**1. Introduction**

Global warming is a problem that could be mitigated by replacing fossil energy sources by renewable energy sources, such as green biomass. [1]. In recent decades numerous research papers have addressed the use of lignocellulose materials to obtain ethanol through biochemical routes [2–4], describing bioprocesses that are mainly composed of three major stages: Biomass pretreatment, cellulose hydrolysis, and sugars fermentation. Acid prehydrolysis at high temperature (around 200 ◦C) is one of the most efficient pretreatments to remove hemicelluloses and extracts present in lignocellulose materials, obtaining a cellulose and lignin-rich pretreated solid [5,6]. This type of pretreatment can be carried out with low concentrations of acid and short reaction times (few minutes), being able to generate liquid prehydrolysates with high concentrations of hemicellulose sugars, provided that suitable operating conditions are used. Thus, for a fixed acid concentration, there will be optimal temperature and reaction time conditions that will lead to complete hydrolysis of the hemicellulose and will maximize the concentration of hemicellulose sugars in the liquid prehydrolysate. The search for these conditions could be approached using response surface methodology as previously described

elsewhere [7,8]. In relation to the type of acid used, sulfuric acid is generally used instead of hydrochloric acid due to its low volatility, lesser equipment corrosion and lower cost per mole of protons [9], so it is frequently used at this stage of the bioprocess.

To improve the biomass fractionation, pretreated solids could be subjected to enzymatic hydrolysis with cellulases, which would allow a selective conversion of cellulose into D-glucose under mild operating conditions (pH 4.8, 50 ◦C temperature). It is of great importance to assess the effect of enzyme loading in the enzymatic hydrolysis, since a low concentration of the same would lead to low monosaccharide yields, while a high concentration would excessively increase the operating costs.

Concerning sugars fermentation, the use of non-traditional yeasts, such as *Pachysolen tannophilus*, would allow converting both hexoses and pentoses into ethanol [10], provided that fermentative inhibitors concentrations are cut down. In this sense, the levels of furfural and acetic acid in the acid prehydrolysates could be reduced by vacuum distillation because of the higher vapor pressures of these compounds with respect to those of monosaccharides. A variable of great interest in the fermentative stage is the inoculum concentration since this is responsible, among other factors, for the bioprocess rate.

Olive stones are a lignocellulose material that is obtained in great quantities in the olive oil mills [11]. Therefore, this biomass is a potential source of biofuels in countries with large olive production, such as Spain, Italy, Greece, or Portugal. The biochemical conversion of olive stones into ethanol, using hydrothermal or acid pretreatments, has been partially studied by different authors. Thus, Miranda et al. [12] applied a hydrothermal pretreatment to olive stones (130 ◦C, 30 min) achieving a good enzymatic digestibility of the cellulose fraction. However, the further use of liquid prehydrolysates is usually not addressed. The liquid prehydrolysates are rich in D-xylose and xylooligosaccharides as described in other works in which water was used as a hydrolytic agent [13,14]. On the other hand, Romero-García et al. [15] applied 2% (wt.) sulfuric acid (130 ◦C, 60 min) to olive stones, achieving a high production of hemicellulose sugars (mainly D-xylose) that were fermented to ethanol after undergoing detoxification by overliming. Notwithstanding, this work did not address the use of the pretreated solids, thus remaining as waste. Saleh et al. [16] applied dilute sulfuric acid (0.025 M H2SO4) to olive stones (195 ◦C, 5 min) to hydrolyze the hemicellulose and obtain a D-xylose-rich prehydrolysate, which was subsequently detoxified and fermented. The pretreated solids were subjected to enzymatic hydrolysis, but the obtained hydrolysates were not fermented despite their high sugars content.

The present study aims to develop a complete scheme of ethanol production from olive stones by applying two consecutive hydrolytic stages (acidic and enzymatic) followed by the subsequent stage of sugars fermentation. The acid pretreatment was optimized to maximize the recovery of hemicellulose sugars, while the effect of enzyme loading on D-glucose production in the subsequent stage of enzymatic hydrolysis was studied. The non-traditional yeast *Pachysolen tannophilus* was used for the fermentation of both the acid prehydrolysates and the enzymatic hydrolysates, and the effect of inoculum concentration on the fermentation performance was assessed. The fractionation applied to the biomass was envisaged by determining the mass macroscopic balances at the different stages of the process. In this way, a rapid description of the conversion of olive stones into ethanol and other bioproducts was achieved.

#### **2. Materials and Methods**

#### *2.1. Raw Material*

Olive stones (fragmented endocarps) were supplied by the olive mill S.C.A. San Juan (Jaén, Spain, UTM coordinates: 37◦47 58.52 N, 3◦47 09.42 W). Once at laboratory, the raw material was washed and then dried at room temperature for three weeks. Afterwards, the olive stones were screened using a vibratory screen (Restch, Mod. Vibro). The solids used in this research had diameters lower than 3 mm while those of diameter lower than 0.85 mm represented less than 3% of the total sample weight. Finally, the dry solids were stored in sealed plastic bags at room temperature until used.

#### *2.2. Dilute Acid Pretreatment*

Dilute acid hydrolysis was carried out in a 2 dm<sup>3</sup> Parr reactor, Series 4522 (Moline, IL, USA). Fo experiments, this reactor was loaded with 50 g of dry solids and 300 cm3 of 0.010 M sulfuric-acid solution. The chosen H2SO4 concentration was significantly lower than that used in a previous work published by us (0.025 M, [16]) to reduce the corrosion that the acid causes on the reactor. The mixture was stirred at 250 rpm and quickly heated up to work temperature, which was maintained over the selected reaction time (see Section 2.3). Subsequently, the reactor was cooled to room temperature (in less than 10 min) by circulating cold water through an internal coil. The liquid prehydrolysate was separated from the pretreated solid by vacuum filtration. The latter one was water-washed and dried at room temperature. The water used for washing the pretreated solid was added to the prehydrolysate until reaching a final volume of 2 dm3. The two separate phases were stored for later characterization and used in the following stages of the scheme.

#### *2.3. Response Surface Methodology (RSM)*

Response surface methodology (RSM) was applied to data to optimize the acid pretreatment conditions by multiple regression analysis. To do this, the Modde 7.0 (Umetrics AB, Umeå, Sweden) software was used. The 22 central composite circumscribed design (CCCD) with two independent variables (temperature and time) at two different levels, four star (axial) points and three central points (total 11 runs, Table 1) was assayed to find linear, quadratic and interaction effects of the independent variables (operational parameters) on the experimental responses (experimental results). The temperature (180–220 ◦C) and time (2–8 min) ranges were selected from the results obtained in preliminary experiments (data not shown) with the aim of achieving as D-xylose recovery as possible in the liquid prehydrolysates. The statistical validation was carried out by one-way ANOVA test (95% confidence), and the optimal conditions values were determined from the response surfaces using the SIMPLEX method.



#### *2.4. Detoxification and Fermentation of Acid Prehydrolysates*

The acid prehydrolysates were detoxified using a Buchi R-114 rotary evaporator (BÜCHI Labortechnik AG, Flawil, Switzerland) at 40 ◦C, thus removing inhibitor compounds (acetic acid, furfural, and 5-hydroxymethylfurfural) and achieving a D-xylose concentration close to 15 g/dm3.

Fermentations of acid prehydrolysates were carried out using the yeast *P. tannophilus* ATCC 32,691, supplied by the American Type Culture Collection. The microorganism was stored in a cold room (5–10 ◦C) in 100-cm<sup>3</sup> test tubes on a sterilized solid culture medium with the following composition: 3 g/dm<sup>3</sup> yeast extract (Becton Dickinson Co), 3 g/dm3 malt extract (Merck), 5 g/dm3 peptone from casein (Merck), 10 g/dm<sup>3</sup> D-xylose (≥99% purity, Panreac), and 20 g/dm3 agar–agar (Panreac). For pre-inocula, cells were transferred to a sterile medium with the same aforementioned composition and kept in

an incubator at 30 ◦C for 60 h in order to obtain cells at the same growth stage at the beginning of each fermentation. Afterwards, the pre-inocula were transferred to 250-cm3 Erlenmeyer flasks along with 100 cm3 of sterile liquid culture made of 4 g/dm3 yeast extract (Becton Dickinson Co), 3.6 g/dm3 peptone from casein (Merck), 3 g/dm3 (NH4)2SO4 (99% purity, Panreac), 2 g/dm3 MgSO4·7H2O (99,5% purity, Carlo Erba), 2 g/dm3 KH2PO4 (99% purity, Panreac), and 25 g/dm3 D-xylose (≥99% purity, Panreac). Cultures were incubated at 30 ◦C for 24 h in an orbital shaker (150 rpm). Then, yeast cells were recovered by centrifugation at 7000 rpm for 10 min, washed with a dilute NaCl solution, and suspended in the fermentation medium to obtain initial inoculum concentrations of 0.5, 1.0, 2.0, and 4.0 g/dm3. These concentrations (*x*, kg/m3) were calculated from the absorbances of the cultures at 620 nm using a previously obtained absorbance versus dry-weight calibration line [10].

Fermentations were carried out with 30 cm<sup>3</sup> of prehydrolysate inside 100 cm3 Erlenmeyer flasks. The prehydrolysates were supplemented with 2 g/dm<sup>3</sup> yeast extract (Becton Dickinson Co), 1.8 g/dm<sup>3</sup> peptone from casein (Merck), 1.5 g/dm3 (NH4)2SO4 (99% purity, Panreac), 1 g/dm3 MgSO4·7H2O (99,5% purity, Carlo Erba), and 1 g/dm3 KH2PO4 (99% purity, Panreac). The resulting cultures were sterilized using a glasswool pre-filter and a 0.2-μm pore-size cellulose nitrate filter. Temperature (30 ◦C) and pH (4.5) were chosen according to previous works [10,17] and kept constant over fermentations. The aeration was only supplied by the stirring vortex (microaerobic conditions). The cultures were sampled at fixed intervals to analyze the biomass, D-glucose, D-xylose, acetic acid, ethanol, and xylitol concentrations. Two replicas of each fermentation were performed.

#### *2.5. Enzymatic Hydrolysis*

The washed and dried water-insoluble solids obtained in the dilute acid hydrolysis of olive stones were submitted to enzymatic hydrolysis in order to obtain D-glucose from the cellulose. To do this, 3 g of dry solids were suspended in 30 cm<sup>3</sup> of 0.05 M citrate buffer solution (pH 4.8) inside 125 cm3 Erlenmeyer flasks. Enzymatic hydrolyses were carried out at 50 ◦C for 72 h on an orbital shaker (150 rpm). A commercial preparation of *Trichoderma reesei* cellulases (Celluclast 1.5L, Novo Nordisk Bioindustrial, Madrid, Spain) was used throughout this research. Enzyme loadings of 10, 20, 40 and 60 FPU per g dried solid were added to the Erlenmeyer flasks. 1-cm3 samples were withdrawn from the reaction media at 4, 10, 24, 48, 72, and 120 h to analyse the D-glucose concentration. The D-glucose yield was calculated as g of D-glucose per 100 g of initial pretreated solid. All the enzymatic hydrolyses were performed in duplicate.

#### *2.6. Fermentation of Enzymatic Hydrolysates*

The fermentation of the enzymatic hydrolysates was performed with the yeast *P. tannophilus* ATCC 32691 following the procedure described in Section 2.4 but with two modifications: Enzymatic hydrolysates were not detoxified, and the initial inoculum concentrations assayed were 0.5, 1.0, 1.5, and 3.0 g/dm3. The cultures were sampled at fixed intervals to analyse the biomass, D-glucose and ethanol concentrations. Two replicas of each fermentation were performed.
