Chitin-acetate/DMSO as a supramolecular green CO 2 -phile †

The supramolecular chemisorption of CO 2 by the oligomeric chitin-acetate (CA) in DMSO as a green solvent o ﬀ ers a novel eco-friendly approach for CO 2 capture. Since the amino groups in the sorbent material are blocked either by protonation or acetylation, the multi-armed hydroxyl based oligosaccharide captures CO 2 through the formation of an organic carbonate species as con ﬁ rmed by 13 C NMR, in situ ATR-FTIR spectroscopy and conductivity. DFT calculations veri ﬁ ed the formation of the CA-CO 2 adduct, in which the organic carbonate groupis stabilizedthrough supramolecularionic interaction and hydrogen bonding with the neighboring ammonium ion and hydroxyl functional group along the oligomer backbone. The use of other polar aprotic solvents ( N , N -dimethylformamide (DMF), aceto-nitrile, and acetone) was not successful due to solubility issues.


Introduction
Since the early days of the twentieth century, the use of monoethanolamine (MEA) as a scrubbing agent for CO 2 capturing has become a mature technology 1 with known drawbacks upon regeneration, viz., cross-linking, evaporation, and limited recoverability. Several alternatives have been evaluated, ranging from solid sorbents (porous, non-porous) to different liquidbased candidates (chemi-or physisorption). 2 Most materials act as nitrogen-bearing donors rather than oxygen-based sorbents due to the greater basicity of the former that can react chemically with CO 2 following carbamate formation. 3 From a practical industrial point of view, any competitive technology needs to full several critical parameters in terms of cost, eco-friendliness, recoverability, working temperature, efficiency of sorption and minimizing the use of sacricial bases. One approach to increase the efficiency is the use of macromolecular, multi-reactive, and ideally bio-renewable materials to improve the stoichiometry of CO 2 capture and reduce costs. 4 Chitin and its hydrolyzed form, viz., chitosan, are potential candidates to replace MEA-based sorbents for CO 2 capturing. What makes them appropriate as sorbents are their commercial availability, sustainability, mechanical strength and much more importantly is their nitrogen content. Solubility in common organic solvents and gelation are critical issues that hinder their mass use. Therefore, efforts are gathered to explore their CO 2 capturing ability in ionic liquids (ILs). 5,6 Away from ILs, we directed our focus towards the use of oligosaccharide rather than their corresponding parent polymers. A careful compromise between solubility, concentration, and prevention of gelation were taken into consideration by tailoring the degree of deacetylation (DDA). As a continuation of our research on green sorbents 7 for the capture of CO 2 , we have discovered a simple method which utilizes chitin-acetate (CA, Scheme 1) as a scrubbing agent in green solvent, namely, DMSO, by-passing the use of ILs 8-10 and/or superbases. 11 Scheme 1 Structural representation for CA as deduced by nuclear magnetic resonance, elemental analysis and potentiometric titrations.

Results and discussion
Two physical parameters regarding chitin and chitosan, viz., solubility and gelation (vide supra) are taken into consideration as a proper choice of a scrubbing material with reduced viscosity. 12,13 In fact, MEA solutions are always prepared with concentrations that are 20-40% (v/v) to diminish their corrosive character and increased viscosities upon capturing CO 2 . The search for scrubbing agents that can overcome such barriers, would be of great importance for both industrial and academic sectors. For the requested task, we were biased towards the use of low molecular weight chitin derivative as a green sorbent. Fig. 1 shows partial 13 C NMR spectra before and aer bubbling of CO 2 . A peak formed at 157.4 ppm indicates the presence of carbonate 11,14 (no carbamate is formed as evidenced by 13 C NMR). 15 Precedent for chemisorption of CO 2 through carbonate formation has been obtained by Stoddart and coworkers 14 for cyclodextrins within extended metal-organic frameworks (MOFs). Herein, the choice of solvent proved to be critical. In particular, no reaction took place in aqueous solution. Presumably, the formation of the carbonate is facilitated in a polar aprotic solvent for kinetic (nucleophilicity) reasons. Herein, DMSO was used as a model solvent. The formation of carbonate at C-6 could be directly followed through the induced shi of the C-6 peaks at 60 ppm ( 13 C NMR, Fig. 1), in which a clear splitting was observed aer the CO 2 bubbling (there was no shi for C-2 before and aer bubbling which conrms the absence of carbamates. C-2 peaks were 57.4 and 57.3 ppm, respectively).
To differentiate between carbamates and carbonates, in situ ATR-FTIR measurements were carried out in the presence of CO 2 in DMSO as a green solvent. 16 Fig. 2 supports the assumption of chemisorption via two major peak changes. On one hand, the appearance of a new peak assigned to 1555 cm À1 that can be ascribed to the vibrational mode of carbonate. 17 On the other hand, the presumed formation of carbamates (if any) should be accompanied with the emergence of a newly formed peak centered at ca. 1690 cm À1 (absent in our case). 18 The peak centered at 1610 cm À1 is assigned to a bending mode of amine, together with the stretching mode of amide (I) at 1675 cm À1 that is overshadowed upon capturing of CO 2 . 19 Further, the red shi in the (C]O) peak of acetate anion centered at 1720 cm À1 conrms the presumed interaction/metathesis of ammoniumacetate into ammonium-carbonate upon chemisorption of CO 2 . 18 A supramolecular stabilization of the organic carbonate species by ionic and hydrogen-bonding interactions of carbonate with neighbouring groups is presumed to be responsible for the observation that the formation of carbonates (-OH nucleophilic attack) was favoured over that of carbamates (amine nucleophilic attack). 13 C NMR indicated that upon capturing CO 2 , the addition of water to the CA/ carbonate adduct showed that the system is a thermodynamically stable due to the persistence of the organic carbonate peak. In addition, a newly formed peak at about 166.2 ppm indicated the formation of HCO 3 À because of interaction between physisorbed CO 2 with water.
To prove the ion-pair formation in CA/CO 2 adduct, the electrical conductivity of CA dissolved in DMSO was measured as a function of CO 2 bubbling time (Fig. 3). The dropdown in conductivity is explained by the formation of neutral ionic aggregates rather than the presence of individually solvated ions (ammonium and acetate). 20 Doubling of the polymer concentration resulted in an increased conductivity keeping up the same trend. This is considered as another proof of concept for the chemisorption as reported by Mu and co-workers. 6 This notion was supported by DFT-calculated structures of the presumed carbonates in a trimer model, which assumed a comparable alternating conformation as in the biopolymer. 21 The obtained geometry afforded close interaction distances between carbonate-ammonium ions (1.52Å) and carbonatehydroxyl groups (1.66Å, Fig. 4).
The results obtained from the above-mentioned analyses (NMR, in situ IR and DFT calculations) are described in the reaction shown in Scheme 2. The usage of DMSO would activate the hydroxyl group at C-6 of the amino pyranose ring toward nucleophilic attack, which result in chemisorption of CO 2 . The formed organic carbonate is stabilized via ionic interaction and  hydrogen bonding along the oligomer backbone. To our knowledge, activation of alcohols by DMSO is a rst of its kind. We tried to carry out the bubbling experiment in other polar aprotic solvent such as acetonitrile, acetone and N,N-dimethylformamide (DMF). Unfortunately, the insolubility of CA (in the former two solvents) or the limited solubility (in DMF) prohibited further investigation of carbonate formation (see Fig. S1, (ESI †)). This implies that solubility, dielectric constant and H-bond accepting character should be taken in consideration to carry out such reaction.
To measure the amount of sorbed CO 2 by CA in DMSO volumetrically, pressurized vessels controlled with digital manometers were used. An arbitrary pressure was set at 4.0 bars to make sure it would remain in excess to achieve the equilibrated sorption capacities. The following conditions were adopted, namely 10.0% (w/v sorbent), 25.0 C, and 4.0 bar CO 2 . Gas sorption was measured relative to neat DMSO, which is known to absorb one bar of CO 2 at the same conditions, and a value of 3.63 mmol CO 2 g sorbent À1 was determined for CA. This value is competitive to other scrubbing agents, 2 which renders CA an interesting alternative from ecological and economic points of view. CO 2 adsorption isotherm experiments at various pressure and temperature will be reported in an upcoming study to compare between the volumetric and gravimetric methods.
The reversibility of the process depends mainly on perturbing the supramolecular stabilized adduct. We tried several parameters e.g. pH change, sonication, thermal/electrical stimuli. Bubbles of CO 2 were evolved upon applying the stress (vide supra). Therefore, the material is recyclable as an effective CO 2 sorbent.

Materials and methods
Unless otherwise stated, all chemicals were used without further purication. Chitin-acetate (CA) was made by G.T.C. Bio Corporation, Qingdao, China. For experimental manipulations, CA was dried in a vacuum oven at 50 C, overnight. Number average molecular weight (M n ) was determined to be 7 kDa, by multiangle light scattering (MALS) analysis using a Wyatt Dawn Heleos II in combination with a Wyatt Optilab rEX as concentration Source (Wyatt, USA). Based on M n value, CA comprised of 15 units. CO 2 (99.95%, food grade) was purchased from advanced technical gases, Amman, Jordan. Dimethyl sulfoxide and DMSO-d 6 were purchased from M-TEDIA and Sigma-Aldrich, respectively. Solution 13 C-NMR spectra were collected at room temperature using AVANCE-III 400 MHz FT-NMR spectrometer NanoBay (Bruker, Switzerland) in either DMSOd 6 or D 2 O. In situ IR measurements were carried out using a MMIR45m RB04-50 (Mettler-Toledo, Switzerland) with an MCT Detector, with silicon windows probe connected via pressure vessel; sampling 3500 to 650 cm À1 at 8 wavenumber  resolution; scan option: 64; gain: 1Â. Conductivity measurements were carried out using 712 Conductormeter (Metrohm, Switzerland). Calculations were performed within Gaussian 09. The full optimization was performed using DFT method (B3LYP/6-31G*). Minima were characterized by the absence of imaginary frequencies.
A. NMR manipulation (CO 2 bubbling). In an NMR tube, 30 mg of the substrate was dissolved in 0.5 mL DMSO-d 6 . Upon dissolution, CO 2 was bubbled into the NMR tube via a long cannula for 20 minutes. CO 2 saturation was ensured using NMR spectroscopy.
B. In situ IR manipulation. In a 50 mL ATR-FTIR autoclave, 10% (w/v) CA was prepared using DMSO as a solvent. The autoclave was charged with 4.0 bars CO 2 and le to scan for 4 hours (with a 15 seconds interval per scan). The initial and nal drops in pressure were collected with a digital manometer. For control purposes, neat DMSO was charged under the same conditions.
C. Conductivity measurements. Conductivity was used to conrm the formation charged species upon bubbling CO 2 . A proper amount of CA was dissolved in DMSO using a 50 mL beaker to prepare 0.33 and 0.67% (w/v). The conductivity was recorded every 30 seconds aer bubbling CO 2 using a conductor until no drastic changes took place.

Conclusions
In conclusion, we propose a new method to absorb atmospheric CO 2 (without the use of stoichiometric amounts of the superbases or ionic liquids) in the presence of DMSO as a green solvent. Whether the CO 2 chemisorption by carbonate formation is favored for kinetic reasons (higher nucleophilicity in polar aprotic solvents) or thermodynamic reasons (supramolecular stabilization of the carbonate by interaction with neighboring groups) requires further investigations. Regardless of the mechanistic details, the formation of stable adducts in the capture of CO 2 by green sorbents is of considerable practical interest. The commercial availability of this oligosaccharide and cheap cost for its processing are additional assets. The use of other polar aprotic solvents (N,N-dimethylformamide (DMF), acetonitrile, and acetone) was not successful due to solubility issues.