Effect of processing conditions on the improvement of properties and recovering yield of Moroccan oil shale

. In the present work, Moroccon Tarfaya oil shale was treated by acids and different solvents under supercritical conditions, successively. Experimental results showed clearly that residual mineral matter had a significant effect on the yield and composition of the resulting organic fraction. Indeed, the oil yields obtained from some samples, 43% and 56%, respectively, were much higher than that from the sub-layer, 18%. In addition, the yield of recuperation and quality of extracted oils were largely dependent on the nature of solvents (toluene, water, shale oil). Thus, phenol was shown to be a very efficient modifier for the supercritical extraction of organic matter from Tarfaya oil shale with toluene, affording a good yield of recovery and a suitable maturation of organic matter. The pitches prepared by mixing phenol and toluene contained more aromatics and had a high char yield (46%) at 950 °C compared to those obtained by extraction with supercritical toluene alone.


Introduction
Oil shale has been considered for a long time a great economic hope for those countries that possess large reserves.Thus, programs to operate oil shale as an energy source were launched by several countries such as Scotland, Australia, France and China [1][2][3][4].These programs demonstrate that oil shale is far from substituting the products of oil, because of high costs and non-competitiveness against conventional petroleum.
Studies have demonstrated that the extraction yield and quality of oils extracted depend on several parameters like nature of solvent, modifier, temperature, pressure, duration of treatment, heating rate, grain size.
The effect of modifiers on the performance of supercritical extraction has been reviewed by several authors [18][19][20].Solvents used were pentane for the extraction of food products [21], polycyclic aromatic hydrocarbons [22], acetone [23] and hexane [24] to extract pesticide residues.Other authors [25,26] have also used reagents in order to degrade the solute molecules.
Lanças and Rissato [21] studied the effect of certain solvents (acetone, hexane, methanol) on the extraction of the herbicide Diuron in sugar cane and orange samples with supercritical CO 2 .They concluded that the addition of methanol as a modifier enhanced significantly the extraction yield compared to other modifiers.
Dariva et al. [27] reported that the addition of organic solvents in the extraction of oil shale with supercritical water improved the yield and quality of the extracted oils.
Many researchers determined the thermal characteristics and kinetic parameters of oil shale samples by thermogravimetry (TG) and differential thermogravimetry (DTG) under non-isothermal heating conditions for both pyrolysis and combustion processes [28][29][30][31][32].They also used differential scanning calorimetry (DSC) to determine the combustion kinetics of oil shale samples by an ASTM method.It was observed that higher heating rates resulted in higher reaction temperatures and higher heat of reactions.
Özgür et al. [33] studied the effect of co-firing of biomass fuels with oil shale on combustion, using a thermogravimetric analysis (TGA) and DSC.They observed that the biomass (10% and 20% by weight) blended with oil shale was a good option for co-firing owing to the formation of a sufficient amount of volatile matter to maintain stability in ignition and combustion.
Other studies [34][35][36] have also shown that the addition of CO or H 2 when extracting shale and coal with water improves the efficiency and quality of the oils produced.
In this work, we chose phenol as a modifier in the extraction of oil shale with supercritical toluene.This choice was dictated by the compound's thermal stability because phenols are associated by intermolecular hydrogentype bonds.In addition, the highly reactive hydroxyl group could react with some organic functions present in oil shale.Therefore, two principal objectives were fixed for the present research: (a) to study the effect of three factors (mineral matter, solvent and phenol as modifier) on the supercritical extraction of Moroccan oil shale to establish the optimal operating conditions likely to give a good yield of recovery with the best oil quality; (b) to demonstrate that phenol was a very efficient modifier for oil shale, giving a good yield of recovery and a suitable maturation of organic matter.

Materials
The oil shale used in this work was from Tarfaya deposit located in South Morocco.This deposit consists of several layers that are in turn subdivided into sub-layers, each comprising a different amount of organic matter.The samples were obtained from sub-layer R 3 , which is characterized by a high content of organic matter [37].The chemical composition of R 3 is given in Table 1 [38].The carbonate-free oil shale (RH) was obtained by dissolution of carbonates with HCl [39,40].The powdered R 3 shale (20 g, grain size 0.063-0.08mm) and 80 mL of concentrated HCl (7 M) were introduced in an Erlenmeyer flask.The mixture was then subjected to magnetic stirring for 4 h.The formed CO 2 was trapped by bubbling in a solution of barium hydroxide.After filtration, the solid residue (referred to as RH) was washed carefully with distilled water, dried at 100 °C and stocked in a sealed plastic bag.
The carbonate-free shale was stirred with concentrated hydrogen fluoride (4 ml g -1 RH) for 4 h.After filtration, the residue (K 3 ) obtained was first washed with dilute HCl solution and then with water and dried at 100 °C and stocked in a closed plastic sachet.

Preparation of the pitches
The supercritical extraction of sample RH (10 g) with toluene as solvent was conducted in a 120 mL stainless steel autoclave equipped with a stirrer and heated in a tubular furnace whose temperature, as well as heating rate, were controlled.The temperature of extraction was 390 °C with a heating rate of 16 °C min -1 [41].The maximum pressure reached during this treatment of 120 min was 8 MPa.After being cooled to room temperature, the mixture was extracted in a Soxhlet apparatus with chloroform for 12 h.After removing the solvent under reduced pressure, the organic material was dried for 12 h at 40 °C and weighed.The recovered oil was treated with n-hexane in a 1/10 oil to solvent mass ratio [42] in order to precipitate the high-molecular fraction (asphaltenes) that constitutes the pitch.After stirring for 12 h, the two fractions, soluble (maltenes) and insoluble (asphaltenes), were separated by filtration through Whatman paper, and then dried for 12 h at 40 °C and 80 °C, respectively.The scheme of the autoclave is shown in Figure 1.

Preparation of carbon fibers
The pitch was melted at 180 °C, in a laboratory-made spinning device equipped with a single spinneret and a filter to remove solid particles from the melt [43].After extrusion and mechanical stretching, the fiber was thermoset by being slowly heated in dry air at atmospheric pressure to temperatures in the range of 60-280 °C.The oxidized fibers were pyrolyzed under nitrogen flow, the temperature was progressively raised up to 950 °C at a heating rate of 5 °C min −1 [44].

Analyses
Elemental analysis was performed at the Service Central d'Analyse du CNRS, France.
The analyses by X-ray photoelectron spectroscopy (ESCA-XPS) were carried out in an Escalab VG220i-XL apparatus.
Infrared (IR) spectra were recorded between 400 and 4000 cm -1 with a Nicolet 205 FT-IR spectrometer by summing 32 scans at 2 cm -1 resolution.Pellets were prepared by dispersing 2 mg of sample into 198 mg of KBr.
Thermogravimetric analyses were performed with a TGS-2 Perkin-Elmer analyzer under high purity argon flow (40 mL min −1 ).Samples of about 12 mg were heated from 50 to 950 °C at a heating rate of 5 °C min −1 .
The surface areas of the fiber were measured by nitrogen adsorption at 77 K with a Micromeritics Flowsorb II 2300 instrument using the Brunauer-Emmet-Teller (BET) method.Before adsorption measurements the samples were outgassed at 160 °C for 24 h.
Scanning electron microscopy (SEM) micrographs were taken with a Hitachi TM-1000 microscope.Fragments of about 0.5 cm 2 cut from the corresponding samples were mounted on a carbon tab, which ensured a good conductivity.A thin layer of gold-palladium was sputtered prior to analysis.
Size exclusion chromatography (SEC) was carried out on a Waters analytical system (pump 510, refractometer 410 and UV/visible detector 486) with teterahydrofuran (THF) as eluent at 1 mL min -1 and using a TSK GMHXL column with a mixed porosity of 1500-10 7 Å.Relative average molecular weights were calculated with a cubic, 12-point calibration curve obtained from monodisperse polystyrene standards. 1Hydrogen nuclear magnetic resonance ( 1 H NMR) spectra were recorded on a Bruker AC 250 spectrometer (250 MHz).The samples were placed in 5 mm-ID tubes with CDCl 3 as solvent.The chemical shifts are given in ppm relative to TMS δ = 0 ppm.

Effect of mineral matter on the yield of recuperation of the obtained oil
In order to show the effect of mineral matter on the yield of recuperation and composition of the extracted oils, we conducted extractions of three different samples according to the protocol described in the experimental part: sample 1, sub-layer R 3 ; sample 2, sub-layer R 3 freed of carbonates RH; sample 3, sublayer R 3 freed of carbonates and silicates K 3 .
Before carrying out the supercritical extraction of the samples, X-ray Photoelectron Spectroscopy (XPS) analyses were performed to identify the composition of the starting materials and to follow the evolution of this composition throughout the demineralisation process.By comparing the results given in Table 2, it can be noted that sample RH is characterized by the absence of Ca and Mg, which confirms the fact that the treatment of the rock by HCl dissolved completely the carbonates.We also noticed that HF dissolved silicates and clays (sample K 3 ).Extraction yields obtained for samples R 3 , RH and K 3 are given in Table 3.The yields of recuperation were calculated from the total amount of organic matter contained in each sample.Results in Table 3 show that the yield of recuperation varies considerably among the different samples.Indeed, the yields obtained for samples RH and K 3 are more significant than for sample R 3 , which shows the effect of mineral matter.The poor yield obtained for sample R 3 can be explained by the high porosity of carbonates, which allows organic matter to be retained in the mineral matrix and consequently delays its extraction.To study the effect of mineral matter on the composition of the oils, elemental analyses of the oils obtained from samples R 3 , RH and K 3 were conducted.The results are presented in Table 4. Atomic ratios H/C and O/C of the three oils are shown in the Van Krevelen diagram (Fig. 2) demonstrating the effect of mineral matter on the degree of maturation of the obtained oils.From this diagram it can be noticed that the maturation degree of the oil shale's organic matter varies considerably with the samples used in supercritical fluid extraction.Indeed, the oil K 3 exhibits a considerably higher maturation degree, compared to oils R 3 and RH.The oils obtained range from an immature organic matter of I-a type (oil R 3 ) to an immature organic matter of II-a type (oil RH), or a more mature organic matter, rich in carbon and which is related to standard catagenesis II-b (oil K 3 ).Those results were confirmed by SEC analyses (Table 5).
The comparison of those values shows that the oil obtained with sample K 3 has the highest value of the average molecular weight (M ̅ w ), compared to oils R 3 and RH.This suggests that, under these conditions, toluene is able to extract large molecules from kerogen.According to the results above sample RH presents several advantages, not so much on the level of preparation (less consumed products), but on the level of the results obtained (high yield of recuperation).Moreover, the presence of fluorine in the organic matter resulting from sample K 3 makes its use difficult.Oumam [44] has previously reported that during the carbonization of the organic matter resulting from K 3 , the walls of the silica tube were gradually attacked by carbonization gases rich in fluorine.These reasons led us to use RH as the starting material to carry on our study.

Effect of the solvent
In order to study the effect of solvent on the yield of recuperation and composition of the oil, the extraction was carried out with 10 g of sample RH by using 60 mL of different solvents (toluene, water, shale oil) under the conditions described above.The yields and compositions of the oils obtained under different conditions are given in Table 6.The results demonstrate that the yield and composition of the oil obtained by extraction with water are obviously different from those of the oil obtained by extraction with toluene or shale oil.The oil's higher yield and maltene content indicate that water not only acts as an extraction solvent, but also reacts with the molecules of kerogen.It also reacts with the double and triple bonds of the decomposition products, resulting in alcohols and aldehydes [45,46].The results of elemental analysis of the obtained oils are presented in Table 7.Those results show clearly that water is an optimal solvent, offering the highest maturation of the organic matter of oil shale.This reinforces a hypothesis that, under these conditions, water is able to degrade large molecules contained in kerogen.The disadvantage of using water is the high pressure generated during the supercritical extraction of oil shale with water (Table 6).To overcome this problem, we proceeded to the extraction of oil shale with toluene by adding phenol as a modifier.
In order to study the effect of phenol on the yield of recuperation and composition of the obtained oil, the extraction was carried out with 10 g of sample RH by using different volumes of toluene and 5 g of phenol according to the same conditions as described previously.The results obtained are shown in Figure 3.
By comparing the two curves (Fig. 3) it can be seen that their appearance is the same, but the mass of the insoluble material is lower in the case of a mixture (toluene + phenol).The decrease in the mass of the chloroform-insoluble material with the addition of phenol can be explained by the fact that phenol can react with the molecules of kerogen which causes the decomposition of large molecules and therefore leads to molecules of smaller sizes [47].It may be noted, for example, that the transalkylation reactions were observed by Leach [47] to take place between 350-550 °C.Moreover, the transformation of phenol to phenolate, in the presence of traces of water or bases in the medium gives it a character of good nucleophile which can attack different carbonyl groups.A previous work by Koel et al. [48] on using neoteric solvents in oil shale studies shows that solvents attack kerogen and its chemical modification or degradation occurs.

Analysis of the pitches by SEC
The results of the SEC analysis of the pitches are presented in Table 8.The pitch obtained using 70 mL of toluene (PHT70) has a high value of M ̅ n , i.e. toluene is able to recover large molecules without degradation of kerogen.In contrast, in the case of extraction by mixing toluene and phenol, the M ̅ w value obtained for the PHT70PH5 pitch is very low.This suggests that phenol has a degrading action, thus providing a pitch formed of low molecular weight molecules.The results of thermal analysis (TGA/DTA) of the pitches (Fig. 4) show that pitch PHT70 is characterized by a mass loss that starts around 100 °C, with a peak around 420 °C, and ends at 550 °C.The residue yield at 950 °C is about 32%.It is also noted that the thermogram of pitch PHT70PH5 is completely different from that of the former.It exhibits a mass loss that spreads over the same temperature range as that of pitch PHT70, but with the appearance of a complex phenomenon consisting of 3 successive mass losses at temperatures of 150 °C, 300 °C and 450 °C.Furthermore, we note that the residue yield at 950 °C for this pitch (36%) is somewhat higher than that of PHT70.A possible explanation for the agreement with the measurements by SEC could be that pitch PHT70PH5 consists of light, but thermally quite stable molecules compared to pitch PHT70.These results are in good agreement with SEC measurements.

Elemental analysis
The elemental analysis of the obtained pitches (Table 9) showed that pitch PHT70PH5 possessed a very low sulfur content and a higher aromaticity factor (low H/C) compared to pitch PHT70.The preceding pitches were also characterized by 1 H NMR spectroscopy (Fig. 5).In the spectrum of pitch PHT70, signals corresponding to aliphatic protons located between 0.5 and 3 ppm are prevalent.On the contrary, aromatic protons are very marginal.On the other hand, the spectrum of pitch PHT70PH5 reveals the presence of aliphatic and aromatic protons.Presumably, aromaticrich compounds were extracted from heavy molecules in that case, which corroborates the results of elemental analysis of this pitch.This confirms that phenol was an efficient modifier likely to give a good yield of recovery and a better quality of oil.

Characterization of the filaments
At 950 °C the carbonization yield was about 49%.The elemental analysis of the pyrolyzed filaments (C 98.83, S 1.08, Fe 0.03, Si 0.06%) showed the extracted organic matter to contain a minimum of metal impurities.Mechanical tensile tests performed on monofilaments of about 20 µm in diameter revealed a tensile strength of 320 MPa and a Young's modulus of 30 GPa.These values are modest when compared to those found for fibers elaborated from precursors with oriented textures.They are however comparable to those of fibers obtained from coal tar or petroleum isotropic pitches [49][50][51].
In order to improve the mechanical properties of the fibers, we tried to increase the pitches softening temperature by adding a supplement step in their preparation.This step consisted firstly in precipitation of oils with ethanol followed by washing with hexane.This was intended for elimination of the majority of light ends and, of the polar functional groups since the presence of heteroatoms makes the materials obtained either crosslinkable or graphitizable [52].So we used this method to prepare a new pitch from HT70PH5 oil.The softening temperature rose to 260 °C.
We analyzed this new pitch, PHT70PH5M, by TGA/DTA (Fig. 6) in order to compare the results to those obtained for pitch PHT70PH5.The mass loss corresponding to molecules of low mass observed in Figure 4b between 180 and 300 °C for pitch PHT70PH5 has completely disappeared.We also noted an increase in the residual rate from 36% for pitch PHT70PH5 to 46% for the new pitch PHT70PH5M.This result confirms that the new method of preparation of pitches leads to materials containing heavier molecules.
We tried to spin the new pitch on the same principle of the spinning described above.The fibers obtained in this case are very thick, d = 850 mm, and they have the shape of rods.This can be explained by the fact that the spinning temperature is the decomposition temperature of the pitches.Furthermore, we noticed that those rods had a very clear porous structure.Therefore, analyses by SEM were performed on the rods (Fig. 7).The SEM pictures show highly   porous structures.The micrographs (Fig. 7a, 7b) demonstrate the distribution of pores throughout the wand and Figure 7c indicates the presence of pores in the heart of the stick.The SEM analysis displays a higher homogeneity of the pores within our sticks.It therefore seemed to us very interesting to determine the specific surface of these rods.To this end, we determined the surface area of rods by using nitrogen sorption measurements.The resulting value S BET is of the order of 300 m 2 g -1 , which suggests that activation of these fibers can result in activated carbon of good quality that could be used as adsorbent or filter in environmental applications.

Conclusions
In conclusion, we can remark that the yield and composition of the obtained oils depend on the solvents and the nature of material used to extract the organic matter from oil shale.In addition, the yield and composition of the oil obtained by extraction with water are markedly different from those of the oil obtained by extraction with toluene or shale oil.The oil's higher yield and maltene content indicate that water is an optimal solvent, offering the highest maturation of the organic matter of oil shale.The disadvantage of using water is the high pressure generated during the supercritical extraction of oil shale with water.To overcome this problem, we proceeded to the extraction of oil shale with toluene by adding phenol as a modifier.The current study showed that phenol was a very efficient modifier for the supercritical extraction of the organic matter from Tarfaya oil shale with toluene, giving a good yield of recovery and a suitable maturation of the organic matter.The pitches prepared by mixing phenol and toluene contain more aromatics and have a high char yield at 950 °C compared to those obtained by extraction with supercritical toluene alone.We also noticed that the chemical modification of the pitches afforded very thick fibers with a highly porous surface, which suggests that activation of these fibers can result in activated carbon of good quality that could be used as adsorbent or filter in environmental applications.

Fig. 2 .
Fig. 2. Classification of the oils extracted from the shale materials by the Van Krevelen diagram.

Fig. 3 .
Fig. 3. Variation of the insoluble material in CHCl 3 as a function of toluene volume.

Table 4 . Elemental analyses of the oils, atomic %
* O is determined by difference

Table 7 . Elemental analyses of the oils obtained from RH, atomic %
* O is determined by difference