Introduction

Decades after Rånby (1949) first reported that strong acid treatments could liberate native cellulose monocrystals from cellulosic materials in the form of stiff nanorods, the advent of nanotechnology has stimulated much research and development activities around cellulose nanowhiskers (CNWs) as testifies the number of recent reviews on the topic (Eichhorn et al. 2010; Eichhorn 2011; Klemm et al. 2011). With their low density, renewable nature, high mechanical properties (Young’s modulus around 140 GPa, and strength of 10 GPa) (Sakurada et al. 1962; Nishino et al. 1995; Sturcova et al. 2005; Moon et al. 2011; Rusli et al. 2011), potential for organization into liquid crystalline phases (Araki and Kuga 2001), CNWs are foreseen as the drivers of new technologies in the forest products and bio-based materials industries (Wegner and Jones 2006). For example, applications have been demonstrated in nanopaper, coatings, adhesives, nanocomposites, optical sensors, biomedical scaffolds, filtration membranes, electronic devices, foams and aerogels (Habibi et al. 2010; Eichhorn et al. 2010; Eichhorn 2011). Parameters of importance for their application comprise morphology, crystallinity, aspect ratio (L/d), thermal stability, surface chemistry and propensity for self-assembly. Additionally material homogeneity and in particular uniformity in nanoparticle dimensions is desirable for utilization and processing into end-products. Most commonly, concentrated acid hydrolysis viz. sulfuric, hydrochloric or phosphoric acid of pulp or microcrystalline cellulose (MCC) is used to produce CNWs in yields ranging from 20 to 40 % (Dong et al. 1998; Roman and Winter 2004; Beck-Candanedo et al. 2005; Bondeson et al. 2006; Elazzouzi-Hafraoui et al. 2008; Fan and Li 2012). Dilute acid conditions generate lower yields of CNWs than concentrated acid conditions (Beck-Candanedo et al. 2005; Bondeson et al. 2006; Sadeghifar et al. 2011). As strong acids require corrosion-resistant reactors on scale up and treatment or recycling, challenges for the large scale production of nanowhiskers with this route remain to be addressed (Li and Zhao 2007).

With their swelling potential, possible roles as reactants and catalysts, ionic liquids (IL) show promise for hydrolyzing polysaccharides into nanofibers, as recently demonstrated for chitin (Kadokawa et al. 2011) and cellulose (Man et al. 2011). In both instances an entangled morphology of relatively thick nanofibers (20–60 nm), alike nanofibrillated materials has been obtained rather than nanowhiskers. In contrast Li and Taubert (2009) apparently produced rod-like nanowhiskers-based materials by treating a mixture of HAuCl4 and cellulose with ([Bmim]Cl). Other approaches with IL have involved regeneration of MCC and cotton into cellulose II nanowhiskers (Han et al. 2013). In these studies, solvent recovery, reaction efficiency and mechanism have not been investigated. Here we report an optimized method for the preparation of cellulose I nanowhiskers from microcrystalline cellulose in relatively high yields with aqueous 1-butyl-3-methyl imidazolium hydrogen sulfate [Bmim]HSO4, whereby the whiskers exhibit enhanced properties and the reactant can be recycled.

Experimental

Preparation of cellulose nanowhiskers (CNWs) by IL-mediated hydrolysis

Microcrystalline cellulose (MCC, Avicel PH-101) and 1-butyl-3-methylimidazolium hydrogen sulfate ([Bmim]HSO4) were purchased from Sigma Aldrich. In a first trial the dry IL (4.8 g) was mixed with 0.48 g MCC in 6.86:1 molar ratio in a 50 ml round bottom reaction flask and maintained at 120 °C under reflux and magnetic stirring for 24 h. In a second trial, the IL was pre-mixed with water in [Bmim]HSO4: H2O molar ratio of 1:14.3 mol/mol, reaching pH 1. Then 4.8 g of the IL aqueous solution was mixed with 0.48 g of MCC and the reaction proceeded under similar conditions as in trial 1. For both trials, the reaction product was centrifuged (Thermo Scientific F15 6X100Y Multifuge at 12,000 rpm for 10 min) and washed repeatedly until the supernatant became turbid. The turbid supernatant was dialyzed against DI water until neutral pH (Bondeson et al. 2006). Trial 1 (pure [Bmim]HSO4) yielded lower volumes of turbid supernatant than trial 2 (aqueous [Bmim]HSO4) suggesting that larger amounts of nanoparticles were generated in the latter case. Flow birefringence of the supernatant was observed under cross-polarization filters and reaction yields gravimetrically determined on freeze-dried samples (ALPHA 1 2 LD plus freeze dryer, Christ GmbH, Germany) from four batches.

Recovery and 1H NMR characterization of the IL

The aqueous IL reaction medium was collected from the first transparent supernates and from the washing effluents after filtration of cellulose residues (Nylon filter membranes, pore size 0.45 microns). After removing the water from the collected liquid with rotary evaporation, purity of the recovered IL was assessed with 1H NMR in DMSO-d6 (Bruker Avance 300 Spectrometer, Billerica, MA, USA) and recovery yields were gravimetrically determined.

Morphological characterization of the produced cellulose nanoparticles

Transmission electron microscopy (TEM)

A droplet of the CNW suspension was placed on a carbon-coated copper grid and allowed to dry overnight before observation under a TEM (Zeiss LEO CEM 912) operating at 100 kV accelerating voltage.

Atomic force microscopy (AFM)

A droplet of CNW suspension was allowed to dry overnight on freshly cleaved mica before observation in tapping mode with an AFM (Nanoscope III) equipped with a tube scanner from Digital Instruments (Veeco Santa Barbara, CA USA) using silicon tips (PPP-NCH, Nanoandmore, Germany) with resonance frequency and spring constant of 360 kHz and 50 N m−1, respectively. Height and phase images were analyzed with a V 5.3r1 software and CNW dimensions determined.

Wide angle X-ray diffraction (WAXD)

WAXD measurements were carried out on MCC and CNW powders in transmission mode on a diffractometer STOE (Darmstadt, Germany, STADI-P) using Cu Kα1 as a radiation at 40 kV and 30 mA. Patterns were recorded in the 2θ range of 3°–70° with a 0.02° resolution. Cellulose crystallinity index (CrI) was calculated from the crystalline (I002) and amorphous (Iam) signal intensities, at 2θ = 22.5° and at 2θ = 18°, respectively (Segal et al. 1959):

$$ CrI = \frac{{(I_{002} - I_{am} )}}{{I_{002} }} $$

Surface and thermal properties of CNWs

Sulfur content and water wettability

Free standing CNWs films were prepared by casting a 2 wt% CNWs aqueous suspension on a Petri dish and then drying overnight in a vacuum oven at 60 °C (Rämänen et al. 2012). The surface and cross-section of the cast films were imaged with a scanning electron microscope (JOEL JSM-6300F) operating at an accelerating voltage of 5 kV. Energy dispersive spectrometry (EDS) was conducted on the films with a CAMECA SX100 electron probe micro analyzer (EPMA) equipped with an Oxford Penta FET detector and operating at an accelerating voltage of 8 kV. Water advancing and receding contact angles, θa, and θr, were measured with a goniometer (OCA 20 Data Physics Instruments, Germany) connected with a CCD camera by advancing and receding a 2 μL drop of DI-1 water on the films.

Thermogravimetric analysis (TGA)

Approximately 10 mg of sample (MCC as control and produced CNWs) was heated at 10 °C min−1 up to 600 °C under 20 mL min−1 air flow in a Pyris 1 thermogravimetric analyzer (Perkin Elmer, Germany).

Results and Discussion

Preparation of cellulose nanowhiskers and recovery of the ionic liquid

For both preparation conditions (pure [Bmim]HSO4 and aqueous [Bmim]HSO4), flow birefringence of the turbid supernatant (Fig. 1a-b) was evidenced, suggesting the presence of anisotropic rod-like cellulose nanowhiskers (Marchessault et al. 1959); furthermore the colloidal suspension was stable for several months (Fig. 1d). Birefringence was stronger when the aqueous ionic liquid was used, indicating a more efficient production of nanowhiskers in aqueous conditions. Further investigations therefore focused on the products of the aqueous IL. The rod-like morphology of the produced cellulose particles was confirmed by TEM (Fig. 1c). In this case, reaction yields and IL recovery reached 48 ± 2 % and ca. 90 %, respectively. Purity of the recovered IL was ascertained with 1H NMR. (Fig. 2). Upon use and recovery of the IL, the 1H NMR spectrum is unchanged with chemical shifts at δ 0.90 (3H, t), 1.23–1.31 (2H, m), 3.87 (3H, s), 4.18 (2H, t), 7.72 (H, s), 7.79 (H, s), 9.15 (H, s) and signal integration and splitting pattern are consistent with those of pure BMIM (Fig. 2). In both fresh and recovered IL the 1H signal for the hydrogen sulfate, expected around 5.1 ppm, is too weak to be observed (SDBS Web 2013). The activity of the recycled IL was confirmed from the maintained high yield of whisker production upon reuse.

Fig. 1
figure 1

Birefringent patterns of aqueous suspensions of CNWs observed between cross-polarizers obtained by a pure ionic liquid [Bmim]HSO4 and b the aqueous ionic liquid [Bmim]HSO4/H2O; c transmission electron micrograph (TEM) of the CNWs obtained after treatment of the MCC with the aqueous [Bmim]HSO4 d colloidal suspension of the CNWs obtained with the aqueous [Bmim]HSO4 still stable after 6 months of storage

Full size image
Fig. 2
figure 2

1H NMR spectra of the original and recycled [Bmim]HSO4

Full size image

Morphology of IL-mediated whiskers

The AFM images confirmed the production of rod-like nanoparticles from the treatment of MCC with aqueous IL (Fig. 3a, b). This contrasts with the microscopic images previously reported (Man et al. 2011), in which a mixture of nanofibrillated cellulose and whiskers is apparent after treatment with the pure ionic liquid. As expected, the addition of water to the IL system favors the production of cellulose nanowhiskers rather than the peeling off of cellulose nanofibrils. Close examination of the CNW surface topography by AFM height measurements reveals that next to the typical flat surface (Fig. 3c), a minute amount of whiskers exhibit an undulated surface (Fig. 3d), similar to that recently observed after dilute sulfuric acid hydrolysis and ascribed to cellulose II re-crystallization (Sèbe et al. 2012). As these structures are observed in minute amounts (<1 %), swelling/regeneration of the native cellulose with the aqueous IL appears practically insignificant. AFM height measurements (500 datapoints) yielded CNW lateral dimension or number average width—MCC-derived whiskers are expected to have a square cross-section (Elazzouzi-Hafraoui et al. 2008). Whiskers width and length distributions were slightly asymmetrical (Fig. 4), as previously reported with concentrated sulfuric acid hydrolysis (Elazzouzi-Hafraoui et al. 2008). Width ranged from 1 to 10 nm and averaged 3.6 ± 1.8 nm (Fig. 4) demonstrating a good uniformity of lateral size. Length, at 146.8 ± 62 nm, was on average smaller and as broadly distributed as that obtained with the optimized sulfuric acid route (Fig. 4). Note however that whiskers’ length can likely be adjusted by optimizing the reaction conditions in the aqueous IL route. With WAXD it was demonstrated that CNWs exhibited similar diffraction peaks as MCC at 2θ equal 15.1, 16.4, 20.6, 22.5 and 34.6° corresponding to the crystallographic planes (1–10), (110), (012), (200) and (004) of cellulose I, respectively (Fig. 5) (Nishiyama et al. 2002, Nishiyama et al. 2003). Perhaps a more intense shoulder at 20.3° of 2θ corresponding to the (012) crystallographic reflections of cellulose II can be discerned in the CNW diffraction pattern (Fig. 5). However the absence of a cellulose II peak around 12° of 2θ rules out cellulose regeneration into cellulose II as a significant contribution. More likely minor surface rearrangements might have been promoted by swelling as proposed in other dilute acid systems (Sèbe et al. 2012). The CrI of the nanowhiskers was in the expected range (77 %) although smaller than for MCC (87 %), perhaps due to swelling in the reaction medium and repeated centrifugation, processes known to reduce cellulose crystallinity index (Modi et al. 1963; Tang et al. 1996; Roman and Winter 2004). Overall it was confirmed that treatment of MCC with aqueous [Bmim]HSO4 predominantly liberates cellulose I monocrystals, in a similar fashion than treatment with concentrated acid but with the particular advantage to occur in higher yields and under mildly acidic conditions.

Fig. 3
figure 3

AFM topography (a) and phase (b) images of CNWs obtained after treatment of MCC with the aqueous ionic liquid; and height measurements along single whiskers demonstrating flat (c) and ribbon-like (d) cellulose nanowhiskers

Full size image
Fig. 4
figure 4

Widths and lengths distributions evaluated from AFM images of CNWs obtained after treating MCC with the aqueous ionic liquid

Full size image
Fig. 5
figure 5

X-ray diffraction patterns of a the original MCC and b CNWs produced after treatment of the MCC with the aqueous ionic liquid

Full size image

Surface and thermal properties of CNWs and film forming abilities

Casting an aqueous suspension of the nanowhiskers yielded 5 μm thick films with high transparency as expected for nanosized cellulose particle films (Fig. 6) (Yano et al. 2005). Under SEM, the cross-section of the film had a layered, sheet-like porous structure similar to that of nanopaper (Sehaqui et al. 2011).

Fig. 6
figure 6

Transparent film obtained by casting the CNWs suspension produced from the aqueous ionic liquid route (left) and its cross-section by SEM (right)

Full size image

With EDS small amounts of sulfur, S (0.46 wt%), corresponding to a DS of 0.1 were evidenced (Table 1) as expected from surface esterification of cellulose into sulfate esters (SO4 2−). The measured sulfur content is slightly below that reported for whiskers after hydrolysis with concentrated sulfuric acid (S: 0.53–0.75 wt%) (Dong et al. 1998) or with anhydrous IL (0.68 wt%) (Man et al. 2011). This confirms however that the reaction mechanism with the aqueous [Bmim]HSO4 is similar to that with concentrated sulfuric acid. Gratifyingly nitrogen was not detected in the samples, demonstrating that purification had efficiently removed all [Bmim]HSO4. The relatively low degree of sulfation in this system suggests that a lower proportion of glycosidic bonds are cleaved compared to the concentrated acid method. The thermal stability of cellulose nanowhiskers is critical for further processing into nanocomposites. It is also well known to decrease with the content of sulfate ester groups (Roman and Winter 2004). In Fig. 7, the TGA thermogram of the whiskers exhibit significant differences compared to that of MCC. While one-step degradation in the 300–400 °C temperature window is observed for MCC, the IL-derived whiskers experience a two-stage degradation. The first stage within the 200–300 °C temperature window exhibits a maximum rate at 285 ± 1 °C, while the second stage between 300 and 400 °C reaches its maximum rate at 346 ± 2 °C. A two-stage degradation process is in line with the thermal behavior reported for whiskers prepared using concentrated sulfuric acid (Roman and Winter 2004, Wang et al. 2007, Yue et al. 2012), with the notable difference that thermal degradation starts at much higher temperature for the IL-derived whiskers. In other words, the whiskers prepared with the aqueous IL are more thermally stable than those prepared by concentrated sulfuric acid. For example, for bacterial cellulose whiskers a first peak degradation temperature was reported between 226° and 262° (Roman and Winter 2004), that is 25°–60° lower than in the present study. As the first stage of degradation relates to the amorphous and esterified cellulose fraction, the higher thermal stability observed for the IL-derived whiskers likely stems from the smaller extent of surface derivatization in line with the lower sulfur content. Similarly, the second degradation temperature appears much higher (346 ± 2 °C) for the IL-derived whiskers in this study compared to earlier reports for bacterial CNWs, in the 250°–300° range (Roman and Winter 2004). This is however likely related to the inherent differences in thermal stability of the starting cellulose materials.

Table 1 Elemental composition of cellulose nanowhiskers obtained by aqueous [Bmim]HSO4
Full size table
Fig. 7
figure 7

Thermogravimetric (TG) and DTG curves of MCC and CNWs produced after treatment of MCC in aqueous ionic liquid (heating rate of 10 °C/min)

Full size image

The advancing and receding contact angles averaged 39.6 ± 3.3° and 7.3 ± 1.6°, evidencing a large contact angle hysteresis (32.3°). These values are higher than those previously reported for cellulose nanowhisker films. For example advancing contact angles in the 17°–24° range were reported for 100 s of nanometer thick films (Eriksson et al. 2007, Aulin et al. 2009), and at 29.2° (wetting hysteresis of 13.7) for free standing 100 micron thick cellulose I films (Edgar and Gray 2003). While allomorph type and film thickness are known to increase contact angles, this could not explain the high contact angle and hysteresis observed in this study, in which thin, cellulose I films were produced. Therefore, the higher contact angles and wetting hysteresis observed for the IL-derived whiskers indicate that these are less hydrophilic and exhibit greater chemical/topographical heterogeneity. This observation is consistent with the slightly lower amount of surface sulfonic groups and thus surface charges of the ionic liquid-derived whiskers. In addition, root mean square roughness of the whisker films increased from 4.7 ± 0.9 nm to 8.2 ± 0.8 nm upon enlarging the probe surface area from 1 to 25 microns with AFM. This confirms that the IL-derived whiskers exhibit higher surface roughness and heterogeneity compared to those prepared by concentrated sulfuric acid (Eriksson et al. 2007, Aulin et al. 2009).

Discussion: advantages and disadvantages of the aqueous IL route and insights on the hydrolysis mechanism

In the present route nanowhisker yields averaged 48 ± 2 % over four batches, significantly above those obtained from MCC with the commonly used concentrated acid route (Table 2). Of particular interest, this favorable yield is obtained under very mild conditions; the acid to anhydroglucose molar ratio, [H+]/[AGU] is 0.24 mol/mol (pH 1) in the aqueous IL route, that is 2–3 orders of magnitude lower than that with strong acid hydrolysis (Table 2). Altogether this indicates that this aqueous IL-mediated route allows a more efficient hydrolysis reaction. Higher yields under milder acidic environment together with the ability to recycle the IL are clear advantages of this preparation route. On the other hand, the long reaction time and high reaction temperature compared to those needed with strong acid hydrolysis, reveal that this reaction requires high energy. The reaction conditions are currently under optimization in our laboratory with the goal to diminish the reaction energetic requirements. The IL-produced whiskers exhibit favorable properties; they have a more uniform distribution of width and are generally thinner albeit shorter than those obtained with concentrated acid, reaching a favorable aspect ratio of 41 (Table 2). The lower extent of surface sulfation from the IL route affords whiskers with enhanced thermal stability; meanwhile, the produced suspension displays good stability. Overall, the aqueous IL route is a promising alternative to the concentrated acid approach in terms of whiskers quality, chemical recovery and corrosiveness, but requires optimization in terms of energetic needs. This however warrants further study of the reaction mechanism and reasons behind the reaction efficiency.

Table 2 Reaction conditions for extracting native cellulose whiskers from cellulosic sources and properties of resulting whiskers
Full size table

The acid hydrolysis of cellulose is a heterogeneous process. It involves the diffusion of acid into the cellulose fibers and the subsequent cleavage of glycosidic bonds (Dong et al. 1998), which first occurs at a fast rate in the amorphous fraction and then at a much slower rate on the reducing end and surface of the residual crystalline regions (Millet 1954). Expectedly then, treating cellulose with a swelling agent prior to acid hydrolysis increases reactivity while decreasing its crystallinity (Modi et al. 1963). As cellulose is not soluble in dilute acid, the heterogeneous hydrolysis in this medium may explain the low success (low yield) to produce nanowhiskers under mild acidic hydrolysis conditions (Dong et al. 1998; Li and Taubert 2009; Fan and Li 2012). In contrast, the IL-mediated route, also a mildly acidic route, allows a more efficient acid catalysis and the generation of smaller and more homogeneous nanocrystals. The solvating power of the aqueous IL system in comparison to that of concentrated acid might be one reason for the better reaction efficiency with the aqueous IL. To test this hypothesis, we compared cellulose swelling/solvation behavior (and pulp fibers- see S-1 Supporting information) in the aqueous IL and in concentrated sulfuric acid. Drastic differences in the solvation power of the aqueous IL and concentrated sulfuric acid are observed (Fig. 8). While the aqueous IL slowly swells MCC eventually causing some cracks perhaps the indication of the onset of a slow dissolution or degradation process (Fig. 8b, S1), in concentrated sulfuric acid MCC readily disintegrates (Fig. 8c, S1). The solvation power of the aqueous IL is alike a Mode 4 swelling/dissolution mechanism for cellulose (Cuissinat and Navard 2006; Cuissinat et al. 2008). Namely, Mode 4 occurs for poor solvents, which induce homogeneous swelling but no dissolution of the fiber. In contrast, fast disintegration of MCC observed on the microscopic scale together with the ballooning in concentrated sulfuric acid (S-1) suggest a Mode 1 swelling/dissolution mechanism, that is a fast dissolution by disintegration into rod-like fragments or a Mode 2 mechanism, that is a large swelling by ballooning followed by dissolution of the whole fiber. In concentrated sulfuric acid (S-1) the dissolution process, being through Mode 1 or Mode 2, is extremely fast; within 2 min the structure of the pulp fibers and of MCC disappears. With the aqueous IL the homogeneous swelling and slow (if at all) dissolution suggest that swelling is distributed throughout the MCC and fiber structure, locally enhancing the accessibility to the amorphous sub-micron domains for H+ catalyzed hydrolysis, while perhaps better preserving the crystalline domains. Such a mechanism would explain the improved efficiency of the few protons available in the system and the obtention of more homogeneously distributed nanocrystals. The swelling observed is also consistent with the noted crystallinity decrease upon hydrolysis and the minute amounts of possible cellulose II allomorphs. In this system, water plays the critical role of inducing H+ release necessary for the hydrolysis reaction but might also serve to control solvation power.

Fig. 8
figure 8

Real-time observations by optical microscopy of dry MCC (a) and MCC upon immersion in aqueous [Bmim]HSO4 (b) and 64 wt% H2SO4 (c) at room temperature for 3 min. The color scales for all images were normalized to provide the proper contrast

Full size image

Conclusions

An aqueous ionic liquid system based on 1-butyl-3-methylimidazolium hydrogen sulfate ([Bmim] HSO4) allowed producing cellulose nanowhiskers in relatively high yields (48 ± 2 %) and under mildly acidic conditions. In addition, the IL liquid could be recycled and maintained its activity upon reuse. Compared to whiskers obtained by concentrated sulfuric acid hydrolysis, the IL-mediated nanowhiskers presented lower amounts of sulfonic groups, significantly higher thermal stability and could be cast in transparent layered films of slightly lower hydrophilicity. Microscopic observations of the solvating power of the aqueous IL system suggested the importance of the swelling behavior on reaction efficiency. We propose that the IL-mediated preparation route can be further optimized by decoupling the swelling and hydrolysis steps in the production of cellulose nanowhiskers from cellulose. Such approaches are currently investigated in our laboratory with the goal to further improve reaction yield while decreasing the energetic requirements of reaction.