SYNTHESIS ROUTES FOR POTENTIAL POUR POINT DEPRESSANTS FROM CASHEW NUT SHELL LIQUIDS
Quach Thi Mong Huyen1, Nguyen Ngoc Minh Chau1, Nguyen Bui Huu Tuan1,2, Dao Thi Kim Thoa1,2, Nguyen Vinh Khanh1,2,* 1
Faculty of Chemical Engineering, HCMC University of Technology, Ho Chi Minh City, Vietnam 2 Key Lab for Chemical Engineering and Petroleum Processing, Vietnam National University, Ho Chi Minh City, Vietnam 268 Ly Thuong Kiet Str., Distr. 10, Ho Chi Minh City, Vietnam * Corresponding author, e-mail:
[email protected], CP: 0918 425 497
ABSTRACT Several synthetic routes are presented for modifying cashew nut shell liquid (CNSL) to obtain pour point depressant (PPD). They include saturation of C=C double bonds in the side chains of CNSL by epoxylation and bromination, as well as cationic polymerization of the side chains and polycondensation of CNSL with formaldehyde. Experimental results show that polarity of side chains of CNSL may be a crucial factor in pour point depressing ability of modified CNSL. On mixing with brominated CNSL, pour point of crude oil can be depressed up to 6oC, while with epoxylated CNSL, a 3oC decrease is observed. In case of CNSL-based polymer, the CNSL-formaldehyde resins seem to be a better PPD, compared to the cationic polymerized product. Best performance is recorded to CNSL-formaldehyde novolac resin, which was able to decrease the pour point of crude oil by 18oC. Moreover, it is also showed that there is an optimum molecular weight value of the synthesized polymer, with which the ability for depressing pour point of crude oil is expected as maximum. Thus, it is believed that with proper molecular weight control, CNSL could be polymerized into a good PPD for local crude oil. I. INTRODUCTION Vietnam crude oil is classified as paraffinic crude oil, with high content of paraffin, c.a. 17 – 30 wt%, of which 50% or more are high (solid) paraffin. The high paraffin content contributes to the high pour point of Vietnam crude oil, typically in the range of 27 – 38oC, causing some serious problems in production, transportation, and storage; and posing the need for use of an effective pour point depressant (PPD) additive. In general, a PPD is a high molecular weight substance that may effectively disperse and hinder the growth of oil’s paraffin crystals when adsorbing on them. Ideally, the structure of PPD consists of (i) a paraffin-like part, which can co-crystallize with paraffin components of crude oil; (ii) a polar part, which limits the level of co-crystallization [1, 2]. Copolymers of alkyl acrylates and maleic anhydride are typical common PPD used in practice [3-5]. Cashew nut shell liquid (CNSL), a by-product of the cashew industry, is a cheap, renewable and abundantly available raw material in Vietnam. Anacardic acid and cardol, two major alkyl phenolic components of CNSL, can act as reagents for a variety of reactions for the manufacture of a multitude of valuable products, including inhibitors, detergents, dispersants, and EP additives [6]. It is widely accepted that anacardic acid with proper thermal treatment yields four types of cardanols, which have almost ready-made chemical structure of an additive. As can be seen in Fig. 1, the C15 linear side chains of cardanols offer excellent solubility in diesel oils, and light lubricating oils and the strongly polar phenol group will induce anti-oxidant characteristics.
1
Fig. 1. Cardanols obtained from distillation of anarcadic acids. [7] Basically, without any modification, CNSL can be utilized as PPD, since the alkyl phenols of CNSL have a linear long alkyl chain and polar hydroxyl group, matching the structural requirements for a good PPD, as described above. However, the major drawback of CNSL is the un-saturation of the linear side chains, which causes thermal- and oxidation instability and also self-condensation and polymerization, forming sludge and gum. Although hydrogenation and alkylation are capable of preserving the non-polarized ability of the linear chain, their main drawbacks include a requirement for expensive equipments and severe reaction conditions. Therefore, a simple and less energy intensive method to saturate double bonds of linear chains of CNSL is worth of notice. Another interesting feature of CNSL is that the alkyl side chains contain C=C double bonds, which can be converted into polymers. These polymers are expected to cocrystallize with crude oil paraffin or to adsorb onto the surface of paraffin molecules. Consequently, the formation of large and strong crystalline network is inhibited, resulting in a decrease of pour point of the crude oil. Alternatively, alkyl phenols in CNSL can react with formaldehyde, in a presence of an acidic catalyst, to form a novolac type resin, which has an ability of dispersing paraffin molecules [6]. The synthesis of novolac type resin from cardanol and formaldehyde is well documented in relevant literature [8-9]. This paper presents our preliminary results on different reaction routes to produce potential PPDs from CNSL, and the performances of the obtained products on depressing of Vietnam crude oil’s pour point.
II. EXPERIMENTALS 2.1. Materials All reagents and analytical compounds were purchased from China and used without further purification, unless otherwise noted. CNSL was bought from a local exporting company, with stable and reproducible quality. 2.2. Methods CNSL was first decarboxylated at 160oC for 1hr, in order to convert most of its anarcadic acid component to cardanol. Ideally, a vacuum distillation step would be desired to obtain pure cardanol for further reaction. However, this would make the process more complex and costly in terms of technology and energy. Thus, the decarboxylated CNSL (deca-CNSL) was directly used as starting material in several consequent synthetic routes, described below. 2.2.1. Epoxylation of linear chain 2
A mixture of 7.82g of formic acid (0.17 moles), 50g of deca-CNSL (0.17 moles) and 1.5g H2SO4 was added to a glass vessel and the mixture under stirring was cooled to 0ºC. 38.5g (0.34 moles) of 30% H2O2 was then added drop-wise to the mixture. The addition time required 5-10 hours. After completion of the addition, the reaction temperature was increased to 35ºC, whereby the epoxy groups were converted to hydroxylformoxy ester. The half ester produced was hydrolyzed with sodium acetate at 80ºC for 4 hours. The product was neutralized, washed with excess water until neutral to litmus paper and dried over anhydrous sodium sulfate. 2.2.2. Bromination of linear chain Bromination was carried out in accordance with the procedure proposed by Attanasi et al [10]. A mixture of 5.565 g of KBr and 19.835 g of KBrO3 dissolved in carbon tetrachloride (80 ml) was well stirred and cooled to 0ºC. Then, 7 ml of 98% H2SO4 was added drop-wise for 10 minutes. After completion of the addition, the mixture was further stirred for 60 minutes. A solution of deca-CNSL (15.1 g) dissolved in carbon tetrachloride (40 ml) was then added to the prepared solution of bromine in carbon tetrachloride (100 ml) and the mixture was magnetically stuffed in ice bath at 0ºC for 60 minutes. The reaction mixture was washed with saturated aqueous sodium carbonate (5x20 ml) and then with water (2x20 ml). The product was dried over anhydrous sodium sulfate. Epoxidized and brominated CNSLs are termed saturated CNSLs. 2.2.3 Cationic polymerization of CNSL side chain 100g of deca-CNSL was fed into a 4-neck reactor with condenser. The CNSL was then heated to 180oC, under nitrogen atmosphere. Then 1.5g of a catalyst consisting of H2SO4 and (C2H5)2SO4 (with volume ratio of 1/3) was slowly added to the heated CNSL, on which the cationic polymerization of CNSL was taking place. Samples were taken every 1 hr for viscosity measurement in order to evaluate the progress of polymerization, and termed as cationic polymerized CNSL. 2.2.4 Synthesis of CNSL-formaldehyde novolac resin Novolac resins were prepared from deca-CNSL and formaldehyde with oxalic acid as catalyst, with molar ratio 1:0.8 of cardanol (aproximately 80% of deca-CNSL) to formaldehyde. The amount of oxalic acid used was 0.25mol% based on the amount of cardanol. Polycondensation was carried out in dimethylformamide (DMF) solvent. Typically, 100g deca-CNSL, 100ml DMF, 0.05g of methyl hydroquinone (MEHQ) and 0.84g oxalic acid were fed into a 500 ml four-necked flask fitted with contact thermometer, stirrer, dropping funnel and water separator. With stirring and nitrogen blanketing, the reaction mixture was heated to 120° C followed by addition of 21.3 ml of 37% aqueous formaldehyde solution. Samples were taken every 30 min for viscosity measurement in order to evaluate the progress of polymerization. Expected chemical structures of cationic polymerized CNSL and CNSLformaldehyde novolac resin are given in Fig. 2. 2.3. Analysis of products C=C saturation was analyzed by iodine index, employing Kaufmann method, in order to determine the effectiveness of saturation reaction (routes 2.2.1 and 2.2.2). The conversion of C=C bonds can be calculated from iodine index, using the following formula: H
IV0 IV 100 IV0
where: IV0: Iodine index of material; IV: Iodine index of product; H: Conversion of C=C bonds. 3
Ubbelhod capillary viscometer (SYD-265B Petroleum Products Kinematic Viscosity Tester) was used for viscosity measurement at 40oC. CNSL-based polymers were analyzed by GPC (Agilent 1100 GPC) to determine the average molecular weights. Mic A separation column with capillary size of 5×102 – 105 Ao was used for measurement at 30oC. SEM (JEOL – JSM 7401F) observation was carried out with 5kV electric potential to investigate the morphological of paraffin crystalline with and without PPD additives. The samples were undergoing crystallization at 0oC followed by a deposition of a thin gold layer on the surface prior to measurement.
Fig. 2. Chemical structure of (a) cationic polymerized CNSL and (b) CNSL-formaldehyde novolac resin. 2.4. Determination of pour point of FO and FO / PPD mixtures The saturated deca-CNSLs and polymers were used for decreasing pour point of “Black Lion” crude oil (BLCO). Each PPD product was mixed with various solvents to form 50wt% PPD solutions. PPD solution was then mixed with crude oil to reach the final content of PPD in the mixture of 0.5wt%. After that, pour points of the mixtures were determined in accordance with ASTM D97. “Blank PPD solution” containing only solvent, was also prepared and tested in order to exclude the pour point decreasing effect of pure solvent. III. RESULTS AND DISCUSSION 3.1. The saturation level of C=C and pour point depressing ability of the products Products obtained from different methods are subjected to iodine index evaluation in order to determine the C=C bonds saturation level. The iodine index (g Iodine/100g) of each modified CNSL product was measured twice to calculate the average value. The results are presented in Table 1. Table 1. Saturation level of neat and saturated CNSLs Calculated Iodine conversion of C=C Samples index bonds (%) CNSL 158.49 Epoxylated CNSL 84.45 46.61 Brominated CNSL 119.79 24.27 The decrease in iodine index of reaction products are evidences of formation of desired products. It is found that, among the saturation methods applied, the epoxylation leads to the highest conversion (46.6%). A higher conversion could be obtained with optimization of reaction conditions. Saturation level obtained with bromination is 24.3%. It should be noted that, drop-wise addition of reagents and avoiding of direct sunlight are crucial to the conversion of bromination. As with the presence of direct light, the substitution reaction can occur, which reduces the level of saturation. 4
3.2. Cationic polymerized CNSL and synthesis of CNSL-formaldehyde novolac resin. Fig. 3 shows the change in viscosity of reaction mixture as the cationic polymerization of deca-CNSL proceeds.
Fig. 3. Change in (▀) viscosity and (●) molecular weight of deca-CNSL-based polymer product obtained by cationic polymerization vs. reaction time. As can be seen, the viscosity slowly increases from 62.7cSt (0h) to 3526 cSt (8h), implying a slow increase in molecular weight of the polymer products and a slow rate of reaction, which might be attributed to self-retarding ability of cardanol molecules, as well as some chain transfer reaction induced by other components in deca-CNSL, most likely water and those components with –COOH functional groups. GPC measurements show the increase of molecular weights from 387g/mol, corresponding to neat deca-CNSL up to 2200 g/mol after 8h of reaction, which is in agreement with viscosity measurement results. As for the polycondensation of deca-CNSL to form novolac type resin, it was observed during the experiment that the reaction was exothermic, with relatively fast increase in viscosity of the product. Eventhough MEHQ was used to prevent selfpolymerization of unsaturated alkyl side chains of CNSL, the viscosity of reaction mixture was found to increase hugely after 2 hrs of reaction time, which was calculated after the addition of final drop of formaldehyde. This seems to suggest that the polymerization of CNSL side chains is inevitable as the polycondensation is prolonged at high temperature. Since it was aimed to determine the effect of polycondensation product in decreasing pour point of crude oil, thus, only the sample products obtained before 2 hrs of reaction were further used for pour point measurements. 3.3. Pour point depressing ability of modified CNSLs Testing with “blank PPD solutions” shows no change in pour point of the crude oil, which implies that addition of low freezing point solvents could not lower the crude oil’s pour point. This could be explained by the fact that solvent molecules, with much more smaller size compared to crude oil molecules, could not prevent the crystallization and growth of large crystalline of paraffin molecules of the crude oil. Also, the small size molecules cannot induce any spatial arrangement to separate paraffin crystalline or to prevent the formation of a paraffin crystalline network in crude oil. a) Epoxylated and brominated CNSLs Results of pour point measurements for mixtures of crude oil and epoxylated / brominated CNSLs are given in Fig. 4.
5
Fig. 4. Pour points of BLCO and epoxylated / brominated CNSL mixtures As is shown, although the bromination give the quite low conversion, the products of this method showed the highest capacity of reducing the pour point of the crude oil. Eventhough, the presence of bromine atoms in the carbon chain can boost the polarity of the alkyl chain, the brominated CNSL can still solubilize the parrafin existing in crude oil, and thus, reduce the pour point of oil. The importance of polarity of linear chain, in respect to pour point depression ability, is evident in the case of epoxylated CNSL. It can be clearly seen that, despite of relatively high conversion, epoxylated CNSL can not effectively decline the pour point of crude oil. It will be rationalized by accepting that the presence of highly polar OH groups with high concentration in the product seems to inhibit strongly its ability of solubilizing paraffin in oil. It should also be noted that there are two other factors, which impact critically the ability of reducing the pour point of crude oil. They are the concentration of additive and the level of dispersion of additive in crude oil. With this argument, the concentration of 0.5wt% applied to test might not be high enough or the level of dispersion of additive in oil might not be good enough to achieve more favorable results. On the other hand, use of PPD with higher concentration (than 0.5wt%) seems to be uneconomical. b) Cationic polymerized CNSL Two cationic polymerized CNSL samples with Mw of 870 and 2200 g/mol, respectively, were employed for decreasing pour point of BLCO. Results are summarized in Fig. 5.
Fig. 5. Pour points of BLCO and cationic polymerized CNSLs. 6
Results in Fig. 5 show that the higher the Mw, the better the pour point decreasing ability is. In another words, there is a correlation between Mw and pour point of the mixture. It is noted that, the magnitude of the decrease is limited for all the samples tested. It is postulated that after polymerized, the alkyl side chains of CNSL become able to cocrystallize with paraffin molecules in crude oil, due to their similarity in structure. Nevertheless, the polymer seems uanble to effectively prevent the growth of large crystalline network. As described in Fig. 2a, the structure of cationic polymerized CNSL is consisting of branches of 6 C atoms, which is not large enough to induce an accountable spatial hindrance effect to prevent the further crystallization of paraffin molecules. An effective hindrance is usually obtained with chains of 16 ~ 21 C atoms [11-12]. Thus, the overall pour point decreasing ability of this type of polymerized CNSL is somehow underachieved. c) CNSL – formaldehyde novolac resin Fig. 6 gives pour points of the mixtures from BLCO and prepared novolac resins. As mentioned in section 3.2 above, only samples obtained before 2hrs of reaction were used.
Fig. 6. Pour points of BLCO and CNSL-formaldehyde novolac resin mixtures It can be seen that the CNSL-formaldehyde novolac resins have remarkably high pour point depressing ability compared to the other modified CNSLs prepared in this study. The pour point depressing ability of these novolac resins seem to be the consequence of their ability in adsorbing on and then dispersing the paraffin microcrystalline [13-14] . Chemical stucture of the resin, shown in Fig. 2b, suggests that linear alkyl chain (R) may well adsorb on the surface of paraffin microcrystalline, due to structure similarity; while phenolic groups connected by methylene bridges may act as a bulk spacer to keep paraffin microcrystallines away from each other. This argument is supported by SEM observation of the crystallized surfaces of mixtures, shown in Fig. 7. It can be seen that, the surface of crude oil without PPD is covered with large paraffin crystals, while in case of crude oil with PPD, paraffin crystals have much smaller size with lower numbers. The smaller size of the crystal indicates that the paraffin molecules in crude oil / PPD mixture are well separated, thus, can not be formed into a network. On the other hand, lower numbers of crystals seem to suggest that the paraffin molecules can not easily crystallize at all. As the result, the fluidity of crude oil/PPD mixture is maintained, at much lower temperature, compared to neat crude oil. It is also noteworthy that sample obtained after 60 mins of reaction time gives the best performance among those tested sample, which is a direct evidence for a dependence of pour point depressing ability on the Mw of the resin. Therefore, it is believed that with 7
proper Mw control during the polycondensation, CNSL can be tailored into an effective PPD for crude oil.
(a)
(b)
Fig. 7. SEM photographs of surface of (a) freezed neat crude oil and (b) freezed crude oil/CNSL-novolac resin.
IV. CONCLUDING REMARKS CNSL was subjected to several synthesis routes in attempt to produce effective PPD for Vietnam crude oil. While all the products might decrease the pour point in some extent, the most promising method seemed to be polycondensation with formaldehyde, in the presence of oxalic acid catalyst, to form novolac resin, which showed remarkable performance in decreasing the pour point of crude oil. A CNSL-formaldehyde novolac resin, obtained after only 1hr of reaction, could reduce the pour point of paraffinic crude oil by 18oC. A dependence of pour point on the molecular weight of the PPD resin was also confirmed, thus, proper molecular weight control measures would be applied during polycondensation process, in order to achieve a good PPD. The optimization of polycondensation condition, including utilized catalyst’s type and content, and detail investigation to clarify the pour point – molecular weight relationship are on-going and will be reported in future papers. Acknowledgement: The financial supports from Vietnam National University - Ho Chi Minh City, under grant No. B2011-20-20, and from Ho Chi Minh City University of Technology under University R&D Program 2012 are greatly appreciated.
1.
2. 3. 4.
REFERENCES Holder G. A., Winkler J., Wax crystallization from distillate fuels. I: Cloud and pour point phenomena exhibited by solutions of binary n-paraffin mixtures, Journal of the Institute of Petroleum, 1965, 51, 228-234. Holder G. A., Winkler J., Wax crystallization from distillate fuels. II: Mechanism of pour depression, Journal of the Institute of Petroleum, 1965, 51, 235-252. Hemant P. S., Dinakar P. B., Synthesis and Evaluation of Polymeric Additives as Flow Improvers for Indian Crude Oil, Iranian Polymer Journal, 2006, 15(12), 943-954. Al-Sabagh A. M., Noor El-Din M. R., Morsi R. E., Elsabee M. Z., Styrene-maleic anhydride copolymer esters as flow improvers of waxy crude oil, Journal of Petroleum Science and Engineering, 2009, 65(3–4), 139-146.
8
5. Deshmukh S., Bharambe D. P., Synthesis of polymeric pour point depressants for Nada crude oil (Gujarat, India) and its impact on oil rheology, Fuel Processing Technology, 2008, 89(3), 227-233. 6. Lubi M. C., Thachil E. T., Cashew nut shell liquid (CNSL) - a versatile monomer for polymer synthesis, Desinged Monomers and Polymers, 2000, 3(2), 123 - 153. 7. Patil A. S., Pattanshetti V. A., Dwivedi M. C., Functional fluids and additives based on vegetable oils and natural products: a review of the potential, Journal of Synthetic Lubrication, 1988, 15(3), 193 - 211. 8. Yadav R., Devi A., Tripathi G., Srivastava D., Optimization of the process variables for the synthesis of cardanol-based novolac-type phenolic resin using response surface methodology, European Polymer Journal, 2007, 43(8), 3531-3537. 9. Sultania M., Rai J. S. P., Srivastava D., Studies on the synthesis and curing of epoxidized novolac vinyl ester resin from renewable resource material, European Polymer Journal, 2010, 46(10), 2019-2032. 10. Attanasi O. A., Buratti S., Filippone P., Regioselective Bromination of Cardanol Derivatives, Organic Preparations and Procedures International - The New Journal for Organic Synthesis, 1995, 27(6), 645 - 650. 11. Misra S., Baruah S., Singh K., Paraffin problems in crude oil production and transportation, SPE Production & Facilities, 1995, 10(1), 50-54. 12. Borthakur A., Laskar N. C., Mazumdar R. K., Rao K. V., Subrahmanyam B., Synthesis and evaluation of alkyl fumarate-vinyl acetate copolymers in combina-tion with alkyl acrylates as flow improvers for Borholla crude oil, Journal of Chemical Technology and Biotechnology, 1995, 62, 75-80. 13. Leinweber D., Feustel M., Wasmund E., Grundner H., Use of cardanol aldehyde resins as asphalt dispersants in crude oil, US Patent No. US20040050752, 2004. 14. Moreira L. F., Lucas E. F., González G., Stabilization of asphaltenes by phenolic compounds extracted from cashew-nut shell liquid, Journal of Applied Polymer Science, 1999, 73(1), 29 - 34.
9
