Hydrothermal hydrolysis of grape seeds to produce bio-oil

Abstract

In the present work, the hydrothermal hydrolysis of grape seeds focused on the production of bio-oil was studied. The grape seeds composition in terms of lignin, sugars, ash, extractives and bio-oil was determined. The composition of grape seeds was: 17.0 wt% of extractives; 36.8 wt% of sugars (hemicellulose and cellulose); 43.8 wt% of lignin and 2.4 wt% of ash. The grape seeds were hydrothermally treated using three different temperatures: 250 °C, 300 °C and 340 °C employing a semi-continuous reactor. The solid residue varied from 25.6–35.8 wt% depending on the hydrolysis temperature. The maximum yields of light (15.7 wt%) and heavy bio-oil (16.2 wt%) were achieved at 340 °C. The Arrhenius parameters for the kinetics of grape seeds hydrolysis in our system were k₀ = 0.995 g min⁻¹ and E_a = 13.8 kJ mol⁻¹. The increment of the flow rate favoured the mass transfer in the system and so, the hydrolysis rate. However, the maximum hydrolysis rate was found at a water surface velocity of 2.3 cm min⁻¹.

---

1. Introduction

The industry based on fossil raw materials will not be sustainable for many decades, and it is necessary to start the shift towards natural materials as the main source of carbon-based materials.¹ This novel economic paradigm is known as bioeconomy.² Vegetal biomass is a renewable source of carbon material that can be used as feedstock for the production of chemicals and fuels. Nowadays, vegetal biomass is one of the most abundant resources of carbon on the planet, being the primary production around 1 × 10¹⁴ tons C per year.³ The three main products that are considered “bio-based products” are: bio-fuels, bio-energy and bio-chemicals, which would be produced in biorefineries.^4,5

In recent decades, the burning of fossil fuels has contributed to increase the level of CO₂ in the atmosphere generating the global warming observed. Because of this and the uncertainty in the petroleum market, the governments of many countries around the world have promoted the production of biofuels. To achieve this change in the production philosophy, it is necessary to develop new technologies for the sustainable and efficient conversion of natural resources into fuels and chemicals. In the last two-three years the use of shale gas in USA and other countries has stopped the use of bioresources for energy but not the interest of added-value products.

The main advances in the production of liquid fuels from biomass was achieved converting corn and sugar cane crops into ethanol. The main disadvantage of these first-generation biofuels was the competition with the production of food, which has caused an increase of food prices under several conditions. For this reason the second-generation biofuels produced from surplus lignocellulosic feedstock can be more feasible. Lignocellulosic biomass has three major components: cellulose, hemicellulose and lignin with minor contents of extractives and ash. The relative quantity of hemicellulose, cellulose and lignin depends on the biomass used as well as variations in methods of cultivation, climatic factors, physiographic variability, vintage year, harvest date, etc.^6,7

There are a large number of methods to fractionate biomass using ionic liquids, organic solvents, supercritical fluids, compressed water, steam pre-treatment, several alkaline treatments, etc.^8–11 The most popular supercritical fluids are CO₂, water and propane. Other of the methods used to fractionate the three lignocellulosic components is the hydrothermal treatment using hot compressed water as solvent. Water is a non-toxic, environmental friendly and inexpensive reaction medium.¹² In addition, the hydrothermal biomass conversion require lower temperatures than other processes such as pyrolysis or gasification.^13–15

Viticulture is an important agricultural activity in Spain, producing a massive amount of stalks and grape seeds as by-product. In the last decade, the production of grapes was close to 6 million tons produced annually.¹⁶ This trend consolidated Spain to be the top wine worldwide producer in 2014, together with Italy and France.

Grape seeds contain an important concentration of lignin (43.8 wt%), extractives (17 wt%) and cellulose and hemicellulose (36.8 wt%).

The production of biofuels from different raw materials have been studied by several authors. Tekin and Alkalin et al. have studied the production of light bio-oil (LBO) and heavy bio-oil (HBO) from beech wood and cornelian cherry stones with and without catalysts (i.e. colemanite).^17,18 Huang et al. maximized the LBO production by recycling the HBO to the hydrolysis process obtaining also, higher yields of char.¹⁹ The pyrolysis of coconut shells;²⁰ rice hulk²¹ and corn stover²² was studied as alternative to hydrothermal treatment for the production of LBO and HBO.^20–22

The use of supercritical CO₂ as co-solvent did not alter significantly the microscopic morphology of biomass. The best advantage is that it can be easily removed by depressurization without generating by-products. The CO₂-assisted process can be carried out at similar temperatures than the common autohydrolysis process.^23,24 Thus, Magalhães da Silva et al. have demonstrated that CO₂ may assist the autohydrolysis via formation of carbonic acid promoting the production of xylo-oligosaccharides. They tested the autohydrolysis of wheat straw from 180 to 210 °C and initial CO₂ pressures of 60 bar. The water/biomass ratio was 10 : 1 w/w.^23,25

In this work, grape seeds were hydrolysed at different temperatures in liquid phase using the autohydrolysis process. The production of HBO, LBO, total amount of sugars and solid residue were determined. The extraction of the oil and the recovery of the liquid bio-oil and solid residue after the treatment was the aim of this work.

2. Materials and methods

2.1. Materials

Grape seeds from Vitis vinifera L (Tempranillo) were provided by Matarromera S.A. winery (Valbuena de Duero, Spain) campaign 2011. The raw seeds were crushed and sieved; a size 0.5–1.0 mm was selected for experiments.

The reagents used were: sulphuric acid (96%), acetone (99.5%) purchased from Panreac and diethyl ether (+99%) purchased from Sigma-Aldrich. Distilled water and Milli-Q water were used in the experiments.

2.2. Pilot plant

A diagram of experimental set-up is shown in Fig. 1. The experimental apparatus consisted of a pump (model: Jasco PU-2080), a preheater (E-01, 200 cm of 1/8′′ AISI 316 piping) and a reactor (R-01, 20 cm length, 1/2′′ O.D. SS316 piping). The reactor was set inside a former chromatographic oven HP5680 (F-01). The outlet of the reactor was connected to a heat exchanger in order to cool down the sample to 20 °C (E-02, 15 cm of concentric tube heat exchanger 1/4′′-3/8′′ counter current operation). The cooling fluid used in E-02 was cold water in the outer tube of the heat exchanger. The pressure of the system was controlled by a Go-backpressure valve (BPV-01).

Fig. 1 Schematic flow diagram of the experimental system. Equipment: D-01 Feeder, P-01 Pump, E-01 Preheater, R-01 Reactor, F-01 Chromatographic oven, E-02 Heat exchanger, BPV-01 Go-backpressure valve, D-02 exit.

The operation mode was semi-continuous (i.e. batch for the solid and continuous for the liquid). The reactor (R-01) was filled with 4.00 ± 0.06 g of grape seeds (not dried before extraction, moisture 6.00 ± 0.20 wt%) and then the reactor was closed with two metallic filters (0.1 mm) to avoid material losses. Once the reactor was tightened the flow was fixed at 5 ml min⁻¹ and the pressure was set to 10 bar over the bubble point at the desire temperature. After the reaction time has elapsed, the reactor was gradually cooled down to room temperature by changing the set point in the F-01. Finally, the pumping was stopped and the reactor was untightened. The remaining solids in the reactor after the hydrolysis process were collected for later analysis of lignin and heavy bio oils content. The liquid obtained during the experiments were collected to determine the light bio oil production.

Three trials were performed at each experimental condition to check repeatability of the results.

2.3. Bio-oil analysis

The sequence of grape seeds analysis after hydrothermal treatment are schematised in Fig. 2.

Fig. 2 Procedure for the analysis of solid and liquid products from hydrolysis.

The remaining solid inside the reactor after extraction, was oven-dried for 24 hours at 100 °C (SR0). Then, it was subjected to solid liquid extraction with acetone at 25 °C. This mixture was filtered and the solid was dried for 24 hours at 100 °C (SR1). On the other hand, the acetone was removed from the liquid sample with a rotary evaporator under reduced pressure (−0.8 barg) at 80 °C. This fraction was called heavy bio-oil 1 (HBO1).

An aliquot from the liquid samples collected during the hydrothermal treatment was acidified using H₂SO₄ (96%) to improve the precipitation of solids (if any). They were filtered using a cellulose membrane with a pore diameter of 100 μm. The solids were dried at 100 °C (SRL0). Then they were subjected to a solid–liquid extraction with acetone to recover the heavy bio-oil trapped (HBO2). Finally, these solids were dried in an oven at 100 °C (SR2). On the other hand, the acidified liquid was exposed to a liquid–liquid extraction with diethyl ether (DEE) in equal volumes. The two phases obtained were then separated in a separation funnel. The organic phase was dosed with Na₂SO₄ and then, it was filtered. The DEE was removed in a rotary evaporator and the soluble light bio-oil (LBO) from the organic phase was recovered. The aqueous phase was dried to compute the sugars (SG).

2.4. Analytical methods

The analytical methods applied in the analysis procedure presented in Section 2.3 are reported below.

2.4.1. Acid insoluble lignin (Klason lignin). Klason lignin was determined according to the protocol developed by the National Renewable Energy Laboratory²⁶ in agreement with Liu and Wyman,²⁷ Toledano et al.^28,29 and Wang et al.³⁰

The Klason lignin content of the raw material and the solid residue inside of reactor after hydrolysis was determined as follows. An amount of solid (aprox. 300 mg) was weighted and placed into a hydrolysis flask with 3.00 ± 0.01 ml of sulphuric acid (72%). The mixture was incubated for 30 ± 5 min at 30 ± 3 °C in a convection oven. After this time, the mixture was taken out from the oven and it was diluted with 84.00 ± 0.04 ml of deionized water. Then, the mixture was kept at 121 °C for 1 hour in a convection oven. After the incubation, the solution was cooled to room temperature and it was filtered under vacuum. Hot deionized water was used to wash any particles adhered to the bottle and also to wash the filter residue in order to remove the remaining acid until pH neutral. The solid was dried at 105 ± 3 °C for at least 24 hours or until a constant weight and then it was weighted. Then, the crucible residue was introduced in a calcination oven at 550 ± 25 °C for 24 hours or until constant weight, and it was then weighted. The quantity of Klason lignin in each sample was determined using the eqn (1), where ‘KL’ is Klason lignin in wt%; ‘AIR’ is the acid insoluble residue after acid hydrolysis in mg; ‘A’ is the ash content in mg and; ‘S’ is the mass of the sample.

2.4.2. Thermogravimetric methods. Thermogravimetric analysis (TGA) was carried out in a TGA/SDTA RSI analyser of Mettler Toledo. The TGA analysis indicate how the hydrothermal treatment has hydrolysed the oil, hemicelluloses, celluloses and lignin. This analysis consists in analyzing the behaviour of a solid sample under gasification conditions over an inert atmosphere (N₂). The samples of approximately 10 mg were heated from 50 °C to 800 °C at a rate of 20 °C min⁻¹ under a N₂ atmosphere (60 ml min⁻¹) to determine the carbonization. In order to determine the final ash content, the sample was heated from 650 °C to 800 °C under an air atmosphere (60 ml min⁻¹) to promote to oxidation of the remaining organic carbonaceous forms.

2.4.3. Spectroscopy FT-IR. Fourier Transform Infrared FT-IR is a quick method that could be used almost in situ to determine functional groups in lignin after the treatment. The FT-IR experiments were conducted using a Bruker Tensor 27. Samples were recorded in the range of 4000–600 cm⁻¹ at 4 cm⁻¹ resolution and 32 scans per sample. The scanner velocity was 10 KHz and interpherogram size was 14220 points.

2.4.4. Microscopy, SEM. Scanning electron microscopy (SEM) experiments were conducted to identify the physical morphology of the samples. A JSM-820 (JEOL, Japan) operated at a 20 kV accelerating voltage and gold evaporator Balzers SCD003 were used. Gold Thickness used was 25–30 nm.

3. Results and discussion

Grape seeds were hydrothermally treated using water as a solvent for one hour at constant temperature. The experiments were performed at three different temperatures (see Table 1): 250 °C, 300 °C and 340 °C. The amount of bio-oil, solid residue and hydrolysed sugars was determined for each experiment.

Table 1 Experimental conditions for one-step fractionation of grape seeds

Experiment	N°	#01	#02	#03
Flow	ml min^−1	5.0	5.0	5.0
Temperature	°C	250	300	340
Pressure	barg	50	95	155
Time	min	60	60	60
Initial mass grape seeds	g	4.002	4.006	4.005
Moisture	wt%	6.00	6.00	6.00

---

The analysis (Klason lignin) of the grape seeds gave a total lignin content of 43.8 wt% similar to values reported in literature.^31,32 Considering that the ash content was 2.4 wt%, the extractable/hydrolysable components represented near to 53.8 wt% of the raw material. The maximum amount of grape seed oil to be extracted was 17.0 wt% dry basis (determined by Soxhlet extraction with hexane). The typical amount of oils in the grape seeds is between 11.0–20.0 wt%.³³ This indicates that the hemicelluloses and celluloses were approx. 36.8 wt% that can be hydrolyse into sugars.

3.1. Hydrolysis products

The solid residue (SR) was determined as the sum of the residues obtained after acetone extraction: SR1 and SR2. The total HBO produced was calculated as the sum of the oil content determined after acetone extraction: HBO1 and HBO2. The total BO produced was obtained as the sum of the LBO and the HBO. The total measured mass (R) was calculated as the sum of the SR, BO and the rest of the components determined in the hydrolysed liquid phase (SG). Each one of previously explained amounts was divided by the initial quantity of raw material in order to obtain the yield of the produced fraction (Table 2).

Table 2 Yields (wt%) of the obtained fractions after one-step fractionation of grape seeds

wt%	250 °C	300 °C	340 °C
SR1	30.9 ± 1.3	18.6 ± 0.9	12.3 ± 0.9
SRL0	10.7 ± 2	22.6 ± 4.4	28.6 ± 2.3
SR2	4.8 ± 3.1	9.7 ± 1.6	13.3 ± 2.5
SR	35.8 ± 4.3	28.3 ± 2.5	25.6 ± 3.4
LBO	8.1 ± 1.6	15.6 ± 3.1	15.7 ± 3.2
HBO1	7.9 ± 2	0.5 ± 0.4	0.3 ± 0.2
HBO2	5.3 ± 1.3	10.1 ± 0.8	15.9 ± 8.3
HBO	13.2 ± 3.3	10.6 ± 1.3	16.2 ± 8.5
BO	21.3 ± 4.9	26.1 ± 4.4	32 ± 11.7
SG	23.2 ± 1.5	26.8 ± 1.5	28.8 ± 1.7
R	80.2 ± 10.7	81.2 ± 8.4	86.3 ± 16.8

---

The amount of solid residue inside the reactor (SR1) decreased exponentially when the temperature was increased. The analysis of the solid obtained after the treatment at 250 °C revealed that more than 98 wt% was Klason lignin, so all the hemicellulose and cellulose fractions were hydrolysed. Although lignin was slightly degraded in water, it was observed that the total recovery of lignin as solid decreased with increasing the temperature. Under hydrothermal treatment, the residual lignin has an increased active area that may enhance the decomposition, as it can be inferred later from the carbonization photographs.³⁴ The amount of solid in the liquid phase (SR2) increased along with temperature (see Fig. 3). The total amount of solids decreased by increasing temperature, suggesting that at higher temperatures the kinetics of hydrolysis were faster. Similarly, Mehmet et al. found that the solid residue yields decreased when the residence time was increased at all experimental temperatures.¹⁷

Fig. 3 Variation of hydrolysis products with the temperature.

Similarly, the total obtained amount of bio-oil (BO = LBO + HBO) increased along with temperature. The same behaviour was observed by Mehmet et al. The highest yield of BO was 32.0 wt% at 340 °C. Cellulose and hemicellulose are firstly hydrolysed into sugars and then, the sugars can be converted into different compounds such as ketones and aldehydes via retro-aldol condensation and dehydration reactions.³⁵ These reactions are favoured at higher temperatures (increment of HBO) improving the yield of BO. It was also observed that the highest yield of HBO1 (bio-oil retained inside the solid) was obtained at the lowest experimental temperature. The hydrolysis process produced LBO yield between 8.1% and 15.7%. Tekin et al. studied the hydrothermal liquefaction of beech wood without and with colemanite. They reported that the amount of LBO increased with increasing temperatures (from 250 °C to 300 °C). The results of our study are in good agreement with this previous research.¹⁸ The aqueous-soluble products were obtained by filtration and subsequent drying (fraction SG). The yield of SG was increased from 23.2 wt% to 28.8 wt% when temperature was increased from 250 °C to 340 °C. Considering that, the aqueous products (SG) were mainly composed of sugars, and the yield obtained was close to the cellulose and hemicellulose content in the raw material. The mass balance of the experiments (R) was between 80–86 wt% of the initial product (see R in Table 2 and Fig. 3). These mass balance values are acceptable considering the small scale used (ca. 4.0 g of grape seeds) and the difficulty of accounting for the sugars provided: (a) only an aliquot was dried, (b) it was difficult to homogenise the hydrolysed product.

The solid samples obtained in SR1 as well as the raw material were analysed by TGA (Fig. 4). The TGA of grape seeds is labelled as ‘grape seeds’ and the TGA for the obtained solids at 250 °C, 300 °C and 340 °C are labelled ‘SR 250 °C’, ‘SR 300 °C’ and ‘SR 340 °C’ respectively. TGA is referred to the mass of the sample after the treatment, for this reason all the curves start from 100% (the values of the final solid residue are listed in Table 2). The curves have a sigmoidal shape and the most degraded samples by hydrothermal treatment were less gasified in the TGA. At about 625 °C all the curves exhibited a plateau (see ‘Grape seeds’ curve) indicating that the remaining biomass have produced char, that was non-degradable by gasification. The ash content was determined by totally oxidising the sample. For this, the gas phase was switched from N₂ (inert) to air. The approx. ash content was 2.4 wt%. The TGA results of the samples SR 250 °C and SR 300 °C were similar and the curves overlap. This phenomenon suggests that the solid composition after hydrolysis may not be altered at temperatures between 250 °C and 300 °C. On the contrary, when the hydrolysis was carried out at 340 °C, almost all the extractable and hydrolysable components were gone, and the remaining material accounts for ca. 12 wt% of the initial seeds. This value was low compared to the previous lignin contents in grape seeds reported in literature (43 wt%)^31,32 and also the value obtained in this work of 43.8 wt%. This indicates that part of the lignin was probably extracted or hydrolysed too. The grape seeds exhibited a plateau around 400 °C during gasification with a yield between 43–45 wt%.

Fig. 4 TGA analysis of the solids from hydrolysis and the raw material.

The inflexion point observed near 400 °C is related to the lignin content of the grape seeds. The extraction with water might modify the structure of the grape seeds so that the final char produced in TGA varied from ca. 17 wt% absolute at 250 °C (54 wt% relative Fig. 4) to around 10 wt% absolute (74 wt% relative Fig. 4) at 300–340 °C. The HBO1 was 7.90 wt% at 250 °C and less than 0.10 wt% in both 300 °C and 340 °C. This can be detected also in the TGA curve at 250 °C from 275 °C and 420 °C where the mass loss was similar to the HBO1 value (ca. 8–10 wt%). At 340 °C, less than 1 wt% of mass loss was observed in the same range of temperature. This does not mean that the TGA step indicates directly the HBO trapped in the solid residue, but they are related. At 300 °C on the HBO content was almost negligible. This was probably because the HBO was dissolved by the flowing water. There are not specific studies in the solubility of bio-oil in water at elevated temperatures presented in literature up to the best of our knowledge. Nevertheless, as an estimation from works related to fatty acids, the solubility at 250 °C may round 2–20 g l⁻¹ (0.20%), while at 300 °C will be close to 30–100 g l⁻¹ (3.00–10.0%) and at 340 °C would be completely soluble if they have not been hydrolysed yet.^36,37

Fig. 5 shows the SEM images of the solid residue obtained after the hydrothermal treatment. The hydrothermal carbonization (HTC), with and without catalysts (such as KOH, etc.), is a process in which nanostructures are created in the carbonized material increasing its adsorption capacity considerably.³⁴ Unur et al. have recently demonstrated the effectiveness of the hydrothermal treatment to produce high capacity adsorbents for batteries at temperatures up to 600 °C.³⁸

Fig. 5 SEM images of the solid residue when grape seeds were submitted to different temperatures for one hour. Solid residue (SR) with reaction at 250 °C (a and b), SR at 300 °C (c and d) and SR at 340 °C (e and f).

The samples obtained at 250 °C presented structures similar to those reported and illustrated by other authors.^39,40 The micrographs of the samples did not show sugar crystals probably because most of the sugars were hydrolysed and dissolved. Thus, a solid residue with a high purity in Klason lignin was obtained. In the all cases shown in Fig. 5, it can be observed a disorganization in the fibres, indicating that almost all the hemicellulose was removed. These results indicated that the solid residue was depleted in hemicellulose, which agrees with FTIR and TGA analyses. Also, the formation of carbon spheres in the solid residue was observed at 340 °C. Sevilla and Fuertes indicated that the presence of carbon spheres depend on the temperature of hydrolysis process, the reaction time, the concentration of the saccharides solution, etc.⁴¹ Recently, Reddy et al. have presented in a conference that these spheres may come from the dissolved and re-precipitated lignin.⁴²

The effect of the treatment over the solid samples was also analysed by FT-IR essays. The FT-IR spectrum of the raw material as well as the SR 300 °C and SR 340 °C are shown in Fig. 6. The main FT-IR bands detected are listed in Table 3.

Fig. 6 FTIR analysis of the raw material and solid product.

Table 3 Assignment of bands in FT-IR spectra of the grape seeds and solid residues at 300 °C and 340 °C

Wavelength cm^−1	Assignment^43,46
777	C–H deformation out of plane, aromatic ring
1037	C–O stretching vibration
1317	Aryl ring breathing with C–O stretch
1375	Existence of guaiacyl and syringyl groups
1402	C–H deformation
1424	C–C bounds and aromatic ring vibration of the phenylpropane groups
1465	C–H vibration of CH2 and CH3 groups and deformations and aromatic ring vibrations
1513	C–C bounds and aromatic ring vibrations of the phenylpropane groups
1608	Aromatic skeletal modes
1632	C=C benzene stretching ring
1717	C=O stretch, unconjugated ketone, carboxyl, and ester groups
1747	Ester-linked acetyl, feruloyl and p-coumaroyl
2868	C–H stretch in methyl and methylene groups
2933	C–H stretch methyl and methylene groups
3340	O–H stretch, H-bonded

---

As shown in Fig. 6, the spectra display several absorption peaks indicating the complex nature of the raw material and solid residue. Significant differences can be distinguished in the three samples. The bands in the region 2882–2942 cm⁻¹ are associated with ν(C–H) stretch in methyl and methylene groups. A reduction in the absorbance intensity of these bands located between in the spectra of the sample SR 340 °C with respect to the others samples were noted. The band at 1747 cm⁻¹ is characteristic of ester-linked acetyl, feruloyl and p-coumaroyl groups between hemicellulose and lignin.^43,44 The lack of this band in the treated samples (it was only observed in the untreated grape seeds) suggest that the links between hemicellulose and lignin were broken during the hydrothermal treatment.

The band at 1717 cm⁻¹ is characteristic of ν(C=O) of ketone, carboxyl and ester groups of hemicellulose. Thus, the band at 1717 cm⁻¹ was observed with more intensity in the grape seeds that in SR 300 °C and 340 °C indicating that the linkage between lignin and hemicellulose was broken, similarly.

The typical bands at 1515, 1465, 1424 and 1375 cm⁻¹ are characteristics of lignin⁴⁵ and they were observed in the three cases.

The bands at 1608 cm⁻¹ and 1632 cm⁻¹ are associated with the presence of lignin.⁴⁶ The presence of these bands was observed in all cases. Similarly, the aromatic ring bands at 777 cm⁻¹ were also identified.

The band at 2924 cm⁻¹ is associated with aliphatic –CH₂– which is a typical band of cellulose. This band can be observed in the three solid residual samples, but it is observed with less intensity as the operating temperature was increased.¹⁷

As it can be seen in Fig. 6, reduction in the absorbance of the typical bands of hemicellulose and cellulose at 2868, 2933 and 3340 cm⁻¹ after the treatment at 300 and 340 °C suggest that these fraction were hydrolysed.

Similarly, the aromatic ring bands are kept in the treated biomass.

3.2. Determination of Arrhenius parameters

To determine the kinetics in process of hydrolysis of lignocellulosic biomass is common to use simplified models. For instance, many authors use the severity factor.^47–50 Others, as it has been done in this work, use directly the zero or first order kinetics.^51–54

The effect of temperature in the hydrolysis rate was studied by determining the solid after 60 min of hydrothermal treatment. The analysed temperatures were: 150, 175, 200, 275, 300, 325 and 340 °C.

It was assumed that the observed reaction rate behaved following a zero order reaction (not depending on the concentration of the remaining biomass). Thus, the reaction rate was calculated as the mass of biomass degraded per time (in this case 60 min). The results are depicted in an Arrhenius plot in Fig. 7. The pre-exponential factor of Arrhenius relationship was k₀ = 0.995 g min⁻¹ and the activation energy was E_a = 13.8 kJ mol⁻¹. The regression coefficient R² = 0.98 shows a good relation between experimental data and predicted data, indicating that this model is representing the process behaviour.

Fig. 7 Arrhenius plot for the observed hydrolysis rate.

A direct comparison of the kinetic parameters is difficult due to the differences in substrate materials, kinetics models and differences in the process.

The study of production of xylose from sugar cane bagasse by acid hydrolysis was carried by Aguilar et al.⁵² They used temperatures between 100 to 128 °C and concentrations of sulphuric acid between 2% to 6%. The activation energy average values reported were between 110.9 to 159.6 kJ mol⁻¹. Those values are similar to others authors for other lignocellulosic biomass.^51,53,54 The difference of values of energy activation between the literature and this study is significant, however the difference in the biomass used and the presence of acid in the process can modify considerable the behaviour. Also, the existence of big amount of extractives in the raw material can influence in the reduction of activation energy, as in the case of grape seeds.

3.3. Effect of the flow rate

The effect of the flow rate was analysed at 250 °C and 50 barg varying the flow rate from 2 to 10 ml min⁻¹. The reaction time was 60 min. The results for the observed reaction rate (r_obs) is shown in Fig. 8.

Fig. 8 Effect of the flowrate in the hydrolysis rate and final solid residue at 250 °C.

Chaogang Liu and Charles E. Wyman studied the effect of flow on hydrolysis of hemicellulose from corn stover at 180, 200 and 220 °C and at flow rates of 0, 1 and 10 ml min⁻¹ in a tubular reactor. In this study, they concluded that the solubilization of hemicellulose increased with flow.²⁷

The effect of flow rate in reaction kinetic is related to the mass transfer. It was found that a flow rate of 4 ml min⁻¹ is the best alternative to maximize the reaction rate of hydrolysis. The rate of reaction increased when the flow rate was incremented from 2 ml min⁻¹ to 4 ml min⁻¹. However, the reaction rate decreased when the flow rate was increased at values higher than 4 ml min⁻¹.

The decrease in the r_obs could be observed because of back-mixing due to the excess of velocity and also due to preferential ways in the fixed bed.

4. Conclusions

The production of bio-oils from grape seeds using a hydrothermal medium was studied at temperatures between 250 °C and 340 °C. The hydrolysis process produced LBO yield between 8.1–15.7 wt%, HBO yield between 10.6–16.2 wt% and the solid residue was between 25.6–35.8 wt% referred to the mass initial of grape seeds. The mass balance or the system was ca. 80.2–86.3 wt%.

The Arrhenius parameters determined for kinetics of hemicelluloses and celluloses hydrolysis between (TT) were k₀ = 0.995 g min⁻¹ with an activation energy E_a = 13.8 kJ mol⁻¹.

The largest amount of extractable and hydrolysable compounds was obtained at 340 °C. The HBO obtained from the solid residue inside the reactor decreased as the temperature was increased. It was probably because it was dissolved by the flowing water (solubility increases with temperature). The total amount of solid residue decreased when temperature was increased, this would be because of lignin degradation.

TGA analysis showed that the structure of the grape seeds were modified after the treatment. The FT-IR spectra revealed that the main aromatic groups were preserved in the solid residue, while the linkage between hemicellulose and lignin was broken.

In the next works of our research group, the hydrothermal treatment of several types of lignocellulosic biomass with a heat recovery system will be tested. In addition, the combination of semi continuous hydrothermal process with a continuous ultrafast hydrolysis in supercritical water will be analysed.

Acknowledgements

The authors thank the Spanish Economy and Competitiveness Ministry (former Science and Innovation Ministry) Project Reference: CTQ2011-23293 and ENE2012-33613 (FracBioFuel) and Junta de Castilla y León Project Reference: VA254B11-2 for funding. Florencia M. Yedro wish to thank Erasmus Mundus Programme 2012–2015 for the scholarship. The authors also wish to thank Bodega Matarromera S. L. for the raw material. The authors thank Eng. Laura Gutiérrez for technical assistance.

References

1. J. H. Clark, V. Budarin, F. E. I. Deswarte, J. J. E. Hardy, F. M. Kerton, A. J. Hunt, R. Luque, D. J. Macquarrie, K. Milkowski, A. Rodriguez, O. Samuel, S. J. Tavener, R. J. White and A. J. Wilson, Green Chem., 2006, 8, 853–860.
2. T. E. P. S. Organisation, THE EUROPEAN BIOECONOMY IN 2030. Delivering Sustainable Growth by addressing the Grand Societal Challenges, http://www.epsoweb.org, 2011.
3. C. B. Field, M. J. Behrenfeld, J. T. Randerson and P. Falkowski, Science, 1998, 281, 237–240.
4. J. J. Bozell, Clean: Soil, Air, Water, 2008, 36, 641–647.
5. S. Cheng and S. Zhu, BioResources, 2009, 4, 456–457.
6. J. V. Galván, J. J. J. Novo, A. G. Cabrera, D. Ariza, J. García-Olmo and R. M. N. Cerrillo, Eur. J. For. Res., 2012, 131, 893–904.
7. G. Gea-Izquierdo, I. Cañellas and G. Montero, Investigacion Agraria: Sistemas y Recursos Forestales, 2006, 15(3), 339–354.
8. B. Acharya, I. Sule and A. Dutta, Biomass Convers. Biorefin., 2012, 2, 349–369.
9. P. Mäki-Arvela, T. Salmi, B. Holmbom, S. Willför and D. Y. Murzin, Chem. Rev., 2011, 111, 5638–5666.
10. A. A. Peterson, F. Vogel, R. P. Lachance, M. Fröling, M. J. Antal Jr and J. W. Tester, Energy Environ. Sci., 2008, 1, 32–65.
11. L. Tock, M. Gassner and F. Maréchal, Biomass Bioenergy, 2010, 34, 1838–1854.
12. L.-P. Xiao, Z.-J. Shi, F. Xu and R.-C. Sun, Bioresour. Technol., 2012, 118, 619–623.
13. Z. Liu, A. Quek, S. Kent Hoekman and R. Balasubramanian, Special Issue of The Second International Symposium on Gasification and its Application, Fuel, 2013, 103, 943–949.
14. A. V. Bridgwater and G. V. C. Peacocke, Renewable Sustainable Energy Rev., 2000, 4, 1–73.
15. B. M. Wagenaar, W. Prins and W. P. M. van Swaaij, Chem. Eng. Sci., 1994, 49, 5109–5126.
16. FAOSTAT, Production of grapes from 2003-2013, Food and Agriculture Organization of the United Nations, http://faostat3.fao.org, 2014.
17. M. K. Akalin, K. Tekin and S. Karagöz, Bioresour. Technol., 2012, 110, 682–687.
18. K. Tekin, S. Karagöz and S. Bektaş, J. Supercrit. Fluids, 2012, 72, 134–139.
19. Y. Huang, S. Kudo, O. Masek, K. Norinaga and J. I. Hayashi, Energy Fuels, 2013, 27, 247–254.
20. U. Hasanah, B. Setiaji, T. Triyono and C. Anwar, J. Pure App. Chem. Res., 2012, 1, 26–32.
21. X. Guo, S. Wang, Q. Wang, Z. Guo and Z. Luo, Chin. J. Chem. Eng., 2011, 19, 116–121.
22. J. A. Capunitan and S. C. Capareda, Fuel, 2013, 112, 60–73.
23. F. M. Gírio, C. Fonseca, F. Carvalheiro, L. C. Duarte, S. Marques and R. Bogel-Łukasik, Bioresour. Technol., 2010, 101, 4775–4800.
24. C. Schacht, C. Zetzl and G. Brunner, J. Supercrit. Fluids, 2008, 46, 299–321.
25. S. P. M. Da Silva, A. R. C. Morais and R. Bogel-Łukasik, Green Chem., 2014, 16, 238–246.
26. J. B. Sluiter, R. O. Ruiz, C. J. Scarlata, A. D. Sluiter and D. W. Templeton, J. Agric. Food Chem., 2010, 58, 9043–9053.
27. C. Liu and C. E. Wyman, Ind. Eng. Chem. Res., 2003, 42, 5409–5416.
28. A. Toledano, L. Serrano and J. Labidi, J. Taiwan Inst. Chem. Eng., 2013, 44, 552–559.
29. A. Toledano, L. Serrano and J. Labidi, Chem. Eng. J., 2012, 193–194, 396–403.
30. K. Wang, H. Yang, X. Yao, F. Xu and R. C. Sun, Bioresour. Technol., 2012, 116, 99–106.
31. D. Moldes, P. P. Gallego, S. Rodriguez Couto and A. Sanroman, Biotechnol. Lett., 2003, 25, 491–495.
32. M. Spanghero, A. Z. M. Salem and P. H. Robinson, Anim. Feed Sci. Technol., 2009, 152, 243–255.
33. N. G. Baydar and M. Akkurt, Turkish Journal of Agriculture and Forestry, 2001, 25, 163–168.
34. B. Hu, K. Wang, L. Wu, S. H. Yu, M. Antonietti and M. M. Titirici, Adv. Mater., 2010, 22, 813–828.
35. M. Sasaki, K. Goto, K. Tajima, T. Adschiri and K. Arai, Green Chem., 2002, 4, 285–287.
36. A. L. Milliren, J. C. Wissinger, V. Gottumukala and C. A. Schall, Fuel, 2013, 108, 277–281.
37. P. Khuwijitjaru, S. Adachi and R. Matsuno, Biosci., Biotechnol., Biochem., 2002, 66, 1723–1726.
38. E. Unur, S. Brutti, S. Panero and B. Scrosati, Microporous Mesoporous Mater., 2013, 174, 25–33.
39. C. He, A. Giannis and J.-Y. Wang, Appl. Energy, 2013, 111, 257–266.
40. S. M. Heilmann, L. R. Jader, M. J. Sadowsky, F. J. Schendel, M. G. von Keitz and K. J. Valentas, Biomass Bioenergy, 2011, 35, 2526–2533.
41. M. Sevilla and A. B. Fuertes, Chem.–Eur. J., 2009, 15, 4195–4203.
42. P. Reddy, P. Lekha, W. Hüppop and C. Kirsch, Structural characterization of pretreated solids from liquid hot water treatment of South African sugarcane bagasse in a flow-through reactor, Valladolid (Spain), 2014.
43. C. Heitner, D. R. Dimmel and J. A. Schmidt, Lignin and lignans. Advances in chemistry., CRC Press, New York, 2010.
44. K. Bilba and A. Ouensanga, J. Anal. Appl. Pyrolysis, 1996, 38, 61–73.
45. P. Manara, A. Zabaniotou, C. Vanderghem and A. Richel, Catal. Today, 2014, 223, 25–34 , http://www.sciencedirect.com/science/article/pii/S0920586113005518.
46. K. Bilba and A. Ouensanga, J. Anal. Appl. Pyrolysis, 1996, 38, 61–73.
47. S. Mateo, J. G. Puentes, S. Sánchez and A. J. Moya, Carbohydr. Polym., 2013, 93, 416–423.
48. Y. Kim, T. Kreke, N. S. Mosier and M. R. Ladisch, Biotechnol. Bioeng., 2014, 111, 254–263.
49. P. Khuwijitjaru, A. Pokpong, K. Klinchongkon and S. Adachi, Int. J. Food Sci. Technol., 2014 DOI:10.1111/ijfs.12524.
50. I. N. Ahmed, S. Sutanto, L. H. Huynh, S. Ismadji and Y. H. Ju, Biochem. Eng. J., 2013, 78, 44–52.
51. M. T. Maloney, T. W. Chapman and A. J. Baker, Biotechnol. Bioeng., 1985, 27, 355–361.
52. R. Aguilar, J. A. Ramírez, G. Garrote and M. Vázquez, J. Food Eng., 2002, 55, 309–318.
53. N. Eken-Saraçoğlu, S. F. Mutlu, G. Dilmaç and H. Çavuşoğlu, Bioresour. Technol., 1998, 65, 29–33 , http://www.sciencedirect.com/science/article/pii/S0960852498000327.
54. N. Bhandari, D. G. Macdonald and N. N. Bakhshi, Biotechnol. Bioeng., 1984, 26, 320–327.

---