Analyst May 1983 Vol. 108 pp. 597-602 597 Substitution in Cellulose Ethers Part 11." Substituents on the Glucose Units Using High-performance Liquid Determination of the Distribution of Alkoxyl ~ ~~ Chromatography Konrad Sachse and Klaus Metzner and Thomas Welsch Kombinat V E B Chemische Werke Buna DDR-42 12 Schkopau German Democratic Karl-Marx- Universitat Leipzig Saktion Chemie Analytisches Zentrum Liebigstr German Democvatic Republic Republic 18 DDR-7010 Leipzig, A procedure is described for the quantitative characterisation of the distribu-tion of alkoxyl groups in methyl- and ethylcelluloses. The total hydrolysate is derivatised by reaction with benzoyl chloride in sodium hydroxide solution. Several favourable implications of pre-column derivatisation for the subsequent separation by high-performance liquid chromatography are discussed.Isocratic separations are carried out on LiChrosorb Si100 10 pm using chloroform with 0.1 and 0.2% V / V ethanol as the eluent. All eight individual glucose ethers present in the hydrolysates including the positional isomers of alkyl- ancl dialkylglucoses can be determined quantitatively from the chroma-tograms. Thus the procedure provides data on the distribution of substituents according to their number and position in the anhydroglucose units. Keywords A lkylcellulose ; substituent distribution ; position of substituents ; @re-column derivatisation ; high-performance liquid chromatography In Part I of this series results of the use of quantitative thin-layer chromatography (TLC) for the characterisation of the substitution in cellulose ethers have been rep0rted.l The proced-ure was shown to provide detailed information on the proportions of tri- di- mono- and unsubstituted anhydro-glucose units in the polymer.For a better understanding of the colloidal chemical properties of alkylcelluloses it is desirable to obtain additional information concerning the positions of the alkoxyl substituents in the monomeric unit. The problem of characterising quantitatively this particular distribution has not yet been solved. The time-consuming procedure of fractionation of a methylcellulose h ydrolysate by classical liquid chromatography as reported by Croon and Manley in 1963,2 does not represent a comprehensive analytical procedure. Although with the development of high-performance liquid chromatography (HPLC) in the last decade many workers have achieved fast separa-tions of mixtures of sugars and/or their deri~atives,~-ll there is no evidence in recent literature of a more efficient solution to this separation problem.The aim of this work was to determine by HPLC all of the components including the positional isomers in alkylcellulose hydrolysates thus providing an analytical procedure for the quantitative characterisation of the substituent distribution according to the positions in the glucose ring. * For details of Part I of this series see reference list p. 602 598 SACHSE et al. SUBSTITUTION Experimental Samples Analyst Vol. 108 The cellulose ethers investigated were technical-grade samples of methylcellulose and ethylcellulose.Reagents grade (VEB Laborchemie Apolda G.D.R.). chloroform was passed through a column of activated silica gel prior to use in HPLC. amount of residual ethanol in the purified solvent was determined by gas chromatography. from Koch-Light Laboratories Ltd. Colnbrook Buckinghamshire. Glucose chloroform benzoyl chloride and sodium hydroxide were all of analytical-reagent For the removal of polar components the The 3-O-Methyl-a-~-glucopyranose and 2,3-di-0-methyl-a-n-glucopyranose were purchased Hydrolysis For total hydrolysis of 100-mg amounts the treatment in 7 ml of 2% V/V sulphuric acid at 140 "C for 1 h in a sealed glass ampoule was sufficient. Hydrolytic cleavage of the samples has been described in detail previous1y.l Derivatisation For perbenzoylation 100 mg of the dry cellulose ether hydrolysate are reacted with 1 ml of benzoyl chloride in 4 ml of 15% m/m sodium hydroxide solution in a 20-ml reaction vial.While the mixture is shaken vigorously a perceptible temperature rise indicates the beginning of the reaction. Subsequently the system is cooled down in ice - water and again shaken until self-heating is detected. These two steps are repeated several times over 10 min until no more heat is liberated. To be certain that the reaction has run to completion the mixture can be stored at 0-5 "C overnight. The benzoates formed appear as a very viscous liquid or semi-solid particles. They are removed from the aqueous phase by extraction with 10ml of chloroform. After evaporating off the solvent at 60 "C a 0.5% m/V solution of the reaction products in the eluent is prepared 5-10 pl of which are injected on to the column.High-performance Liquid Chromatography Most of the separations were carried out with an apparatus consisting of a micro-dosing pump (MC 706L Mikrotechna Prague Czechoslovakia) a damping unit laboratory-built 5-and 10-p1 sample loops and a variable wavelength ultraviolet detector (VSU-2P spectrophoto-meter VEB Carl Zeiss Jena G.D.R.). Quantitative evaluation was accomplished using the commercially available chromatographic data system AS 4 (VEB Chromatron Berlin G.D.R.). For part of the work a liquid chromatograph Model 1084B (Hewlett-Packard Avondale, PA USA) was used. The columns (250 x 4 mm) were made in the laboratory from stainless-steel tubing with a low dead volume design of the inlet and outlet parts.LiChrosorb 5100 10 pm (E. Merck, Darmstadt F.R.G.) served as the stationary phase. The columns were packed using a slurry of the sorbent in carbon tetrachloride - dioxane (1 + 1 V / V ) that was pumped with heptane into the column. Packing quality was checked via the parameter A as proposed by KnoxI2 and other workers. A values of between 1 and 1.7 confirmed that good packing had been achieved by this technique. Results and Discussion Derivatisation After total hydrolysis of the cellulose ether a mixture of glucose 2- 3- and 6-O-alkylglucoses, 2,3- 2,6- and 3,6-0-dialkylglucoses and 2,3,6-0-trialkylglucose remains to be separated. Possible anomerisation during the hydrolytic process and/or subsequent reactions is not relevant in the present investigations.To solve this separation problem it appeared useful to consider pre-column ultraviolet derivatisation; there were two main reasons for this. Firstly it allows the ultraviolet detector to be employed which would be specific and more sensitive towards these substances after th May 1983 IN CELLULOSE ETHERS. PART I1 599 introduction of chromophores and generally is easier to handle in comparison with the refractive index detector normally used for underivatised carbohydrates. Secondly if all free hydroxyl groups are converted as a result of derivatisation so that the polarities of the sample molecules are decreased the alkoxyl groups would in this instance give relatively larger contributions to the over-all retention on silica gel than they would be able to provide in the non-derivatised molecules where the interaction of hydroxyl groups with the sorbent would be dominant.From this point of view the introduction of less polar groups into the glucose ether molecules can be expected to facilitate a separation of positional isomers at appropriate chromatographic conditions. Furthermore this step leads to better solubility in hydro-phobic solvents which opens up a wider range of possibilities in the choice of the mobile phase. Several known procedures for the ultraviolet derivatisation of monosaccharides have been tested.lo~l1~l3 Results have shown that perbenzoylation does fulfil all the requirements demanded for a pre-column derivatisation reaction in HPLC if the conditions are chosen according to the classical Schotten - Baumann acylation.Esterification of the glucoses is then easily carried out giving the highest yields of all the reactions tested (80% of the theoretical value) and only very small amounts of by-products are formed. Moreover the performance of the derivatisation reaction in an aqueous medium offers several advantages over benzoyla-tion in pyridine,ll as the mixture need not be heated and the final products are simply isolated by extraction. Chloroform has been used as the extracting solvent in accordance with the choice of the eluent described below. 3 Time/min Fig. 1. Chromatogram of a perbenzoylated test mixture 1 ethyl benzo-ate; 2 methyl benzoate; 3 and 4 glucose; 5 , 3-0-methylglucose ; and 6,2,3-di-O-methylglucose.Column 250 x 4 mm, LiChrosorb Si100 10 pm; eluent chloroform con-taining 0.2% V / V of ethanol; flow-rate 1.0 ml min-l; and detection a t 278 nm. The formation of a single derivatisation product of each sample component is illustrated by the chromatogram of a perbenzoylated test mixture in Fig. 1. lH nuclear magnetic resonance spectra of the benzoylation product of glucose supported this and confirmed that all five hydroxyl groups had been converted. The double peak for glucose in Fig. 1 is caused by a partial separation of Q- and P-anomers. Small amounts of ethyl and methyl benzoates can appear as by-products owing to the presence of trace amounts of ethanol in the chloroform use 600 SACHSE et al. SUBSTITUTION Analyst Vol.108 for extraction and methanol in the hydrolysate. However these substances are almost un-retained under the chromatographic conditions used here and therefore do not interfere with the determination of the components of interest. Separation of the Hydrolysates Preliminary TLC tests on silica gel suggested that chloroform should be a suitable eluent for the separation of the glucose - ether benzoates mixture. Further investigations have shown that isocratic separations on LiChrosorb Silo0 are possible if chloroform with defined small amounts of ethanol is used as the mobile phase. The selectivity of the chromatographic system depends heavily on the percentage of alcohol present. Thus 0.2% V/V of ethanol in chloroform was found to be optimum for the separation of the eight species of the methyl-cellulose hydrolysate (Fig.2) whereas for ethylcellulose 0.1 yo V/V was appropriate (Fig. 3). The alcohol’s function in the system is that of a typical moderator as described for example by Engelhardt.14 Special attention must be paid to maintain the exact eluent composition, as a 0.1% change of the ethanol concentration may cause shifts of capacity factors amounting to several units. Another necessary condition for good resolution of all eight sample components under the selected chromatographic conditions involves column performance. I t has been established that the critical pair of peaks (peaks 7 and 8 in Fig. 2) will be resolved sufficiently (&> l) if at least 3 500 theoretical plates are generated for 2,3,6-0-trimethylglucose (capacity factor 9).The complete assignment of the peaks in Fig. 2 is complicated by the fact that for most of the methyl- and dimethylglucose isomers no reference substances are available. In order to establish the number of methoxyl groups present in the various components a preparative TLC separation has been carried out (for the conditions see reference 1). The isolated fractions of glucose methyl- dimethyl- and trimethylglucoses have been perbenzoylated in the described manner and finally injected on to the column. As expected glucose (peak 1) and trimethylglucose (peak 7) showed only one peak while the methyl- and dimethylglucose fractions were both separated into three components (peaks 2 3 and 5 and 4 6 and 8 respec-ively).For the assignment of those peaks representing positional isomers we used the relative velocity constants of methoxylation of the various hydroxyl groups. Knowing that position 2 is the most reactive followed by 6 and 3,l5$l6 it should be justified to ascribe the largest of the three methylglucose peaks to 2-O-methylglucose the second largest to 6-0-methylglucose and the smallest to 3-0-methylglucose. The assignment of the latter has been confirmed by running the reference substance. The identity of the dimethylglucose peaks has been estab-lished on the same basis again supported by one reference substance (peak 4). 1 I Time/min Fig. 2. Separation of a perbenzoylated hydrolysate of inethylcellulose 1 glucose (two anomers) ; 2 3-0-methylglucose ; 3 2-0-methyl-glucose ; 4 2,3-di-O-methylglucose ; 5 6-0-methylglucose; 6 2,6-cli-O-niethylglucose ; 7, 2,3,6-tri-O-niethylglucose ; arid 8 3,6-di-0-methylglucose.Conditions as in Fig. 1 May 1983 IN CELLULOSE ETHERS. PART I1 601 The analysis of the non-derivatised samples on phase systems usually employed for carbo-hydrate separations (LiChrosorb NH, E. Merck F.R.G. ; acetonitrile - water mixtures) was attempted but no separation of the positional isomers could be achieved. Moreover it be-came obvious t hat pre-chromatographic derivatisation according to our procedure resulted in a decrease in the detection limits by 2-3 orders of magnitude. Thus for instance a few nano-grams of glucose were still detected in the ultraviolet region after perbenzoylation. I 23 0 5 10 15 20 25 Time/min Fig.3. Separation of a perbenzoylated hydrolysate of ethylcellulose 2 glucose; 1,3 and 5 ethylglucoses ; 4,6 and 8 diethylglucoses ; and 7 triethylglucose. Eluent chloroform containing 0.1 % V / I' of ethanol ; other condi-tions as in Fig. 1. At 7 min the sensitivity of the plotter was increased. Quantitative Determination For the quantitative evaluation of the chromatograms the different ultraviolet absorbances, which must be expected because of the different numbers of benzoyl groups attached to the various species had to be taken into account. The relative responses given in Table I have been determined by chromatographing derivatised test mixtures containing defined amounts of glucose 3-O-methylglucose and 2,3-0-di-methylglucose. As a result a linear dependence of the logarithm of relative responses on the number of benzoyl groups has been established.It has been assumed that the ultraviolet response is identical for positional isomers and depends ultimately on the number of benzoyl groups attached to the molecule. As for 2,3,6-0-tri-methylglucose no reference substance was available the relative response of this component has been established by extrapolation of the above mentioned dependence. In Table I1 the results of a series of determinations with a methylcellulose sample are presented. The values obtained give a detailed picture of the alkoxyl groups' distribution and TABLE I EXPERIMENTAL VALUES OF RELATIVE ULTRAVIOLET RESPONSES FOR PERBENZOYLATED COMPONENTS Component Relative response at 278 nm Glucose .. 1 Methylglucose . . 0.63 Dimethylglucose . . 0.37 Trimethylglucose . . 0.23* * Extrapolated value 602 SACHSE METZNER AND WELSCH confirm the reproducibility of our procedure as a whole as well as that of all the steps involved, i.e. hydrolysis derivatisation and HPLC. The average degree of substitution DS is calcu-lated according to the equation DS = c2 + c3 + c5 + qc + cg + c,) + 3c7 100 where c2 . . . c8 are the molar percentages of the components numbered as in Fig. 2. The calculated DS of the methylcellulose sample is in good agreement with the value of 1.86 obtained by quantitative TLC. It is again higher than the corresponding Zeisel value of 1.76. Possible reasons for this difference have been discussed e1~ewhere.l.~~ For ethylcellulose analogous information has been obtained by this procedure.TABLE I1 QUANTITATIVE CHARACTERISATION OF THE SUBSTITUENT DISTRIBUTION I N A METHYLCELLULOSE SAMPLE Component Glucose unsubstituted . . . . . . Methylglucoses-2-O-Methylglucose . . . . 3-O-Methylglucose . . . . 6-O-Methylglucose . . . . 2,3-O-Dimethylglucose . . 2,6-O-Dimethylglucose . . . . . . . . 3,6-O-Dimethylglucose . . 2,3,6-0-Trimethylglucose . . Dimethylglucoses-Trimethylglucose-Mass proportion, 10.3 % mlm 11.6 1.5 6.8 9.3 19.2 13.5 27.8 Standard deviation (n = 9) yo mlm 0.7 0.8 0.2 0.8 0.7 0.5 1.4 0.9 Degree of substitution . . 1.82 0.02 Conclusions It has been shown that HPLC offers new possibilities for a very detailed characterisation of the substitution in alkylcelluloses.The reproducibility of the whole procedure described in this paper has been confirmed. It is possible to determine quantitatively the proportions of monomeric units according to number and position of the ether substituent. The proposed procedure will serve as a valuable tool for obtaining more detailed information about the chemical structure of cellulose ethers. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. References Sachse K. Metzner K. and Welsch T. AnaZyst 1982 107 53. Croon B. I. and Manley R. S. J. in Whistler R. L. Editor “Methods in Carbohydrate Chemistry, Palmer J. K. Appl. Polym. Symp. 1975 28 237. Jones A. D. Burns I. W. Sellings S. G. and Cox J . A. J. Chromatogr. 1977 144 169. McGuinnis G. D. and Fang P. J. Chromatogr. 1978 153 107. Ugrinovits M. Chromatographia 1980 13 386. Yang M. T. Milligan L. P. and Mathison G. W. J. Chromatogr. 1981 209 316. Orth P. and Engelhardt H. Chromatographia 1982 15 91. Sinner M. J. Chromatogr. 1976 121 122. Nachtmann F. Fresenius 2. Anal. Chem. 1976 282 209. Lehrfeld J. J. Chromatogr. 1976 120 141. Knox J. H J. Chromatogr. Sci. 1977 15 352. Meyer H. “Analyse und Konstitutionsermittlung Organischer Verbindungcn,” Springer Vienna, Engelhardt H. J. Chromatogr. Sci. 1977 15 380. Croon I. Sven. Papperstidn. 1960 63 247. Rogowin S. and Deriwitzkaja W. Faserforsch. Textiltech. 1957 8 61. Sachse K. Dissertation Leipzig 1982. Volume 111 Cellulose,” Academic Press New York and London 1963 p. 281. 1938 p. 587. NOTE-Reference 1 is to Part I of this series. Received November 15th 1982 Accepted November 24th 198