Synthesis of regioselectively substituted cellulose derivatives and applications in chiral chromatography

Synthesis of Regioselectively Substituted CelluloseDerivatives and Applications in

Chiral ChromatographyMURAT ACEMOGLU,* ERNST KUSTERS, JURGEN BAUMANN, IVAN HERNANDEZ, AND

CHING PONG MAKChemical Process Research and Development, Novartis Pharma AG, Basel, Switzerland

ABSTRACT Various cellulose-2,3-bis-arylcarbamate-6-O-arylesters and cellulose-2,3-bis-arylester-6-O-arylcarbamates, designed to test the possible combined effects of theknown tris-arylcarbamate and tris-arylester classes, were synthesized with high regios-electivity at O-C(6), and their use as CSPs in liquid chromatography for enantiomericseparations was investigated. The separations obtained with the synthesized CSPs werecompared to the separations achieved on a self-packed reference column, consisting ofcellulose-tris-(3,5-dimethylphenyl-carbamate) as CSP standard. Among the synthesized,regioselectively substituted cellulose derivatives, 2,3-bis-O-(3,5-dimethylphenyl-carbamate)-6-O-benzoate-cellulose and 2,3-bis-O-(benzoate)-6-O-(3,5-dichlorophenyl-carbamate)-cellulose gave the best CSPs for the separation of the test racemates. CSPsfrom regioselectively substituted cellulose derivatives seem to exhibit higher selectivi-ties than cellulose-tris-(3,5-dimethylphenylcarbamate) for certain classes of racemiccompounds. Chirality 10:294–306, 1998. © 1998 Wiley-Liss, Inc.

KEY WORDS: cellulose; regioselective derivatization; chiral stationary phases; liquidchromatography; enantioseparation

Homogenously substituted cellulose derivatives arecommon materials and have been used as chiral stationaryphases (CSPs) now for decades. Cellulose itself is an in-expensive raw material and relatively easy to derivatize. Bychoosing the proper derivatives, quite different propertiesregarding crystallinity and solubility can be achieved. Themost frequently used cellulose derivatives for enantiomericseparations were introduced in 1984 by Ichida et al.1 Theyare now well known as the commercially available Chiral-celt columns.

Among the cellulose derivatives reported so far as CSPsfor the chromatographic resolution of racemates, cellulose-tris-benzoates2–4 and cellulose-tris-carbamates5,6 are of par-ticular interest. Besides their excellent capabilities for theresolution of racemates, they exhibit different but comple-mentary selectivities. Out of the series of carbamates, cel-lulose-tris(3,5-dimethylphenylcarbamate) exhibits thehighest and broadest resolving power. Several designshave meanwhile been studied to further improve thesetype of CSPs and to widen their applicability: The effect ofparticle size and pore size of the aminopropylated silicasupport was investigated,7,8 as well as the mixing of twodifferent cellulose derivatives, before or after being coatedon macroporous silica gel.9 In the latter case, the physicalmixing of two ‘‘good’’ selectors was intended to enhancethe scope and selectivity of the CSP. The main character-istic of all approaches cited above was the synthesis andutilization of tris-homosubstituted cellulose derivatives.CSPs based on regioselective heterosubstitution of cellu-

lose and amylose were first introduced by Kaida et al.,10

who prepared cellulose and amylose derivatives bearingdifferent carbamoyl substituents at O-C(2,3) and O-C(6),respectively. Francotte and Zhang11,12 utilised the samestrategy for the preparation of cellulose derivatives havingdifferent ester substituents at O-C(2,3) and O-C(6) respec-tively. However, it would be interesting to test the effect ofselective combinations of carbamate and benzoate substitu-ents on cellulose, e.g., to synthesize cellulose derivativessubstituted as carbamates at positions 2 and 3 and as ben-zoate at position 6, or vice versa (regioselective heterofunc-tional substitution). This might lead to more generally use-ful CSPs and also contribute to a better understanding ofthe mechanism of chromatographic resolution of race-mates. CSPs based on regioselective heterofunctional sub-stitution of celluloses have very recently been reported byChassaing et al.13 in moderate regioselectivities, and nocomplete selectivity at O-C(6). Heterogenous substitutionat O-C(6) is known to influence the resolution capability ofthe products in the cellulose series.10 In this paper, resultsof our research program on these same objectives will bedescribed; this work was carried out independently, for thefirst time yielding products with up to complete regioselec-

Abbreviations used: NMP, 1-methyl-2-pyrrolidone; Ph, phenyl; Py, pyri-dine; RT, room temperature; THF, tetrahydrofuran.*Correspondence to: Dr. M. Acemoglu, Chemical Process Research andDevelopment, Novartis Pharma AG, CH-4002 Basel, Switzerland.Received for publication 2 April 1997; Accepted 7 July 1997

CHIRALITY 10:294–306 (1998)

© 1998 Wiley-Liss, Inc.

tivity at O-C(6). The synthesized polymers were screenedas CSPs on silica gel and their chiral discrimination prop-erties will be reported.

EXPERIMENTAL PROCEDURESMaterials and Methods for Synthesis

1H NMR spectra were recorded on a Brucker (Karls-ruhe, Germany) AM 360 spectrometer using tetramethyl-silane as internal standard. UV/VIS spectra were measuredon a Perkin Elmer (Oak Brook, IL) Lambda 19 spectropho-tometer. IR spectra were measured on a Brucker IFF 66spectrometer. Dynamic viscosities (hdyn) were measuredat 20°C in THF (1% solution) on an AVS 350 instrument(Hosheim, Germany) with a Ubbelohde capillary. Solventsfor synthesis were purchased from Fluka (Buchs, Switzer-land) or Merck (Darmstadt, Germany) and were dried onmolecular sieves. Avicel (cellulose-powder) and 3-aminopropyltriethoxysilane were purchased from Merck.Benzoic acid, 4-methylbenzoic acid, dicyclohexylcarbodi-imide (DCC), pyridine, 4-dimethylaminopyridine (DMAP),R(+)- and S(−)-phenylethylisocyanate and trityl chloridewere purchased from Fluka. 3,5-dichlorophenyl-isocyanateand 3,5-dimethylphenyl-isocyanate were purchased fromLancaster (Strasbourg, France).

Materials and Methods for Liquid ChromatographyA Kontron (Zürich, Switzerland) HPLC pump (Model

420) was used in conjunction with a Kontron variable-wavelength UV detector (Model 430). Chiralcel-ODt (250× 4.6 mm) was purchased from Daicel (Tokyo, Japan). Forstandard packing procedures, the stationary phases wereslurry packed using a suspension in hexane/2-propanol(9:1) for 120 min at a back pressure of 300 bar. The com-position of mobile phases and flow rates are given in thecorresponding tables.

Racemates rac-1 (pindolol), rac-2-4, rac-10-13, rac-14(quinagolide) and rad-15 are Novartis (Basel, Switzerland)proprietary substances. Racemate rac-9 (bupivacaine) wasa kind gift from Astra Chemicals (Deggendorf, Switzer-land). Racemates rac-7 and rac-8 were purchased fromFluka, rac-5 was purchased from EGA (Steinheim, Ger-many), and rac-6 from Aldrich (Milwaukee, WI). 1,3,5-Tri-tert.-butylbenzene was purchased from Fluka. The solventswere of Lichrosolvt grade from Merck. Nucleosil 1000-10from Macherey-Nagel (Duren, Germany) was used for thepreparation of the CSPs.

Dead times (t0) were estimated by injection of 1,3,5-tri-tert.-butylbenzene. Capacity factors (k8) were calculatedaccording to the equation k8 = ((tr − t0)/t0), enantioselec-tivities (a) according to a = k28/k18, and the resolutionfactors (Rs) according to Rs = 1.18(t2 − t1)/(W1(0.5) +W2(0.5)), where t1 and t2 refer to the retention time of thefirst and second eluted enantiomer and W1(0.5) and W2(0.5)represent the peak width of the corresponding peak at 50%of its height. The theoretical plate number N has beencalculated according to N = 5.54 (tr/Wr(0.5))2 after injectionof the individual enantiomers.

CarbamoylationsCarbamoylations were carried out by treatment of the

cellulose derivative with an excess of the isocyanate in

pyridine at 100°C for 17–24 h. Methanol was added to thesolution at 55–60°C to precipitate the crude products.

De-TritylationsIn a typical procedure, compound 2 (44.27 g, 63 mmol)

was dissolved in CH2Cl2 (1.7 liter) and the solution wascooled in an ice bath to 0–5°C. A slight stream of HCl gaswas bubbled into the solution for 5 min at this temperatureand, subsequently, argon gas was passed through the so-lution for 10 min, while always keeping the temperature at0–5°C. Na2CO3 (88.54 g, 835 mmol) and CH2Cl2 (1.8 liter)were added sequentially and the suspension was stirred for30 min at RT. The solvent was evaporated under reducedpressure, the residual was taken up in THF (3.5 liter), andthe formed suspension was filtrated. After concentration ofthe filtrate to a final volume of approximately 200 ml,methanol (3 liter) was slowly added and the precipitate(product) was isolated by filtration. The precipitate waswashed thoroughly with water (2.5 liter) to yield the crudeproduct, which was reprecipitated where appropriate. Inthe case of the more sensitive products 14 and 15, thework-up was slightly modified: After HCl and argon gaswere passed through the solution, it was poured onto sat.NaHCO3, was stirred at RT for 15 min and the organicsolvent was evaporated. The formed precipitate was iso-lated by filtration, dissolved in THF, and reprecipitatedfrom methanol to give the crude product.

Esterifications of 2,3-Bis-O-Arylcarbamate-CellulosesThe 2,3-bis-O-arylcarbamate-celluloses were esterified at

C-(6)-O by treatment with excess acyl chloride in pyridineat 100°C for 17 h. The crude products were precipitated byaddition of methanol to the solution at 55–60°C.

Esterifications of 6-O-Trityl-Cellulose (1)

6-O-trityl-cellulose (1) (15.0 g, 37.1 mmol) was sus-pended in CH2Cl2 (1 liter). The corresponding aromaticacid (741.6 mmol), DCC (153.01 g, 741.6 mmol), andDMAP (9.06 g, 74.2 mmol) were added and the suspensionwas heated at 40°C for 68 h. The reaction mixture wasallowed to cool to RT and was subsequently cooled in anice bath for 1 h. The precipitate was filtered off and washedtwice with CH2Cl2. The combined filtrates were evaporatedto a final volume of 500 ml and the solution was pouredslowly into ethanol (5 liter). The crude precipitate was fil-tered, washed twice with ethanol, dissolved in THF (1 li-ter), and reprecipitated from ethanol (5 liter) to give thepure products in 90–95% yields.

RESULTS AND DISCUSSIONSynthesis of Regioselectively Substituted

Cellulose DerivativesFor the synthesis of regioselectively heterosubstituted

cellulose derivatives, several factors were expected to becrucial: (1) The regioselectivity and completeness of theprotection step of cellulose at position O-C(6), (2) completesubsequent derivatization of the free hydroxyl groups atC(2) and C(3), and (3) mild cleavage of the protectinggroup at O-C(6). Complete tritylation of the hydroxyl groupat C(6) of cellulose was achieved in very good regioselec-

REGIOSELECTIVELY SUBSTITUTED CELLULOSE DERIVATIVES 295

Scheme 1. Synthesis of cellulose-2,3-bis-O-arylcarbamate-6-O-ester derivatives. (a) Aryl isocyanate,Py, 100°C, 20–22 h, 90%; (b) CH2Cl2, HCl(gas), 0°C, 5min., 90–94%; (c) Acid chloride, Py, 100°C, 17 h, 70–90%;(d) Aryl isocyanate, Py, 100°C, 17h, 80–82%.

TABLE 1. Chemical and analytical data for the cellulose-2,3-bis-O-arylcarbamate series (2–13)

Compound R1 R2

Composition(molmass)

Elemental analysiscalcd. (found) (%)

C H N Cl

2 3,5-Dimethylphenylcarbamoyl- n.a. (C43H42N2O7)n 73.91 6.06 4.01 —(698.87)n (73.94) (6.08) (4.06) —

3 3,5-Dichlorophenylcarbamoyl- n.a. (C39H30N2O7Cl4)n 60.02 3.87 3.59 18.17(780.51)n (59.92) (3.97) (3.76) (17.90)

4 3,5-Dimethylphenylcarbamoyl-/ n.a. (C23.1H28.1N1.9O6.9) 62.66 6.40 6.01 —H- (95:5) (442.79)n (62.04) (6.15) (6.14) —

5 3,5-Dichlorophenylcarbamoyl-/ n.a. (C19.3H15.7O6.9N1.9Cl3.8)n 44.63 3.05 5.12 25.94H- (95:5) (519.37)n (45.32) (3.30) (5.81) (25.45)

6 3,5-Dimethylphenylcarbamoyl-/ 4-Methylbenzoyl- (C31.9H33.7N1.9O8)n 67.02 5.94 4.65 —4-Methylbenzoyl- (95:5) (571.73)n (66.62) (5.89) (4.66) —

7 3,5-Dimethylphenylcarbamoyl-/ Benzoyl- (C30.8H31.5N1.9O8)n 66.50 5.71 4.78 —Benzoyl- (95:5) (556.30)n (66.49) (5.77) (4.95) —

8 3,5-Dimethylphenylcarbamoyl-/ Camphanoyl- (C33.0H38.1N1.9O10.2)n63.46 6.15 4.26 —

Camphanoyl- (95:5) (624.58)n (63.74) (6.41) (3.97) —9 3,5-Dichlorophenylcarbamoyl-/ 4-Methylbenzoyl- (C28.1H22.3N1.9O8Cl3.8)n 51.98 3.46 4.10 20.75

4-Methylbenzoyl- (95:5) (649.32)n (55.01) (3.70) (3.48) (18.70)10 3,5-Dimethylphenylcarbamoyl-/ 3,5-Dichlorophenyl- (C30.8H30.4N3.0O8.0Cl2.2)n 57.04 4.72 6.48 12.02

3,5-Dichlorophenylcarbamoyl- carbamoyl (648.60)n (55.80) (4.57) (6.62) (13.90)(95:5)

11 3,5-Dimethylphenylcarbamoyl-/ (R)-(1)-Phenyl- (C33H37N3O8)n 65.66 6.18 6.96 —(R)-(1)-Phenylethylcarbamoyl- ethylcarbamoyl- (603.68)n (65.45) (6.19) (6.91) —(95:5)

12 3,5-Dimethylphenylcarbamoyl-/ (S)-(1)-Phenyl- (C33H37N3O8)n 65.66 6.18 6.96 —(S)-(1)-Phenylethylcarbamoyl- ethylcarbamoyl- (603.68)n (65.35) (6.21) (6.95) —(95:5)

13 3,5-Dichlorophenylcarbamoyl-/ 3,5-Dimethylphenyl- (C29.2H25.6N3O8Cl3.8)n 51.48 3.79 6.17 19.783,5-Dimethylphenylcarbamoyl- carbamoyl- (681.27)n (52.32) (4.02) (6.28) (18.85)(95:5)

296 ACEMOGLU ET AL.

tivity on treatment of cellulose with excess trityl chloride ina mixture of N-ethylpyridinium chloride/pyridine (2:1) assolvent at 80°C, according to the method of Pfannemullerand Berg.14 The degree of substitution (d.s., ratio substitu-ent/monomer unit for a given substituent) of 6-O-trityl-cellulose (1) was determined by UV/VIS15 and by 1H NMRspectroscopy.16 Both methods gave comparable results in-

dicating d.s. = about 1.0 for 6-O-trityl-cellulose (1), whichwas completely carbamoylated at positions O-C(2) and O-C(3) on subsequent treatment with the corresponding ar-ylisocyanates in pyridine at 100°C, to give the compounds2 and 3 (Scheme 1). Complete carbamoylation reactionwas demonstrated by the disapperance of hydroxy absorb-tion in the IR spectra of the products 2 and 3, and was

Scheme 2. Synthesis of cellulose-2,3-bis-O-arylester-6-O-arylcarbamate derivatives. (a) Ar-CO2H, DCCl,DMAP, CH2Cl2, reflux, 68 h, 90–95%; (b) HCl(gas),CH2Cl2, 0°C, 5 min. 87–96%; (c) Aryl isocyanate, Py,100°C, 24 h, 93–98%.

TABLE 2. Chemical and analytical data for the cellulose-2,3-bis-O-arylester series (14–23)

Compound R1 R2

Composition(molmass)

Elemental analysiscalcd. (found) (%)

C H N Cl

14 4-Methylbenzoyl-/ H- (90:10) n.a. (C39.4H34.8O6.8)n 76.69 5.68 — —(617.11)n (76.31) (5.85) — —

15 Benzoyl- n.a. (C39H32O7)n 76.46 5.26 — —(612.69)n (75.83) (5.33) — —

16 4-Methylbenzoyl-/ H- (85:15) n.a. (C19.6H20.2O6.7)n 64.86 5.61 — —(362.98)n (63.63) (5.74) — —

17 Benzoyl-/ H- (95:5) n.a. (C19.3H18.6O6.9)n 64.22 5.19 — —(360.96)n (62.89) (4.94) — —

18 4-Methylbenzoyl-/ 3,5-Dimethyl- 3,5-Dimethylphenyl- (C31.3H31.9N1.3O8)n 67.82 5.80 3.28 —phenylcarbamoyl- (85:15) carbamoyl- (554.31)n (67.21) (5.84) (3.28) —

19 4-Methylbenzoyl-/ 3,5-Dichloro- 3,5-Dichlorophenyl- (C28.7H24.1N1.3O8Cl2.6)n 56.75 4.00 3.00 15.18phenylcarbamoyl- (85:15) carbamoyl- (607.39)n (56.00) (4.03) (3.14) (15.70)

20 4-Methylbenzoyl-/ (R)-(1)-Phenyl- (R)-(1)-Phenylethyl- (C31.3H31.9N1.3O8)n 67.82 5.80 3.28 —ethylcarbamoyl- (85:15) carbamoyl- (554.31)n (67.15) (6.01) (3.16) —

21 4-Methylbenzoyl-/ (S)-(1)-Phenyl-/ (S)-(1)-Phenylethyl- (C31.3H31.9N1.3O8)n 67.82 5.80 3.28 —ethylcarbamoyl- (85:15) carbamoyl- (554.31)n (67.14) (5.86) (3.14) —

22 Benzoyl-/ 3,5-Dimethylphenyl- 3,5-Dimethylphenyl- (C29.2H27.5N1.1O8.0)n 67.21 5.31 2.95 —carbamoyl- (95:5) carbamoyl (521.85)n (66.48) (5.40) (3.08) —

23 Benzoyl-/ 3,5-Dichlorophenyl- 3,5-Dichlorophenyl- (C27.0H20.9N1.1O8Cl2.2)n 57.22 3.72 2.72 13.76carbamoyl- (95:5) carbamoyl- (566.77)n (56.57) (3.99) (2.90) (13.90)


confirmed by elemental analysis, which was in agreementfor a substitution pattern of carbamate:trityl (2:1) (Table 1).In the following step, the trityl protecting group wascleaved under water-free conditions with HCl gas inCH2Cl2 in good yield. Complete removal of the tritylgroups was confirmed by an almost colourless solution of

the products in 60% HClO4 and by UV spectra.15 as well asby 1H NMR: The signals of H2C(6) were shifted to approxi-mately 3.35 and 3.55 ppm, respectively, and the lack of anysignals in the range of 3.9–4.5 ppm clearly showed theabsence of any substitution at O-C(6). Some minor decar-bamoylation must have occurred during this step, since the

TABLE 3. 1H NMR data for the cellulose-2,3-bis-O-arylcarbamate series in d6-DMSO (chemical shifts d in ppm)

Compound H-C(1)H-C(2)/H-C(3)

H-C(4)/H-C(5) H2C(6) R1(8), R2(9)

2 (150°C) 2.80–5.10 2.80–5.10 2.80–5.10 2.80–5.10 1.85–2.25 (4 CH3), 2.80–5.10 (2 NH),6.40–7.50 (21 arom. H).

3 (140°C) 2.90–5.00 2.90–5.00 2.90–5.00 2.90–5.00 5.25 (2NH), 6.70–7.60 (21 arom. H).4 (150°C) 4.95 4.58, 4.67 3.75 3.35, 3.55 1.90–2.20 (4 CH3), 6.48, 6.57 (2 H-C(48)), 6.94 (2 H-C(28),

2 H-C(68)), 8.40 (NH), 8.57 (NH).5 (100°C) 4.96 4.55, 4.69 3.15–3.97 3.15–3.97 6.92, 7.03 (2 H-C(48)), 7.20–7.50 (2 H-C(28), 2 H-C(68)),

9.36 (NH), 9.75 (NH).6 (120°C) 4.98 4.10–4.90 3.40–3.90 4.10–4.90 1.90–2.40 (5 CH3), 6.37, 6.53 (2 H-C(48), 6.55–7.25 (2

H-C(28), 2 H-C(68), H-C(39), H-C(59)), 7.65 (2 H-C(29)),8.20 (NH), 8.44 (NH).

7 (100°C) 4.97 4.15–4.85 3.63 4.15–4.85 1.98 (CH3), 2.08 (CH3), 6.35, 6.55 (2 H-C(48)), 6.7–7.0 (2H-C(28), 2 H-C(68)), 7.24 (H-C(39), H-C(59)), 7.36 (H-C(49)),7.72 (H-C(29), H-C(69)), 8.42 (NH), 8.65 (NH).

8 (80°C) 5.00 4.15–4.90 3.56, 3.76 4.15–4.90 0.6–1.15 (3 CH3), 1.41, 1.76 (2 CH2), 2.08 (2 H3C-C(38), 2H3C-C(58)), 6.40–6.64 (2 H-C(48)), 6.75–7.10 (2 H-C(28), 2H-C(68)), 8.20 (NH), 8.82 (NH).

9 (80°C) 5.00 3.90–4.85 3.68 3.90–4.85 2.19 (CH3), 6.40–7.65 (10 arom. H), 9.00–9.90 (2 NH).10 (80°C) 4.96 4.55, 4.70 3.70 4.03, 4.27 1.92 (2 CH3), 2.00 (2 CH3), 6.23, 6.47 (2 H-C(48)), 6.60–

6.94 (2 H-C(28), 2 H-C(68)), 6.98 (H-C(48)), 7.26 (H-C(29)),H-C(69)), 8.40 (NH), 8.70 (NH), 9.32 (NH).

11 (80°C) 4.82 4.15–4.70 3.38, 3.62 4.00, 4.15–4.70 1.20 (CH3), 2.07 (2 H3C-C(38), 2 H3C-C(58)), 4.15–4.70(H-C(19)), 6.35–6.65 (2 H-C(48)), 6.75–7.35 (9 arom H, 1NH), 8.40–8.80 (2 NH).

12 (80°C) 4.87 3.20–4.76 3.20–4.76 3.20–4.76 1.25 (CH3), 1.80–2.20 (2 H3C-C(38), 2 H3C-C(58)), 6.25–6.65 (2 H-C(48)), 6.65–7.40 (9 arom H, 1 NH), 8.62 (NH),8.96 (NH).

13 (80°C) 4.95 4.53, 4.65 3.61, 3.77 4.00, 4.20 2.10 (CH3), 6.50 (H-C(48)), 6.55–7.40 (8 arom H)8.83 (NH), 9.37 (2 NH).

TABLE 4. 1H NMR data for the cellulose-2,3-bis-O-arylester series in d6-DMSO (chemical shifts d in ppm)

Compound H-C(1) H-C(2)/H-C(3) H-C(4)/H-C(5) H2C(6) R1(8), R2(9)

14 (150°C) 4.93 2.60–5.50 2.60–5.50 2.60–5.50 1.90–2.45 (2 CH3), 6.40–7.90 (23 arom H).15 (150°C) 3.17–5.30 3.17–5.30 3.17–5.30 3.17–5.30 6.60–8.20 (arom. H).16 (150°C) 2.90–5.50 2.90–5.50 2.90–5.50 2.90–5.50 2.10–2.45 (2 CH3), 6.95–7.32 (2 H-C(38), 2 H-C(58)), 7.40–7.90

(2 H-C(28), 2 H-C(68)).17 (150°C) 4.85 4.85, 5.35 2.55–3.75 2.55–3.75 7.00–8.00 (10 arom H).18 (150°C) 4.90 5.33, 4.50–5.20 3.30–4.20 3.30–4.20 1.90–2.27 (2 H3C-C(48), H3C-C(39), H3C-C(59)), 6.37–7.20

(H-C(49), 2 H-C(38), 2 H-C(58), H-C(29)), H-C(69)), 7.35–7.75(2 H-C(28), 2 H-C(68)), 7.90–8.40 (NH).

19 (150°C) 4.92 5.38, 4.65–5.20 3.40–4.20 3.40–4.20 1.90–2.40 (2 CH3), 6.70–7.75 (11 arom. H), 8.7–9.1 (NH).20 (150°C) 4.30–5.40 4.30–5.40 3.05–4.30 3.05–4.30 1.00–1.43 (H3C-C(19)), 2.10–2.43 (2 H3C-C(48)), 3.05–5.40

(H-C(19)), 6.35 (NH), 6.75–7.40 (7 arom. H), 7.40–7.90(2 H-C(28), 2 H-C(68)).

21 (150°C) 3.00–5.40 3.00–5.40 3.00–5.40 3.00–5.40 1.00–1.43 (H3C-C(19)), 2.05–2.40 (2 H3C-C(48)), 3.00–5.40(H-C(19)) 6.10–6.60 (NH), 6.75–7.40 (7 arom. H), 7.40–7.90(2 H-C(28), 2 H-C(68)).

22 (150°C) 4.92 4.75, 5.35 3.43, 3.88 3.73 1.90–2.30 (2 CH3), 6.60 (H-C(49)), 6.92 (H-C(29), H-C(69)),7.05–7.50 (2 H-C(38), 2 H-C(58), 2 H-C(48)), 7.65 (2 HC-(28),2 H-C(68)), 8.17 (NH).

23 (150°C) 4.94 5.42, 4.83 3.40–4.20 3.40–4.20 6.95–7.48 (9 arom H), 7.62 (2 H-C(28), 2 H-C(68)), 8.86 (NH).

298 ACEMOGLU ET AL.

TABLE 5. Additional chemical, physical, and spectroscopical data of the synthesized cellulose derivatives

Comp. RecrystallizationYield(%) hdyn(MPa*s) IR (KBr, cm−1)

2 NMP/MeOH 91 — 3394 (NH), 3053 (=C-H), 3021, 2915 (C-H), 1753 (C=O), 1615 (=C-C),1539 (NH), 1215 ((O)C-O), 1082 (C-O-C), 838 (trisubst. Ph),706 (monosubst. Ph).

3 CH2Cl2/MeOH 90 — 3383 (NH), 3051 (=C-H), 2926, 2872 (C-H), 1761 (C=O), 1585 (arom.),1525 (NH), 1211 ((O)C-O), 1114 (=C-Cl), 1075 (C-O-C), 840 (trisubst. Ph),706 (monosubst. Ph).

4 THF/MeOH 87 — 3392 (br. OH, NH), 3096, 3011, 2919 (C-H), 1750 (C=O), 1617 (=C-C),1542 (NH), 1223 ((O)C-O), 1079 (C-O-C), 839 (trisubst. Ph).

5 NMP/MeOH 94 — 3410 (OH), 3320 (NH), 3074 (=C-H), 2947, 2872 (C-H), 1755 (C=O), 1594(=C-C), 1532 (NH), 1217 ((O)C-O), 1080 (=C-Cl), 1072 (C-O-C), 842(trisubst. Ph).

6 CH2Cl2/MeOH 88 1.25 3348 (NH), 3010 (=C-H), 2925 (C-H), 1755 (C=O amide), 1726 (C=O ester),1614 (=C-C), 1541 (NH), 1453, 1272, 1217 ((O)C-O), 1178, 1084 (C-O-C),1019, 839 (trisubst. Ph), 752 (disubst. Ph).

7 CH2Cl2/MeOH 80 0.85 3351 (NH), 3011 (=C-H), 2915 (C-H), 1753 (C=O amide), 1732 (C=O ester),1615 (=C-C), 1603, 1541 (NH), 1451, 1271, 1218 ((O)C-O), 1085 (C-O-C),840 (trisubst. Ph), 713 (monosubst. Ph).

8 CH2Cl2/MeOH 68 1.10 3370 (NH), 2969 (C-H), 1797 (C=O lactone), 1751 (C=O amide, ester), 1616(=C-C), 1544 (NH), 1219 ((O)C-O), 1062 (C-O-C), 841 (trisubst. Ph).

9 CH2Cl2/MeOH 77 1.20 3320 (NH), 3075 (=C-H), 2950 (C-H), 1759 (C=O amide), 1710 (C=O ester),1590 (=C-C), 1527 (NH), 1209 ((O)C-O), 1068 (C-O-C), 839 (trisubst. Ph),751 (disubst. Ph).

10 THF/MeOH 81 1.50 3415 and 3330 (NH), 3085 (=C-H), 2920 (C-H), 1734 (C=O amide), 1616(=C-C), 1593, 1539 (NH), 1213 ((O)C-O), 1082 (C-O-C), 841 (trisubst. Ph).

11 THF/MeOH 82 1.25 3395 and 3340 (NH), 2925 (C-H), 1749 (C=O amide I), 1717 (C=O amide II),1615 (=C-C), 1541 (NH), 1223 ((O)C-O), 1083 (C-O-C), 841 (trisubst. Ph),700 (monosubst. Ph).

12 THF/MeOH 85 1.25 3395 and 3340 (NH), 2925 (C-H), 1749 (C=O amide I), 1717 (C=O amide II),1615 (=C-C), 1541 (NH), 1223 ((O)C-O), 1083 (C-O-C), 841 (trisubst. Ph),700 (monosubst. Ph).

13 THF/MeOH 80 1.50 3415 and 3330 (NH), 3080 (=C-H), 2960 and 2920 (C-H), 1758 (C=O amide I),1732 (C=O amide II), 1594 (=C-C), 1536 (NH), 1213 ((O)C-O), 1075 (C-O-C)841 (trisubst. Ph).

14 THF/EtOH 90 — 3500 (very weak, OH), 3085 (=C-H), ca. 2915 (C-H), 1735 (C=O), 1612 (=C-C),1271 ((O)C-O), 1090 (C-O-C), 835 (disubst. Ph), 706 (monosubst. Ph).

15 THF/EtOH 95 — 3060 (=C-H), ca. 2920 (C-H), 1738 (C=O), 1602 (=C-C), 1271 ((O)C-O), 1069–1090 (C-O-C), 706 (monosubst. Ph).

16 THF/MeOH 88 — 3500 (OH), 3040 (=C-H), 2924 (C-H), 1733 (C=O), 1612 (=C-C), 1272 ((O)C-O),1093 (C-O-C), 836 (disubst. Ph).

17 THF/MeOH 96 — 3456 (OH), 3060 (=C-H), 2949 (C-H), 1734 (C=O), 1602 (=C-C), 1272 ((O)C-O),1026–1094 (C-O-C), 706 (monosubst. Ph).

18 THF/MeOH 94 0.96 3340 (NH), 3040 (=C-H), 2920 (C-H), 1735 (C=O ester and amide), 1613 (=C-C),1540 (NH), 1271 ((O)C-O), 1091 (C-O-C), 838 (disubst. Ph.).

19 THF/MeOH 98 0.95 3350 (NH), ca. 3100 (=C-H), 2950 (C-H), 1737 (C=O ester and amide), 1592(=C-C), 1528 (NH), 1269 ((O)C-O), 1096 (C-O-C), 839 (disubst. Ph).

20 THF/MeOH 94 0.87 3360–3420 (NH), ca. 3031 (=C-H), 2971 (C-H), 1734 (C=O ester and amide),1612 (=C-C), 1509 (NH), 1271 ((O)C-O), 1094 (C-O-C), 838 (disubst. Ph), 700(monosubst. Ph).

21 THF/MeOH 93 0.90 3360–3420 (NH), ca. 3031 (=C-H), 2971 (C-H), 1734 (C=O ester and amide),1612 (=C-C), 1509 (NH), 1271 ((O)C-O), 1094 (C-O-C), 838 (disubst. Ph), 700(monosubst. Ph).

22 THF/MeOH 98 0.97 3330 (NH), 3060 (=C-H), 2920 (C-H), 1739 (C=O ester and amide), 1610 (=C-C),1541 (NH), 1271 ((O)C-O), 1070–1093 (C-O-C), 841 (trisubst. Ph), 709(monosubst. Ph).

23 THF/MeOH 98 0.95 3340 (NH), ca. 3080 (=C-H), 2960 (C-H), 1739 (C=O ester and amide), 1592(=C-C), 1528 (NH), 1268 ((O)C-O), 1070–1094 (C-O-C), 844 (trisubst. Ph),709 (monosubst. Ph).


substitution pattern changed to 1.9:1.1 (carbamate:H) forthe products 4 and 5, indicating that about 5% of the car-bamate functions at O-C(2) and O-C(3) were also cleaved,yielding the parent hydroxy groups. This substitution pat-tern was found in all the following products, prepared bysubsequent derivatization at O-C(6) of cellulose-2,3-bis-O-(3,5-dimethylphenyl)carbamate (4) (Table 1). Using com-pound 4 or 5 as starting material, different substituentscould easily be introduced at the O-C(6), yielding productswith 100% homosubstitution at O-C(6) (Scheme 1 andTable 1). Treatment of 4 with the corresponding acid chlo-rides in pyridine at 100°C gave the desired cellulose-2,3-bis-O-carbamate-6-O-ester derivatives 6 and 7 in goodyields. As an example for esterifications with a chiral ali-phatic acid, 4 was treated with (−)-camphanoyl chlorideunder the same conditions to give the compound 8. Fi-nally, several regioselectively heterosubstituted tris-carbamates (compounds 10–12) were prepared by treat-ment of 4 with 3,5-dichlorophenylisocyanate, (R)-(1)-

phenylethylisocyanate, and (S)-(1)-phenylethylisocyanate,respectively. The 3,5-dichlorophenylcarbamoyl substitu-ents were quite sensitive to reaction conditions, yieldingseveral products with unsatisfactory elemental analysis,e.g., products 9 and 10, which were not considered for theanalytical tests as CSPs.

Contrary to carbamoylation reactions, complete esterifi-cations at positions O-C(2) and O-C(3) of 6-O-trityl-cellulose (1) proved to be difficult to achieve. If the corre-sponding benzoyl chlorides were used as reagents, partialdetritylation and subsequent esterification at position O-C(6) occurred, leading to products also benzoylated at po-sition O-C(6), in accordance with the results reported byChassaing et al.13 The de-tritylation occurred even in thepresence of excess triethylamine or pyridine. Base cata-lyzed transesterification reactions with NaOMe or Ti-(iPrO)4 as catalysts also gave no satisfactory conversions.The best results were obtained on esterifications using thefree acids with DCC as activating agent. On treatment of

Fig. 1. Structures of investigatedracemates.

300 ACEMOGLU ET AL.

6-O-trityl-cellulose (1) with an excess of 4-methylbenzoicacid in the presence of DCC and DMAP in CH2Cl2, at least90% of the free hydroxy groups at C(2) and C(3) could bederivatized according to 1H NMR and elemental analysis ofproduct 14. Complete esterification of the hydroxy groupsat C(2) and C(3) was achieved with benzoic acid, as indi-cated by the disappearance of the OH absorption in the IRof product 15. In both cases, negligible (<3%) de-tritylationand esterification at O-C(6) were observed, as could bedetected by 1H NMR. Esterifications in THF gave similarresults, whereas NMP and DMF were not efficient as sol-vent. As in the previous cases, the trityl protecting group atO-C(6) was removed completely using HCl gas in CH2Cl2.Unfortunately, under these conditions both the 4-methylbenzoate and the benzoate ester substituents at O-C(2) and O-C(3) were labile and varying degree of hydro-lysis occurred, leading to a smaller degree of substitutionin the products than in the corresponding 6-O-trityl com-pounds (16, d.s. = 1.7 vs. 14, d.s. = 1.8; 17, d.s. = 1.9 vs.15, d.s. = 2.0). The substitution patterns of 16 and 17remained unchanged during subsequent derivatizations atposition O-C(6), to give the corresponding cellulose-2,3-bis-O-ester-6-O-carbamates 18–23 (Scheme 2 and Table 2). Itis noteworthy to mention that, due to the mild and selec-tive methods chosen for protection-deprotection steps, aswell as for derivatizations, high (>97%) regioselectivity atO-C(6) of the products was also achieved in these cases.However, the selectivities at O-C(2) and O-C(3) were lowerthan in the 2,3-bis-O-carbamate-6-O-ester series, due tohigher acid sensitivity of ester substituents.

The 1H NMR data of the synthesized compounds aregiven in Tables 3 and 4. Some assignments were madetentatively and may be interchanged in these cases. Due tothe polymeric structure combined with a complex substi-tution pattern, many of the synthesized cellulose deriva-tives showed broad, overlapping signals in their 1H NMRspectra, some of which could not be assigned to individualprotons, even at 150°C in DMSO. In these cases, the rangeof overlapping multiplets is given, in which the correspond-ing signal is present. However, the 1H NMR spectra con-firmed the structure of the compounds in all cases. Fromthe integrals of the methyl groups, of the chain protons andof the aromatic protons, which were well separated, theapproximate substitution pattern could be calculated foreach compound, except for compound 15, in which thesignals of the trityl group and the benzoate ester substitu-ent overlapped. Since the OH absorbtion disappeared inthe IR of 15, complete benzoylation was concluded to givea substitution pattern of 2:1 (benzoate:trityl) for this com-pound. Some degradation of the polymer chain may occurduring the acidic de-tritylations. Therefore, dynamic vis-cosities were measured to confirm the polymeric structureof the products (Table 5). Additional chemical data as wellas characteristic IR absorbtions of the products are alsosummarized in Table 5.

Tests of Regioselectively Substituted CelluloseDerivatives as CSP’s in Chiral Chromatography

The structures of the racemates, used for the evaluationof the selectively substituted cellulose derivatives, are

shown in Figure 1. The selection of the racemates wasarranged in a way to ensure that a broad range of struc-turally different racemates is covered. Racemates with oneor more asymmetric centers were included as well as ra-cemates with a heteroatom as the stereogenic center andatropisomers. The synthesized cellulose derivatives werecoated on silica gel to obtain the corresponding CSPs. In afirst set of screenings, the resolution properties of CSPs1–5, prepared from different cellulose-2,3-bis-O-carbamate-6-O-arylesters, were evaluated for the resolutionof the racemates rac-12, rac-15, and rac-16 (Table 6).Although none of the CSPs 1–5 was able to resolve allthree test racemates, CSP 2 (showing highest enantiose-lectivity for rac-12) and CSP 4 (having almost identicalresolution properties to CSP 5) were selected for furthertesting under more optimized conditions.

In a second set of screenings, the resolution propertiesof CSPs 6–11, prepared from different cellulose-2,3-bis-O-arylester-6-O-carbamate derivatives (compounds 18–23),were evaluated for the same test compounds (rac-12, rac-15, and rac-16). The results of this preliminary screeningare summarized in Table 7. The chromatograms obtainedshowed broad peaks, indicating a non-optimized packingprocedure for the columns. Nevertheless, with CSP 7, anexcellent baseline resolution is achieved for rac-15,whereas the resolution of rac-16 was obtainable on CSP 7,CSP 10, and CSP 11. Only CSP 11 exhibited the abilityfor the separation of all three racemates and, therefore,was chosen for further testing.

TABLE 6. Retention times for the second elutedenantiomer (t2), capacity factors (k2*), enantioselectivities(a), and resolution factor (Rs) for investigated racemates

rac-12, rac-15, and rac-16 oncellulose-bis-2,3-O-carbamate-6-O-benzoate CSP 1-5

coated aminopropylfunctionalized silica gel*

Racemate% 2-propanol

in hexaneTemp.(°C) t2(min) k28 a Rs

CSP1 (from comp. 6)12 10 25 16.60 2.32 1.16 0.6315 10 25 23.74 3.75 1.00 0.0016 10 25 23.01 3.60 1.00 0.00

CSP 2 (from comp. 7)12 10 25 17.82 2.49 1.17 0.5615 10 25 27.43 4.38 1.00 0.0016 10 25 23.67 3.64 1.00 0.00

CSP 3 (from comp. 8)12 10 25 13.11 1.43 1.00 0.00a

15 10 25 24.01 3.45 1.00 0.0016 10 25 30.74 4.69 1.00 0.00

CSP 4 (from comp. 11)12 10 25 14.73 2.03 1.00 0.0015 10 25 27.48 4.64 1.00 0.0016 10 25 39.35 7.08 1.17 0.82

CSP 5 (from comp. 12)12 10 25 14.72 2.19 1.10 0.4915 10 25 28.41 5.15 1.00 0.0016 10 25 29.03 5.28 1.22 1.21

*Flow rates for all mobile phases: 0.5 ml/min; UV-detection for rac-12 (220nm), rac-15 (210 nm), and rac-16 (280 nm).aA slight shoulder is observed.


As was pointed out by Okamoto et al.3,17,18 and Francotteand Zhang,11,12 the coating and packing procedure for thepreparation of a CSP from a chiral selector is as essentialfor its performance in the enantiomeric separation as the

chiral selector itself. Therefore, the selected CSPs CSP 2,CSP 4, and CSP 11 were re-packed under standard con-ditions,19 with THF as solvent for the coating procedure.For adequate comparison of the selectors, a reference col-umn was packed under the same conditions, using CSP12 with the known cellulose-tris-(3,5-dimethylphenylcarba-mate) as selector. For the validation of our column packingprocedure, the performance of each column was evaluatedinitially using 1,3,5-tri-tert.-butylbenzene for the estimationof plate numbers and dead times, and using rac-12 asracemic standard. The results are summarized in Table 8,and visualised in the overlay plot of Figure 2 for the sepa-ration of rac-12 on CSP 12 and Chiralcel-OD. It is obviousfrom Table 8 that our packing procedure is not as efficientas the packing of commercially available columns. The re-solving power of some of our new CSPs should be inter-

Fig. 2. Comparison of commercially available Chiralcel-OD (curve 1)vs. home-made and packed cellulose-tris(3,5-dimethylphenylcarbamate)(CSP 12) coated aminopropylfunctionalized silica gel (curve 2) for thecharacterization of the column packing (with 1,3,5-tri-t-butyl-benzene) andthe enantiomeric separation of rac-12; chromatographic parameters aregiven in Table 7, except flow rate being 0.5 ml/min.


rac-12, rac-15, and rac-16 oncellulose-bis-2,3-O-arylester-6-O-carbamate CSP 6-11

coated aminopropylfunctionalized silica gel*



CSP 6 (from comp. 18)12 10 25 23.09 2.50 1.00 0.0015 20 25 21.90 2.85 1.00 0.0016 20 25 26.31 3.62 1.00 0.00

CSP 7 (from comp. 19)12 10 25 17.26 1.91 1.00 0.0015 10 25 34.80 5.58 1.55 2.0416 15 25 21.20 6.07 1.26 1.03

CSP 8 (from comp. 20)12 10 25 23.93 3.21 1.00 0.0015 20 25 20.57 2.71 1.00 0.0016 15 25 21.64 2.56 1.00 0.00

CSP 9 (from comp. 21)12 10 25 23.99 3.16 1.00 0.0015 20 25 21.06 2.81 1.00 0.0016 15 25 16.41 1.75 1.00 0.00

CSP 10 (from comp. 22)12 10 25 20.31 2.58 1.00 0.0015 20 25 21.09 2.60 1.00 0.0016 15 25 20.24 2.38 1.25 0.49

CSP 11 (from comp. 23)12 5 27 25.32 3.13 1.11 0.6115 20 25 31.55 4.18 1.44 1.3916 15 25 19.20 2.20 1.16 0.60

*Flow rates for all mobile phases: 0.5 ml/min; UV-detection for rac-12 (220nm), rac-15 (210 nm), and rac-16 (280 nm).

TABLE 8. Comparison of column packing forChiralcel-OD, CSP 12, CSP 2, and CSP 11 for the elutionof 1,3,5-tri-tert-butylbenzene and the enantioseparation of

rac-12*

Chiralcel-OD CSP 12 CSP 2 CSP 11

N (1,3,5-tri-tert-butylbenzene)

2,171 881 389 1,042

N (1st elutingenantiomerof rac-12)

1,596 524 185 343

N (2nd elutingenantiomerof rac-12)

2,016 571 193 363

k28 2.31 2.52 2.48 2.22a 1.64 1.66 1.15 1.09Rs 3.42 1.94 0.51 0.40

*Mobile phase: 2-propanol/hexane = 1/9 (V/V); flow rate: 1.0 ml/min;temperature: 20°C; UV-detection: 220 nm.


on CSP 12 (reference column)*



1 20 40 55.0 18.88 6.40 9.082 20 40 12.1 3.36 2.06 3.553 10 20 74.2 25.74 1.00 0.004 10 20 35.2 11.73 1.00 0.005 10 20 7.0 1.53 1.25 1.026 10 20 13.1 3.75 1.32 1.087 10 20 34.0 11.28 1.00 0.008 10 20 3.8 0.38 1.00 0.009 10 20 4.9 0.78 1.13 0.53

10 10 20 7.5 1.70 1.41 1.5311 10 20 5.5 0.98 1.65 1.4112 10 20 9.6 2.47 1.62 2.1913 10 20 22.5 7.11 1.00 0.0014 10 20 22.7 7.21 1.01 0.0515 10 20 12.5 3.47 1.00 0.00

*Flow rate: 1.0 ml/min; UV-detection: 210 nm.

302 ACEMOGLU ET AL.

preted in this context. We purposely omitted the resultsobtained with CSP 4 since the stability of the columnturned out to be extremely poor. Evaporation of the mobilephase after each series of separations revealed a significantbleed of the chiral selector, which is likely to be the resultof the moderate solubility of compound 11 in the mobilephase. The bleed increased with increasing amounts of

isopropanol in the mobile phase. Nevertheless, the separa-tion of rac-5 on a freshly prepared column from CSP 4using 10% isopropanol in hexane as mobile phase wasachieved with an enantioselectivity of a = 2.73 (R8 = 1.3)and a resolution factor of Rs = 2.02, illustrating the potentialof this selector. Comparing this result with that obtained byChassaing et al.,13 it is noteworthy to mention that our CSP4 showed significantly higher enantioselectivity (a = 2.73vs. a = 1.72); this is likely to be the result of homosubsti-tution at O-C(6) of our corresponding selector, compound11. It is conceivable that immobilisation of this selector onsilica gel by means of covalent bonding should help tocircumvent its solubility problem and allow repeatedly re-producible enantioseparations.

The results obtained with the reference column fromCSP 12 are summarized in Table 9, confirming the out-standing resolving power of cellulose-tris(3,5-dimethylphe-nylcarbamate) as a chiral selector in liquid chromatogra-phy. For a better comparison with CSP 2, CSP 4, andCSP 11, the whole screening was performed under iden-tical conditions. Typical examples of enantiomeric separa-tions with CSP 12 are shown in Figure 3 (separation ofrac-2, rac-6, rac-11, and rac-12). An interesting observa-tion was the temperature dependence of the enantiomericseparation for rac-4 on this column: separation signifi-cantly improved at higher temperatures. Under standardconditions, not even a partial separation was obtained,

Fig. 3. Chiral separation of rac-2 (curve 1), rac-6 (curve 2), rac-11 (curve 3) and rac-12 (curve 4) on CSP 12; chromatographic parameters aregiven in Table 8.

Fig. 4. Effect of temperature on enantioselectivity for the chiral sepa-ration of rac-4 on Chiralcel-OD; chromatographic parameters are given inTable 8, except temperature being 20°C (curve 1), 40°C (curve 2), and50°C (curve 3).


while a baseline separation was achieved at 50°C. For bet-ter visualisation, the temperature dependency was repeat-edly demonstrated on Chiralcel-ODt and is shown in theoverlay plot of Figure 4. A similar phenomenon had alreadybeen reported in the literature, e.g., for the separation ofracemic methyl phenyl sulphoxide on Chiralcel-OBt20 andfor rolipram on Chiralcel-ODt.21

The results of the detailed investigation of CSP 2 understandard conditions are summarized in Table 10, demon-strating a broad applicability for this CSP. Good enantios-electivities are observed for the separations of rac-2 (a =1.18), rac-3 (a = 1.22), rac-5 (a = 2.00), rac-10 (a = 1.18),and rac-12 (a = 1.19). Noteworthy is the separation ofrac-3 with CSP 2 (Fig. 5), since this racemate has so faronly been separated on protein bonded phases and was notseparable on cellulose-derived stationary phases. Never-theless, the best enantioseparation is achieved for rac-5,indicating the excellent suitability of this CSP for the sepa-rations of racemates having relatively rigid conformation.Of particular interest again is the comparison of our CSP 2with that reported by Chassaing et al.:13 While higher reso-lution factors are observed by Chassaing et al. for the sepa-ration of rac-5 (Rs = 1.71 vs. Rs = 1.50), indicating a betterpacking procedure for the column, the enantioselectivity ofour CSP 2 is significantly higher (a = 2.00 vs. a = 1.37),underlining the importance of homosubstitution at O-C(6)of the selector for chiral recognition.

The results of the detailed investigation of CSP 11 un-

Fig. 5. Chiral separation of rac-2 (curve 1), rac-3 (curve 2), rac-5 (curve 3), and rac-12 (curve 4) on CSP 2; chromatographic parameters are givenTable 9.


on CSP 2*



1 10 25 39.95 12.06 1.00 0.002 10 25 12.35 3.04 1.18 0.56

3 15 33.04 9.67 1.17 0.473 10 25 56.95 17.61 1.22 0.634 10 25 44.28 12.97 1.00 0.005 10 25 6.73 1.20 1.95 1.50

3 15 10.41 2.36 2.00 1.516 10 25 13.29 3.34 1.00 0.007 10 25 35.69 10.66 1.00 0.008 10 25 5.02 0.64 1.00 0.009 10 25 5.94 0.94 1.00 0.00

10 10 25 6.99 1.29 1.16 0.483 15 11.45 2.70 1.18 0.57

11 10 25 4.05 0.32 1.00 0.0012 10 25 9.17 2.00 1.17 0.59

3 15 14.79 3.77 1.19 0.6713 10 25 26.44 7.64 1.00 0.0014 10 25 22.88 6.48 1.00 0.0015 10 25 13.20 3.31 1.00 0.0016 10 25 12.31 3.02 1.00 0.00a

3 15 13.06 3.22 1.10 0.60

*Flow rate: 1.0 ml/min; UV-detection: 210 nm.aA slight shoulder is observed.

304 ACEMOGLU ET AL.

der standard conditions are summarized in Table 11. Evenunder these conditions, which were not optimized, a chiralseparation could be achieved for approximately 50% of alltest racemates. The most outstanding result was obtainedin the case of rac-15, for which a baseline separation of theenantiomers (Rs = 1.15) was observed. Not even a partialresolution could be achieved on the reference column(CSP 12) for this racemate. Typical results for the sepa-rations with CSP 11 are shown in Figure 6. From theresults obtained so far, it seems likely that the applicationof CSP 11 can be mainly envisaged for enantiomeric sepa-rations of amino acid derivatives and racemates containinga heteroatom as stereogenic center.

CONCLUSIONS

The synthesis of rationally designed, regioselectivelyheterosubstituted cellulose derivatives was achieved withup to complete regioselectivity at O-C(6). A selection of thesynthesized compounds was successfully utilized as CSPsin chiral chromatography. The importance of homosubsti-tution at O-C(6) of regioselectively heterosubstituted cel-lulose derivatives for high enantioselectivity was con-firmed. Regioselectively substituted heterofunctional cellu-loses having 2,3-bis-O-carbamate-6-O-ester functions orvica versa do not combine the resolution power of theknown cellulose-tris-carbamate and cellulose-tris-benzoate

Fig. 6. Chiral separation of rac-2 (curve 1, with 20% 2-propanol), rac-5 (curve 2), rac-6 (curve 3), and rac-15 (curve 4) on CSP 11; chromato-graphic parameters are given in Table 10.


on CSP 11*



1 20 40 7.2 1.52 1.00 0.0010 20 35.8 11.56 1.15 0.69

2 20 40 10.3 2.60 1.77 1.7610 20 16.6 4.83 1.70 1.19

3 10 20 32.7 10.47 1.00 0.004 10 20 13.5 3.74 1.00 0.005 10 20 8.1 1.83 1.57 1.006 10 20 28.6 9.02 1.08 0.437 10 20 29.9 9.49 1.00 0.008 10 20 3.7 0.31 1.00 0.009 10 20 7.5 1.62 1.00 0.00

10 10 20 11.4 2.99 1.00 0.0011 10 20 5.1 0.79 1.00 0.0012 10 20 10.4 2.63 1.00 0.00a

3 20 14.1 3.94 1.11 0.5113 10 20 17.8 5.24 1.00 0.0014 10 20 36.2 11.74 1.00 0.0015 10 20 30.8 9.79 1.51 1.15

*Flow rate: 1.0 ml/min; UV-detection: 210 nm.aA slight shoulder is observed.


phases. However, they seem to be more selective for cer-tain classes of racemates. They must be considered ratheras complementary CSPs than replacements for the widelyused commercial CSPs.

LITERATURE CITED1. Ichida, A., Shibata, T., Okamoto, I., Yuki, Y., Namikoshi H., Toga, Y.

Resolution of enantiomers by HPLC on cellulose derivatives. Chro-matographia 19:280–284, 1984.

2. Francotte, E., Zhang, T. Supramolecular effects in the chiral discrimi-nation of meta-methylbenzoyl cellulose in high-performance liquidchromatography. J. Chromatogr. A 718:257–266, 1995.

3. Okamoto Y., Aburatani, R., Hatada, K. Cellulose tribenzoate derivativesas chiral stationary phases for high-performance liquid chromatogra-phy. J. Chromatogr. 389:95–102, 1987.

4. Francotte E., Wolf, R.M. Modulation of the chiral recognition by varia-tion of the position of the methyl group on the aromatic ring. J. Chro-matogr. 595:63–75, 1992.

5. Okamoto Y., Kawashima, M., Hatada, K. Controlled chiral recognitionof cellulose triphenylcarbamate derivatives supported on silica gel. J.Chromatogr. 363:173–186, 1987.

6. Okamoto Y., Kaida Y. Polysaccharide derivatives as chiral stationaryphases in HPLC. J. High Resol. Chromatogr. 13:708–712, 1990.

7. Grieb, S.J., Matli, S.A., Phillips, J.G., Belenguer, A.M., Ritchie, H.J.Chiral HPLC with carbohydrate carbamates: influence of supportstructure and enantioselectivity. Chirality 6:129–134, 1994.

8. Okamoto Y., Cao, Z., Aburatani, R., Hatada, K. Optical resolution ofalcohols as carbamates by HPLC on cellulose tris(phenylcarbamate)derivatives. Bull. Chem. Soc. Jpn. 60:3999–4003, 1987.

9. Zhang, T., Francotte, E. Chromatographic properties of composite chi-ral stationary phases based on cellulose derivatives. Chirality 7:425–433, 1995.

10. Kaida, Y., Okamoto, Y. Optical resolution on regioselectively carbam-oylated cellulose and amylose with 3,5-dimethylphenyl and 3,5-dichlorophenyl isocyanates. Bull. Chem. Soc. Jpn. 66:2225–2232, 1993.

11. Francotte, E.R., Zhang, T. Effects moleculaires en chromatographiesur phases stationnaires chirales a base de cellulose. Analusis Maga-zine 23:13–20, 1995.

12. Francotte, E.R., Zhang, T. Influence of the molecular and supramo-lecular structure of cellulose based CSPs on their chiral recognitionability. GIT Laborfachzeitschrift, 40:15–19, 1996.

13. Chassaing, C., Thienpont, A., Felix, G. Regioselective carbamoylatedand benzoylated cellulose for the separation of enantiomers in high-performance liquid chromatography. J. Chromatogr. A 738:157–167,1996.

14. Pfannemuller, B., Berg, A. Zur frage des abbaus der polysaccharid-hauptkette bei verzweigung uber glykosidische bindungen. Makro-mol. Chem. 180:1183–1199, 1979.

15. Cashion, P.J., Fridkin, M., Agarwal, K.L., Jay, E., Khorana, G.H. Use oftrityl- and a-naphtylcarbamoylcellulose derivatives in oligonucleotidesynthesis. Biochemistry 12:1985–1990, 1973.

16. Harkness, B.R., Gray, D.G. Preparation and chiroptical properties oftritylated cellulose derivatives. Macromolecules 23:1452–1457, 1990.

17. Okamoto, Y., Kawashima, M., Hatada, K. Useful chiral packing mate-rials for high-performance liquid chromatographic resolution of enan-tiomers: Phenylcarbamates of polysaccharides coated on silica gel. J.Am. Chem. Soc. 106:5357–5359, 1984.

18. Okamoto, Y., Kawashima, M., Hatada, K. Controlled chiral recognitionof cellulose triphenylcarbamate derivatives supported on silica gel. J.Chromatogr. 363:173–186, 1986.

19. Okamoto, Y., Kawashima, M., Yamamoto, Y., Hatada, K. Useful chiralpacking materials for high performance liquid chromatographic reso-lution. Cellulose triacetate and tribenzoate coated on macroporoussilica gel. Chem. Lett. 1984:739–742.

20. Kusters, E., Loux, V., Schmid, E., Floersheim, Ph. Enantiomeric sepa-ration of chiral sulphoxides: screening of cellulose-based sorbents withparticular reference to cellulose tribenzoate. J. Chromatogr. A 666:421–432, 1994.

21. Kusters, E., Spondlin, Ch. Influence of temperature on the enan-tioseparation of rolipram and structurally related racemates on Chiral-cel-OD. J. Chromatogr. A 737:333–337, 1996.

306 ACEMOGLU ET AL.

Synthesis of regioselectively substituted cellulose derivatives and applications in chiral chromatography

Documents