Fermentation of saccharified HFW

Composition of food wastes can present a wide variety. Zhang and Richard [27] utilized a food-waste sample from a composting site of a University with a composition of 23.3% w/w total reducing sugars, 34.8% w/w starch and 1.6% w/w fibers and employed mainly amylases for its saccharification. Moon et al.[6] mentioned a high starch (30.1% w/w) and fiber content (14.9% w/w) with total reducing sugars of 17.6% w/w making it necessary to utilize both amylases and cellulases to treat it, whereas a high starch content (63.9% w/w) combined with low cellulose amounts was reported by Yan et al.[12] for the HFW sample that was used in their experiments.

The liquefaction/saccharification process was performed for 8 h at an initial DM content of 45% w/v followed by fermentation at two different initial DM contents (35% and 45% w/v). As can been evidenced from Figure 1, HFW was fully liquified after 8 hours of enzymatic treatment. This fact is also supported by the difference in the viscosity measured at an angular velocity of 10 rad/s before and after the enzymatic treatment, which decreased from 2790 Pa · s to 67.5 Pa · s, respectively. Maximum ethanol production in both DM contents was observed after 15 h of fermentation (Figure 2) and found to be 34.85 g/L (35% w/v DM) and 42.78 g/L (45% w/v DM) with a volumetric productivity of 2.32 g/L · h (35% w/v DM) and 2.85 g/L · h (45% w/v DM) (Table 2). Cellulose hydrolysis at the end of the fermentation reached 50.27% of the initial cellulose content in raw material. Considering this, the obtained yields (Y _p/s ) were 0.443 g/g and 0.423 g/g at 35% and 45% w/v DM respectively. The highest ethanol yield obtained when the fermentation was performed at 35% w/v DM could be attributed to the better mixing conditions.

With no use of the separate liquefaction step the ethanol production reduced by 28.98% and 8.49% when fermentation performed at 35% and 45% w/v initial DM, respectively. The use of the liquefaction/saccharification step was also associated with a significant increase in ethanol volumetric productivities (Table 2), mainly due to the partial cellulose hydrolysis which enabled reduction of viscosity and better mixing conditions of the fermenting raw material [28]. Same enhancement in ethanol production efficiency was also demonstrated by Kim et al.[25], who observed an increase in ethanol yield from 0.31 g/g to 0.43 g/g total solid when applying a Separate Hydrolysis and Fermentation (SHF) process instead of Simultaneous Saccharification and Fermantation (SSF) process on cafeteria food waste. Manzanares et al.[29] also found that with increasing initial DM content, a separate saccharification step improves fermentation of liquid hot water-pretreated olive-pruning biomass. Finally, Hoyer et al.[30] reported that with increasing DM content, even 4 h of saccharification could significantly improve fermentation of softwoods.

Pretreatment and fermentation of the solid residue

At the end of the fermentation there is a remaining solid fraction that was not converted to ethanol. This solid fraction contains unhydrolyzed cellulose which practically is lost from the ethanol production process. In order to increase the overall biofuel yield of the raw material’s mass, these solids could be further utilized. Some research has proposed the utilization of the remaining solids after fermentation and ethanol distillation for the production of biogas on kitchen waste [35], oat straw [36], wheat straw [37, 38] and corn stover [39].

As it is described in the Methods section, after the end of the fermentation at 45% initial DM the solids were removed from the fermentation broth. This solid fraction contains the unhydrolyzed cellulose fraction, which could be further utilized for the production of ethanol in order to increase the overall production yield. The high degree of recalcitrance of this fraction [40] makes a pretreatment process prior to liquefaction/saccharifiacation necessary. During this study, hydrothermal pretreatment with the presence of acetic acid as a catalyst was applied [41, 42]. After the pretreatment, solids were separated from the liquid fraction and washed with distilled water in order to remove the catalyst and other inhibitors formed during pretreatment. Inhibitor removal is necessary in order to decrease the stress to the fermenting microorganism, allowing higher fermentation rates and ethanol production efficiency [43–47].

As has been previously discussed cellulose hydrolysis reached 50.27% of the initial presented cellulose in HFW whereas the cellulose content of residue was 14.75% w/w. During the hydrothermal pretreatment process 42.73% of the initial mass of the residue was solubilized. Cellulose content of the pretreated residue was 16.31%, indicating a 36.68% of cellulose solubilization during the pretreatment.

During fermentation of the pretreated residue at 35% and 45% w/v, maximum ethanol concentrations of 11.44 g/L and 15.92 g/L, respectively, were observed after 15 h (Figure 3). Moreover, increasing substrate concentrations of the fermented residue was associated with an increase in ethanol productivity (Table 4). From the initial cellulosic fraction of the pretreated residue, 42.67% was hydrolyzed. The obtained yields (Y_p/s) reached 0.423 g/g and 0.458 g/g at 35% and 45% DM respectively, which were almost identical with the yields obtained during fermentation of HFW.

Overall ethanol yields

Figure 4 presents the overall obtained ethanol yield after the fermentation of 1 kg of raw material using the two-stage sequential fermentation procedure (concerning the fermentations at 45% DM). In the first stage, 95.07 g of ethanol and 617.2 g of residue were obtained. The remaining residue was hydrothermally pretreated and the solid fraction was 353.5 g, whereas the other solids were dissolved to the liquid fraction. In the second stage after fermentation of the pretreated residue 12.51 g of ethanol could be obtained, thus the ethanol production yield could be increased from 95.07 g/kg DM to 107.58 g/kg DM, corresponding to an increase of 13.16%. The ethanol yield comparing to the maximum theoretical increased from 57.12% to 63.64%.

Raw material

HFW utilized as raw material during this work were source-separated from houses in Papagos-Cholargos Municipality, Athens, Greece. The wastes were dried in situ in a prototype house dryer designed and developed by the Unit of Environmental Science and Technology (UEST), School of Chemical Engineering, NTUA [48]. Dried HFW were milled with a small laboratory mill at an average particle size less than 3 mm. The composition of dried HFW is presented in Table 1.

Reagents and enzyme solutions

All chemicals were of analytical grade. During this work a mixture of the commercial enzyme solutions from Novozymes A/S (Bagsværd, Denmark) Celluclast® 1.5 L (cellulases) and Novozym 188 (β-glucosidase) at a ratio of 5:1 v/v has been applied for the liquefaction/saccharification process. The activity of the mixture was measured according to the standard filter paper assay [49] and found to be 83 FPU/mL.

Analytical methods

Total reducing sugars were measured by the dinitro-3,5-salicilic acid (DNS) method [50]. Monomeric sugars and ethanol were analyzed by HPLC (Shimadzu LC-20 AD, Kyoto, Japan) equipped with a refractive index detector (Shimadzu RID 10A, Kyoto, Japan). Monomeric sugar were analyzed utilizing an Aminex HPX-87P (300 × 7.8 mm, particle size 9 μm, Bio-Rad, Hercules, California) chromatography column, operating at 70°C with HPLC-water as a mobile phase at a flow rate of 0.6 mL/minute. Ethanol was determined by an Aminex HPX-87H (300 × 7.8 mm, particle size 9 μm, Bio-Rad, Hercules, California) chromatography column at 40°C, with a mobile phase of 5 mM sulfuric acid (H₂SO₄) at a flow rate of 0.6 mL/minute.

Soluble fraction was analyzed according to the official method of the National Renewable Energy Laboratory (NREL) [51]. The liquid fraction was further analyzed for sugars and proteins [52]. Moisture was analyzed according to Sluiter et al.[53], whereas crude fat, ash, protein and total starch content determination were conducted according to standard Association of Official Agricultural Chemists (AOAC) methods [54]. Pectin was determined according to Phatak et al. [55]. Finally, the cellulose, hemicellulose and (acid-insoluble) lignin content was determined according to Sluiter et al.[53]. It is worth mentioning that the HFW utilized during this work had no starch content. Analysis was carried out in triplicate. Apparent viscosity of the HFW before and after enzymatic treatment was determined by an Anton Paar Physica MCR rheometer apparatus (Anton Paar GmbH, Ashland, USA) as previously described [56].

Hydrothermal pretreatment of residue

The remaining solids after fermentation of HFW were hydrothermally pretreated by microwave digestion equipment at 200°C for 10 minutes as previously described [41]. After the pretreatment, the solids were removed from the liquid fraction through vacuum filtration and washed in order to remove inhibitors formed during pretreatment. Finally, solids were dried at 60°C until constant weight reached.

Enzymatic liquefaction and saccharification of HFW

Enzymatic liquefaction/saccharification of untreated HFW and hydrothermally pretreated residue was conducted in a liquefaction reactor which was designed and manufactured in-house. More specifically, the reactor consists of two vertical cylindrical chambers which are 6 cm wide and 25 cm diameter, with a rotating shaft driven by a 0.37-kW motor for the mixing of the material. The mixing shaft was programmed to shift from clockwise to anti-clockwise rotation every minute in order to achieve better mixing. Finally, the temperature was controlled by an external oil jacket.

The liquefaction/saccharification process was performed at initial DM concentration of 45% w/v for 8 h. The pH was adjusted to 5.0 by using 50 mM citrate-phosphate buffer and the enzyme load applied was 10 FPU/g DM. Finally, the temperature of the liquefaction/saccharification was set at 50°C. At the end of the liquefaction/saccharification process the whole slurry (also containing unhydrolyzed solids) was utilized for the fermentation experiments.

Ethanol fermentation

Fermentations at 35% and 45% (w/v) DM of non-sterilized liquefied HFW or pretreated residue were performed in 100-mL Erlenmeyer flasks in an orbital shaker at 30°C with an agitation of 100 rpm. The fermenting microorganism was dry baker’s yeast (Yiotis, Athens, Greece), which was added at a concentration corresponding to 15 mg/g of initial DM. To evaluate the importance of the separate liquefaction/saccharification step, untreated HFW were fermented under the same conditions. Samples were taken at certain time intervals, centrifuged and analyzed for ethanol. All trials were carried out in duplicate.

When the HFW fermentation process was completed, the broth was filtrated under vacuum in order to remove the solids which were further washed with distilled water. The solids (residue) were dried at 60°C until constant weight reached and were further utilized for ethanol production after being hydrothermally pretreated (as previously described).