45), to convert estolides into FAME. Different trials were made varying TMSCl to castor oil ratio, reaction times and temperatures (Table 1, entries 8–12).

The optimal conditions for the first step involved a 3 : 1 TMSCl/castor oil ratio and heating at 60 °C for 8 hours (Table 1, entry 10). In fact a lower TMSCl/castor oil ratio was not able to convert completely glycerol (Table 1, entry 8), while a higher ratio just increased the percentage of 1,3-DCH with respect to 1-MCH, reducing the selectivity of chlorination, either operating at 60 or 100 °C (Table 1, entries 11 and 12).

Concerning reaction time, 8 hours turned out to be necessary for the first step. For lower reaction times (Table 1, entry 9), GC analyses revealed the presence of peaks likely associated to products of partial transesterification, such as mono- and diglycerides, simply reported in Table 1 under the generic term of glycerides. Comparing GC chromatograms and ¹³C-NMR spectra (Fig. 4) of the reaction mixtures deriving from the first and second step of trial 9, the new peaks (in the frames) were likely associated to esters of fatty acids with 1,3-DCH. Nevertheless any further analyses were avoided, since these compounds disappeared increasing reaction times and/or methanol to oil ratios.

Fig. 4 ¹³C-NMR spectra (100 MHz) in DMSO-d₆ (right) of two sample coming from trial 9 respectively after 5 and 8 hours of reaction. On the left chromatograms of both samples compared with distilled methyl esters and estolides (obtained as residue of distillation).

The optimized reaction conditions (Table 1, entry 10) were applied to the transesterification of castor oil with different alcohols. All the reactions were performed at 60 °C in two steps (8 + 6 hours, see experimental section) and involved in the second one a further addition of the alcohol. After removal of the volatiles compounds, yellow-brownish liquids were obtained. Simple Kugelrohr distillation at reduced pressure allowed to recover α-monochlorohydrin (3), BD and estolides (see experimental section). The results obtained for the transesterification of castor oil with respectively methanol, ethanol and n-butanol are summarized in Table 2.

Table 2 Molar yields of the products isolated from the reaction between castor oil and different alcohols

	FAAE^a	1-MCH	1,3-DCH	Estolides
a FAAE – Free Acid Alkyl Esters.
Methanol	81%	37%	2%	12%
Ethanol	80%	73%	11%	19%
n-Butanol	82%	63%	3%	20%

---

4. Conclusions

In conclusion, a facile process for producing chlorohydrins and fatty acid alkyl esters from castor oil has been disclosed. Thanks to the presence of an hydroxyl group, ricinoleic acid provides to castor oil a high solubility in alcohols that allows transesterification reactions in homogeneous phase. In this particular conditions, an acid mediator like TMSCl, which was already found to be a good chlorinating agent for glycerol, drives the reaction towards BD and chlorohydrins. The sustainability of the process in relation to the use of large amounts of TMSCl (that, should be stressed here, behaves as a transesterification mediator and a chlorinating agent at the same time) is warranted by the complete recovery of hexamethyldisiloxane (TMSOTMS), the hydrolysis product of the mediator, which can be converted back into TMSCl by literature simple procedures.²⁵

Moreover, by controlling the alcohol to oil ratio it's possible to achieve a good selectivity for α-monochlorohydrin (3) formation with respect to α,γ-dichlorohydrin (4) as well as to limit the amount of estolides.

The present method does not allow an efficient production of estolides from castor oil. However, some preliminary results show the possibility to enhance the estolides yields by treatment of methyl ricinoleate with TMSCl and a little amount of acetic acid at 60 °C for 24 hours. Further studies on this topic are ongoing in our laboratories.

Notes and references

1. F. Ma and M. A. Hanna, Bioresour. Technol., 1999, 70, 1–15.
2. J. V. Gerpen, Fuel Process. Technol., 2005, 86, 1097–1107.
3. M. Balat, Energy Convers. Manage., 2011, 52, 1479–1492.
4. E. Santacesaria, G. M. Vicente, M. Di Serio and R. Tesser, Catal. Today, 2012, 195, 2–13.
5. F. Yang, M. A. Hanna and R. Sun, Biotechnol. Biofuels, 2012, 5, 1–10.
6. A.-M. Al-Lal, J.-E. García-González, A. Llamas, A. Monjas and L. Canoira, Fuel, 2012, 93, 632–637.
7. A. Kazmi, RSC Green Chem. Ser., 2013, 24, 1–24.
8. E. Santacesaria, R. Vitiello, R. Tesser, V. Russo, R. Turco and M. Di Serio, Ind. Eng. Chem. Res., 2014, 53, 8939–8962.
9. D. Giomi, M. Malavolti, O. Piccolo, A. Salvini and A. Brandi, RSC Adv., 2014, 4, 46319–46326.
10. J. M. Encinar, N. Sánchez, G. Martínez and L. García, Bioresour. Technol., 2011, 102, 10907–10914.
11. S. M. P. Meneghetti, M. R. Meneghetti, C. R. Wolf, E. C. Silva, G. E. S. Lima, L. de Lira Silva, T. M. Serra, F. Cauduro and L. G. de Oliveira, Energy Fuels, 2006, 20, 2262–2265.
12. N. de Lima da Silva, C. Benedito Batistella, R. Maciel Filho and M. R. W. Maciel, Energy Fuels, 2009, 23, 5636–5642.
13. G.-T. Jeong and D.-H. Park, Appl. Biochem. Biotechnol., 2009, 156, 1–11.
14. J. M. Encinar, J. F. González and A. Pardal, Fuel Process. Technol., 2012, 103, 9–15.
15. V. E. Efeovbokhan, A. Ayodeji, A. P. A. Lucky and O. Ogheneofego, Int. J. Eng. Technol., 2012, 2, 1734–1742.
16. J. M. Dias, J. M. Araújo, J. F. Costa, M. C. M. Alvim-Ferraz and M. F. Almeida, Energy, 2013, 53, 58–66.
17. F. Halek, A. Delavari and A. Kavousi-rahim, Clean Technol. Environ. Policy, 2013, 15, 1063–1068.
18. M. McCoy, Chem. Eng. News, 2013, 91, 56.
19. H. Mutlu and M. A. R. Meier, Eur. J. Lipid Sci. Technol., 2010, 112, 10–30.
20. L. Canoira, J. García Galeán, R. Alcántara, M. Lapuerta and R. García-Contreras, Renewable Energy, 2010, 35, 208–217.
21. S. M. P. Meneghetti, M. R. Meneghetti, C. R. Wolf, E. C. Silva, G. E. Lima, M. d. A. Coimbra, J. I. Soletti and S. H. Carvalho, J. Am. Oil Chem. Soc., 2006, 83, 819–822.
22. N. de Lima da Silva, M. R. W. Maciel, C. B. Batistella and R. Maciel Filho, Appl. Biochem. Biotechnol., 2006, 130, 405–414.
23. T. M. Serra, D. R. de Mendonça, J. P. V. da Silva, M. R. Meneghetti and S. M. Plentz Meneghetti, Fuel, 2011, 90, 2203–2206.
24. A. Salvini, D. Giomi, G. Cipriani, G. Bartolozzi, R. Alfini and A. Brandi, RSC Adv., 2012, 2, 4864–4868; A. Brandi, A. Salvini, G. Cipriani, D. Giomi and G. Bartolozzi, Process for the production of biodiesel by transesterification of triglycerides, WO2011023712A1, 2011.
25. B. Zhang and L. Shi, Hebei Gongye Keji, 2007, 24, 63–65.
26. W. L. F. Armarego and D. D. Perrin, Purification of Laboratory Chemicals, Pergamon, 4th edn, 1997.
27. E. Castillo, I. Regla, P. Demare, A. Luviano-Jardón and A. López-Munguía, Synlett, 2008, 18, 2869–2873.
28. A. Orfanakis, E. Hatzakis, K. Kanaki, S. A. Pergantis, A. Rizos and P. Dais, J. Am. Oil Chem. Soc., 2013, 90, 39–51.
29. S. C. Cermak and T. A. Isbell, J. Am. Oil Chem. Soc., 2001, 78, 557–565.
30. T. A. Isbell, M. R. Edgcomb and B. A. Lowery, Ind. Crops Prod., 2001, 13, 11–20.
31. T. A. Isbell and S. C. Cermak, J. Am. Oil Chem. Soc., 2002, 79, 1227–1233.
32. C. L. Ruddick, P. Hodge, A. Cook and A. J. McRiner, J. Chem. Soc., Perkin Trans. 1, 2002, 629–637.
33. M. G. Kulkarni and S. B. Sawant, Eur. J. Lipid Sci. Technol., 2003, 105, 214–218.
34. S. C. Cermak, K. B. Brandon and T. A. Isbell, Ind. Crops Prod., 2006, 23, 54–64.
35. J. Salimon, N. Nallathamby, N. Salih and B. M. Abdullah, J. Autom. Methods Manage. Chem., 2011, 2011, 1–4.
36. H. Horchani, A. Bouaziz, Y. Gargouri and A. Sayari, J. Mol. Catal. B: Enzym., 2012, 75, 35–42.

---

Footnote

† The mixture of MeOH, TMSOTMS, and TMSCl can be quantitatively recovered and recycled for the transesterification of oils after restoring the appropriate molar ratio of MeOH and TMSCl (see ref. 24).

---