Novel Approach for the Separation of Shape-Constrained Isomers with

Anal. Chem. 2010, 82, 3320–3328

Novel Approach for the Separation of Shape-Constrained Isomers with Alternating Copolymer-Grafted Silica in Reversed-Phase Liquid Chromatography Abul K. Mallik, Tsuyoshi Sawada, Makoto Takafuji, and Hirotaka Ihara* Department of Applied Chemistry and Biochemistry, Faculty of Engineering, Kumamoto University, 2-39-1 Kurokami, Kumamoto 860-8555, Japan This paper describes a novel packing material for high selective reversed-phase high-performance liquid chromatography (RP-HPLC). The organic phase on silica is chemically designed in a way that the weak interaction sites are integrated with high orientation along the polymer main chain and high selectivity can be realized by multiple interactions with solutes. For the above purpose, we synthesized poly(octadecyl acrylate-alt-Noctadecylmaleimide)-grafted silica (Sil-poly(ODA-alt-OMI)) stationary phase. The alternating copolymerization was carried out from 3-marcaptopropyltrimethoxysilane (MPS)modified silica via surface-initiated radical-chain transfer reaction. Elemental analysis, diffuse reflectance infrared Fourier transform (DRIFT),1H NMR, solidstate 13C cross polarization magic angle spinning (CP/MAS) NMR, and suspended-state 1H NMR were used to characterize the new organic phase. Aspects of shape selectivity was evaluated with Standard Reference Material (SRM 869b), Column Selectivity Test Mixture for Liquid Chromatography. Enhanced molecular shape selectivity was observed, that lead to the separation of SRM 1647e (16 polycyclic aromatic hydrocarbons, PAHs) in an isocratic elution. The effectiveness of this phase was also demonstrated by the separation of several β-carotene and tocopherol isomers. The complete baseline separation of the tocopherol isomers was achieved using the Sil-poly(ODA-alt-OMI) phase. Chromatographic study revealed that Sil-poly(ODA-alt-OMI) has extremely high separation ability compared to monomeric and polymeric C18 columns. Higher shape selectivity of the new RP material can be explained by a π-π and dipole-dipole interaction mechanism. The separation of a certain class of isomers that are important in the disciplines of environmental, clinical, and food science is one of the most challenging issues facing the analyst, because isomerism gives structurally similar compounds that sometimes differ negligibly in their physicochemical properties. However, their separation is essential. For example, carotenoids * Corresponding kumamoto-u.ac.jp.

3320

author.

Tel./Fax:

+81-96-342-3661.

E-mail:

Analytical Chemistry, Vol. 82, No. 8, April 15, 2010

ihara@

are thought to diminish the incidence of certain degenerative diseases,1 and though in common foods they are mostly found in the form of all trans-isomers,2 significant amounts of geometrical cis-isomers can be produced during processing.3 The cis-isomers of β-carotene have attracted attention because their bioavailability is absolutely different from that of their corresponding trans-isomers,4,5 and they may have specific functions.6-9 Separation of the isomers of polycyclic aromatic hydrocarbons (PAHs), vitamin E (tocopherol), retinoids, vitamin A derivatives, etc., also possess similar challanges.10-17 One of the easiest methods of separating these isomers may be the direct use of biosystems such as enzymes,18 antibodies,19 (1) Tyssandier, V.; Reboul, E.; Dumas, J.-F.; Demange, C. B.; Armand, M.; Marcand, J.; Sallas, M.; Borel, P. Am. J. Physiol. Gastrointest. Liver Physiol. 2003, 284, G913–G923. (2) Lessin, W. J.; Catigani, G. L.; Schwartz, S. J. Agric. Food Chem. 1997, 45, 3728–3732. (3) Shi, J.; Le Maguer, M. Crit. Rev. Biotechnol. 2000, 20, 293–334. (4) Gaziano, J. M.; Johnson, E. J.; Russell, R. M.; Manson, J. E.; Stampfer, M. J.; Ridker, P. M.; Frei, B.; Hennekens, C. H.; Krinsky, N. I. Am. J. Clin. Nutr. 1995, 61, 1248–1252. (5) Stahl, W.; Schwarz, W.; Vonlaar, J.; Sies, H. J. Nutr. 1995, 125, 2128– 2133. (6) Hieber, A. D.; King, T. J.; Morioka, S.; Fukushima, L. H.; Franke, A. A.; Bertram, J. S. Nutr. Cancer 2000, 37, 234–244. (7) Lavy, A.; Amotz, A. B.; Aviram, M. Eur. J. Clin. Chem. Clin. Biochem. 1993, 31, 83–90. (8) Levin, G.; Mokady, S. Free Radical Biol. Med. 1994, 17, 77–82. (9) Levin, G.; Yeshurun, M.; Mokady, S. Nutr. Cancer 1997, 27, 293–297. (10) Poster, D. L.; Sander, L. C.; Wise, S. A. In The Handbook of Environmental Chemistry: Vol. 3. Anthropogenic Compounds; Hutzinger, O., Ed.; SpringerVerlag: Berlin, 1998; Chapter 3. (11) Jacob, J.; Karcher, W.; Belliardo, J. J.; Wagsttffe, P. J. Fresenius’ J. Anal. Chem. 1986, 323, 1–10. (12) Wise, S. A.; Sander, L. C.; May, W. E. J. Chromatogr., A 1993, 642, 329– 349. (13) Leinster, P.; Evans, M. J. Ann. Occup. Hyg. 1986, 30, 481–495. (14) Nubbe, M. E.; Adams, V. D.; Watts, R. J.; Robinet Clark, Y. S. J. Water Pollut. Control Fed. 1988, 60, 773–796. (15) Strohschein, S.; Pursch, M.; Lubda, D.; Albert, K. Anal. Chem. 1998, 70, 13–18. (16) Henry, C. W.; Fortier, C. A.; Warner, I. M. Anal. Chem. 2001, 73, 6077– 6082. (17) Pursch, M.; Strohschein, S.; Ha¨ndel, H.; Albert, K. Anal. Chem. 1996, 68, 386–393. (18) Krenkova, J.; Lacher, N. A.; Svec, F. Anal. Chem. 2009, 81, 2004–2012. (19) Evans, D. M.; Williams, K. P.; McGuinness, B.; Tarr, G.; Regnier, F.; Afeyan, N.; Jindal, S. Nat. Biotechnol. 1996, 14, 504–507. (20) Fassina, G.; Ruvo, M.; Palombo, G.; Verdoliva, A.; Marino, M. J. Biochem. Biophys. Methods 2001, 49, 481–490. 10.1021/ac1001178  2010 American Chemical Society Published on Web 03/25/2010

and proteins;20 however, this method has its disadvantages, such as a too narrow selectivity, low stability, and limited scope of use. Another approach to isomer separation may be based on biomimetic approaches. For example, biomembrane systems are very attractive to mimic for a separation because their selfassembling structures yield various supramolecular functions. Pidgeon and Venkataram have introduced phospholipids onto silica as packing material for liquid chromatography in order to endow silica with specific characteristics that mimic biological membranes.21 However, the main drawback of the direct immobilization of membrane-forming components onto silica is their prohibition of the lateral diffusion of lipids, which produce highly ordered structures leading to supramolecular functions of lipid membrane systems. In order to achieve both stability and fluidity of this organic phase on silica, this drawback must be overcome. On the basis of these facts, we have designed and synthesized poly(octadecyl acrylate)-grafted silica (Sil-ODAn)22 as a lipid membrane analogous organic phase. This organic phase does not resemble lipid membranes; however, it exhibits not only an ordered-to-disordered phase transition similar to that in lipid membranes but also exceptionally enhanced selectivity for polcyclic aromatic hydrocarbons (PAHs) via multiple carbonyl-π interactions at lower temperature or ordered state.23-25 However, the separation of the shape-constrained isomers of large molecules like β-carotene and tocopherol cannot be realized. Therefore, the development of new types of synthetic organic phases has long been desired in order to achieve complex isomer separation. It has been a long-standing goal of analysts to find high selective systems with a new organic phase. In order to realize the concept behind our solution, we use an alternating copolymer, which can be spontaneously obtained by the copolymerization of an electron-rich (donor type) monomer and an electron-deficient (acceptor type) monomer through the formation of charge transfer complexes (CTCs).26-28 Alternating copolymerization mediated by electron donor-acceptor interactions remains a popular approach for achieving special materials with different functionalities in a wide range of scientific fields including nanotechnology29-32 and drug delivery.33,34 In this paper, we report a new class of high-performance liquid chromatography (HPLC) stationary phase as the first application (21) Pidgeon, C.; Venkataram, U. V. Anal. Biochem. 1989, 176, 36–47. (22) Hirayama, C.; Ihara, H.; Mukai, T. Macromolecules 1992, 25, 6375–6376. (23) Chowdhury, M. A. J.; Boyson, R. I.; Ihara, H.; Hearn, M. T. W. J. Phys. Chem., B 2002, 106, 11936–11950. (24) Ihara, H.; Goto, Y.; Sakurai, T.; Takafuji, M.; Sagawa, T.; Nagaoka, S. Chem. Lett. 2001, 1252–1253. (25) Ihara, H.; Sagawa, T.; Goto, Y.; Nagaoka, S. Polymer 1999, 40, 2555–2560. (26) Butler, G. B.; Olson, K. G.; Tu, C.-L. Macromolecules 1984, 17, 1884–1887. (27) Hill, D. J. T.; O’Donnell, J. H.; O’Sullivan, P. W. Macromolecules 1985, 18, 9–17. (28) Prementine, G. S.; Jones, S. A.; Tirrel, D. A. Macromolecules 1989, 22, 770–775. (29) Huck, W. T. S.; Yan, L.; Stroock, A.; Haag, R.; Whitesides, G. M. Langmuir 1999, 15, 6862–6867. (30) Blomberg, S.; Ostberg, S.; Hart, E.; Bosman, A. W.; Van Horn, B.; Hawker, C. J. J. Polym. Sci., Part A: Polym. Chem. 2002, 40, 1309–1320. (31) Tian, Y.; He, Q.; Tao, C.; Cui, Y.; Ai, S.; Li, J. J. Nanosci. Nanaotechnol. 2006, 6, 2072–2076. (32) Wang, M.; Braun, H.-G.; Meyer, E. Chem. Mater. 2002, 14, 4812–4818. (33) Harrisson, S.; Wooley, K. L. Chem. Commun. 2005, 3259–3260. (34) Henry, S. M.; El-Sayed, M. E. H.; Pirie, C. M.; Hoffman, A. S.; Stayton, P. S. Biomecromolecules 2006, 7, 2407–2414.

of alternating copolymerization on silica. This approach can be used to enhance the selective interaction via multiple interactions based on the precise arrangement of weak interaction groups. As a result, we can demonstrate exceptionally good selectivity toward extremely difficult separation of 16 PAHs (SRM1647e),35,36 isomers of β-carotene,37,38 and isomers of tocopherol.15,16 Here, we also report the first complete baseline separation of β- and γ-isomers of tocopherol with alternating copolymer, poly(octadecyl acrylate-alt-N-octadecylmaleimide)-grafted silica (Sil-poly(ODA-altOMI)) in reversed-phase high-performance liquid chromatography (RP-HPLC). EXPERIMENTAL SECTION Materials and Reagents. Alternating copolymers, poly(octadecyl acrylate-alt-N-octadecylmaleimide)-grafted silica (Sil-poly(ODAalt-OMI)) and poly(methyl acrylate-alt-methylmaleimide)-grafted silica (Sil-poly(MA-alt-MMI)) stationary phases were synthesized, characterized, and packed into a stainless steel column (150 × 4.6 mm i.d.). A YMC silica (YMC SIL-120-S5 having diameter 5 µm, pore size 12 nm, and surface area 300 m2 g-1) was used. Another phase of (Sil-poly(ODA-alt-OMI)) was prepared with 3 µm particle size silica and packed into a stainless steel column (150 × 4.6 mm i.d.). The silica having a diameter of 3 µm, pore size of 10 nm, and surface area of 313 m2 g-1 was a gift from Fuji Silysia Chemical Ltd. (Aichi, Japan). The columns with 5 and 3 µm particle size silicas were packed by the companies, Masis, Inc. (Aomori, Japan) and Fuji Silysia Chemical Ltd. (Aichi, Japan), respectively. The column efficiencies with 5 µm particle size packed column for uracil, toluene, and naphthalene were calculated to be 21 300/m, 55 400/m, and 57 000/m, respectively, using 70:30 (volume fraction) methanol/water at 25 °C and a flow rate of 1 mL/min. Very high column efficiencies were obtained with 3 µm particle size silica, such as 31 900/m, 117 900/m, and 122 300/m for uracil, ethylbenzene, and propylbenzene, respectively, using 80:20 (volume fraction) methanol/water at 25 °C and a flow rate of 0.8 mL/ min. For comparing chromatographic results, we have used two commercial monomeric and polymeric C18 columns. The monomeric C18 column (Inertsil, ODS 3, column size 250 × 4.6 mm i.d. with particle size 5.5 of µm, pore size of 10 nm, and surface area of silica particles of 450 m2 g-1) was purchased from G. L. Sciences (Tokyo, Japan). This contains 13.8% C in the bonded octadecyl phase. The polymeric C18 column (250 × 4.6 mm i.d., Shodex, C18 P, particle size of 5 µm, pore size of 10 nm, and surface area of 300 m2 g-1 with end-cap of the unreacted silanol group) containing 17.5% C was obtained from Shodex (Tokyo, Japan). N-Octadecylamine and β-carotene were purchased from Sigma (St. Louis, MO). Maleic anhydride, methylmaleimide (MMI), and methyl acrylate (MA) were purchased from TCI (Kyoto, Japan). Octadecyl acrylate (ODA) was obtained from Tokyo Kasei Kogyo (Tokyo, Japan) and used after removing polymerization inhibitor. Standard Reference (35) Wegmann, J.; Albert, K.; Pursch, M.; Sander, L. C. Anal. Chem. 2001, 73, 1814–1820. (36) Rimmer, C. A.; Sander, L. C.; Wise, S. A. Anal. Bioanal. Chem. 2005, 382, 698–707. (37) Sander, L. C.; Sharpless, K. E.; Craft, N. E.; Wise, S. A. Anal. Chem. 1994, 66, 1667–1674. (38) Meyer, C.; Skogsberg, U.; Welsch, N.; Albert, K. Anal. Bioanal. Chem. 2005, 382, 679–690.


3321

Figure 1. Reaction scheme for the synthesis of monomer (Noctadecylmaleimide) and alternating copolymer-grafted silica phases.

Material (SRM) 869b, Column Selectivity Test Mixture for Liquid Chromatography, and SRM 1647e, Priority Pollutant Polycyclic Aromatic Hydrocarbons, were obtained from the Standard Reference Materials Program (NIST, Gaithersburg, MD). The tocopherol isomers were obtained from Calbiochem (Darmstadt, Germany). For the sample preparation, trans-βcarotene was photoisomerized based on a literature procedure.39 HPLC-grade solvents were used in chromatographic separations. Synthesis of N-Octadecylmaleimide (OMI). The monomer N-octadecylmaleimide (OMI) was synthesized according to the previously reported method,40 and the reaction scheme is shown in Figure 1. A solution of N-octadecylamine (6.496 g, 6.14 mmol) in chloroform (50 mL) was added dropwise to a solution of maleic anhydride (9.904 g, 36.74 mmol) in chloroform (100 mL), and the mixture was stirred for 30 min. The mixture was vacuum filtered, and the solid was washed with CHCl3 (100 mL) and hexane (100 mL) to yield 1 as a white solid: yield 14.13 g (86%); mp 102-104 °C; νmax (KBr)/cm-1 3247 (br), 3077 (br), 2919, 2857, 1709, 1645, 1586, 1528, 1469, 1405, 1294, 1154; 1H NMR (400 MHz, DMSO) δ 0.84-0.87 (t, 3H), 1.15-1.48 (m, 30H), 2.49-2.51 (q, 2H), 3.14-3.19 (t, 2H), 6.22-6.25 (d, 1H), 6.38-6.42 (d, 1H). A mixture of 1 (10.820 g, 29.40 mmol), sodium acetate (4.100 g, 50.00 mmol), and acetic anhydride (200 mL) was heated to 100 °C for 1 h. Then, the mixture was cooled to room temperature and poured into an ice/water slurry. The obtained solid was vacuum filtered and purified by column chromatography (CHCl3, silica gel) to get N-octadecylmaleimide (OMI) as a white solid: yield 7.46 g (50%); mp 72-73 °C; νmax (KBr)/cm-1 2919, 2848, 1771, 1697, 1405; 1H NMR (400 MHz, CDCl3) δ 0.86-0.90 (t, 3H), 1.25 (m, 32H), 1.55-1.59 (m, 6H), 3.48-3.52 (t, 2H), 6.68 (s, 2H). Immobilization of MPS onto Silica. Dried silica gel (4.00 g) was placed to a dried flask and dispersed in dry toluene (30 (39) Zechmeister, L.; Polger, A. J. Am. Chem. Soc. 1943, 65, 1522–1528. (40) Vargas, M.; Kriegel, R.-M.; Collard, D. M.; Schiraldi, D. A. J. Polym. Sci., Part A: Polym. Chem. 2002, 40, 3256–3263.

3322


mL). Then, 3-marcaptopropyltrimethoxysilane (MPS; 1.2 g, 6.11 mmol) was added, and the reaction mixture was refluxed for 72 h. The suspension was filtered, and the solid was washed with toluene, methanol, water, methanol, and diethyl ether successively to obtain the MPS-modified silica (Sil-MPS) as shown in Figure 1. After drying, the particles were characterized by elemental analysis. Copolymerization of ODA and OMI by Surface-Initiated Chain Transfer Reaction. To a reaction flask containing MPSmodified silica (3.50 g), a mixture of ODA (4.65 g, 14.32 mmol), OMI (5.00 g, 14.32 mmol), 100 mg of radical initiator 2,2′azobisisobutyronitrile (AIBN), and dry toluene (10 mL) was added. The polymerization mixture was gently stirred for 24 h at 60 °C. The slightly yellow-white reaction mixture was filtered and washed successively with toluene, chloroform, and methanol. To remove nongrafted polymer, Soxhlet extraction was carried out with chloroform for 24 h and again washed with methanol and finally with diethyl ether and dried in vacuo to obtain Sil-poly(ODA-altOMI) for both 5 and 3 µm silica particles as shown in Figure 1. The same procedure was used for the monomers MA and MMI to obtain Sil-poly(MA-alt-MMI). Proton NMR. 1H NMR were recorded with a JEOL JNMLA400 (Japan) instrument at 400 MHz in CDCl3 solutions at 25 °C. Chemical shifts (δ) of 1H were expressed in parts per million (ppm) with use of the internal standard Me4Si (δ ) 0.00 ppm). FT-IR, DRIFT, and Elemental Analysis. FT-IR measurements were conducted with JASCO FT-IR-4100 (Japan). For diffuse reflectance infrared Fourier transform (DRIFT) measurement, accessory DR PRO410-M (JASCO, Japan) was used. Elemental analyses were carried out on a Yanaco CHN Corder MT-6 Apparatus (Japan). Solid-State 13C CP/MAS NMR. Adamantane was used for adjusting the magic angle before each experiment. NMR frequency referencing was performed by adjusting carbon peak of adamantine to 38.5 ppm. Representative samples of 200-250 mg were spun at 4000-4500 Hz using 7 mm double bearing ZrO2 rotors. Solid-state 13C cross polarization magic angle spinning (CP/MAS) NMR spectra were measured at temperatures: 20, 25, 30, 35, 40, 45, and 50 °C using a line-broadening factor of 5 (lb ) 5). Other important parameters were relaxation delay of 4.0 s, pulse of 85.5 degrees, acquisition time of 0.05 s, and spectral width of 40282.0 Hz. High power proton decoupling of 63 db with fine attenuation of dipole r ) 2500 was used only during detection periods. Suspended-State 1H NMR. Suspended-state 1H NMR Spectra were measured at temperatures: 20, 25, 30, 35, 40, 45, and 50 °C using GHX Varian AS400 nanoprobe. Four milligrams of samples with 40 µL of CD2HOD was taken in a nanoprobe. The parameters used for measurement were relaxation delay of 1.5 s, pulse of 45.0 degrees, acquisition time of 3.5 s, and spectral width of 6000.6 Hz. For assigning peaks, after determination of pulse with 90°, simple RELAY COSY (correlation spectroscopy) was done and the chemical shifts of the terminal methyl and methylene of octadecyl groups were determined. Shimming was adjusted for each temperature using a standard semiautomatic method.

Chromatography. The chromatographic system consists of a Gulliver PU-1580 intelligent HPLC pump a Rheodyne sample injector having a 20 µL loop. The chromatograph included a JASCO 1580 pump and a JASCO MD-1510 UV-vis photodiode array detector. As the sensitivity of UV detector is high, 5 µL of sample was injected through a Reodyne Model 7125 injector. The column temperature was maintained using a column jacket with a circulator having a heating and cooling system. A personal computer connected to the detector with JASCO-Borwin (Ver 1.5) software was used for system control and data analysis. Methanol was used as a mobile phase for the separation of SRM 869b and β-carotene isomers at a temperature of 5 °C. Separations of SRM 1647e and other PAHs were carried out using a 90:10 (volume fraction) methanol/water mobile phase at 15 and 20 °C, respectively. The separation of tocopherol isomers was performed using 90:10 (volume fraction) methanol/water at 35 °C (at 40 °C for 3 µm particle size silica). All separations were at a flow rate of 1 mL/min. UV detection for the separation of SRM 869b and 16 PAHs (SRM 1647e) was at 254 nm and for β-carotene and tocopherol isomers was at 450 and 285 nm, respectively. The retention factor (k) measurement was done under isocratic elution conditions. The separation factor (R) is the ratio of the retention factor of 2 solutes that are being analyzed. The chromatography was done under isocratic elution conditions. The retention time of D2O was used as the void volume (t0) marker. (The absorption for D2O was measured at 400 nm.) All data points were derived from at least triplicate measurements; with the retention time (tR) value varying ±1%. The water/1-octanol partition coefficient (log P) was measured by the retention studies with octadecylsilylated silica, C18 (monomeric) (Inertsil, ODS, column size 250 × 4.6 mm i.d., G. L. Sciences, Tokyo, Japan): log P ) 3.579 + 4.207 log k (r ) 0.999997).41 Calculation. The dipole moments of aromatic protons of βand γ-tocopherols were calculated by Hyperchem Ver 5.1 with molecular mechanics (until the energy changes were below 0.001 kcal/mol) and following the semiempirical AM1 method.

Table 1. Elemental Analysis Data of Sil-MPS and Alternating Copolymer-Grafted Stationary Phases

Sil-MPS Sil-poly(ODA-alt-OMI) Sil-poly(ODA-alt-OMI)a Sil-poly(MA-alt-MMI) a

%C

%H

%N

C/N

2.28 15.68 21.64 15.01

1.02 3.52 3.55 2.35

0 0.45 0.62 1.82

35 35 8.2

Sil-poly(ODA-alt-OMI) phase prepared with 3 µm particle size silica.

RESULTS AND DISCUSSION We chose two important monomers (octadecyl acrylate and N-octadecylmaleimide) because not only can these monomers act as electron-donating and electron-accepting moieties, respectively, but also the resultant copolymer would display precise arrangement of carbonyl groups as interaction sources along the rigid polymer main chain. Two other similar monomers that do not possess long alkyl chains (methyl acrylate and N-methylmaleimide) were used as references. For the immobilization of the alternating copolymers onto silica, 3-mercaptopropyltrimethoxysilane (MPS) was initially grafted and, then, the copolymerization was carried out through a surfaceinitiated radical-chain transfer reaction (Figure 1). The copolymer-grafted silicas obtained from octadecyl acrylate and N-octadecylmaleimde (Sil-poly(ODA-alt-OMI)) and from methyl acrylate and N-methylmaleimide (Sil-poly(MA-alt-MMI)) were characterized by elemental analysis, DRIFT, solid-state 13C CP/MAS, and suspended-state 1H NMR spectroscopy. Carbon, hydrogen, and nitrogen elemental analyses of the bonded silica gels are summarized in Table 1. From the elemental analysis results, the surface coverage of the bonded phases were calculated

according to the literature method.42 The surface coverage for Sil-MPS, Sil-poly(ODA-alt-OMI), Sil-poly(ODA-alt-OMI)(3 µm particle size silica), and Sil-poly(MA-alt-MMI) were calculated to be 2.18, 2.66, 3.97, and 10.10 µmol m-2 respectively. The formation of alternating structures could be identified due to the fact that not only can N-alkylmaleimides be polymerized without ODA and MA but also the 1-to-1 monomer composition was confirmed by an elemental analysis and 1H NMR spectroscopy. The elemental analysis results indicate that the copolymer grafted onto silica surface and the type of the copolymer can be estimated from the C/N ratio, because one monomer (ODA) has no nitrogen atom. The initial composition of the monomers for each polymerization was 1-to-1 and C/N ) 36 and 7.7 for Sil-poly(ODA-alt-OMI) and Sil-poly(MA-alt-MMI), respectively. After copolymerization, almost the same C/N ratios were obtained, as shown in Table 1. It indicates that 1-to-1 monomer composition was obtained in both the cases. It is known that homopolymerization of the maleimide type monomer is very difficult, and therefore, 1-to-1 composition ratio of the monomers proved the alternating nature of the copolymers. For further confirmation, we prepared the alternating copolymer of octadecyl acrylate and N-octadecylmaleimide by one step telomerization with 3-mercaptopropyltrimethoxysilane. After purification of the telomere, we measured 1H NMR in CDCl3 to identify the copolymer and to confirm the 1-to-1 composition of the monomers. The characteristic peaks (Supporting Information, Figure S1) clearly indicate the formation of the copolymer. In addition, the peak area of 7 and 13 (Supporting Information, Figure S1) is almost the same, revealing 1-to-1 composition of the monomers. Presence of grafted copolymer on silica particles can also be confirmed by diffuse reflectance infrared Fourier transform (DRIFT) spectroscopy. Figure 2 shows the DRIFT spectra, for (a) bare silica, (b) Sil-poly(ODA-alt-OMI), and (c) Sil-poly(MAalt-MMI). DRIFT spectrum of alternating copolymer-grafted silica (Sil-poly(ODA-alt-OMI)) demonstrated characteristics signals at 2922, 2852, 1773, 1732, and 1700. The signals at 2922 and 2852 cm-1 arise from C-H stretches. The peaks near 1773 and 1700 cm-1 are attributed to the symmetrical CdO stretch and the asymmetrical CdO stretch of the imide linkage of the copolymer, respectively. The band at 1732 (partially overlapped) arises from ester carbonyl stretching for both of the alternating copolymers. For Sil-poly(MA-alt-MMI), the characteristic signals are at 1777, 1732, and 1700. The peaks near 1777 and 1700 cm-1 are attributed to the symmetrical CdO stretch and the asymmetrical CdO stretch of the imide linkage of the polymer, respectively. The signals at 2922 and 2852 cm-1 are absent in

(41) Goto, Y.; Nakashima, K.; Mitsuishi, K.; Takafuji, M.; Sakaki, S.; Ihara, H. Chromatographia 2002, 56, 19–23.

(42) Qiu, H.; Jiang, Q.; Wei, Z.; Wang, X.; Liu, X.; Jiang, S. J. Chromatogr., A 2007, 1163, 63–69.


3323

Figure 2. DRIFT spectra for (a) bare silica particles, (b) Silpoly(ODA-alt-OMI), and (c) Sil-poly(MA-alt-MMI).

Sil-poly(MA-alt-MMI), indicating no long alkyl chain present in the copolymer. The above results indicate both of the copolymers grafted onto silica, which are in good agreement with the elemental analyses data. Solid-state 13C CP/MAS NMR gives useful information about the chemical composition of the polymer grafted silica. In addition, it offers evidence about the conformation and dynamics of immobilized alkyl chains.43 Under the condition of magicangle spinning and dipolar coupling of protons, the chemical shift of methylene groups in 13C CP/MAS NMR spectroscopy depends largely on the conformation of alkyl chains -(CH2)n-. It is reported that the 13C signals for alkyl chains is observed at two resonances; one is at 32.6 ppm attributed to trans conformation, indicating crystalline and rigid state, and the other is at 30.0 ppm corresponding to gauche conformation, indicating disordered and mobile state.17 The NMR signal for the main methylene group occurs at 32.54 ppm corresponding to trans and at 30.29 ppm attributed to gauche conformation for Sil-poly(ODA-alt-OMI). Solid-state 13C CP/MAS NMR spectroscopy reveals that in Sil-poly(ODA-alt-OMI) the conformation of alkyl chains -(CH2)n- can mainly be attributed to gauche with a low content of trans conformation,43 while a slight temperature-dependent trans-gauche transition was observed (Figure 3). Previously, we have reported that the alkyl chains in polymeric C18 are almost all trans (ordered) and in monomeric C18 are almost all gauche (disordered) conformations.43 Suspended-state 1H NMR can be used to determine the mobility of the alkyl chains on silica surface of a stationary phase.43 The intensity of the molecular mobility of the -(CH2)n- methylene groups (in methanol) was increased little with increasing temperature but not dramatically, indicating alkyl chains remain in a liquid state due to solvation at lower and higher temperature, and the results agree with the results of solid-state 13C CP/MAS NMR spectroscopy. However, the mobility of the alkyl chains gradually increased with temperature (Supporting Information, Figure S2). (43) Ansarian, H. R.; Derakhsan, M.; Rahman, M. M.; Sakurai, T.; Takafuji, M.; Taniguchi, I.; Ihara, H. Anal. Chim. Acta 2005, 547, 179–187.

3324


Figure 3. Partial solid-state 13C CP/MAS NMR spectra of Silpoly(ODA-alt-OMI) at variable temperatures.

The molecular recognition ability of Sil-poly(ODA-alt-OMI) was evaluated by liquid chromatographic retention with the packed column. The selectivity for two-dimensional shape was studied with a molecular shape descriptor, such as molecular length and length-to-breadth (L/B) ratio. The first examination was carried out with four nonlinear and linear PAHs. Sil-poly(ODA-alt-OMI) showed unusually high selectivity for benzo[a]anthracene (1) and naphthacene (2), which have the same number of carbon atoms and π-electrons but differ only in their molecular shape such as the length and aspect ratio (L ) 11.1 and 12.1 Å, respectively). The separation factor (R2/1) reaches 2.28 with Sil-poly(ODA-altOMI). On the contrary, C18 phase, which are the most widely used adsorbents for HPLC showed only R2/1 ) 1.16 and 1.38 in their monomeric and polymeric C18 types (C18 (mono), C18 (poly)), respectively, as shown in Figure 4a. Similar results were observed for five ring PAHs: e.g., dibenz[a,c]anthracene (3, L ) 9.8 Å) and dibenz[a,h]anthracene (4, L ) 12.1 Å) (Figure 4b). We also investigated the selectivity for o-terphenyl (5) and triphenylene (6), which have the same number of carbon atoms and π-electrons and a similar molecular length (L ) 10.5 Å and 11.5 Å, respectively) but differ in their molecular planarity. As a result, Sil-poly(ODA-alt-OMI) showed much better selectivity (R6/5 ) 3.15) than monomeric C18 (R6/5 ) 1.54) and polymeric C18 (R6/5 ) 2.28). These results indicate that Sil-poly(ODA-alt-OMI) can recognize not only molecular length (aspect ratio) but also molecular planarity better than conventional C18 phases. The enhanced molecular shape selectivity of Sil-poly(ODA-altOMI) enabled us to separate SRM 869b as a Column Selectivity Test Mixture for Liquid Chromatography.44 This material consists of three PAHs with planar and nonplanar shapes. The test mixture was originally developed to facilitate classification of (the most available octadecylsilylated silica) columns in terms of stationary phase bonding chemistry (monmeric vs polymeric) and to evaluate separation selectivity of shape-constrained solutes. Generally, late elution of benzo[a]pyrene (BaP) relative to tetrabenzonaphthalene, (TBN) indicates enhanced column selectivity toward geometrical isomers. The selectivity coefficient RTBN/BaP can be used as a (44) Sander, L. C.; Wise, S. A. SRM 869b, Column Selectivity Test Mixture for Liquid Chromatography (Polycyclic Aromatic Hydrocarbons). Certificate of Analysis; NIST: Gaithersburg, MD, 2008.

Figure 4. Chromatograms for the mixtures of (a) benzo[a]anthracene and naphthacene and (b) dibenz[a,c]anthracene and dibenz[a,h]anthracene with Sil-poly(ODA-alt-OMI), monomeric, and polymeric C18 phases.

Figure 5. Separation of SRM 869b with Sil-poly(ODA-alt-OMI).

measure of this property. Values of RTBN/BaP < 1.0 are indicative of “polymeric-like” retention behavior with enhanced shape recognition abilities, and values of RTBN/BaP > 1.7 are indicative of “monomeric-like” retention behavior with reduced molecularshape selectivity. Figure 5 shows the chromatogram of SRM 869b test mixture on Sil-poly(ODA-alt-OMI), indicating polymer-like retention behavior with enhanced shape selectivity (RTBN/BaP ) 0.5). Similarly, better separation with Sil-poly(ODA-alt-OMI) was also realized for the other important standard mixtures: e.g., SRM 1647e (16 PAHs are listed as priority pollutants by the EPA) were completely separated into 16 peaks (Figure 6) within 50 min. Figure 6 shows typical chromatograms with distinct separations for 3 and 4 with R ) 1.15, 9, and 10 with R ) 1.11, and 15 and 16 with R ) 1.10 on Sil-poly(ODA-alt-OMI) although almost no separation was observed in one of the most useful columns, monomeric C18 or polymeric C18. In addition, the advantage of Sil-poly(ODA-alt-OMI) was emphasized by the fact that the separation of SRM 1647e was realized in an isocratic mode, which represented one of the most important challenges; it is generally analyzed in a gradient elution mode35 because an isocratic elution is much more attractive due to its better

interlaboratory reproducibility, increased column life, and the less-complex apparatus it requires, among other reasons. Better overall resolution and peak shape were achieved with Silpoly(ODA-alt-OMI) using 3 µm particle size silica compared to 5 µm (data not shown). On the basis of the above-mentioned critical advantages in terms of molecular-shape selectivity, we attempted the challenging separation of some biologically important compounds such as carotenoids and vitamin E (tocopherol) that are difficult to separate with conventional C18 phases. It is known that carotenoids are highly conjugated with 11 double bonds, so that trans-cis isomerization yields many geometrical isomers that have different biological activities. The difficulty in the separation of these isomers is that they differ only in the characteristics of their molecular-shape, such as linearity or some bending. In this study, isomers of β-carotene were obtained by photoirradiation based on a method in the literature.39 Figure 7 shows the chromatograms of all-E β-carotene with Sil-poly(ODA-alt-OMI) before and after photoirradiation. They clearly show that before irradiation almost one peak of all-E, observed with some impurities, was recorded; however, after irradiation, several isomers were separated completely. The peak of the most linear isomer (all-E) was identified by comparing the retention time of injected samples before and after isomerization. The other peaks were estimated according to the molecular linearity of the isomers. On the other hand, almost no separation was observed with monomeric and polymeric C18 phases (Supporting Information, Figure S3). This separation of β-carotene isomers was only realized due to the very high molecular linearity selectivity of Sil-poly(ODA-alt-OMI), a result which agreed with the fundamental results shown in Figure 4. Among the R-, β-, γ-, and δ-isomers of tocopherol (Vitamin E), the separation of β- and γ-tocopherol has long presented a special challenge. The baseline separation of these two homologues especially for β- and γ-isomers could not be achieved in conventional RP-HPLC and is completely impossible on C18 phases.15,16,45 (45) Abidi, S. L. J. Chromatogr., A 2000, 881, 197–216.


3325

Figure 6. Separation of the 16 priority pollutant PAHs (SRM 1647e) on Sil-poly(ODA-alt-OMI) and monomeric and polymeric C18 phases. Key: 1, naphthalene; 2, acenaphthylene; 3, acenaphthene; 4, fiuorene; 5, phenanthrene; 6, anthracene; 7, fluoranthene; 8, pyrene; 9, benzo[a]anthracene; 10, chrysene; 11, benzo[b]fluoranthene; 12, benzo[k]fluoranthene; 13, benzo[a]pyrene; 14, dibenz[a,h]anthracene; 15, bezo[ghi]perylene; 16, indeno[1,2,3-cd]pyrene.

Figure 8b shows the complete baseline separation of β- and γ-isomers with Sil-poly(ODA-alt-OMI) using 3 µm particle size silica. To the best of our knowledge, this is the first instance of the complete baseline separation of β- and γ-tocopherols with very good resolution (2.2). Herein, we must discuss why Sil-poly(ODA-alt-OMI) shows unique and excellent selectivity toward PAHs, tocopherols, and carotenoids than conventional octadecylsilylated silicas such as C18 (mono) and C18 (poly). In general, the molecular shape selectivity in the C18 phase increases with increasing carbon loading.46,47 This phenomenon has been attributed to slight increase of alkyl chain ordering but not for direct interaction with guest molecules.48 Sil-poly(ODA-alt-OMI), however, showed better selectivity, regardless of the fact that not only does it have lower carbon loading than the C18 (poly) phase (%C 17.5; Table 1) but also the alkyl chains of Sil-poly(ODA-alt-OMI) are not ordered completely and are rather flexible, as indicated by the NMR results. In order to thoroughly understand the difference between Sil-poly(ODA-alt-OMI) and simply hydrophobized silica, we propose the direct interaction with carbonyl groups from the poly(ODAalt-OMI), as schematically illustrated in Figure 9a. There are some experimental and theoretical supports for this proposal: (1) unlike C18 phases, the higher selectivity of Sil-poly(ODA-alt-OMI) cannot be explained by the hydrophobic effect with solute molecules. It is known that conventional alkyl phases (C18) can recognize the hydrophobicity of solutes in HPLC, and this hydrophobicity is measured by the methylene group selectivity of the stationary phases. This reflects the possibility of the phase being able to separate two molecules Figure 7. Separation of β-carotene isomers using Sil-poly(ODA-altOMI).

Figure 8a shows the separation ability of Sil-poly(ODA-alt-OMI) compared with that obtained on C18 (mono) and C18 (poly). 3326


(46) Wise, S. A.; Sander, L. C. In Chromatographic Separation Based on Molecular Recognition; Jinno, K., Ed.; Wiley-VCH, Inc.: New York, 1996; pp 1-64. (47) Mallik, A. K.; Rahman, M. M.; Czaun, M.; Takafuji, M.; Ihara, H. J. Chromatogr., A 2008, 1187, 119–127. (48) Rahman, M. M.; Takafuji, M.; Ansarian, H. R.; Ihara, H. Anal. Chem. 2005, 77, 6671–6681.

Figure 9. (a) Schematic illustration to explain the molecular linearity recognition of Sil-poly(ODA-alt-OMI) through carbonyl-π interaction. A multiple interaction effect will be expected for a linear and planar substance such as naphthacene than for benzo[a]anthracene. (b) Agreement of higher selectivity of linear isomers of PAHs and β-carotene than of nonlinear ones with Sil-poly(ODA-alt-OMI).

that differ only in methylene groups, e.g., amylbenzene and butylbenzene or ethylbenzene or toluene. The retention mode as well as the extent of hydrophobic interaction among the solutes and the packing materials in HPLC can be determined by retention studies of alkylbenzenes as solutes.49 The correlation between log k (capacity factor) and log P (water/1octanol partition coefficient) for Sil-poly(ODA-alt-OMI) and polymeric C18 phase showed the retention mode of Silpoly(ODA-alt-OMI), a reversed-phase mode to that of C18 phase (Supporting Information, Figure S4). The relationship between log k and log P is used to determine the hydrophobicity recognition ability or retention mode of a stationary phase. It was observed that log k and log P plots of alkylbenzenes and PAHs in polymeric C18 were parallel and almost coincided with each other, providing evidence that C18 phase can recognize only the hydrophobicity of analytes. On the other hand, it has been found that Sil-poly(ODA-alt-OMI) showed higher retention for PAHs compared to its values for alkylbenzenes (Supporting Information, Figure S4). For instance, the log P of naphthacene (5.71) is smaller than that of octylbenzene (6.29), while log k value

of naphthacene (0.92) is higher than that of octylbenzene (0.25). The increase of log k for PAHs was accompanied by selectivity enhancement which provides specific interactive sites for PAHs that can recognize aromaticity besides molecular hydrophobicity. Therefore, other interactions are certainly involved in the unique selectivity of Sil-poly(ODA-alt-OMI) toward PAHs. A carbonyl-π interaction has been discussed in our previous calculation works.50 This interaction in a model complex of HCHO-benzene is much stronger (1.87 kcal mol-1) than a CH4-benzene interaction (0.53 kcal mol-1) and a plane-to-plane interaction between two benzenes (0.49 kcal mol-1).50 On the basis of this calculation, when acetone with a carbonyl group was added to the mobile phase, both the retention time and selectivity decreased remarkably. This indicates that acetone functions as an inhibitor for a carbonyl-π interaction.48 (2) Selectivity enhancement through a carbonyl-π interaction has been also discussed in relation to homopolymers from octadecyl acrylate (ODAn). When ODAn was grafted onto silica and then evaluated by the retention time of PAHs, the resultant selectivity was higher, especially at crystalline temperatures of ODAn than for C18 columns.23-25 This is attributed to the fact that ODAn has a crystalline-to-isotropic phase transition, so that a multiple carbonyl-π interaction with PAHs becomes possible through the ordering of carbonyl groups at lower temperatures, where the polymers are in a crystalline state. In contrast, poly(ODA-alt-OMI) has no such phase transition as well as a

(49) Claessens, H. A.; Van Straten, M. A.; Cramers, C. A.; Jezierska, M.; Buszewski, B. J. Chromatogr., A 1998, 826, 135–157.

(50) Sakaki, S.; Kato, K.; Miyazaki, T.; Ohkubo, K.; Ihara, H.; Hirayama, C. J. Chem. Soc. Faraday Trans. 1993, 89, 659–664.

Figure 8. Separation of tocopherol isomers on (a) Sil-poly(ODAalt-OMI) and monomeric and polymeric C18 phases. (b) Sil-poly(ODAalt-OMI) using 3 µm particle size silica.


3327

slight ordering of the alkyl chain (low content of trans conformation), but the alternating copolymerization containing a rigid maleimide ring can definitely provide a rigid polymer main chain that leads to the linear ordering of carbonyl groups derived from both maleimide and acrylate. Additionally, a maleimide ring concentrates the carbonyl groups, and an alternating nature can integrate the carbonyl groups along the polymer main chain, leading to an increase in the interaction area for linear and planar PAHs with a high molecular aspect ratio. As a result, multiple carbonyl-π interaction is more advantageous to linear planar molecules than to nonlinear planar molecules (Figure 9a) and, therefore, to enhance molecular linearity selectivity for PAHs as well as β-carotene isomers (Figure 9b). (3) An alternating copolymer with methyl acrylate and Nmethylmaleimide (poly(MA-alt-MMI) without long alkyl chains as a reference of poly(ODA-alt-OMI) did not show any clear selectivity toward tocopherol or carotenoids and showed less molecular linearity and planarity selectivity toward PAHs than Silpoly(ODA-alt-OMI). This indicates the importance of long alkyl chains in Sil-poly(ODA-alt-OMI), even though they are not completely ordered. It is certain, however, that the role of the long alkyl chains is not involved in the hydrophobic effect, as mentioned above. One possibility is that slightly ordered long alkyl chains may assist in the orientation of the carbonyl groups along the polymer main chain and create a microenvironment which favors multiple π-π interaction with the bulk guest molecules. (4) In regard to the separation of β- and γ-isomers of tocopherol, a slight difference in their dipoles is very important for poly(ODA-alt-OMI). The calculated dipole moments of β- and γ-tocopherol are 2.71 and 2.24, respectively. It is reasonable to assume that poly(ODA-alt-OMI) as a carbonyl group-rich organic phase can recognize this difference and that the polymeric

3328


structure can then promote multiple carbonyl-π (dipole-dipole) interactions. On the other hand, no such effect can be expected in simply hydrophobized silicas. CONCLUSION We have described the separation of some biologically important shape-constrained isomers on Sil-poly(ODA-alt-OMI) in an RP-HPLC mode and compared them with commercially available octadecylsilylated silicas such as monomeric and polymeric C18 phases. The result presented above documents the first synthesis of an alternating copolymer with the monomers ODA and OMI and the first application of an alternating copolymer in the preparation of a liquid chromatographic adsorbent. We have found that Sil-poly(ODA-alt-OMI) is clearly one of the most promising adsorbents that displays a high degree of molecular recognition ability. Application of alternating copolymers in the science of separation will open the way to the solution of many of its challenges, which can be accelerated by research into the system of cooperation between monomers. ACKNOWLEDGMENT This work was partially supported by a Grant-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology, Japan. We thank Dr. I. Nobuhara of Fuji Silysia Chemical LTD (Aichi, Japan) for providing 3 µm particle size silica. SUPPORTING INFORMATION AVAILABLE Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org. Received for review January 15, 2010. Accepted March 10, 2010. AC1001178

Novel Approach for the Separation of Shape-Constrained Isomers with

Recommend Documents