3.1. Production of sugars and ethanol from corncob
Table 1 summarizes the chemical composition of the untreated and pretreated corncob samples. In general, the chemical compositions of the untreated corncob samples were similar. This behavior was expected since milling in a knife mill does not imply structural changes to the lignocellulosic matrix. Similar reports are shown in Liu et al. [30]. Regardless of the initial particle size, acid pretreatment effectively removed hemicellulose. In corncobs with an initial size of < 20 mesh, hemicellulose decreased from 34.37–7.87% after pretreatment. The hemicellulose reduction was from 27.63–8.12% in the pretreated sample > 20 mesh. Hemicellulose is an amorphous polysaccharide with glycosidic bonds susceptible to acid attack at mild temperatures (> 120°C), generating pentoses and organic acids in the liquid phase. The ash was also sensitive to acid pretreatment, reducing values below 1.2%. Alkaline metals (such as Na and K cations) and alkaline earth metals (such as Mg and Ca cations) are present in the ash of lignocellulosic biomasses, and they are easily removed in an acidic environment [31]. As a result, there was a proportional increase in the other components in the present study. Klason lignin values increased slightly in the pretreated corn cob from 14.54–24.74% in the > 20 mesh sample and 17.23–20.72%. Based on Schutyser et al. [32] and He et al. [33], using autoclave and acid was probably able to break down the lignin, but it eventually repolymerizes and deposits on the surface of the pretreated biomass. In turn, cellulose contents in pretreated samples increased almost twice compared to cellulose values in untreated samples. Enriching cellulose in biomass to levels such as ~ 60% is attractive for achieving high glucose concentrations in enzymatic hydrolysis.
The application of pretreatment generated solid yields of ~ 50%, which can be explained by the dissolution of hemicellulose and loss of ash in an acidic environment. In turn, the initial size of the untreated cob had little effect on this metric. The solid yield with corncob < 20 mesh was 51.4%, while the solid yield with corncob > 20 mesh was 53.5%. The reduction in particle size is generally favorable for solid-liquid systems with a reaction due to increased surface area. Here, the slight drop in yield suggests pretreatment success but implies a smaller quantity of solids destined to generate sugars and ethanol.
The crystallinity values reinforce the effectiveness of the acid pretreatment. For samples > 20 mesh, the ICr values of untreated and pretreated corncobs were 42.1% and 44.8%, respectively. The untreated sample < 20 mesh presented an ICr of 39.3%, and the pretreated sample < 20 mesh presented an ICr of 45.1%. This behavior can be explained by the proportional increase in cellulose in the corn cob after pretreatment. Costa Filho et al. [4] and Nogueira et al. [11] also recorded a significant increase in corncob crystallinity after acid pretreatments.
Table 1
Solid yield, chemical composition, and crystallinity of untreated and acid pretreated corncob samples
Samples | Solid yield (%) | Cellulose (%) | Hemicellulose (%) | Klason lignin (%) | Extractives (%) | Ash (%) | ICr (%) |
Untreated CC (< 20 mesh) | - | 32.15 ± 2.15 | 34.37 ± 1.13 | 17.23 ± 1.67 | 9.23 ± 0.63 | 2.47 ± 0.1 | 39.3 |
Untreated CC (> 20 mesh) | - | 30.28 ± 0.97 | 27.63 ± 1.06 | 14.54 ± 0.16 | 11.17 ± 0.63 | 3.16 ± 0.1 | 42.1 |
Acid pretreated CC (< 20 mesh) | 51.4 | 63.92 ± 1.56 | 7.87 ± 0.24 | 20.72 ± 1.91 | 11.46 ± 1.91 | 0.48 ± 0.00 | 45.1 |
Acid pretreated CC (> 20 mesh) | 53.5 | 61.96 ± 3.78 | 8.12 ± 0.51 | 24.74 ± 3.40 | 11.00 ± 0.84 | 1.2 ± 0.00 | 44.8 |
Figure 2 shows the sugar release in the enzymatic hydrolysis of untreated and pretreated corncobs under different enzyme dosages. Unlike pretreatment experiments, particle size was important at this stage. For untreated samples, reducing particle size from > 20 mesh to < 20 mesh increased glucose release from 4.44 to 6.40 g/L. The difference between the responses is 44.1%, although it is not statistically significant. Similar behavior was observed for the release of xylose in these samples, increasing from 1.77 to 2.66 g/L. After pretreatment, the corncob became less recalcitrant, and its polysaccharides were more accessible to enzymatic attack, which explains the increased production of glucose and xylose. The increase in xylose release is curious since the pretreated sample has a lower hemicellulose content than the untreated sample. Glucose release from pretreated corncob reached 18.79 g/L in the < 20 mesh sample and 18.90 g/L in the > 20 mesh sample using 15 FPU/g. However, the granulometry was again impactful when lower dosages of cellulases were applied. Using 5 FPU/g (lowest dosage of cellulases), the pretreated sample < 20 mesh released 15.23 g/L, while the sample released only 11.10 g/L.
When cellulases are in excess, as possibly in the 15 FPU/g condition, mass transfer aspects can be left aside in our study. Even because the experiments were conducted at 50°C (the optimum cellulase temperature) and the final incubation was 48 h. However, difficulties linked to using larger particles, such as low surface area and higher viscosity, become more relevant when there is a lack of cellulases in the medium (5 FPU/g). In Lu et al. [34], corn stover samples with average sizes of 286.7 and 75.2 µm showed 16.8% and 18.1% glucose release using 5 FPU/g. Using rice straw as substrate, Kapoor et al. [35] reported identical glucose release (~ 88 g/L) in experiments with a particle size of 5 mm and 5 FPU/g and a particle size of 20 mm and 10 FPU/g.
To maximize the sugar release, pretreated corncob with < 20 mesh was chosen in the following steps. Choosing samples with > 20 mesh would also be interesting since they consumed less energy for comminution, but we adopted another criterion. About cellulases, the dosage of 5 FPU/g was chosen to maintain the glucose produced at the same level, consuming three times less enzyme, an expensive input for second-generation bioethanol technology.
Based on the previous graph (Fig. 2), glucose release was good from pretreated corn cob with < 20 mesh using 5 FPU/g, but we suspected it could be even higher. Higher ethanol titers are required to ensure the economic viability of second-generation bioethanol technology, which requires higher solid loadings. Solid loading of 5% (w/v) is interesting for screening operational conditions; however, the technology must be tested at solid loadings superior to 15% (w/v). The cellulosic conversion in experiments with 5% (w/v) was reasonable (~ 44.3%), but it indicates that more cellulose could be digested. To increase the sugars generated, PEG 1500 was added to mitigate non-productive adsorption in enzymatic hydrolysis and its results are presented in Fig. 3.
Increasing the solid loadings increased the release of glucose and xylose. In experiments without PEG 1500, glucose release increased from 15.23 g/L at 5% (w/v) solids to 23.61 g/L at 10% (w/v) solids. The same goes for the release of xylose, which increased from 3.32 to 4.07 g/L. This behavior is directly associated with raw material availability in the reaction environment. When more solids were added, there was a slight increase in glucose release with 20% (w/v) solids (26.2 g/L) and a substantial increase in glucose production with the gradual addition of solids (10% + 10%, w /v). Although these conditions have the same final amount of solids, gradually adding biomass causes less stress. High amounts of solids increase the viscosity of the environment and reduce free water, compromising the performance of enzymes. The practice of gradually adding biomass in enzymatic hydrolysis was successful in the work of Gao et al. [36], Mukasekuru et al. [37], and Nogueira et al. [38].
The addition of PEG 1500 also played a critical role in enzymatic hydrolysis. Enzymatic hydrolysis with 5% (w/v) solids and PEG 1500 achieved 24.39 g/L glucose and 4.27 g/L xylose, a performance increase of 60% and 29% compared to the same experiment without additive. These results were superior to those obtained at 10% (w/v) solids without PEG 1500. The cellulose conversion in enzymatic hydrolysis with 5% (w/v) solids was 68.7%, a result comparable to other papers that used low dosage in the enzymatic hydrolysis of cob [4, 39, 40]. Using the gradual addition of biomass, the experiment with PEG 1500 achieved a glucose release of 56.73 g/L, the best volumetric production result among the experiments. A disclaimer is necessary when compared to our previous study. Nogueira et al. [11] investigated the effects of PEG on ethanol production from corncobs. In this study, unlike what was presented here, adding PEG is not relevant to boosting enzymatic digestibility. These discrepancies may be linked to enzyme dosage. The paper by Nogueira et al. [11] operated at 20 FPU/g, four times higher than the enzyme dosage used in the present study. The excess enzymes likely masked any effects of non-productive adsorption onto residual lignin and deactivation of cellulase at the air-liquid interface.
The addition of PEG 1500 and changes in solids load were investigated in saccharification and semi-simultaneous fermentation (SSSF), as seen in Fig. 4. Tests with saccharification and simultaneous fermentation (SSF) were also carried out, but ethanol production was negligible. (data not shown). The low enzyme dosage of 5 FPU/g in SSF could not generate sugars to maintain yeast cells; therefore, ethanol production was compromised. This hypothesis is corroborated by HPLC analysis that showed high glucose concentrations at the end of the experiment. As expected, ethanol production increased with the increase in solid loadings. Ethanol values reached 3.00, 6.67, 20.18 and 26.71 g/L without PEG 1500 with 5%, 10%, 20% (w/v) abrupt addition, and 20% (w/v) gradual addition, respectively. PEG 1500 boosted ethanol production in the SSSF, mainly in the condition with 10% (w/v) solids (17.73 g/L and ethanol yield of 48.8%). Conceição Gomes et al. [41] reported the benefits of adding PEG in SSSF experiments. Statistical differences were not observed in experiments with 20% (w/v) solids. Experiments with abrupt addition without and with PEG 1500 obtained 20.18 and 21.65 g/L, respectively. In experiments with gradual addition, PEG 1500 promoted an advantage in average ethanol production (31.64 g/L vs. 26.71 g/L) compared to the additive-free condition. The enzyme dosage of 5 FPU/g is low; therefore, reducing the temperature to 40°C in fermentation step and restricting free water reduced the efficiency of cellulases. Probably, better ethanol results at high solid loadings would be achieved by increasing the cellulase dosage (perhaps increasing it to 7.5 or 10 FPU/g). Still, the present study results are surprising compared to other reports with corncob. According to Paschos et al. [42], 30–40 g/L ethanol titers are already capable of ensuring the economic viability of second-generation bioethanol technology.
3.2. Production and application of lignosulfonate
Figure 5 shows images of the lignosulfonate produced. Lignosulfonate has a light brown hue, typical of lignin, and has broad solubility in water. The lignosulfonate yield was 23.3%, considered low compared to studies on surfactant preparation based on isolated lignin. However, it is worth highlighting that the fermentation residue still contains relevant levels of cellulose (~ 25%), which could still be linked to lignin.
Figure 6 presents the FTIR spectra of lignosulfonate and alkaline lignin. Similarities between the spectra are evident since aromatic structures form both substances. However, some differences must be highlighted. Lignosulfonate reduced the peaks at 1685 and 1150 cm− 1, corresponding to the C = O vibration in non-conjugated moieties [43, 44]. Lignosulfonate presented a peak in the region of 1373 cm− 1, attributed to the vibration of in-plane deformation of phenolic hydroxyl, according to Sameni et al. [45]. These differences may arise from the different processing to which each lignin was exposed. The fermentation residue originated from an acid pretreatment, in which the lignin structure is disrupted and repolymerized, which hindered the subsequent extraction/functionalization of the biopolymer during lignosulfonate synthesis.
The formation of lignosulfonate was confirmed by the presence of the peak located at 1208 cm− 1. According to Wibowo et al. [46], this peak is attributed to the vibration of S = O bonds, which results from the attachment of the sulfomethyl group in the ortho position of phenolic groups. The lack of a peak in lignosulfonate at 573 cm− 1 may also reflect the success of functionalization. Stretching vibrations of aromatic -CH are assigned at wavenumbers in the region of 618 − 542 cm− 1 [47].
The emulsification index values of the different hydrophobic solvent/water systems are shown in Table 2. The height of the emulsified layer was directly affected by the lignosulfonate concentration. Just like any surfactant, lignosulfonate has hydrophobic groups (in this case, aromatic groups) and hydrophilic groups (i.e., sulfonate and hydroxyl groups), and their increase guarantees greater dispersion of hydrophobic solvent in the aqueous phase. Remarkable results were achieved in the n-octanol/water system, in which the E24 value reached up to 33.33% using 1 g/L lignosulfonate. The positive effects of increasing concentration on the stabilization of n-octanol/water systems are also shown in Fig. 7. Most droplets generated (q = 89%) were smaller than 30 µm using 1 g/L lignosulfonate. Droplets smaller than 30 µm do not correspond to even 30% (q = 30%) of the droplet population in the n-octanol/water system without lignosulfonate. According to Uchida et al. [48], smaller droplets are less susceptible to instability phenomena.
The worst performance of lignosulfonate was observed in the toluene/water system. This result was intriguing since toluene is a planar molecule and π-π interactions between it and lignin were expected to drive stabilization. In Padilha et al. [29], the authors achieved good emulsification rates of around 68% using 1 g/L alkaline lignin nanoparticles as stabilizers.
Comparing our results with other papers involving lignosulfonate was a complicated task. Although lignosulfonates have been used for decades as a dispersant, data in the literature still need to be provided regarding emulsification index experiments. To our knowledge, the only work that investigated the role of lignosulfonate in stabilizing oil/water systems was Ruwoldt et al. [49], but another hydrophobic solvent (xylene) was used. There are other more popular purposes for lignosulfonates, such as cement pastes or biotechnological applications [50–52]. However, this type of experiment is quick and practical to confirm the formation of lignosulfonate. Compared to commercial surfactants, such as sodium dodecyl sulfate (SDS) [53], the performance of lignosulfonate was low and raises the possibility of improving surfactant synthesis in future work. Extraction of alkaline lignin from fermentation residue and oxidation could disrupt aromatic rings and increase functionalization sites.
Table 2
Emulsification indexes of oil/water systems (1:1, v/v) with lignosulfonate (LS) obtained from corncob fermentation residue.
Systems | Emulsification index (E24, %) |
0.2 g/L LS | 0.5 g/L LS | 1.0 g/L LS |
Toluene/Water | 9.17 ± 1.67c | 9.33 ± 1.88c | 18.99 ± 0.34b |
n-octanol/Water | 8.00 ± 0.00c | 28.56 ± 1.62a | 33.33 ± 1.81a |
Soybean oil/Water | 8.94 ± 2.16c | 9.33 ± 1.89c | 30.77 ± 0.00a |
3.3. Alkaline hydrolysis performance
In addition to lignosulfonate, phenolic acids were produced from the fermentation residue and the yield results are presented in Table 3. The production of p-coumaric acid was prioritized over ferulic acid in all conditions investigated. Using only sodium hydroxide in the liquid phase, the yield of p-coumaric acid reached 2770.7 mg/100 g, while ferulic acid reached 496.0 mg/100 g. Both responses increased with the addition of hydrogen peroxide. The p-coumaric acid values were 5178.6 and 8045.3 mg for every 100 g of fermentation residue in alkaline hydrolysis with 2% and 5% (v/v) hydrogen peroxide, respectively. Ferulic acid yields were 891.3 and 1429.4 mg/100 g using 2% and 5% (v/v) hydrogen peroxide, respectively.
Sodium hydroxide breaks ether and ester bonds in biomass, releasing phenolic acids into the liquid phase during alkaline hydrolysis. Hydrogen peroxide releases active oxygen to break C-C bonds in the lignin structure, an additional mechanism for generating phenolic acids [32]. Ferulic acid is linked to lignin and hemicellulose [54], which explains the loss of the phenolic acid in the acid pretreatment step. In a study on the generation of phenolic acids from corn fiber, Valério et al. [55] showed a loss of almost 20% of ferulic acid in the pretreatment stage.
p-Coumaric acid is a constituent of lignin and its high yield in the present study is due to the lignin accumulation in the fermentation residue. High yields of p-coumaric acid are reported in the literature, such as in Mussatto et al. [23] (922 mg/100 g) or Timokhin et al. [22] (4800 mg/100 g); however, these values are still several times lower than the results of the present study. The p-coumaric acid generated is also in high purity, as seen from the HPLC chromatogram (Fig. 8). Given these results, obtaining p-coumaric acid from fermentation residue must be praised. In addition to its antioxidant properties offered via ingestion, p-coumaric acid is used as a precursor for new functional materials, including polymers, adhesives, pharmaceuticals, and nutritional products [22]. It is also noteworthy that p-coumaric acid is estimated at $50/kg, at least ten times more expensive than ethanol [56, 57].
Table 3
Yields of phenolic acids generated from alkaline hydrolysis of solid saccharification residue and semi-simultaneous fermentation. The experiments were carried out with 1% (w/v) solids, a temperature of 80°C, and incubation for 4 h.
Conditions | p-coumaric acid (mg/100g) | Ferulic acid (mg/100 g) |
4% (w/v) Sodium hydroxide | 2,770.7 ± 201.7c | 496.0 ± 28.9c |
4% (w/v) Sodium hydroxide + 2% (w/v) Hydrogen peroxide | 5,178.6 ± 131.9b | 891.3 ± 33.3b |
4% (w/v) Sodium hydroxide + 5% (w/v) Hydrogen peroxide | 8,045.3 ± 442.7a | 1,429.4 ± 64.3a |