Introduction

The global demand of sustainable and environmentally friendly renewable energy is increasing significantly because of greenhouse gas emissions from fossil fuels and their associated impacts on the climate change [1]. Lignocellulosic biomass has the potential to play a key role in fulfilling the high energy demand due to its abundancy and the prospect of converting it into fuels and other value-added products. Lignocellulosic biomass consists of semicrystalline cellulose, amorphous hemicellulose and aromatic polymers of lignin [2]. It is estimated that Australia has the potential of generating approximately 80 million tonnes lignocellulosic biomass per year and this number is expected to increase up to 110–115 million tonnes per year in the next 20–40 years [3].

River red gum (RRG) (Eucalyptus camaldulensis), a species of eucalypt family, is one such example of woody lignocellulosic biomass. Eucalyptus is native to Australia and approximately three quarter of the Australian forest, equivalent to 92 million hectares, is eucalypt forest where majority trees are RRG [4]. In South Australia, almost 50% of the total annual firewood consumption (410,000 t) is met by RRG [4]. RRG can be considered as an energy crop as it requires low water and fertiliser for its growth. RRG biomass mainly comprises of cellulose (46–49%), hemicellulose (18–23%), lignin (29–33%), ash (0.1–0.2%) and extractives (2–5%) [5]. The major advantage of RRG is that it contains very low amount of ash and extractives compared to many other agricultural lignocellulosic wastes such as sugarcane bagasse [6]. Therefore, it is considered to be one of the most promising candidates for biofuel and biochemical production [7, 8]. It is estimated that RRG tree planted in one hectare area has the potential to produce approximately 7000 l of bioethanol [9].

Recent research on lignocellulosic biomass treatment [10, 11] is focusing on separating lignin and cellulose-rich fractions and further processing them to produce value-added materials as highlighted in Fig. 1.

Fig. 1
figure 1

Production of value-added materials from lignin and cellulose-rich fractions

Full size image

Several chemical pre-treatment methods such as dilute acid treatment, alkali treatment and organic solvent treatment have been reported in the literature for separating lignin- and cellulose-rich fractions [11]. In the case of organic solvent treatment, known as OrganoSolv pre-treatment, chemicals such as ethanol, methanol, ethylene glycol, glycerol, acetic acid and formic acid have been extensively studied with and without catalysts [12,13,14]. Among these organic solvents, low boiling point alcohols such as ethanol and methanol are considered to be the most favourable solvents for OrganoSolv pre-treatment due to their low cost [15]. However, there are several shortcomings reported for OrganoSolv pre-treatment in the literature such as low solvent recovery as well as environmental, health and safety hazards associated with the high volatile and flammability properties of organic solvents [15].

Recently, the IL-based pre-treatment of lignocellulosic biomass, known as IonoSolv pre-treatment, has gained notable interests [16,17,18,19,20,21,22,23]. IonoSolv pre-treatment can effectively extract lignin from lignocellulosic biomass. Additionally, acidic IL treatment of lignocellulosic biomass for the production of platform chemicals such as furfural, 5-hydroxymethylfurfural (5-HMF) and levulinic acid is also attractive to the scientific community. For instance, the direct hydrolysis of cellulose and hemicellulose fractions using acidic ILs is an attractive approach for platform chemical production [24,25,26,27]. ILs are low temperature organic salts and these organic salts have a wide range of benefits including negligible vapour pressure, non-flammability and chemical and thermal stability. In addition, their properties can be tuned by combining various anions and cations from a large range of selections, their recovery is easy as they are low vapour pressure compounds and they are environmentally sustainable [28, 29]. To date, a wide numbers of ILs have been synthesised from the combinations of trillions of cations and anions [30] and employed in the biomass dissolution study. Among the imidazolium based ILs, 1-butyl-3-methylimidazolium chloride ([Bmim][Cl]), 1-ethyl-3-methylimidazolium acetate ([Emim][OAc]) and 1-butyl-3-methylimidazolium acetate ([Bmim][OAc]) showed promising results on lignin separation from lignocellulosic structure [31,32,33].

While there are numerous studies focusing on various approaches of biomass pre-treatments, the studies published on IonoSolv pre-treatment for separating cellulose-rich fraction from biomass are still very limited when compared to OrganoSolv pre-treatment [31, 34,35,36,37]. Moreover, RRG, despite its huge potential in many parts of the world including Australia, has not been studied extensively with organic solvents and ILs for its pre-treatment. It is still unclear which pre-treatment method (i.e. OrganoSolv treatment or IonoSolv treatment) works better for recovering cellulose-rich fraction from RRG. It is yet not known which ILs can work better for the pre-treatment of RRG and how comparable those ILs will be with organic solvents. Additionally, morphological transformations observed in the cellulose-rich material obtained from IL and organic solvent pre-treatments are also not explored in detail. In this work, the IonoSolv pre-treatment of RRG is compared with the OrganoSolv pre-treatment of the same biomass on the basis of several measurable factors including the degree of delignification, morphological changes in the cellulose-rich material structure, thermal stability and cellulose crystallinity index (CrI). The comprehensive study on both OrganoSolv and IonoSolv pre-treatments presented in this work discloses several interesting findings as reported further in the “Results and Discussion” sections.

Experimental

Materials

The biomass sample employed in this investigation was from RRG tree and it was sourced from a local Victorian supplier. The material was grinded and sieved to obtain particles in the range between 75 and 150 μm. The proximate and ultimate analyses of the untreated RRG were carried out using TGA Q500IR and LECO CHNS 932, and the results are tabulated in Table 1.

Table 1 Proximate and ultimate analyses of untreated RRG
Full size table

Before initiating any pre-treatment experiment, RRG was dried at 90 °C for 4 h. The ILs, employed in this investigation, are [Emim][OAc] (purity ≥ 95%), [Bmim][OAc] (purity ≥ 95%) and [Bmim][Cl] (purity ≥ 98%). These ILs were sourced from Sigma Aldrich. A couple of organic solvents were used in this investigation and they are methanol (purity ≥ 99.9%, Sigma Aldrich) and ethanol (purity ≥ 99.5%, Sigma Aldrich). Other chemicals employed in the experimental investigations include sulphuric acid (72 w/w%, RCI Labscan) and acetone (purity ≥ 99.9%, Sigma Aldrich).

Biomass Pre-Treatment

IonoSolv Pre-Treatment

The IonoSolv pre-treatment was carried out in a 20ml vial. Initially, 0.5g oven dried RRG sample and 5 g IL were taken into the vial (RRG and IL in a 1:10 w/w ratio). In addition, 1 g distilled water was mixed with 5.5 g biomass-IL mixture when solid IL; [Bmim][Cl] was used. The weight of oven dried RRG sample was used as basis on the estimation of regenerated cellulose-rich material (RCRM).

The vial was immersed into a silicon oil bath as shown in the Supplementary Material (Fig. S1). The oil bath was placed on a magnetic stirrer (MR Hei-Standard, Heidolph Instruments) operating at 70 °C while providing a magnetic element stirring effect of 500 rpm in the oil. The pre-treatment was run for 3 h.

The pre-treated solution was then diluted with 50% (v/v) aqueous acetone. The ratio of pre-treated solution and aqueous acetone was 1:10 (v/v). The diluted solution was stirred at room temperature with a magnetic element at a speed of 500 rpm for 1 h. This stirring effect as well as chemical effect of acetone led to precipitate cellulose from the solution. The solution was centrifuged at 3000 rpm for 30 minutes and filtered through Whatman filter paper (pore size 6 μm) to collect the RCRM. While in filter paper, the solid RCRM was washed with distilled water to remove the residual IL. A conductivity meter (FiveEasy F30, Mettler Toledo) was employed in this investigation and the removal of ILs was confirmed by conductivity test of filtered water. When the conductivity of filtered water and the conductivity value of distilled water were very close, the washing of RCRM was terminated. The RCRM was dried overnight at 90 °C in a drying oven. It was then weighed employing a laboratory scale (Ohaus Pioneer PA114C, USA; readability 0.0001 g). The RCRM yield was calculated using Eq. (1).

$$ \mathrm{RCRM}\ \mathrm{yield}\ \left(\%\right)=\frac{M_{\mathrm{RCRM}}}{M_{\mathrm{URB}}}\times 100 $$
(1)

where MRCRM and MURB indicate the mass of RCRM and the mass of untreated RRG, respectively.

The filtrate is a lignin-rich solution and lignin was separated from this solution by evaporating acetone. Once the evaporation of acetone is completed, the solution was centrifuged at 3000 rpm for 30 minutes and filtered through Whatman filter paper (pore size ≤2 μm). The separated lignin contains IL which needs to be removed. Similar to previous procedure, while in filter paper, lignin was washed with distilled water and the removal of IL was ensured by conductivity test. The lignin was dried overnight at 90 °C in a drying oven and its weight was measured gravimetrically. The degree of delignification was estimated by the following equation:

$$ \mathrm{Degree}\ \mathrm{of}\ \mathrm{delignification}\ \left(\%\right)=\frac{M_{\mathrm{RL}}}{M_{\mathrm{LURB}}}\times 100 $$
(2)

where MRL and MLURB indicate the mass of recovered lignin and the mass of lignin in untreated RRG respectively.

In this study, the TAPPI (Technical Association of the Pulp and Paper Industry) test method was applied on the determination of lignin content in the untreated RRG. This test procedure was established in a prior work [38]. The lignin content in the untreated RRG was found to be 29% and this value was used as MLURB for the degree of delignification calculations.

The second filtrate (obtained after lignin separation) was constituted of hemicellulose, water and IL. This filtrate was initially heated to evaporate water from the solution. Ethanol (95%, v/v) was then added to this saturated solution in a ratio of 3:1 (v/v). Finally, the solution was centrifuged at 3000 rpm for 30 minutes and the solid hemicellulose was separated from the solution by filtration (Whatman filter paper, pore size ≤2 μm). The filtrate, obtained from this filtrated, was evaporated to recover IL. The recovered IL was kept overnight in a vacuum drying oven at 80 °C. The oven dried IL was ready to use for the next pre-treatment process. Once the IonoSolv pre-treatment of RRG was completed with one IL, the process was repeated with the other two ILs.

OrganoSolv Pre-Treatment

A solvent–water solution was initially prepared in a 50:50 volume ratio. In a 20ml vial, 0.5g oven dried RRG sample was mixed with the solvent–water solution in a ratio of 1:10 (w/w). The mixture-containing vial was heated for 3 h using an oil bath and a magnetic stirrer as described in the IonoSolv pre-treatment process. After the completion of pre-treatment, the mixture was filtered through Whatman filter paper (pore size 6 μm) and the solid material was washed with distilled water ten times to ensure that the solvent was removed completely.

The filtrate obtained from this filtration contained lignin. In order to separate lignin from the filtrate solution, it was diluted with distilled water in a 3:1 (w/w) ratio and kept overnight to precipitate lignin. The temperature of distilled water was kept at 4 °C as suggested by a previous work [39]. The solution was centrifuged at 3000 rpm for 30 minutes and the solid lignin was separated from the solution by filtration using a Whatman filter paper (pore size ≤2 μm). The separated lignin was then washed with distilled water. The filtrate was concentrated by water evaporation. Hemicellulose was precipitated by adding 95% ethanol in a ratio of 3:1 (v/v) to this concentrated solution. The solution was centrifuged at 3000 rpm for 30 minutes and the solid hemicellulose was then separated by filtration (Whatman filter paper, pore size ≤2 μm). Both RCRM and lignin were dried overnight at 90 °C in a drying oven and their weights were measured. The organic solvent was not recovered. Once the RRG pre-treatment was completed by one organic solvent, the procedure was repeated with the other organic solvent.

Each experiment was repeated for three times. Thus, the data presented in the article represent the average values of the three experiments performed at identical conditions. Error bars are included where possible. These bars illustrate the maximum and minimum values from the average values.

Characterisation Methods

XRD Analysis

The crystalline and amorphous structures of untreated RRG and RCRM samples were analysed by a Bruker Axs D4 Endeavour Wide Angle X-Ray Diffraction instrument. The X-ray diffraction patterns of various samples were obtained at 2θ = 6 − 40° with a step size of 0.02°, a counting time of 0.3 s/step and divergent split of 1°. The instrument was operated at 40 kV voltage and 35 mA current. The CrI was estimated by using the XRD data and applying the following equation [6, 7, 40]:

$$ \mathrm{CrI}\ \left(\%\right)=\frac{I_{002}-{I}_{\mathrm{am}}}{I_{002}}\times 100 $$
(3)

where I002 refers to the maximum intensity of crystalline portion and Iam indicates the minimum intensity of amorphous region. Iam was estimated while capturing a XRD spectrum in the instrument. It is the intensity of the valley between I101 and I002 peaks with respect to baseline.

Thermogravimetric Analysis

Thermogravimetric analysis (TGA) of untreated RRG and RCRM samples was conducted by a STA 6000 Simultaneous Thermal Analyser (PerkinElmer, USA). The instrument was integrated with PerkinElmer Software Pyris®. This software was used in operating the instrument and collecting data. The sample was heated in a temperature range of 50–700 °C with a heating rate of 20 °C/min. Nitrogen was used as the TGA purge gas with a flow rate of 20 ml/min.

Scanning Electron Microscope Imaging and Analysis

The surface morphologies of untreated RRG and RCRM samples were examined by a scanning electron microscope (Philips XL30, USA). Prior to capture any SEM image, the specimen was placed on an aluminium stub with carbon tape and coated with gold using sputtered coating instrument. With the aim of comparing the surface morphology of untreated RRG and RCRM samples, the SEM images were obtained at the same spot size (5.0) and magnification (2000×). The SEM images were analysed using ImageJ software (version 1.46r), a public domain software developed by Wayne Rasband, National Institutes of Health, USA. In the software platform, the scale was initially set with a line of known length. Once the scale was set, it was used to determine pore diameter and pore area.

Fourier-Transform Infrared Spectroscopic Analysis

FTIR spectroscopic analyses of untreated RRG and RCRM samples as well as pure and recovered ILs were performed in a PerkinElmer FTIR Spectrometer (Spectrum 100, USA). FTIR spectra were captured in the scanning range of 4000–600 cm−1.

For each experimental condition, the characterisation of materials was replicated. Thus, for each experimental condition, the characterisation of materials was performed for six times. This approach ensured that various figures and images presented in this article from material characterisation are reproducible.

Results and Discussion

RRG Dissolution and Degree of Delignification

The RRG dissolution capabilities of various ILs and organic solvents are illustrated in the Supplementary Material (Fig. S2). The darker colour IL pre-treated solutions suggest that ILs are more capable in RRG dissolution than organic solvents. It is expected that lignin, hemicellulose and cellulose of RRG can be dissolved in ILs while only lignin and hemicellulose are dissolved in organic solvents. Hence, ILs show higher RRG dissolution capabilities compared to organic solvents. Similar findings are reported for other lignocellulosic biomass OrganoSolv pre-treatments in a previous work [13].

The IonoSolv pre-treatment processes transformed RRG into more fluffy structured materials when compared to OrganoSolv pre-treatments and this was confirmed by the apparent density measurements of RCRM samples as highlighted in Fig. 2. It can be seen that apparent densities of IonoSolv RCRM samples were reduced significantly when compared to untreated RRG or OrganoSolv RCRM samples. The gradual development of fluffy structured materials in IL pre-treatment of RRG assisted on the enhanced cellulose dissolution and lignin fragmentation [41].

Fig. 2
figure 2

Effect of pre-treatment method on RCRM apparent density

Full size image

We hypothesised that the dissolution of RRG in ILs might be attributed to the splitting of inter-unit linkages of lignin and dissolution of its constituents (i.e. cellulose, lignin and hemicellulose). It is previously established that both anions and cations play significant roles in the dissolution of biomass constituents [41]. During the pre-treatment process, ILs dissociate into cations such as [Bmim]+ and [Emim]+ and anions such as [Cl]− and [OAc]−. Cations from ILs interact with polymeric molecules of lignin. In this process, electron-rich aromatic π-systems of IL cations involve in the interactions with π-systems of aromatic groups of lignin [42, 43]. In addition, IL anions can form H-bonds with hydrogen atoms of OH− groups of lignin. The π–π interactions of IL cation–lignin and the formation of H-bonds by IL anions thus dissolve lignin polymers during IL pre-treatment processes. In the case of cellulose, it contains OH− groups and IL anions interact with hydrogen atoms of OH− groups of cellulose forming H-bonds. Hydrogen atoms of IL cations can also form H-bonds with the electronegative oxygen atoms of OH− groups of cellulose [44]. The mechanism of dissolving hemicellulose is believed to be similar to that of cellulose dissolution [44].

As described in the methodology section, the dissolved cellulose in the IL pre-treatments was recovered by using antisolvent (aqueous acetone). In the case of OrganoSolv pre-treatment, the cellulose remained undissolved as described elsewhere [13] and therefore, no further treatment was required to recover the cellulose component.

The delignification performance of ILs and organic solvents is summarised in Fig. 3. The degree of delignification employing ILs was in the range of 21.4 to 26.2%. This pre-treatment result is in agreement with the previous works reported in the literature and described as follows. Doherty et al. studied the pre-treatment of maple wood flour at 90 °C employing [Emim][OAc] [45]. In their study, the degrees of delignification were approximately 25, 35 and 49% for operating times of 6, 12 and 24 h respectively. When the wood flour was pre-treated with [Bmim][OAc] at same temperature, the degrees of delignification were approximately 26, 32 and 37% for 6, 12 and 24 h operating times, respectively. As evident in these results, the duration of pre-treatment is an important factor in the delignification process. Li et al. found that [Bmim][OAc] and [Bmim][Cl] can remove lignin from a hybrid biomass (Eucalyptus urophylla x Eucalyptus grandis) by 16.97 and 7.50%, respectively [32]. Their experiments were carried out at 120 °C for 30 min. The pre-treatment of southern yellow pine employing [Emim][OAc] at 110 °C for 16 h showed 26.1% reduction of lignin in the pre-treated material [38], and this degree of delignification was nearly similar to the present study (26.2%). Heggset et al. pre-treated Norway spruce chips with [Emim][OAc] at 100 °C for 6 h and the study showed nearly 31.7% lignin reduction in the pre-treated material [46]. Mohtar et al. observed approximately 54% lignin reduction in pre-treated material obtained from the [Bmim][Cl] pre-treatment of oil palm biomass for 8 h at 110 °C in N2 environment [33]. The process temperature is also an important factor that impacts the degree of delignification. When relatively higher process temperature (175 °C) was applied on the pre-treatment of southern yellow pine with [Emim][OAc], a higher value of lignin (49.4%) was separated within a short pre-treatment duration of 30 min (comparative result in that study 26.1% lignin separation at 110 °C for 16 h) [47]. The lignin content in the RCRM, obtained from the pre-treatment of maple wood flour with [Emim][OAc], was found to be reduced by more than 85% when the raw biomass material was pre-treated for 70 h at 90 °C in N2 environment [16]. It can be seen from the above discussions that the degrees of delignification values obtained in the present study are comparable with several previous studies. However, variations are also observed on the degrees of delignification values. Those variations of delignification efficiencies may be attributed to the differences in pre-treatment conditions and biomass structures.

Fig. 3
figure 3

Delignification performance of ILs and organic solvents on RRG

Full size image

In the case of organic solvents, ethanol removed more lignin from RRG than methanol. The degree of delignification values for methanol and ethanol pre-treatments was 21 and 26.1%, respectively and these values are in agreement with a recent study on the pre-treatment of eucalyptus chips with ethanol (concentration 60%) [14]. The eucalyptus chips pre-treatment study was conducted at 160 °C for 90 min and an extraction of 21.7% lignin was reported.

When comparing IonoSolv and OrganoSolv pre-treatments, it can be concluded that the delignification efficiency of ethanol solvent was very similar to that of [Emim][OAc] pre-treatment (26.2%). According to the degree of delignification values obtained from the current investigation, the employed ILs can be ordered as follows: [Emim][OAc] > [Bmim][OAc] > [Bmim][Cl]. This finding suggests that ILs having acetate ([OAc]−) anions are superior in lignin dissolution than that of chloride ([Cl]−) anions. This was mainly because the hydrogen bond basicity of [OAc]− anions is higher than that of [Cl]− anions [44, 45, 48]. Hydrogen bond basicity of a particular molecule refers to the ability of that molecule to accept protons [49]. The higher hydrogen bond basicity of [OAc]− anions enhances the formation of H-bonds between IL anions and hydroxyl ([OH]−) groups of lignin. Between two IL cations, employed in this investigation, [Bmim]+ is heavier than [Emim]+. The shorter [Emim]+ ion is more mobile than [Bmim]+ ion in the reaction environment [17, 50]. Thus, [Emim]+ ion-containing ILs can show higher delignification capability than that of [Bmim]+ ion-containing ILs.

Figure 4 shows the average mass compositions of hemicellulose, lignin and RCRM components. It can be seen from the figure that the percentage of lignin removal was ranged from 6.2 to 7.6% while the RCRM yields were between 90.4 and 92.4%. Additionally, the pre-treatment processes removed hemicellulose with values between 1.0 to 1.8% of untreated RRG. The corresponding mass values of the separated hemicellulose were too low for further analysis. A small mass loss was observed in each pre-treatment experiment and the values of mass loss varied from 0.1 to 0.3% of untreated RRG. This level of mass loss is supported by previous works [51, 52]. Brandt-Talbot et al. [51] reported noticeable mass loss (> 10%) during the pre-treatment of Miscanthus x giganteus biomass using triethylammonium hydrogen sulphate. Similar observations are reported by a different study on the discrepancy of mass balance [52].

Fig. 4
figure 4

Mass balance of various fractions obtained from IonoSolv and OrganoSolv pre-treatments of RRG

Full size image

Characterisation of Untreated RRG and RCRM Samples

XRD and CrI Analysis

Figure 5 illustrates XRD spectra of untreated RRG and RCRM samples. The CrI values presented in this figure are average CrI values from three experiments. In the untreated sample spectra, three diffraction peaks were observed with 2θ values of 15.6°,  22.5° and 34.5° corresponding to lattice plane I101, I002 and I040, respectively. These three peaks represent the native crystalline cellulose which is also known as cellulose I. These findings are supported by a previous work which reported the presence of cellulose I at 2θ = 15.8°, 2θ = 22.3° and 2θ = 34.3° [32]. According to the XRD measurements obtained in the present investigation, the CrI value of untreated RRG was 46.3%. When the RRG was pre-treated with ILs, the CrI values of IonoSolv RCRM samples dropped to 38.2, 41.9 and 42.3% for [Emim][OAc], [Bmim][OAc] and [Bmim][Cl], respectively. As a result of IL pre-treatments, the peaks at the lattice planes I002 and I101 shifted towards lower angles and the intensities of those peaks were reduced. The highest change in peak shift (i.e. 22.5 to 22.1° and 15.6 to 15.3°) was observed in the case of [Emim][OAc] treatment. The peak shift and the reduction of the peak intensity were attributed to lead the disruption of cellulose structure as well as the transformation of cellulose I to cellulose II [53].

Fig. 5
figure 5

XRD spectra of untreated RRG and RCRM samples

Full size image

Two new weak XRD peaks were observed at 11.7° and 20.2° angles in the case of IonoSolv pre-treatments. These peaks refer to cellulose II illustrating the transformation of cellulose I to cellulose II [53, 54]. The transformation of cellulose I to cellulose II was responsible for the decrease in CrI for IonoSolv RCRM samples. When the RRG was pre-treated with organic solvents, the values of CrI were found to be increased in the RCRM samples. The increase in CrI in OrganoSolv RCRM was due to the separations of amorphous lignin and hemicellulose from RRG as supported by a previous work [14]. The separations of lignin and hemicellulose from biomass also occur in the case of IonoSolv pre-treatments; however, the impact of the transformation of cellulose I to cellulose II on CrI values is higher than that of the separations of lignin and hemicellulose from biomass in the IL pre-treatments. It can be seen from Fig. 5 that there is no new development of XRD peaks for OrganoSolv RCRM samples and this indicates that cellulose I may be unchanged in the OrganoSolv RCRM samples.

FTIR Spectroscopic Analysis

Figure 6 shows FTIR spectra of untreated RRG and RCRM samples. In the case of untreated sample, the peaks appeared at 3342, 2900, 1734, 1505, 1422, 1228 and 1028 cm−1 indicate OH stretching in cellulose, C–H stretch in cellulose, C=O unconjugated stretching in hemicellulose, C=C aromatic symmetrical stretching in lignin, C=C stretching in lignin and hemicellulose, C=O stretching in lignin and hemicellulose and C–O stretching in cellulose, respectively [53,54,55]. The peak intensity at 1505 and 1422 cm−1 decreased for the all RCRM samples indicating the partial removal of lignin and hemicellulose during pre-treatment. In the case of IonoSolv RCRM samples, two new peaks are observed at 3265 and 1267 cm−1 and their appearance is due to the results of the disruptions of H-bonds in cellulose and the transformation of cellulose I to cellulose II [53, 56]. This result is in accordance with the XRD measurements as discussed earlier. This effect was highly pronounced in the [Emim][OAc] pre-treatment. Similar to the XRD measurements, no cellulose II structure was observed in the OrganoSolv RCRM samples.

Fig. 6
figure 6

FTIR spectra of untreated RRG and RCRM samples

Full size image

Relative Change in CrI

Figure 7 shows the relative change in CrI (with respect to CrI of untreated RRG) for the IonoSolv and OrganoSolv RCRM samples. In contrast to OrganoSolv RCRM samples, CrI was found to be decreased in the case of IonoSolv pre-treatment. [Emim][OAc] exhibited the maximum reduction in CrI with a value of 17.4%. Similar findings are reported in the literature for the [Emim][OAc] treatment of wood flour with a relative reduction in CrI value of 17.5% [16]. As described earlier, OrganoSolv pre-treatment increased the CrI value of RRG. For example, the relative change in CrI of RCRM for ethanol pre-treatment was found to be 6.6%. A similar finding was reported for the pre-treatment of eucalyptus with ethanol (concentration 60%), where CrI of eucalyptus increased 5.2% with respect to CrI of untreated eucalyptus [14]. In the present study, ethanol pre-treatment showed higher increase in CrI than that of methanol.

Fig. 7
figure 7

Relative change of CrI for IonoSolv and OrganoSolv RCRM samples with respect to untreated RRG CrI

Full size image

Overall, it can be summarised that an increase in the degree of delignification is observed for IonoSolv pre-treatments and this is linked to the decrease in the CrI. This suggests that a fraction of crystalline cellulose I converts to amorphous cellulose II during IL pre-treatment process. It has been reported previously that the effectiveness of enzymatic hydrolysis of biomass is strongly correlated with CrI of biomass material [16, 57]. For instance, Lee et al. reported that the enzymatic hydrolysis of maple wood was increased when the maple wood was pre-treated with [Emim][OAc] leading to the reduction of CrI values [16]. Figure 8 shows the relative increase in enzymatic hydrolysis (with respect to enzymatic hydrolysis of untreated maple wood) against the relative reduction of CrI (with respect to CrI of untreated maple wood). Taking into account the hydrolysis trend of Lee et al. [16], CrI value of the present study can be linked to the hydrolysis efficiency (Fig. 8). Accordingly, it may be predicted that the low temperature (70 °C) [Emim][OAc] pre-treatment of RRG employed in this study can increase the hydrolysis efficiency of RCRM. As can be seen from the Fig. 8, the relative increase in hydrolysis of RCRM is expected approximately 26.1% with respect to that of untreated RRG. The lower value of CrI is also favourable for thermochemical conversion of biomass [58].

Fig. 8
figure 8

Correlation between enzymatic hydrolysis and CrI of biomass [16]

Full size image

Thermogravimetric Analysis

Figure 9a, b show the thermogravimetric analysis (TGA) and derivative thermogravimetric (DTG) analyses respectively for untreated RRG and RCRM samples (TGA profiles with repeat data are presented in the Supplementary Material in Fig. S3). Notable changes are observed in the thermal profiles of untreated RRG and RCRM samples as shown in Fig. 9a. These changes were attributed to the physicochemical transformations occurred during the IonoSolv and OrganoSolv pre-treatments. To gain deeper understanding, DTG plots provided in Fig. 9b are analysed. In general, three peaks are observed for both untreated RRG and RCRM samples. The first peak is believed to be the dehydration phase (70–80 °C) and it indicates mainly moisture removal. The second and third peaks are assumed to be the thermal decomposition of hemicellulose (250–290 °C) and cellulose (350–380 °C), respectively [59]. The decomposition of lignin is expected to extend throughout the hemicellulose and cellulose region (200 to 700 °C) as reported previously [59]. Therefore, no separate peak is observed for lignin in Fig. 9b. Noticeable differences are observed between the IonoSolv and OrganoSolv RCRM samples in the decomposition temperatures of cellulose and hemicellulose. With respect to untreated RRG and OrganoSolv RCRM samples, the decomposition temperatures are found to be shifted slightly towards a lower temperature for IonoSolv RCRM samples. This observation is highly pronounced in the RCRM sample obtained from [Emim][OAc] pre-treatment which has shown the lowest CrI and the highest intensity. This further confirms the physicochemical alterations, decrease in CrI and conversion of cellulose I to cellulose II in IonoSolv RCRM samples

Fig. 9
figure 9

Thermal decomposition behaviour of untreated RRG and RCRM samples a TGA profiles b DTG thermograms

Full size image

Surface Morphology Analysis

The morphological changes in RRG during IonoSolv and OrganoSolv pre-treatments are presented in Fig. 10 (SEM images with repeat data are presented in the Supplementary Material in Fig. S4). As described in the “Experimental” section, the SEM images were captured at the same spot size and magnification. This provides a visual comparison amongst untreated RRG and RCRM samples. The untreated RRG was observed to be compact, regularly ordered and intact structure with no pores on the surface (Fig. 10a). This was due to the presence of lignin polymers as these polymers act as glue coating and surround cellulose and hemicellulose forming rigid structure [2]. The IonoSolv and OrganoSolv pre-treatments removed lignin and hemicellulose from lignocellulosic structure and modified the surface morphologies of RCRM samples. Both organic solvents exhibited mainly thinning effects as shown in Fig. 10b and c. In contrast, IL pre-treatment, along with thinning, developed pores in the RRG structures as evident in Fig. 10d–f. The maximum pore area in RCRM samples yielded from [Emim][OAc], [Bmim][OAc] and [Bmim][Cl] pre-treatments was approximately 182.5, 172.2 and 53.4 μm2, respectively (the estimated pore areas considered the top flat area of a pore). The deconstruction of the lignocellulosic structure and washing out of materials is responsible for the thinning of the structure in both IonoSolv and OrganoSolv pre-treatments while the pore development in IL pre-treatment is attributed to the transformation of cellulose I to cellulose II.

Fig. 10
figure 10

SEM images of a untreated RRG, bRCRM from methanol pre-treatment, cRCRM from ethanol pre-treatment, dRCRM from [Bmim][Cl] pre-treatment, eRCRM from [Bmim][OAc] pre-treatment and fRCRM from [Emim][OAc] pre-treatment

Full size image

Figure 11 represents a comparative analysis on pore size (the maximum diameter of the top flat area of a pore) in the IonoSolv RCRM samples. The porous structures of RCRM samples favour their hydrolysis and biodigestibility for biofuel production [12]. In terms of pore development, the ILs can be organised as follows: [Emim][OAc] > [Bmim][OAc] > [Bmim][Cl]. As explained earlier, such trend is observed due to higher hydrogen bond basicity of [OAc]− anions compared to [Cl]− anions and higher mobility of shorter [Emim]+ cations compared to [Bmim]+ cations.

Fig. 11
figure 11

The size variation of developed pores in IonoSolv RCRM samples

Full size image

Recovery of Ionic Liquid

The cost of IL is considered to be the main impediment for the commercial applications of ILs in biomass pre-treatments [60]. Nevertheless, ILs, due to their low vapour pressures, can be regenerated which can significantly reduce the cost of biomass pre-treatment. In this study, ILs were recovered after the pre-treatments and the structural changes were analysed by FTIR spectrometer. The recovery percentage of [Emim][OAc], [Bmim][Cl] and [Bmim][OAc] were 88.2, 83.6 and 87.4%, respectively. These values are in agreement with previous studies that reported the recovery of [Emim][OAc] and [Emim][DEP] to be 89 and 90% respectively for the pre-treatments of corn stover and oil palm frond [61, 62]. However, the percentage recovery of ILs and performance of recovered ILs on biomass dissolution is largely dependent on the method employed for the recovery [63, 64]. For instance, Liang et al. recovered [Amim][Cl] (1-allyl-3-methylimidazolium chloride) using membrane-based electrodialysis method and used the recovered IL for 5 cycles in the pre-treatment of Eucalyptus globulus [65]. They observed that the recovered IL showed almost the same dissolution capability when compared that to fresh IL. Hou et al. observed a minor reduction in lignin separation capability of [Ch][Ly] (cholinium lysine) when the IL was recovered and reused over five times [66]. These highlight the recyclability of ILs and the commercial suitability of IonoSolv pre-treatment processes.

The FTIR spectra of pure and recovered ILs, obtained from the present study, are provided in the Supplementary Material (Fig. S5). It can be seen from Fig. S5 that pure ILs show three peaks in the spectral region of 3000–2800 cm−1 indicating the stretching of alkyl C–H bands [67,68,69,70]. These peaks can be traced at 2863, 2917 and 2977 cm−1 for pure [Emim][OAc]; 2868, 2934 and 2963 cm−1 for pure [Bmim][OAc] and 2870, 2932 and 2957 cm−1 for pure [Bmim][Cl], respectively. In the spectral region of 3200–3000 cm−1, two imidazolium C-H stretching bands can be observed at 3055 and 3144 cm−1 for pure [Emim][OAc]. Similar imidazolium C–H stretching bands are identified at 3073 and 3142 cm−1 for pure [Bmim][OAc] and 3090 and 3149 cm−1 for pure [Bmim][Cl] [67,68,69,70]. The broad peak in the region of 3400 cm−1 can be assigned to O–H stretching bands [67]. The absorption peaks at 1175, 1377 and 1563 cm−1 for pure [Emim][OAc]; 1171, 1379 and 1563 cm−1 for pure [Bmim][OAc] and 1170, 1461 and 1566 cm−1 for pure [Bmim][Cl] are assigned to symmetric O–C–O, asymmetric O–C–O and C–O stretching, respectively [71]. In the recovered [Bmim][Cl], the imidazolium C–H stretching band is shifted from 3090 to 3045 cm−1 and it is due to changes in local geometry caused by H-bonding [68] as illustrated in Fig. S5c. In the case of recovered [Emim][OAc], two broad O–H stretching peaks are observed at 3396 and 1660 cm−1 wavenumbers indicating the presence of moisture in recovered IL. Similarly, O–H stretching peak is identified at 1640 cm−1 wavenumber for the recovered [Bmim][Cl]. Similar to the FTIR analysis of this study, the presence of O–H group was observed in the FTIR spectrum of the recovered [Bmim][Cl] employed in the pre-treatment of coal [72]. Briefly, the FTIR spectral analysis suggests that there are no noticeable differences in the FTIR spectra of the recovered and the pure ILs.

Conclusion

The low temperature pre-treatments (i.e. 70 °C for 3 h) of RRG were studied employing ILs and organic solvents with the aim of separating lignin from cellulose-rich material (i.e. delignification).

By applying IL and organic solvent pre-treatments, more than 20% lignin was removed from RRG in this investigation. According to the delignification capabilities, the ILs can be ordered as follows: [Emim][OAc] > [Bmim][OAc] > [Bmim][Cl]. The higher delignification performance of [Emim][OAc] is observed due to higher hydrogen bond basicity of [OAc]− anions compared to [Cl]− anions and higher mobility of shorter [Emim]+ cations compared to [Bmim]+ cations. For organic solvents, ethanol was found to be working better than methanol for the delignification of RRG.

The IL pre-treatments were able to reduce CrI in the IonoSolv RCRM samples while CrI was found to be increased in OrganoSolv RCRM samples. Such differences are attributed to the conversion of crystalline cellulose I to amorphous cellulose II during IL pre-treatments and this phenomenon has not been observed in the case of organic solvent pre-treatments.

IonoSolv pre-treatments are found to produce pores in the regenerated cellulose-rich material fractions. According to pore size (from large to small) in RCRM samples, the order of ILs is as follows: [Emim][OAc] > [Bmim][OAc] > [Bmim][Cl]. In the case of OrganoSolv pre-treatments, no noticeable pore is observed in the RCRM samples.

[Emim][OAc], [Bmim][Cl] and [Bmim][OAc] were recovered with recovery values of 88.2, 83.6 and 87.4%, respectively. According to FTIR spectral analysis, a minor structural alteration was observed in the recovered [Bmim][Cl] when compared to the pure [Bmim][Cl]. No structural change was observed in the case of recovered [Emim][OAc] and [Bmim][OAc]. This establishes their recyclability and indicates that the ILs can be used cost effectively for low temperature RRG pre-treatments.