A Technical Analysis of Solid Recovered Fuel from Torreﬁed Jatropha Seed Residue via a Two-Stage Mechanical Screw Press and Solvent Extraction Process

: This study investigated the torrefaction of de-oiled Jatropha seed residue after a two-stage sequential process consisting of mechanical screw pressing and solvent extraction using n-hexane (denoted as JMS). The optimal torrefaction temperature (T r ) and torrefaction time (t r ) were determined in the ranges of 260–300 ◦ C and 10–60 min, respectively, so to achieve a better heating value and satisfactory energy densiﬁcation (E D ) with acceptable mass loss. Thermogravimetric analysis was employed to elucidate the thermal decomposition behaviors of JMS. By comparison with the torrefaction of Jatropha seed residue after mechanical oil extraction by screw pressing only (namely, JME T ), the results indicated that the E D of the torrefaction of JMS yielding the torreﬁed product JMS T (two-stage product) was higher than that of the torrefaction of JME giving the torreﬁed product JME T (single-stage product). Further, it was found that JME T contained some tar, which was attributed to a thermal reaction in the residual oil in JME during torrefaction. The tar/oil content of JME T was about 1.0–1.8 wt.% in the determined optimal conditions. Thus, the enhanced recovery of the residual oil is advantageous not only because it allows obtaining more oil from Jatropha seed residue with a positive net energy gain but also because it prevents the formation of tar in torreﬁed biomass products. Caloriﬁc, proximate, and ﬁber analyses, mass and energy yields, and energy densiﬁcation were determined to elucidate the characteristic of the torreﬁed JMS products. The results were further compared with those obtained using mechanical extraction only (symbolized as JME) and other oil extraction methods. The ﬁndings are of interests for practical consideration of the torrefaction process in relation to different types or combinations of oil extraction processes.


Introduction
Building a global energy sector with biomass crops is highly recognized as a sustainable and economically viable pathway for reaching net-zero emission [1]. Among various types of biomass energy feedstock, non-food or second-generation energy crops have the potential to provide benefits such as consuming waste residues, making use of abandoned land, and promoting rural development [2][3][4]. For instance, Jatropha curcas L. is a non-edible The seeds samples were from Jatropha cultivated in southern Taiwan. They were directly mechanically screw-pressed at a moderately high temperature of 170 • C to enhance the oil extraction. The obtained JME was further subjected to Soxhlet extraction (FOSS model 2043, FOSS Worldwide Co., Birchwood, Warrington, Cheshire, UK) using n-hexane solvent (95%, Avantor Performance Material, Inc., Phillipsburg, PA, USA) to recover the residual oil. Then, the de-oiled residue JMS was obtained after a 2-stage sequential mechanical and solvent extraction. The JMS was subjected to torrefaction, obtaining solid recovered fuel of torrefied biomass of JMS (called JMS T ) with upgraded heating value.

Torrefaction Experiments
The pre-drying of the raw samples of JMS was conducted at 25 • C for 24 h in an oven (FW40, Channel Business Co., Taipei, Taiwan) to gently remove the surface moisture. For woody plants or lignocellulosic biomass, a higher temperature would reduce the required time for pre-drying, e.g., 40 • C for 7 h or 50 • C for 6 h, etc. [18]. Torrefaction was performed in a muffle furnace (DF-40, Deng Yng Co., Taipei, Taiwan). To ensure the absence of oxygen in the furnace, nitrogen (N 2 ) gas was introduced for 15 min. In each run, about 40 ± 1 g (on a dry basis) of sample was loaded on an aluminum disk circle.
For identifying the proper T r , a thermogravimetric analyzer (TGA-51, Shimadzu, Kyoto, Japan) was used for the pyrolysis analysis of JMS with N 2 purging of 50 mL/min from 25 to 850 • C at the heating rate of 30 • C/min. Three different T r of 260, 280, and 300 • C with acceptable mass loss were thus chosen for further torrefaction runs. When reaching the set temperature, torrefaction was continued at a constant temperature for six different times (t r ) of 10,20,30,40,50, and 60 min. This study further compared the findings for JMS T with those of previous work on JME T obtained with mechanical extraction only and subjected to similar torrefaction procedures [8].

Raw and Torrefied Product Characteristics
The properties of raw and torrefied samples of JMS and JMS T were determined as follows. Proximate analyses on a wet basis of moisture (M W ), ash (M A ), volatile matters (M VM ), and fixed carbons (M FC ) were performed according to the NIEA R205.01C method of the National Institute of Environmental Analysis (NIEA), Taiwan, where the combustibles (M C ) are the sum of M VM and M FC . HHV or calorific value was determined by the ASTM D2015 method of the American Society for Testing and Materials (ASTM), using a plain jacket oxygen bomb calorimeter (Model 1341, Parr Instrument Co., Moline, IL, USA). Chemical elemental analyses (C, H, O, N and S) were performed using an elemental analyzer (Elementar Vario EL-III, Hanau, Germany) following the NIEA R409.21C method. The oil content (C T ) was determined employing a Soxhlet extractor and using n-hexane, as described previously. In this study, all experiments were performed in duplicate to validate the composition, mass residual fraction, and HHV of the samples.

Pyrolysis Characteristics of JME and JMS
The proximate analysis of JME and JMS was performed, and the results were compared in this study. The mechanical oil extraction efficiency of JME via screw press was about 89.6%, with 5.5 wt.% oil of JME being retained, as reported by previous work [8]. Soxhlet extraction of JMS using n-hexane recovered most of the residual oil from JME. Hence, the proximate analysis of JME and JMS revealed that the M C (sum of M FC and M VM ) was slightly reduced from 87.7% for JME to 85.0 wt.% for JMS on a wet basis; the FC and VM consisted of hemicellulose, cellulose, and lignin along with residual oil and other organics. In addition, M W and M A were 8.4% and 6.6% for JMS and 6.1% and 6.2% for JME. The moisture of JMS was lower than that of JME, mainly due to the second stage of the process, i.e., solvent extraction.
The results of fiber analysis indicated that the contents of hemicellulose, cellulose, and lignin were 24.5, 20.7, and 14.5 wt.% for JMS and 21.7, 18.3, and 12.8 wt.% for JME on a dry basis. The relative content ratios for hemicellulose, cellulose, and lignin (obtained by dividing each content by the total content of these three fibers) for JME and JMS were the same, with the relative ratio of hemicellulose, cellulose, and lignin of 41.0%, 34.7%, and 24.3% (sum of 100%). Since these fibers (in particular lignin) are rich in carbon and hydrogen, the HHV was slightly decreased from 20.78 MJ/kg for JME to 18.8 MJ/kg for JMS, mainly owing to the de-oiling process of JME.
The TGA curves of JME and JMS from 105 to 850 • C at the heating rate of 30 • C/min under nitrogen purging (50 mL/min) are presented in Figure 1. The JMS began to crack at a lower pyrolytic temperature (T P ) with respect to JME. At the residue mass fraction during pyrolysis (M P ) of 95 wt.% or at 5 wt.% mass loss, the T P of JMS and JME were around 290-300 • C. T r of 260, 280, and 300 • C could be selected as the onset or triggering temperature according to the TGA curves, with acceptable mass loss of JMS of 5-10 wt.% in the high heating rate condition.

Torrefaction Performance of JMST
The residual mass fraction (Mr) of JMST during torrefaction at torrefaction temperatures (Tr) of 260, 280, and 300 °C and a torrefaction time (tr) from 10 to 60 min when using a muffle furnace is shown in Figure 2. As can be seen, a longer holding time tr resulted in mass loss for JMS (or Mr decreases). At the same tr, a higher Tr promoted a vigorous mass loss, as expected. At 20 min, the Mr of JMST were 82.0%, 66.8%, and 56.7 wt.% at 260, 280, and 300 °C, respectively. The effect of the holding time on Mr was very slight for tr of 40 the main difference during pyrolysis between JME and JMS would be attributed to the pyrolysis of residual oil. As can be seen from the difference of the DTG curves of JME and JMS, the residual oil contained in JME was firstly pyrolyzed into a variety of liquid bio-oils and gas by-products at 350-550 • C. For example, Jourabchi et al. [20] showed that no bio-oil was obtained at a T P below 300 • C from Jatropha seed oil cake, suggesting that the heat only cracked hemicellulose and produced mainly CO and CO 2 . In the range of 350-500 • C, a yield of about 32-50% of bio-oil was obtained. Kanaujia et al. [21] reported that the organic fraction of bio-oil from the pyrolysis of Jatropha seed oil cake at 550 • C for 30 min consisted mainly of 48% hydrocarbons, 12% aldehydes and ketones, 10% phenols, 9% guaiacols, 8% esters, and 8% of other chemicals. As T P reached 600 • C, the M P of JME were higher than those of JMS. This result can be explained by the fact that the retained oil in JME was potentially carbonized, thus aromatic growth and polymerization occurred. During the gasification of biomass above 600 • C, a decreased yield of bio-oil byproducts was observed due to cracking or secondary tar reactions on the char surface [22]. Similar findings of declining bio-oil byproducts were reported for the pyrolysis of JME at higher temperatures [23]. Furthermore, these results imply that further de-oiling of JME could improve the production of solid fuel by pyrolysis with less tars or by-products.

Torrefaction Performance of JMS T
The residual mass fraction (M r ) of JMS T during torrefaction at torrefaction temperatures (T r ) of 260, 280, and 300 • C and a torrefaction time (t r ) from 10 to 60 min when using a muffle furnace is shown in Figure 2. As can be seen, a longer holding time t r resulted in mass loss for JMS (or M r decreases). At the same t r , a higher T r promoted a vigorous mass loss, as expected. At 20 min, the M r of JMS T were 82.0%, 66.8%, and 56.7 wt.% at 260, 280, and 300 • C, respectively. The effect of the holding time on M r was very slight for t r of 40 min or longer, while 60.5%, 52.1%, and 48.9 wt.% of M r for 260, 280, and 300 • C were observed, respectively. Unlike the TGA curve in Figure 1, the torrefaction experiment was conducted under a constant temperature to mildly pyrolyze the biomass. The findings indicated that the torrefaction of JMS in the studied conditions reduced the volatile matter content of hemicellulose, cellulose, and moisture and consequently increased the fixed carbon content of lignin.

Torrefaction Performance of JMST
The residual mass fraction (Mr) of JMST during torrefaction at torrefaction temperatures (Tr) of 260, 280, and 300 °C and a torrefaction time (tr) from 10 to 60 min when using a muffle furnace is shown in Figure 2. As can be seen, a longer holding time tr resulted in mass loss for JMS (or Mr decreases). At the same tr, a higher Tr promoted a vigorous mass loss, as expected. At 20 min, the Mr of JMST were 82.0%, 66.8%, and 56.7 wt.% at 260, 280, and 300 °C, respectively. The effect of the holding time on Mr was very slight for tr of 40 min or longer, while 60.5%, 52.1%, and 48.9 wt.% of Mr for 260, 280, and 300 °C were observed, respectively. Unlike the TGA curve in Figure 1, the torrefaction experiment was conducted under a constant temperature to mildly pyrolyze the biomass. The findings indicated that the torrefaction of JMS in the studied conditions reduced the volatile matter content of hemicellulose, cellulose, and moisture and consequently increased the fixed carbon content of lignin.   Figure 3a, higher T r and t r values facilitated the increase of HHV as calorific value per mass. The obtained HHV values were higher than those of JMS (18.8 MJ/kg). The rate of HHV gradually decreased with t r . After 30 min, the HHV was higher than 24.0 MJ/kg, and thus than the HHV of hard black coal of 23.9 MJ/kg, as recommended by International Energy Agency [24]. As compared to JME T in the same torrefaction conditions [8], the JMS T exhibited a HHV generally lower than that of JME T because of the removal of residual oil. For instance, higher HHV values of volatile matter content of hemicellulose and cellulose within the Tr of 260-300 °C. T trends of MFC were similar to those of HHV in the same torrefaction conditions. As show in Figure 3d, the increase of MA for JMST above the initial amount (ash in JMS) was simp because of the organics loss (MVM) during torrefaction. A 3% fluctuation of MVM and M was observed at tr above 40 min. This study prepared a J-cake sample by grinding who seeds, which consisted of 59.0 ± 0.52 wt.% of kernel and 40.66 ± 0.12 wt.% of shell. reported [8], the MA of kernel (9.0 wt.%) is higher than that of shell (4.3 wt.%). Therefo this fluctuation could be attributed to minor changes in kernel and shell content in J-ca In gasification, a low VM content (including oil) of a solid biofuel is favorable to avoid tar generation and a high FC content to enhance carbon enrichment and energy densification. As shown in Figure 3c, the increment of HHV for the torrefied products can be attributed to the removal of the M VM of raw biomass while retaining the M FC of lignin. Further, an increase of M FC with the corresponding reduction of M VM was observed, as shown in Figure 3b,c. The values of M FC substantially increased from 21.76% to 47.55%, while M VM decreased from 71.08% to 40.63% under a more severe torrefaction conditions (t r = 30 min and T r = 300 • C). At t r over 30 min, the rates of change of M FC and M VM were moderate, suggesting that a residence time of above 30 min was sufficient for the elimination of the volatile matter content of hemicellulose and cellulose within the T r of 260-300 • C. The trends of M FC were similar to those of HHV in the same torrefaction conditions. As shown in Figure 3d, the increase of M A for JMS T above the initial amount (ash in JMS) was simply because of the organics loss (M VM ) during torrefaction. A 3% fluctuation of M VM and M A was observed at t r above 40 min. This study prepared a J-cake sample by grinding whole seeds, which consisted of 59.0 ± 0.52 wt.% of kernel and 40.66 ± 0.12 wt.% of shell. As reported [8], the M A of kernel (9.0 wt.%) is higher than that of shell (4.3 wt.%). Therefore, this fluctuation could be attributed to minor changes in kernel and shell content in J-cake.

Energy Densification of JMS T
The mass yield (Y M ), energy yield (Y E ), and energy densification (E D ) were determined to elucidate the performance of torrefaction for producing solid recovered fuel (SRF) of torrefied biomass from JMS, as follows [8,25,26]: Y E = m tor HHV tor /(m raw HHV raw ) where m raw , m tor = mass of dried raw JMS and JMS T , HHV raw , HHV tor = HHV of dried raw JMS and JMS T . E D is defined as the HHV ratio of torrefied and dried raw biomass, which can be also named energy ratio or enhancement factor [12][13][14]. The calculated E D of JMS T is presented in Figure 4. A higher T r as well as a longer t r generally increased the E D of JMS T . The Ed increased significantly in the first 30 min over 260-300 • C. For t r = 30 min, the E D values were 1.25, 1.29, and 1.39 at T r of 260, 280, and 300 • C, respectively, while E D at 10 min was 1.04, 1.06, and 1.18 respectively. Thermal decomposition of the high-heating-value components would occur at high T r of 280-300 • C and long t r of 30 min. Consequently, the values of HHV and the E D of JMS T declined in this temperature range. YE = mtor HHVtor/(mraw HHVraw) (2) ED = YE/YM = HHVtor/HHVraw where mraw, mtor = mass of dried raw JMS and JMST, HHVraw, HHVtor = HHV of dried raw JMS and JMST. ED is defined as the HHV ratio of torrefied and dried raw biomass, which can be also named energy ratio or enhancement factor [12][13][14]. The calculated ED of JMST is presented in Figure 4. A higher Tr as well as a longer tr generally increased the ED of JMST. The Ed increased significantly in the first 30 min over 260-300 °C. For tr = 30 min, the ED values were 1.25, 1.29, and 1.39 at Tr of 260, 280, and 300 °C, respectively, while ED at 10 min was 1.04, 1.06, and 1.18 respectively. Thermal decomposition of the high-heating-value components would occur at high Tr of 280-300 °C and long tr of 30 min. Consequently, the values of HHV and the ED of JMST declined in this temperature range.
According to Lloyd and Wyman [27], the severity factor (SF) can be defined as an integrated index for the effect of temperature and holding time during the torrefaction of biomass. A high value of SF represents a vigorous reaction which requires more energy input. Therefore, this study compared the JMST by using SF, defined as: where tr is in min, Tr is in °C, and Tref is 100 °C. The results of logSF numbers are shown in Figure 4.
To obtain the commonly accepted ED of 1.3, suitable conditions for JMST were as follows: 1) ED of 1.37 and HHV of 25.7 MJ/kg at logSF = 6.90, Tr = 280 °C ,and tr = 40 min and 2) ED of 1.35 and HHV of 25.4 MJ/kg at logSF = 7.19, Tr = 300 °C, and tr = 20 min. Although satisfactory torrefaction in terms of ED was achieved in both cases, a lower logSF to reach higher ED would be advantageous. Therefore, Case 1 is better than Case 2 when considering the ED and logSF numbers. According to Lloyd and Wyman [27], the severity factor (SF) can be defined as an integrated index for the effect of temperature and holding time during the torrefaction of biomass. A high value of SF represents a vigorous reaction which requires more energy input. Therefore, this study compared the JMS T by using SF, defined as: where t r is in min, T r is in • C, and T ref is 100 • C. The results of logSF numbers are shown in Figure 4.
To obtain the commonly accepted E D of 1.3, suitable conditions for JMS T were as follows: (1) E D of 1.37 and HHV of 25.7 MJ/kg at logSF = 6.90, T r = 280 • C, and t r = 40 min and (2) E D of 1.35 and HHV of 25.4 MJ/kg at logSF = 7.19, T r = 300 • C, and t r = 20 min. Although satisfactory torrefaction in terms of E D was achieved in both cases, a lower logSF to reach higher E D would be advantageous. Therefore, Case 1 is better than Case 2 when considering the E D and logSF numbers.

Elemental Analysis of JMS T
The results of the elemental analysis of JMS and JMS T , reported in Table 1, show a significant decrease in oxygen content and an increase in carbon content when increasing T r . Meanwhile, hydrogen content decreased slightly. After t r of 60 min, the oxygen content of JMS T significantly decreased from 38.63 wt.% for JMS to 21.18-23.77 wt.%, while the carbon content of JMS T significantly improved from 46.79 wt.% for JMS to 59.51-61.50 wt.%. Based on these data, the relative ratios of O/C and H/C ratio for JMS T and JMS were calculated and are shown in Figure 5a,b, respectively. As revealed, the relative ratio of O/C was about 0.42-0.48, and the relative reduction of the O/C ratio was about 52-58% at t r = 60 min, while a relative reduction of 38-52% of the H/C ratio was obtained by torrefaction. The declining trend of the O/C ratio was more significant than that of the H/C ratio. The results suggest that oxygen-containing molecules (e.g., CO 2 and H 2 O) would be eliminated more easily than hydrogen-containing molecules (e.g., CH 4 and H 2 ) during torrefaction of JMS. Further, the reduction in oxygen and hydrogen content in all the biomass types can be specifically attributed to the removal of hydroxyl groups (OH) via hemicellulose decomposition [28]. Therefore, lowering O/C and H/C tends to produce more hydrophobic biomass [29,30]. Table 1. Chemical elemental analyses of JMS at t r = 0 min and of torrefied JMS (JMS T ). The effect of T r and t r on torrefaction performance can be also examined by the van Krevelen diagram, as shown in Figure 5c. In the conditions of 280 • C and 40 min (E D = 1.37, logSF = 6.90) for Case 1 and of 300 • C and 20 min (E D = 1.35, logSF = 7.19) for Case 2, the atomic ratios of O/C of JMS T were 0.23 and 0.29, while those of H/C were 1.06 and 0.97, respectively. These values of JMS T at E D above 1.28 were all within the range of those for lignite, were close to those for sub-bituminous coal, and superior to those for other torrefied wood, giving that the values of O/C and H/C were 0.52-0.68 and 1.01-1.41 for torrefied wood, 0.22-0.38 and 0.78-1.26 for lignite, and 0.01-0.25 and 0.34-0.98 for coal [31]. Furthermore, the van Krevelen diagram also revealed that JMS T at T r = 280 • C and t r = 40 (Case 1) and at longer times (50 and 60 min) were all close to those for coal, suggesting better carbon enrichment and the removal of oxygen and hydrogen in JMS T .  Table 2 lists the torrefaction properties of torrefied Jatropha biomass by various oil extraction processes. As can be seen, the E D and van Krevelen diagram (H/C vs O/C) in the optimal torrefaction conditions for JME T [8] were alike to those of JMS T . For instance, JME T at T r = 280 • C and t r = 50 min had E D of 1.28, H/C of 1.01, and O/C of 0.24, while those values for JMS T at T r = 280 • C and t r = 40 min (Case 1) were 1.37, 1.06, and 0.23, respectively. Nevertheless, the HHV of JME T was 26.7 MJ/kg, higher than that of JMS T (25.7 MJ/kg). The difference of HHV between JMS T and JME T is highly attributed to the residual oil depending on the oil extraction method, with the calorific value of Jatropha oil being about 39.63 MJ/kg [5].   Table 2 lists the torrefaction properties of torrefied Jatropha biomass by various oil extraction processes. As can be seen, the ED and van Krevelen diagram (H/C vs O/C) in the optimal torrefaction conditions for JMET [8] were alike to those of JMST. For instance, JMET at Tr = 280 °C and tr = 50 min had ED of 1.28, H/C of 1.01, and O/C of 0.24, while those values for JMST at Tr = 280 °C and tr = 40 min (Case 1) were 1.37, 1.06, and 0.23, respectively. Nevertheless, the HHV of JMET was 26.7 MJ/kg, higher than that of JMST (25.7 MJ/kg). The difference of HHV between JMST and JMET is highly attributed to the residual oil depending on the oil extraction method, with the calorific value of Jatropha oil being about 39.63 MJ/kg [5].

Comparison of JMST and JMET
As displayed in Table 2, the retained oils were 5.5 and 5.7 wt.% after mechanical screw press only and ultrasonic solvent extraction for 15 min, respectively. On the other hand, the retained oil after solvent extraction pretreatment is usually not determined because the residual oil is exactly measured by the extraction efficiency achieved by extrac- As displayed in Table 2, the retained oils were 5.5 and 5.7 wt.% after mechanical screw press only and ultrasonic solvent extraction for 15 min, respectively. On the other hand, the retained oil after solvent extraction pretreatment is usually not determined because the residual oil is exactly measured by the extraction efficiency achieved by extraction with n-hexane for 4-8 h, corresponding to 99%. The estimated input energy for solvent extraction was 3.67 MJ for 1 kg of soybean oil production [32]. Given that 1 ton of JME will supply 55 kg oil, the second stage of solvent extraction requires an input energy of 201.8 MJ, which is largely lower than the heating value of 2180 MJ from the recovery of Jatropha oils. Therefore, a two-stage sequential process consisting of mechanical screw pressing and solvent extraction allows the recovery of more Jatropha oil with and a gain of energy, while the torrefaction properties of JME T and JMS T are quite compatible. a JSK: Jatropha seed kernel; JSS: Jatropha seed shell; JFH: Jatropha fruit husk. b M: Mechanical screw press; M + S: a two-stage sequential process of mechanical screw pressing and solvent extraction; S: solvent extraction by n-hexane; US: Ultrasonic solvent extraction for 15 min and mesh of 0.5-1.0 mm by n-hexane.
In fact, the combination of mechanical screw pressing and solvent extraction has been widely used for the production of edible oil, namely, extra-virgin and extraction oils. The input energy for the second-stage solvent extraction process should be economically compensated by the obtained oil product. Nevertheless, the mass and energy balance for mechanical screw pressing only and two-stage mechanical screw pressing and solvent extraction is worthy to be determined for clarifying the economic benefit of the torrefaction of non-edible de-oiled Jatropha biomass.
It should be pointed out that the residual oil of 5.5 wt.% in JME retarded and delayed the torrefaction of biomass of JME compared to that of JMS containing no residual oil, as revealed in Figure 1. Theoretically, the torrefaction of residual oil containing JME above 300 • C will produce solid fuel with tar due to a thermal reaction in the residual oil. The tar in biomass fuel is a challenge for air pollution control, particularly for smallscale decentralized gasification-to-power systems [33]. The fine particle, soot, tar ball and black carbon emitted from the incomplete combustion of biomass fuel also contribute to climate change, ozone formation and other air quality issues [34]. Hence, this study further examined the tar and residual oil content of JME T , as displayed in Figure 6. The results indicated that JME T containing tar was observed in all cases. The tar/oil content was about 2.1 wt.% for 280 • C at 40 min and 2 wt.% for 300 • C at 20 min. A higher T r and longer t r facilitated the thermal decomposition of residual oil and tar. In addition, the tar generated from JMS would be theoretically lower than that from JME. Nevertheless, the tar/oil content of JMS T should be interesting to be examined in a future study.

Conclusions
In this study, the torrefaction performance at different Tr and tr was investigated for de-oiled pressed cakes of Jatropha seeds following a two-stage sequential process consisting of mechanical screw pressing and solvent extraction using n-hexane (denoted as JMS).

Conclusions
In this study, the torrefaction performance at different T r and t r was investigated for de-oiled pressed cakes of Jatropha seeds following a two-stage sequential process consisting of mechanical screw pressing and solvent extraction using n-hexane (denoted as JMS). The optimal operation conditions were examined at fixed T r of 260, 280, and 300 • C and t r of 10-60 min. The results showed that the increase of T r and t r upgraded the HHV and E D of the torrefied products, with acceptable mass loss. The enhancement of HHV was attributed to the increase of fixed carbons. The mass loss was mainly attributed to the decomposition of hemicelluloses and celluloses and the elimination of volatile matters. The satisfactory E D of about 1.3 was achieved: 1) E D of 1.37 at logSF = 6.90, T r = 280 • C, and t r = 40 min and 2) E D of 1.35 at logSF = 7.19, T r = 300 • C, and t r = 20 min. Although both Cases 1 and 2 reached satisfactory torrefaction in terms of E D , a lower logSF to gain higher E D should be encouraged. Furthermore, better carbon enrichment and the elimination of hydrogen and oxygen in the torrefaction of JMS were obtained with T r = 280 • C and t r = 40 (Case 1) and at longer times (50 and 60 min) because the atomic ratios of O/C and H/C for these JMS T were close to those of coal.
By comparing JMS with Jatropha seed residue biomass undergoing screw pressing only (defined as JME), the residual oil had a mixed effect for the torrefied biomass fuel. The tar content of JME T was 2-5 wt.%, which is positive for HHV but negative as regards the low gasification. The other solid fuel characteristics of JME T and JMS T remained comparable. Therefore, the current findings indicate that the enhanced recovery of residual oil by a two-stage sequential process of mechanical screw pressing and solvent extraction can prevent the formation of tar in the torrefied biomass products without damaging their E D and solid fuel characteristics. More importantly, the two-stage process would probably allow not only the recovery of more oil from Jatropha seed residue with a net energy gained but also the reduction of hazardous air pollutant emission.