Pressure, temperature and processing time in enhancing Camelina sativa oil extraction by Instant Controlled Pressure-Drop (DIC) texturing pre-treatment ; Presión, temperatura y tiempo de procesamiento para mejorar la extracción de aceite de Camelina sativa mediante pretratamiento texturizado de descompresión instantánea controlada (DIC)

K. Bouallegue, T. Allaf, R. Ben Younes, C. Téllez-Pérez, C. Besombes and K. Allaf Gafsa University; Research Unit of Physics, Computers Science and Mathematics, Faculty of Science; University of Gafsa (Tunisia) ABCAR-DIC Process; 17000 La Rochelle, France Tecnologico de Monterrey, Escuela de Ingeniería y Ciencias, Campus Querétaro. Epigmenio González 500, Fracc. San Pablo. Querétaro, Querétaro, 76130, México University of La Rochelle, Intensification of Transfer Phenomena on Industrial Eco-Processes, Laboratory of Engineering Science for Environment LaSIE UMR-CNRS 7356, 17042 La Rochelle, France Corresponding author: kallaf@univ-lr.fr


INTRODUCTION
Native to different Romanian areas, northern Europe, the Mediterranean region, and Central Asia, Camelina (Camelina sativa L.) is an ancient annual oilseed crop that belongs to the Cruciferae family (Brassicaceae, Mustard) (Hurtaud and Peyraud, 2007;Carciumaru, 2007). Also named 'False flax' because its fruits resemble flax bolls (Linum usitatisimum L.), and 'Gold of Pleasure' a popular name invented by the Romans in the early centuries (Berti et al., 2011), nowadays Camelina plays an important role in oilseed biotechnology. As an example, in 2008, for the first time, Chile introduced camelina seeds as a potential feedstock for biodiesel. It has also drawn the attention of the pharmaceutical industry thanks to its high omega-3-content. (Berti et al., 2011).
Furthermore, thanks to its high oil content, with about 0.40 g oil/g db (dry matter basis), which is 0.67 g oil/g ddb (dry-dry matter basis: this means a mass basis of raw material rid of water and oil), and its healthy oil properties, with up to 90% of unsaturated fatty acid, camelina seeds have seen an increased interest from different industries (Budin et al., 1995;Li et al., 2014).
Camelina seed vegetable oil is extracted by cold mechanical pressing, by solvent extraction, or by a combination of both methods. In fact, combinations of both methods are most often used for economic reasons since the pressing process leaves a significant amount of residual oil in the oilcakes and meal, which can be extracted by solvent extraction (Gunstone, 2006). Indeed, camelina oilcakes/meal remaining from seed pressing typically contains 10 to 15% residual oil, 40% crude protein, about 13% fibers, 5% minerals, and minor amounts of other substances such as vitamins (Li et al., 2014;Zubr, 1997). Moreover, thanks to its high crude protein content, oilcakes can be used as nutritive supplements in animal feed formulations (Moloney et al., 2001).
Mechanical pressing and solvent extraction are the most commonly applied industrial techniques for camelina oil recovery, although they present some drawbacks such as low yields, long extraction periods, toxicological risks, excessive solvent residues and high costs (Belayneh et al., 2015;Kartika et al., 2010). To overcome these constraints, several oilseed pre-treatments have been evaluated at laboratory scale, such as pre-heating oilseeds using a hot air oven and microwaves (Yusuf, 2018). However, although these pre-treatments increase oil yield, their use at the industrial level remains difficult because of the drawbacks of their scaling-up and high costs, and the difficult preservation of the oil's quality. In this respect, the Instant Controlled Pressure Drop (DIC) and high-temperature short-time texturing pretreatment makes the seed oil more readily available. DIC technology is based on thermo-hydromechanical processing induced by subjecting a product to a rapid transition from high-steam pressure to a vacuum, triggering an instant autovaporization of a quantity of material water, a fast cooling and a well-controlled expansion of the product. The change in the structural characteristics of the product is generally revealed through the expansion rates of the product, which depend on the operating conditions. Various studies have shown that the most influential DIC operating parameters are saturated steam pressure, which is strictly correlated with temperature, and processing time (Mounir and Allaf, 2008;Allaf and Allaf, 2013).
On the other hand, previous studies on numerous crops (Allaf et al., 2014;Bouallegue et al., 2015), have shown that this innovative technology can intensify oil solvent extraction by giving higher yields in a shorter time, and can be easily scaled-up at industrial level by using the same processing conditions optimized at laboratory scale. Specifically, in the case of camelina seeds, it has been shown that DIC could be used to intensify the in-situ transesterification of seeds, leading to a significant increase in biodiesel production yield (Bamerni et al., 2017). Moreover, the composition of camelina oil from DIC textured seeds was similar to raw material oil, which means a good preservation of the oil's quality (Bamerni, 2018).
Therefore, the goal of this study was to enhance the oil extraction of Camelina sativa seeds and meals by coupling Instant Controlled Pressure Drop (DIC) texturing pre-treatment to mechanical pressing and solvent extraction. To optimize oil yield extraction while preserving the quality of both oil and meals, the impacts of saturated steam pressure and processing time of DIC texturing pre-treatment were studied.

Solvent and pressing extraction of Camelina sativa oil
Camelina sativa seeds were sorted, cleaned and divided into two groups: 1) Raw material seeds, and 2) DIC-textured seeds. For oil extraction, four individual operations were studied: 1) Cold Pressing Extraction; 2) n-hexane Accelerated Solvent Extraction ASE; 3) Dynamic-Maceration extraction DM extraction, and 4) a coupled cold pressing of seeds and DM solvent extraction of meals. The processing protocol of DIC pre-treatment of Camelina sativa seeds and the mechanical and solvent extractions of oil are illustrated in Figure 1.
Solvent extraction. Two methods were used for solvent extraction; Accelerated Solvent Extraction (ASE) and Dynamic Maceration Method (DM). Before any solvent extraction, camelina seeds were ground at the rate of 10000 rpm for 15 s (Grindomix, GM200 -F. Kurt Retsch GmbH & Co. KG, Haan, Germany), and the obtained powders were sieved (Vibratory Sieve Shaker, Fritsch, Germany) into 0.4 mm particle size fractions. Sieving was carried out for 10 min under an amplitude of 1.5 mm and the camelina powders were kept at 4 °C in the dark until analysis.
To define ASE suitable conditions for camelina, preliminary tests based on the study of Kraujalis et al., (2013) were performed on a Dionex ASE 350 system (Thermo Fisher scientifique, Sunnyvale, CA, USA). 7 g of camelina seed powder were mixed with 1 or 2 g of diatomaceous earth and placed in a 34 ml Dionex stainless-steel cell (2.9 cm diameter). Hexane was used as a solvent, which represented 60% of the cell volume. ASE was initiated by 5 min pre-heating time to reach 100 ºC at 10 MPa, followed by 4 cycles of extraction of 10 min each. Then, cell content was purged during 150 s with nitrogen to remove impurities and the extracts were collected in a vial. Hexane was removed by a rotary vacuum evaporator under reduced pressure (10 mbar; 40 °C), and the extracted oils were dried under a stream of nitrogen. Extractions were conducted in triplicate. Finally, the oil extracts were weighed and stored at 4 °C for subsequent chemical analysis. Oil yield was expressed in g oil/kg ddb (Y ASE; seeds); ddb concerns material which excludes both water and oil contents.
DM was performed with extraction batches of 2 g of grain powder with 20 ml of n-Hexane. To assure a good contact between the phases, the entire extraction operation was conducted under magnetic stirring at 400 rpm. The extraction was carried out at ambient temperature (25 °C), in triplicate. To evaluate the extraction performance, oil yield was determined at different interval times (0, 15, 30, 45, 60, 90, 120, 150, 180, 240, 300, 360, 420, 480 and 1440 min). To measure the oil contents, the extracts were filtered with 0.2 μm PTFE filters (Sartorius Stedim Biotech GmbH/Germany), and the mixtures (hexane/oil solutions) were separated under vacuum by nitrogen flow (Liebisch Mini evaporator, Germany). The extracted oil was dried until constant weight. Oil extraction yield (Y DM; seeds) was calculated as shown in equation 1. Oil yield was expressed in g oil/kg ddb. Pressing extraction. Pressing extraction was carried in a single-screw press machine ("OMEGA 20" type "Taby Orebro", Germany). First, without any sample, the screw-press barrel was heated by an electrical resistance to 80 °C. Subsequently, 300 g of camelina seeds were placed in the screw press, and oil and meals were recovered after 30 to 60 s of pressing. The diameter of pressing nozzle was 8 mm, and fine particles in the expressed oil were separated by filtration. Oil content was gravimetrically determined (Y pressing 1 ), and it was expressed as g oil/ kg of dried seeds ( To measure the total residual oil contents of the meals, the ASE method was applied by using 10 g of meal powder (Y ASE; meals), and the recovered oils were stored at 4 °C for further analysis. To establish the extraction kinetics of camelina pressed meals, the DM method was applied as described previously (Y DM; meals).
DIC texturing pre-treatment. The DIC (French for Détente Instantanée Contrôlée) is a thermo-mechanical process induced by subjecting the product to a saturated steam pressure (about 0.05 -1 MPa, according to the product) for a short period of time (some seconds), followed by an abrupt pressure drop towards a vacuum (up to 1.5 kPa). This abrupt pressure drop (ΔP/Δt > 0.2 MPa s -1 ) triggers an autovaporization of volatile molecules which induces a cooling and texturing effect (Allaf and Allaf, 2013). Figure 2A shows the schematic time-temperature-pressure profiles of a DIC processing cycle.
DIC texturing pre-treatment was carried on 300 g of camelina seeds, and selected DIC processing parameters were saturated steam pressure (from 0.2 to 0.7 MPa) and treatment time (from 20 to 120 s). The DIC equipment was composed of three main elements: i) the processing vessel (1), where samples were treated; ii) the vacuum system, which consists of a vacuum tank (2) with a volume 130 times greater than the processing reactor, and an adequate vacuum pump (3) and iii) the pneumatic valve (V2) with a large diameter (more than 200 mm). To ensure an abrupt/instant connection between the vacuum tank and the processing reactor, V2 was opened in a very short time (less than 0.2 s). Figure 2B presents the schematic diagram of the DIC equipment.
To determine the impact of the DIC treatment on the oil extraction yield and quality, treated samples were submitted to both extraction methods: a) Solvent Extraction (ASE and DM) and b) Pressing Extraction. Untreated raw material was used as a control (RM), and in both cases oil yields were expressed in g oil/kg ddb.
The experiments were run randomly to minimize the effects of unexpected variability in the observed responses due to extraneous factors.
To identify the significant differences of the effects between independent variables, the analysis of variance (ANOVA) was performed (p < 0.05). Moreover, a second-order polynomial function was employed to relate each response variable (Y) to the operating parameters (χ) as shown in equation 3: Where Y is the response, β 0 ,β i ,β ii and β ij , are the regression coefficients, χ i and χ j are the independent variables of DIC, ε is random error and i and j are the indices of factors.
The significance and the adequacy of the model were interpreted by estimating the lack of correspondence, R 2 and Fisher test value (F-value). The Pareto chart was used to determine the effects that were statistically significant with a p-value of 0.05. Response surface methodology (RSM) was used to analyze the experimental design results and optimize the treatment parameters through a multi-criteria procedure. Experimental results were statistically analyzed by Statgraphics Centurion Software (MANUGISTICS Inc., Rockville, USA).

Accelerated solvent extraction of non-treated (RM) and treated (DIC) camelina seeds
The averages oil yields of treated and non-treated C. sativa seeds are shown in Table 2. For RM, the average oil content obtained after ASE was 555.5 g oil/kg ddb, and in the case of DIC-treated seeds, the results varied from 569.4 to 615.9 g oil/kg ddb. The highest oil yield value for DIC samples corresponded to the experimental point DIC 5 (P: 0.63 MPa and t: 105 s) and the lowest to DIC 8 (P: 0.27 MPa and t: 35 s). Furthermore, to evaluate the impact of DIC parameters on ASE oil yield, the Analysis of Variance (ANOVA) and the response surface methodology (RSM) were applied. Table 3 and Figures 3a and 3b illustrate the impact of P (steam pressure) and t (treatment time) on the ASE oil yield of camelina seeds. The obtained results showed that under the selected domain values, the pressure (P), the time (t) and the interaction of both variables (P and t), presented positive effects on the increase in oil yield; the higher P and t, the higher the oil yield.
By expressing the "P" in MPa and "t" in s, the statistical analysis allowed to obtain a regression model for the ASE oil yield (Y ASE seeds ), with R 2 of 89.8%: Eqn. 4 To optimize (maximize) the ASE oil yield of seeds, the optimal conditions for DIC treatment for this response were P =0.7 MPa and t = 120 s (638.49 g of oil/100 g dry-dry basis).

Dynamic maceration extraction of non-treated (RM) and treated (DIC) camelina seeds
For RM, the average oil content obtained after 2, 8 and 24 h of DM were 284.8, 405.2 and 550.0 g oil/kg ddb, respectively. In the case of DIC-treated seeds, after 1, 2 and 8h of DM yields were 407.8, 462.8 and 520 g oil/kg ddb, respectively. All these results corresponded to experimental point DIC 5 (P: 0.63 MPa and t: 105 s). In the specific case of DM after 2h (DM 2h ), DIC-treated seeds varied from 439.5 to 462.8 g oil/kg ddb. At this extraction time (2 h), in any of the selected DIC-treatment conditions, the oil yield kinetics of camelina seeds were always better than that of untreated samples. Figure 4A presents the extraction kinetics of seed oil by DM of RM and DIC 5 (0.63 MPa and 105 s).
To evaluate the impact of DIC parameters on DM 2h oil yield, the ANOVA and the RSM were applied. Table 3 and Figures 3C and 3D illustrate the impact of P (steam pressure) and t (treatment time) on the DM 2h oil yield of camelina seeds. The results showed that under the selected domain values, the pressure (P) and the quadratic effect of the time (t 2 ) had a significant effect on the oil yield. The higher the pressure, the higher the DM 2h oil yield.
Equation 5 shows the regression model for the DM 2h oil yield (DM 2h; seeds ), with R 2 of 83%: To optimize (maximize) the DM 2h oil yield of seeds, the optimal conditions for DIC treatment for this response were P =0.7 MPa and t = 91 s (462.68 g of oil/100 g dry-dry basis).

Pressing extraction of non-treated (RM) and treated (DIC) camelina seeds
The average camelina seed oil pressing yield (Y pressing ) from RM was 444.7 g oil/kg ddb. And in the case of DIC-treated seeds Y pressing values varied from 449.8 to 490.9 g oil/kg ddb. The highest oil yield value for the DIC samples corresponded to experimental point DIC 5 (P: 0.63 MPa and t: 105 s) and the lowest to DIC 8 (P: 0.27 MPa and t: 35 s).
The ANOVA and the RSM allowed to evaluate the effect of DIC parameters on oil Y pressing and the results showed that under the selected domain  values, the pressure (P), time (t) and quadratic effect of this factor (t 2 ) had a significant effect on the pressing oil yield. The higher the pressure and the time, the higher the oil Y pressing . Table 3 and Figure 5 illustrate the impact of P (steam pressure) and t (treatment time) on oil Y pressing of camelina seeds. Equation 6 shows the regression model for the oil Y pressing , with R 2 of 97%: To optimize (maximize) the oil Y pressing of seeds, the optimal conditions for DIC treatment for this response were P =0.7 MPa and t = 120 s (493.59 g of oil/100 g dry-dry basis).

Accelerated solvent extraction of non-treated (RM) and treated (DIC) camelina meals
To determine the performance of pressing extraction, the residual oil yields of the meals were determined through ASE (Y ASE; meals ). The average oil meal content for RM was 110 g oil/kg ddb, while for DIC-treated samples, yields varied from 130.7 to 133.4 g oil/kg ddb. The highest oil yield value for DIC samples corresponded to experimental point DIC 5 (P: 0.63 MPa and t: 105 s) and the lowest to DIC 11 (P: 0.20 MPa and t: 70 s).
To evaluate the effect of DIC parameters on Y ASE; meals, the ANOVA and the RSM were applied. The results showed that under the selected domain values, the pressure (P), the time (t) and the quadratic effect of this factor (t 2 ) had a significant positive effect on the pressing oil yield. The higher the pressure and the time, the higher the oil Y ASE; meals . To optimize (maximize) the oil Y pressing of seeds, the optimal conditions of DIC treatment for this response were P =0.7 MPa and t = 120 s (134.15 g of oil/100 g dry-dry basis).

Dynamic maceration extraction of non-treated (RM) and treated (DIC) camelina meals
The average oil contents obtained after 2, 8 and 24 h of dynamic maceration of untreated camelina meals were 51.6, 63.5 and 110 g oil/kg ddb, respectively. In the case of DIC-treated meals, the best performance was shown by DIC 5 (0.63 MPa and 105 s). For this treatment, after 1, 2 and 8 h of DM, oil extraction oil yields were 91.3, 115.8 and 130 g oil/kg ddb, respectively. As it can be observed in Table 2, in only two hours all DIC-treated samples achieved the same oil yield as RM after 24 h. DM oil yields from DIC meals varied from 111.4 to 115.8 g oil/kg ddb. Figure 4B presents the extraction kinetics of meal oil by DM of RM and DIC 5 (0.63 MPa and 105 s).
To evaluate the effect of DIC parameters on the oil yield of meals after 2h of DM (DM 2h; meals ), the ANOVA and the RSM was applied. The results showed that under the selected domain values, the pressure (P) and the quadratic effect of the time (t 2 ) had a significant positive effect on the pressing oil yield. The higher the pressure and the time, the higher the DM 2h; meals oil yield s . To maximize the of Y DM2h of meals, the optimal conditions of DIC treatment were P =0.7 MPa and t = 120 s (121.02 g of oil/100 g dry-dry basis).

DISCUSSION
Thanks to its environmental adaptability, its satisfactory seed yields, and its multiple oil applications (i.e, biofuels, oleochemical compounds, animal feed, and food applications), Camelina sativa has attracted the attention of both research and industry (Zanetti et al., 2017). Besides, as showed in this study, camelina seeds contains more than 50% oil, 555.5 g oil/ kg ddb for ASE of RM, which increase its feasibility for use in several industries. Similar results were found by Moslavac et al., (2014) after 8 h of Soxhlet extraction, where they determined 42.40 ± 0.73% of oil from RM camelina seeds. Then, to extract the total oil content of camelina seeds, a variety of different extraction methods were applied, being the most used techniques: a) mechanical pressing, b) solvent extraction, c) or a combination of both methods Moslavac et al., 2014;Stroescu et al., 2015). However, the main drawback of mechanical pressing is the high percent of residual oil in the cakes. Moreover, even if solvent extraction helped to recover the remaining oil from the cakes, long extraction periods would be needed to perform at appropriate efficiency.
Then, to make the extraction of camelina oil more affordable, it is therefore important to redefine industrial methods that allow for recovering the largest amount of this oil in the shortest time. In this study, DIC texturing pre-treatment was applied to camelina seeds before mechanical pressing and solvent extraction, and the results showed that this technology systematically enhanced the oil extraction (Table 2).
Camelina oil solvent extraction was studied through ASE and DM, the first one aimed to determine the total oil content of seeds and meals, and the second one to study the different transfer mechanisms during oil solvent extraction. In the case of ASE of seed oil, DIC treatment allowed to extract 10.8% (615.9 g oil/kg ddb) more than RM samples, which means that through ASE it was not feasible to obtain the total oil amount from the seeds. In fact, DIC treatment allowed the seeds to attain higher porosity which triggered the rupture of the oil-containing glands. Moreover, in figure 4B, it can be observed that the DIC treatment of seeds produced a clear improvement in the reduction of extraction time by DM, reaching (DIC 5: 410.5 g oil/kg ddb) a higher yield in 1 h than RM after 8 h of extraction (RM 8h: 405.2 g oil/kg ddb). According to Allaf and Allaf, (2013), most of the kinetics are highly dependent on the porosity and tortuosity of the material. Then, the structure modification of seeds generated by DIC enhanced the solute-insolvent transfer within the holes of the solid matrix, which allowed to reduce the extraction period and to increase oil yield.
In the specific case of 2h of DM of seeds, DIC 5 presented 1.6 times (462.8 g oil/kg ddb) better oil yield than RM (284.8 g oil/kg ddb); which meant that thanks to the DIC treatment more than 80% of the total RM seed oil (555.5 g oil/kg ddb) could be obtained after 2 h of DM extraction. Furthermore, in both cases, ASE and DM, it was observed that the higher the pressure of the DIC treatment, the higher the improvement in the seed oil yield; the optimum experimental pressure value was 0.63 MPa.
Due to the fact that conventional solvent extraction methods represent 80% of the total processing time, 90% of the required energy, and more than 99% of the solvent used for the whole analysis procedure, pressing extraction has become an interesting solvent-free extraction technique to study and to ameliorate (Chemat et al., 2015). In this work, seed pressing was carried in a single-screw pressing machine, and results showed that compared to the total RM seed oil, this technique allowed for the recovery of a little more than 80% of camelina seed oil in a few minutes, 444.7 g oil/kg ddb in the case of RM seeds and 490.9 g oil/kg ddb for DIC 5. Just the same as in solvent extraction, DIC treatment conditions allowed for the disruption of cell membranes and the cell walls of seeds, which increased the efficiency of pressing extraction though the new porous matrix.
Though pressing extraction allowed for obtaining good oil yields, at best there was around 20% of seed oil that remained in meals. For this reason, in order to attain total oil recovery, industries apply a second solvent extraction step (Uitterhaegen and Evon, 2017). In this study, to determine the residual oil in pressed meals, ASE was applied. The results showed that by comparing the total oil amount from RM seeds (555.5 g oil/kg ddb) to recovered meal oil, it was possible to recover 19.8% of the oil (110 g oil/kg ddb) in the case of RM, and up to 24% from DIC-treated meals (133.4 g oil/kg ddb). On the other hand, by looking at the results of DM from camelina meals, we could point out that after 2 h of extraction, DIC 5 treatment allowed for obtaining 20.8% oil (115.8 g oil/kg ddb), which represented 2.24 times more oil than from RM meals (51.6 g oil/ kg ddb). It should be noted that even after 24 of DM for RM meals (110 g oil/kg ddb), DIC 5 showed the highest oil yield. Similar results were found by Moslavac et al., (2014), who, after 8h of Soxhlet extraction of camelina pressed meals, obtained 15.7% of recovered oil. Furthermore, in all cases, it was observed that for pressed seeds and meals, the higher the pressure and the time of DIC treatment, the higher the improvement in oil yield. In both cases, optimum experimental values were found at P= 0.63 MPa and t= 105 s.
When comparing the performance of the different studied extraction methods, it can be concluded that by coupling DIC treatment (P: 0.63 MPa and t: 105 s) to pressing extraction followed by the DM of meals for 2 h it was possible to reach 606.7 g oil/ kg ddb of oil yield, which meant that thanks to the DIC treatment it was possible to obtain 9.21% more than the initial total oil content recovered by ASE of camelina RM seeds (555.5 oil/kg ddb of oil yield). Moreover, by comparing these results to RM oil yield after pressing coupled to DM for 2 h (496.3 g oil/kg ddb), we could highlight an improvement of 22% in the final oil yield as a result of the DIC pre-treatment.
DIC is a convenient texturing pre-treatment to increase the oil yield from seeds, to reduce the oil extraction time, to ensure the final oil quality, to valorized pressing meals, and to increase the industrial processing capacities. Moreover, one of the main advantages of the DIC process is its ease of use at industrial level. In fact, the obtained optimal laboratory parameters could be scaled up without any problem. DIC reactors are currently operating at laboratory, pilot, and industrial scales. Nowadays, different DIC reactors are operating worldwide, e.g., in France, Spain, Italy, Mexico, Malaysia, and China.

CONCLUSIONS
Camelina sativa L. Crantz is a reemerging oilseed crop with a high oil content. This work focused on the enhancement of the oil extraction of Camelina sativa seeds and meals through the Instant Controlled Pressure Drop (DIC) texturing pretreatment. Results showed that compared to RM, DIC pre-treatment coupled to solvent and pressing extraction increased the oil yields of camelina seeds and meals. In the case of solvent extraction, DIC pre-treatment coupled to ASE allowed for obtaining 10.8% more oil than from untreated camelina seeds. Furthermore, through coupling DM to DIC it was possible to reduce the extraction time of oil seeds from 8 h to only 1 h. On the other hand, compared to raw seeds, DIC coupled to pressing allowed for a 10.3% increase in oil yields. Additionally, DIC improved extraction from oil meals by recovering 2.2 times more oil than untreated meals after two hours of DM. The optimal experimental DIC treatment conditions were 0.63 MPa and 105 s. DIC pretreatment allowed for increasing camelina oil yields, reducing extraction time and valorizing pressing meals, which makes it the ideal process for camelina oil extraction.