The effect of switchable ionic liquid (SIL) treatment on the composition and crystallinity of birch chips (Betula pendula) using a novel alkanol amine-organic superbase-derived SIL

2.1 Reagents

DBU (99%) and DEA (99%) were used as received from (Sigma Aldrich, Germany). CO₂ (99.999%, H₂O<0.5 ppm) was provided by AGA Oy, Finland. Methanol (99%) used as an antisolvent and for washing, was purchased from Merck, USA and used as received. Hexamethyldisilazane (HMDS) 150 µl (99%, Fluka, USA), 80 µl trimethylchlorosilane (TMCS) (98%, Fluka, USA) and 100 µl of pyridine (99%, Sigma, Germany) were used as received.

2.2 Raw material

Air-dried industrial size birch chips (approximately 2.5×2.5×0.5 cm) (Betula pendula) were provided by the Finnish Forest Research Institute (Metla).

2.3 SILs preparation

SILs were prepared from DBU, DEA and CO₂ by methods described in detail previously [9, 10]. The mixture containing 2:1 molar amounts of DBU and DEA was charged into a dry flask in a glove box under N₂ atmosphere. A narrow gauge glass tube was inserted and CO₂ was bubbled through the liquid at a constant rate, until there was no more weight increase. The reaction was rather exothermic and the solution was stirred mechanically throughout the bubbling cycle. During the course of the reaction, the liquid became more viscous.

2.4 Treatment of birch wood chips in SIL (DBU-DEA-CO₂-SIL)

For the treatment of birch chips, air dried chips (moisture content ~5%) were charged into a glass bottle. Thereafter, SIL was added to give a wood to SIL weight ratio of 1:5. The sealed glass bottle containing the wood SIL mixture was placed in the oven for 24 h at 100°C; no form of mechanical agitation was applied. After the treatment, the undissolved chips remaining were washed thoroughly with hot water (approximately 60°C) until they were free from SIL. The washed undissolved chips were dried in an oven at 105°C overnight and mass differences were recorded (first treatment cycle). For the second treatment cycle, fresh SIL was used to treat the wood recovered after the first treatment cycle. Thereafter the samples were subjected to various analyses methods.

2.5 X-ray diffraction

Untreated and SIL-treated wood chip samples were analyzed by powder X-ray diffraction (XRD) using a PANalytical X’Pert PRO diffractometer equipped with primary beam Johansson monochromator, to generate pure Cu K_α1radiation (1.5406 Å; 45 kV, 30 mA). Samples were prepared in a steel-made sample holder with a 16 mm radius sample cavity. An X’Celerator detector in a continuous scanning mode was used to collect the data in 2θ range of 3–71° with a step size of 0.017° and a counting time of 200 s per step (overall time of ~2 h). A programmable divergence slit was used in automatic mode to set the irradiated length on the sample to 10 mm, together with a fixed 15 mm incident beam mask. Soller slits of 0.02° rad were used on both incident and diffracted beam sides, together with antiscatter slits of 4° and 8.7 mm, respectively. The diffraction data were converted from the automatic slit mode to the fixed slit mode data in an X’pert Highscore Plus v. 2.2d software package, before further analyses. The ICDD PDF-2 powder diffraction database implemented in Highscore Plus was used for the qualitative phase analyses of the XRD patterns. Average crystallite sizes of the identified structure form were evaluated using the Scherrer equation (shape factor K=0.94).

2.6 Carbohydrate analysis

2.7 Lignin content determination

The lignin content was determined using the Klason lignin method with slight modification. Boiling of the reaction mixture for 4 h to complete hydrolysis of the polysaccharides was replaced with autoclave treatment at 125°C at 1.4 bar for 90 min [15, 16].

3.1 DBU-DEA-CO₂-SIL treatment of birch chips

The weight reduction of the undissolved fraction of the wood chips after SIL treatment was 23% after the first cycle and the weight reduction recorded after the second SIL treatment cycle was 32%. The SIL treatment in this study reduced the wood lignin content by 20% after the first cycle. Then, the lignin content was reduced further by 6%, after the second cycle of SIL treatment. The SIL used for the second cycle was fresh, an indication that the SIL affinity towards lignin was not very high. The trend in the hemicelluloses removal was similar to that for the lignin removal, since the hemicelluloses content (sugars yield) was reduced to 36% lower after two cycles compared to that of the native birch; it was reduced by 18% after the first SIL treatment (Figure 1A–C). Upon addition of methanol to the spent SIL, the solid precipitates recovered were mainly hemicelluloses, of which about 90% was xylose, as expected, since the treated wood was hardwood (Figure 2). A summary of the results is presented in Table 1.

Figure 1

(A) Lignin content; (B) cellulose content (glucose yield); (C) hemicelluloses content (sugar yield) for native birch chip and diethanolamine (DEA)-CO2-switchable ionic liquid (SIL)-treated birch once-treated and twice-treated, using fresh SIL in both treatments at 100°C for 24 h and in the absence of stirring.

Figure 2

Hemicelluloses content (sugar yield) for the recovered material from the spent switchable ionic liquid (SIL) upon addition of methanol for the once-treated and twice-treated, using fresh SIL in both treatments at 100°C for 24 h and in the absence of stirring.

3.2 XRD analysis of birch

Treatment of birch chips using DBU-DEA-CO₂-SIL, resulted in a decrease in the cellulose crystallinity, as well as a transformation of cellulose I (Iα and Iβ) to cellulose II, which is a typical result in IL treatment of biomass/wood, depending on the pretreatment conditions used [3, 17, 18]. As is generally known, the native state cellulose exists as a semicrystalline polymer that has, depending on the origin of the samples, varying fractional distributions of two cellulose structure modifications (allomorphs); triclinic Iα and monoclinic Iβ [19]. Both allomorphs have the cellulose chains aligned along the c-axis, so that the unit cell length of each crystal form is about 10.3−10.4 Å in length in the direction of the chains, and the chains have settled into stacked sheets [20]. The main deviation between the Iα and Iβ forms is related to the relative displacement of these sheets along the c-axis, which is dissimilar due to a different C—H···O bonding network occurring between sheets [21] In contrast to cellulose form I, in form II, polymer chains with opposite polarity are stacked to form corrugated sheets [22]. For higher plants such as wood, the Iβ form is generally the major component, whereas on primitive plants, the Iα form prevails [23, 24] and the cellulose form II exists mainly as a result of pretreatment of native cellulose by, for instance, mercerization and regeneration [25]. Subcritical water treatment [26] and ball-milling of cellulose in water [27] have also been shown to transform cellulose I to cellulose II, at least partially.

The XRD patterns of one untreated and two SIL treated wood chip samples are shown in Figure 3. In all three XRD patterns, two major peaks (at about 15.8° and 22.1° 2θ) along with one weaker one (34.5° 2θ) can be observed. These peaks originate from the cellulose form Iβ having a monoclinic crystal system and space group P2₁ with unit cell settings a=7.784, b=8.201, c=10.38 and γ=96.5° [21].The strongest peak at 22.1° originates from the (200) lattice place and it is indicative for d-spacing between the hydrogen bonded sheets. The second strongest broad peak at 15.8° consisted of at least two overlapping peaks of form Iβ residing at 14.9° (1–10) and 16.7° (110). All of these three peaks represent directions perpendicular to the fiber axis (c-axis), whereas the third weaker peak at 34.5° (004) represents a direction parallel to the fiber axis and is indicative for a quarter of the length of the cellobiose unit along the c-axis. The crystallite diameter and length can be estimated using peak widths of the given peaks [28]. It can also be noted that depending on the origin of the wood sample, less fraction of form Iα may be present, adding its contribution in the observed diffraction intensities, as it has a similar peak distribution at the angular range (14.3, 16.8 and 21.8° 2θ) corresponding to form Iβ. However, as discussed above, the main component of wood is typically cellulose form Iβ.

Figure 3

The X-ray diffraction (XRD) patterns of untreated, once-treated and twice-treated birch chip using diethanolamine (DEA)-CO2-switchable ionic liquid (SIL). Fresh SIL was applied for the first as well as the second treatment at 100°C.

The SIL treatment-induced changes can mainly be observed in the overall crystallinity of the wood chips. By comparing the XRD patterns of untreated and treated samples (Figure 3), it can be seen that a major change occurred on the baseline level between the angular range of 10–20° 2θ, indicating partial removal of the amorphous content, such as amorphous cellulose, lignin and hemicelluloses. The change is most significant between the untreated and once-treated samples, whereas only minor differences exist between once-treated and twice-treated samples. In both the once-treated and twice-treated XRD patterns (slightly more extensively on the latter), peak widths of the cellulose form Iβ seemed to be tapered in to some extent, thus indicating either equal or slightly increased crystal size/crystallinity of cellulose form Iβ. Contradictory literature can be found on this subject, depending on the pretreatment conditions, origin of the cellulosic material (neat cellulose, wood, switchgrass, etc.) and IL type, as in some cases cellulosic materials subjected to the IL treatment resulted in decreased crystallinity [17, 20–22, 24], whereas in some cases, it has been shown to function inversely [2, 29]. In this particular case, one additional observation may suggest that slight crystallization of the crystallite zones may have occurred. By examining the peak position of the main peak at 22.1° in all three patterns, it is obvious that its position is shifted from 22.1° (untreated) via 22.2° (once-treated) to 22.4° (twice- treated) together with a slight narrowing in the peak width. This may indicate that the d-spacing between the sheets has shrunk thereby inducing a somewhat higher crystallinity. Also, a flicker of an increase in intensity and narrowing of the peak at 34.5° may indicate a slight increase in length of the crystallite zones along the fiber axis. The crystal sizes calculated using the Scherrer equation also follow the suggested trend, as the average crystal sizes calculated from the (200) peak are 26, 32 and 38 Å, for untreated, once-treated and twice-treated, respectively. Several groups have demonstrated that, for instance, pure cellulose (e.g., Avicel) and biomasses such as switchgrass treated in IL or conventional solvents undergo structural transformation from cellulose form I to form II at rather modest temperatures and in relatively short treatment times. [2, 28, 29–31]. By contrast, wood chips of, e.g., eucalyptus and pine, require higher temperatures and longer treatment times to show even partial transformation. In our case, no evidence of transformation to form II can be seen, because the most likely observable change of a moderately strong diffraction peak of form II at 12.2° (-110) is absent in both XRD patterns of the treated samples.

Finally, it is noted that when the observed changes are rather modest, it is quite difficult to distinguish whether the tapering of diffraction peak widths is just caused by more coherent scattering from less amorphous samples or true crystallization of the crystallite zones. From the perspective of crystal size evaluations, this “invalid” interpretation originates partly from the difficulty to deconvolute the diffraction/scattering intensities of fully amorphous content and intensities caused by the nanosized cellulose crystallites. The integrated intensities and the determined peak widths are always strongly influenced by the scattering power of the amorphous content that tends to falsely broaden the peaks originating from the crystallites, hence increasing the overall gain of integrated intensity and resulting in lower crystallinity and/or crystallite size for untreated samples. Therefore, the removal of the amorphous part also removes its scattering contribution from the examined peak intensities and widths, causing apparent increases in crystallinity and crystal size, while the crystallites may have remained unchanged. This could be one of the reasons for the contradictions found in reports dealing with crystallization properties of cellulosic materials under analogous conditions.