Ind. Eng. Chem. Res. 2010, 49, 11809–11813
11809
Alkanolamine Ionic Liquids and Their Inability To Dissolve Crystalline Cellulose Andre´ Pinkert, Kenneth N. Marsh,* and Shusheng Pang Department of Chemical and Process Engineering, UniVersity of Canterbury, Christchurch, New Zealand
The purpose of this research note is to report on the inability of alkanolammonium ionic liquids to dissolve crystalline cellulose, supporting the claims we make in an associated commentary paper, “Reflections on the Solubility of Cellulose” (Pinkert et al., Ind. Eng. Chem. Res. 2010, 49, DOI: 10.1021/ie1006596). Four organic acids (formic acid, acetic acid, malonic acid, and citric acid) and five amines (ethanolamine, diethanolamine, triethanolamine, propan-1-olamine, and diallylamine) were selected for the synthesis of 17 protic ionic liquids. The products were characterized by 1H nuclear magnetic resonance spectroscopy, infrared spectroscopy, thermogravimetric analysis, and Karl Fischer titration. Introduction
2. Experimental Section
“Reflections on the Solubility of Cellulose” is the title of a commentary article that presents new perspectives on the dissolution process of cellulose in ionic liquids (ILs), offering explanations why only certain IL ion combinations are potent cellulose solvents.1 Our reflections in the commentary are mainly based on literature data, but additional experiments were necessary to clarify a criticism on why the article did not endorse the ability of an alkanolammonium ionic liquid (AAIL), ethanolammonium formate, to dissolve cellulose.2 It was reported that dispersions of up to 0.3 mass fraction of cellulose can be prepared, but without obtaining a clear solution. However, no information was given regarding the IL water content, the mixing conditions, or the crystallinity of the cellulose, and all those factors are known to influence an IL’s ability to dissolve the biopolymer.3 This research note is being published in association with our commentary,1 and its purpose is to report on qualitative experiments regarding the ability of a range of alkanolammonium-based ILs to dissolve crystalline cellulose.
2.1. Materials and Instruments. 2.1.1. Materials. Formic acid, acetic acid, malonic acid, 3-hydroxypropylamine, microcrystalline cellulose (MCC), and chemicals for Karl Fischer titrations were obtained from Merck. Citric acid and diallylamine were purchased from Sigma Aldrich. 2-Hydroxyethylamine and tris(2-hydroxyethyl)amine were from Unilab, and bis(2-hydroxyethyl)amine from AnalaR. The commercial IL 1-ethyl-3methylimidazolium acetate ([EMIM]Ac) was from Fluka and dried at 353 K for 48 h prior to use, resulting in a water mass fraction of 8 × 10-4. Molecular sieve type 4A was obtained from Fluka, and nylon syringe filters with a pore size of 0.45 µm and a diameter of 13 mm were obtained from Raylab. All reagents were analytical-grade (AR) purity, except 2-hydroxyethylamine, and were used without further purification. Liquid reagents were dried with a molecular sieve prior to use. Microcrystalline cellulose, with a particle size of (20 to 160) µm and a degree of polymerization (DP) of 163,40 was dried at 353 K for 72 h to reduce the initial water content by 90 % to a final water mass fraction of 2.9 × 10-3. 2.1.2. Instruments. Fourier transform infrared (FT-IR) spectra were recorded on a Perkin-Elmer Spectrum One FT-IR spectrometer over a range of (4000 to 450) cm-1. A total of 20 scans at a resolution of 4 cm-1 was taken for each sample. Liquid samples were sandwiched between two polished KBr plates, and solid samples were pelleted with KBr containing a sample mass fraction of 0.01. 1H NMR spectra were recorded on a Varian INOVA 500 spectrometer (500 MHz), using the signals of the residual solvent protons as internal references (δH 2.6 for DMSO-d6). Water content was determined with a Radiometer Analytical TIM550 Volumetric Karl Fischer Titrator, using Karl Fischer Reagent 5 from Merck. The titrant was calibrated with a 0.01 % Apura water standard. Thermogravimetric analysis (TGA) was conducted on a SDT Q600 from TA Instruments by heating (5 to 10) mg of sample in a platinum pan under a nitrogen atmosphere (N2 flow rate ) 100 mL min-1). The decomposition temperature (TD) values were determined at a fractional mass loss of 0.1 at two different heating rates (TD,1K ) 1 K min-1 and TD,10K ) 10 K min-1). Microscopy was performed with an Olympus PX60 microscope. 1 H NMR spectra, FT-IR spectra, and the decomposition temperatures of each IL are given as Supporting Information. 2.2. Synthesis. The ILs were prepared according to the synthesis of 2-hydroxyethylammonium formate described in the literature.2 A three-necked 500 mL flask was equipped with a reflux condenser, a dropping funnel, a nitrogen gas source, and
The first AAIL, ethanolammonium acetate, was synthesized in 1955.4 Since then, many more studies on AAILs have been reported, mainly focusing on ethanolammonium formate and ethanolammonium acetate.2,5-20 Less data are available on the diethanolammonium and triethanolammonium cations.5,8,9,17,19-23 Several patents that mention alkanolammonium malonates and diallylammonium or propan-1-olammonium ILs have been issued, but no articles have been published on these compounds.24-33 Angell and co-workers have shown that the ionicity of protic ILs can differ from the ionicity of aprotic ILs, because of the possibility of proton back-transfers.34-36 It has been argued that ∆pKa values of the aqueous educts can be used to estimate the Gibbs energy of the proton transfer between acid and base,34 but other authors have raised concerns, arguing that the aqueous pKa data are only of remote relevance to the IL state.37 However, 15 N nuclear magnetic resonance (15N NMR) studies on monoethanolammonium, diethanolammonium, and triethanolammonium acetates and formates strongly suggest that the proton transfer is complete for stoichiometric educt ratios.22 Our studies on the ionicity of these AAILs support these findings, and have been reported elsewhere.38,39 * To whom correspondence should be addressed: E-mail:
[email protected].
10.1021/ie101250v 2010 American Chemical Society Published on Web 10/08/2010
11810
Ind. Eng. Chem. Res., Vol. 49, No. 22, 2010
Table 1. Acronyms for IL Ions acronym
explanation
[HEA] [HPA] [DEA] [TEA] [DAA] Fmt Ac Mal Cit
2-hydroxyethylammonium 3-hydroxypropylammonium bis(2-hydroxyethyl)ammonium tris(2-hydroxyethyl)ammonium diallylammonium formate acetate malonate citrate
a magnetic stirring bar. Moisture-free conditions in the flask were ensured by repeating the following cycle three times: (i) apply vacuum, (ii) heat flask surface with a heat gun while vacuum is applied, and (iii) flush with nitrogen. Subsequently, the flask was kept under a dry nitrogen gas atmosphere at all times. The flask was mounted in an ice bath to control the exothermic reaction, and the alkanolamine was stirred vigorously while the acid was added dropwise. In the case of the acid being a solid at ambient temperatures, the alkanolamine was added dropwise to the acid. Very viscous products required the use of an oil bath to control the temperature. After mixing, stirring was continued at the lowest suitable temperature for at least 12 h. Finally, the ILs were stirred for at least 12 h with a molecular sieve, to reduce the water content to a minimum. Prior to use, the ILs were filtered through nylon syringe filters under nitrogen atmosphere to remove the molecular sieve residues. Table 1 explains the acronyms used for the description of IL ions in this study. 2.3. Cellulose Solubility. All experiments were conducted in a dry N2 atmosphere. Five grams of dried IL was transferred to a moisture-free 25 mL flask. Microcrystalline cellulose was added in mass fractions of either 0.01, 0.05, or 0.1 and stirred at 353 K for 12 h. The stirring temperature was kept below the thermal decomposition temperature (TD,1K) of the IL (if TD,1K < 353 K). If no dissolution occurred after 12 h with decreased stirring temperatures, the experiment was repeated with a stirring temperature of 353 K. To determine whether any cellulose has dissolved, the turbid mixtures were centrifuged at 10 000 rpm for 30 min at 313 K in a dry atmosphere. The supernatant was separated and centrifuged again. This process was repeated until no further residue was formed after centrifugation. The obtained IL supernatant was then mixed with distilled water to regenerate and identify any dissolved cellulose. In addition, all IL + water mixtures were analyzed with an Olympus PX 60 microscope with varying magnification (10×, 50×, 100×) for the existence of regenerated cellulose. 3. Results and Discussion 3.1. Synthesis. The objective of this work is to investigate whether protic AAILs can dissolve microcrystalline cellulose and, if so, to determine the influence of the ion structure on the dissolution ability. Figure 1 shows the organic acids and amines used for the synthesis. Cation variations include (i) the number of carbon atoms between the hydroxy group and the nitrogen atom ([HEA] < [HPA]), (ii) the number of alkanol substituents on the nitrogen atom ([HEA] < [DEA] < [TEA]), and (iii) the replacement of alkanol substituents with allyl groups ([DAA]). Anions vary mainly in their number of -CO2H groups (Fmt ) Ac < Mal < Cit). All the investigated AAILs are protic and are formed by proton exchange reactions between amines and organic acids (see Figure 2). These reactions are exothermic, and the temperature must be controlled to avoid the formation of condensa-
Figure 1. Organic acids and amines used for the synthesis of alkanolammonium ionic liquids (AAILs). Relevant atoms are labeled to assist 1NMR assignment. Three-lettered acronyms are introduced to facilitate the description of the resulting ionic liquid (IL) products.
Figure 2. Synthesis of AAILs: proton exchange reactions between organic acids and amines result in ammonium carboxylates.
tion products. Figure 3 compares the 1H NMR spectra of the IL diethanolammonium acetate ([DEA]Ac) with the spectra of its educts, acetic acid and diethanolamine. The chemical shift differences of the nonlabile protons between the educts and the resulting IL are attributed to the ionic nature of the product. Peaks of rapidly exchanging protons are found further downfield. The lack of additional signals in the spectra, which would result, for example, from the asymmetric condensation product or unreacted educts, shows that the reaction is complete and without byproduct. The increased coupling constants of the triplets in [DEA]Ac, compared to its educts, are attributed to the increased viscosity of the resulting IL. In general, the viscosity increased with the number of hydroxy groups present in the product. Some syntheses required increased stirring temperatures to counteract the high viscosities and to ensure a complete reaction. As a consequence, negligible amounts of condensation byproduct could be detected in [DEA] and [TEA] malonates. A pH value of 7 for all but one of the synthesized AAILs is an additional indication that the educts have reacted completely. Diallylammonium formate ([DAA]Fmt) had a pH value of 5, suggesting incomplete reaction. This assumption was confirmed by the lack of H-1 proton peaks in the 1H NMR spectra of [DAA]Fmt, and the fact that the signal of the acid proton H-a appeared further upfield, in comparison with that of the other alkanolammonium formates. The number of hydroxy groups present in an IL significantly influences its properties. In general, the higher the number of hydroxy groups, the higher the viscosity, melting temperature, and hygroscopic behavior of the ILs. This hygroscopic trend can be seen in Table 2, which summarizes the water content and the optical appearance of the dried ILs at 298 K. We observed the condensation of the volatile acids at cooler areas at the neck of the flask during water removal at elevated temperatures (353 K) and decreased pressures (
