Determining optimal cashew nut shell liquid extraction method

D.O.C.E.N.T (Determining Optimal Cashew nut shell liquid ExtractioN meThod) Final report project By: L. van de Laar, M. Martinez Aguilera, J. Nelis, B. de Raad & M. Vergeer Avans University of applied sciences, Lovensdijkstraat 61 - 63, 4818 AJ Breda Abbreviation list AA Anacardic acid CNS Cashew nut shells CNSL Cashew nut shell liquid CoE BBE Centre of Expertise Biobased Economy FT-IR Fourier TransformInfrared spectroscopy GC Gas chromatography HPLC High-performance liquid chromatography RBR Rotating bed reactor TGA Thermogravimetric Analysis Abstract The production of cashew nuts (2E9 kg/y) generates an annual waste stream of 3E9 kg, in the form of cashew nut shells (CNS). These CNS contain a valuable liquid component: cashew nut shell liquid (CNSL). This brown viscous liquid contains, amongst others, anacardic acids and cardonol, which find use in the medical and coating industry, respectively. The goal of this project was to determine the optimal extraction conditions of cashew nut shell liquid (CNSL) out of cashew nut shells (CNS), while maintaining an economically attractive and technically feasible process. The desired yield for the extraction was ≥90 m%. The project started with three extraction techniques in mind; a rotating bed reactor, accelerated solvent extraction and supercritical fluid extraction. Since the bed reactor showed promising results in early stages, a yield of over 70 m%, in relation to the total amount of CNSL present, this was chosen as the extraction method to optimize. The parameters used for the project were extraction temperature, particle size distribution, stirring speed, extraction time, rotating bed or propeller stirrer, solvent and solid/solvent ratio. The resulting extract was examined using HPLC, TGA, FTIR and GC. The parameters that were found to influence the extraction yield: extraction time, solvent use, solid/solvent ratio and extraction temperature. It was found that the optimal extraction time, in relation to the yield gained afterwards, is one hour with a yield of 0,40 g CNSL/g CNS. After two hours of extraction, there was an increase of yield below one percent, which determined the optimum. It was also found, that removing the rotating bed, and using a mixer in the reactor, did not result in deviations in yield. The influence of temperature was visible, with 50 ºC and 70 ºC showing similar results, with a yield of 0,40 g CNSL/g CNS. This was while 20 ºC showed a result of 0,33 g CNSL/g CNS, therefore 50 ºC was found to be optimal. When comparing the solid/solvent ratios 1:7, 1:13 and 1:20, it was found to give a yield of 0.28, 0.33 and 0.39 g CNSL/g CNS, respectively. Finally, the use of the solvents heptane and ethanol were examined, which showed the following: Heptane showed a yield of 0,36 g CNSL/g CNS, while ethanol yielded 0,40 g CNSL/g CNS under similar conditions. In summary, the optimal extraction conditions, when taking influentory parameters in mind, is using ethanol at 50 ºC, with 1:20 solid/solvent ratio, for one hour. TGA analysis showed that, when comparing ‘fresh’ CNS and CNS that underwent extraction, an extraction rate of 78.2 m% was achieved, therefore not obtaining the 90 m% goal of the project. In order to assess the ethanol fraction left in the CNSL after vacuum distillation, GC-analysis was executed, which showed an average of 10 % (v/v) remaining in the CNSL. The yield was not adjusted for ethanol content, since there was no data on the amount of solvent in the extractions, where heptane was used. Further analysis, in order to determine the composition of the CNSL, in the form of HPLC and FTIR, was also performed. Literature shows [2] an AA content in CNSL differing between 50-70 m%, when extracted under mild conditions. HPLC-analysis showed the absence of saturated AA, while AA triene was present in concentrations around 83 mg/ml. However, due to the lack of monoand diene AA analytical standards, the content of these substances in the CNSL could not be determined. AA’s were extracted out of CNSL, by dissolving it in 5% aqueous methanol and excess calcium hydroxide, which formed the precipitate calcium anacardate. This was then further purified, by suspending it in hydrochloric acid (1.5M) and washing it with ethyl acetate and distilled water. The solids were then dried and analyzed by FTIR and HPLC. The data of the extract corresponded with that of magnesium sulphate, which was used as a drying agent during the extraction, therefore, its effectiveness was called into question. D.O.C.E.N.T (Determining Optimal Cashew nut shell liquid ExtractioN meThod) Authors: Lauri van de Laar [email protected] Mario Martinez Aguilera [email protected] Jarno Nelis [email protected] Bart de Raad [email protected] Max Vergeer [email protected] Chemical engineering students (Biobased Technology) Avans Hogeschool Breda ATGM Assigned by: Qian Zhou [email protected] Centre of Expertise Biobased Economy (CoE BBE) & Avans Lovensdijkstraat 61 - 63, 4818 AJ Breda Supervisor: Nathalie Marquez Luzardo [email protected] Project details: Version: 2.1 Date: 24 June 2019 Contents 1. Introduction .......................................................................................................................... 1 2. Theoretical background ....................................................................................................... 2 2.1. Cashew nut shell liquid (CNSL) .................................................................................... 2 2.2. Extraction set-ups ......................................................................................................... 3 2.1.1. Rotating bed reactor (RBR) ....................................................................................... 3 2.1.2. Agitator ...................................................................................................................... 3 2.3. General concept of solid-liquid extraction ..................................................................... 3 2.4. Thermogravimetric Analysis (TGA) ............................................................................... 4 2.5. Anacardic acid .............................................................................................................. 4 3. Methods .............................................................................................................................. 5 3.1. CNS preparation ........................................................................................................... 5 3.2. Extraction of CNSL ....................................................................................................... 5 3.3. CNSL composition ........................................................................................................ 6 3.3.1 Anacardic acid content by HPLC ................................................................................. 6 3.3.2. Ethanol content of CNSL ........................................................................................... 7 3.3.3. Isolation of anacardic acid ......................................................................................... 7 3.3.4. FT-IR analysis ........................................................................................................... 7 4. Results ................................................................................................................................ 8 4.1. Extraction performance ................................................................................................. 8 4.2. Mass balance ............................................................................................................. 14 4.3. Compound analyses ................................................................................................... 15 4.3.1. High-performance liquid chromatography (HPLC) ............................................... 15 4.3.2. Fourier-transform infrared spectroscopy (FT-IR) .................................................. 19 4.3.3. Gas chromatography (GC) .................................................................................. 20 4.3.4. Thermogravimetric analysis (TGA) ...................................................................... 21 5. Discussion ......................................................................................................................... 23 6. Conclusion......................................................................................................................... 25 7. Recommendations ............................................................................................................. 26 Literature .................................................................................................................................. 27 Appendix 1 Mass balance extractions ..................................................................................... 1 Appendix 2: HPLC results ....................................................................................................... 7 Appendix 3: GC analysis calibration curve .............................................................................. 9 Appendix 4: TGA analysis results .......................................................................................... 10 1 1. Introduction With the combat against global warming and pollution, research towards bio-based sources of valuable components has increased over the last decades [1]. Bio-based sources come in many forms, like sugar-based feedstock such as sucrose and starch, but also microbial feedstock, like bacteria and fungi. This way, the dependence of the finite fossil resources can be decreased and the distorting effects towards the climate can be decreased. This project is done on behalf of, and in collaboration with the research group Center of Expertise, BioBased Economy, (CoE BBE), which has previously put effort towards the research around Cashew nut shell liquid (CNSL). Previous work involved extraction by pyrolysis, while this project lays the focus on solvent extractions. The composition of the CNSL is mostly phenolic acids (~80 m%) [2] but is dependent on the extraction route [3]. The relevant liquid composition for this report is CNSL extracted by (natural) solvent, which contains mostly anacardic acids (saturated, mono-, dior triene) and cardol [3]. AA triene is known for its anti-inflammatory and anti-oxidizing properties and is most prominently (~44 m% of AA content) available in CSNL [4], while the versatile liquid as a whole can also be used as fuel source [5], additive [6] or paint. Even though this research mainly focused on extraction, the market potential for CNSL and its compounds are promising. The main goal of this project was to design an extraction process, which could extract CNSL from cashew nut shells (CNS) with a yield of ≥90 m%. The prerequisite was that the process was technically feasible and economically attractive. In order to achieve this, mild conditions were chosen, such as temperatures between 20 and 70 °C and (retractable) solvents, like ethanol and heptane. Several extraction methods were selected: a reactor with a rotating bed (RBR) or agitator (propeller stirrer), accelerated solvent extraction (ASE) and supercritical CO2 extraction (SC-CO2). Due to the effectiveness of the RBR, the other methods were not executed. This report consists of seven chapters, first the background information and the motivation for this project will be construed. Then in the theoretical background, additional information, requisite for the progression of the report is given. Afterwards, the used materials and methods will be illustrated. Subsequently, all results with accompanying visual guidance will be interpreted, after which these results will be discussed in the discussion. Following, the conclusions will be presented, and corresponding recommendations will be outlined afterwards. 2 2. Theoretical background 2.1. Cashew nut shell liquid (CNSL) A cashew nut comes from a tree that produces a cashew seed and a cashew apple. A shell surrounds the cashew nut, or kernel, (see figure 1). Only the cashew nut itself is used for consumption. The shell of the cashew nut seed is, amongst others, used for paints and lubricants. The nuts are sold without the shell because the shell contains compounds that can cause an allergic reaction or an irritation. The cashew nut shell, CNS, is leathery and contains a brown viscous liquid, CNSL or the pericarp fluid, which amounts for ~67 m% of the nut weight. Depending on the method of obtainment, CNSL may present a different chemical composition and can therefore be classified into two main types: solventextracted CNSL and technical CNSL. The CNSL contains, among others, phenolic lipids, anacardic acids, and cardanol. This project focuses on the natural CNSL, the solvent extracted CNSL. [7] Natural CNSL is obtained by using some solvent extraction techniques (like Soxhlet, supercritical carbon dioxide, or subcritical water) in order to obtain the components without inducing any chemical modifications. With these techniques, natural CNSL represents the original composition found in nature, which is composed by AAs (60-70%), cardols (10-20 m%), cardanols (3-10 m%), 2-methylcardols (2-5 m%), and other minor components.[8] Figure 1: Profile of a cashew nut. [7] 3 2.2. Extraction set-ups 2.1.1. Rotating bed reactor (RBR) A rotating bed reactor holds a rotating packed bed with a solid phase (CNS) as displayed in figure 2. The solid phase remains in the bed by a filter. By rotating the bed at high speed, the extraction solvent will be aspirated from the bottom and top of the vessel to the middle and through centrifugal force pushed outward through the solid phase. This maximizes axial mixing and convective transport. The resulting efficient mass transfer minimizes extraction time and enables a high yield. [9] Figure 2: Rotating bed reactor with a cooling system for batch process. [9] 2.1.2. Agitator For checking the advantages of the rotating bed compared to a regular propeller stirrer (hereafter: ‘regular stirrer’) for agitation, extractions will also be executed using a propeller stirrer (figure 3). The CNS will be on the bottom of the reactor, therefore the solids need to be separated from the liquid after the extraction by filtration. Figure 3: Regular propeller stirrer 2.3. General concept of solid-liquid extraction In solid-liquid extraction, molecules from a solid particle are extracted by the liquid phase. The speed of reaching the extraction equilibrium depends on many variables, such as temperature, particle size, solvent, solid/solvent ratio and the concentration difference. Transport of molecules to the liquid is dependent on the affinity of the molecules to the solvent. With a higher temperature, the mass transfer increases and is more efficiently due to a higher solubility of CNSL in the solvent. If the rotation speed increases, convection increases, resulting in faster mass transfer through the external film layer [10]. When particles are smaller, there is more contact area for the solvent, which should also result in a higher yield. When there is more liquid than solid, the extraction should also proceed faster, because the concentration of the extract is lower, thus creating a bigger concentration difference [11]. 4 2.4. Thermogravimetric Analysis (TGA) Thermogravimetric analysis, TGA (figure 3), is an analytical technique that can measure the loss of mass of a sample over time, while the temperature is increased. When using this technique, the temperature is increased, ranging from 20 to 1000°C. The boiling point of CNSL (215 °C) is lower than the other components in the CNSL, like hemicellulose (277 ºC), cellulose (327 ºC), lignin (387 ºC) and polymeric material (430 ºC). At the specific boiling points of the components, the components start to evaporate. As a result, the mass changes, which is measured by a precision balance. By using TGA, the amount of CNSL in the shell can be determined. Therefore, the extraction yield can be determined.[12] Figure 4: Schematic drawing of the inside of a TGA.[12] 2.5. Anacardic acid AA is a compound that can be used for medicinal purposes. In a pure form it is an expensive ($1000/10 mg for AA triene) product. AA can be present in 4 different forms: saturated, with one (mono-), two (di-) and three (tri) double bonds (ene). The mono-, diand triene are the ones present in the CNSL. AA’s structure formula looks similar to cardol, 2-methylcardol and cardanol, three molecules that are also present in CNSL (figure 5). Figure 5: AA and its similar forms in CNSL: cardanol, cardol and 2-methylcardol. The AA is recognized by its carboxylic acid group (COOH) attached to the benzene ring. This acid group can be easily detected using FT-IR analysis. Using HPLC analysis, the different compounds can be detected due to their different polarities, i.e. different affinity with the column. This is also the case for the three forms of AA present in the CNSL. These will be detected separately due to the mono-, diand triene bonds resulting in different polarities. 11 Figure 11 shows the influence of rotating on the yield of CNSL per g CNS. For rotating speeds of 0, 300 and 1050 rpm a yield of 0,35, 0,39 and 0,36 were found with standard deviations of 0,010, 0,009 and 0,018, respectively. There is no clear correlation between the rotating speed and the CNSL yield. Figure 11: g CNSL/g CNS found for extraction performed with different rotation speeds (for a 1-hour extraction at 50°C in RBR with 7,5 g CNS, particle size 2-4 mm and 150 ml ethanol as solvent. 12 Figure 12 shows the influence of temperature on the yield of g CNSL per g CNS. For temperatures of 20, 50 and 70 ºC a yield of 0,33, 0,39 and 0,40 were found with standard deviations of 0,009, 0,012 and 0,037, respectively. This shows that lowering the temperature has a negative effect on the CNSL yield per gram CNS. However, increasing the temperature from 50 to 70 ºC did not give a clear raise in CNSL yield per g CNS. . Figure 12: g CNSL/g CNS found for extractions performed at different temperature (for a 1-hour extraction at 50°C in RBR with 7,5 g CNS, particle size 2-4 mm, rotating speed 300 rpm and 150 ml ethanol as solvent.) 13 The rotating bed showed different extraction kinetics when compared to using a regular stirrer as displayed in figure 13. In this figure, the yield of g CNSL per g of CNS is set out against time. The kinetics for the rotating bed is expressed in equation 2, with t as time in minutes. Equation 2: 𝑔 𝐶𝑁𝑆𝐿/𝑔 𝐶𝑁𝑆 = 0,20𝑡0,16 (𝑅2 = 0,96) The kinetics of the extraction with a regular stirrer is expressed in equation 3, with t as time in minutes. In early stages, the yield of g CNSL per g CNS is higher than using a rotating bed reactor, but for extractions times after 45 minutes the RBR performs better than the regular stirrer does. Equation 3: 𝑔 𝐶𝑁𝑆𝐿/𝑔 𝐶𝑁𝑆 = 0,27𝑡0,076 (𝑅2 = 0,89) Figure 13: g CNSL/g CNS found at different extraction times (for a extraction at 50°C in RBR with 7,5 g CNS, particle size 2-4 mm, rotating speed 300 rpm and 150 ml ethanol as solvent.) 14 4.2. Mass balance The mass balance was drafted over three steps: extraction, evaporation and drying. The tables with the total difference mass balances can be found in appendix 1. With the extraction mass balance, the losses in the reactor (Losses 1) can be determined. These losses are mainly ethanol that evaporated during the extraction. The value of the ‘Losses 1’ is calculated as following: ‘Losses 1’ = (CNS + solvent) - (L1 + S1). The lowest value of Losses 1 (2,85 g) is with a particle size of 6-10 mm. The highest value of the ‘Losses 1’ (17,2 g) is with the regular stirrer. The losses of the regular stirrer are higher, due to filtration, where (pure) ethanol easily evaporates. With the evaporation mass balance, the losses during the evaporation (Losses 2) can be determined. The value of ‘Losses 2’ is calculated as: ‘Losses 2’ = (L1) - (CNSL + SO-REC). Some samples have a negative loss during evaporation, which is expected to be caused by an improper emptied recovery flask or due to measurement errors. The drying mass balance was performed for the experiments that were supported by TGA-analysis. In this balance, the vapour that is going to be solvent, is calculated as V1 = S2 – S1. In summary, the total amount of losses is described as ‘Total losses’ = Losses 1 + Losses 2 + V1. This calculation is only possible for the experiments supported by TGA-analysis, which are shown in table 2. Table 2: Total losses during the extraction process parameter condition Duplo L1 (g) L2 (g) V1 (g) L TOTAL (g) none standard 1 4,6 3,4 5,3 13,3 none standard 2 5,1 0,1 5,2 10,4 set-up regular stirrer 1 16,9 4,9 2,9 24,7 set-up regular stirrer 2 17,5 25,1 2,5 45,1 time 10 min 1 2,6 2,4 3,3 8,3 time 10 min 2 2,7 1,9 3,9 8,5 Table 2 shows that the highest loss are with the regular stirrer setup, with an average of 34,9 g. Per expectation, decreasing extraction time results in the lowest loss with an average of 8,4 g. The average recovery of ethanol, after solvent removal in all experiments performed with 150 ml ethanol and 7,5 g CNS, averaged at 89,4 m% with a 95% confidence interval between 9,8-11,4 m%. 15 4.3. Compound analyses 4.3.1. High-performance liquid chromatography (HPLC) HPLC analysis was executed on the following samples: CNSL gained by solvent extraction, AA triene , technical CNSL and cardanol. The chromatogram shows prominent peaks in the first minute. Using uracil as a known standard, the peaks in the first minute were recognized as so called “injection peaks”. This injection peak could be caused by the solvent or the mobile phase [14]. Therefore, in all chromatograms, the first peak was ignored. Figure 14: Sample 30 minutes, 7,5 grams CNS, 150 ml ethanol and 50 °C in RBR Figure 14 shows clear differentiation between distinct peaks, which allows for easy distinction of the compounds present, further in this chapter . Appendix 2 shows the graphs, retention times and areas of the AA triene peaks of 0,05, 0,1, 0,2, 0,5 mg/ml. There is a prominent peak at 6,48 min in the CNSL chromatogram, that matches with the peak of the sample at 6,47 min. Therefore, it was concluded that AA triene is present in the CNSL. As displayed per figure 15, a calibration line of AA triene was drafted with the concentrations 0, 0.05, 0.1, 0.2 and 0.5 g/L. Figure 15: Calibration line of AA triene The trend line of the calibration line makes it possible to detect the relation between the area of the peak and the concentration of the sample. 16 Equation 4: 𝑷𝒆𝒂𝒌 𝒂𝒓𝒆𝒂 =𝟐𝟏𝟒𝟗, 𝟐 ∗ 𝒄𝒐𝒏𝒄𝒆𝒏𝒕𝒓𝒂𝒕𝒊𝒐𝒏 𝑨𝒏𝒂𝒄𝒂𝒓𝒅𝒊𝒄 𝒂𝒄𝒊𝒅 𝑹𝟐= 𝟎, 𝟗𝟑 For the sample nº 16, the concentration of AA triene is 0,083 mg/ml. However, it must considered that the sample is diluted 1000x. Therefore, the real concentration of AA triene is 83,02 mg/ml. To identify other compounds, additional HPLC-analyses were executed. Figure 16 shows the chromatogram of technical CNSL. Figure 16: HPLC-chromatogram of technical CNSL. 17 The main compounds of technical CNSL are cardanol and cardol. In figure 17, the chromatogram of cardanol is shown to exclude other larger peaks in the chromatograms. The higher peaks are at 7.3, 11 and 18 min, these are presumed to be the cardanol tri, di and monoene, respectively. Figure 17: HPLC chromatogram of cardanol. Figure 18 shows that the retention time of saturated AA is at 23,5 min. Figure 18: Saturated AA analytical standard In figure 19, the HPLC chromatogram of the AAs, isolated from the CNSL gained by solvent extraction is shown. No high peaks are visible in the chromatogram. The peaks that are visible (retention time 6,026; 7,458; 11,857), are also evident in the chromatogram of the technical CNSL. 18 Figure 19: HPLC chromatogram of the AA, isolated from the CNSL gained by solvent extraction. No AA can be present in technical CNSL, since due to the high pyrolysis temperatures, AA decarboxylates to cardanol. When comparing this chromatogram with the chromatogram of technical CNSL, all peaks match, therefore it can be concluded that no AA is present in the sample. With this data, all peaks in figure 20 (14) can be identified. Figure 20: Identification of the peaks:(I) injection peak; (II) Cardol at 3 min; (III) AA triene at 6,5 min; (IV) Cardanol at 7,3 min; (V) AA diene at 9,7 min and (VI) AA monoene at 16,2 min. Now, according to literature [15] and HPLC-analysis of the AA triene standard, cardanol and the technical CSNL, it is possible to identify the peaks. The (I) is the injection peak, while (II) is cardanol because when comparing to the retention times of the technical CNSL, there is a peak that matches at 3 min. The (III) is AA triene according to the standard. The (IV) is cardanol according due to equal retention times when analysing pure cardanol. The other two peaks are AA diene (V) and AA monoene (VI), the AA diene leaves the column before the monoene, due to its polarity. Additionally, HPLC-analysis of saturated AA standard was performed, showing a retention time of 23 min, therefore it can be concluded that no saturated AA is present in the CNSL. 19 4.3.2. Fourier-transform infrared spectroscopy (FT-IR) In figure 20, strong peaks at wavenumber 2900 and 2800 (cm-1) indicate the presence of carboxylic acid groups, which AAs contains. The FT-IR analysis also indicates there is an acid present in the CNSL. The peak at 1700 (cm-1), indicates C=O bonds which match the double bonds in the diene AA. Figure 20: FT-IR chromatogram of CNSL sample gained by ethanol solvent extraction. In figure 21, an FT-IR chromatogram of CNSL is represented, which used heptane as a solvent. Figure 21: FT-IR chromatogram of CNSL sample obtained by heptane solvent extraction. The same strong peaks appear at wavenumbers 2900 and 2800 (cm-1), which again indicate the presence of carboxylic acid groups. The chromatogram is similar to the CNSL derived with ethanol. The strong peak around 1100 (cm-1) is the only peak that differs. The peak is most likely indicating an alcohol, which was the solvent used. 20 Figure 22, shows the FT-IR of the AA, obtained by isolating AA from CNSL. This shows there are small peaks, again at wavenumber 2900 and 2800 (cm-1). These again indicate the presence of carboxylic acid groups. Additionally, an OH-peak at 3400 (cm-1) is visible. These OH-groups can be related to AA as well as to cardol and cardanol since these molecules all contain OH-groups. Figure 22: the FT-IR chromatogram of the AA, obtained after using the method for isolation of AA from CNSL. 4.3.3. Gas chromatography (GC) Using GC-analysis, the amount of ethanol left in the CNSL after evaporation is determined. The concentration is approximately 10 % (v/v) ethanol. After a multi-stage distillation, the amount of ethanol is reduced to approximately 5 %(v/v) ethanol. These concentrations are determined by comparing the area of the ethanol peak in the GC-chromatogram with the calibration curve shown (R² of 0,92) in appendix 3 figure 1. Therefore, the g CNSL/ g CNS derived from the 1-hour extraction at 50°C in the RBR with 7,5 g CNS, particle size 2-4mm, rotating speed 300 rpm and 150 ml ethanol as solvent, is 0,35 instead of 0,39 g CNSL/ g CNS. 27 Literature [1] S. Y. L. e. al, "A comprehensive metabolic map for production of bio-based chemicals," Elsevier, 2019. [2] M. Y. e. al, "Effect of extraction methods on characteristic and composition of Indonesian cashew nut shell liquid," Elsevier, vol. 35, no. 1, pp. 230-236, 2012. [3] F. H. A. R. e. al, "Comparison between physico-chemical properties of the technical Cashew Nut Shell Liquid (CNSL) and those natural extracted from solvent and pressing," polimeros, vol. 21, no. 2, 2011. [4] S. K. e. al, "Efficient synthesis of AA analogues and their antibacterial activities," Elsevier, vol. 23, no. 6, pp. 1667-1670, 2013. [5] M. e. al, "Cleaner production in Burkina Faso: Case study of fuel briquettes made from cashew industry waste," elsevier, vol. 195, pp. 1047-1056, 2018. [6] A. V. e. al, "experimental investigations on combustion, performance and emission characteristics of thermal cracked cashew nut shell liquid (TC-CNSL)–diesel blends in a diesel engine," fuel, vol. 132, pp. 236-245, 2014. [7] M. Tejas Gandhi, „Studies on effect of various solvents on extraction of cashew nut shell liquid,” 2012. [8] P. Anilkumar, Cashew Nut Shell Liquid, Vancouver: Springer International Publishing, 2017. [9] Spinchem, „Rotating bed reactor applications and products,” 2018. [Online]. Available: http://www.spinchem.com/application/rotating-bed-reactor-applications-and-products/. [Geopend 19 February 2019]. [10] R. Wongkittipong, L, Prat, S. Damronglerd,, Solid–liquid extraction of andrographolide from plants—experimental study, kinetic reaction and model’’ 2004 [11] Brown, T. E. (sd). Chemistry:The central science. [12] Perkinelmer ‘’ Frequently asked question TGA’’ [13] R. Paramashivappa, P. Phani Kumar, P. J. Vithayathil, A. Srinivasa Rao ‘’ Novel Method for Isolation of Major Phenolic Constituents from Cashew (Anacardium occidentale L.) Nut Shell Liquid’’ 2001 [14] Sigma-Aldrich Co ‘’HPLC Troubleshooting Guide’’ 2009 [15] F. Oiram Filho’’Development and Validation of a Reversed Phase HPLC Method for Determination of AAs in Cashew (Anacardium occidentale) Nut Shell Liquid’’ 2017 Appendixes Appendix 1 Mass balance extractions Figure 1: overview of the process with stream names. Table 1 shows all mass streams for the extraction step documented according figure 1. Table 2 shows all the mass streams for the evaporation step documented according figure 2. Table 3 shows all mass streams for the drying step documented according to figure 3. Figure 2: Extraction step overview with stream names Table 1: Total Mass balance of the extraction step. According mass stream names from figure 2. parameter condition duplo CNS (g) Solvent (g) L1 (g) S1 (g) Losses 1 (g) none standard 1 7,5 115,4 108,3 10 4,6 none standard 2 7,5 115,7 108,1 10 5,1 Particle size 6-10 mm 1 7,5 115,5 110,4 9,5 3,1 Particle size 6-10 mm 2 7,5 115,5 111,2 9,2 2,6 Solvent Heptane 1 7,5 99,4 95,5 7,9 3,5 Solvent Heptane 2 7,5 99,4 91,9 8,7 6,3 stirring speed 0 rpm 1 7,5 115,2 108,1 5,2 9,4 stirring speed 0 rpm 2 7,5 115,1 107,6 5,1 9,9 stirring speed 1050 rpm 1 7,5 115,8 110 8,2 5,1 stirring speed 1050 rpm 2 7,5 115,7 108,5 9,8 4,9 set-up regular stirrer 1 7,5 115,4 98 8 16,9 set-up regular stirrer 2 7,5 115,4 97,9 7,5 17,5 solvent/solid ratio 6,67 ml/g 1 7,5 39 24,2 6,1 16,2 solvent/solid ratio 6,67 ml/g 2 7,5 38,8 24,9 6,3 15,1 solvent/solid ratio 13,33 ml/g 1 7,5 78,3 68,5 5,2 12,1 solvent/solid ratio 13,33 ml/g 2 7,5 78,2 67,8 5,3 12,6 temperature 20 °C 1 7,5 115,5 110,4 9,5 3,1 temperature 20 °C 2 7,5 115,6 110,3 9,6 3,2 temperature 70 °C 1 7,5 115,7 110 8,6 4,6 temperature 70 °C 2 7,5 115,4 111,5 9,5 1,9 time 10 min 1 7,5 115,7 111,6 9 2,6 time 10 min 2 7,5 115,6 110,9 9,5 2,7 time 20 min 1 7,5 115,8 105,4 9,3 8,6 time 20 min 2 7,5 115,6 107,5 9,6 6 time 30 min 1 7,5 115,5 108,8 7,6 6,6 time 30 min 2 7,5 115,7 108,9 7,9 6,4 time 40 min 1 7,5 115,7 108,4 10,2 4,6 time 40 min 2 7,5 115,2 109,9 10,5 2,3 time 50 min 1 7,5 115,4 104 18,9 time 50 min 2 7,5 115,5 100 9,6 13,4 time 120 min 1 7,5 115,7 123,2 time 120 min 2 7,5 115,5 123 time and set-up 2 min 1 7,5 115,3 104 5,1 13,7 time and set-up 2 min 2 7,5 115,2 104,5 5 13,2 time and set-up 4 min 1 7,5 115,4 105,6 4,8 12,5 time and set-up 4 min 2 7,5 115,6 105,4 5,1 12,6 time and set-up 8 min 1 7,5 115,5 105,3 5,1 12,6 time and set-up 8 min 2 7,5 115,3 104,6 4,9 13,3 time and set-up 10 min 1 7,5 115,3 110 5,9 6,9 time and set-up 10 min 2 7,5 115,1 107,8 5,1 9,7 time and set-up 20 min 1 7,5 115,2 108,4 5,1 9,2 time and set-up 20 min 2 7,5 115,2 107,5 4,9 10,3 time and set-up 30 min 1 7,5 115,3 106,8 5 11 time and set-up 30 min 2 7,5 115,5 107,1 5 10,9 time and set-up 40 min 1 7,5 115,5 107,6 5,1 10,3 time and set-up 40 min 2 7,5 115,7 107,41 4,9 10,89 time and set-up 50 min 1 7,5 115,6 101,2 9,5 12,4 time and set-up 50 min 2 7,5 115,3 102,6 11,6 8,6 Figure 3: Overview of the evaporation step Table 2: Total Mass balance of the evaporation step. According to mass stream names from figure 3. parameter condition duplo L1 (g) CNSL (g) SO-REC (g) Losses 2 (g) none standard 1 108,3 2,9 102 3,4 none standard 2 108,1 3 105 0,1 Particle size 6-10 mm 1 110,4 3 104 3,4 Particle size 6-10 mm 2 111,2 3 100 8,2 Solvent Heptane 1 95,5 2,8 89 3,7 Solvent Heptane 2 91,9 2,6 85 4,3 stirring speed 0 rpm 1 108,1 2,6 104 1,5 stirring speed 0 rpm 2 107,6 2,6 104 1 stirring speed 1050 rpm 1 110 3,5 102 4,5 stirring speed 1050 rpm 2 108,5 2,6 107 -1,1 set-up regular stirrer 1 98 3,1 90 4,9 set-up regular stirrer 2 97,9 2,8 70 25,1 solvent/soli d ratio 6,67 ml/g 1 24,2 2,2 20 2 solvent/soli d ratio 6,67 ml/g 2 24,9 2 20 2,9 solvent/soli d ratio 13,33 ml/g 1 68,5 2,5 65 1 solvent/soli d ratio 13,33 ml/g 2 67,8 2,4 64 1,4 temperatur e 20 °C 1 110,4 2,4 70 38 temperatur e 20 °C 2 110,3 2,5 80 27,8 temperatur e 70 °C 1 110 3,2 103 3,8 temperatur e 70 °C 2 111,5 2,8 105 3,7 time 10 min 1 111,6 2,2 107 2,4 time 10 min 2 110,9 2 107 1,9 time 20 min 1 105,4 2,2 107 -3,8 time 20 min 2 107,5 2,5 104 1 time 30 min 1 108,8 2,6 104 2,2 time 30 min 2 108,9 2,5 105 1,4 time 40 min 1 108,4 2,7 104 1,7 time 40 min 2 109,9 2,8 105 2,1 time 50 min 1 104 2,6 97 4,4 time 50 min 2 100 2,8 98 -0,8 time 120 min 1 3,1 -3,1 time 120 min 2 3 95 -98 time and set-up 2 min 1 104 2,1 103 -1,1 time and set-up 2 min 2 104,5 2,2 102 0,3 time and set-up 4 min 1 105,6 2,3 103 0,3 time and set-up 4 min 2 105,4 2,3 104 -0,9 time and set-up 8 min 1 105,3 2,3 101 2 time and set-up 8 min 2 104,6 2,4 103 -0,8 time and set-up 10 min 1 110 2,3 104 3,7 time and set-up 10 min 2 107,8 2,4 104 1,4 time and set-up 20 min 1 108,4 2,5 104 1,9 time and set-up 20 min 2 107,5 2,4 105 0,1 time and set-up 30 min 1 106,8 2,7 103 1,1 time and set-up 30 min 2 107,1 2,5 104 0,6 time and set-up 40 min 1 107,6 2,7 104 0,9 time and set-up 40 min 2 107,41 2,6 105 -0,19 time and set-up 50 min 1 101,2 2,7 99 -0,5 time and set-up 50 min 2 102,6 2,6 103 -3 Figure 4: Overview of the drying step Table 3: Total Mass balance of the drying step. According mass stream names from figure 4. parameter condition duplo S1 (g) S2 (g) V1 (g) none standard 1 10 4,7 5,3 none standard 2 10 4,8 5,2 set-up regular stirrer 1 8 5,1 2,9 set-up regular stirrer 2 7,5 5 2,5 time 10 min 1 9 5,7 3,3 time 10 min 2 9,5 5,6 3,9 Appendix 2: HPLC results Figure 1: 0,1 mg/ml AA triene standard on HPLC Figure 2: 0,05 mg/ml AA triene standard on HPLC Figure 3: 0,2 mg/ml AA triene standard on HPLC Figure 4: 0,5 mg/ml AA triene standard on HPLC Figure 5: retention time CNSL sample 30 minutes, 150 ml ethanol, 7,5 gram CNS, RBR Appendix 3: GC analysis calibration curve Figure 1: Calibration curve for determination of ethanol (v/v%) left in sample after evaporation. Table 1: Calibration line and sample nº 4 before and after the distillation