Design and Synthesis of Thermoresponsive Ionic Liquid Polymer in

Feb 6, 2015 - We report the design and development of a series of novel ionic liquid polymers (PILs) which possess ... Temperature-Responsive Ionic Li...
0 downloads 0 Views 3MB Size
Download PDF
Article pubs.acs.org/Macromolecules

Design and Synthesis of Thermoresponsive Ionic Liquid Polymer in Acetonitrile as a Reusable Extractant for Separation of Tocopherol Homologues Yangyang Lu,† Guoqiang Yu,† Wen-Jun Wang,*,†,‡ Qilong Ren,‡ Bo-Geng Li,† and Shiping Zhu*,§ †

State Key Lab of Chemical Engineering, College of Chemical and Biological Engineering, Zhejiang University, Hangzhou, China 310027 ‡ Key Lab of Biomass Chemical Engineering of Ministry of Education, College of Chemical and Biological Engineering, Zhejiang University, Hangzhou, China 310027 § Department of Chemical Engineering, McMaster University, Hamilton, Ontario Canada L8S 4L7 S Supporting Information *

ABSTRACT: We report the design and development of a series of novel ionic liquid polymers (PILs) which possess thermoresponsive properties in organic solvent. The PILs were synthesized via reversible addition−fragmentation chain transfer (RAFT) copolymerization of 1-vinyl-3-butylimidazolium bromide and N-isopropylacrylamide, followed by anion exchange of bromide to amino acid Dalanine. The PILs possessed 0.36−0.76 molar fraction of ionic liquid with number-average molecular weight of 2.70−8.17 kg/mol and polydispersity index ranging between 1.12 and 1.25. The copolymerizations followed first-order reaction kinetics, and they were well-controlled, as indicated by the linear increase of molecular weight with monomer conversion. The PILs were thermoresponsive in acetonitrile with upper critical solution temperatures (UCST) varying from 25.7 to 34.8 °C, owing to the introduction of anion of amino acid. The PILs could be completely precipitated out by lowering the solution temperature. The PIL/acetonitrile solutions were used as extract phase for separation of tocopherol homologues in hexane. The distribution coefficient of δ-tocopherol between the extract and raffinate phases (Dδ) and the selectivity coefficient of δ-tocopherol to α-tocopherol (Sδ/α) reached as high as 7.86 and 13.0, respectively, while Dβ&γ was 3.63 and Sβ&γ/α was 6.0. The mole ratio of α-tocopherol in the raffinate phase increased from 0.08 to 0.27 after one stage extraction. The PILs could be reused for multiple extraction cycles with negligible change in the tocopherol distribution and selectivity coefficients. The thermoresponsivity of the PILs is of great benefit by eliminating the normally required back extraction steps. This work demonstrates the potential of thermoresponsive polymers for use in high performance separation of natural products.

■

sorbents,10 carbon precursors,11 responsive materials,12 porous materials,13 and energy storage battery.14 PILs are generally synthesized by direct polymerization of ILs or postmodification of polymeric precursor.15,16 The majority of the reported PILs are prepared by conventional radical polymerization.10,17 However, living radical polymerization of ILs allows synthesis of PILs with controlled molecular weight and narrow polydispersity. Many living radical polymerization techniques have been applied to polymerization of ILs, such as reversible addition−fragmentation chain transfer (RAFT) polymerization,18,19 atom transfer radical polymerization,20,21 nitroxidemediated polymerization,22 and organotellurium-mediated living radical polymerization.23 Among these techniques,

INTRODUCTION Ionic liquids (ILs), which are entirely composed of ions under 100 °C, have attracted great interest in recent years because of their special and desirable properties, such as high polarity, negligible volatility, good ionic conductivity, and remarkable chemical and thermal stability. ILs are emerging as greener solvents, replacing volatile organic compounds in numerous processes.1−3 They have been used in fuel cells and batteries,4 as solvents for organic reactions5 and extractive processes,6 and as components in composite materials.7 Ionic liquid polymers (PILs) refer to a new class of polyelectrolytes which contain ionic liquid units in the polymer chains. There have been significant progresses in the areas of PILs research. In PILs, the unique properties of ILs are combined with those of polymers for possible synergetic effects in enhancing the mechanical performance, processability, and electrochemical and thermal stability.8,9 These research efforts have widely expanded the potential applications of PILs as CO2 © XXXX American Chemical Society

Received: December 29, 2014 Revised: January 26, 2015

A

DOI: 10.1021/ma502611s Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules Scheme 1. Structure of Four Tocopherol Homologues

organic solvents, hence severely restricting the commercial development of ILs.43,44 Although ILs have been used as extractant and solvents simultaneously in many experimental cases, they are often diluted by other solvents in industrial practices to reduce viscosity and costs. It is also critical to have good strategies to separate ILs from these used solvents. In addition, toxicity of ILs are also of some concerns, particularly when ILs are used to separate natural products.45,46 Recently, our group launched a research program, aiming at the design and development of smart PILs for use in separation of bioactive compounds. The grounds behind this program are twofold. First, PILs are much safer than ILs, since high molecular weight polymers have lower toxicity than their small monomers.47 Even though the small molecule monomers are highly toxic and hazardous, upon polymerization, they become less toxic or even nontoxic. Second, it is much easier to separate PILs, than ILs, from solvent mixtures. Polymer chains, to a certain degree, have already “collected” small molecule units. The recent advent of stimuli-responsive polymers provides a great opportunity in developing smart polymers triggered by external stimulus, such as pH, CO2, temperature, light, etc.48−57 Since organic solvents are commonly used in conventional LLE, the stimuli-responsiveness of PILs in the solvents, such as thermoresponsive PILs having upper critical solution temperatures (UCST) at ambiance temperature, can be utilized to eliminate the back-extraction process. The objectives of this work are to design a class of thermoresponsive PILs in acetonitrile, a typical solvent for extraction of tocopherol, and to demonstrate their use as extractant in selective separation of tocopherol homologues, without the need of a back-extraction process. The extract phase can be cooled to precipitate the PIL extractant, which can be recycled and reused in subsequent extraction cycles, after efficient separation of tocopherol homologues. This work offers an innovative cost- and energy-saving method for LLE process. To the best of our knowledge, this is the first work reporting the use of PILs as extractant to separate natural products by LLE. A series of thermoresponsive PILs were synthesized through RAFT copolymerization of 1-vinyl-3-butylimidazolium bromide and N-isopropylacrylamide (NIPAM), followed by anion exchange of bromide to amino acid D-alanine. The 1vinyl-3-butylimidazolium alanine units, randomly distributed along chain backbone, are responsible for the extraction and the thermoresponsivity in acetonitrile, while the NIPAM units improve the solubility of polymer in acetonitrile. The PIL/ acetonitrile solutions were used as extractant for the separation of tocopherol homologues in hexane. The extraction performance, thermoresponsivity, and reusability of the PILs were investigated.

RAFT polymerization has proved to be tolerant of various reaction conditions and monomers.24,25 Natural products with improved pharmacological and/or biological activities are considered to be high-value-added products, which have played significant roles in developing medical and health products over the past decades.26 However, most bioactive compounds in herbal plants are in mixtures of very low concentrations. Moreover, these compounds often have very similar structures, adding difficulty in their separations. For example, vitamin E is an effective antioxidant, which is considered as a common and essential nutrient required for maintaining normal physiological functions in the life process of human beings.27,28 Vitamin E consists of tocopherols and tocotrienols, of which α-tocopherol is identified to have the highest biological activity.29 Unfortunately, α-tocopherol in vegetables and fruits is usually accompanied by other three tocopherol homologues: β-, γ-, and δ-tocopherols. The molecular structures of the four homologues differ in the number and position of methyl groups on the benzene ring, as shown in Scheme 1. It is a great challenge to separate the targeted α-tocopherol from the other tocopherols because of their structural similarities. Chromatographic technology has been used for the separation of tocopherol homologues,30,31 even for β- and γtocopherol separation.30 However, its application is limited due to low throughput. Liquid−liquid extraction (LLE) is a wellknown method for efficient product separation.32 This unit operation technique has been widely utilized in industry due to its relatively simple equipment and ease of operation. Conventional LLE generally requires back-extraction to separate the extractant from the extract phase. This step is an energy intensive process and requires a large amount of volatile solvent, which raises environmental and energy concerns. Therefore, the elimination of back-extraction step will be highly beneficial in LLE processes. There have been numerous investigations on the use of ILs as extractant in the separation of biomolecules, such as extraction of lignin from lignocellulose, 33 chitin from crustacean,34 polysaccharides from bran,35 α-tocopherol from tocopherol homologues,36 L-tryptophan from water solution,37 and proteins from biological fluids.38 A common challenge with using ILs in practice is the difficulty to separate the product from ILs in order to regenerate and recycle ILs. Immobilization of ILs or IL derivatives as stationary phase has been applied for natural product separation.39,40 Supercritical CO2 has also been used to back extract products from ionic liquids.41,42 Other separation techniques that have been used in the purification of ILs are membrane and distillation.6,43 However, these aforementioned methods cost large amounts of energy and B

DOI: 10.1021/ma502611s Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules Scheme 2. Synthetic Route of VBIM-Br and Poly(VBIM-Ala-r-NIPAM) (PIL)

■

found in the Supporting Information. For comparison, an amino acid IL monomer VBIM-Ala was also synthesized following the same anion exchange procedure as follows. RAFT Copolymerization of VBIM-Br Monomer and NIPAM. A typical copolymerization of VBIM-Br monomer and NIPAM was conducted as follows: VBIM-Br monomer (4.0 g, 17.3 mmol), NIPAM (1.96 g, 17.3 mmol), CTA (84.0 mg, 0.432 mmol), AIBN (28.4 mg, 0.173 mmol), and dried ethanol (17.88 g) were added into a 50 mL rubber septum sealed flask with a magnetic stirrer. The mixture was bubbled with nitrogen for 30 min to remove oxygen and followed by stirring at 65 °C for 30 h. The resulting polymer was concentrated and precipitated into a large amount of acetone. It was then dried under vacuum for 4 h. The total conversion of VBIM-Br and NIPAM was determined gravimetrically. The ratio of VBIM-Br units in the copolymer was determined by 1H NMR. Anion Exchange of Poly(VBIM-Br-r-NIPAM) (PIL-Br) to Poly(VBIM-Ala-r-NIPAM) (PIL). A copolymer sample PIL-Br of 2.5 g was dissolved in 100 mL of deionized water. Ambersep-OH ionexchange resin (50 g) was added. The mixture was stirred for 4 h at room temperature and then filtered to remove the resin. No further precipitation was observed when the resulting solution was tested by AgNO3/HNO3 solution, which ensured a complete exchange. A slight excess equimolar of D-alanine (0.963 g) was dissolved in 20 mL of deionized water and added dropwise into the polymer solution. The mixture was stirred at room temperature for 16 h. Water was then removed by distillation under reduced pressure. Following the distillation, 18 mL of acetonitrile and 2 mL of methanol were added to remove the unreacted D-alanine. The filtrate was concentrated and dried at 60 °C under vacuum for 4 h. Extraction Experiments. The extraction experiments were conducted as follow: A known amount of tocopherol was dissolved in 3 mL of hexane as raw liquid. The mixture was added into an Erlenmeyer flask. An equal volume mixture of PIL and acetonitrile was slowly added into the flask. The flask was shaken for 40 min at 35 °C in an orbital shaker and left undisturbed for 30 min at the same temperature. The samples were taken from each phase and diluted with methanol for HPLC analysis. Recycling Procedure. The raffinate phase after extraction was first removed from the mixture. A small amount of water was added into the extract phase. The flask was cooled to 20 °C, which results in precipitation of the PIL extractant from acetonitrile. The concentrations of tocopherols in acetonitrile before and after the removal of PIL extractant were measured by HPLC. The mass of regenerated PIL

EXPERIMENTAL SECTION

Materials. 1-Vinylimidazole (≥99%), 1-bromobutane (≥99%), Dalanine (≥98%), and NIPAM (≥99%) were purchased from J&K Chemical. AIBN (≥99%) was purchased from Aladdin Reagent Co. and recrystallized with ethanol twice. Ambersep−OH ion-exchange resin was purchased from Alfa Aesar and washed with water. DMF (AR, J&K Chemical), acetonitrile (ACN) (AR, Sinopharm), and ethanol (AR, Sinopharm) as solvents were distilled before use. The chain transfer agent (CTA) O-ethyl-S-(1-carboxy)methyl dithiocarbonate was synthesized according to the literature.58 Characterization. Nuclear magnetic resonance (NMR) spectra were recorded on Bruker Advance 400 MHz spectrometer, using DMSO-d6 as solvent. Polymer molecular weights and distributions were measured by the gel permeation chromatography (GPC) system with a Waters 1515 isocratic HPLC pump, two PLgel 5 μm Mixed-C, 300 × 7.5 mm columns, and a Wyatt Technology Optilab DSP refractive index detector at 25 °C. DMF containing 0.05 M LiBr was used as solvent at a flow rate of 1 mL/min. Poly(methyl methacrylate) (PMMA) standards (molecular weight from 2710 to 176 000 g mol−1) were used for calibration. Tocopherol concentrations were determined with a high performance liquid chromatograph (HPLC) equipped with an autosampler, an Atlantis C18 column (5 μm, 250 × 4.6 mm), a Waters 1525 binary pump, and a Waters 2487 dual λ absorbance detector. A mixture of methanol and water (95/5 v/v) was used as solvent at a flow rate of 1 mL/min at 40 °C. The detection of tocopherols was performed at 292 nm. Synthesis of 1-Vinyl-3-butylimidazolium Bromide Monomer. The ionic liquid monomer 1-vinyl-3-butylimidazolium bromide (VBIM-Br) was prepared according to the procedure reported in the literature.59 For convenience of the readers, the synthesis procedure for VBIM-Br is briefly explained here (illustrated in Scheme 2): A slight excess of 1-bromobutane (40.0 g, 0.292 mol) was slowly added into 1-vinylimidazole (25.0 g, 0.265 mol). The mixture was stirred for 20 h at 60 °C and then washed with ethyl acetate for several times. The product was dried for 24 h at room temperature under vacuum to obtain a white solid (53.3 g 87%). 1H NMR (DMSO-d6, 400 MHz): δH 9.84 (s, 1H), 8.32 (s, 1H), 8.04 (s, 1H), 7.36 (dd, 1H), 6.06 (dd, 1H), 5.41 (dd, 1H), 4.26 (t, 2H), 1.82 (m, 2H), 1.27 (m, 2H), 0.87 (t, 3H). 13C NMR (DMSO-d6, 400 MHz): δC 135.2 (−N−CH−N−), 128.7 (CH2CH−), 123.2 (−N−CHCH−N−), 119.2 (−N− CHCH−N−), 108.6 (CH2CH−), 48.8 (−N−CH2−), 31.0 (−CH2−CH2−CH2), 18.7 (−CH2−CH2−CH3), 13.2 (−CH2− CH3). 1H (Figure S1) and 13C NMR spectra (Figure S2) could be C

DOI: 10.1021/ma502611s Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules Table 1. Experimental Conditions and Results of RAFT Copolymerization of VBIM-Br and NIPAMa xc (%)

Mn (kg/mol)

run

[VBIM-Br]0/[NIPAM]0/[CTA]0

X (%)

VBIM-Br

NIPAM

NMRd

GPCe

PDI

VBIM-Br:NIPAM

f

20/20/1 30/30/1 40/40/1 30/15/1 30/60/1 40/40/1 40/40/1

36.4 42.1 56.1 54.5 34.4 57.9 53.7

42.2 50.4 65.6 53.4 47.5 68.0 63.3

24.5 25.2 36.8 32.0 41.7 37.4 34.1

2.70 4.55 7.92 4.44 6.32 8.17 7.59

6.05 9.14 5.87 7.21 9.18 8.04

1.12 1.15 1.13 1.25 1.22 1.17

8:5 15:8 26:15 16:5 14:25 27:15 25:14

1 2 3 4 5 6g 7h

b

a 25 wt % of total monomer in polymerization system, [CTA]0/[AIBN]0 = 2.5, 65 °C, 30 h, DMF as solvent for runs 1−5. bX: overall monomer conversion was measured gravimetrically. cx: VBIM-Br and NIPAM conversions, calculated from X and the mole ratio of VBIM-Br units in copolymer F. dMn,NMR calculated from 1H NMR spectra. eMeasured by GPC using PMMA standards in DMF (containing 0.05 M LiBr). fGPC curve of the copolymer obtained from run 1 was unavailable because it was overlapped with the peak of mobile phase. gEthanol as solvent. hAcetonitrile as solvent.

was measured gravimetrically. The obtained PIL was dried at 60 °C for 4 h to be used in subsequent extraction experiments, which were conducted following the same procedure.

■

RESULTS AND DISCUSSION RAFT Synthesis of PIL Extractants. The PIL-Brs were synthesized via RAFT copolymerization of VBIM-Br and NIPAM as shown in Scheme 2. All copolymerization runs were carried out with AIBN as initiator and O-ethyl-S-(1carboxy)methyl dithiocarbonate as RAFT chain transfer agent (CTA). The polymerization conditions and results are summarized in Table 1. The random copolymer PIL-Brs were soluble in acetonitrile, but VBIM-Br homopolymer had poor acetonitrile solubility. The RAFT copolymerizations of VBIM-Br with methyl methacrylate, acrylamide, or oligo(ethylene glycol) methacrylate were also carried out. The resulting copolymers had poor solubilities in acetonitrile and thus could not be used as extract phase for separation of tocopherol homologues in hexane. The VBIM-Br content in the random copolymer PIL-Br was determined from the 1H NMR spectra. As shown in Figure 1a, the peak at 9.0−10.2 ppm was ascribed to the protons of N− CH−N of the imidazolium ring in VBIM-Br units, with the integral area denoted as S1. The peak at 7.4−8.6 ppm was attributed to the protons of N−CHCH−N in VBIM-Br units and C−NH−C in NIPAM units, with the integral area of S2. The mole ratio of VBIM-Br units in the copolymer PIL-Br, F, was calculated from F = S1/(S2 − S1), where 1 − F is the mole ratio of NIPAM units in PIL-Br. The conversions of the two monomers were calculated from the mass balances based on the overall conversion and VBIM-Br mole ratio data. The copolymer PIL-Br was also confirmed by the 13C NMR, as shown in Figure 1b. After copolymerization of VBIM-Br with NIPAM at 65 °C for 30 h, the overall conversions were calculated to be 36.4− 57.9%. The [VBIM-Br]:[NIPAM] ratio in the copolymers PILBr varied from 16:5 to14:25. The GPC traces of samples from runs 2−5 are shown in Figure 2. The PIL-Br possessed number-average molecular weight (Mn,NMR) of 2.70−8.17 kg/ mol with low polydispersity index (PDI) between 1.12 and 1.25. All the molecular weights measured by GPC were higher than the calculated theoretical values. The deviation could be attributed to the different structures of PIL-Br and the PMMA standards used for calibration, which is a well-known unsolved problem in the GPC characterization of polyelectrolytes.

Figure 1. NMR spectra of poly(VBIM-Br15-r-NIPAM8) (run 2 PILBr): (a) 1H NMR and (b) 13C NMR.

Figure 2. GPC traces of PIL-Br samples (runs 2−5).

The PIL-Br samples having different molecular weights and compositions were prepared by varying the total monomer to CTA ratio, ([VBIM-Br]0 + [NIPAM]0):[CTA]0, and the D

DOI: 10.1021/ma502611s Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

Figure 3. Evolution of VBIM-Br and NIPAM conversions with polymerization time and number-average molecular weights and PDI with overall monomer conversion in run 2 (a, c) and run 5 (b, d).

comonomer ratio, [VBIM-Br]0:[NIPAM]0. The first three runs in Table 1 differ in the monomer to CTA ratio in order to target copolymers with the same composition but different molecular weights. The resulting polymer products had the compositions of [VBIM-Br]:[NIPAM] = 8:5, 15:8, and 26:15, which translate to approximately 63% VBIM-Br and 37% NIPAM units. The comonomer composition of 50% VBIM-Br resulted in the copolymer composition of 63% VBIM-Br, suggesting IL monomer VBIM-Br had higher reactivity than NIPAM. In other words, the IL monomer was the fast monomer in this copolymerization. The molecular weight (Mn) of the PIL-Br increased with the monomer to CTA ratio. The GPC curve of run 1 sample was unavailable because it was overlapped with the peak of the mobile phase. The PIL-Br copolymer having a similar number of VBIM-Br units but different numbers of [NIPAM] units were also synthesized. The polymerizations were conducted with comonomer compositions of [VBIM-Br]:[NIPAM]:[CTA] = 30:15:1 (run 4), 30:30:1 (run 2), and 30:60:1 (run 5), with the resulting copolymers compositions of [VBIM-Br]:[NIPAM] = 16:5, 15:8, and 14:25, respectively. These two sets of copolymer samples provided a basis for evaluation of the effects of polymer chain parameters on tocopherol extraction performance. To gain a better understanding on the reaction kinetics of RAFT copolymerization of VBIM-Br and NIPAM, the evolution of monomer conversions and molecular weights with polymerization times for run 2 and run 5 was investigated. Figure 3a,b shows the plots of −ln(1 − x) versus time for each

monomer in runs 2 and 5. The linear relationships observed suggest first-order kinetics in the RAFT copolymerization of VBIM-Br and NIPAM. It also further verified that VBIM-Br was more reactive than NIPAM during the copolymerization. The number-average molecular weights and PDIs of the samples in runs 2 and 5 against overall conversions are plotted in Figure 3c,d. The polymerizations were controlled and living, as indicated by the low PDIs of all samples (within 1.1−1.3) and the linear relationships between the number-average molecular weights and the overall monomer conversions. The effect of various solvents, DMF (run 3), ethanol (run 6), and acetonitrile (run 7), on the RAFT copolymerization of VBIM-Br and NIPAM was also investigated. Different solvents did not significantly affect the polymer molecular weight, as shown by the GPC curve comparison in Figure S4 of the Supporting Information. Moreover, different solvents resulted in PIL-Br copolymer with the same molar fraction of VBIM-Br at 0.64. Thus, the influence of solvent types on the copolymerization was negligible. All of the PIL-Br polymer samples synthesized in this work could be easily dissolved in acetonitrile, DMSO, DMF, alcohols, and H2O. An anion exchange of Br− to amino acid was carried out, based on the procedure reported in the literature.60 Complete exchange from Br− to amino acid was evident by conducting AgNO3/HNO3 test. The result of anion exchange was also confirmed by 1H and 13C NMR spectra as shown in Figure 4. The peak at 2.94−3.04 (ascribed to the proton of N−CH(C)− COO− of the tertiary carbon) and the peak at 1.04−1.12 (for the methyl protons of the anion) appeared in 1H NMR spectra. E

DOI: 10.1021/ma502611s Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

solutions containing PILs of different molecular weights, as determined by UV−vis spectroscopy. The upper critical solution temperature (Tc) is commonly defined as the temperature at which the transmittance dropped below 90%.61 The Tc values of poly(VBIM-Ala8-r-NIPAM5), poly(VBIM-Ala15-r-NIPAM8), and poly(VBIM-Ala26-r-NIPAM15) were 25.7, 31.4, and 34.8 °C. This trend can be attributed to the intermolecular hydrogen bonding, which is easier to form with longer polymer chains, thereby resulting in higher Tc values. Figure 6b shows that the Tc value decreased (32.4, 31.4, and 28.8 °C) with increasing NIPAM segment (from 5 to 25). Extraction of Tocopherol Homologues. The natural product of tocopherols is a mixture composed of α-, β-, γ-, and δ-tocopherols. The β- and γ-tocopherols have the same number of methyls and their peak areas in HPLC are overlapped. However, a slight difference in the methyl position results in different hydrogen bond acidity.62 The PILs synthesized in this work have hydrogen-bond basicity, with the anions interacting with −OH of tocopherols, forming hydrogen bonds in the order of δ-tocopherol > β- and γ-tocopherol > α-tocopherol. The distribution coefficient (Di) of tocopherol i and the selectivity coefficient (Sij) of tocopherol i to tocopherol j are defined as Di = Cie/Cir and Sij = Di/Dj, respectively, where Ci is the mass fraction of tocopherol i (i.e., the mass of tocopherol i over the total mass of the phase) with the superscripts “e” and “r” representing the extract and raffinate phases, respectively. The extraction efficiency (Ei) of tocopherol i is defined as Ei = Cir/(Cir + Cie) (note: this definition only applies to cases with equal masses of the two phases). The distribution coefficients differed greatly among the different tocopherols, as shown in Table 2. The coefficients decreased in the order of δ-tocopherol > β- and γ-tocopherol > α-tocopherol. That is, β-, γ-, and δ-tocopherols were favorably transferred to the extract phase, while α-tocopherol preferred to remain in the raw liquid. As a result, α-tocopherol was concentrated in the raffinate phase. Taking run 5 as an example, the polymer extractant contained 8 VBIM-Ala units and 5 NIPAM units. The mole ratio of VBIM-Ala units to acetonitrile was 2:98. The distribution coefficients of δ-, β&γ-, and αtocopherols were 7.86, 3.63, and 0.60, respectively, which translate to 88.7%, 78.4%, and 37.5% of these tocopherols were transferred from the raw liquid to the extract phase. The mole ratio of α-tocopherol in the total tocopherols in hexane increased from 0.08 to 0.27 after one stage extraction. A higher purity α-tocopherol could thus be obtained after multiple stages of the extraction. The effects of PIL’s molecular weight and composition on the extraction efficiency were investigated. PILs with different molecular weight (poly(VBIM-Ala8-r-NIPAM5), poly(VBIMAla15-r-NIPAM8), and poly(VBIM-Ala26-r-NIPAM15)) but similar molar ratios of VBIM-Ala to NIPAM (around 0.63) were tested. The distribution coefficients of tocopherols slightly decreased with polymer chain length, while the selectivity coefficients remained similar. Moreover, as shown in Table 2, all the PILs gave much higher distribution and selectivity coefficients than pure ACN. For example, the Dδ value of poly(VBIM-Ala8-r-NIPAM5) was 7.86, i.e., about 15.4 times higher than that of pure ACN. The Sδ value was also 3.3 times higher than that of pure ACN. The ionic liquid 1-butyl-3methylimidazolium chloride (BMIM-Cl) has been applied for the tocopherol homologue separation.36 The BMIM-Cl/ methanol solution (1:2.7 in mol) as extract phase possessed Dδ = 2.58 and Sδ = 16.1. The PILs had higher distribution

Figure 4. NMR spectra of poly(VBIM-Ala15-r-NIPAM8) (run 2 PIL): (a) 1H NMR and (b) 13C NMR.

No peaks in these two positions were present in 1H NMR spectra of the PIL-Br copolymers before anion modification. The result was further confirmed by 13C NMR as shown in Figure 4b. The peaks at 179.6 (COO−), 49.1 (due to tertiary carbon of N−CH(C)−COO−), and 22.5 (methyl) were observed in the copolymer samples after anion modification. Thermo-Responsivity of PILs. The synthesized copolymers having VBIM-Ala and NIPAM showed a good thermoresponsivity in acetonitrile. As shown in Figure 5, the

Figure 5. Reversible phase separation of PILs in acetonitrile upon temperature change.

PIL was well dissolved in acetonitrile at a relatively high temperature, but the solution separated into two phases when it was cooled to ambient temperature. The phase separation was caused by the hydrogen bondings between the anions of VBIMAla units as well as those between the anions of VBIM-Ala units and the secondary amines of NIPAM units. The amino acid anion played an essential role, since there was no thermoresponsive behavior observed for the halide anion copolymers, PIL-Br in acetonitrile. The phase change of PILs in acetonitrile upon cooling depends on the polymer molecular weight and composition. Figure 6a shows the change in transmittance of the acetonitrile F

DOI: 10.1021/ma502611s Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

Figure 6. Transmittance (A) change of different PILs in acetonitrile upon cooling: the molar ratio of VBIM-Ala:acetonitrile was 2:98: (a) VBIMAla:NIPAM = 8:5 (purple), 15:8 (red), and 26:15 (blue); (b) VBIM-Ala:NIPAM = 14:25 (black), 15:8 (red), and 16:5 (green).

Table 2. Results of the Extractive Separation of Tocopherol Homologues with PILsa distribution coefficient (D)/extraction efficiency (E) run

extractantb

E1 E2 E3 E4 E5 E6 E7 E8 E9

pure ACN VBIM-Ala poly(VBIM-Ala26-r-NIPAM15) poly(VBIM-Ala15-r-NIPAM8) poly(VBIM-Ala8-r-NIPAM5) poly(VBIM-Ala16-r-NIPAM5) poly(VBIM-Ala14-r-NIPAM25) poly(VBIM-Ala15-r-NIPAM8) poly(VBIM-Ala15-r-NIPAM8)

selectivity coefficient (S)

[VBIM-Ala]0:[ACN]0

Dδ/Eδ

Dβ&γ/Eβ&γ

Dα/Eα

Sδ/α

Sβ&γ/α

2:98 2:98 2:98 2:98 2:98 2:98 1:99 4:96

0.48/32.4% 12.6/92.7% 6.82/87.2% 7.20/87.8% 7.86/88.7% 6.91/87.4% 7.31/88.0% 6.24/86.1% 9.66/90.6%

0.28/21.9% 6.58/86.8% 3.15/75.9% 3.41/77.3% 3.63/78.4% 3.29/76.7% 3.66/78.5% 3.2/76.2% 3.74/78.9%

0.17/14.5% 1.01/50.2% 0.54/35.2% 0.59/37.1% 0.60/37.5% 0.57/36.3% 0.61/37.9% 0.52/34.2% 0.65/39.4%

2.9 12.6 12.6 12.2 13.0 12.1 12.0 12.0 11.9

1.6 6.6 5.8 5.8 6.0 5.8 6.0 6.1 5.8

a Initial tocopherol concentrations in hexane (mg/mL): δ 1.1, β&γ 1.15, α 0.22, volume ratio of the two phases: 1:1, 35 °C, mixture shaken for 40 min. bIn run 1: pure acetonitrile used as the extract phase; in run 2: VBIM-Ala used as the extractant; in all other runs, poly(VBIM-Alai-r-NIPAMj) represents the random copolymer containing i VBIM-Ala units and j NIPAM units on average.

Scheme 3. Representative Cyclic Extractive Separation for Tocopherol Homologues Using Poly(VBIM-Ala8-r-NIPAM5) as Extractant

coefficients with slightly lower selectivity coefficients than the BMIM-Cl at a higher concentration (2 mol % versus 27 mol %), leading to less PIL extractant required for the separation.

The separation efficiencies between the VBIM-Ala monomer and the PIL as extractants were also compared. Taking run 3 as an example, the Dδ value of poly(VBIM-Ala8-r-NIPAM5) was G

DOI: 10.1021/ma502611s Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules about half of the monomer; that is, only 5.5% less δ-tocopherol (extraction efficiency from 92.7% to 87.2% in Table 2) was transferred from the raw liquid to the raffinate phase. The extraction performances of three PIL extractants having the same number of VBIM-Ala but different numbers of NIPAM were compared. It is evident from runs 4, 6, and 7 in Table 2 that increasing the NIPAM units slightly decreased the distribution coefficient but had no significant change to the selectivity coefficient. The primary function of NIPAM units was to improve solubility of the copolymers in acetonitrile. With more NIPAM units, better solubility was achieved. Without NIPAM, the PILs were not soluble in acetonitrile. It is plausible that NIPAM did not adversely affect the distribution and selectivity coefficients. The effect of the PIL concentrations (runs 8, 4, and 9) on the separation efficiency was also investigated, as shown in Table 2. The distribution coefficients increased with increasing PIL concentration, while no change was observed for the selectivity coefficients. Regeneration and Recycle of the PIL Extractants. Recovery of extractant from the extract phase through backextraction is one of the key steps in LLE processes. The backextraction requires a large amount of organic solvents and highenergy consumption. The major advantage of the PILs synthesized here stems from their thermoresponsivity. These PILs possessed UCST (Tc) around 25−35 °C, depending of the polymer molecular weight and composition. The cyclic extractive separation is shown in Scheme 3. During the extraction stage, the PILs were well dissolved at 35 °C in acetonitrile to achieve a high extractive efficiency, resulting in high-purity α-tocopherol in the raffinate phase. After removing the raffinate phase, the tocopherols in the extract phase were separated from the PILs extractants by lowering the temperature of the solution. Temperature was lowered by 15 °C to ensure complete separation and regeneration of the PIL extractants. Possible loss of the extracted tocopherols in the PIL regeneration step was considered, i.e., whether the tocopherols precipitated out with the PILs. To ensure tocopherols stay in the acetonitrile, 2% water was added into the extract phase to break hydrogen bonds between tocopherols and VBIM-Ala units of the extractant, prior to cooling.16 The recovery efficiency (Ri) of tocopherol i was defined as Ri = Cie/Cie′, where Cie and Cie′ are the mass fractions of tocopherol i in the extract phase before and after removal of the extractant, respectively. It was found that all the recovery efficiencies of the extracted tocopherols in the regeneration of PIL extractants were very high (>95% in most cases). In other words, almost all of the tocopherols remain in the acetonitrile. It is noteworthy that if VBIM-Ala monomer was used as extractant instead of the polymeric form, more than 20 times of back-extraction by fresh hexane are theoretically required for a recovery efficiency >90%. Eight times of back extraction are needed for the BMIMCl extractant system to achieve recovery efficiency over 90%.36 This clearly displays the major advantage in using PIL as extractant for tocopherol separation. There was some inevitable loss of the PIL in the regeneration step due to limited solubility of the PILs in acetonitrile, even below their UCST’s. The mass losses for random copolymer poly(VBIM-Ala8-r-NIPAM5), poly(VBIM-Ala15-r-NIPAM8), and poly(VBIM-Ala26-r-NIPAM15) after three cycles were 7.2%, 4.5%, and 4.6%, respectively. This level of losses is acceptable and comparable to any extractant recycling. It becomes clear that increasing the polymer molecular weight at

the same copolymer composition yielded better regeneration of the PIL extractants. On the other hand, an adequate number of NIPAM is essential to ensure PILs is soluble and have UCST in acetonitrile. However, too many NIPAM would result in significant losses of PIL during regeneration. Indeed, the mass loss of poly(VBIM-Ala14-r-NIPAM25) after three cycles increased to 15.8%, due to its high solubility in acetonitrile. The recyclability of the regenerated PIL extractants was also investigated. Table 3 shows the experimental data of polyTable 3. Change in Distribution, Selectivity, and Recovery Coefficients in the Three Extraction Cycles of Tocopherol Homologues Using Poly(VBIM-Ala15-r-NIPAM8) as Extractanta distribution coefficient

selectivity coefficient

recovery efficiency

cycle

Dδ

Dβ&γ

Dα

Sδ/α

Sβ&γ/α

Rα (%)

Rβ&γ (%)

Rδ (%)

1 2 3

7.20 7.16 7.12

3.41 3.38 3.38

0.59 0.56 0.58

12.2 12.8 12.2

5.8 6.0 5.8

100 93.3 96.8

96.9 93.7 93.7

96.2 95.7 93.9

Initial tocopherol concentrations in hexane (mg/mL): δ 1.1, β&γ 1.15, α 0.22, volume ratio of the two phases: 1:1, 35 °C, mixture shaken for 40 min. Molar ratio of VBIM-Ala units to ACN was 2:98.

a

(VBIM-Ala15-r-NIPAM8) used for three extraction cycles. No significant changes in both distribution and selectivity coefficients were observed in each cycle. The recovery efficiency of tocopherols in the regeneration step of each cycle also remained satisfactory.

■

CONCLUSIONS In summary, a series of ionic liquid copolymers with 0.36−0.76 molar fraction of ionic liquid and number-average molecular weight of 2.70−8.17 kg/mol were synthesized by RAFT copolymerization of VBIM-Br and NIPAM, followed by an anion exchange of bromide to amino acid D-alanine. The RAFT copolymerizations showed good living characteristics. The PILs possessed narrow PDI of 1.12−1.25 with UCST in acetonitrile between 25.7 and 34.8 °C. The UCST increased with molecular weight of PIL and VBIM-Ala content. The PILs were dissolved in acetonitrile and used to extract tocopherols from hexane. The distribution coefficient of δtocopherol in the acetonitrile and hexane phases reached 7.86, and the selectivity coefficient of δ-tocopherol to α-tocopherol reached 13.0, while Dβ&γ was 3.63 and Sβ&γ/α was 6.0, respectively. The mole ratio of α-tocopherol in tocopherols increased from 0.08 to 0.27 in hexane after one stage extraction. The thermoresponsive property greatly facilitated the separation of extracted tocopherols and PIL extractants. The PILs were precipitated out from the extract phase by lowering the temperature by 15 °C and released the extracted tocopherols to acetonitrile. A negligible amount of tocopherols was precipitated together with the PIL chains. Moreover, the loss of the PIL extractants was minimal and depended on the polymer molecular weight and composition. Only with a high NIPAM content was the loss of PIL in the extractant regeneration step significant because of its high solubility in acetonitrile. The regenerated extractants have good recyclability. After three cycles, the recycled PILs extractant did not show significant change in the distribution and selectivity coefficients as well as in the recovery efficiency. H

DOI: 10.1021/ma502611s Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

(16) Carrasco, P. M.; Ruiz de Luzuriaga, A.; Constantinou, M.; Georgopanos, P.; Rangou, S.; Avgeropoulos, A.; Zafeiropoulos, N. E.; Grande, H. J.; Cabanero, G.; Mecerreyes, D.; Garcia, I. Macromolecules 2011, 44, 4936−4941. (17) Zhang, H. W.; Hong, K. L.; Mays, J. W. Macromolecules 2002, 35, 5738−5741. (18) Vijayakrishna, K.; Jewrajka, S. K.; Ruiz, A.; Marcilla, R.; Pomposo, J. A.; Mecerreyes, D.; Taton, D.; Gnanou, Y. Macromolecules 2008, 41, 6299−6308. (19) Mori, H.; Yahagi, M.; Endo, T. Macromolecules 2009, 42, 8082− 8092. (20) He, H. K.; Luebke, D.; Nuwala, H.; Matyjaszewski, K. Macromolecules 2014, 47, 6601−6609. (21) Texter, J.; Vasantha, V. A.; Crombez, R.; Maniglia, R.; Slater, L.; Mourey, T. Macromol. Rapid Commun. 2012, 33, 69−74. (22) Weber, R. L.; Ye, Y.; Schmitt, A. L.; Banik, S. M.; Elabd, Y. A.; Mahanthappa, M. K. Macromolecules 2011, 44, 5727−5735. (23) Nakamura, Y.; Nakanishi, K.; Yamago, S.; Tsujii, Y.; Takahashi, K.; Morinaga, T.; Sato, T. Macromol. Rapid Commun. 2014, 35, 642− 648. (24) Moad, G.; Rizzardo, E.; Thang, S. H. Aust. J. Chem. 2005, 58, 379−410. (25) Wang, W.-J.; Wang, D.; Li, B.-G.; Zhu, S. Macromolecules 2010, 43, 4062−4069. (26) Newman, D. J.; Cragg, G. M. J. Nat. Prod. 2007, 70, 461−477. (27) Azzi, A.; Stocker, A. Prog. Lipid Res. 2000, 39, 231−255. (28) Brigelius-Flohe, R.; Traber, M. G. FASEB J. 1999, 13, 1145− 1155. (29) Traber, M. G.; Stevens, J. F. Free Radical Biol. Med. 2011, 51, 1000−1013. (30) Valdivielsoa, I.; Bustamanteb, M. A.; Gordoab, J. C. R.; Nájeraa, A. I.; Renobalesb, M.; Barrona, L. J. R. Food Chem. 2015, 173, 709− 717. (31) Gornas, P.; Soliven, A.; Seglina, D. Eur. J. Lipid Sci. Technol. 2014, DOI: 10.1002/ejlt.201400566. (32) Anthemidis, A. N.; Ioannou, K. I. Talanta 2009, 80, 413−421. (33) Tan, S. S. Y.; MacFarlane, D. R.; Upfal, J.; Edye, L. A.; Doherty, W. O. S.; Patti, A. F.; Pringle, J. M.; Scott, J. L. Green Chem. 2009, 11, 339−345. (34) Qin, Y.; Lu, X.; Sun, N.; Rogers, R. D. Green Chem. 2010, 12, 968−971. (35) Abe, M.; Fukaya, Y.; Ohno, H. Green Chem. 2010, 12, 1274− 1280. (36) Yang, Q.; Xing, H.; Cao, Y.; Su, B.; Yang, Y.; Ren, Q. Ind. Eng. Chem. Res. 2009, 48, 6417−6422. (37) Pereira, J. F. B.; Lima, Á . S.; Freire, M. G.; Coutinho, J. A. P. Green Chem. 2010, 12, 1661−1669. (38) Du, Z.; Yu, Y. L.; Wang, J. H. Chem.Eur. J. 2007, 13, 2130− 2137. (39) Berthod, A.; Ruiz-Angel, M. J.; Carda-Broch, S. J. Chromatogr. A 2008, 1184, 6−18. (40) Bi, W.; Tian, M.; Row, K.-H. J. Chromatogr. B 2013, 913, 61−68. (41) Brown, R. A.; Pollet, P.; McKoon, E.; Eckert, C. A.; Liotta, C. L.; Jessop, P. G. J. Am. Chem. Soc. 2001, 123, 1254−1255. (42) Kroon, M. C.; van Spronsen, J.; Peters, C. J.; Sheldon, R. A.; Witkamp, G. J. Green Chem. 2006, 8, 246−249. (43) Haerens, K.; Van Deuren, S.; Matthijs, E.; Vander Bruggen, B. Green Chem. 2010, 12, 2182−2188. (44) Claudio, A. F. M.; Marques, C. F. C.; Boal-Palheiros, I.; Freire, M. G.; Coutinho, J. A. P. Green Chem. 2014, 16, 259−268. (45) Garcia, M. T.; Gathergood, N.; Scammells, P. Green Chem. 2005, 7, 9−14. (46) Petkovic, M.; Seddon, K. R.; Rebelo, L. P. N.; Pereira, C. S. Chem. Soc. Rev. 2011, 40, 1383−1403. (47) Vihola, H.; Laukkanen, A.; Valtola, L.; Tenhu, H.; Hirvonen, J. Biomaterials 2005, 26, 3055−3064. (48) Zhang, Q.; Wang, W.; Lu, Y.; Li, B.; Zhu, S. Macromolecules 2011, 44, 6540545.

This innovation in extract regeneration is of great benefit in saving energy and reducing the use of volatile organic solvents, which are normally associated with the multistage backextraction using small molecular organic solvents. In contrast to the conventional liquid−liquid extraction, which used ionic liquids as the extractant, the use of thermoresponsive PILs eliminates the need of any back-extraction process. In addition, polymers are generally less toxic than small molecules like monomers. This work is believed to represent a significant progress in designing smart polymers in organic solvent for separation of natural products through greener unit operations, such as liquid−liquid extraction.

■

ASSOCIATED CONTENT

S Supporting Information *

Details about 1H and 13C NMR spectra of ionic liquid VBIMBr, GPC traces of PIL-Brs at different conversions, and GPC traces of PIL-Brs synthesized at different solvents. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Authors

*E-mail [email protected] (W.-J.W.). *E-mail [email protected] (S.Z.). Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The authors thank the National Natural Science Foundation of China (Grants 21376211 and 20976153, Key Grant 21420102008), Chinese State Key Laboratory of Chemical Engineering at Zhejiang University (Grants SKL-ChE-12T05 and SKL-ChE-14D01), and the Program for Changjiang Scholars and Innovative Research Team in University in China (IRT0942) for financial support. We are grateful to Prof. Huabing Xing and Prof. Qiwei Yang for their helpful discussions.

■

REFERENCES

(1) Huddleston, J. G.; Willauer, H. D.; Swatloski, R. P.; Visser, A. E.; Rogers, R. D. Chem. Commun. 1998, 16, 1765−1766. (2) Wilkes, J. S. Green Chem. 2002, 4, 73−80. (3) Greaves, T. L.; Drummond, C. J. Chem. Rev. 2008, 108, 206−237. (4) Armand, M.; Endres, F.; MacFarlane, D. R.; Ohno, H.; Scrosati, B. Nat. Mater. 2009, 8, 621−629. (5) Hallett, J. P.; Welton, T. Chem. Rev. 2011, 111, 3508−3576. (6) Han, X.; Armstrong, D. Acc. Chem. Res. 2007, 40, 1079−1086. (7) Bara, J. E.; Camper, D. E.; Gin, D. L.; Noble, R. D. Acc. Chem. Res. 2010, 43, 152−159. (8) Yuan, J.; Mecerreyes, D.; Antonietti, M. Prog. Polym. Sci. 2013, 38, 1009−1036. (9) Mecerreyes, D. Prog. Polym. Sci. 2011, 36, 1629−1648. (10) Tang, J.; Sun, W.; Tang, H.; Radosz, M.; Shen, Y. Macromolecules 2005, 38, 2037−2039. (11) Fellinger, T. P.; Thomas, A.; Yuan, J.; Antonietti, M. Adv. Mater. 2013, 25, 5838−5855. (12) Yoshimitsu, H.; Kanazawa, A.; Kanaoka, S.; Aoshima, S. Macromolecules 2012, 45, 9427−9434. (13) Sui, X. F.; Hempenius, M. A.; Vancso, G. J. Am. Chem. Soc. 2012, 134, 4023−4025. (14) Appetecchi, G. B.; Kim, G. T.; Montanino, M.; Carewska, M.; Marcilla, R.; Mecerreyes, D.; De Meatza, I. J. Power Sources 2010, 195, 3668−3675. (15) Yoshizawa, M.; Hirao, M.; Ito-Akita, K.; Ohno, H. J. Mater. Chem. 2001, 11, 1057−1062. I

DOI: 10.1021/ma502611s Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules (49) Zhang, Q.; Yu, G.; Wang, W.; Yuan, H.; Li, B.; Zhu, S. Langmuir 2012, 28, 5940−5946. (50) Zhang, Q.; Yu, G.; Wang, W.; Li, B.; Zhu, S. Macromol. Rapid Commun. 2012, 33, 916−921. (51) Zhang, Q.; Yu, G.; Wang, W.; Yuan, H.; Li, B.; Zhu, S. Macromolecules 2013, 46, 1261−1267. (52) Liu, P.; Lu, W.; Wang, W.; Li, B.; Zhu, S. Langmuir 2014, 30, 10248−10255. (53) Jessop, P. G.; Heldebrant, D. J.; Li, X.; Eckert, C. A.; Liotta, C. L. Nature 2005, 436, 1102. (54) Liu, Y.; Jessop, P. G.; Cunningham, M.; Eckert, C. A.; Liotta, C. L. Science 2006, 313, 958−960. (55) Su, X.; Jessop, P. G.; Cunningham, M. Macromolecules 2012, 45, 666−670. (56) Pinaud, J.; Kowal, E.; Cunningham, M.; Jessop, P. ACS Macro Lett. 2012, 1, 1103−1107. (57) Yu, G.; Lu, Y.; Liu, X.; Wang, W.; Yang, Q.; Xing, H.; Ren, Q.; Li, B.; Zhu, S. Ind. Eng. Chem. Res. 2014, 53, 16025−16032. (58) Marcilla, R.; Alberto Blazquez, J.; Rodriguez, J.; Pomposo, J. A.; Mecerreyes, D. J. Polym. Sci., Part A: Polym. Chem. 2004, 42, 208−212. (59) Nese, A.; Kwak, Y.; Nicolaÿ, R.; Barrett, M.; Sheiko, S. S.; Matyjaszewski, K. Macromolecules 2010, 43, 4016−4019. (60) Fukumoto, K.; Yoshizawa, M.; Ohno, H. J. Am. Chem. Soc. 2005, 127, 2398−2399. (61) Kohno, Y.; Deguchi, Y.; Ohno, H. Chem. Commun. 2012, 48, 11883−11885. (62) Yang, Q.; Xing, H.; Su, B.; Bao, Z.; Wang, J.; Yang, Y.; Ren, Q. AIChE J. 2013, 59, 1657−1667.

J

DOI: 10.1021/ma502611s Macromolecules XXXX, XXX, XXX−XXX