Sustainable Jatropha Oil-Based Membrane with Graphene Oxide for Potential Application in Cu(II) Ion Removal from Aqueous Solution

: More recent attention has been focused on the utilization of Jatropha curcas in the field of water treatment. The potential of Jatropha oil in the synthesis of membrane for water filtration had been explored, its performance compared to the addition of graphene oxide (GO) in the polymer matrix. Jatropha oil was modified in a two-step method to produce Jatropha oil-based polyol (JOL) and was blended with hexamethylene diisocyanate (HDI) to produce Jatropha polyurethane membrane (JPU). JPU was synthesized in different conditions to obtain the optimized membrane and was blended with different GO loading to form Jatropha/graphene oxide composite membrane (JPU/GO) for performance improvement. The synthesized pristine JPU and JPU/GO were evaluated and the materials were analyzed using fourier transform infrared spectroscopy (FTIR), differential scanning calorimetry (DSC), thermogravimetric analysis (TGA), contact angle, water flux, and field emission scanning electron microscopy (FESEM). Results showed that the ratio of HDI to JOL for optimized JPU was obtained at 5:5 (v/v) with the cross-linking temperature at 90 °C and curing temperature at 150 °C. As GO was added into JPU, several changes were observed. The glass transition temperature (Tg) and onset temperature (To) increased from 58 °C to 69 °C and from 170 °C to 202 °C, respectively. The contact angle, however, decreased from 88.8° to 52.1° while the water flux improved from 223.33 L/m2·h to 523.33 L/m2·h, and the pore distribution in JPU/GO became more orderly. Filtration of copper ions using the synthesized membrane was performed to give rejection percentages between 33.51% and 71.60%. The results indicated that GO had a significant impact on JPU. Taken together, these results have suggested that JPU/GO has the potential for use in water filtration. Abstract: More recent attention has been focused on the utilization of Jatropha curcas in the ﬁeld of water treatment. The potential of Jatropha oil in the synthesis of membrane for water ﬁltration had been explored, its performance compared to the addition of graphene oxide (GO) in the polymer matrix. Jatropha oil was modiﬁed in a two-step method to produce Jatropha oil-based polyol (JOL) and was blended with hexamethylene diisocyanate (HDI) to produce Jatropha polyurethane membrane (JPU). JPU was synthesized in di ﬀ erent conditions to obtain the optimized membrane and was blended with di ﬀ erent GO loading to form Jatropha / graphene oxide composite membrane (JPU / GO) for performance improvement. The synthesized pristine JPU and JPU / GO were evaluated and the materials were analyzed using fourier transform infrared spectroscopy (FTIR), di ﬀ erential scanning calorimetry (DSC), thermogravimetric analysis (TGA), contact angle, water ﬂux, and ﬁeld emission scanning electron microscopy (FESEM). Results showed that the ratio of HDI to JOL for optimized JPU was obtained at 5:5 (v / v) with the cross-linking temperature at 90 ◦ C and curing temperature at 150 ◦ C. As GO was added into JPU, several changes were observed. The glass transition temperature (T g ) and onset temperature (T o ) increased from 58 ◦ C to 69 ◦ C and from 170 ◦ C to 202 ◦ C, respectively. The contact angle, however, decreased from 88.8 ◦ to 52.1 ◦ while the water ﬂux improved from 223.33 L / m 2 · h to 523.33 L / m 2 · h, and the pore distribution in JPU / GO became more orderly. Filtration of copper ions using the synthesized membrane was performed to give rejection percentages between 33.51% and 71.60%. The results indicated that GO had a signiﬁcant impact on JPU. Taken together, these results have suggested that JPU / GO has the potential for use in water ﬁltration.


Introduction
There has been an increasing interest in exploring environment-friendly membrane filters for water treatment from green and sustainable alternative materials. This scenario has instigated a change in the conventional approach to inorganic sources by exploiting biodegradable and fully sustainable materials such as those coming from plants and microorganisms due to the multitude uses of petroleum-derived materials [1][2][3][4]. This compound was prepared by adapting the procedure used by Salaah et al. [21]. Water was added in a 1 L flask at 10% (w/w) of total EJO. Methanol and sulfuric acid were then added into the same flask with a molar ratio of methanol to water at 5:1 while sulfuric acid was added at 0.3 wt% of the total mixture. The mixture was then heated in water bath (BUCHI B-100) with continuous stirring by using WiseStir HT50AX until the temperature reached 64 • C. EJO was then added for another 30 min before the addition of Na 2 CO 3 for quenching. The polyol was left to be cooled down before discarding the deposited residues. Methanol and water were removed through vacuum rotary evaporation at 45 • C.

Preparation of JPU Membranes
JPU membrane was prepared by modifying some procedure used by Marlina et al. [35]. Jatropha oil-based polyol (JOL) was reacted with hexamethylene diisocyanate (HDI) at three different variables as presented in Table 1. JOL and HDI were added in a 250 mL flask and stirred by using magnetic stirrer at 600 rpm while heated in water bath within 20 min. The formed polyurethane solution was cast on glass petri dish that was marked at 0.5 mm thickness and was cured for 8 h before the formed sheet of membrane was exfoliated using a spatula.

Preparation of JPU/GO Composite Membrane
From the elementary studies, the optimum ratio of JOL to HDI was obtained at 5:5 (v/v). The amount of GO was measured at 0.35 wt%, 0.50 wt%, and 0.65 wt% from the total polymer, respectively to produce three different samples. At different loadings of GO, JOL was mixed with GO through sonication. Next, three samples of GO/JOL blending were mixed with HDI separately for 20 min at 90 • C-100 • C in water bath and stirred at 600 rpm. Jatropha polyurethane with GO solution (JPU/GO) was then cast on glass petri dish (marked at 0.5 mm thickness) and was thermally cured at 150 • C for 8 h. The formed JPU/GO composite membrane was subsequently exfoliated using a spatula.

Characterization
Various JPU membranes were prepared and optimized with respect to a different ratio (v/v) of JOL to HDI, reaction temperature and curing temperature. At the same time, the membrane's texture, dryness, and solidification were observed. The optimized JPU was then introduced with a different loading of GO. The functional groups were determined by Fourier-transform infrared spectrometer (FTIR), while thermal stabilities of the membrane were determined by differential scanning calorimetry (DSC) and thermogravimetric analysis (TGA). Apart from that, field emission scanning electron microscopy (FESEM) was done to identify the morphology of the membrane, while contact angle was determined to verify the hydrophilicity of the membrane.
Porosity was studied through image processing from ImageJ, percentage water content (%), and porosity (%) calculation. Each sample was processed individually through ImageJ (version 1.5b for Windows, 64 bit, free software, National Institutes of Health, Bethesda, MD, USA). For the thresholding process, the threshold was adjusted manually through the default setting to achieve the highest degree of visualized porosity. The percentage of porosity values in terms of its Feret diameter was plotted in histogram.
The membrane samples were further analyzed by calculating the water content and the porosity percentages. To obtain the water content, both the wet membrane and dry membrane weight was measured using the following procedures. The membrane sample was soaked in water for 24 h, wiped off with tissue paper, and weighed for wet membrane weight. The membranes were then dried in an oven at 40 • C for 24 h, and then weighed to obtain dry membrane weight. Membrane water content (%) and porosity (%) were calculated by following the below equation [36]: where W w is the weight of wet membrane, W d is the weight of dry membrane, d is density of water (g/cm 3 ), A is the membrane effective area (cm 2 ), and L is the thickness of membrane (cm).

Water Flux and Heavy Metal Removal
A customized cross-flow filtration system was used to evaluate the performance of both the JPU and JPU/GO membrane, as shown in Figure 1. The membrane was measured to have an effective surface area 9 cm 2 for pristine JPU and JPU/GO samples. Initially, feed water was pumped into the membrane module containing the membrane sample using a peristaltic pump at 40 rpm. Once the module was completely filled with water, the pressure was set at 0.5 bar and the experiments were monitored in 10 min time. Water flux was determined by measuring the total water permeated through the membrane samples at selected area, time, water speed, and pressure. The flux rate was calculated by referring to the equation below [36]: where J w is the water flux (L/m 2 ·h), Q is the permeate (L), A is the effective surface area (m 2 ), and ∆t is the duration time (h).
The experiment was then continued by using heavy metal solution from copper (Cu(II)) to further confirm the performance of the membranes. 100 ppm of copper sulfate pentahydrate (CuSO 4 ·5H 2 O) was prepared as the feed to the filtration process. The experiments was conducted at 1.5 bar to give a substantial heavy metal rejection [37]. The amount of Cu(II) ions in the feed solution and in permeate were measured using atomic absorption spectroscopy (AAS) and further used in Equation (4) in order to calculate the heavy metal rejection [36,38]: where C p and C f are the heavy metal concentrations in permeate and the feed, respectively.
Processes 2020, 8, x FOR PEER REVIEW 6 of 22 1.5 bar to give a substantial heavy metal rejection [37]. The amount of Cu (II) ions in the feed solution and in permeate were measured using atomic absorption spectroscopy (AAS) and further used in Equation 4 in order to calculate the heavy metal rejection [36,38]: (4) where Cp and Cf are the heavy metal concentrations in permeate and the feed, respectively.

JPU Membrane Synthesis and Optimization
The present study was designed to determine the optimum conditions for the synthesis of dried and solidified JPU. Three factors had been selected as being potentially important: polyol to HDI ratio, reaction temperature, and curing temperature. Further analysis was done to investigate the effect of GO on the properties of JPU for the purpose of improvement.
Mainly, JPU membrane was prepared by cross-linking method between JOL and HDI in different ratios at 90°C with stirring rate at 600 rpm. The solution was then cured at 150°C for 8 h. Based on Table 2, JPU solutions at 5:1 and 5:3 (v/v) ratio were not dried and were not solidified as the amount of HDI was not sufficient enough to produce a dried membrane. This was due to the low concentration of isocyanate (-NCO) that reacted to hydroxyl (-OH) which formed primarily unstable carbamic acid (urethane link). In addition to the reaction, the primary reaction could be reverted at an elevated temperature due to the unstable bond [39][40][41][42][43][44]. By adding more HDI, at 5:7 (v/v), a solid membrane was produced. However, it was too dried and brittle. The membrane broken into small hard fragments when exfoliated from the petri dish. This could be due to the excess isocyanate (aliphatic isocyanate) loading which might form other linkages when heating to give an extremely stable and irreversible isocyanurate linkage and produce allophanate and trimerize [39,40,[43][44][45][46]. Generally, excess isocyanates are needed to form sufficient cross-linkages during secondary reactions to synthesize a dried and solidified JPU. However, over excessive isocyanate generate undesirable surplus linkages which resulted in a too dried and brittle membrane. [47][48][49].
At a ratio of 5:5 (v/v), the synthesized JPU membrane dried favorably as desired and became solidified. This view was supported by Rehman [39] who suggested that an adequate and appropriate excess amount of isocyanates was crucial to promote the cross-linking reaction during polyurethane synthesis. Thus it can be concluded that the optimum JOL:HDI ratio, in order to obtain dried, elastic, and solidified JPU membrane, was at 5:5 (v/v). Table 2. Physical appearance of JPU and JPU/GO membrane prepared at different conditions.

JPU Membrane Synthesis and Optimization
The present study was designed to determine the optimum conditions for the synthesis of dried and solidified JPU. Three factors had been selected as being potentially important: polyol to HDI ratio, reaction temperature, and curing temperature. Further analysis was done to investigate the effect of GO on the properties of JPU for the purpose of improvement.

Effect of Different JOL:HDI Ratio
Mainly, JPU membrane was prepared by cross-linking method between JOL and HDI in different ratios at 90 • C with stirring rate at 600 rpm. The solution was then cured at 150 • C for 8 h. Based on Table 2, JPU solutions at 5:1 and 5:3 (v/v) ratio were not dried and were not solidified as the amount of HDI was not sufficient enough to produce a dried membrane. This was due to the low concentration of isocyanate (-NCO) that reacted to hydroxyl (-OH) which formed primarily unstable carbamic acid (urethane link). In addition to the reaction, the primary reaction could be reverted at an elevated temperature due to the unstable bond [39][40][41][42][43][44]. By adding more HDI, at 5:7 (v/v), a solid membrane was produced. However, it was too dried and brittle. The membrane broken into small hard fragments when exfoliated from the petri dish. This could be due to the excess isocyanate (aliphatic isocyanate) loading which might form other linkages when heating to give an extremely stable and irreversible isocyanurate linkage and produce allophanate and trimerize [39,40,[43][44][45][46]. Generally, excess isocyanates are needed to form sufficient cross-linkages during secondary reactions to synthesize a dried and solidified JPU. However, over excessive isocyanate generate undesirable surplus linkages which resulted in a too dried and brittle membrane [47][48][49].
At a ratio of 5:5 (v/v), the synthesized JPU membrane dried favorably as desired and became solidified. This view was supported by Rehman [39] who suggested that an adequate and appropriate excess amount of isocyanates was crucial to promote the cross-linking reaction during polyurethane synthesis. Thus it can be concluded that the optimum JOL:HDI ratio, in order to obtain dried, elastic, and solidified JPU membrane, was at 5:5 (v/v). Curing temperature and the JOL:HDI ratio for the optimum JPU membrane were at 150°C and 5:5 (v/v), respectively. In this part, the reaction temperature for blending JOL and HDI was conducted at different values of 50°C, 70°C, and 90°C. As shown in Table 2, the JPU solution still remains as liquid at 50°C and was also not completely solidified at 70°C. It seemed possible that these results were caused by the primary reactions. At low temperature, the hydroxyl groups started to react with isocyanate to form primarily urethane linkage that was unstable. These findings corroborated the ideas of Szycher [50], who highlighted that the unstable primarily urethane linkage had started to form from room temperature. At 90°C, the JPU membrane was completely dried and easily removed as the back surface was not fractured. This was attributed to the higher cross-linking and branching through secondary reactions that occurred at the elevated temperature [50]. Since the reactions and curing process happened in a very quick overlapping succession process [50], the reaction temperature was monitored by heating the solution on a hot plate stirrer. This is also very important to avoid rapid reactions that converted the polymer into solid end products. Hence, the findings here confirmed that the optimum reaction temperature for JPU membrane was 90°C.
3.1.3. Effect of different curing temperature.
The selected JOL: HDI ratio for subsequent investigation in the synthesis of JPU membrane was at 5:5 (v/v). At this stage, the JPU membrane was produced with constant JOL: HDI ratio and heated at 90°C during cross-linking reaction but was cured at different temperature. As shown in Table 2, JPU solutions at 90°C and 120°C were dried, but the texture was brittle when exfoliated from the petri dish. These results suggested that the curing temperature was not sufficient to obtain the peak Curing temperature and the JOL:HDI ratio for the optimum JPU membrane were at 150°C and 5:5 (v/v), respectively. In this part, the reaction temperature for blending JOL and HDI was conducted at different values of 50°C, 70°C, and 90°C. As shown in Table 2, the JPU solution still remains as liquid at 50°C and was also not completely solidified at 70°C. It seemed possible that these results were caused by the primary reactions. At low temperature, the hydroxyl groups started to react with isocyanate to form primarily urethane linkage that was unstable. These findings corroborated the ideas of Szycher [50], who highlighted that the unstable primarily urethane linkage had started to form from room temperature. At 90°C, the JPU membrane was completely dried and easily removed as the back surface was not fractured. This was attributed to the higher cross-linking and branching through secondary reactions that occurred at the elevated temperature [50]. Since the reactions and curing process happened in a very quick overlapping succession process [50], the reaction temperature was monitored by heating the solution on a hot plate stirrer. This is also very important to avoid rapid reactions that converted the polymer into solid end products. Hence, the findings here confirmed that the optimum reaction temperature for JPU membrane was 90°C.
3.1.3. Effect of different curing temperature.
The selected JOL: HDI ratio for subsequent investigation in the synthesis of JPU membrane was at 5:5 (v/v). At this stage, the JPU membrane was produced with constant JOL: HDI ratio and heated at 90°C during cross-linking reaction but was cured at different temperature. As shown in Table 2, JPU solutions at 90°C and 120°C were dried, but the texture was brittle when exfoliated from the petri dish. These results suggested that the curing temperature was not sufficient to obtain the peak Curing temperature and the JOL:HDI ratio for the optimum JPU membrane were at 150°C and 5:5 (v/v), respectively. In this part, the reaction temperature for blending JOL and HDI was conducted at different values of 50°C, 70°C, and 90°C. As shown in Table 2, the JPU solution still remains as liquid at 50°C and was also not completely solidified at 70°C. It seemed possible that these results were caused by the primary reactions. At low temperature, the hydroxyl groups started to react with isocyanate to form primarily urethane linkage that was unstable. These findings corroborated the ideas of Szycher [50], who highlighted that the unstable primarily urethane linkage had started to form from room temperature. At 90°C, the JPU membrane was completely dried and easily removed as the back surface was not fractured. This was attributed to the higher cross-linking and branching through secondary reactions that occurred at the elevated temperature [50]. Since the reactions and curing process happened in a very quick overlapping succession process [50], the reaction temperature was monitored by heating the solution on a hot plate stirrer. This is also very important to avoid rapid reactions that converted the polymer into solid end products. Hence, the findings here confirmed that the optimum reaction temperature for JPU membrane was 90°C.
3.1.3. Effect of different curing temperature.
The selected JOL: HDI ratio for subsequent investigation in the synthesis of JPU membrane was at 5:5 (v/v). At this stage, the JPU membrane was produced with constant JOL: HDI ratio and heated at 90°C during cross-linking reaction but was cured at different temperature. As shown in Table 2, JPU solutions at 90°C and 120°C were dried, but the texture was brittle when exfoliated from the petri dish. These results suggested that the curing temperature was not sufficient to obtain the peak Curing temperature and the JOL:HDI ratio for the optimum JPU membrane were at 150°C and 5:5 (v/v), respectively. In this part, the reaction temperature for blending JOL and HDI was conducted at different values of 50°C, 70°C, and 90°C. As shown in Table 2, the JPU solution still remains as liquid at 50°C and was also not completely solidified at 70°C. It seemed possible that these results were caused by the primary reactions. At low temperature, the hydroxyl groups started to react with isocyanate to form primarily urethane linkage that was unstable. These findings corroborated the ideas of Szycher [50], who highlighted that the unstable primarily urethane linkage had started to form from room temperature. At 90°C, the JPU membrane was completely dried and easily removed as the back surface was not fractured. This was attributed to the higher cross-linking and branching through secondary reactions that occurred at the elevated temperature [50]. Since the reactions and curing process happened in a very quick overlapping succession process [50], the reaction temperature was monitored by heating the solution on a hot plate stirrer. This is also very important to avoid rapid reactions that converted the polymer into solid end products. Hence, the findings here confirmed that the optimum reaction temperature for JPU membrane was 90°C.
3.1.3. Effect of different curing temperature.
The selected JOL: HDI ratio for subsequent investigation in the synthesis of JPU membrane was at 5:5 (v/v). At this stage, the JPU membrane was produced with constant JOL: HDI ratio and heated at 90°C during cross-linking reaction but was cured at different temperature. As shown in Table 2, JPU solutions at 90°C and 120°C were dried, but the texture was brittle when exfoliated from the petri dish. These results suggested that the curing temperature was not sufficient to obtain the peak Curing temperature and the JOL:HDI ratio for the optimum JPU membrane were at 150°C and 5:5 (v/v), respectively. In this part, the reaction temperature for blending JOL and HDI was conducted at different values of 50°C, 70°C, and 90°C. As shown in Table 2, the JPU solution still remains as liquid at 50°C and was also not completely solidified at 70°C. It seemed possible that these results were caused by the primary reactions. At low temperature, the hydroxyl groups started to react with isocyanate to form primarily urethane linkage that was unstable. These findings corroborated the ideas of Szycher [50], who highlighted that the unstable primarily urethane linkage had started to form from room temperature. At 90°C, the JPU membrane was completely dried and easily removed as the back surface was not fractured. This was attributed to the higher cross-linking and branching through secondary reactions that occurred at the elevated temperature [50]. Since the reactions and curing process happened in a very quick overlapping succession process [50], the reaction temperature was monitored by heating the solution on a hot plate stirrer. This is also very important to avoid rapid reactions that converted the polymer into solid end products. Hence, the findings here confirmed that the optimum reaction temperature for JPU membrane was 90°C.

Effect of different curing temperature.
The selected JOL: HDI ratio for subsequent investigation in the synthesis of JPU membrane was at 5:5 (v/v). At this stage, the JPU membrane was produced with constant JOL: HDI ratio and heated at 90°C during cross-linking reaction but was cured at different temperature. As shown in Table 2, JPU solutions at 90°C and 120°C were dried, but the texture was brittle when exfoliated from the petri dish. These results suggested that the curing temperature was not sufficient to obtain the peak Curing temperature and the JOL:HDI ratio for the optimum JPU membrane were at 150°C and 5:5 (v/v), respectively. In this part, the reaction temperature for blending JOL and HDI was conducted at different values of 50°C, 70°C, and 90°C. As shown in Table 2, the JPU solution still remains as liquid at 50°C and was also not completely solidified at 70°C. It seemed possible that these results were caused by the primary reactions. At low temperature, the hydroxyl groups started to react with isocyanate to form primarily urethane linkage that was unstable. These findings corroborated the ideas of Szycher [50], who highlighted that the unstable primarily urethane linkage had started to form from room temperature. At 90°C, the JPU membrane was completely dried and easily removed as the back surface was not fractured. This was attributed to the higher cross-linking and branching through secondary reactions that occurred at the elevated temperature [50]. Since the reactions and curing process happened in a very quick overlapping succession process [50], the reaction temperature was monitored by heating the solution on a hot plate stirrer. This is also very important to avoid rapid reactions that converted the polymer into solid end products. Hence, the findings here confirmed that the optimum reaction temperature for JPU membrane was 90°C.

Effect of different curing temperature.
The selected JOL: HDI ratio for subsequent investigation in the synthesis of JPU membrane was at 5:5 (v/v). At this stage, the JPU membrane was produced with constant JOL: HDI ratio and heated at 90°C during cross-linking reaction but was cured at different temperature. As shown in Table 2, JPU solutions at 90°C and 120°C were dried, but the texture was brittle when exfoliated from the petri dish. These results suggested that the curing temperature was not sufficient to obtain the peak Curing temperature and the JOL:HDI ratio for the optimum JPU membrane were at 150°C and 5:5 (v/v), respectively. In this part, the reaction temperature for blending JOL and HDI was conducted at different values of 50°C, 70°C, and 90°C. As shown in Table 2, the JPU solution still remains as liquid at 50°C and was also not completely solidified at 70°C. It seemed possible that these results were caused by the primary reactions. At low temperature, the hydroxyl groups started to react with isocyanate to form primarily urethane linkage that was unstable. These findings corroborated the ideas of Szycher [50], who highlighted that the unstable primarily urethane linkage had started to form from room temperature. At 90°C, the JPU membrane was completely dried and easily removed as the back surface was not fractured. This was attributed to the higher cross-linking and branching through secondary reactions that occurred at the elevated temperature [50]. Since the reactions and curing process happened in a very quick overlapping succession process [50], the reaction temperature was monitored by heating the solution on a hot plate stirrer. This is also very important to avoid rapid reactions that converted the polymer into solid end products. Hence, the findings here confirmed that the optimum reaction temperature for JPU membrane was 90°C.

Effect of different curing temperature.
The selected JOL: HDI ratio for subsequent investigation in the synthesis of JPU membrane was at 5:5 (v/v). At this stage, the JPU membrane was produced with constant JOL: HDI ratio and heated at 90°C during cross-linking reaction but was cured at different temperature. As shown in Table 2, JPU solutions at 90°C and 120°C were dried, but the texture was brittle when exfoliated from the petri dish. These results suggested that the curing temperature was not sufficient to obtain the peak Curing temperature and the JOL:HDI ratio for the optimum JPU membrane were at 150°C and 5:5 (v/v), respectively. In this part, the reaction temperature for blending JOL and HDI was conducted at different values of 50°C, 70°C, and 90°C. As shown in Table 2, the JPU solution still remains as liquid at 50°C and was also not completely solidified at 70°C. It seemed possible that these results were caused by the primary reactions. At low temperature, the hydroxyl groups started to react with isocyanate to form primarily urethane linkage that was unstable. These findings corroborated the ideas of Szycher [50], who highlighted that the unstable primarily urethane linkage had started to form from room temperature. At 90°C, the JPU membrane was completely dried and easily removed as the back surface was not fractured. This was attributed to the higher cross-linking and branching through secondary reactions that occurred at the elevated temperature [50]. Since the reactions and curing process happened in a very quick overlapping succession process [50], the reaction temperature was monitored by heating the solution on a hot plate stirrer. This is also very important to avoid rapid reactions that converted the polymer into solid end products. Hence, the findings here confirmed that the optimum reaction temperature for JPU membrane was 90°C.
3.1.3. Effect of different curing temperature.
The selected JOL: HDI ratio for subsequent investigation in the synthesis of JPU membrane was at 5:5 (v/v). At this stage, the JPU membrane was produced with constant JOL: HDI ratio and heated at 90°C during cross-linking reaction but was cured at different temperature. As shown in Table 2, JPU solutions at 90°C and 120°C were dried, but the texture was brittle when exfoliated from the petri dish. These results suggested that the curing temperature was not sufficient to obtain the peak Curing temperature and the JOL:HDI ratio for the optimum JPU membrane were at 150°C and 5:5 (v/v), respectively. In this part, the reaction temperature for blending JOL and HDI was conducted at different values of 50°C, 70°C, and 90°C. As shown in Table 2, the JPU solution still remains as liquid at 50°C and was also not completely solidified at 70°C. It seemed possible that these results were caused by the primary reactions. At low temperature, the hydroxyl groups started to react with isocyanate to form primarily urethane linkage that was unstable. These findings corroborated the ideas of Szycher [50], who highlighted that the unstable primarily urethane linkage had started to form from room temperature. At 90°C, the JPU membrane was completely dried and easily removed as the back surface was not fractured. This was attributed to the higher cross-linking and branching through secondary reactions that occurred at the elevated temperature [50]. Since the reactions and curing process happened in a very quick overlapping succession process [50], the reaction temperature was monitored by heating the solution on a hot plate stirrer. This is also very important to avoid rapid reactions that converted the polymer into solid end products. Hence, the findings here confirmed that the optimum reaction temperature for JPU membrane was 90°C.
3.1.3. Effect of different curing temperature.
The selected JOL: HDI ratio for subsequent investigation in the synthesis of JPU membrane was at 5:5 (v/v). At this stage, the JPU membrane was produced with constant JOL: HDI ratio and heated at 90°C during cross-linking reaction but was cured at different temperature. As shown in Table 2, JPU solutions at 90°C and 120°C were dried, but the texture was brittle when exfoliated from the petri dish. These results suggested that the curing temperature was not sufficient to obtain the peak Curing temperature and the JOL:HDI ratio for the optimum JPU membrane were at 150°C and 5:5 (v/v), respectively. In this part, the reaction temperature for blending JOL and HDI was conducted at different values of 50°C, 70°C, and 90°C. As shown in Table 2, the JPU solution still remains as liquid at 50°C and was also not completely solidified at 70°C. It seemed possible that these results were caused by the primary reactions. At low temperature, the hydroxyl groups started to react with isocyanate to form primarily urethane linkage that was unstable. These findings corroborated the ideas of Szycher [50], who highlighted that the unstable primarily urethane linkage had started to form from room temperature. At 90°C, the JPU membrane was completely dried and easily removed as the back surface was not fractured. This was attributed to the higher cross-linking and branching through secondary reactions that occurred at the elevated temperature [50]. Since the reactions and curing process happened in a very quick overlapping succession process [50], the reaction temperature was monitored by heating the solution on a hot plate stirrer. This is also very important to avoid rapid reactions that converted the polymer into solid end products. Hence, the findings here confirmed that the optimum reaction temperature for JPU membrane was 90°C.
3.1.3. Effect of different curing temperature.
The selected JOL: HDI ratio for subsequent investigation in the synthesis of JPU membrane was at 5:5 (v/v). At this stage, the JPU membrane was produced with constant JOL: HDI ratio and heated at 90°C during cross-linking reaction but was cured at different temperature. As shown in Table 2, JPU solutions at 90°C and 120°C were dried, but the texture was brittle when exfoliated from the petri dish. These results suggested that the curing temperature was not sufficient to obtain the peak Curing temperature and the JOL:HDI ratio for the optimum JPU membrane were at 150°C and 5:5 (v/v), respectively. In this part, the reaction temperature for blending JOL and HDI was conducted at different values of 50°C, 70°C, and 90°C. As shown in Table 2, the JPU solution still remains as liquid at 50°C and was also not completely solidified at 70°C. It seemed possible that these results were caused by the primary reactions. At low temperature, the hydroxyl groups started to react with isocyanate to form primarily urethane linkage that was unstable. These findings corroborated the ideas of Szycher [50], who highlighted that the unstable primarily urethane linkage had started to form from room temperature. At 90°C, the JPU membrane was completely dried and easily removed as the back surface was not fractured. This was attributed to the higher cross-linking and branching through secondary reactions that occurred at the elevated temperature [50]. Since the reactions and curing process happened in a very quick overlapping succession process [50], the reaction temperature was monitored by heating the solution on a hot plate stirrer. This is also very important to avoid rapid reactions that converted the polymer into solid end products. Hence, the findings here confirmed that the optimum reaction temperature for JPU membrane was 90°C.
3.1.3. Effect of different curing temperature.
The selected JOL: HDI ratio for subsequent investigation in the synthesis of JPU membrane was at 5:5 (v/v). At this stage, the JPU membrane was produced with constant JOL: HDI ratio and heated at 90°C during cross-linking reaction but was cured at different temperature. As shown in Table 2, JPU solutions at 90°C and 120°C were dried, but the texture was brittle when exfoliated from the petri dish. These results suggested that the curing temperature was not sufficient to obtain the peak Curing temperature and the JOL:HDI ratio for the optimum JPU membrane were at 150°C and 5:5 (v/v), respectively. In this part, the reaction temperature for blending JOL and HDI was conducted at different values of 50°C, 70°C, and 90°C. As shown in Table 2, the JPU solution still remains as liquid at 50°C and was also not completely solidified at 70°C. It seemed possible that these results were caused by the primary reactions. At low temperature, the hydroxyl groups started to react with isocyanate to form primarily urethane linkage that was unstable. These findings corroborated the ideas of Szycher [50], who highlighted that the unstable primarily urethane linkage had started to form from room temperature. At 90°C, the JPU membrane was completely dried and easily removed as the back surface was not fractured. This was attributed to the higher cross-linking and branching through secondary reactions that occurred at the elevated temperature [50]. Since the reactions and curing process happened in a very quick overlapping succession process [50], the reaction temperature was monitored by heating the solution on a hot plate stirrer. This is also very important to avoid rapid reactions that converted the polymer into solid end products. Hence, the findings here confirmed that the optimum reaction temperature for JPU membrane was 90°C.
3.1.3. Effect of different curing temperature.
The selected JOL: HDI ratio for subsequent investigation in the synthesis of JPU membrane was at 5:5 (v/v). At this stage, the JPU membrane was produced with constant JOL: HDI ratio and heated at 90°C during cross-linking reaction but was cured at different temperature. As shown in Table 2, JPU solutions at 90°C and 120°C were dried, but the texture was brittle when exfoliated from the petri dish. These results suggested that the curing temperature was not sufficient to obtain the peak Curing temperature and the JOL:HDI ratio for the optimum JPU membrane were at 150°C and 5:5 (v/v), respectively. In this part, the reaction temperature for blending JOL and HDI was conducted at different values of 50°C, 70°C, and 90°C. As shown in Table 2, the JPU solution still remains as liquid at 50°C and was also not completely solidified at 70°C. It seemed possible that these results were caused by the primary reactions. At low temperature, the hydroxyl groups started to react with isocyanate to form primarily urethane linkage that was unstable. These findings corroborated the ideas of Szycher [50], who highlighted that the unstable primarily urethane linkage had started to form from room temperature. At 90°C, the JPU membrane was completely dried and easily removed as the back surface was not fractured. This was attributed to the higher cross-linking and branching through secondary reactions that occurred at the elevated temperature [50]. Since the reactions and curing process happened in a very quick overlapping succession process [50], the reaction temperature was monitored by heating the solution on a hot plate stirrer. This is also very important to avoid rapid reactions that converted the polymer into solid end products. Hence, the findings here confirmed that the optimum reaction temperature for JPU membrane was 90°C.
3.1.3. Effect of different curing temperature.
The selected JOL: HDI ratio for subsequent investigation in the synthesis of JPU membrane was at 5:5 (v/v). At this stage, the JPU membrane was produced with constant JOL: HDI ratio and heated at 90°C during cross-linking reaction but was cured at different temperature. As shown in Table 2, JPU solutions at 90°C and 120°C were dried, but the texture was brittle when exfoliated from the petri dish. These results suggested that the curing temperature was not sufficient to obtain the peak Curing temperature and the JOL:HDI ratio for the optimum JPU membrane were at 150°C and 5:5 (v/v), respectively. In this part, the reaction temperature for blending JOL and HDI was conducted at different values of 50°C, 70°C, and 90°C. As shown in Table 2, the JPU solution still remains as liquid at 50°C and was also not completely solidified at 70°C. It seemed possible that these results were caused by the primary reactions. At low temperature, the hydroxyl groups started to react with isocyanate to form primarily urethane linkage that was unstable. These findings corroborated the ideas of Szycher [50], who highlighted that the unstable primarily urethane linkage had started to form from room temperature. At 90°C, the JPU membrane was completely dried and easily removed as the back surface was not fractured. This was attributed to the higher cross-linking and branching through secondary reactions that occurred at the elevated temperature [50]. Since the reactions and curing process happened in a very quick overlapping succession process [50], the reaction temperature was monitored by heating the solution on a hot plate stirrer. This is also very important to avoid rapid reactions that converted the polymer into solid end products. Hence, the findings here confirmed that the optimum reaction temperature for JPU membrane was 90°C.

Effect of different curing temperature.
The selected JOL: HDI ratio for subsequent investigation in the synthesis of JPU membrane was at 5:5 (v/v). At this stage, the JPU membrane was produced with constant JOL: HDI ratio and heated at 90°C during cross-linking reaction but was cured at different temperature. As shown in Table 2, JPU solutions at 90°C and 120°C were dried, but the texture was brittle when exfoliated from the petri dish. These results suggested that the curing temperature was not sufficient to obtain the peak

Effect of Different Reaction Temperature
Curing temperature and the JOL:HDI ratio for the optimum JPU membrane were at 150 • C and 5:5 (v/v), respectively. In this part, the reaction temperature for blending JOL and HDI was conducted at different values of 50 • C, 70 • C, and 90 • C. As shown in Table 2, the JPU solution still remains as liquid at 50 • C and was also not completely solidified at 70 • C. It seemed possible that these results were caused by the primary reactions. At low temperature, the hydroxyl groups started to react with isocyanate to form primarily urethane linkage that was unstable. These findings corroborated the ideas of Szycher [50], who highlighted that the unstable primarily urethane linkage had started to form from room temperature. At 90 • C, the JPU membrane was completely dried and easily removed as the back surface was not fractured. This was attributed to the higher cross-linking and branching through secondary reactions that occurred at the elevated temperature [50]. Since the reactions and curing process happened in a very quick overlapping succession process [50], the reaction temperature was monitored by heating the solution on a hot plate stirrer. This is also very important to avoid rapid reactions that converted the polymer into solid end products. Hence, the findings here confirmed that the optimum reaction temperature for JPU membrane was 90 • C.

Effect of Different Curing Temperature
The selected JOL:HDI ratio for subsequent investigation in the synthesis of JPU membrane was at 5:5 (v/v). At this stage, the JPU membrane was produced with constant JOL:HDI ratio and heated at 90 • C during cross-linking reaction but was cured at different temperature. As shown in Table 2, JPU solutions at 90 • C and 120 • C were dried, but the texture was brittle when exfoliated from the petri dish. These results suggested that the curing temperature was not sufficient to obtain the peak performance from the cross-linking reactions. As stated by Zhao [49], other side reactions were insignificant during branching reactions but were likely to be crucial during curing session to have a notable polyurethane properties. In comparison to result at 150 • C, the JPU solution became dried, solidified, and yellowish in color due to the increased cross-linking reactions at elevated curing temperature. Curing process could enhance the achievement of polyurethane synthesis [49] as the slight excess of isocyanate had promoted higher formation of branches [39]. However, as the curing temperature was increased to 170 • C, the JPU solution was excessively dried and formed a scorch mark due to the tendency of thermo-oxidative degradation reaction [42]. Taken together, these results indicated that the optimum JPU membrane was at 150 • C during curing reaction.

Effect of Different Amount of GO
As the optimized pristine JPU was obtained, GO was then introduced in the JPU matrix to enhance its properties. A recent study on the effect of GO in polyethersulfone (PES) membranes by Zinadini et al. [51] had reported that 1 wt% of GO loading, resulted in an increment of the contact angle due to the aggregation of GO. Hence, lower amount of GO loading was selected. JPU/GO samples were prepared with different percentage loading of GO from the total polymer at 0.35 wt%, 0.50 wt%, and 0.65 wt%. From the observations, there was no significant difference when comparing the three JPU/GO samples with pristine JPU, as shown in Table 2. Further data collections were done in characterizations to determine exactly how GO had affected the properties of JPU membrane.

Membrane Characterization
Properties of EJO, JOL, and pristine JPU were studied to verify the synthesis of optimum Jatropha polyurethane, while JPU/GO samples were studied to determine the effect of GO loading on the properties of Jatropha polyurethane in three different parameters: 0.35 wt%, 0.50 wt%, and 0.65 wt% of GO from the total polymer. The significant changes of the properties were verified with a few characterizations that comprised of FTIR, DSC, TGA, FESEM, and contact angle. The functional groups were determined from FTIR, while thermal stabilities of the membrane were determined from DSC and TGA. Apart from that, FESEM was done to identify the changes in membrane morphology. The hydrophilicity changes of the membrane were analyzed from the contact angle. For the porosity, it was studied through image processing from ImageJ, water content (%), and porosity (%) calculation. The experiment was duplicated, and average values were taken.

Fourier-Transform Infrared Spectroscopy (FTIR)
The chemical structures of EJO, JOL, and JPU were characterized by identifying the functional groups through FTIR analysis. The FTIR results had shown the changes of FTIR peaks and patterns for every synthesis due to further reactions, as recorded in Figure 2. The synthesis of epoxy was observed from the absence spectra of the double bond at 3008 cm −1 [52][53][54] and the appearance of a new peak (-COC) at 835 cm −1 [18,21]. The undesirable side reactions in JPU were prompted by the presence of water and acid [16,21]. In EJO spectra, a minimum side reaction of ring opening was observed at 3464 cm −1 with weak and broad tip due to the formation of hydroxyl (-OH). Carbonyl group (C=O) in ester was also found at 1742 cm −1 in EJO [18,[52][53][54]. While in JOL spectra, the hindrance of hydroxyl (-OH) band in EJO with the appearance of higher intensity absorbance band at 3431 cm −1 corresponded to the formation of hydroxyl group (-OH) due to the ring opening reaction. The carbonyl group (C=O) of ester in JOL was found to be shifted at 1741cm −1 . With the absence of (N=C=O) group band in the range at 2254 cm −1 to 2310 cm −1 , the isocyanate group was confirmed to be completely reacted [55]. In JPU spectra, the functional groups were observed in four main peaks: N-H group stretching at 3336 cm −1 , urethane carbonyl group (C=O) for urethane linkage at 1705 cm −1 , (C-O) stretching at 1254 cm −1 , and (N-H) bending at 1542 cm −1 . The finding of hydrogen-bonded carbonyl group (C=O) at 1705 cm −1 seemed to be consistent with other researches [25,56]. Apart from that, a sharp peak at 2922 cm −1 and 2854 cm −1 corresponded to the (-CH) stretching as other bands of (-CH) vibrations at 1463 cm −1 and 1373 cm −1 [25]. As reported by Ahmad et al. [16], bands from 1100 cm −1 to 1200 cm −1 were the spectral ranges for polyether and polyester.
The functional group of JPU/GO samples was also studied through FTIR to determine the effect of GO in Jatropha polyurethane properties. Apparently, isocyanate was prompted to have reaction with the functional groups of GO as side reaction, whether from the edge carboxylic acid, surface hydroxyl, or epoxide in producing amides or carbamate esters [57]. As stated by Sun et al. [58], the sequence of isocyanate reactivity with different functional group was in the sequence as primary amine > secondary amine > hydroxyl > acid > anhydride > epoxide. Due to the reactivity of isocyanate versus different functional group, isocyanate had higher reactivity with the hydroxyl group compared to the other functional group of GO. In JPU/GO samples, the hydroxyl (-OH) band was While in JOL spectra, the hindrance of hydroxyl (-OH) band in EJO with the appearance of higher intensity absorbance band at 3431 cm −1 corresponded to the formation of hydroxyl group (-OH) due to the ring opening reaction. The carbonyl group (C=O) of ester in JOL was found to be shifted at 1741cm −1 . With the absence of (N=C=O) group band in the range at 2254 cm −1 to 2310 cm −1 , the isocyanate group was confirmed to be completely reacted [55]. In JPU spectra, the functional groups were observed in four main peaks: N-H group stretching at 3336 cm −1 , urethane carbonyl group (C=O) for urethane linkage at 1705 cm −1 , (C-O) stretching at 1254 cm −1 , and (N-H) bending at 1542 cm −1 . The finding of hydrogen-bonded carbonyl group (C=O) at 1705 cm −1 seemed to be consistent with other researches [25,56]. Apart from that, a sharp peak at 2922 cm −1 and 2854 cm −1 corresponded to the (-CH) stretching as other bands of (-CH) vibrations at 1463 cm −1 and 1373 cm −1 [25]. As reported by Ahmad et al. [16], bands from 1100 cm −1 to 1200 cm −1 were the spectral ranges for polyether and polyester.
The functional group of JPU/GO samples was also studied through FTIR to determine the effect of GO in Jatropha polyurethane properties. Apparently, isocyanate was prompted to have reaction with the functional groups of GO as side reaction, whether from the edge carboxylic acid, surface hydroxyl, or epoxide in producing amides or carbamate esters [57]. As stated by Sun et al. [58], the sequence of isocyanate reactivity with different functional group was in the sequence as primary amine > secondary amine > hydroxyl > acid > anhydride > epoxide. Due to the reactivity of isocyanate versus different functional group, isocyanate had higher reactivity with the hydroxyl group compared to the other  [59][60][61][62]. However, (-OH) stretching band in JPU/GO 0.65 wt% sample was reduced with the overlapping of (-NH) stretching. This was due to the higher GO functionalized as GO loading increased [63]. The appearance of another peak was observed near 867 cm −1 , 864 cm −1 , and 870 cm −1 for JPU/GO 0.35 wt%, JPU/GO 0.50 wt%, and JPU/GO 0.65 wt% samples, respectively, which was predicted to be (-CH) vibration. The reactivity of isocyanate toward hydroxyl (-OH) groups also depended on whether they were free or hydrogen bonded due to the higher activation energies required for the reaction as provided by hydrogen-bonded alcohols than free alcohols [63][64][65]. Free hydroxyl groups could be observed in the spectra range near 3821 cm −1 , 3852 cm −1 , and 3843 cm −1 for JPU/GO 0.35 wt%, JPU/GO 0.50 wt%, and JPU/GO 0.65 wt% samples, respectively [59].

DSC
DSC was used to analyze the thermal stability of the polymer membrane. DSC technique was able to spot the changes of glass transition temperature (T g ) in the polymer membrane. From DSC thermogram, all JPU samples were showing a low-intensity peak in the range of 58 • C to 69 • C as shown in Figure 3. JPU/GO membranes had higher T g compared to pristine JPU and T g increased with the increased of GO loading as in the following order: JPU/GO 0.65 wt% > JPU/GO 0.50 wt% > JPU/GO 0.35 wt% > JPU ( Table 3). The thermal stabilities of the JPU/GO samples increased and were compatible with the decomposition temperature of TGA thermograms. As reported by Avishek et al. [66], the uniform distribution of the nanofillers within the polymer had caused the increment of thermal stability due to the restraining of the flexibility and mobility of the polymer chains.
predicted to be (-CH) vibration. The reactivity of isocyanate toward hydroxyl (-OH) groups also depended on whether they were free or hydrogen bonded due to the higher activation energies required for the reaction as provided by hydrogen-bonded alcohols than free alcohols [63][64][65]. Free hydroxyl groups could be observed in the spectra range near 3821 cm −1 , 3852 cm −1 , and 3843 cm −1 for JPU/GO 0.35 wt%, JPU/GO 0.50 wt% , and JPU/GO 0.65 wt% samples, respectively [59].

DSC.
DSC was used to analyze the thermal stability of the polymer membrane. DSC technique was able to spot the changes of glass transition temperature (Tg) in the polymer membrane. From DSC thermogram, all JPU samples were showing a low-intensity peak in the range of 58°C to 69°C as shown in Figure 3. JPU/GO membranes had higher Tg compared to pristine JPU and Tg increased with the increased of GO loading as in the following order: JPU/GO 0.65 wt% > JPU/GO 0.50 wt% > JPU/GO 0.35 wt% > JPU ( Table 3). The thermal stabilities of the JPU/GO samples increased and were compatible with the decomposition temperature of TGA thermograms. As reported by Avishek et al. [66], the uniform distribution of the nanofillers within the polymer had caused the increment of thermal stability due to the restraining of the flexibility and mobility of the polymer chains.

Thermogravimetric Analysis (TGA)
Results from TGA provided the degradation peaks versus temperature with residue weight of the JPU membrane. Thermal stability is defined as the stability of a molecule at high temperature which is affected by the mobility of the polymeric chains. With the formation of side chain, the decomposition of material became difficult due to the steric hindrance. At high cross-linking, the polymeric membrane became thermally stable at elevated temperature with the char formation [16]. TGA curve displayed four degradation peaks for pristine JPU, as shown in Figure 4. In the first step, the initial low weight loss began with the removal of water, while the second step was from the easily altered urethane group. The third and fourth steps indicated the degradation of other already formed compounds such as primary or secondary amine as well as olefin and isocyanurate with the continuous evolution of the volatile compound due to the elevated temperature [67,68]; lastly was the char formation [16,[69][70][71]. However, degradation of polyurethane usually involved complex multistep of consecutive [72]. The general decomposition of urethane can be explained based on the equation below [73] with R' as aromatic constituent and R" and R"' as aliphatic constituents: Processes 2020, 8, x FOR PEER REVIEW 11 of 22

Thermogravimetric analysis (TGA).
Results from TGA provided the degradation peaks versus temperature with residue weight of the JPU membrane. Thermal stability is defined as the stability of a molecule at high temperature which is affected by the mobility of the polymeric chains. With the formation of side chain, the decomposition of material became difficult due to the steric hindrance. At high cross-linking, the polymeric membrane became thermally stable at elevated temperature with the char formation [16]. TGA curve displayed four degradation peaks for pristine JPU, as shown in Figure 4. In the first step, the initial low weight loss began with the removal of water, while the second step was from the easily altered urethane group. The third and fourth steps indicated the degradation of other already formed compounds such as primary or secondary amine as well as olefin and isocyanurate with the continuous evolution of the volatile compound due to the elevated temperature [67,68]; lastly was the char formation [16,[69][70][71]. However, degradation of polyurethane usually involved complex multistep of consecutive [72]. The general decomposition of urethane can be explained based on the equation below [73] with R' as aromatic constituent and R'' and R''' as aliphatic constituents: Interestingly, a new peak from JPU/GO samples was detected between peak 2 and peak 3 when compared to pristine JPU, as shown in Figure 5. These results matched with those observed in earlier studies on other modified GO derivative [63] and suggested that new degradation at peak 3 was formed due to the pyrolysis of surface oxygen functional groups from GO; namely, epoxide, hydroxyl, and carboxylic acid. As recorded in Table 4, both JPU/GO 0.65 wt% and JPU/GO 0.50 wt% had higher degradation at step 3 than JPU/GO 0.35 wt% as the GO loading increased. By observing the degradation rate at step 4, the weight loss of JPU/GO samples were higher than pristine JPU due to the pyrolysis of covalently grafted isocyanate to GO, as well as other evolution of volatile compound in JPU composite. During reaction, GO was functionalized as a side reaction between hydroxyl groups of GO when exposed with isocyanate [29], forming a composite with polymer membrane [30][31][32][33][34]. There were similarities between the degradation at step 4 described in this result Interestingly, a new peak from JPU/GO samples was detected between peak 2 and peak 3 when compared to pristine JPU, as shown in Figure 5. These results matched with those observed in earlier studies on other modified GO derivative [63] and suggested that new degradation at peak 3 was formed due to the pyrolysis of surface oxygen functional groups from GO; namely, epoxide, hydroxyl, and carboxylic acid. As recorded in Table 4, both JPU/GO 0.65 wt% and JPU/GO 0.50 wt% had higher degradation at step 3 than JPU/GO 0.35 wt% as the GO loading increased. By observing the degradation rate at step 4, the weight loss of JPU/GO samples were higher than pristine JPU due to the pyrolysis of covalently grafted isocyanate to GO, as well as other evolution of volatile compound in JPU composite. During reaction, GO was functionalized as a side reaction between hydroxyl groups of GO when exposed with isocyanate [29], forming a composite with polymer membrane [30][31][32][33][34]. There were similarities between the degradation at step 4 described in this result and those expressed from the previous studies [63,74]. The remaining step 5 showed the char formation. and those expressed from the previous studies [63,74]. The remaining step 5 showed the char formation.  Other notable results were based on the initial onset temperatures (To). To for JPU/GO composites were higher than the pristine JPU due to the more stable thermal stability of JPU/GO composites, as shown in Table 5. For maximum temperature (Tmax), Tmax from JPU/GO samples were shifted to a higher temperature compared to pristine JPU by following the increment of GO loading in the JPU matrix. The introduction of GO attributed to the rigidity of the polymer, which led to the restriction motion of the cross-linked chain [75].

Sample
Degradation Rate (%) Step 1 Step 2 Step 3 Step 4 Step 5 Residue Other notable results were based on the initial onset temperatures (T o ). T o for JPU/GO composites were higher than the pristine JPU due to the more stable thermal stability of JPU/GO composites, as shown in Table 5. For maximum temperature (T max ), T max from JPU/GO samples were shifted to a higher temperature compared to pristine JPU by following the increment of GO loading in the JPU matrix. The introduction of GO attributed to the rigidity of the polymer, which led to the restriction motion of the cross-linked chain [75].  Figure 6a,b, pristine JPU and JPU/GO 0.35 wt% samples had a disorderly sponge's pores with huge different sizes of pores. As the loading of GO in JPU was increased to 0.50 wt% from the total polymer, the arrangement of pores was orderly altered by GO nanofillers, as shown in Figure 6c. Figure 6d shows that the higher loading of GO up to 0.65 wt% from the total polymer was beyond the optimum level as the pores decreased.
Processes 2020, 8, x FOR PEER REVIEW 13 of 22 0.35 wt% samples had a disorderly sponge's pores with huge different sizes of pores. As the loading of GO in JPU was increased to 0.50 wt% from the total polymer, the arrangement of pores was orderly altered by GO nanofillers, as shown in Figure 6(c). Figure 6(d) shows that the higher loading of GO up to 0.65 wt% from the total polymer was beyond the optimum level as the pores decreased. It should be noted that well dispersion of GO in the polymer matrix was crucial for enhancing membrane properties. Ionita et al. [76] had found that high loading of GO to the polymer matrices only initiated poor dispersion of GO due to aggregation. Thereby, higher loading of GO has caused low porosity for the JPU/GO 0.65 wt% sample as well as increased the membrane surface roughness [77]. Ammar et al. [78] stated that the viscosity of the polymer solution could be affected by the addition of GO. The pore size of the membrane was reduced due to the higher viscosity of the polymer solutions as it exceeded the optimum loading of GO. The findings were supported by Chien et al. [79] in their studies for the addition of sulfonated graphene oxide (SGO) to Nafion polymer matrices. They stated that the solution viscosity of the polymer was increased with the increased SGO loading. This is due to the cooperative bonds interaction that occurred between the Nafion and SGO functional group.
3.2.5. Contact angle. It should be noted that well dispersion of GO in the polymer matrix was crucial for enhancing membrane properties. Ionita et al. [76] had found that high loading of GO to the polymer matrices only initiated poor dispersion of GO due to aggregation. Thereby, higher loading of GO has caused low porosity for the JPU/GO 0.65 wt% sample as well as increased the membrane surface roughness [77]. Ammar et al. [78] stated that the viscosity of the polymer solution could be affected by the addition of GO. The pore size of the membrane was reduced due to the higher viscosity of the polymer solutions as it exceeded the optimum loading of GO. The findings were supported by Chien et al. [79] in their studies for the addition of sulfonated graphene oxide (SGO) to Nafion polymer matrices. They stated that the solution viscosity of the polymer was increased with the increased SGO loading. This is due to the cooperative bonds interaction that occurred between the Nafion and SGO functional group.

Contact Angle
Data analysis was based on the angle of water droplet's shape from the surface of material. Lower contact angle indicates the tendency of the surface to be more hydrophilic. Contact angle data was recorded to monitor the changes of the surface wettability from the introduction of GO as illustrated  Figure 7. According to this result, the contact angle for pristine JPU did not change for the whole 8 min and remained constant at 88.8 • . After the loading of GO in JPU, the contact angle for JPU/GO samples decreased within 8 min, as in Table 6. 0.50 wt% JPU/GO sample with contact angle of 52.1 • became the most hydrophilic sample. As explained earlier, the addition of GO during the cross-linking of JOL and HDI have caused side reactions that decreased the hydrophilicity. Typically, high reaction temperature led to lower degree of functionalization between GO and isocyanate since the reaction of n-alcohol with phenyl isocyanate is inversely proportional to temperature [80]. In these experiments, the reaction between alcohol and isocyanate was exothermic. The increased of reaction temperature resulted in the decreased yield of functionalized isocyanate-GO products [63] that led to more untreated hydroxyl from GO surface at the end of the reaction. The hydrophilicity (decreasing contact angle) was seen to improve as the loading of GO to the polymer matrix increased to give a more orderly porous surface [81] since the oxygen groups in GO reduced the van der Waals forces and enabled water molecules to penetrate through the interlayer [29]. Data analysis was based on the angle of water droplet's shape from the surface of material. Lower contact angle indicates the tendency of the surface to be more hydrophilic. Contact angle data was recorded to monitor the changes of the surface wettability from the introduction of GO as illustrated in Figure 7. According to this result, the contact angle for pristine JPU did not change for the whole 8 min and remained constant at 88.8°. After the loading of GO in JPU, the contact angle for JPU/GO samples decreased within 8 min, as in Table 6. 0.50 wt% JPU/GO sample with contact angle of 52.1° became the most hydrophilic sample. As explained earlier, the addition of GO during the cross-linking of JOL and HDI have caused side reactions that decreased the hydrophilicity. Typically, high reaction temperature led to lower degree of functionalization between GO and isocyanate since the reaction of n-alcohol with phenyl isocyanate is inversely proportional to temperature [80]. In these experiments, the reaction between alcohol and isocyanate was exothermic. The increased of reaction temperature resulted in the decreased yield of functionalized isocyanate-GO products [63] that led to more untreated hydroxyl from GO surface at the end of the reaction. The hydrophilicity (decreasing contact angle) was seen to improve as the loading of GO to the polymer matrix increased to give a more orderly porous surface [81] since the oxygen groups in GO reduced the van der Waals forces and enabled water molecules to penetrate through the interlayer [29].  On the other hand, as the GO loading of the membrane was continued to increase to 0.65 wt%, the contact angle surprisingly increased. It should be noted that the occupancy of GO in the polymer matrix might affect the pore size and the pore distribution inside the membrane framework [78,82]. According to Zhao et al. [83] in their studies about the loading of isocyanate-treated graphene oxide (iGO) to polysulfone (PSf) membrane, the water flux decreased with the increased iGO. This was due to the decrease in pore size and the altered pore structure. The hydrophilic properties were still apparent for JPU/GO 0.65 wt%, but more time is required to decrease the contact angle at the optimum value. In accordance with the present results, previous studies on polyethersulfone/GO composite membrane by Zinadini et al. [51] had demonstrated that the contact angle increased with the loading of GO beyond the optimum level. These findings further supported that the increased contact angle was due to aggregation of nanoparticles.
As shown in Figure 8, larger and disordered pores were observed from pristine JPU and JPU/GO 0.35 wt% membrane samples, while more orderly and smaller pores were observed from JPU/GO 0.50 wt% membrane samples. However, as the GO loading increased to 0.65 wt%, the porosity reduced. From ImageJ analysis, the Feret diameter of the porosity was recorded and exported into  On the other hand, as the GO loading of the membrane was continued to increase to 0.65 wt%, the contact angle surprisingly increased. It should be noted that the occupancy of GO in the polymer matrix might affect the pore size and the pore distribution inside the membrane framework [78,82]. According to Zhao et al. [83] in their studies about the loading of isocyanate-treated graphene oxide (iGO) to polysulfone (PSf) membrane, the water flux decreased with the increased iGO. This was due to the decrease in pore size and the altered pore structure. The hydrophilic properties were still apparent for JPU/GO 0.65 wt%, but more time is required to decrease the contact angle at the optimum value. In accordance with the present results, previous studies on polyethersulfone/GO composite membrane by Zinadini et al. [51] had demonstrated that the contact angle increased with the loading of GO beyond the optimum level. These findings further supported that the increased contact angle was due to aggregation of nanoparticles.

Porosity
As shown in Figure 8, larger and disordered pores were observed from pristine JPU and JPU/GO 0.35 wt% membrane samples, while more orderly and smaller pores were observed from JPU/GO 0.50 wt% membrane samples. However, as the GO loading increased to 0.65 wt%, the porosity reduced. From ImageJ analysis, the Feret diameter of the porosity was recorded and exported into histogram.
As shown in Figure 9, the visualized porosity of PU/GO 0.50 wt% membrane sample revealed the optimum level of desired porosity with the higher amount of Feret diameter below 300 µm at 22 counts, followed by PU/GO 0.35 wt% at 19 counts, while PU/GO 0.65 wt% and JPU are both at 18 counts with the presence of pores at 600-700 µm Feret diameter for pristine JPU. As indicated in Figure 10, membrane water content and porosity increased significantly with the higher content of GO loading, but the optimum result was achieved by JPU/GO 0.50 wt% with 39.42% of water content and 93.14% of porosity. Membrane sample with the optimum amount of GO loading had shown the greatest water content. This phenomenon proved that the increment of GO loading in the membrane structure had led to the higher porosity of the membrane. However, as the loading of GO exceeded the optimum level, the water content and porosity percentage decreased. This degraded porosity might be due to the agglomeration of extra GO loading in the polymer matrix [84][85][86]. histogram. As shown in Figure 9, the visualized porosity of PU/GO 0.50 wt% membrane sample revealed the optimum level of desired porosity with the higher amount of Feret diameter below 300 m at 22 counts, followed by PU/GO 0.35 wt% at 19 counts, while PU/GO 0.65 wt% and JPU are both at 18 counts with the presence of pores at 600-700 m Feret diameter for pristine JPU. As indicated in Figure 10, membrane water content and porosity increased significantly with the higher content of GO loading, but the optimum result was achieved by JPU/GO 0.50 wt% with 39.42% of water content and 93.14% of porosity. Membrane sample with the optimum amount of GO loading had shown the greatest water content. This phenomenon proved that the increment of GO loading in the membrane structure had led to the higher porosity of the membrane. However, as the loading of GO exceeded the optimum level, the water content and porosity percentage decreased. This degraded porosity might be due to the agglomeration of extra GO loading in the polymer matrix [84][85][86].  histogram. As shown in Figure 9, the visualized porosity of PU/GO 0.50 wt% membrane sample revealed the optimum level of desired porosity with the higher amount of Feret diameter below 300 m at 22 counts, followed by PU/GO 0.35 wt% at 19 counts, while PU/GO 0.65 wt% and JPU are both at 18 counts with the presence of pores at 600-700 m Feret diameter for pristine JPU. As indicated in Figure 10, membrane water content and porosity increased significantly with the higher content of GO loading, but the optimum result was achieved by JPU/GO 0.50 wt% with 39.42% of water content and 93.14% of porosity. Membrane sample with the optimum amount of GO loading had shown the greatest water content. This phenomenon proved that the increment of GO loading in the membrane structure had led to the higher porosity of the membrane. However, as the loading of GO exceeded the optimum level, the water content and porosity percentage decreased. This degraded porosity might be due to the agglomeration of extra GO loading in the polymer matrix [84][85][86].

Membrane performance
Membrane performance was studied through water flux and heavy metal rejection from aqueous solutions. The concentration of Cu in feedstock and permeate was analyzed with AAS. The experiments were done in a customized cross-flow system membrane filtration and were duplicated for the average values.

Water flux.
As recorded in Figure 11, JPU/GO samples gave higher water flux than pristine JPU. These findings were consistent with those of Hu et al. who reported the fabrication of GO membrane with layer-by-layer method on a polydopamine-coated polysulfone and 1,3,5-benzenetricarbonyl trichloride as cross-linker which had higher water flux than most commercial nanofiltration (NF) membranes [87]. Similarly, Yin and coworkers found that GO-enhanced polyamide thin-film nanocomposite (TFN) membrane had higher water flux compared to TFN membrane containing MCM-41 silica nanoparticles with 59.4 ± 0.4 L/m 2 .h and 46.6 L/m 2 .h, respectively [88,89]. The

Membrane Performance
Membrane performance was studied through water flux and heavy metal rejection from aqueous solutions. The concentration of Cu in feedstock and permeate was analyzed with AAS. The experiments were done in a customized cross-flow system membrane filtration and were duplicated for the average values.

Water Flux
As recorded in Figure 11, JPU/GO samples gave higher water flux than pristine JPU. These findings were consistent with those of Hu et al. who reported the fabrication of GO membrane with layer-by-layer method on a polydopamine-coated polysulfone and 1,3,5-benzenetricarbonyl trichloride as cross-linker which had higher water flux than most commercial nanofiltration (NF) membranes [87]. Similarly, Yin and coworkers found that GO-enhanced polyamide thin-film nanocomposite (TFN) membrane had higher water flux compared to TFN membrane containing MCM-41 silica nanoparticles with 59.4 ± 0.4 L/m 2 ·h and 46.6 L/m 2 ·h, respectively [88,89]. The amphiphilic nature of GO had formed water channel as water molecules were first absorbed onto the hydrophilic sites to allow water diffusion onto the hydrophobic carbon sites [84,90]. Similarly, in the case of polysulfone-nanostructured reduced graphene oxide (Ps-nRGO) composite UF membrane, loading of nRGO in Ps matrix had formed automatically smooth interconnected interlayer that caused higher water flux in Ps-nRGO than pristine Ps NF membrane [66].
However, as loading of GO in JPU matrix increased to 0.65 wt%, water flux decreased to 260 L/m 2 ·h, which is almost similar to pristine JPU at 223.33 L/m 2 ·h. In contrast to JPU/GO samples with 0.35 wt% and 0.50 wt% GO loading in JPU matrix, higher water flux was recorded at 406.66 L/m 2 ·h and 523.33 L/m 2 ·h, respectively. These results were due to the GO sheet agglomeration in the polymer matrix that reduced the hydrophilicity and defected the membrane porosity [84][85][86]. Thus, the loading of GO in polymer matrix was crucial for enhancing JPU/GO membrane. The increase of water flux for JPU/GO samples compared to pristine JPU was consistent with the contact angle analysis as discussed earlier. amphiphilic nature of GO had formed water channel as water molecules were first absorbed onto the hydrophilic sites to allow water diffusion onto the hydrophobic carbon sites [84,90]. Similarly, in the case of polysulfone-nanostructured reduced graphene oxide (Ps-nRGO) composite UF membrane, loading of nRGO in Ps matrix had formed automatically smooth interconnected interlayer that caused higher water flux in Ps-nRGO than pristine Ps NF membrane [66]. However, as loading of GO in JPU matrix increased to 0.65 wt%, water flux decreased to 260 L/m 2 .h, which is almost similar to pristine JPU at 223.33 L/m 2 h. In contrast to JPU/GO samples with 0.35 wt% and 0.50 wt% GO loading in JPU matrix, higher water flux was recorded at 406.66 L/m 2 .h and 523.33 L/m 2 .h, respectively. These results were due to the GO sheet agglomeration in the polymer matrix that reduced the hydrophilicity and defected the membrane porosity [84][85][86]. Thus, the loading of GO in polymer matrix was crucial for enhancing JPU/GO membrane. The increase of water flux for JPU/GO samples compared to pristine JPU was consistent with the contact angle analysis as discussed earlier.

Cu (II) ions removal
The performance of membranes for Cu (II) ions removal was shown in Figure 12. JPU/GO exhibited higher rejection of heavy metals than pristine JPU due to the presence of GO. Various oxygenated functionalities on GO surface have increased the effectiveness in heavy metal removal [91] through the sharing of the lone electron pair on oxygen that binds with the metal ion to form a metal complex [92] compared to pristine membrane. A recent study reported the ability of GOimpregnated mixed matrix membrane using polysulfone (PSF)  . However, as loading of GO in JPU matrix increased to 0.65 wt%, Cu (II) rejection decreased. In membrane studies, pore size and electrostatic interaction with membrane surface were the two factors affecting the solute rejections [93]. By increasing the loading of GO, pore sizes of the membrane decreased due to the agglomeration of GO and became the influencing factor in lowering membrane performance. Thus, the solute rejection decreased with higher loading of GO beyond the optimum level. In the present study, the order of Cu (II) ions rejection is as follows: JPU/GO 0.50 wt% > JPU/GO 0.35 wt% > JPU/GO 0.65 wt% > JPU with rejections values of 71.60 %, 67.56 %, 63.58 %, and 33.51 %, respectively. Taken together, the results indicated that JPU/GO 0.50 wt% have the potential to be used as a membrane filtration for heavy metal rejection.

Cu(II) Ions Removal
The performance of membranes for Cu(II) ions removal was shown in Figure 12. JPU/GO exhibited higher rejection of heavy metals than pristine JPU due to the presence of GO. Various oxygenated functionalities on GO surface have increased the effectiveness in heavy metal removal [91] through the sharing of the lone electron pair on oxygen that binds with the metal ion to form a metal complex [92] compared to pristine membrane. A recent study reported the ability of GO-impregnated mixed matrix membrane using polysulfone (PSF)  . However, as loading of GO in JPU matrix increased to 0.65 wt%, Cu(II) rejection decreased. In membrane studies, pore size and electrostatic interaction with membrane surface were the two factors affecting the solute rejections [93]. By increasing the loading of GO, pore sizes of the membrane decreased due to the agglomeration of GO and became the influencing factor in lowering membrane performance. Thus, the solute rejection decreased with higher loading of GO beyond the optimum level. In the present study, the order of Cu(II) ions rejection is as follows: JPU/GO 0.50 wt% > JPU/GO 0.35 wt% > JPU/GO 0.65 wt% > JPU with rejections values of 71.60%, 67.56%, 63.58%, and 33.51%, respectively. Taken together, the results indicated that JPU/GO 0.50 wt% have the potential to be used as a membrane filtration for heavy metal rejection.

Conclusions
Jatropha oil-based polyol (JOL) was obtained from the modification of CJO through epoxidation and ring opening reaction. JPU was obtained from the blending of JOL and HDI. The addition of GO into JPU matrix remarkably improved the membrane properties in JPU/GO. JPU/GO composite membrane showed better thermal stability, porosity, hydrophilicity, and higher water flux, as well as Cu (II) ions removal compared to pristine JPU membrane. However, excessive addition of GO adversely affected the properties of JPU/GO membrane. This study is the preliminary experiment in revealing the capability of JPU/GO membrane for water treatment. Other further studies are required in the future to ensure good water permeation. Nonetheless, these results indicate JPU/GO samples have the potential to be exploited for membrane filtration in water treatment.

Conflicts of Interest:
The authors declare that there is no conflict of interest regarding the publication of this paper.

Conclusions
Jatropha oil-based polyol (JOL) was obtained from the modification of CJO through epoxidation and ring opening reaction. JPU was obtained from the blending of JOL and HDI. The addition of GO into JPU matrix remarkably improved the membrane properties in JPU/GO. JPU/GO composite membrane showed better thermal stability, porosity, hydrophilicity, and higher water flux, as well as Cu(II) ions removal compared to pristine JPU membrane. However, excessive addition of GO adversely affected the properties of JPU/GO membrane. This study is the preliminary experiment in revealing the capability of JPU/GO membrane for water treatment. Other further studies are required in the future to ensure good water permeation. Nonetheless, these results indicate JPU/GO samples have the potential to be exploited for membrane filtration in water treatment.