The role of cyclodextrin-tetrabutylammonium complexation on the cellulose dissolution

Carbohydrate Polymers 140 (2016) 136–143

Contents lists available at ScienceDirect

Carbohydrate Polymers journal homepage: www.elsevier.com/locate/carbpol

The role of cyclodextrin-tetrabutylammonium complexation on the cellulose dissolution Bruno Medronho a,∗ , Hugo Duarte a , Luis Alves b , Filipe E. Antunes b , Anabela Romano a , Artur J.M. Valente b a b

Faculty of Sciences and Technology (MEDITBIO), University of Algarve, Campus de Gambelas, Ed. 8, 8005-139 Faro, Portugal CQC, University of Coimbra, Department of Chemistry, 3004-535 Coimbra, Portugal

a r t i c l e

i n f o

Article history: Received 25 September 2015 Received in revised form 3 December 2015 Accepted 10 December 2015 Available online 24 December 2015 Keywords: Cellulose dissolution Tetrabutylammonium hydroxide Hydrophobic interactions Cyclodextrins Host–guess complex

a b s t r a c t Cellulose dissolution is a challenging process which is typically very sensitive to the solvent characteristics such as pH, temperature or presence of additives. Regarding the later aspect, it is here reported the interaction between ␣-cyclodextrin (␣-CD) and ␤-cyclodextrin (␤-CD) with the tetrabutylammonium cation (TBA+ ) by 1 H NMR titration experiments. The analysis by the continuous variation method suggests the formation of 1:1 CD:TBA+ complexes. However, the computed apparent association constants reveal that the interaction of TBA+ with the ␤-CD (K = 1580 M−1 ) is unexpectedly stronger than with ␣-CD (K = 106 M−1 ). In both CD cases, the formation of CD:TBA+ complexes decrease the dissolution efficiency of the solvent and this has been rationalized as an effective decrease in the concentration of the amphiphilic cation and concomitant weakening of the hydrophobic interactions in solution influencing the overall performance of the solvent. Additionally, the data also supports the fact that amphiphilic species in solution are beneficial for the enhancement of cellulose solubility. © 2015 Elsevier Ltd. All rights reserved.

1. Introduction Cellulose is structurally an anisotropic molecule displaying amphiphilic properties (Biermann, Hadicke, Koltzenburg, & MullerPlathe, 2001; Diddens, Murphy, Krisch, & Müller, 2008; Medronho et al., 2015; Medronho & Lindman, 2014b; Miyamoto et al., 2009; Yamane et al., 2006). Such amphiphilicity has not always been recognized and its role in cellulose solubility is often neglected in detriment to the more conceptual view which attributes to hydrogen bonding the leading responsibility in the cellulose solubility pattern (Glasser et al., 2012; Lindman, Karlström, & Stigsson, 2010; Medronho & Lindman, 2014a; Medronho, Romano, Miguel, Stigsson, & Lindman, 2012). Cellulose solvents are of vast nature regarding the composition, pH, operating temperature, etc. However, the majority are limited to lab scale applications and share similar problems and concerns such as the costs of production, safeness, recyclability, environmental impact and capacity of dissolution (Heinze & Koschella, 2005; Medronho & Lindman, 2014b). Among them the alkaline systems based on NaOH have claimed more attention in a hypothetical replacement of more hazardous

∗ Corresponding author. Tel.: +351 289800910; fax: +351 289818419. E-mail address: [email protected] (B. Medronho). http://dx.doi.org/10.1016/j.carbpol.2015.12.026 0144-8617/© 2015 Elsevier Ltd. All rights reserved.

routes used nowadays worldwide such as the viscose process (Cai & Zhang, 2005; Isogai & Atalla, 1998). The NaOH aqueous system draws attention mainly due to its simplicity and inexpensive chemical composition, low environmental impact and an apparent easy implementation in the existing large scale facilities. The use of different additives has further improved the dissolution efficiently on NaOH aqueous solutions (i.e. increasing the kinetics of dissolution and stability of the dopes formed) (Cai et al., 2007; Liu, Budtova, & Navard, 2011a; Wan & Li, 2015; Yan & Gao, 2008; Zhang, Ruan, & Gao, 2002). However, some important limitations still persist in these systems such as the sub-zero temperatures needed to efficiently dissolved cellulose and the rather low molecular weights (i.e. DP < 400) and cellulose contents (i.e. <10 wt%) that the solvents are capable to handle. An interesting alternative alkaline system is the aqueous tetrabutylammonium hydroxide solvent (TBAH) (Abe, Fukaya, & Ohno, 2012). This system is capable to dissolve high amounts of cellulose with high molecular weight at room temperature reasonably fast. Additionally, while cellulose in the TBAH appears to be dissolved down to the molecular level, the dissolution in the analogous NaOH aqueous system is far of being complete where fairly large and crystalline colloidal aggregates are found to be stable in solution. This difference in the dissolution performance was suggested to arise from the amphiphilicity of the tetrabutylammonium (TBA+ ) cation which supposedly facilitates

B. Medronho et al. / Carbohydrate Polymers 140 (2016) 136–143

the dissolution of cellulose via the weakening of the hydrophobic interactions (Alves et al., 2015). In this work the effect of two different natural (␣- and ␤-) cyclodextrins (CDs) on the aqueous TBAH performance is studied. The question of how CDs interact and affect the availability of the TBA+ specie and how that influences the cellulose dissolution will be addressed by means of NMR methodologies. Further X-ray, infrared and microscopy studies elucidate the effect of CDs in the dissolution performance of the solvent and some thermodynamic parameters will be discussed. 2. Materials and methods 2.1. Chemicals and solution preparation Dissolving pulp was supplied from Buckeye Tech. Inc., USA with an averaged degree of polymerization (estimated by intrinsic viscosity measurements) of ca. 630. Tetrabutylammonium hydroxide (TBAH) of chromatographic grade (40 wt% solution in water), ␣-cyclodextrin (>98%), ␤-cyclodextrin (>98%), 3(trimethylsilyl)propionic-2,2,3,3-d4 acid sodium salt and urea were purchased from Sigma-Aldrich. All chemical were used as received. The NMR samples were prepared using D2 O (99.8%) supplied by EURISO-TOP (France). NaOD was also supplied by EURISO-TOP (France). The cellulose pulp solutions were prepared by simply mixing a known amount of cellulose with a 40 wt% TBAH/H2 O solution at room temperature for 30 min (Abe et al., 2012). 2.2. Characterization 2.2.1. X-ray diffraction (XRD) The X-Ray diffraction experiments were performed on a Siemens D5000 X-ray diffractometer, capable of identifying crystalline phases down to 3% of the bulk. This equipment consists of a /2 diffraction instrument operating in the reflection geometry. ˚ focused by a CuK␣1 is used as radiation source with  = 1.54056 A, primary Ge crystal monochromator. The detector is a standard scintillation counter. The Cu tube runs at 40 mA and 40 kV. The cooling is supplied by an internal water-filled recirculation chilling system, running at approximately 16 ◦ C with a flow rate of 4–4.5 L/min. The slit arrangement is a 2 mm pre-sample slit, 2 mm post-sample slit and a 0.2 mm detector slit. 2.2.2. Fourier transform infrared spectroscopy (FTIR) The infrared spectra were recorded at 25 ◦ C with a ATR-FTIR spectrophotometer Thermo Nicolet, IR300 (USA), using a universal ATR sampling accessory. FTIR spectral analysis was performed within the wave number range of 400–4000 cm−1 . A total of 256 scans were run to collect each spectrum at a resolution of 1 cm−1 in the transmission mode. The CrI (also referred to as “lateral order index”—LOI) was estimated from the ratio between the absorption band at 1430 cm−1 and the absorption band at 890 cm−1 (Hurtubise & Krasig, 1960; O’Connor, DuPre´ı, & Mitcham, 1958). 2.2.3. Turbidimetry A T70 UV–vis spectrophotometer (PG Instruments Ltd) was used for the optical transmittance measurements. Essentially, the cellulose solutions were placed in a proper cell and the transmittance was measured at a wavelength of 600 nm. 2.2.4. Polarized light microscopy A Linkam LTS 120 microscope equipped with a Q imaging (Qicam) Fast 1394 camera was used to observe the cellulose dissolution in the solvent systems used. Samples were kept between

137

cover slips and illuminated with linearly polarized light and analysed with a linear polarizer under crossed position. Images were captured and analyzed using Qcapture software. 2.2.5. 1 H NMR measurements 1 H NMR spectra were recorded at 25.0 (±0.1) ◦ C on a Varian 500 MHz spectrometer using a 5 mm NMR probe. Spectra were obtained with residual solvent (HOD) presaturation and the acquired parameters included 24k data points covering a spectral width of 8 kHz, a radiofrequency excitation pulse of 45◦ and a scan repetition time of 15 s to allow for full magnetic relaxation of proton nuclei. The resonance at 0 ppm due to Si (CH3 )3 signal, from 3(trimethylsilyl)propionic-2,2,3,3-d4 acid sodium salt (TSP) at tracer concentration (below 3 ␮M), was used as internal reference. The method of continuous variation has been used to determine the stoichiometry of the CD:TBAH interaction; for that, samples were prepared by dissolving an amount of the solid in D2 O to achieve a concentration of CD and TBAH of 1.01 and 1.02 mM, respectively. The TBAH was previously lyophilized overnight just before using and the pH* of D2 O was adjusted to 12.08 with the addition of NaOD, in order to keep the pH of CD:TBAH solutions constant. The binding constant were computed by using experimental 1 H NMR chemical shifts of TBA+ nuclei for mixed solutions with different [␣-CD]/[TBAH] and [␤-CD]/[TBAH] molar ratios, and keeping the [TBAH] constant and equal to 1.13 and 1.20 mM, respectively. 3. Results and discussion 3.1. Characterization of the dissolved and regenerated cellulose pulp: Microscopy, turbidimetry, XRD and FTIR analysis Cellulose dissolution is an intriguing process but essential in numerous applications due to the fact that this biopolymer does not melt. The addition of different additives (typically with amphiphilic like properties) has been observed to enhance cellulose dissolution (Medronho & Lindman, 2014b). On the other hand, simple salts tend to compromise the dissolution performance (Medronho et al., 2015). For instance, in the TBAH system it has been observed that alkali metal ions (e.g. LiCl, NaCl and KBr) when added to the solvent decrease its dissolution performance (Ema, Komiyama, Sunami, & Sakai, 2014). Ema et al. suggest an interpretation based on the salt effect on the disruption of the hydrogen bonding network. On the other hand, we believe that at such extreme high pH cellulose is, if not fully, at least, partially protonated. Therefore, the addition of simple salts is expected to screen the electrostatic repulsion between the charged cellulose chains. As cellulose molecules become less hydrophilic with the salt addition, the hydrophobic interactions start to dominate triggering the chain aggregation. In Fig. 1, polarizing optical micrographs are shown for 2 wt% cellulose dissolved in TBAH with different additives. Regardless the concentration of urea added, no visible effect on the dissolution in observed (Fig. 1b and c) when compared with cellulose dissolved in the pure solvent (Fig. 1a). Some substances, such as surfactants and urea, have intermediate polarity and, as recently reviewed, are known to eliminate hydrophobic association in aqueous media (Medronho & Lindman, 2014b). It is here noted that urea increases the kinetics of cellulose dissolution and slightly decreases the viscosity of the dopes. On the other hand, the addition of low concentrations of ␣-CD (Fig. 1d) and ␤-CD (Fig. 1g) does not change considerably the dissolution efficiency although a residual fraction of undissolved fibers are observed. The typical initial swelling step preceding dissolution, the formation of “balloons” (Cuissinat, Navard, & Heinze, 2008), is also clearly visible for the lower concentrations of CD. On the other hand, if the CD concentration is further increased no balloons are observed and the fraction of undissolved

138

B. Medronho et al. / Carbohydrate Polymers 140 (2016) 136–143

Fig. 1. Optical micrographs (cross-polarization mode) of 2 wt% cellulose pulp in different solvents at 25 ◦ C. (a) 40 wt% TBAH/H2 O, (b) 40 wt% TBAH/H2 O + 2 wt% urea, (c) 40 wt% TBAH/H2 O + 20 wt% urea, (d) 40 wt% TBAH/H2 O + 2 wt% ␣-CD, (e) 40 wt% TBAH/H2 O + 10 wt% ␣-CD, (f) 40 wt% TBAH/H2 O + 20 wt% ␣-CD, (g) 40 wt% TBAH/H2 O + 2 wt% ␤-CD, (h) 40 wt% TBAH/H2 O + 10 wt% ␤-CD, (i) 40 wt% TBAH/H2 O + 20 wt% ␤-CD. The micrographs were taken 30 min after mixing the solvent with cellulose.

fibres increases considerably (Fig. 1e and h). Finally, for the highest concentration of CDs used (10 and 20 wt%, Fig. 1f and i), almost no dissolution is detected and the micrographs exhibit a high density of birefringent undissolved cellulose fibers. No improvement in the dissolution is observed for at least two weeks. Simultaneously to the polarized light microscopy, the turbidity of the samples was also measured. While urea has no perceptible influence on the transparency of cellulose dopes (Fig. 2a), the progressive addition of CDs turns the solutions turbid (Fig. 2b). At low concentrations of CDs samples are still highly transparent and TBAH is still a good solvent for cellulose pulp while at higher amounts of CDs TBAH is a poor solvent for cellulose pulp and the dopes are turbid. Although the difference in transmittance between ␣CD and ␤-CD is rather small the transition from a “good-to-bad” solvent is anticipated for ␤-CD suggesting that ␤-CD is slightly more efficient in decreasing the quality of the solvent. In principle, from a mechanistically point of view, this would mean a stronger

interaction between ␤-CD and the solvent and this hypothesis will be later discussed with the NMR data. So far it has been seen that when CDs are added the quality of the dissolution is worsened, in agreement with our previous report on similar systems (Medronho et al., 2013). This can also be inferred from the properties of the regenerated product, that is, analysing the coagulated material after being dissolved. In Fig. 3 the X-ray diffraction patterns of regenerated cellulose in water are shown for both CDs and urea. Regardless the concentration of urea used, the regenerated material always exhibits a polymorph form resembling the cellulose II crystalline structure (Fig. 3a): a 1 0 −1 (ca. 20.1◦ ) peak overlapping with a 0 0 2 (ca. 21.9◦ ) peak, and a separation between 1 0 −1 and 1 0 1 (ca. 12◦ ) peaks (Liang, 1972; p 59). On the other hand, the native cellulose material displays a characteristic cellulose I diffraction pattern with a primary 0 0 2 (ca. 22.5◦ ) lattice plane peak and a secondary overlapped 1 0 1 (ca. 14.9◦ ) and 1 0 −1 (ca.

Fig. 2. Turbidimetry measurements of 2 wt% cellulose pulp dissolved in the 40 wt% TBAH/H2 O solvent doped with urea (left) and cyclodextrins (right). The gray (␤-CD) and black (␣-CD) curves represent the sigmoidal fits of the data.

B. Medronho et al. / Carbohydrate Polymers 140 (2016) 136–143

Fig. 3. XRD spectra of regenerated cellulose pulp in water after being previously dissolved in (a) 40 wt% TBAH/H2 O + urea, (b) 40 wt% TBAH/H2 O + ␣-CD and (c) 40 wt% TBAH/H2 O + ␤-CD. The concentrations of additives are 2 wt% (black curves), 10 wt% (dark gray curves) and 20 wt% (light gray curves). The dashed back curve on the top of figure (a) represents the XRD spectrum of the native undissolved pulp.

139

16.2◦ ) peaks (Liang, 1972; p 59). Another characteristic reflection from cellulose I type structure can be found at 34.7◦ (0 4 0). The regenerated cellulose materials from the solutions doped with CDs exhibit similar features (Fig. 3b and c): At low CD concentrations the cellulose dissolution is feasible and therefore the regenerated material presents typical Bragg diffractions patterns (black curves) for a cellulose II crystalline organization with 2 angles similar to the ones in Fig. 3a. On the other hand, the progressive increase of CD decreases the solvent quality and the regenerated materials displays a characteristic cellulose I diffraction pattern similar to the native cellulose material. In Fig. 4, FTIR spectra are displayed highlighting the regions of interest for cellulose pulp. All spectra are very similar presenting the characteristic vibration bands for a cellulosic material regardless the presence or not of CDs and urea (Poletto, Pistor, Santana, & Zattera, 2012; Siroky, Blackburn, Bechtold, Taylor, & White, 2010). On the high wavenumber region (Fig. 4a) the broad absorption band in the range 3100–3600 cm−1 can be ascribed to the stretching of the OH groups (with typical sharpening at 3300 cm−1 ) (Jahan, Saeed, He, & Ni, 2011; Sun, Sun, Liu, Fowler, & Tomkinson, 2002) while the peak at 2900 cm−1 appears due to C H stretching (Jahan et al., 2011; Sun et al., 2002). On the other hand, on the low wave number region (Fig. 4b) an intense band between 1600 and 1650 cm−1 originates from the absorbed moisture (i.e. bending mode of water absorbed to cellulose) (Ren, Sun, Liu, Chao, & Luo, 2006). The deformation, wagging and twisting modes of anhydroglucopyranose vibration are shown from 600 to 1800 cm−1 . More specifically, the absorbance at around 900 cm−1 can be assigned to the C H deformation mode of the glycosidic linkage between the glucose units (Kacurakova, Ebringerova, Hirsch, & Hromadkova, 1994; Liu, Ni, Fatehi, & Saeed, 2011b), while the absorbance bands between 1000 cm−1 and 1200 cm−1 are attributed mainly to the C-O stretching in major ether bands (Liu et al., 2011b). It is possible to assign specific absorption bands to crystalline and amorphous regions and from it estimate a crystallinity index (CrI). O’Connor et al. (1958) established that the absorption band at around 1430 cm−1 is characteristic of crystalline areas in the polymer and the absorption band at 890 cm−1 typical of amorphous regions; the ratio of these two bands was established as a “crystallinity index”, later referred to as the “lateral order index”. This is one of the simplest methods used and thus one should keep in mind that the extracted values are not absolute and comparisons between CrI of different polymorphs should be avoided. The regenerated samples which a cellulose II crystal organization have about the same CrI, ca. 0.6, while the cellulose I samples present a slightly higher CrI, when compared with the native cellulose samples, (ca. 0.8) after being dispersed in the solvent system. This higher CrI might be related to some partial dissolution of the amorphous areas

Fig. 4. FTIR spectra of regenerated cellulose pulp in water after being previously dissolved in 40 wt% TBAH/H2 O (dashed black curve), 40 wt% TBAH/H2 O + 2 wt% urea (dark gray curve), 40 wt% TBAH/H2 O + 2 wt% ␣-CD (gray curve) and 40 wt% TBAH/H2 O + 2 wt% ␤-CD (light gray curve). The full black curve represents the FTIR spectrum of the native undissolved pulp. The high wavenumber region is displayed in figure (a) while the low wavenumber region is displayed in figure (b).

140

B. Medronho et al. / Carbohydrate Polymers 140 (2016) 136–143

that occurs despite the low quality of the solvent and, overall, the CrI increases. Additionally, it is also possible that after the swelling of the fibers, and during the coagulation and drying, some disordered areas eventually crystallize, thus enhancing CrI. 3.2. The interaction between CD and TBAH: 1 H NMR analysis Urea has no visible effects on the cellulose pulp dissolution in TBAH while CDs clearly influence the dissolution performance of the solvent. Additionally, the turbidity measurements suggested a preferential interaction between ␤-CD and the solvent molecules (i.e. TBA+ cation). In this section NMR is explored to clarify the interaction between TBA+ and CDs. Fig. 5 shows representative 1 H NMR spectra of TBAH (a), ␤-CD (b) and a mixed ␤-CD:TBAH (c) solution, at pH* equal to 12.08. It can be seen that the 1 H NMR spectra for ␤-CD and TBAH are similar to those previously reported (Carvalho, Correia, Valente, Soderman, & Nilsson, 2011; Schneider, Hacket, Rudiger,

Fig. 5. 1 H NMR spectra of solutions of (a) TBAH, 1.02 mM; (b) ␤-CD, 1.01 mM; and (c) TBAH and ␤-CD, xCD = 0.5.

& Ikeda, 1998; Wang, Messman, Mays, & Baskaran, 2010). Briefly, the 1 H NMR spectrum of ␤-CD in D2 O (Fig. 1a) shows (from upfield to downfield) a triplet (assigned to the H4 protons) at ıH4 = 3.534 ppm and a doublet of doublets assigned to H2 protons at ıH2 = 3.596 ppm. These protons are located outside the cavity of CD, near the narrow and wide sides, respectively. The H5 and H6 (including the anomeric H6 ) protons overlap at the chemical shift around 3.83 ppm. The H3 protons, located inside the cavity, at the wide side, show a triplet at ıH3 = 3.920 ppm. The doublet located at ı = 5.025 ppm is assigned to H1 protons, which are located outside of CD cavity and in between H4 and H2 protons. These chemical shifts show a slight downfield displacement when compared with those ones obtained for non-buffered CD solutions (Medronho, Valente, Costa, & Romano, 2014). The 1 H NMR spectrum of TBAH (Fig. 5b) shows a triplet assigned to methyl group (H␦ ) at 0.955 ppm, and the 1 H assigned to methylene groups show resonances at 1.370, 1.660 and 3.206 ppm for H␥ , H␤ and H␣ , respectively. The 1 H NMR spectrum of a ␤-CD:TBAH mixed solution, with a molar ratio r = [TBAH]/[␤-CD]=1.01, and with [␤-CD] = 0.506 mM, is given in Fig. 5c. The characteristic resonances of CD and TBAH do not overlap. In fact, upfiled shifts are observed for the tetrabutylammonium protons while in case of ␤-CD protons, a downfield shift is observed (due to the interaction with TBA+ ). To have an assessment on the possible interaction between TBA+ and ␤-CD, the stoichiometry and the corresponding binding constants were determined using the continuous variation, CVM, (or Job’s plots) and titration methods, respectively (for details see, for instance, Valente and Soderman (2014). The CVM is based on the analysis of 1 H NMR spectra for a series of ␤-CD:TBAH mixtures, in which the total concentration of the two species is kept constant (ca.1.0 mM). The stoichiometries are determined by plotting ı × [␤-CD] or ı × [TBAH] against xi (where i = TBAH or CD) and finding the xi values corresponding to the maximum (or minimum) of these distributions (Al-Soufi, Cabrer, Jover, Budal, & Tato, 2003). In Fig. 6 the Job’s plots are represented. As it has been pointed out, the CD protons, in particular the inner cavity protons H3 , suffer a downfield shift in the presence of TBAH. This may indicate the incorporation of the alkyl chains of TBA+ in the CD cavity and the formation of inclusion complexes. Additionally, the “deshielding effect” (i.e. downfield shift) is probably a consequence of such supramolecular interaction where the ammonium group of the TBA+ becomes closer to the CD inner cavity protons affecting their chemical environment. It is noted, however, that no significant

Fig. 6. Job’s plots for TBAH and ␤-CD protons for mixed TBAH: ␤-CD solutions. 1 H TBAH: (䊏) H␣ , (䊉) H␤ , () H␥ , () H␦ . 1 H ␤-CD: (䊐) H4 , () H3 , () H1 , (♦) H5 . The sum of the concentrations of TBAH and ␤-CD is kept constant at 1.0 mM.

B. Medronho et al. / Carbohydrate Polymers 140 (2016) 136–143 Table 1 Molar fractions and corresponding stoichiometry of association between TBAH (m) and ␤-CD (n) obtained from the CVM.

xCD m:n xTBAH m:n

H4

H3

H1

H5

0.499 (0.007) 1.03 (0.01) H␣

0.463 (0.004) 1.16 (0.01) H␤

0.454 (0.004) 1.20 (0.01) H␥

0.463 (0.004) 1.16 (0.01) H␦

0.616 (0.006) 1.61 (0.02)

0.590 (0.006) 1.44 (0.01)

0.607 (0.008) 1.55 (0.02)

0.537 (0.006) 1.16 (0.01)

differences between the different protons of either CD or TBAH, in terms of chemical shift displacements, are observed. On the other hand, an upfield shift is characterizing all protons of tetrabutylammonium, probably indicating that the presence of CD is inducing an anisotropic magnetic field on those protons. An interesting point that comes out from the analysis of Fig. 6 is that the molar fraction of CD at which occurs a maximum (i.e. x around 0.5) does not perfectly match with the molar fraction of TBAH at which the minimum is observed (i.e. x ca. 0.6). In order to have an a accurate determination of the inflexion points, a Gauss peak function analysis has been performed (not shown) and the data are resumed in Table 1. Taking into account the analysis of CD protons, we can conclude that the stoichiometry of interaction TBAH and ␤-CD is mainly 1:1. Comparing these values with those obtained from TBA+ resonances some inconsistency seems to occur. However, it should be stressed that if we consider a pyramidal geometry for TBA+ , with one chain interacting with the CD, it should be expected that the remaining non-complexed ␣-protons are affected for the CD environment and, consequently, the apparent stoichiometry (m:n) increases. Besides, m:n decreases in the order H␣ > H␥ > H␤ > H␦ which is consistent with a trans-configuration of TBA+ alkyl chain and with the previous explanation. Taking into account the stoichiometry of association found by the CVM, the equilibrium formation of the complex can be written as: K=

[CD − G] [CD]f [G]f

(1)

141

where [CD]f and [G]f represent the concentration of free (noncomplexed) species, ␤-CD and TBAH, respectively, and [CD − G] is the concentration of the 1:1 complex. Considering the corresponding mass balance equations, Eq. (1) can be re-written as: K=

f (1 − f ) ([CD]T − f [G]T )

(2)

where f is the fraction of TBAH complexed with the ␤-CD, and [CD]T and [G]T correspond to total concentrations of ␤-CD and TBAH, respectively. Assuming fast exchange conditions, the observed chemical shift for a guest specie is expressed as: ıexp = (1 − f ) ıG + fıCD−G

(3)

where ıG and ıCD−G , represent the chemical shift of a given nucleus when free and complexed, respectively. The chemical shift displacement of a given nucleus of the TBAH, can be expressed as: ıexp =

ıCD−G [CD − G] [G]T

(4)

where [G]T and [CD − G] are the initial concentration of TBAH and of complex, respectively; for the 1:1 complex, after some algebraic manipulation and simplification results in ıexp

ıCD−G = 2[G]T



−



[G]T + [CD]T +

1 [G]T + [CD]T + K

2

1 K

0.5  − 4 ([G]T + [CD]T )

(5)

where [CD]T is the initial concentration of the ␤-CD. The experimental data (see Fig. 7) can be perfectly fitted to the corresponding Eq. (5), using a non-linear least-square algorithm, to obtain the fitting parameters K and ıCD−G (Table 2). The analysis of Fig. 7 and summarized data in Table 2 shows that Eq. (5) fits reasonably well to the ı experimental data. The 1 H NMR spectroscopy rely on direct measurements of the free

Fig. 7. Experimental chemical shifts (ıexp ) of TBAH (1.20 mM) protons as a function of ␤-CD concentration, at 25 ◦ C.

142

B. Medronho et al. / Carbohydrate Polymers 140 (2016) 136–143

Table 2 Fitting parameters of Eq. (5) to experimental data (see Fig. 7).

H␦ H␥ H␤ H␣ H␦ H␥ H␤ H␣

CD

[CD] (mM)

ımax (ppm)

K

R2

␤ ␤ ␤ ␤ ␣ ␣ ␣ ␣

1.20 1.20 1.20 1.20 1.13 1.13 1.13 1.13

−0.124 (±0.003) −0.122 (±0.002) −0.124 (±0.003) −0.123 (±0.003) −0.15(±0.02) −0.11(±0.01) −0.14(±0.02) −0.11(±0.01)

1547 (±155) 1582 (±132) 1573 (±166) 1616 (±189) 81.2 (±13.5) 119.1 (±16.9) 94.6 (±19.7) 130.5 (±19.6)

0.99734 0.99818 0.99704 0.99639 0.99774 0.99748 0.99582 0.99689

R2 is the correlation coefficient of the fitting of Eq. (5) to experimental data shown in Fig. 7.

and bound ligand in a solution containing a known amount of the CD and guest (in this case TBAH); consequently, it is possible to determine microscopic association constants (Connors & Pendergast, 1984) once the affinity of the different protons from the guest to the CD environment (or vice-versa) is different. In the present case, the estimated association constants for all protons are rather similar within the corresponding standard deviation. Nevertheless, it is worth noticing that K values decrease in the order K(H␣ ) > K(H␥ ) > K(H␤ ) > K(H␦ ). As it was suggested before, the H␣ , positioned in the vicinity of ammonium group, shows a higher stability for the complexation, whilst the H␦ (distant from the ammonium) possesses the lowest K. This is also in agreement with the previous discussion where it has been pointed out that, apart from the hydrophobic interactions, the ammonium group is also expected to play a relevant role in the supramolecular interaction. The interaction of TBA+ with ␣-CD was also studied by 1 H NMR and the more relevant information is also summarized in Table 2. It is striking that the averaged K values are more than one order of magnitude lower than the corresponding K values for the ␤-CD:TBA+ association. Although such result is qualitatively in agreement with the turbidimetry measurements (see Fig. 2) this is a rather unusual behavior because the narrower cavity of ␣-CD ˚ (diameter ca. 4.7–5.3 A˚ while ␤-CD has a diameter of ca. 6.0–6.5 A, Szejtli, 1998) is expected to maximize the interactions between the host and the guest (Cabaleiro-Lago, Nilsson, & Soderman, 2005). Surfactants are among the vast list of guest molecules that present a stronger interaction with ␣-CD then with ␤-CD (Szejtli, 1998). If one considers a 1:1 complex where an alkyl tail of TBAH is threading into the CD cavity while the remaining three butyl chains are outside the cavity (displaying a pyramidal geometry), than it is reasonable to assume that the narrower the cavity is, the more energetically unfavorable will be (geometrical constrains) to bend the three butyl chains outside in order to fully insert one butyl chain and maximize the interactions. In other words, TBA+ inclusion in ␣-CD is probably sterically hindered while the TBA+ inclusion in a wider cavity (i.e. ␤-CD) does not require a considerably change in the spatial disposition of the three remaining alkyl chains of the TBA+ . Molecular dynamics will be performed in the future to further elucidate the interaction between TBA+ and ␣-CD and ␤-CD. 4. Discussion Cellulose dissolution has to be regarded as a process which depends on the critical balance between different interactions. In this respect, hydrophobic interactions are also expected to play an important role particularly due to the patent amphiphilicity of cellulose molecules. We have previously suggested that the good performance of the TBAH solvent (in comparison with, for instance, the NaOH aqueous system) arises from the amphiphilicity of the tetrabutylammonium cation which supposedly facilitates the dissolution of cellulose via the weakening of the hydrophobic interactions (Alves et al., 2015). From the data presented here it

is clear that the addition of CDs considerably affects the performance of the TBAH solvent. Moreover, the data strongly suggests an interaction between the TBA+ cation and CDs which occurs via the inclusion of the hydrophobic alkyl moieties in the less polar cavity of the CDs. Therefore, it is thus reasonable to assume that such CD-TBA interactions are weakening the hydrophobic contribution of the TBA ion, which we believe to play a major role in the dissolution of an amphiphilic molecule such as cellulose. In other words, the amphiphilicity of the TBA ion is being compromised due to the formation CD-TBA complex which is manifested by a progressive worsening of the solvent quality with CD addition. This indicates a critical role of hydrophobic interactions between cellulose molecules and provides support for the picture that cellulose molecules have pronounced amphiphilic properties. Another important aspect worth mention about of these tetraalkyllamonium hydroxide-based systems is related to their dissolution efficiently which is intimately dependent on the size of the hydrophobic moieties. If the ammonium (the hydrophilic part) is kept constant, systems with small alkyl chains do not dissolve cellulose (e.g. tetrapropylammonioum hydroxide) and systems with bigger aliphatic chains, such as tetraoctylammonium hydroxide, also do not dissolve cellulose. Although the hydrophilic part does not change, the dissolution performance is changing considerably by changing the size of the aliphatic moieties which we believe to be another supportive indication on the major role of hydrophobic interactions in dissolution. Although the data strongly suggests that the CD-TBA interaction is mainly affecting the hydrophobic interactions of TBA+ , we do not neglect that the formation of such complex may also contribute to disturb, to some extent, the hydrophilic interactions. This hydrophilic/hydrophobic balance is generally highly relevant for solubility issues and particularly important for the class of systems studied in this work. 5. Conclusions The interaction between ␣-CD and ␤-CD with the tetrabutylammonium cation (TBA+ ) was followed by 1 H NMR titration experiments. The data does suggest the formation of 1:1 CD:TBA+ complexes with association constants higher for the ␤-CD (K = 1580 M−1 ) than for ␣-CD (K = 106 M−1 ). It is believed that such unexpected difference in association constant may be due to sterical hindrance which occurs for ␣-CD while does not occurs for ␤-CD. In any case, the formation of CD:TBA+ complexes are found to decrease the dissolution efficiency of the solvent and this is suggested to be a consequence of an effective decrease in the concentration of the available amphiphilic cation in solution. The worsening of the solvent with CD addition is also striking from the polarized light microscopy and turbidimetry measurements. The data does suggest that the amphiphilic properties of the TBA+ are determinant for an efficient dissolution of cellulose. When such properties are reduced or even absent (with progressive addition of CD) the solvent performance is seriously compromised. This provides good support for the view that cellulose molecules have both polar and nonpolar regions and have a strong tendency to associate by hydrophobic interactions. Acknowledgements This work was financially supported by the Portuguese Foundation for Science and Technology, FCT, via the projects PTDC/ AGR-TEC/4049/2012, PTDC/AGR-TEC/4814/2014 and Pest-OE/QUI/ UI0313/2014, researcher grant IF/01005/2014 and PhD grant SFRH/BD/80556/2011. NMR data was collected at the UC-NMR facility which is supported through grants REEQ/481/QUI/2006, RECI/QEQ-QFI/0168/2012, CENTRO-07-CT62-FEDER-002012.

B. Medronho et al. / Carbohydrate Polymers 140 (2016) 136–143

References Abe, M., Fukaya, Y., & Ohno, H. (2012). Fast and facile dissolution of cellulose with tetrabutylphosphonium hydroxide containing 40 wt% water. Chemical Communications, 48(12), 1808–1810. Al-Soufi, A., Cabrer, P. R., Jover, A., Budal, R. M., & Tato, J. V. (2003). Determination of second-order association constants by global analysis of H-1 and C-13 NMR chemical shifts, Application to the complexation of sodium fusidate and potassium helvolate by beta- and gamma-cyclodextrin. Steroids, 68(1), 43–53. Alves, L., Medronho, B., Antunes, F. E., Romano, A., Miguel, M. G., & Lindman, B. (2015). On the role of hydrophobic interactions in cellulose dissolution and regeneration: Colloidal aggregates and molecular solutions. Colloids and Surfaces A—Physicochemical and Engineering Aspects, 483, 257–263. Biermann, O., Hadicke, E., Koltzenburg, S., & Muller-Plathe, F. (2001). Hydrophilicity and lipophilicity of cellulose crystal surfaces. Angewandte Chemie-International Edition, 40(20), 3822–3825. Cabaleiro-Lago, C., Nilsson, M., & Soderman, O. (2005). Self-diffusion NMR studies of the host–guest interaction between beta-cyclodextrin and alkyltrimethylammonium bromide surfactants. Langmuir, 21(25), 11637–11644. Cai, J., & Zhang, L. (2005). Rapid dissolution of cellulose in LiOH/Urea and NaOH/Urea aqueous solutions. Macromolecular Bioscience, 5(6), 539–548. Cai, J., Zhang, L. N., Zhou, J. P., Qi, H. S., Chen, H., Kondo, T., et al. (2007). Multifilament fibers based on dissolution of cellulose in NaOH/urea aqueous solution: Structure and properties. Advanced Materials, 19(6), 821–830. Carvalho, R. A., Correia, H. A., Valente, A. J. M., Soderman, O., & Nilsson, M. (2011). The effect of the head-group spacer length of 12-s-12 gemini surfactants in the host–guest association with beta-cyclodextrin. Journal of Colloid and Interface Science, 354(2), 725–732. Connors, K. A., & Pendergast, D. D. (1984). Microscopic binding constants in cyclodextrin systems—Complexation of alpha-cyclodextrin with Sym-1, 4-disubstituted benzenes. Journal of the American Chemical Society, 106(24), 7607–7614. Cuissinat, C., Navard, P., & Heinze, T. (2008). Swelling and dissolution of cellulose, Part IV: Free floating cotton and wood fibres in ionic liquids. Carbohydrate Polymers, 72(4), 590–596. Diddens, I., Murphy, B., Krisch, M., & Müller, M. (2008). Anisotropic elastic properties of cellulose measured using inelastic X-ray scattering. Macromolecules, 41(24), 9755–9759. Ema, T., Komiyama, T., Sunami, S., & Sakai, T. (2014). Synergistic effect of quaternary ammonium hydroxide and crown ether on the rapid and clear dissolution of cellulose at room temperature. RSC Advances, 4(5), 2523–2525. Glasser, W. G., Atalla, R. H., Blackwell, J., Brown, R. M., Burchard, W., French, A. D., et al. (2012). About the structure of cellulose: Debating the Lindman hypothesis. Cellulose, 19(3), 589–598. Heinze, T., & Koschella, A. (2005). Solvents applied in the field of cellulose chemistry—A mini review. Polímeros: Ciência e Tecnologia, 15, 84–90. Hurtubise, F. G., & Krasig, H. (1960). Analytical Chemistry, 32, 177–181. Isogai, A., & Atalla, R. H. (1998). Dissolution of cellulose in aqueous NaOH solutions. Cellulose, 5(4), 309–319. Jahan, M. S., Saeed, A., He, Z. B., & Ni, Y. H. (2011). Jute as raw material for the preparation of microcrystalline cellulose. Cellulose, 18(2), 451–459. Kacurakova, M., Ebringerova, A., Hirsch, J., & Hromadkova, Z. (1994). Infrared study of arabinoxylans. Journal of the Science of Food and Agriculture, 66(3), 423–427. Liang, C. Y. (1972). In R. T. O’Connor (Ed.), Instrumental analysis of cotton cellulose and modified cotton cellulose (p. 59). New York, NY: Marcell Dekker Inc. Lindman, B., Karlström, G., & Stigsson, L. (2010). On the mechanism of dissolution of cellulose. Journal of Molecular Liquids, 156(1), 76–81. Liu, W. Q., Budtova, T., & Navard, P. (2011). Influence of ZnO on the properties of dilute and semi-dilute cellulose-NaOH-water solutions. Cellulose, 18(4), 911–920.

143

Liu, Z. H., Ni, Y. H., Fatehi, P., & Saeed, A. (2011). Isolation and cationization of hemicelluloses from pre-hydrolysis liquor of kraft-based dissolving pulp production process. Biomass & Bioenergy, 35(5), 1789–1796. Medronho, B., Andrade, R., Vivod, V., Ostlund, A., Miguel, M. G., Lindman, B., et al. (2013). Cyclodextrin-grafted cellulose: Physico-chemical characterization. Carbohydrate Polymers, 93(1), 324–330. Medronho, B., Duarte, H., Alves, L., Antunes, F. E., Romano, A., & Lindman, B. (2015). Probing cellulose amphiphilicity. Nordic Pulp and Paper Research Journal, 30(1), 58-56. Medronho, B., & Lindman, B. (2014a). Brief overview on cellulose dissolution/regeneration interactions and mechanisms. Advances in Colloid and Interface Science, http://dx.doi.org/10.1016/j.cis.2014.05.004 Medronho, B., & Lindman, B. (2014b). Competing forces during cellulose dissolution: From solvents to mechanisms. Current Opinion in Colloid & Interface Science, 19, 32–40. Medronho, B., Romano, A., Miguel, M. G., Stigsson, L., & Lindman, B. (2012). Rationalizing cellulose (in)solubility: Reviewing basic physicochemical aspects and role of hydrophobic interactions. Cellulose, 19(3), 581–587. Medronho, B., Valente, A. J. M., Costa, P., & Romano, A. (2014). Inclusion complexes of rosmarinic acid and cyclodextrins: Stoichiometry, association constants, and antioxidant potential. Colloid and Polymer Science, 292(4), 885–894. Miyamoto, H., Umemura, M., Aoyagi, T., Yamane, C., Ueda, K., & Takahashi, K. (2009). Structural reorganization of molecular sheets derived from cellulose II by molecular dynamics simulations. Carbohydrate Research, 344(9), 1085–1094. O’Connor, R. T., DuPre´ı, E. F., & Mitcham, D. (1958). Applications of infrared absorption spectroscopy to investigations of cotton and modified cottons: Part I: Physical and crystalline modifications and oxidation. Textile Research Journal, 28, 382. Poletto, M., Pistor, V., Santana, R. M. C., & Zattera, A. J. (2012). Materials produced from plant biomass, Part II: Evaluation of crystallinity and degradation kinetics of cellulose. Materials Research-Ibero-American Journal of Materials, 15(3), 421–427. Ren, J. L., Sun, R. C., Liu, C. F., Chao, Z. Y., & Luo, W. (2006). Two-step preparation and thermal characterization of cationic 2-hydroxypropyltrimethylammonium chloride hemicellulose polymers from sugarcane bagasse. Polymer Degradation and Stability, 91(11), 2579–2587. Schneider, H. J., Hacket, F., Rudiger, V., & Ikeda, H. (1998). NMR studies of cyclodextrins and cyclodextrin complexes. Chemical Reviews, 98(5), 1755–1785. Siroky, J., Blackburn, R. S., Bechtold, T., Taylor, J., & White, P. (2010). Attenuated total reflectance Fourier-transform Infrared spectroscopy analysis of crystallinity changes in lyocell following continuous treatment with sodium hydroxide. Cellulose, 17(1), 103–115. Sun, R., Sun, X. F., Liu, G. Q., Fowler, P., & Tomkinson, J. (2002). Structural and physicochemical characterization of hemicelluloses isolated by alkaline peroxide from barley straw. Polymer International, 51(2), 117–124. Szejtli, J. (1998). Introduction and general overview of cyclodextrin chemistry. Chemical Reviews, 98(5), 1743–1753. Valente, A. J. M., & Soderman, O. (2014). The formation of host–guest complexes between surfactants and cyclodextrins. Advances in Colloid and Interface Science, 205, 156–176. Wan, C. C., & Li, J. (2015). Cellulose aerogels based on a green NaOH/PEG solution: Preparation, characterization and influence of molecular weight of PEG. Fibers and Polymers, 16(6), 1230–1236. Wang, X. J., Messman, J., Mays, J. W., & Baskaran, D. (2010). Polypeptide grafted hyaluronan: Synthesis and characterization. Biomacromolecules, 11(9), 2313–2320. Yamane, C., Aoyagi, T., Ago, M., Sato, K., Okajima, K., & Takahashi, T. (2006). Two different surface properties of regenerated cellulose due to structural anisotropy. Polymer Journal, 38(8), 819–826. Yan, L. F., & Gao, Z. J. (2008). Dissolving of cellulose in PEG/NaOH aqueous solution. Cellulose, 15(6), 789–796. Zhang, L. N., Ruan, D., & Gao, S. J. (2002). Dissolution and regeneration of cellulose in NaOH/thiourea aqueous solution. Journal of Polymer Science, B—Polymer Physics, 40(14), 1521–1529.