**Enhanced Enzymatic Hydrolysis and Structural Features of Corn Stover by NaOH and Ozone Combined Pretreatment**

#### **Wenhui Wang <sup>1</sup> , Chunyan Zhang <sup>1</sup> , Shisheng Tong <sup>2</sup> , Zhongyi Cui <sup>1</sup> and Ping Liu 1,\* ID**


Received: 19 March 2018; Accepted: 8 May 2018; Published: 29 May 2018

**Abstract:** A two-step pretreatment using NaOH and ozone was performed to improve the enzymatic hydrolysis, compositions and structural characteristics of corn stover. Comparison between the unpretreated and pretreated corn stover was also made to illustrate the mechanism of the combined pretreatment. A pretreatment with 2% (*w*/*w*) NaOH at 80 ◦C for 2 h followed by ozone treatment for 25 min with an initial pH 9 was found to be the optimal procedure and the maximum efficiency (91.73%) of cellulose enzymatic hydrolysis was achieved. Furthermore, microscopic observation of changes in the surface structure of the samples showed that holes were formed and lignin and hemicellulose were partially dissolved and removed. X-ray Diffraction (XRD), Fourier Transform Infrared Spectroscopy (FTIR) and Cross-Polarization Magic Angle Spinning Carbon-13 Nuclear Magnetic Resonance (CP/MAS <sup>13</sup>C-NMR) were also used to characterize the chemical structural changes after the combined pretreatment. The results were as follows: part of the cellulose I structure was destroyed and then reformed into cellulose III, the cellulose crystal indices were also changed; a wider space between the crystal layer was observed; disruption of hydrogen bonds in cellulose and disruption of ester bonds in hemicellulose; cleavage of bonds linkage in lignin-carbohydrate complexes; removal of methoxy in lignin and hemicellulose. As a result, all these changes effectively reduced recalcitrance of corn stover and promoted subsequent enzymatic hydrolysis of cellulose.

**Keywords:** corn stover; alkali; ozone; combined pretreatment; enzymatic hydrolysis; surface morphology; structural characteristics

### **1. Introduction**

In an effort to reduce the energy crisis and the environmental pollution, preparation of recycled lignocellulosic biomass for the use of energy, materials and chemicals has become the focus of today's research. Due to the low degree of lignification, high carbohydrate content and easy absorption of carbohydrate, corn stover has high value in comparison with other lignocellulosic biomass [1]. Pretreatments with physical, chemical and biological methods, however, are necessary to change complex network structure among cellulose, hemicellulose and lignin in corn stover, ascertain pretreatment can release the closure and reduce the strong interchain between lignin and cellulose, compromise the crystalline structure of cellulose, enhance accessibility of enzyme and make corn stover fully utilized [2]. With existing pretreatment methods, sodium hydroxide can rupture the interchain between lignin and other carbohydrates significantly, saponify the inter-molecular ester bonds between hemicellulose and other components, make lignocellulosic swell to remove lignin effectively [3,4]. In addition, this pretreatment not only increases the porosity and internal specific surface area of fibrous materials to ensure effective contact of the enzyme with fibrous materials and degrade it but also

changes the structure of lignocellulose and improves its digestibility of polysaccharide by increasing cellulose conversion rate [5]. Because of strong oxidation, low solubility and selective oxidation of ozone, ozone oxidation technology has some limitations in gas-liquid transfer such as the slow rate, high cost and low ozone utilization rate, which makes it difficult to be used alone [6,7]. Ben'koet et al. used ozone to pretreat aspen wood and found that the efficiency of enzymatic hydrolysis was determined by the absorption rate of ozone [8]. Panneerselvam et al. used different ozone concentrations of 40 mg/L, 50 mg/L, 58 mg/L to treat energy grass [9]. Pretreatment conditions and results showed that ozone treatment can remove lignin effectively without cellulose degradation. Bule et al. used ozone to pretreat wheat stover, the particle size of which was less than 60 mesh, for 2 h and the results showed that the lignin structure was modified significantly and the sugar recovery rate increased from 13.11% to 63.17% in comparison with untreated samples [10]. The previous experiment made corn stover treated in 2% NaOH solution at a temperature of 80 ◦C for 2 h. The specific surface area diameter of corn stover particles was reduced from 189.9 µm to 132.2 µm, and the specific surface area of stover decreased after ozone treatment at pH 5 for 50 min up to 93.11 µm, compared with the specific surface area diameter of the non-alkaline control group decreased by 51%, indicating that alkali combined with ozone made the stover particles smaller by removing lignin. This result was shown in Supplementary Materials.

In this work, a two-step pretreatment using NaOH and ozone was performed on corn stover to improve its enzymatic hydrolysis and changes in compositions and structural characteristics compared to unpretreated. Pretreated corn stover was also analyzed to illustrate the mechanism of the combined pretreatment.

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

#### *2.1. Materials and Sample Preparation*

The corn stover was obtained from the farm research fields at the Jilin Agricultural University (Changchun, Jilin, China). After the corn stover sample was cut into small pieces, it was oven-dried to bring down the moisture content, then milled and screened to particle size of less than 1 mm. The dry sample was kept at −20 ◦C for future use.

#### *2.2. Methods*

#### 2.2.1. NaOH Treatment

Two gram of dry corn stover and 30 mL of 2% (*w*/*v*) NaOH were mixed completely in a 50 mL centrifugal tube reactor, which was then incubated in a water bath for 2 h–4 h at 40 ◦C, 60 ◦C and 80 ◦C. The pretreatment conditions including NaOH treatment temperatures and times, shown in Table 1. When the reaction was over, the tube reactor was cooled to room temperature and filtered via a 300-mesh sieve to separate the mixture into the solid residue and liquid hydrolysate. The solid residue was rinsed with deionized water or saturated carbon dioxide water until it reached neutral pH.



Note: In order to facilitate the description of the structure of typical pre-treated samples, a special nomenclature for NaOH pre-treatment conditions were as follows: A: 2% NaOH at 80 ◦C for 2 h; B: 2% NaOH at 80 ◦C for 3 h; C: 2% NaOH at 60 ◦C for 2 h; D: 2% NaOH at 80 ◦C for 4 h.

#### 2.2.2. Ozone Treatment

Two gram of sample and 30 mL deionized water were placed in a 60 mL of beaker to prepare for ozone pretreatment. 2 mol/L of dilute sulphuric acid was also added to adjust the initial pH of the reaction liquid. Ozone was generated by an ozone generator (CF-10F, Beijing, China). During the reaction, ozone concentration maintained at 78 mg/L for different time with magnetic stirring (85-2, Shanghai, China) at a room temperature. Ozonation experimental conditions are shown in Table 2.


**Table 2.** Level of Ozone Pretreatment factors.

#### 2.2.3. Combined Sodium Hydroxide and Ozone Pretreatment

The experiment is divided into two processes: NaOH Pretreatment and Ozone Pretreatment. The experimental design of the alkali treatment stage is a two-factor three-level, as shown in Table 1. Each group of experiments after alkali treatment was further treated with ozone. The experimental design of the ozone treatment stage is a two-factor two-level, as shown in Table 2. Combined pretreatment samples were prepared for enzymatic hydrolysis and other analysis. The pretreatment conditions, including NaOH and ozone treatment, were optimized for high delignification and high cellulose composition. For convenient description of structure characterization, the special nomenclature for combined pre-treatment conditions were as follows: A-5-25: 2% NaOH at 80 ◦C for 2 h and the ozone initial pH 5 for 25 min. A-9-25: 2% NaOH at 80 ◦C for 2 h and the ozone initial pH 9 for 25 min. A-9-35: 2% NaOH at 80 ◦C for 2 h and the ozone initial pH 9 for 35 min. B-9-25: 2% NaOH at 80 ◦C for 3 h and the ozone initial pH 9 for 25 min. C-9-35: 2% NaOH at 60 ◦C for 2 h and the ozone initial pH 9 for 35 min. D-9-35: 2% NaOH at 80 ◦C for 4 h and the ozone initial pH 9 for 35 min. Blank: untreated degreased stover.

#### 2.2.4. Enzymatic Hydrolysis

Pretreated stover samples, of 0.2 g in 100 mL, were placed in an Erlenmeyer flask and added to 10 mL acetate buffer (0.1 mol/L, pH 4.8), which was prepared with sterile water and contained 40 µg/mL tetracycline, 30 µg/mL cycloheximide and 40 µL xylanase solution. The mixture was incubated in shaking bath (120 rpm) at 70 ◦C for 24 h. After the reaction, the sample was cooled down to room temperature, 40 µL cellulase and 30 µL β-glucosidase were added and it was then incubated at 50 ◦C for 72 h. Cycloheximide could inhibit the DNA translation of eukaryotes to stop cell growth or even cause death. The purpose of adding cycloheximide and tetracycline hydrochloride was to inhibit the growth of microorganism which influenced the pH value during the enzymatic process and affected enzyme activity. Enzymatic hydrolysate was filtered through 0.22 µm membrane and then analyzed by HPLC to determine the glucose content and calculated the cellulase hydrolysis rate.

#### 2.2.5. Determination of the Composition Content of the Corn Stover Samples

In this paper, three components (cellulose, hemicellulose and lignin) in the stover before and after pretreatment were determined by two-step acid hydrolysis method (NREL, 2008b). The content of cellulose and hemicellulose, the lignin removal rate was determined by the following equation:

$$\text{Cellulose content} \left( \% \right) \text{ = } \frac{\text{C}\_1 \times \text{V} \times 0.9}{\text{m}} \times 100 \text{ } \tag{1}$$

$$\text{Hemicellulose content} \left( \% \right) = \frac{\text{C}\_2 \times \text{V} \times 0.88}{\text{m}} \times 100 \,\text{J} \tag{2}$$

In the equation, C<sup>1</sup> was the concentration of glucose measured by HPLC (mg/mL); C<sup>2</sup> was the concentration of xylose measured by HPLC (mg/mL); V was the total volume of the reaction system (87 mL); m was the dry weight of the sample (300 mg); 0.9 was the conversion of glucose to cellulose, 0.88 was the conversion of xylose to hemicellulose.

$$\mathbf{W}\_{1} \left( \% \right) = \frac{\mathbf{m}\_{1} - \mathbf{m}\_{2}}{0.3} \times 100 \,, \tag{3}$$

In the equation, W<sup>1</sup> was the acid-insoluble lignin in the stover; m<sup>1</sup> was the total weight of the sand core funnel and the residue; m<sup>2</sup> was the weight of the sand core funnel; 0.3 was the dry weight of the sample. The unit of measurement was g.

$$\text{W}\_2\left(\%\right) = \frac{\text{OD}\_{320} \times \text{V} \times \text{n}}{300} \times 100\text{\textdegree} \tag{4}$$

In the equation, W<sup>2</sup> was the acid-soluble lignin in the stover; V was the total volume of the reaction system (87 mL); 300 was the dry weight (units: mg) of the sample; OD<sup>320</sup> was the absorbance at 320 nm, 30 L/g·cm; n was the dilution factor.

$$\text{W}\_{\text{3}}\left(\%\right) = \frac{\text{m}\_{\text{4}}}{\text{m}\_{\text{3}}} \times 100,\tag{5}$$

In the equation, W<sup>3</sup> was the lignin removal rate, the total weight of acid-soluble lignin and the acid-insoluble lignin were the total lignin content of the stover; m<sup>3</sup> was the total lignin weight of the untreated stover; m<sup>4</sup> was the total lignin weight of the pretreatment stover.

#### *2.3. Structural Analysis*

#### 2.3.1. Scanning Electron Microscope (SEM) Analysis

Measured stover samples were placed in an oven for 24 h at 50 ◦C to remove moisture, imaged with S-3400n scanning electron microscope with a voltage of 20 kV, current of 30 mA and distance of 11.3 mm. Electron microscopy was amplified at different rates to observe the surface morphology of the sample.

#### 2.3.2. X-ray Diffraction (XRD) Analysis

The samples were examined by X-ray diffractometer with CuKa radiation (λ = 0.154 nm). CuKa radiation was eliminated with nickel. The operation voltage and current was 40 kV and 40 mA respectively. The measurement method was θ/2θ linkages scanning. The range of 2θ was 5◦ to 70◦ . The step was 0.02◦ and the time interval was 0.2 s. The sample was pressed at 40 ◦C and subjected to a 2θ intensity curves. Using Origin and MDI jade 5.0 for data analysis.

#### 2.3.3. Fourier Transform Infrared Spectroscopy (FTIR) Analysis

The samples were placed in an oven at 50 ◦C for 24 h to remove moisture. 10 mg of dry sample was mixed with 200 mg KBr, manually ground in an agate mortar and pressed at 20 MPa for 2 min in oil pressure. The tablets were placed on a sample rack for FTIR spectra spectroscopy and the spectra was recorded between 4000 and 400 cm−<sup>1</sup> . The PerkinElmer Spectrum and Origin software were used for data analysis.

#### 2.3.4. Cross-Polarization Magic Angle Spinning Carbon-13 Nuclear Magnetic Resonance (CP/MAS <sup>13</sup>C-NMR) Analysis

Solid-state cross-polarization magic angle spinning was performed on an Agilent 600 M spectrometer operating. The cellulose-rich solid residue sample was packed tightly into the 4 mm ZrO<sup>2</sup> rotor, 150.81 MHz, spun at 12 kHz at 40 ◦C. The contact time for cross-polarization was set to 1 ms and delayed for 3 s.

### **3. Results and Discussion**

#### *3.1. Enzymatic Hydrolysis and Composition of Pretreated Corn Stover*

The stover was co-pretreated by NaOH and ozone and the three compontent content and cellulose enzymolysis were shown in Figure 1. The initial pH at 9 before ozone treatment, which was more conducive to cellulose enzymatic hydrolysis, than pH 5 and the ozone treatment time that conducive to cellulose enzymatic hydrolysis was 25 min > 35 min. When the stover was treated with 2% NaOH at 80 ◦C for 2 h and ozone treatment condition was the initial pH 5 for 25 min, the maximum enzymatic hydrolysis rate was 86.84%. When the stover was treated with 2% NaOH at 80 ◦C for 2 h and ozone treatment conditions were the initial pH 9 for 25 min, the maximum enzymatic hydrolysis rate was 91.73%. The effect of the three components in the pretreated stover on the enzymatic hydrolysis of cellulose was different due to the pretreatment conditions. The relative content of cellulose in the stover was 62.48%, the removal rate of lignin was 84.35% and the relative content of hemicellulose was 13.74% after the best pretreatment combination. The correlation between hemicellulose content and cellulose enzymatic hydrolysis was significant (*p* = 0.037 < 0.05). The correlation between hemicellulose content and cellulose enzymatic hydrolysis was significant (*p* = 0.037 < 0.05). The significant difference between the cellulose content and cellulose enzymatic hydrolysis rate was found to be *p* = 0.000 (<0.05) which meant their relevance was extremely significant. The significant difference between the lignin removal and the cellulose enzymolysis rate was found to be *p* = 0.017 (<0.05), indicating that the enzymatic hydrolysis of cellulose was significantly affected by lignin removal.

**Figure 1.** Results of enzymatic hydrolysis and content of cellulose, lignin and hemicellulose in corn stover after NaOH-ozone pretreatment. (**A**) Cellulose hydrolysis rate; (**B**) Cellulose content; (**C**) Lignin removal rate; (**D**) Hemicellulise content.

### *3.2. SEM Analysis*

The surface structure of the stover before and after pretreatment is shown in Figure 2. It was found that the surface of untreated (blank) degreased stover was smooth, intact, dense. After the synergistic treatment, change in the surface of the stover was obvious. The density structure was damaged to different degrees, the surface of the stover was fluffy and full of holes, depressions and cracks that increased its specific surface area. In addition, a significant peeling phenomenon appeared on the surface, which indicated that the silica protrusions, waxes and bolts on the outer surface of corn stover were basically cleaned up after synergistic treatment. In A-9-25, we could see fluffy, neat and ordered fiber bundles along the fiber, which indicated that the synergistic treatment could effectively remove ingredients wrapped outside cellulose and break the complex network structure of lignocellulos.

**Figure 2.** Scanning electron microscopy (SEM) images of samples before/after the combined pretreatment. A-5-25: 2% NaOH at 80 ◦C for 2 h and the ozone initial pH 5 for 25 min. A-9-25: 2% NaOH at 80 ◦C for 2 h and the ozone initial pH 9 for 25 min. A-9-35: 2% NaOH at 80 ◦C for 2 h and the ozone initial pH 9 for 35 min. B-9-25: 2% NaOH at 80 ◦C for 3 h and the ozone initial pH 9 for 25 min. C-9-35: 2% NaOH at 60 ◦C for 2 h and the ozone initial pH 9 for 35 min. D-9-35: 2% NaOH at 80 ◦C for 4 h and the ozone initial pH 9 for 35 min. Blank: untreated degreased stover.

In comparison with A-9-25 and B-9-25, we could see that the stover surface of A-9-25 had more pores, less fiber bundles filler and larger gap, the surface mechanical tissue outside stover was exposed, the cell wall was relaxed, the outer wall specific surface area increased, indicating that 2% NaOH was capable to expand the fiber structure than 4% NaOH. As a result, the enzymatic hydrolysis had better penetration into the cellulose and improved the accessibility of the enzyme [11]. Cellulose content showed that A-9-25 (62.48%) < B-9-25 (69.64%). This angle indicated that the factors affected the contact of cellulose and the enzyme, such as the swelling of the fiber material, the impact of the pores on the cellulose hydrolysis rate was greater than the increase of the fiber content in the stover. Studies have also shown

that the enzymatic hydrolysis rate of cellulose and cellulose swelling degree had a linear relationship [12]. In comparison with A-9-25, B-9-25 and A-9-35, it could be roughly concluded that the effect of ozone treatment time on cellulose content and subsequent enzymatic hydrolysis was higher than that of NaOH treatment concentration.

In comparison with A-9-35, C-9-35 fiber bundle surface had a translucent thin layer of material and fluffy scaly structure, the degree of damage was less than A-9-35, indicating that 2% NaOH treatment at 80 ◦C for 2 h compared to 60 ◦C treatment 2 h on the stover surface structure damage was greater and it was consistent with the results of low lignin and lower lignin removal rate and lower cellulose enzyme hydrolysis rate (61.45%) in the C-9-35 stover. This may be attributed to the fact that the 80 ◦C solution allowed the NaOH solution to penetrate better into the cellulose crystallization zone, better weakened the intermolecular or intramolecular hydrogen bonding forces of the cellulose, resulting in better defatting of the degreased stover [13].

The degree of destruction of A-9-35 stover was greater than D-9-35. Combined cellulose content A-9-35 (65.76%) > D-9-35 (58.98%), hemicellulose content A (11.17%) > D-9-35 (9.15%) and lignin removal rate A-9-35 (81.65%) < D-9-35 (88.43%), we could see that pretreatment at a high temperature (80 ◦C) with a long time could remove more lignin and reduce cellulose and it was consistent with the result that the enzymatic hydrolysis rate of A-9-35 was about 18% higher than that of D-9-35.

In comparison with A-5-25 and A-9-25, surface structural damage degree was A-9-25 > A-5-25. It was consistent with the result of enzymatic hydrolysis A-9-25 > A-5-25, cellulose content A-9-25 > A-5-25, lignin removal rate was A-9-25 > A-5-25. It also showed that the pretreatment effect at the initial pH 9 for 25 min was better than that at pH 5 for 25 min.

#### *3.3. XRD Analysis*

Both the crystalline structure and crystal grain index of cellulose played an important role in the enzymolysis efficiency. In order to study the structural changes of stover cellulose after co-treatment, X-ray diffraction analysis of stover before and after pretreatment was showed in Figure 3.

**Figure 3.** X-ray diffraction (XRD) patterns of the samples before/after combined pretreatment. A-5-25: 2% NaOH at 80 ◦C for 2 h and the ozone initial pH 5 for 25 min. A-9-25: 2% NaOH at 80 ◦C for 2 h and the ozone initial pH 9 for 25 min. A-9-35: 2% NaOH at 80 ◦C for 2 h and the ozone initial pH 9 for 35 min. B-9-25: 2% NaOH at 80 ◦C for 3 h and the ozone initial pH 9 for 25 min. C-9-35: 2% NaOH at 60 ◦C for 2 h and the ozone initial pH 9 for 35 min. D-9-35: 2% NaOH at 80 ◦C for 4 h and the ozone initial pH 9 for 35 min. Blank: untreated degreased stover.

The peaks at 2θ of 13–17◦ and 20–23◦ in Figure 3 exhibited more homogeneous polycrystalline of cellulose [14,15]. All tested samples had significant cellulose surface peaks at 2θ of 15.2◦ and 22.1◦ . According to the literature, cellulose I had two crystalline forms, named cellulose Iα and Iβ, different XRD spectra depended on the proportion of these two fiber morphology [16]. After pretreatment, the diffraction peak near 15.2◦ changed obviously. The diffusion peak of untreated stover shifted to the lower position and the peak shape became high and sharp, which indicating that the spacing of the cellulose microcrystals increased and the stacking density decreased. Specifically, the peak of untreated stover and A-5-25 treatment group were close to 15.2◦ (100, Iα) but the peak angle after the treatment of B-9-25 and A-9-25 shifted to 14.7◦ (100, Iα), the peak angle after pretreatment of C-9-35 shifted to 14.1◦ (10-1, Iα) and the peak angle of A-9-35 and D-9-35 were reduced to 13.8◦ (011, Iβ). By comparison, NaOH treatment at 80 ◦C, ozone initial pH 9 for 35 min could reduce the intergranular layer spacing but the ozone initial pH 5 did not have this effect. It also showed that synergistic pretreatment had a great effect on the change of stover crystal grain index. The peak intensity in the 100 crystal plane of A-9-25 was stronger than that of 011 crystals in A-9-35, which may be the key reason for the difference of enzymatic hydrolysis effects.

After pretreatment, the crystal diffraction peak amplitude of 020 near 22.1◦ at 2θ was small and the peak intensity reduced obviously, indicating that the pretreatment did not have a significant effect on the distance of the crystal layer of the crystal grain. The pretreatment group showed weak peaks near 26.7◦ (201, Iα), 27.8◦ (20-1, Iα) and 34.7◦ (004, Iβ), showing the characteristic structure of natural cellulose I. The stover sample after co-pretreatment of NaOH-ozone, 201 and 20-1 crystal faces disappeared. The new diffraction peak (022) formed at 2θ of 29.5◦ and proved the presence of cellulose II, which indicated that the stover sample after pretreatment was a mixed crystal structure of cellulose I and II [17]. The change of crystal structure and grain index promoted the hydrolysis of the cellulose and the enzymatic hydrolysis of cellulose occurred more easily in crystal face 100 and newly formed crystal face 022 [18].

D-9-35 had the highest peak intensity at 2θ of 13.8◦ , mainly due to its high lignin removal rate. Zhao et al. pretreated bagasse with peracetic acid and found that CrI increased because of the removal of lignin [18]. This was consistent with the result of our study that peak intensity of D-9-35 at 2θ = 22.1◦ , was higher than A-9-35, indicating that 2% NaOH treatment was more conducive than 4% NaOH to the subsequent increase in the rate of enzymatic hydrolysis. The peak intensity of A-9-25 at 2θ = 14.7◦ and 29.5◦ was significantly higher than that of B-9-25 and the peak intensity at 2θ = 22.1◦ was significantly lower than that of B-9-25, indicating that 2% NaOH was more conducive to the hydrolysis than 4% NaOH.

In summary, the peak intensity of A-9-25 near 14.7◦ and 29.5 at 2θ were much higher than that of other pretreatment groups, the peak intensity of 22.1◦ at 2θ was the lowest. These results indicated that the microcrystalline structure of stover treated after 2% NaOH at 80 ◦C for 2 h and the initial pH 9 of ozone for 25 min had shifted, which was conducive to enzymatic hydrolysis.

#### *3.4. FTIR Analysis*

The characteristic absorption peaks of cellulose, hemicellulose and lignin in infrared spectrum was shown in Table 3. The FTIR spectra of untreated and pretreated stover were measured in Figure 4. The carbonyl at 1737 cm−<sup>1</sup> was esterified (polyxylose C=O conjugate) and came from the ester bond between the acetyl group attached to xylose and glucuronic acid, the peak was much stronger in spectra of untreated stover than that of the treated. This indicated that co-pretreatment could remove hemicellulose ester linkages [19].



**Figure 4. Fourier transform infrared** (FTIR) spectra of the samples before/after combined pretreatment. A-5-25: 2% NaOH at 80 ◦C for 2 h and the ozone initial pH 5 for 25 min. A-9-25: 2% NaOH at 80 ◦C for 2 h and the ozone initial pH 9 for 25 min. A-9-35: 2% NaOH at 80 ◦C for 2 h and the ozone initial pH 9 for 35 min. B-9-25: 2% NaOH at 80 ◦C for 3 h and the ozone initial pH 9 for 25 min. C-9-35: 2% NaOH at 60 ◦C for 2 h and the ozone initial pH 9 for 35 min. D-9-35: 2% NaOH at 80 ◦C for 4 h and the ozone initial pH 9 for 35 min. Blank: untreated degreased stover.

−1 – –

– – Wavelength 1512 cm−<sup>1</sup> belonged to the stretching of lignin aromatic ring –C=C– aromatic skeleton. The peak was characteristic in lignin indicating G > S [20]. The sharp band almost disappeared in spectra of treated stover but had stronger absorbance in spectra of the untreated sample. The decrease or disappearance of peak intensity could be attributed to the removal of aromatic ring lignin and the destruction of the lignocellular structure in the residue under the corresponding pretreatment conditions [21]. This was consistent with the chemical composition of the sample. It showed that ozone treatment could reduce the content of –C=C– in wheat stover [10].

Due to the breakage of the bond between the lignin-carbohydrate after pretreatment, the peak at 1250 cm−<sup>1</sup> was evident in the control group and was weak in the other groups [16]. The peak at 1320 cm−<sup>1</sup> was much stronger in the spectra of the control group than that of other groups, suggesting that the guaiacyllignin (G) structure of the lignin in the residue was destroyed after pretreatment. Compared to the peak intensities of A-9-25 and A-9-25, it showed that 25 min ozone treatment was more favorable for removing the G structure. According to the literature, the toughness of G structure was higher than that of S structure [22,23], so the destruction of G structure was more conducive to subsequent enzymatic hydrolysis.

#### *3.5. <sup>13</sup>C-NMR Analysis*

It could be seen from the figure that most of the signals in the <sup>13</sup>C-NMR spectrum of the samples after the pretreatment were similar. Compared with the non-pretreated samples, the peaks in the 20–35 ppm region were weakened in the A-5-25 group and almost disappeared in the other groups. The disappearance of the peak in the region or the decrease of the peak intensity showed that stover pretreated only trace amounts of residual lignin. It was also noted that the initial treatment with ozone at pH 9 from the initial ozone at pH 5 removed the lignin more. The result was consistent with the 65.40% removal rate of lignin treated in A-5-25 in chapter 3. Compared A-9-25 and A-9-35, C1–6 signal strength corresponding to the former peak to the latter were significantly stronger and sharper peak shape, indicating that the pretreatment of ozone for 25 min compared to 35 min, the greater degree of damage to the stover, the cellulose could be better separated, the higher relative content of cellulose was more conducive to subsequent enzymatic hydrolysis.

Compared A-9-35 and D-9-35, no new peak appeared and no old peak disappeared, indicating that the type of carbon in the carbohydrate compound did not change when the treatment time of 2% NaOH at 80 ◦C increased from 2 h to 4 h. D-9-35 compared with the non-pretreatment group, the carbon signal peak was sharper, indicating that the treatment of 2% NaOH at 80 ◦C for 4 h with the ozone initial pH 9 for 35 min resulted in the high purity separation of the components in the stover.

The peaks in the 106–153 ppm region were significantly higher in A-9-25 than in B-9-25 and the peaks in B-9-25 almost disappeared, indicating that 2% NaOH at 80 ◦C for 2 h had less lignin removal and lower cellulose relative content than 4% NaOH. Compared with the above conclusions, the effect of NaOH concentration on composition of stover was greater than that of NaOH treatment time. Compared B-9-25 and untreated group, the peak intensity of 4% NaOH treatment was higher and sharper than that of untreated group, suggesting that 4% NaOH at 80 ◦C for 2 h with ozone initial pH 9 for 25 min made stover component separated in high purity. The spectra of A-9-35 and C-9-35 were similar and there was no change in peak number and intensity. For <sup>13</sup>C-NMR, it was impossible to determine the difference between NaOH treatment temperatures 80 ◦C and 60 ◦C.

Comparing the spectra in Figure 5, the peak was obvious at 113 ppm in untreated stover. In A-5-25 and A-9-25, the peak at 106–153 ppm area was obvious, the peak in A-5-25 was mainly at 124 ppm, the peak in A-9-25 shifted to 134.2 ppm. It showed that with the initial pH of co-pretreatment changed from 5 to 9, the aromatic ether bond fragmented and free phenolic hydroxyl group formed. But the peak of disappeared in B-9-25 and the peak was not present in all ozone treatments for 35 min, it presumed that the corresponding aromatic lignin in this area was sensitive to the ozone initial pH 9 for 35 min and 4% NaOH. The untreated stover and A-5-25 treatment group had significant levels of aliphatic hydroxy lignin in the 20–35 ppm area but disappeared after ozone treatment at pH 9, indicating that the corresponding aromatic substances in this region were more sensitive to the ozone initial pH 9.

**Figure 5.** Cross-Polarization Magic Angle Spinning Carbon-13 Nuclear Magnetic Resonance (CP/MAS <sup>13</sup>C-NMR) spectra of the samples before/after combined pretreatment. A-5-25: 2% NaOH at 80 ◦C for 2 h and the ozone initial pH 5 for 25 min. A-9-25: 2% NaOH at 80 ◦C for 2 h and the ozone initial pH 9 for 25 min. A-9-35: 2% NaOH at 80 ◦C for 2 h and the ozone initial pH 9 for 35 min. B-9-25: 2% NaOH at 80 ◦C for 3 h and the ozone initial pH 9 for 25 min. C-9-35: 2% NaOH at 60 ◦C for 2 h and the ozone initial pH 9 for 35 min. D-9-35: 2% NaOH at 80 ◦C for 4 h and the ozone initial pH 9 for 35 min. Blank: untreated degreased stover.

#### **4. Conclusions**

The optimal pretreatment condition was found to be 2% (*w*/*w*) NaOH treated at 80 ◦C for 2 h followed by ozone treatment for 25 min with an initial pH 9 and the maximum efficiency (91.73%) of enzymatic hydrolysis of cellulose was achieved. The promoting effect of three components in corn stover on the cellulose enzymatic hydrolysis was different with pretreatment conditions. Under the optimum pretreatment condition, the relative content of cellulose in the treated stover was 62.48%, the removal rate of lignin was 84.35% and the relative content of hemicellulose was 13.74%.

The results of SEM observation showed that the intensive structure of stover fiber changed to different degrees after synergistic treatment, many pores appeared on the surface and the fiber bundle was exposed. All of these increased the substrate accessibility of enzyme. The XRD characterization of the cellulose crystalline state showed that the synergistic pretreatment could change the crystal structure and crystallographic index, expand the interlayer spacing and so that the crystalline state of cellulose was more conducive to enzymatic hydrolysis. The FTIR characterization of chemical bond properties of stover before and after pretreatment showed that the pretreatment could effectively break the hemicellulose bond, the linkage bond between lignin and other carbohydrate and the intra-/inter-molecular hydrogen bond between the cellulose and other carbohydrate. The CP/MAS <sup>13</sup>C-NMR determination of different positions of carbon in stover showed that pretreatment was beneficial to the removal of acetyl groups in hemicellulose and -OCH<sup>3</sup> in lignin. All changes were conducive to the promotion of cellulose enzymatic hydrolysis. Finally, it emphasized that in this paper, corn stover was a representative of lignocellulose and the pretreatment method used in this article was suitable for other lignocellulosic materials [24].

**Supplementary Materials:** This study optimized conditions of sodium hydroxide synergistic ozone pretreatment. Preconditioning conditions of this experiment were determined based on a large number of previous experiments in our laboratory. The most representative is the analysis of surface area and porosity. The following is available in supplementary materials: http://www.mdpi.com/1420-3049/23/6/1300/s1. Figure S1 showed the specific surface area changes with the increase of alkali treatment temperature. Figure S2: the specific surface area of corn stover changed with the NaOH treatment time increases. Figure S3 was about the particle size changes in the different pH value of ozone treatment in NaOH combined with ozone treatment. Figure S4: the effect of ozone treatment time on the specific surface area of corn stover was carried out. The mechanism of ozonation is revealed clearly by the analysis of specific surface area.

**Author Contributions:** W.W. responsible for conducting experiments, data collection and analysis, manuscript writing and revision; C.Z. responsible for the conduct of pretreatment experiments; S.T. responsible for experimental guidance and data analysis guidance; Z.C. responsible for conducting pre-experiment and literature search; P.L. responsible for the design of experimental ideas and manuscript writing instructions. All authors have read and approved the final manuscript.

**Acknowledgments:** The authors are deeply grateful for the support provided by Aidong Sun at the Beijing Forestry University (Beijing, China), as he offered us the experimental materials.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **Nomenclature**


#### **References**


**Sample Availability:** Samples of the compounds corn stover are available from the authors.

© 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Review* **Efficient Anaerobic Digestion of Microalgae Biomass: Proteins as a Key Macromolecule**

**Jose Antonio Magdalena <sup>1</sup> , Mercedes Ballesteros 1,2 and Cristina González-Fernandez 1,\***

	- joseantonio.magdalena@imdea.org (J.A.M.); mercedes.ballesteros@imdea.org (M.B.)

Academic Editor: Ivet Ferrer

Received: 12 March 2018; Accepted: 3 May 2018; Published: 6 May 2018

**Abstract:** Biogas generation is the least complex technology to transform microalgae biomass into bioenergy. Since hydrolysis has been pointed out as the rate limiting stage of anaerobic digestion, the main challenge for an efficient biogas production is the optimization of cell wall disruption/hydrolysis. Among all tested pretreatments, enzymatic treatments were demonstrated not only very effective in disruption/hydrolysis but they also revealed the impact of microalgae macromolecular composition in the anaerobic process. Although carbohydrates have been traditionally recognized as the polymers responsible for the low microalgae digestibility, protease addition resulted in the highest organic matter solubilization and the highest methane production. However, protein solubilization during the pretreatment can result in anaerobic digestion inhibition due to the release of large amounts of ammonium nitrogen. The possible solutions to overcome these negative effects include the reduction of protein biomass levels by culturing the microalgae in low nitrogen media and the use of ammonia tolerant anaerobic inocula. Overall, this review is intended to evidence the relevance of microalgae proteins in different stages of anaerobic digestion, namely hydrolysis and methanogenesis.

**Keywords:** microalgae; anaerobic digestion; proteins; biogas; inhibition

#### **1. Introduction**

Environmental issues and energy self-sufficiency concerns related to fossil fuels have led to research on new approaches to improve renewable energy production to substitute them. Anaerobic digestion is one of those technologies devoted to the production of biofuels, which involves the degradation of organic matter through the action of different microorganisms. Anaerobic digestion exhibits many advantages such as its efficiency for organic matter removal, its applicability at any scale and the wide variety of substrates that can be used as feedstock. Likewise, the multiproduct generation attained during digestion is also a major benefit of this technology. Those end-products, including biogas and digestate, are easy to separate and can be a source of energy and fertilizers, respectively [1].

Among the different substrates that can be employed, microalgae are being recently studied since this biomass can be grown in residual effluents, do not need arable land to be cultivated while contributing to CO<sup>2</sup> mitigation and wastewater bioremediation [2]. Previous studies have demonstrated the technoeconomic and environmental benefits of microalgae biomass for bioenergy purposes when considered as by-product in other technologies [3–8]. In the same manner, out of the bioenergy producing technologies where microalgae can be used as feedstocks, anaerobic digestion is probably the most economically feasible since it does not require highly concentrated biomass [9] and anaerobes can use proteins, carbohydrates and lipids for methane production purposes [10]. Microalgae biomass has

a wide range of compositions, depending on growth conditions and species [11,12]. In general terms, biochemical profile of chlorophytes range 30–60% of proteins, 20–40% of carbohydrates, and 4–57% of lipids [13,14]. Each macromolecule has different achievable methane yields [10]. Thus, in principle, different microalgae compositions produce different methane yields [12]. At the same time, microalgae composition varies depending not only among strains but also on the growth conditions (nutrients availability and operational conditions) [15,16]. In addition to the different macromolecular composition that microalgae might exhibit, this biomass also differs in structural features. Most of the microalgae able to thrive in wastewater effluents have a chemically complex and structurally robust cell wall composed of low biodegradable substances that hinder the anaerobic digestion [17,18]. Some of these compounds are sporopollenin, algaenan, and polymeric carbohydrates that offer a barrier towards anaerobes [19,20]. During anaerobic digestion, cell walls are degraded by extracellular enzymes of hydrolytic bacteria. Nevertheless, this process might be too slow and thus, a limited hydrolysis rate renders the anaerobic digestion as a lengthy and inefficient bioprocess. Pretreatments are used in order to facilitate the accessibility of these extracellular enzymes whereby improving hydrolysis stage. Different microalgae pretreatments have been studied such as thermal, chemical, mechanical or biological. Methane yields improvements achieved with those different pretreatments can be found elsewhere [21–24]. Out of the different pretreatments, biological approach is the most environmentally friendly [25]. Opposite to other pretreatments, the additional benefits of biological pretreatments are the absence of inhibiting by-products [26] and the high selectivity of the reactions [27]. This approach might not only be used for biomass hydrolysis but also to provide crucial information related to the macromolecule that reduces the anaerobic biodegradability of microalgae biomass. In this manner, this review summarizes the main results attained during the last years of research devoted to microalgae pretreatments in the biogas production context. Moreover, this period of research highlighted the importance of proteins on different stages of the digestion. This review attempts to provide comprehensive evidences of the key role of microalgae proteins.

#### **2. Pretreatment of Microalgae Biomass to Improve Biogas Production**

Since low biodegradability is a common issue in anaerobic digestion of different substrates (such as activated sludge, lignocellulose and photosynthetic microorganisms), a wide range of pretreatments are available to enhance the hydrolysis step [28]. Cell wall rupture or hydrolysis is needed to make available microalgae organic matter to anaerobic microorganisms [29]. Pretreatments are classified in four main groups, namely thermal, mechanical (ultrasound and microwave), chemical (acidic, alkaline, solvents and ozonation) and thermo-chemical (acid or alkali reagents addition combined with high temperatures) and biological (enzymes and microorganisms). Those pretreatments have been intensively studied during the last decade to improve biogas production of microalgae biomass (Table 1). Most of them have been assessed in Biochemical Methane Potential (BMP) assays (batch digestion mode).


#### **Table 1.**Studied pretreatments to improve biogas production using microalgae as substrates.

#### *2.1. High Energy Demanding Pretreatments*

Thermal, thermo-chemical and mechanical pretreatments are considered as high energy demanding processes and, in order to evaluate its efficiency, the final energy balance of the pretreatment process has to be addressed. Given that thermal energy is available in biogas production facilities, the most used pretreatment is thermal application. Thermal pretreatments involve biomass heat up in a wide range of temperatures (50–270 ◦C) and reaction time (from minutes to hours). With regard to thermal application, the effect on the biomass depends on the microalgae strain and applied temperature [30]. Passos et al. [31] and Passos and Ferrer [42] applied thermal pretreatment to *Scenedesmus* sp. biomass at 75 ◦C and 95 ◦C for 10 h resulting in methane yield enhancement of 58% and 69%, respectively . Similar values were attained by González-Fernández et al. [43] when treating *Scenedesmus* at 80 ◦C for only 15 min, highlighting the impact of temperature rather than the heating time as the most relevant parameter in thermal pretreatment. Similar temperatures were tested in *Chlorella* biomass (70 and 90 ◦C) for 30 min resulting in an enhanced methane yield of 37% and 48% compared to the raw biomass (322 mL CH4/g VSin) [32]. These results evidenced that thermal pretreatments are strain specific and thus, at the same temperature applied, different methane yields enhancement can be attained among the different biomass used. Higher temperatures (130 ◦C for 15–30 min) were also tested, resulting in 28% methane yield increase when compared to a raw biomass composed by a mixture of green algae (*Stigeoclonium* sp. and *Monoraphidium* sp.) and diatoms (*Nitzschia*) (105 mL CH4/g VSin) [31]. Due to the potential formation of Maillard compounds at higher temperatures, moderate temperatures in the range of 80–120 ◦C are most widely tested. Moreover, thermal pretreatments have been tested not only in batch mode, but also in semicontinuous mode. Méndez et al. [33] reported a methane yield enhancement of 1.5-fold compared to raw *Chlorella* biomass (84 mL CH4/g CODin) when using 120 ◦C for 40 min for feeding a Completely Stirred Tank Reactor (CSTR). Although no common inhibitors were identified, the results obtained in the CSTR were considerably lower (50% less) than the ones obtained in batch mode digestion. This experimentation corroborated the need to test each pretreatment in different feeding modes. Although thermal pretreatments normally present positive results in terms of methane yield, the values attained are very diverse depending on different variables such as the pretreated biomass, temperature, pretreatment time employed and operation mode during the digestion. Moreover, as mentioned above, these methods involved some drawbacks such as the formation of recalcitrant compounds that could potentially decrease the performance of the process [34,35].

Mechanical pretreatments are commonly employed to disrupt different kind of organic substrates in industrial processes [44,45]. Ultrasound treatment has been applied to disrupt microalgae cell wall in different bioprocess devoted to biofuel production, such as ethanol production from *Chlorella* biomass [46] and biodiesel generation from *Spirulina* biomass [47]. In the case of anaerobic digestion, ultrasound pretreatment has also shown positive results in terms of methane yield enhancement. González-Fernández et al. [43] applied 128.9 kJ/g TS at 85 ◦C and 30 min to enhance methane yield of *Scenedesmus* biomass from 81 mL CH4/g CODin to 153 mL CH4/g CODin (87% enhancement). Nevertheless, those authors also pointed out the fact that ultrasound application is having associated an increase in temperature which also acts as a pretreatment. As a matter of fact, when it comes to the pretreatment of *Scenedesmus* sp., the benefits of ultrasound application were rather questionable compared to the enhancement in methane yield attained only with the application of temperature. Ultrasound pretreatment (26.7 KJ/g TS for 30 min) was also applied to *Monoraphidium* sp. and *Stigeoclonium* sp. biomass and their methane yields were enhanced from 105 mL CH4/g CODin to 196 mL CH4/g CODin [42]. When testing different energy inputs (10; 27; 40; 57 KJ/g TS), applied to different mixtures of microalgae biomass (mixture A: 40% *Chlamydomonas*, 20% *Scenedesmus* and 40% *Nannocloropsis*; mixture B: 58% *Acutodesmus obliquus*, 36% *Oocystis* sp., 1% *Phormidium* and 5% *Nitzschia* sp; Mixture C: *Microspora* ≈ 100%), an increase in methane yield ranging from 6 to 24% at 10 MJ/kg TS was determined, while higher energy inputs did not report any significant increase [34]. Despite all those positive results in terms of methane yields enhancement, the main limitation of ultrasound

pretreatment is the high energy input required when compared to thermal, chemical or biological methods [21].

Chemical methods are often combined with heat pretreatment. Thermochemical pretreatments have been less employed than thermal and mechanical pretreatments due to its potential toxicity for the anaerobes. Cell wall disruption with alkali and acid pretreatments has been tested with positive results for the production of ethanol, butanol and biomethane when using microalgae biomass as a feedstock [48,49]. Studies related to microalgae biomass solubilization using thermo-alkaline methods include for instance the use of reagents such as NaOH or CaO. Different doses of CaO (4 and 10% *w*/*w*) and different temperatures (25, 55 and 72 ◦C) resulted in maximum proteins and carbohydrates solubilization of 32.4% and 31.4%, respectively, and methane yield enhancement of 25% compared to the raw biomass (260 mL CH4/g VSin) at the highest temperature and lime dose tested (72 ◦C and 10% *w*/*w*) [50]. When using NaOH (0.5, 2 and 5% *v*/*v*) in *Chlorella* and *Scenedesmus* biomass, the conducted experiments revealed that despite of the biomass solubilisation, the methane yield enhancement was really low (10%, [36]). Thermo-acid pretreatments have been less employed than thermo-alkali. For instance, *Chlorella* biomass was heated at 120 ◦C either for 20 min and 40 min. Sulphuric acid addition combined with 120 ◦C for 40 min enhanced carbohydrates solubilization by 7-fold, although no solubilization of the protein fraction was reported. In terms of methane production, this thermo-acid pretreatment improved the methane yield from the untreated biomass from 139 mL CH<sup>4</sup> g/CODin to 230 mL CH<sup>4</sup> g/CODin [51]. Since anaerobic digestion is taking place at around pH 7, one of the main limitations of chemical pretreatments is the need to readjust the pH previously to the digestion. In this manner, chemical costs limit the use of these pretreatments. Moreover, some of the chemicals need to be removed previously to the anaerobic digestion as they can be toxic for anaerobes [27].

In conclusion, high energy demanding pretreatment methods report high values in terms of methane yield. However, they are energetically unbalanced. This means that the energy required to carry out the pretreatment is higher than the one obtained in form of biogas. This is why research has been directed towards the use of low energy demanding pretreatments

#### *2.2. Low Energy Demanding Pretreatments*

Compared to other pretreatments, the biological approach presents some advantages such as lower energy demand and high specificity [37]. These pretreatments include the use of suitable enzymes or microorganisms to hydrolyze microalgae biomass. Information about the cell wall composition is scarce, but necessary in order to select the most suitable enzyme for the pretreatment. For that reason, a wide range of biocatalysts have been tested. In principle, given the similarities between higher plants and microalgae, the most studied catalysts are carbohydrases. Among them, cellulases, hemicellulases, amylases and pectinases are the most tested ones [37,52]. Some other enzymatic cocktails employed for microalgae biomass hydrolysis include lysozyme (catalyzing the hydrolysis of 1,4-beta-linkages between *N*-acetylmuramic acid and *N*-acetyl-D-glucosamine residues in peptidoglycan [53]), proteases (hydrolyzing peptide bonds [39]) and laccases [25]. Overall, the best results have been evidenced by adding commercial proteases cocktails. For instance, carbohydrases and proteases were compared hydrolyzing *Chlamydomonas reinhardtii* and *Chlorella vulgaris* [38]. Enzyme doses applied for carbohydrases and proteases were 0.3 mL/g DW and 0.2 mL/g DW, respectively. The enzymatic pretreatment lasted for 5 h and results obtained after carbohydrases addition were 86% and 96% carbohydrate solubilization for *C. vulgaris* and *C. reindhartii* while in the case of protease addition both biomass resulted in 96% protein solubilization. However, the authors pointed out that despite of the high carbohydrate solubilization, only a 14% enhancement methane yield was observed in *Chlorella* biomass, whereas no improvement was observed in *Chlamydomonas*. In the case of protease pretreated biomass, methane yield was enhanced by 51% in the *C. vulgaris* and 7% for *C. reindhartii*. The reason for the low methane yield enhancement recorded for *C. reindhartii* was due to the inherent high anaerobic biodegradability of this strain (75%, 263 mL CH<sup>4</sup> g/CODin). Methane yield is limited

by the inherent methane yield that the biomass can attain. However, the kinetics might be enhanced by the use of pretreatments. More specifically, methane yield might be enhanced by protease pretreatment in the range of 1.07 to 6.3 fold depending on the targeted microalgae biomass within 10–15 days of digestion [38,40].

An alternative to improve economically the enzymatic pretreatment and avoid the addition of high cost cocktails is the addition of hydrolytic secretomes released by other microorganisms. For instance, 0.7 g/L of cellulase-secreting bacteria was added to *Chlorella vulgaris* for 48 h resulting in an increase of 18% organic matter solubilization and 2-fold methane yield compared to the raw biomass [54]. Non-specific extracellular enzymes of *Anthracophyllum discolor* were employed to disrupt the cell wall of *Botryococcus braunii*, resulting in an improvement of 60% methane yield, when enzymatic concentration of 1000 U/mL was applied [55]. Likewise, cellulolytic marine bacteria were applied to *Botryococcus braunii* and *Nannochloropsis gaditana* biomass 1:1 ratio DW resulting in a methane enhancement of 140% and 150%, respectively compared to the raw biomass [56].

As it is observed in Table 1, almost all tested pretreatments improved methane production yields although a direct linkage between solubilization and methane enhancement still requires in-depth research in continuous systems to determine the energy balance and costs of the overall process [57]. Even though this pretreatment is economically unfeasible yet, enzymatic pretreatments, targeting at specific molecules, could provide important information in order to identify which is the microalgae macromolecule hampering biogas production when using this biomass [23].

#### **3. Biological Approach to Enhance Biogas Production: Enzymatic Pretreatment**

Opposite to other pretreatments, biological reactions show high selectivity and absence of inhibitory compounds. Biocatalysts do not only disrupt the cell wall, but they also hydrolyze the macromolecules during biological pretreatment. As it was indicated above, these methods are energetically competitive since they require soft temperatures and smooth shaking. Different parameters must be taken into account such as pH, temperature, enzyme dose, and exposure time [21]. Given the different macromolecular composition, structural features and cell wall composition among microalgae strains, a wide range of biocatalysts can be used. Despite of the high economic cost of the enzymatic cocktails, the use of biocatalysts can provide crucial information to identify the macromolecule hampering anaerobic digestion of microalgae biomass. Moreover, the costs could be reduced either by in situ enzymes production [54,58] or by particular enzymes secreted by bacteria and fungi via sludge bioaugmentation [23,59,60].

#### *3.1. Carbohydrases*

Carbohydrases are in charge of hydrolysing carbohydrates polymers present within the cell wall and inside the cells into simple sugars. Since it is believed that carbohydrates are the responsible of cell wall toughness, cellulaseshave been tested in microalgae biomass to enhance the hydrolysis. Cellulases from *Trichoderma reseei* were mixed with metal oxides to treat *Chlorella* biomass resulting in glucose yield of 91% of the theoretical maximum [61]. Furthermore, enzymatic cocktails aimed at degrading the compartmentalized cell material such as amylases and amyloglucosidases have been tested to promote the efficiency of the hydrolysis step. As a matter of fact, a combination of amylases and cellulases was tested to degrade the cell wall and the cell material with acid hydrolysis in *Chlorella sorokiniana, Nannochloropsis gaditana,* and *Scenedesmus.* This treatment produced a sugar release of 128 mg/g DW, 129 mg/g DW and 60 mg/g DW, respectively against control values for the different biomass (70 mg/g DW, 20 mg/g DW and 25 mg/g DW) [62]. Carbohydrases have also been tested to facilitate lipid extraction by using exoglucanase, endoglucanase, xylanase and laccase produced by different biomass-degrading bacteria, improving lipid extraction up to 40% [63]. All those studies are mainly focused on carbohydrates solubilisation while, only recently, the biomass subjected to carbohydrases has been investigated for biogas production purposes. Ometto et al. [9] tested different enzymatic cocktails on three different biomass, namely *Scenedesmus obliquus, Chlorella sorokiniana* and *Arthrospira*

*maxima* [5]. Out of the tested enzymatic cocktails, mixtures of cellulase plus pectinase and esterase plus protease were the most effective catalysts for organic matter hydrolysis of all three biomass. In the same manner, commercial cocktails hydrolyzing the carbohydrate fraction such as Viscozyme, Celluclast and Pectinase (from Novozymes, Bagsværd, Denmark) have been employed in *C. vulgaris* and *Scenedesmus*. The use of Viscozyme provided carbohydrate fraction solubilization of 84% and 36% for *C. vulgaris* and *Scenedesmus* respectively, while the methane yield enhancement was 1.2-fold for both of them, despite of the different biomass composition and strain [41]. This experimentation suggested that the carbohydrate fraction cannot be understood as a whole to elucidate the relation between solubilization efficiency and the methane yield achievable. Instead of this, an in-depth research must be done concerning the carbohydrates composition of microalgae cell wall.

#### *3.2. Lipases*

When compared to other macromolecular constituents, lipids could be very useful substrates for anaerobic digestion due to its high potential methane yield. More specifically, theoretical methane yield for lipids is 1.014 L CH4/g VS compared to 0.496 and 0.415 L CH4/g VS for proteins and carbohydrates, respectively [10]. However, instability of the system can easily occur due to the formation of long chain fatty acids when lipids are hydrolyzed [64]. As a matter of fact, studies are mainly focused on the optimal concentration of lipids that makes possible to carry out anaerobic digestion without inhibition. In this way, it has been highlighted that lipid fraction should not be over 30% to avoid process inhibition [65]. To overcome such an inhibition, different strategies have been developed. For instance, Palatsi et al. [66] tested different recovery strategies to reduce the negative effect of long chain fatty acids by using different feeding patterns and adsorbents addition. Despite of the high lipid potential to enhance methane yield, microalgae biomass grown in wastewater does not present high lipid content [67,68]. At this point, it should be stressed that microalgae grown in residual effluents is the only feasible way to produce biofuel using this feedstock. In this manner, really limited information on lipases treatment of microalgae biomass for biogas production can be found in literature. For instance, an enzymatic mixture containing protease, α-amylase, xylanase, lipase and cellulase employed for *Rhizoclonium* biomass (filamentous green algae) hydrolysis resulted in 40% yield enhancement [69]. In this case, the mixture of enzymes made difficult the identification of the enzymatic activity responsible for such an enhancement. Ometto et al. [9] also tested esterases in different lipid rich microalgae biomass. Moreover, this investigation reported the use of esterases alone and the mixture of esterases and proteases. No biogas production was attempted for the biomass pretreated with esterases alone and thus, no conclusion could be withdrawn. Nevertheless, their work revealed that this later enzymatic mixture supported much higher organic matter solubilization than the values attained for esterases application alone, highlighting the importance of microalgae proteins.

#### *3.3. Proteases*

Microalgae biomass is normally prevailing in protein content. As a matter of fact, this polymer might represent approximately 40–60% of the microalgae dry weight [24,70]. Protein fraction might be degraded by proteases since they hydrolyze peptides into amino acids. The use of proteases is receiving particular interest in last years, especially in combination with other pretreatments or other commercial enzymatic cocktails [71,72]. Some examples on the use of proteases in different microalgae biomass were evaluated in terms of organic matter solubilization and methane yields [38–40]. In the context of anaerobic digestion, methane yields of *C. vulgaris* and *Scenedesmus* sp. were enhanced by 2.6-fold and 1.53-fold, respectively, when pretreated with protease [39]. It is important to note that those results were attained with proteins rich biomass. More specifically, *Chlorella vulgaris* exhibited 64% protein and 22% carbohydrate content. When dealing with carbohydrate rich *C. vulgaris* biomass (39.6%), protease hydrolysis efficiency (54%) displayed higher organic matter values than carbohydrolase hydrolysis (approx. 26%). The different effect of both enzymatic cocktails was also observed in the methane yields attained by both pretreated biomass. In that case, methane yield achieved with the biomass pretreated with proteases was 137 mL CH<sup>4</sup> g/CODin while 65 mL CH<sup>4</sup> g/CODin was obtained for the biomass pretreated with carbohydrases [40]. This fact showed that even working with carbohydrate rich *C. vulgaris*, the proteolytic cocktail supported high organic matter hydrolysis and methane yields.

Comparison of different studies regarding enzymatic pretreatments suggested that proteins are the molecules that hindered the access of anaerobic bacteria to microalgae organic matter in the anaerobic digestion process to a greater extent than carbohydrates or lipids. Therefore, the protein fraction has been carefully analyzed during the anaerobic digestion process of microalgae biomass in the subsequent section

#### **4. Biomass Proteins in Anaerobic Digestion of Microalgae**

Anaerobic digestion is divided in four different stages including hydrolysis, acidogenesis, acetogenesis and methanogenesis (Figure 1). When protein rich microalgae are subjected to anaerobic digestion, the bioprocess can be affected at different stages. Molecules 2018, 23, x FOR PEER REVIEW 9 of 16

Figure 1. Reactive scheme for the anaerobic digestion of polymeric microalgal biomass. **Figure 1.** Reactive scheme for the anaerobic digestion of polymeric microalgal biomass.

Anaerobic degradation of proteins and lipids has not been investigated in depth compared to that of carbohydrates. Proteins are hydrolyzed to aminoacids by extracellular enzymes. Anaerobic and facultatively anaerobic bacteria, mainly Clostridium, are responsible of aminoacids fermentation. Clostridia obtain energy by coupled oxidation-reduction reaction between aminoacids via the socalled Stickland reaction. This reaction entails the oxidation (dehydrogetation) of one aminoacid and Anaerobic degradation of proteins and lipids has not been investigated in depth compared to that of carbohydrates. Proteins are hydrolyzed to aminoacids by extracellular enzymes. Anaerobic and facultatively anaerobic bacteria, mainly *Clostridium*, are responsible of aminoacids fermentation. Clostridia obtain energy by coupled oxidation-reduction reaction between aminoacids via the so-called Stickland reaction. This reaction entails the oxidation (dehydrogetation) of one aminoacid and the reduction of a second aminoacids (hydrogenation) (Figure 2).

Figure 2. Stickland reactions scheme.

carboxylic acid with one carbon shorter than the original acid (e.g alanine to acetate) while when acting and electron acceptor, it retains the carbon to form a carboxylic acid with the same chain length

Aminoacids can act as electron acceptors or donors. In the first case, the aminoacid form a

the reduction of a second aminoacids (hydrogenation) (Figure 2).

Figure 1. Reactive scheme for the anaerobic digestion of polymeric microalgal biomass.

Anaerobic degradation of proteins and lipids has not been investigated in depth compared to that of carbohydrates. Proteins are hydrolyzed to aminoacids by extracellular enzymes. Anaerobic and facultatively anaerobic bacteria, mainly Clostridium, are responsible of aminoacids fermentation. Clostridia obtain energy by coupled oxidation-reduction reaction between aminoacids via the so-

Molecules 2018, 23, x FOR PEER REVIEW 9 of 16

Figure 2. Stickland reactions scheme. **Figure 2.** Stickland reactions scheme.

Aminoacids can act as electron acceptors or donors. In the first case, the aminoacid form a carboxylic acid with one carbon shorter than the original acid (e.g alanine to acetate) while when acting and electron acceptor, it retains the carbon to form a carboxylic acid with the same chain length Aminoacids can act as electron acceptors or donors. In the first case, the aminoacid form a carboxylic acid with one carbon shorter than the original acid (e.g alanine to acetate) while when acting and electron acceptor, it retains the carbon to form a carboxylic acid with the same chain length as the original aminoacid (e.g., glycine to acetate). The aminoacid is de-ammonified by anaerobic oxidation, yielding volatile fatty acids and hydrogen, as shown in Table 2 [73].


**Table 2.** Aminoacid products based on Stickland reaction (modified from [73]).

#### *4.1. The Relevance of Microalgae Proteins in the Hydrolysis Stage of Anaerobic Digestion*

The first biological process involved in anaerobic digestion is hydrolysis, which is the limiting step and its effectiveness is crucial for the overall process [9,74]. Focusing on proteins, they are hydrolyzed into amino acids by extracellular enzymes secreted by different bacteria such as *Clostridium*, *Vibrio*, *Peptococcus*, *Bacillus*, *Proteus*, or *Bacteroides* [23]. As reviewed above, research devoted to microalgae digestion conducted over last years showed higher methane production in protease pretreated biomass compared to raw biomass and biomass treated with carbohydrases [40]. Methane production of protease pretreated *C. vulgaris* was enhanced by 51% compared to the raw biomass, showing the benefits of having proteins in the soluble phase. Similarly, methane yield enhancement (37%) of cyanobacteria was also attributed to the proteolytic activity developed upon biomass storage [74]. Even though protease addition has revealed the importance of microalgae proteins in microalgae digestion, it is clear that the use of commercial cocktails would not make biogas production profitable. In this manner, the use of commercial proteases helped in the identification of the macromolecule opposing more resistance to an optimal anaerobic digestion but cheaper alternatives should be investigated for avoiding the addition of commercial enzymes. Two main strategies can be applied for such a purpose. The first one entails the use of in-situ released enzymes by fungi or bacteria. Through the so-called bioaugmentation, microorganisms can be added to the anaerobic sludge used as degradation consortium. In this manner, once identified the microorganisms producing the enzymatic cocktail required for the targeted microalgae biomass, it can be added to the anaerobic sludge. Obviously, the appropriate microbial species should be carefully selected to be effective, not only for microalgae hydrolysis, but also to be viable and present good activity within the anaerobic microbiome. The potential of bioaugmentation, including the main benefits and limitations, has been recently reviewed [75]. This approach has been applied in more conventional substrates while literature available on bioaugmentation strategies devoted to microalgae anaerobic digestion is scarce. Nevertheless, this strategy was successfully applied to improve methane production of *C. vulgaris* biomass [60]. Those researchers showed an enhanced methane yield (18–38%) after adding *Clostridium thermocellum* at various inoculum ratios to degrade the carbohydrate fraction of microalgae biomass. Likewise, the same bacteria, *C. thermocellum*, was reported to enhance methane yield (18–38%) when degrading *Haematococcus pluvialis*. Therefore, this acidogenic phase bacteria is nowadays considered as a promising biotechnological tool to improve anaerobic digestion of microalgae through bioaugmentation.

The second alternative to increase the hydrolytic activity of anaerobic sludge is the use of metals. The addition of trace metals as micronutrients have been proven to stimulate methane production. The dosing needs to be well balanced to support the desired microbial activity or growth rate above which the trace metals become inhibitory or toxic. These metals are essential in the anaerobic reactions, since most of them are part of the active site of enzymes. The effect on different trace metal on anaerobic digestion can be found elsewhere [76]. Even though the use of trace elements is beneficial in most cases, the response of the system is uncertain due to the complexity of the anaerobic digestion process. It is recommended for substrates which initially have low trace element content. For instance, Kim et al. [77] evaluated the effect of trace elements at different range temperatures highlighting the benefits of using Fe, Co. or Ni for the hydrolysis step due to the increase of COD solubilization and organic acids production.

#### *4.2. The Relevance of Microalgae Proteins in the Methanogenesis Stage of Anaerobic Digestion*

Out of the subsequent stages involved in anaerobic digestion, hydrogen and acetic acid are converted to methane gas and carbon dioxide during methanogenesis. This last stage is performed by archaea. When compared to anaerobic bacteria involved in anaerobic digestion, archaea are more sensitive to toxic compounds and also exhibit lower growth rates. Acidifiers present ten to twenty-fold higher growth rates and five-fold conversion rates than methanogens [1,69]. With regard to their sensibility toward toxic compounds, methanogens exhibit low tolerance against ammonium nitrogen. Depending on digester pH and operation temperature, the ammonium/ammonia equilibrium might shift. This latter component has been claimed to be highly toxic for methanogens. Ammonia diffuses freely through the permeable membrane of methanogens cells causing changes in intracellular pH and resulting in potassium deficiency and/or proton imbalance [78]. Moreover, ammonium can also inhibit enzymes that are involved in methane production [79]. Yenigün and Demirel [80] reported inhibition of the methanogenesis stage at total ammonia nitrogen (TAN) and ammonia concentrations of 1700–1800 mg/L and 150 mg/L, respectively. As a result, the high concentration of TAN (NH<sup>3</sup> and NH<sup>4</sup> + ) can lead to volatile fatty acids accumulation. This last process involves acidification of the anaerobic broth, which in turns inhibits methanogens activity. Therefore, the main drawback of protein rich biomass, such as microalgae, during digestion is the high amount of nitrogen released

in form of ammonium that can inhibit methane formation. In fact, this inhibition has been already evidenced by Mahdy et al. [38] during the digestion of protein rich *Chlorella vulgaris*. Those authors attributed the stepwise methane production decrease to the high nitrogen mineralization (77%) taking place during the digestion of protease pretreated microalgae biomass. In this manner, microalgae proteins are not only limiting the hydrolysis stage of the anaerobic digestion but they might also be detrimental in methanogenesis stage. Similar to the developed strategies to overcome the negative effect of microalgae proteins in hydrolysis, some solutions have been proposed to overcome the issues that proteins might cause in methanogenesis during those last years of research.

To avoid inhibition by ammonium, different strategies can be implemented. One of them entails the use of nitrogen poor media for microalgae cultivation. Due to the low nitrogen availability in the medium, proteins accumulation is restricted while lipids and carbohydrates fractions become more abundant in the grown biomass [81,82]. Biogas production was modified using this method in different studies [80,83]. This strategy can be easily applied by using urban wastewater as culture media, which normally contains considerable lower nitrogen concentrations than synthetic salt media (≈60 vs. 300–600 mg N/L). The benefit of this strategy has been evidenced recently using *Spirulina* biomass for biogas production [12]. Similar results were obtained with *C. vulgaris*, where a higher accumulation of carbohydrates (40%) was observed when microalgae was grown in urban wastewater while only 22% was obtained in biomass grown in synthetic medium. Concomitantly with the increase in carbohydrates, protein biomass content was reduced (from 64 to 33%) and thus, methane production was enhanced [40].

A second approach to avoid ammonium inhibition is through sludge bioaugmentation. This approach consists in introducing or enriching specific anaerobic microorganisms with special features. Thus, anaerobic microorganisms that are tolerant to high NH<sup>4</sup> + concentrations should be used within the anaerobic sludge to accomplish this goal. Although it is generally believed that total ammonia levels above 3 g/L have toxic effect on the methanogens, the resistance of methanogens can be increased by exposing the microorganisms to high nitrogen concentrations [83]. The use ammonia tolerant inocula has been recently demonstrated as an efficient option for digestion of *C. vulgaris* and cattle manure [84]. In this study, the effectiveness of adapted methanogens resulted in a 33% methane yield increase. This approach allowed operating the digester at 3.7–4.2 g NH<sup>4</sup> + -N/L. Tian et al. [85] operated an acclimation experiment in continuous anaerobic reactors fed with substrate rich in the protein fraction such as microalgae and cattle slurry manure. Results showed a stable biomethanization process despite of the high ammonium concentration (10 g NH<sup>4</sup> + -N/L). Authors stressed the changes on the anaerobic population taking place as the responsible feature to handle high ammonium concentration. Even though this biological strategy is very promising, it is necessary to do further research due to the challenges that might arise such as the different behavior that the bioaugmented inocula under different operational conditions imposed in the reactors. Attention must be directed to microorganism's population since they might fail to thrive or be washed out from the reactors.

#### **5. Conclusions**

Anaerobic digestion of microalgae has been presented as a promising alternative for generation of bioenergy. The implementation of this process requires pretreatment of the rigid algae cell wall in order to make available the organic matter to anaerobes. Enzymatic pretreatment with proteases showed the best performance in terms of organic matter solubilization and methane production. This feature already highlighted the importance of proteins in the hydrolysis stage of anaerobic digestion. Solving this problem with protease addition could result in methanogens inhibition mediated by high ammonium concentrations reached during nitrogen mineralization. Two solutions are proposed to overcome potential inhibition, namely the reduction of nitrogen levels of microalgae biomass using a low nitrogen concentration culture media and the use of ammonium tolerant anaerobic inocula. This fact showed that protein embedded in microalgae cell wall might be responsible for their

inherent low biodegradability. Microalgae proteins might be crucial not only in the hydrolytic phase but also during methanogenesis.

**Author Contributions:** J.A.M. and C.G.-F. were responsible for the manuscript preparation. M.B. was responsible for revising the manuscript. The final publication was prepared with contribution from all authors.

**Acknowledgments:** The authors wish to thank the Spanish Ministry of Economy and Competitiveness for the financial support provided through the grants ENE2013-45416-R and RYC-2014-16823 and the Community of Madrid for the support offered in the framework of the project INSPIRA-1 (S2013/ABI-2783).

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Article* **Effective Saccharification of Corn Stover Using Low-Liquid Aqueous Ammonia Pretreatment and Enzymatic Hydrolysis**

#### **Nguyen Phuong Vi Truong <sup>1</sup> and Tae Hyun Kim 2,\* ID**


Received: 27 March 2018; Accepted: 28 April 2018; Published: 1 May 2018

**Abstract:** Low-liquid aqueous ammonia (LLAA) pretreatment using aqueous ammonia was investigated to enhance enzymatic saccharification of corn stover. In this method, ground corn stover was simply contacted with aqueous ammonia mist (ammoniation step), followed by pretreatment at elevated temperature (90–150 ◦C) for an extended period (24–120 h) at different solid/liquid (S/L) ratios (0.29, 0.47 or 0.67), termed a pretreatment step. After that, excess (unreacted) ammonia was removed by evaporation, and the pretreated material was immediately saccharified by an enzyme without a washing step. The effects of key reaction parameters on both glucan digestibility and XMG digestibility were evaluated by analysis of variance (ANOVA). Under the best pretreatment conditions [S/L = 0.47, 0.16 (g NH3)/(g biomass), 90 ◦C, 24 h], LLAA pretreatment enhanced enzymatic digestibility from 23.1% for glucan and 11.3% for XMG (xylan + galactan + mannan) of untreated corn stover to 91.8% for glucan and 72.6% for XMG in pretreated solid.

**Keywords:** aqueous ammonia; alkaline pretreatment; enzymatic digestibility; lignocellulosic biomass; cellulosic sugar

#### **1. Introduction**

Limited supplies of fossil resources, climate change due to carbon dioxide accumulation in the atmosphere, and increased demand for fuels and chemicals have triggered an increase in utilization of diverse renewable feedstock. To implement the production of a wide range of fuels, chemicals, and materials from renewable sources, most green research in recent years has focused on the development of renewable fuels and bio-based chemicals as a substitute for conventional fossil fuels (gasoline and diesel) and petroleum-based chemicals. In particular, cellulosic fuel ethanol, a second-generation biofuel, has the potential to solve several problems, including limited feedstock availability and food competition with fuel, that are currently associated with first-generation biofuels such as fuel ethanol from corn starch or sugarcane [1]. Cellulosic ethanol can be produced from inexpensive and abundant lignocellulosic materials such as woody biomass and herbaceous biomass [2]. Therefore, it is currently believed that cellulosic ethanol can meet a larger proportion of global transportation fuel demand in the near future. Production of ethanol from lignocellulosic biomass is still challenging because of the recalcitrant nature of the latter; for example, lignin is an inhibitor of enzymatic and microbial reactions and has high crystallinity and complex chemical composition [3,4]. Unlike sugar and starch, carbohydrates of lignocellulosic biomass consist of five different sugar units (glucose, xylose, arabinose, galactose, and mannose). To utilize lignocellulosic biomass effectively, production of fuels and chemicals from all sugars is necessary [5].

Currently, most of fuel ethanol is being produced from corn starch or sugarcane in many countries, such as China, Brazil, and the United States. Corn stover includes husk, leaves, and stalk that are left in the field after grain harvest and is a co-product of corn grain production. Therefore, manufacture of fuel ethanol from corn stover may be a reasonable approach to commercialization of the first cellulosic ethanol process at present [1].

Because of the aforementioned difficulties with utilization of lignocellulosic biomass, pretreatment is necessary to disrupt the recalcitrant structure of the plant cell walls, thus enabling easy access to production of fermentable sugar, which is then fermented to produce ethanol [6]. Therefore, study in recent years has been focused on the development of effective pretreatment method intended to make the lignocellulosic sugars available for ethanol conversion. Nonetheless, it is known that most of pretreatment methods involving various acids and alkalis at high temperature typically generate inhibitory products such as phenolic compounds, furfural, 5-hydroxymethylfurfural, and aldehydes. Therefore, some alkaline pretreatments under mild reaction conditions are considered viable pretreatment methods for different types of lignocellulosic biomass such as wood biomass and herbaceous biomass with high lignin content [7]. In a large-scale biomass conversion process involving a pretreatment unit, the chemical and water inputs can be a critical factor for the development of a commercially viable biochemical method. Nevertheless, a washing step is typically required in both acid and alkali pretreatment methods for the removal of the remaining chemical reagents from the chemically treated biomass, and the recovery and reuse of water and chemicals significantly affect the total energy cost of the biomass conversion process.

To reduce the water and chemical inputs into biomass processing, our laboratory previously reported that a pretreatment method using anhydrous ammonia (low-moisture anhydrous ammonia; LMAA) effectively improves the enzyme saccharification yield of agricultural biomass [8–10]. Although the LMAA method has been developed to eliminate the washing step, one of the drawbacks of anhydrous (gaseous) ammonia is that it must be stored and handled under high pressure, which requires specially designed and well-maintained high-pressure equipment and systems during biomass processing.

In our present study, low-liquid aqueous ammonia (LLAA) pretreatment was proposed to solve such problems associated with a process using gaseous ammonia. This pretreatment method consists of ammoniation, pretreatment, and evaporation steps; i.e., corn stover is well contacted with aqueous ammonia mist using nozzle spray and tumbler mixer (Figure 1a) (ammoniation step), followed by pretreatment step at an elevated temperature (up to 150 ◦C) for an extended period (up to 120 h) using a tight-sealed batch reactor (Figure 1b). After that, excess (unreacted) ammonia is removed by evaporation, and the resulting material can be immediately saccharified by a commercial cellulase without a washing step. LLAA pretreatment can be expected to lower the operating cost because it requires low input of liquid (reagents and water). Furthermore, aqueous ammonia is easy to handle, making this method a more industrially adoptable process for an upcoming biomass-processing facility.

**Figure 1.** Tumbler mixer (**a**) and batch-type pretreatment reactor (**b**).

#### **2. Results and Discussion**

*2.1. Effects of Reaction Temperature and Time on the Chemical Composition of Pretreated Corn Stover*

The initial composition of the untreated corn stover is summarized in Table 1.


**Table 1.** Chemical composition of untreated corn stover.

Note: All weight percentages were calculated on the basis of oven-dried biomass weight; <sup>1</sup> AIL: acid-insoluble lignin; <sup>2</sup> ASL: acid-soluble lignin SD: standard deviation (*n* = 3).

The effects of reaction temperature and time were evaluated, and Figure 2 presents the changes in chemical composition at various pretreatment temperatures with extended pretreatment periods. Three pretreatment temperatures (90, 120 and 150 ◦C) were applied during extended pretreatment periods (24–120 h) while we kept other conditions constant [0.16 (g NH3)/(g biomass), S/L = 0.47]. As shown in Figure 2a,b, pretreatment at lower temperatures (90 and 120 ◦C) did not result in significant changes in carbohydrates (glucan and XMG) and lignin (acid-insoluble lignin, AIL and acid-soluble lignin, ASL) even with a prolonged reaction period (up to 120 h). On the other hand, there was a marginal change in both XMG and lignin contents at 150 ◦C (Figure 2c), in particular, after 72–96 h of pretreatment. Pretreatment at a high temperature (150 ◦C) for 120 h increased both AIL and ASL contents to 15.8% and 4.1%, respectively, which represented 1.3% and 2.0% increases as compared to untreated corn stover (Figure 2c). On the contrary, XMG content decreased from 20.0% of untreated corn stover to 18.4% after pretreatment at 150 ◦C for 120 h. Glucan content was maintained well at all three temperatures of pretreatment.

**Figure 2.** Effects of pretreatment temperature and time on the composition of pretreated corn stover. Pretreatment: 0.16 (g NH<sup>3</sup> )/(g biomass), S/L ratio = 0.47, 24–120 h. (**a**) 90 ◦C, (**b**) 120 ◦C, (**c**) 150 ◦C. The data in the figure show mean values.

#### *2.2. The Effect of the S/L Ratio on Chemical Composition of Pretreated Corn Stover*

In the above test, various temperatures (90–150 ◦C) were tested while we kept NH<sup>3</sup> loading at 0.16 (g NH3)/(g biomass) and S/L ratio at 0.47. Because it was found that chemical composition was more affected at 150 ◦C than other temperature (90 ◦C and 120 ◦C), another set of experiments to study the compositional changes during pretreatment was conducted at low NH<sup>3</sup> loading [0.08 (g NH3)/ (g biomass)] at various S/L ratios. When the S/L ratio was varied between 0.29 and 0.47, both XMG and lignin were slightly affected; i.e., as the reaction time increased, XMG content gradually decreased from 20.0% of untreated corn stover to 18.7–19.0% of pretreated corn stover, while both AIL and ASL increased accordingly; in particular, AIL increased from 14.5% of untreated corn stover to 15.8–16.6% of pretreated corn stover (Figure 3a,b). Nevertheless, it was found that the increase in the S/L ratio did not result in a considerable change in glucan content under all the tested conditions. Most significant changes in XMG and AIL occurred in case of the pretreatment at the highest S/L ratio (S/L = 0.67) and reaction time >72 h (Figure 3c). XMG content decreased from 20.0% of untreated corn stover to 2.9% of 120-h pretreated corn stover, whereas AIL increased from 14.5 to 28.0%. ASL content was slightly increased by ammonia pretreatment (from 2.1% of untreated corn stover to 4.0% of pretreated corn stover) as reaction time was extended to 120 h. The reason for the lignin upregulation during pretreatment under the harsh conditions (Figure 3c) was not clear at this stage. This observation was consistent with our previous report about the pretreatment of herbaceous biomass using gaseous ammonia; i.e., pretreated corn stover at 130–150 ◦C showed a considerable change in the composition of treated solids [8]. Nevertheless, it could be hypothesized according to the literature that the pretreatment reaction in the presence of water and the chemical depolymerize the linkages in the lignin–carbohydrate complex; this action results in removal of lignin along with other fiber fragments from cellulose and hemicellulose and, if they are not removed promptly, causes its subsequent repolymerization [11]. XMG is the main component of hemicellulose in herbaceous plants [12] and can easily be degraded during chemical pretreatment at a high temperature with a long reaction period [13,14]. This repolymerized lignin contains residual xylan and other degradation products becoming acid-insoluble complexes that are not hydrolyzed by sulfuric acid during chemical composition analysis following standard laboratory analytical procedure (LAP) of the National Renewable Energy Laboratory (NREL; Golden, CO, USA), thus resulting in increased measured lignin amounts [15–18]. In addition, another study indicates that the degraded hemicellulose/cellulose forms pseudo-lignin [19], which can affect lignin analysis.

**Figure 3.** *Cont.*

**Figure 3.** Effects of the S/L ratio on composition of corn stover. Pretreatment: 150 ◦C, 24–120 h, 0.08 (g NH<sup>3</sup> )/(g biomass) (**a**) S/L ratio = 0.29, (**b**) S/L ratio = 0.47, (**c**) S/L ratio = 0.67. The data in the figure show mean values.

#### *2.3. The Effect of NH<sup>3</sup> Loading on Enzymatic Digestibility of Pretreated Solids*

In the above test (Section 2.2), high S/L ratio (0.67) resulted in significant decomposition of sugar, in particular, XMG during pretreatment, which was not desirable feature for an effective pretreatment for high sugar conversion yield [20,21]. To evaluate the effect of NH<sup>3</sup> loading on enzymatic digestibility, three different NH<sup>3</sup> loads [0.08, 0.16, or 0.24 (g NH3)/(g biomass)] were applied while other conditions were kept constant (S/L = 0.47, reaction temperature 90 ◦C, and reaction time 24 h), and Table 2 summarizes the chemical composition data and enzymatic digestibility (at 72 h of the hydrolysis reaction) of the pretreated corn stover. An interesting trend was observed with increased NH<sup>3</sup> loading: glucan digestibility of the pretreated solid sample increased from 71.6 to 91.8% with NH<sup>3</sup> loading up to 0.16 (g NH3)/(g biomass), then decreased to 84.7% at 0.24 (g NH3)/(g biomass) loading. The XMG digestibility showed a similar trend: it increased from 66.7 to 72.6% when NH<sup>3</sup> loading was increased from 0.08 to 0.16 (g NH3)/(g biomass) and decreased again above that NH<sup>3</sup> loading [66.5% at 0.24 (g NH3)/(g biomass)]. Although it was unclear in the present step, it was assumed that a change in chemical composition may play a role in enzymatic saccharification.


**Table 2.** Effects of ammonia loading on composition and enzymatic digestibility.

Note: Pretreatment: 0.08–0.24 (g NH3)/(g biomass), S/L = 0.47, 24 h, 90 ◦C; Enzymatic hydrolysis: 15 (FPU (filter paper unit) CTec2)/(g glucan) loading, 50 ◦C, 150 rpm, 72 h.

To further evaluate the effect of various S/L ratios on enzymatic saccharification, two different S/L ratios (0.29 and 0.47) were applied. Ammonia loading of 0.16 (g NH3)/(g biomass) was used because it resulted in the highest digestibility (91.8% for glucan and 72.6% for xylan in Table 2). In this set of tests, three temperatures (90, 120 and 150 ◦C) with increased pretreatment time (~120 h) were applied to each S/L ratio (0.29 and 0.47). Figure 4 indicates that pretreatment at 150 ◦C for an extended treatment period (>72 h) resulted in lower glucan digestibility (71–85% at S/L = 0.29, 65–72% at S/L = 0.47) in comparison with the samples treated for 24–48 h (88–90% at S/L = 0.29, 82–84% at S/L = 0.47). It was assumed that higher lignin content (AIL) of pretreated corn stover at the high temperature (150 ◦C) contributed to the reduced enzymatic digestibility, in agreement with results from another study [22]. Owing to the improved enzymatic digestibility (Figure 4), 90 ◦C and 24 h were selected as the best pretreatment conditions for a further experiment (described in the following section); these conditions were assumed to be desirable because milder reaction conditions (90 ◦C and 24 h) are preferred for a reduction in the operating cost in a large-scale biomass conversion process.

**Figure 4.** Glucan digestibility at elevated pretreatment temperature. Pretreatment: 0.16 (g NH<sup>3</sup> )/(g biomass), 90–150 ◦C, 24–120 h, (**a**) S/L ratio = 0.29, (**b**) S/L ratio = 0.47. Enzymatic hydrolysis conditions: 15 (FPU CTec2)/(g glucan) loading, 50 ◦C, 150 rpm, hydrolysis time: 72 h. The data in the figure show mean values (standard deviation < 1.5).

e 370

4 0001

#### *2.4. Analysis of Variance (ANOVA)*

To assess possible correlations of the effects between various reaction parameters and enzymatic digestibility, the single and combined effects of various factors on both glucan digestibility and XMG digestibility were evaluated by ANOVA, and the performance data are shown in Table 3. Among various reaction conditions, only the combined coefficient of "Temp × Time" had a *p* value less than 0.05 (*p* = 0.0233 for glucan digestibility and *p* = 0.0370 for XMG digestibility), implying that this coefficient significantly affects both glucan and XMG digestibility levels simultaneously, while other coefficients did not have a significant effect on enzymatic digestibility or influenced on either glucan or XMG digestibility. Therefore, the pretreatment temperature–time may be considered primary factors that can enhance the pretreatment effectiveness. In addition, the reaction temperature (Temp) seemed to have a significant effect on glucan digestibility (*p* = 0.0182) and showed a clear-cut tendency (close to significance) to affect XMG digestibility (*p* = 0.0511). On the other hand, the coefficient of time (reaction time), NH<sup>3</sup> (ammonia loading), and S/L and combined coefficient of "Time × S/L" and "S/L × NH3" had lower influence on both glucan and XMG digestibility (*p* > 0.05). The combined coefficient of "Temp × S/L" had a *p* value less than 0.05, indicating that this coefficient significantly affects the glucan digestibility, whereas the combined coefficient of "Temp × NH3" and "Time × NH3" had a *p* value less than 0.05, suggesting that there is a significant effect on XMG digestibility.



Note: Pretreatment: 0.08, 0.16 and 0.24 (g NH3)/(g biomass), S/L = 0.29, 0.47, and 0.67, 24–120 h, 90–150 ◦C. Enzymatic hydrolysis: 15 (FPU CTec2)/(g glucan) loading, 50 ◦C, 150 rpm, 72 h. The probability level of 0.05 (*p* = 0.05) was used to test the significance.

As discussed previously, the alkaline treatment such as the use of an ammonia solution can remove lignin and thereby increase the digestibility of biomass [18,23,24]. It was assumed that increasing the ammonia loading caused the breakdown of ester bonds in hemicellulose and lignin polymers at the elevated temperature; this situation consequently can improve the enzymatic hydrolysis of hemicellulose (XMG).

#### *2.5. Residual Ammonia*

Although ammonia can be evaporated and removed due to its high volatility, some of the impregnated ammonia cannot be easily removed and was assumed to affect the saccharification of fibers during enzymatic hydrolysis. The effect of residual ammonia content on enzymatic digestibility was evaluated, but it was assumed that residual ammonia content does not solely affect enzymatic digestibility because the level of residual ammonia content can be strongly influenced by other reaction parameters such as ammonia loading, pretreatment temperature, pretreatment time, the S/L ratio, and the combined effects of these parameters.

An evaluation assay of the effect of residual ammonia content on glucan digestibility was conducted for each reaction parameter. The effect of reaction severity on residual ammonia content was evaluated under various reaction conditions and the *R* <sup>2</sup> values as the predicted probability are summarized in Table 4. Because four different reaction parameters were compared, we categorized each different reaction condition into three different severity levels such as low, medium, and high severities. The higher severity means severe treatment conditions (see the note in Table 4). The *R* 2 values in Table 4 indicate that samples treated at S/L ratios corresponding to low and high severity resulted in a relatively strong correlation between residual ammonia content and glucan digestibility (*R* <sup>2</sup> = 0.3950 and 0.5607, respectively). In addition, samples treated with ammonia loading of medium severity showed *R* <sup>2</sup> = 0.4113, which indicated some correlation between residual ammonia content and glucan digestibility. Overall, the coefficients (*R* 2 ) of the trend lines were 0.027–0.5607; therefore, the model equations of the trend lines were not significant.


**Table 4.** Effects of residual ammonia content on glucan digestibility under various reaction conditions.

Note: Low severity: Time = 24 h, temp. = 90 ◦C, NH<sup>3</sup> loading = 0.08 (g NH3)/(g biomass), S/L = 0.29. Medium severity: time = 48–96 h, temp. = 120 ◦C, NH<sup>3</sup> loading = 0.16 (g NH3)/(g biomass); S/L = 0.47. High severity: time = 120 h, temp. = 150 ◦C, NH<sup>3</sup> loading = 0.24 (g NH3)/(g biomass), S/L = 0.67.

Because it was found that the S/L ratio had the strongest effect (Figures 3 and 4), we next evaluated the effect of increasing S/L. Besides, residual ammonia was significantly affected (*p* < 0.05; data not shown) when S/L was changed. Figure 5 presents the relation of enzymatic digestibility (at 72 h of hydrolysis) of glucan with residual ammonia content at different S/L ratios. The *R* <sup>2</sup> values of three different S/L levels indicated that there was no clear trend between glucan digestibility and residual ammonia content even though the treated samples with high concentrations of residual ammonia seemed to have slightly lower digestibility than did the samples with low residual ammonia content. It should also be noted that the residual ammonia in the pretreated biomass can serve as an essential nitrogen source for microbial cell growth during fermentation if it is at an appropriate concentration.

**Figure 5.** Effects of residual ammonia content on glucan digestibility of pretreated corn stover. Pretreatment: 0.08, 0.16 or 0.24 (g NH<sup>3</sup> )/(g biomass), S/L ratio = 0.29, 0.47 or 0.67, 24–120 h, 90–150 ◦C. Enzymatic hydrolysis: 15 (FPU CTec2)/(g glucan) loading, 50 ◦C, 150 rpm, 72 h.

#### *2.6. Mass Balance*

Figure 6 summarizes the overall mass balance for the process of conversion of 100 g of corn stover to fermentable sugar by pretreatment under the best conditions [S/L = 0.47, 0.16 (g NH3)/(g biomass), 90 ◦C, 24 h]. One of the features of LLAA method is that it does not solubilize any component during pretreatment and only modifies lignin and hemicellulose. Therefore, the input and output of the whole sugar conversion process are almost the same. The residual ammonia and ammonia recovery in this calculation were 1.7 wt % residual ammonia [0.16 (g NH3)/(g biomass)] and 98.3% (15.7 g) after pretreatment and evaporation, respectively. Next, the pretreated solids were saccharified by means of 15 (FPU (filter paper unit) CTec2)/(g glucan) at 50 ◦C, 150 rpm, 72 h, and 1.0% (*w*/*v*) glucan loading. The highest glucan and XMG digestibility at an enzyme load of 15 FPU/(g glucan) was 91.8% and 72.6%, respectively. According to the mass balance in Figure 6, 33.7 g of glucose and 16.5 g of xmg were produced from 100 g of corn stover. The residue after enzymatic saccharification mostly consisted of lignin, ash, and unconverted polysaccharides.

**Figure 6.** Schematic diagram and mass balance during conversion of corn stover to sugars.

### *2.7. Comparison of Various Ammonia Pretreatments*

Alkaline pretreatment is considered an effective way to break down the structure of lignin and therefore to enhance the enzymatic hydrolysis of lignocellulosic biomass [25]. Table 5 shows a comparison of the features and reaction conditions of various alkaline pretreatment methods (in particular, methods involving ammonia). Pretreatment methods shown in Table 5 include low-liquid ammonia recycle percolation (LLARP), soaking in aqueous ammonia (SAA), LMAA, and LLAA [8,10,23,26,27].

H ch


**Table 5.** A comparison of various ammonia pretreatment methods.

Note: <sup>1</sup> Water consumption does not include water for washing after pretreatment. ARP: ammonia recycle percolation, LLARP: low-liquid ammonia recycle percolation, SAA: soaking in aqueous ammonia, LMAA: low-moisture anhydrous ammonia, LLAA: low-liquid aqueous ammonia; <sup>2</sup> enzyme loading; 15 FPU/g-glucan, enzymatic digestibility after 72 h of hydrolysis; <sup>3</sup> enzyme loading; 10 FPU/g-glucan.

Among the methods listed in Table 5, LMAA requires the least amount of chemical loading [0.1 (g NH3)/(g biomass)], and LLAA is the next best method [0.16 (g NH3)/(g biomass)] and shows the same water consumption [<1.0 (g H2O)/(g biomass)]. The sugar production process using these two pretreatment methods can be considered more economical than those based on other pretreatment methods [0.5–0.9 (g catalyst)/(g biomass) and 2.8–10 (g H2O)/(g biomass)]. Furthermore, in contrast to other methods (LLARP and SAA) in Table 5, the most desirable feature of LLAA and LMAA is that the washing step after pretreatment is not necessary; this feature can reduce the water consumption and thus reduce total energy cost in the biomass conversion process. In terms of severity of pretreatment conditions, LLAA, LMAA, and SAA processes involve mild reaction conditions. Although LLARP requires a short reaction period (~10 min), it should be carried out at high temperature (170 ◦C), while the other three ammonia pretreatment methods (LLAA, LMAA, and SAA) require more time (12–24 h) at a moderate temperature (60–90 ◦C). On the other hand, the longer pretreatment time and large water input in the SAA method are required even though it involves a mild reaction temperature; these characteristics are not considered desirable for an economically viable process [28].

#### **3. Materials and Methods**

#### *3.1. Materials*

#### 3.1.1. Feedstock

Corn was grown and harvested in China in September 2015, and corn stover was then collected and provided by CJ Cheiljedang Co. (Seoul, Korea). The received corn stover was air-dried at ambient temperature (~25 ◦C), ground up, passed through a sieve with a mesh size of 10–35 mesh (US Standard, 0.5–2.0 mm of nominal sieve opening) sieves, and then stored in sealed plastic containers at ambient temperature. The initial composition of the biomass was determined by a standard LAP of the NREL (Table 1) [29]. It should be noted that glucan, xylan, and lignin are the main components among the various ones shown in Table 1; therefore, an evaluation of pretreatment effects was focused on those three components in this study.

Ammonium hydroxide (28.0–30.0%; lot number A29260I1) and sulfuric acid (ACS grade, 95–98%, lot number SZBF0140V) were purchased from Daejung Chemical & Metals Co., Ltd. (Shehung-si, Gyeonggi-do, Korea) and Sigma-Aldrich (St. Louis, MO, USA), respectively. Avicel® PH-101 (catalog number 900-3-6, lot number BCBJ029V, Sigma-Aldrich) was acquired and served as a control sample in the enzymatic-digestibility test.

#### 3.1.2. Enzymes

Cellic® CTec2 (batch number: VCP10006, Novozymes Inc., Bagsvaerd, Denmark) was used for enzymatic saccharification of untreated and pretreated corn stover. The average activity of the enzyme, as determined by the LAP of the NREL was 88.91 FPU/mL [30].

#### *3.2. Pretreatment*

#### 3.2.1. The First Step: Ammoniation

To apply ammonia loading at different target concentrations [0.08, 0.16 or 0.24 (g NH3)/ (g biomass)], an ammonium hydroxide (NH4OH) solution at various solid/liquid (S/L) ratios (0.29, 0.47 or 0.67) was added in the form of mist using nozzle spray and tumbler mixer. The S/L ratio was calculated as follows:

$$\text{S/L} = \frac{\text{Total solids (g)}}{\text{Total solids (g)} + \text{water \& moisture (g)}}.$$

The initial moisture content of corn stover was approximately 8.5% and was loaded for ammoniation. After spraying of ammonium hydroxide mist, corn stover (100 g, dry basis) was homogenized at 30 rpm for 1 h in the tumbler mixer shown in Figure 1a.

#### 3.2.2. The Second Step: Pretreatment

Ammoniated corn stover treated with aqueous ammonia (10 g, dry basis) was packed in a smaller sealed batch reactor (30.0-cm length, 2.54-cm internal diameter [ID], and 0.21-cm tube wall thickness [internal volume: 105.7 mL]; Figure 1b). Openings of the sealed batch reactor were tightened carefully enough to prevent ammonia leaking. The reactor was placed in the forced convection oven (model no. OF-22GW, Jeio Tech Co., Ltd., Daejeon, Korea) and then heated from ambient temperature to the target temperatures (90–150 ◦C) in 1 h and maintained at the desired temperature for 24–120 h.

#### 3.2.3. The Third Step: Evaporation

After completion of the pretreatment process, the reactors were cooled down to ambient temperature. The reactors were then opened, and the treated sample was transferred into a tray. The collected sample was placed in the fume hood to remove excess ammonia by evaporation for 1 h at 25 ◦C. One portion of the sample was used for analysis of residual ammonia content, and the other portion was used for composition analysis.

#### *3.3. Analytical Methods*

Soxhlet extraction was applied to determine the water- and ethanol-soluble extractives of untreated corn stover. A two-step Soxhlet extraction was conducted; the first step of extraction with de-ionized (DI) water for 8 h was followed by the second step of extraction with ethanol (190 proof) for 24 h.

The chemical composition of untreated and pretreated corn stover was analyzed for carbohydrates, AIL, ASL (on a UV spectrophotometer at 320 nm), and ash (a gravimetric method involving a muffle furnace at 575 ◦C) following the NREL LAP [29]. Carbohydrate contents were determined by means of a high-performance liquid chromatography (HPLC) system (Shimadzu LC-10A, Shimadzu Inc., Kyoto, Japan) equipped with Bio-Rad Aminex HPX-87P (catalog number 1250098; Bio-Rad Inc., Hercules, CA, USA) and an 87H column (catalog number 1260140; Bio-Rad Inc., Hercules, CA, USA) and a refractive

index detector (model RID-10A, Shimadzu Inc., Kyoto, Japan). Analytical conditions for HPLC were as follows: mobile phase of water (0.6 mL/min) at column temperature of 85 ◦C and 0.005 M H2SO<sup>4</sup> (0.6 mL/min) at 65 ◦C for the HPX-87P column and HPX-87H column, respectively.

#### *3.4. Enzymatic Digestibility*

This property of pretreated and untreated corn stover was evaluated in duplicate in rubber-capped 250-mL Erlenmeyer flasks containing 100 mL of a liquid and 1.0 g of a glucan loading (3.0 g of pretreated solid loading, dry basis) according to the NREL-LAP [30]. The recovered solid samples obtained after the evaporation were used directly in the enzymatic digestibility tests without drying. Reaction conditions for the digestibility test were 50 ◦C, pH 4.8, and 150 rpm at 15 FPU/(g glucan) enzyme load in 0.05 M citrate buffer. Each sample in 100-mL working volume was saccharified in a shaking incubator (model number VS 8480SFN, Vision Scientific Co., Ltd., Daejeon, Korea). Total glucose content after 72 h of hydrolysis was used to calculate the enzymatic digestibility. Avicel® PH-101 was also put through the same digestibility test conditions and served as a control sample. The glucan and XMG digestibility values were calculated as follows:

$$\text{Glucan digestibilityity} = \frac{\text{Total released glucose (g)} \times 0.9}{\text{Initial glucose loading (g)}} \times 100\%$$

where 0.9 is the factor for conversion of glucose to equivalents of glucan.

XMG digestibility = Total released XMG (g) × 0.88 Initial XMG loading (g) × 100,

where 0.88 is the factor for conversion of xylose to equivalents of XMG.

#### *3.5. Residual Ammonia Analysis*

One gram of untreated and pretreated samples was placed in a glass bottle with 80 mL of a 1.0% borate buffer solution. These glass bottles were placed in a convection oven at a stable temperature (80 ◦C) and incubated there for 24 h. After that, the glass bottles with residual ammonia in the liquid were removed from the oven. Liquid and solids were separated by filtration through filter paper (Fisher catalog number F2044-090, size: 90 mm Ø, pack: 100 units from CHmlab Group, Barcelona, Spain). Then, the filtrate was diluted to 100-mL working volume. The liquid, which contained ammonia, was reacted with a 10 N sodium hydroxide (NaOH) solution. Residual ammonia content in the liquid was determined by means of an ammonia analyzer (model Accumet®, XL250, Dual Channel pH/mV/Ion, Thermo Fisher Scientific Inc., Tampa, FL, USA) and an ion-selective electrode (ISE, Fisher catalog number 13-620-509).

#### *3.6. ANOVA*

The statistical analysis of the data was performed using SAS® software (version 9.4, SAS Institute Inc., Cary, NC, USA).

#### **4. Conclusions**

LLAA pretreatment can reduce energy use because it requires lesser inputs of ammonia and water as compared to other pretreatment technologies, and can enable economically viable processes. In addition, the LLAA pretreatment has advantages over previously developed ammonia pretreatment methods, e.g., it uses aqueous ammonia without washing. Therefore, this approach can be regarded as a more economically feasible technology for scaling up. Moreover, LLAA shows promise because of the effectiveness of this pretreatment at enhancing enzymatic digestibility of corn stover. The highest glucan and XMG digestibility levels were 91.8% and 72.6%, respectively, at 15 FPU/(g glucan) enzyme loading.

**Author Contributions:** N.P.V.T., the first author, performed all the experiments and analyzed the data. T.H.K., the corresponding author, designed the overall study and experiments, interpreted the results, and finalized the manuscript. All authors have read and approved the final manuscript.

**Acknowledgments:** This work was supported by the R&D program of Korea Institute of Energy Technology Evaluation and Planning (KETEP) grant funded by the Ministry of Trade, Industry & Energy (MOTIE), Republic of Korea (No. 20153010091990).

**Conflicts of Interest:** The authors declare no conflicts of interest.

#### **References**


**Sample Availability:** Samples of the compounds are not available from the authors.

© 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

*Article*
