Ionic liquid pretreatment of stinging nettle stems and giant miscanthus for bioethanol production

Production of ethanol from lignocellulosic biomass is considered the most promising proposition for developing a sustainable and carbon–neutral energy system. The use of renewable raw materials and variability of lignocellulosic feedstock generating hexose and pentose sugars also brings advantages of the most abundant, sustainable and non-food competitive biomass. Great attention is now paid to agricultural wastes and overgrowing plants as an alternative to fast-growing energetic crops. The presented study explores the use of stinging nettle stems, which have not been treated as a source of bioethanol. Apart from being considered a weed, stinging nettle is used in pharmacy or cosmetics, yet its stems are always a non-edible waste. Therefore, the aim was to evaluate the effectiveness of pretreatment using imidazolium- and ammonium-based ionic liquids, enzymatic hydrolysis, fermentation of stinging nettle stems, and comparison of such a process with giant miscanthus. Raw and ionic liquid-pretreated feedstocks of stinging nettle and miscanthus were subjected to compositional analysis and scanning electron microscopy to determine the pretreatment effect. Next, the same conditions of enzymatic hydrolysis and fermentation were applied to both crops to explore the stinging nettle stems potential in the area of bioethanol production. The study showed that the pretreatment of both stinging nettle and miscanthus with imidazolium acetates allowed for increased availability of the critical lignocellulosic fraction. The use of 1-butyl-3-methylimidazolium acetate in the pretreatment of stinging nettle allowed to obtain very high ethanol concentrations of 7.3 g L−1, with 7.0 g L−1 achieved for miscanthus. Results similar for both plants were obtained for 1-ethyl-3-buthylimidazolium acetate. Moreover, in the case of ammonium ionic liquids, even though they have comparable potential to dissolve cellulose, it was impossible to depolymerize lignocellulose and extract lignin. Furthermore, they did not improve the efficiency of the hydrolysis process, which in turn led to low alcohol concentration. Overall, from the presented results, it can be assumed that the stinging nettle stems are a very promising bioenergy crop.


Scientific Reports
| (2021) 11:18465 | https://doi.org/10.1038/s41598-021-97993-y www.nature.com/scientificreports/ cellulose enzymatic digestibility. During the IL pretreatment (dissolution and regeneration with anti-solvent), the crystalline structure of cellulose can be changed to amorphous, which largely increases the bioethanol production process efficiency [28][29][30] . This work aimed to analyze the effectiveness of ethanol production from stinging nettle stems, an innovative cellulose source considered as agricultural waste. For this purpose, at first, pretreatment of stinging nettle stems was performed using imidazolium and ammonium ionic liquids. Afterward, the stems were subjected to enzymatic hydrolysis and alcoholic fermentation. As it is still challenging to compare the efficacy of bioethanol production from various lignocellulosic sources, the results obtained for stinging nettle stems were compared to the giant miscanthus, a well-known energy crop. The study's originality lies in the demonstration that the stinging nettle stems can be a potential raw material for ethanol production for fuel purposes.

Results
Compositional analysis. The qualitative composition of lignocellulose biomass is a crucial aspect that qualifies the raw material for bioethanol production. Another issue is the choice of appropriate pretreatment method and its costs. The use of ionic liquids has many advantages, such as the possibility of their recirculation and reuse and interesting physicochemical properties, e.g., low vapour pressure, thermal and chemical stability, and wide liquid range. The share of fractions of raw and IL-pretreated materials tested in the study is presented in Table 1.
Non-pretreated giant miscanthus contained on average 43.5% cellulose, 25.7% hemicellulose, and 14.1% lignin, while the stalks of stinging nettle contained 42.5% cellulose, 18.7% hemicellulose, and 15.2% lignin. The lignocellulose composition of these two plant species was similar in raw form. It differed depending on the type of ionic liquid used for its pretreatment. It should be pointed out that the pretreatment was carried out in two stages, which was also reflected in differences in the final composition of substrates aimed for hydrolysis. Stinging nettle stalks subjected to the influence of imidazolium ionic liquids caused a decrease in cellulose content by about 10% and lignin content by about 7-9%, with a simultaneous increase in the amount of hemicellulose. In turn, the pretreatment using ammonium ionic liquids did not affect carbohydrate losses and did not cause nettle stalk delignification. For comparison, similar results were obtained for giant miscanthus before and after pretreatment with ionic liquids. However, when imidazolium ionic liquids were used, the lignin content in miscanthus samples decreased to 5-7%, and the cellulose content increased by about 2-4%.
The use of imidazolium ionic liquids to dissolve miscanthus and nettle stalks, although it brings better results, is also troublesome. As cellulose dissolves, the liquid becomes more viscous and hardly miscible. When antisolvent, in that case, deionized water is added, it turns difficult to dissolve the gel. Washing out cellulose fibres from the ionic liquid is challenging and time-consuming due to its dense consistency. This is not the case with ammonium ionic liquids. The viscosity of the ionic liquid and biomass solution was lower, which allowed further mixing. It was also easier to precipitate fibres from these liquids, and the process of ionic liquid washing out took a shorter time. Images from the scanning electron microscope exhibit that miscanthus and nettle cellulose fibres change from crystalline to amorphous forms after the treatment with both imidazolium and ammonium liquids, as presented in Fig. 1. On the other hand, the most significant disadvantage of ammonium ionic liquids is that using them is not possible to depolymerize lignocellulose and extract lignin.
Enzymatic hydrolysis. After pretreatment with ionic liquids, stinging nettle and giant miscanthus were subjected to enzymatic hydrolysis. The application of xylanase aimed to increase the porosity of cellulose fibres and increase the number of contact points for cellulolytic enzymes. The consequence of such action should be hemicellulose crystallization and exposure to cellulose fibres and, as a result, an increase in hydrolysis efficiency. The efficiency of enzymatic hydrolysis was evaluated after 96 h of the process, determining the glucose and xylose concentration. Glucose concentrations are presented in Fig. 2 . In other cases, the glucose concentration did not exceed 2.5 g L −1 , both for stinging nettle and miscanthus. Enzymatic hydrolysis occurred with very low efficiency in the variants where no treatment with ionic liquids was applied. The glucose concentration of 1.9 g L −1 was obtained for miscanthus and 0.025 g L −1 for stinging nettle. The content of xylose determined in samples subjected to enzymatic hydrolysis is presented in Fig. 3.  [OAc] pretreatment, 2.6 g L −1 and 2.5 g L −1 of xylose were obtained from stinging nettle and giant miscanthus, respectively. The xylose concentration in control samples did not exceed 1.5 g L −1 . The lowest xylose content was observed in nettle and miscanthus samples treated with ammonium ionic liquids (results below 0.9 g L −1 ). Considering the influence of the application of individual ionic liquids on glucose and xylose content, it can be concluded that the most effective in the treatment of biomass is dissolution with [emim] [OAc]. On the other hand, the ammonium ionic liquids did not improve the efficiency of the hydrolysis process in comparison to the native material.
Alcoholic fermentation. The obtained hydrolysates were subjected to alcoholic fermentation using Saccharomyces cerevisiae type II yeast. Chromatographic analysis showed that the highest concentration of ethanol was obtained in samples of stinging nettle (7.3 g L −1 ) and giant miscanthus (7.0 g L −1 ), which were pretreated with [bmim][OAc], as presented in Fig. 4.
The results of the statistical analysis include ordering the depending variables according to their level of significance. It was shown that the content of ethanol from biomass of stinging nettle and giant miscanthus is mainly affected by the type of ionic liquid used, then the amount of simple sugars after enzymatic hydrolysis, and the content of lignin in samples intended for hydrolysis and fermentation. Using the Random Forest machine learning algorithm, a model with a determination factor R2 between the estimated and the observed ethanol content of 0.96 was created, as presented in Fig. 5.
The collection of Out of Bag observations made it possible to determine the importance of particular traits for the content of ethanol from biomass. Information about the importance of a given variable is obtained directly from a trained model. Using the internal structure of the Random Forest algorithm, it is possible to determine how important are the traits used for its learning. In a stochastic manner, the algorithm selects the features of

Discussion
The production of ethanol from lignocellulose requires improvements and modifications related to pretreatment, enzymatic hydrolysis, and fermentation to increase the profitability of ethanol production and the transition from the laboratory to the industrial/commercial scale. One of the most important objectives is to increase the efficiency of the fermentation process so that the entire pool of sugars (pentose and hexose) is metabolized to ethanol.
Other barriers related to ethanol production include the variable composition of biomass, the presence of inhibitors as a result of pretreatment, osmotic and oxidative stress. However, the critical element in the production of second-generation bioethanol is the choice of raw material. The raw material from which bioethanol will be produced should contain the highest cellulose content possible, along with high hemicellulose and low lignin content, because it is this fraction of the lignocellulosic complex that is reflected in the concentration of fermenting sugars. A variety of lignocellulose feedstocks have been examined for use in the production of biofuels, including energy crops (e.g., miscanthus or switchgrass), forest-based woody wastes, and forest biomass, agricultural, industrial, municipal, and food wastes 32 . Another essential feature is the high availability of raw materials. Therefore, this article points to the possibility of using stinging nettle stems as a substrate for bioethanol production.
It was shown that the content of ethanol from stinging nettle and giant miscanthus biomass is affected mainly by the type of ionic liquid used for pretreatment, then the amount of simple sugars after enzymatic hydrolysis, and the content of lignin in samples intended for hydrolysis and fermentation. Importantly, not only the lignin content but also its structure influences bioethanol production. After the conventional pretreatment methods, more condensed lignin is generated, hampering the bioconversion efficiency 33 .
A pretreatment method well-suited to the raw material can significantly improve the hydrolysis efficiency of the lignocellulosic substrate. However, it should be pointed out that aggressive methods cause losses in the pool of fermenting sugars and contribute to the formation of process inhibitors 31 . Therefore, mild pretreatment methods are becoming more and more popular, which, as in the case of ionic liquids, can improve the availability of cellulose fibres and remove lignin but also causes the formation of large quantities of insoluble hemicelluloses [34][35][36][37][38] . Moreover, a combination of pretreatment methods is also gaining more and more attention 39 . The separation of lignin from cellulose using ILs depends on several factors. Hart et al. reported that hydrogen bonding strength was not a crucial factor for the lignin dissolution in ILs as it was in the cellulose dissolution, however a minimum hydrogen bonding basicity was still required to solubilize the lignin 40 . The removal of lignin in the pretreatment improves the efficiency of enzymatic hydrolysis but causes changes in the lignin structure. Moreover, even if lignin is not fully removed, its structural change also alters its position to cellulose fibres. It creates pores and free spaces, which ultimately causes the hydrolysis process to be more efficient. Unfortunately, there aren't many reports that would exhibit total fractionation of lignocellulose using ionic liquids into the main constituents.  [41][42][43] . An increase in cellulose content may be caused by the depolymerisation of fibres in ionic liquids, which improves their extraction. It can be assumed that the treatment with ammonium ionic liquids did not significantly influence the composition of the biomass studied (Table 1). In particular, it did not contribute to the removal of lignin, which makes it impossible to carry out further stages aimed at ethanol production effectively. Kumar   www.nature.com/scientificreports/ et al. used various pretreatment methods for the stinging nettle and obtained (in the best case studied) 85% of cellulose, 6% of hemicellulose, and 3% of lignin using water retting and decortication 22 .

Most of the ionic liquids used in biomass fractionation are imidazolium salts. The literature indicates that 1-ethyl-3-methylimidazolium acetate ([emim][OAc]), 1-allyl-3-methylimidazolium chloride ([amim][Cl]) and 1-butyl-3-methylimidazolium chloride ([bmim][Cl]) can serve as effective, non-derivatizing cellulose solvents at temperatures below 100 °C, and out of more than 20 ionic liquids tested, [amim][Cl] has proven to be an excellent wood chip solvent. For example, the addition of [bmim]
[Cl] causes the initial enzymatic hydrolysis rate and the pretreatment efficiency of the cellulose process to increase 50 times for regenerated cellulose compared to the untreated one 45 . Importantly, an increase in the rate of enzymatic hydrolysis of cellulose is associated with an increase in the production of simple sugars, which can be converted to ethanol. Moreover, the process of biomass degradation with the use of ionic liquids is less energy-intensive, easier to carry out, and more environmentally friendly than previously known solutions 24,28,46,47 . On the other hand, limitations for the application of ILs in the pretreatment of lignocellulose biomass are also being identified, with the most important factors being their high cost, high viscosities and moisture sensitivity, which makes it difficult to introduce an industrial scale process with their use 48 . For the process to be economically viable, water consumption must be reduced, and an effective system for ionic liquids recycling must be developed. Attempts were made to reduce the cost of solvent acquisition, replacing imidazolium ionic liquids with liquid obtained from aromatic aldehydes of lignin and hemicellulose, i.e., by-products from biofuel production 49,50  It was reported that both imidazolium and ammonium ionic liquids compete for hydrogen bonds present in cellulose structure, thus disrupting its three dimensions network 26,41 . It was reported that a key reason for this was the high hydrogen bond acceptor capacity (β) of the [OAc] anion (β = 1.201) in comparison to previously mentioned chloride anion (β = 0.83) 52 . Due to this, 1-ethyl-3-methylimidazolium acetate is confirmed to be one of the best and is one of the most commonly used ILs, able to dissolve a large variety of lignocellulosic biomass and to fractionate it into cellulose-and hemicellulose-rich fractions, as well as to produce high pure lignin 53-55 . In the case of the presented cyclohexylammonium ionic liquids, the high performance of [CHDMA-C6][OAc] affected all alkyl groups' fine-tunning. According to the mechanism described previously, two activity categories were fundamental: (1) two methyl groups using its six activated C-H bonds to link with both the acetate and cellulose surface and (2) hexyl and especially cyclohexyl are symmetry breaking substituents. Similarly, as in the case of imidazolium ionic liquids, appropriate cation allows exploiting proton acceptability of carboxylate, which further enables the breakdown of inter-and intramolecular hydrogen bonds 41 .
The loosening of the lignocellulosic complex structure significantly facilitates enzymatic hydrolysis, the effectiveness of which depends on the selection of enzymatic preparations. Studies on the hydrolysis of a specific raw material are closely related to optimizing the preparation dose and process conditions. These arduous activities are usually carried out on selected 2-3 variants, characterized by the highest cellulose concentration after pretreatment. At the initial stage of research on the suitability of a given raw material for ethanol production, it is advisable to select enzymatic preparations known and tested in the context of hydrolysis effectiveness. However, it is worth mentioning that an important element of the lignocellulosic complex is hemicellulose, which may significantly reduce the effectiveness of hydrolysis 56 . Therefore, it is justified to use xylanases, which increase the material's porosity, expose cellulose fibres, and result in higher concentrations of fermenting sugars, which was also performed in this study (Figs. 2 and 3) is justified in the literature 57,58 . The next stage of second-generation bioethanol production is ethanol fermentation. In this study, the biosynthesis of ethanol was carried out with the participation of Saccharomyces cerevisiae yeast. Hydrolysates from giant miscanthus and stinging nettle were compared.
The full potential of stinging nettle has not yet been shown in any experimental work, and importantly it was not without reason that we have decided to use a plant that has not yet been used for ethanol production and compare it with one of the most popular energy plants used in this context. The production of ethanol from giant miscanthus has already been the subject of many studies comparing the methods and effectiveness of pretreatment, the degree of hemicellulose conversion to fermenting sugars, and the efficiency of ethanol fermentation.
This study focused on the potential of novel lignocellulosic wastes and their comparison at the same conditions applied to well-known biomass sources. The ethanol concentration obtained for both investigated raw materials is comparable, as presented in Fig. 4, which means that the stinging nettle stems are a promising alternative to energy crops such as giant miscanthus.

Materials and methods
Raw material. Common nettle stalks used for the research came from an agricultural wasteland with an area of 4.9 ha (Maszkowo, Zachodniopomorskie, Poland) excluded from agricultural production for 15 years. Ionic liquids. For the pretreatment of cellulose-rich material, five ILs from imidazolium and ammonium groups were chosen, as presented in Fig. 7. Three of them were commercially available (Iolitech GmbH, Germany) Raw material pretreatment. 10 g of ground stalks of stinging nettle and giant miscanthus were added to 50 cm 3 of an appropriate ionic liquid, homogenized, and dissolved at 120 °C for 2 h. After incubation, the samples were left to cool, and then deionized water was added to rinse the cellulose fibres and separate the biomass from the IL. During the addition of the deionized water, the IL dissolves in water, and the plant fraction precipitates. The water-IL solution with biomass was filtered on a Shot funnel with a filter (Whatman 1.0 paper). This procedure was repeated four times for imidazolium ILs, where a significant increase in the plant-IL mixture was present, and two times for ammonium ILs. Such purified stalks of nettle and miscanthus were subjected to enzymatic hydrolysis.
Enzymatic hydrolysis. Thermostable xylanase, derived from a modified strain of E. coli bacteria (Sigma Aldrich) and CellicCTec2 enzyme were used for enzymatic hydrolysis of biomass samples (purified and nonpurified ones). The initial cellulose and hemicellulose concentration was 1.0% (w/v) based on 100 mL (50 mM sodium citrate buffer) of total liquid in 250 mL Erlenmeyer flasks. Initially, xylanase (≥ 40 units mg −1 ) in the amount of 8 U mg −1 hemicellulose was added and incubated at 65 °C, pH 5.0. Hydrolysis at this stage was carried out for 24 h with 250 rpm mixing. After this time, the temperature was lowered to 50 °C, and 15 FPU g −1 cellulose of commercial cellulase enzyme Cellic CTec2 (Novozymes, Denmark) was added to the solutions. After 96 h of enzymatic hydrolysis with the use of cellulases complex, the content of glucose and xylose was determined with the use of high-performance liquid chromatography. All experiments were performed three times to establish a standard deviation.
Alcoholic fermentation. The alcoholic fermentation was carried out accordingly to our previous reports [59][60][61] . Hydrolysate solutions, previously filtered to separate the lignocellulose residue, were subjected to alcoholic fermentation. The pH of the fermentation broth was measured at each sampling and adjusted to 5.0 by the addition of either 10 wt% H 2 SO 4 or 20 wt% NaOH. Fermentation was started by adding freeze-dried distiller's yeast Saccharomyces cerevisiae type II (Sigma-Aldrich) (5% v/v). Ethanol fermentation was conducted for four days in anaerobic conditions. Samples were taken and analysed for ethanol concentration after fermentation.
Analytical methods. To examine the influence of ionic liquids on the structure of lignocellulose and the amount of available cellulose, all samples were tested for the content of cellulose, lignin, and hemicellulose (Ankom A200; ANKOM Technology); the crystalline structure of the samples was recorded using a scanning electron microscope (SEM). The morphology of cellulose fibers in buckwheat straw samples before and after ionic liquid pretreatment was recorded using SEM FEI Quanta 200 Mark 2. The content of glucose and ethanol was determined by using high-performance liquid chromatography. Samples were first centrifuged at 4000×g for 10 min at 4 °C (Multifuge 3SR, Germany) and then was filtered through a 0.22 µm membrane filter (Millex-GS, www.nature.com/scientificreports/ Millipore, USA) before analysis using an HPLC system (Merck Hitachi, Germany). Glucose and ethanol were separated on an Aminex HPX-87P (Bio-Rad, USA) at 30 °C using a 5 mM H 2 SO 4 as the mobile phase at a flow rate of 0.6 cm 3 min −1 and then detected with a refractive index detector (Model L-7490, Merck Hitachi, Germany). All the analytical methods have been described in detail before in our previous works [59][60][61] . The Random Forest algorithm implemented by David Lary (https:// david lary. info) in Matlab was used to analyze the results. The Random Forest algorithm generalizes the idea of decision trees and is based mainly on the bagging method. The concept of this algorithm is based on the construction of a group of decision trees, which are created based on a random data set 62 . Classification in this algorithm is based on the voting of classifiers. The assessment of the probability of misclassification, built into the mechanism of the classifier, allows determining the out of bag error (OOB). Thanks to OOB observations, it is also possible to estimate the importance of the observation vector variables from the point of view of the classification 63 based on this property, the vector of traits (f) was constructed, based on which their significance in the process of bioethanol production was determined, as presented in the Eq. (1), where G is the glucose; X is the xylose; C is the cellulose; L is the lignin; H is the hemicellulose; IL is the ionic liquid.

Conclusions
Pretreatment of stinging nettle and giant miscanthus with imidazolium ionic liquids allow for the increase of the availability of a key fraction of lignocellulose which is cellulose. Such application of [bmim][OAc] in the pretreatment of stinging nettle stems (and subsequent enzymatic hydrolysis) allowed us to obtain the highest concentrations of ethanol in the fermentation process, equal to 7.3 g L −1 . In comparison, the ethanol amount achieved for miscanthus was 7.0 g L −1 . Moreover, it was shown that ammonium liquids, although they allow for the more effective dissolution of the raw material, do not increase the concentration of ethanol in the fermentation process. Given the presented results of bioprocesses conducted and literature data related to the common occurrence and characteristics of the raw material, it can be assumed that stinging nettle, which in the case of the used stems is considered an agricultural waste, is a promising raw material for the production of secondgeneration bioethanol.