Anal. Chem. 1981, 5 3 , 1818-1821
1818
Table VI. The Mark-Houwink Parameters Calculated from Equations 8 and 9 and 21 table or eq no. eq no. sample or t and s 8and 9 I1 PVC-1 PVC-2 PVC-3 PVC-4 8and 9 I11 PVC t = 1.140 s = 0.118 8and9 111 PVC t = 1.099 s = 0.189 Band 9 18 PMMA 8and9 19 PVAc 21 I11 PVC-1 and PVC-3 t = 1.140 s = 0.118 21 I11 PVC-1 and PVC-4 t = 1.140 s = 0.118 21 PMMA 21 PS
a
0.482 0.361 0.514 0.492 0.497
KX
lo4
49.8 182
0.552
31.1
37.6 39.2 21.2
calibration curve for the polymer sample is already constructed, parameters K and a can be computed from eq 21 by the search technique. The results, which appear to be in fairly good agreement with the literature values, are shown in Table VI. These parameters are valid at 25 O C in THF. Polymer samples, PMMA and PS, as a pair polymer were prepared in our laboratory and the values of molecular weights and intrinsic viscosity are Hw= 1.20 X lo6, H,, = 7.0 X lo4, and [q] = 0.505 (PMMA) and Mw= 7.5 X lo4,Hn= 3.8 X lo4, and [q] = 0.319 (PS). The Mark-Houwink parameters of any polymers calculated by this method are independent on those of the PS primary standards.
LITERATURE CITED
0.858 0.613 0.753
1.58 1.59
0.706
2.79
0.764 0.705
0.699 1.23
0.455
literature values (a = 0.63, K = 3.5 X lo4 (19)). This calculation technique to obtain the Mark-Houwink parameters involves the Mark-Houwink parameters of PS, which seem to have considerable errors, resulting in inaccurate values. The viscosity-average molecular weight of a polymer sample may be computed from the chromatogram hi of the polymer by the formula Substituting in the Mark-Houwink equation gives where a and K are the Mark-Houwink parameters and [q]is the intrinsic viscosity of the polymer. If there are two polymer samples whose intrinsic viscosity values are known and if a
(1) Grubisic, 2.; Rempp, P.; Benoit, H. J. folym. Scl., Part B 1987, 5 , 753-759. (2) Moore, J. C.; Hendrlckson, J. G. J. Polym. Sci., Part C 1985, 8 , 233-241. (3) Mori, S. J. Chromatogr. 1978, 757, 75-84. (4) Mori, S. J . Appl. folym. Scl. 1974, 18, 2391-2397. (5) Baike, S. T.; Hamielec, A. E.; Leclair, B. P.; Pearce, S. L. Ind. €ng. Chem. Prod'. Res. Dev. 1989, 8 , 54-57. (6) Loy, B. R. J . folym. Sci., folym. Chem. Ed. 1976, 14, 2321. (7) Vrijbergen, R. R.; Soeteman, A. A.; Smit, J. A. M. J. Appl. folym. Sci. 1978, 22, 1287-1276. (8) McCrackin, F. L. J. Appl. folym. Scl. 1977, 21, 191-198. (9) Mahabadi, H. K.; ODriscoli, K. F. J. Appl. folym. Scl. 1977, 21, 1283-1287. (10) Mori, S.; Suzukl, T. J . Llq. Chromatogr. 1980, 3 , 343-351. (11) Welss, A. R.; Cohn-Ginsberg, E. J. fo&m. Sci., Part B 1989, 7 , 379-38 1. (12) Morris, M. C. J . Chromatogr. 1971, 55, 203-210. (13) Hamieiec, A. E.; Omorodion, S. N. E. ACS Symp. Ser. 1980, No. 738, 183-196. (14) Provder, T.; Woodbrey, J. C.; Clark, J. H. S e p . Scl. 1971, 6, 101-136. (15) Mori, S.; Yamakawa, A. J. Llq. Chromatogr. 1980, 3 , 329-342. (16) Hamieiec, A. E.; Ray, W. H. J . Appl. folym. Scl. 1989, 73, 1319-1321, (17) Provder, T.; Rosen, E. M. S e p . Sci. 1970, 5, 437-464. (18) Ouano, A. C.; Dawson, B. L.; Johnson, D. E. "Liquid Chromatography of Polymers and Related Materials"; Cazes, J., Ed.; Marcel Dekker: New York, 1977; pp 1-9. (19) Goedhart, D.; Opshoor, A. J. folym. Sci., folym. Phys. Ed. 1970, 8 , 1227- 1237.
RECEIVED for review April 27,1981. Accepted June 24,1981.
Determination of Alkyl Chain Distribution of Alkylbenzenesulfonates by Liquid Chromatography Atsuo Nakae," Kazuro Tsujl, and Makoto Yamanaka Tochigi Research Laboratories, Kao Soap Co., Ltd., 2606, Akabane, Ichikai-machi, Haga-gun, Tochigi, 32 1-34, Japan
Llnear alkylbenrenesulfonates (LAS) were separated according to thelr alkyl chain length, and phenyl Isomers were also well separated by the reversed-phase hlgh-performance llquld chromatography (HPLC), employlng octadecylsllanlzed slllca gel as a stationary phase and aqueous acetonltrlle solutlon containing sodlum perchlorate as a moblle phase. The results obtalned by HPLC agreed very closely with those by the caplllary-GC analysis of desulfonated LAS with hot phosphoric acid. LAS in detergents was determined directly without any pretreatments prior to the analysis.
Linear alkylbenzenesulfonate (LAS) is an anionic surfactant widely employed in detergent formulations. The commercial
product is generally a mixture of a homologous series of alkyl chain lengths of 10-13 carbon atoms, and the phenyl group is attached to any methylene group of the alkyl chain. The determination of the alkyl chain distribution of LAS has been carried out by gas chromatography (GC). This method, however, requires the conversion of LAS into the volatile derivatives before analysis. Desulfonation with acids (I-4), alkali fusion (5),sulfochlorination (6), methylation (7), reduction to the alkylthiophenol (8),pyrolysis-GC (91, and "acid" pyrolysis-GC (IO, 11)have been well-known methods of prederivatization for the GC determination. High-performance liquid chromatography (HPLC) is the most suitable method for the determination of the alkyl chain distribution of LAS, because it does not require conversion of LAS into the volatile derivatives. In a previous paper (12),
0003-2700/81/0353-1818$01.25/00 1981 American Chemical Society
ANALYTICAL CHEMISTRY, VOL. 53, NO. 12,OCTOBER 1981
we reported that trace amounts of LAS were determined fluorimetrically employing 5-pm LiChrosorb RP-18 as a stationary phase and 0.1 M NaC104 in CH30H/H20 (8:2) as a mobile phase. Under the above chromatographic condition, LAS was separated completely according to the alkyl chain lengths and the %phenyl isomers were separated from the other phenyl isomers. In this paper, we describe the more efficient separations of the phenyl isomers of LAS and the determination of the alkyl chain distribution employing a Hitachi Gel 3053 instead of LiChrosorb RP-18 as a stationary phase and aqueous acetonitrile solution instead of aqueous methanol solution as a mobile phase. Detergent for clothes is a mixture of surfactants and inorganic ingredients, and the surfactants used are also a mixture of various types, such as, LAS, soap, polyoxyethylene alkyl ether, etc. The contents and the types of surfactants and their alkyl chain distributions influence the performance of the detergents. Therefore, it is very important to analyze the surfactants in detergents. The determination of alkyl chain distributions of LAS is one of the interesting subjects for both detergent and surfactant analysis. The proposed HPLC method determines the alkyl chain distributions in detergents without any pretreatments.
EXPERIMENTAL SECTION Apparatus. The liquid chromatograph consisted of a reciprocating piston pump (Model 635, Hitachi Scientific Instrumental Co., Tokyo, Japan), a variable-wavelengthUV detector (UVIDEC 100, Japan Spectroscopic Co., Tokyo, Japan), a variable injector (Hitachi 635), a recorder (Model U-l25M, Nihondenshikagaku Co., Kyoto, Japan), and a circulator (Model FE, Haake Inc., Karlsruhe, West Germany) for column temperature control. Peak areas and retention times were obtained by using a data processor (CHROMA'I'OPAC-E1 A, Shimadzu ScientificInstrumental Co., Kyoto, Japan). The gas chromatograph used was a Hitachi 163 equipped h t h a flame ionization detector (FID).A 150 ft X 0.010 in. i.d. stainless steel capillary column coated with Silicone DC-550 (Perkin-Elmer Co.) was used. Column, injector, and detector temperature were 160,240, and 240 "6,respectively. Chromatograms, peak areas, and retention times were obtained by using a data processor (CHROMATOPAC C-RlA, Shimadzu Scientific Instrumental CO.). Reagents and Samples. Hitachi Gel 3053 was used as a stationary phase, which is a porous spherical octadecylsilanized silica gel of average diameter 5 wm. LAS was obtained from our company and the single alkyl chain lengths prepared previously (13)were used for the identification of the elution peaks. Other reagents were of analytical reagent grade. Procedure. The column, 4.6 mm i.d. X 150 mm long stainless steel, was prepared by the slurry packing procedure described previously (12). The mobile phase used was an aqueous acetonitrile solution containing sodium perchlorate. The column effluent was monitored at 225 nm and the flow rate was maintained at 1.0 mL/min. Desulfonation of LAS for the GC analysis was carried out by a modification of the acid-decomposition method described by Knight and House ( I ) . About 100 mg of LAS, 2 g of tin powder, 1 g of carborumdum boiling chips, and 30 mL of anhydrous phosphoric acid prepared from phosphorus pentoxide and 85% phosphoric acid are placed in a 100-mL flask equipped with the steam distillation trap. The mixture is heated and the temperature of the mixture is maintained at 215 OC for 60 min. After the reaction, the water trap is filled with water and steam distillation is achieved for 10 min. Then the alkylbenzene and water in the trap are quantitatively transferred to the separatory funnel and the alkylbenzene is extracted with n-hexane. LAS in detergenk was analyzed as follows. To 1g of detergent, 100 mL of methanol was added, and the mixture was vigorously stirred for 10 min. Surfactants are soluble in methanol, but inorganic ingredients are insoluble. Ten microliters of the supernatant methanol was injected into the HPLC. The average molecular weights of LAS were calculated by the
wmo
c11
I;@
5- 6/a
clo
c12
I
0
5
I
10
15 min
Figure 1. Typical chromatogram of linear alkylbenzenesulfonatest. Conditions: column, Hitachi Gel 3053 5 pm, 4.6 mm 1.d. X 150 mmi; eluent, 0.1 M sodium perchlorate in acetonibile/water (4555);flow rater, 1.0mL/mln; pressure, 100 kg/cm2; column temperature, 40 O C ; detector, UV 225 nm, 0.16 AUFS; injection volume, 5 pL: sample concentration, 0. i%. 4 is the position of the phenyl group from thrs termlnal methyl group on the chaln.
summation of the multification of the molecular weight of the individual chain length LAS by the mole fraction (peak area percent x W2).
RESULTS AND DISCUSSION Effect of Mobile Phase Composition on the Separation of LAS. Excellent separations of the phenyl isomers of LASS compared with those described previously (12) were obtained by using Hitachi Gel 3053 as a stationary phase and aqueous acetonitrile solution containing sodium perchlorate as a mobile phase. The eluted peaks were identified by using authentic samples and the elution order was the same as that described previously (12). The eodium perchlorate concentration wao held constant a t 0.1 M and the effects of the water content on the separation were investigated. The capacity factors of LAS increased with increasing the water content of the mobile! phase. For the separation of LAS with isocratic condition, 55% (v/v) of the water content was recommended because! the longer separation times were required with the higher water content and the insufficient separations were obtained by using the lower water content. In the absence of sodium perchlorate in the mobile phase (CH&N/H20 (45:55)), LAS was eluted at the solvent front. By addition of sodium perchlorate to the mobile phase, however, LAS was well separated as shown in Figure 1. When the sodium perchlorate concentration of the mobile phase increased from 0.05 to 0.5 M, the capacity factor of 2-phenyl CI3LAS also increased from 5.4 to 19.9. But the separations of the phenyl isomers were not improved. Therefore, 0.1 M sodium perchlorate was chosen for the rapid separation of LAS. If aqueous acetonitrile solution containing other inorganic salts such as sodium chloride, sodium nitrate, ammonium chloride, etc. instead of sodium perchlorate will be used as a mobile phase, similar separations of LAS will be obtained. In this paper, sodium perchlorate is preferred for the separation to other inorganic salts, because aqueous acetonitrile
1820
ANALYTICAL CHEMISTRY, VOL. 53, NO. 12, OCTOBER 1981 2pr 2.0
1.5
. alkyl chain length and isomer distribution
-
CIO 5-phenyl 4-phenyl 3-phenyl 2-phenyl
1
5
--
Table I. Determination of Alkyl Chain Distribution of Alkylbenzenesulfonatea
1.0
.
0.5
11
12
13
Alkyl chain length
Relatlonshlps between the capacity factor of alkylbenzenesulfonates and their alkyl chain length. Chromatographic conditions are as in Figure 1. Flgure 2.
solution containing sodium perchlorate does not absorb light of UV region and does not corrode the stainless steel pipings of the liquid chromatograph. Effect of Column Temperature on the Separation. The capacity factors of LAS decreased with increasing the column temperature. The logarithms of the capacity factors of the each component of LAS were linear with the reciprocal of the column temperature. The results suggest that the column temperature control is necessary in order to achieve reproducible analysis. Relationship between t h e Capacity Factor and the Alkyl Chain Lengths. The logarithms of the capacity factors of 2-phenyl LAS were directly proportional to the alkyl chain lengths, and the similar linear relationships for the other phenyl isomers of LAS were found as shown in Figure 2. These results might be considered that LAS is a mixture of homologous series such as 2-phenyl LAS, 3-phenyl LAS, and so on. Better separations of the phenyl isomers of LAS by these chromatographic conditions compared with those described previously (12)would result in an increase of the selectivity among the each phenyl homologous series. We indicated previously (13)that in liquid-liquid partition chromatography, the two linear relationships are set up between the logarithm of the capacity factor and the carbon number (alkyl chain lengths) of a homologous organic solute and between the reciprocal of the column temperature and the logarithm of the capacity factor. The chromatographic results obtained by the separation of LAS were explainable in terms of the theory of the partition chromatography to satisfy the above two relations. The substituted position of the phenyl group and the alkyl chain lengths of LAS have a significant effect on the separation compared with the sulfonic group. But the sulfonic group also has a significant effect on the LAS retention and the effect appears to be constant. The contribution of the sulfonic group to the retention can be expected to be of similar magnitude to other hydrophilic groups of surfactants such as the sulfate group and the quaternary ammonium group since the chromatographic results are obtained by using porous poly(styrene-divinylbenzene) gel as a stationary phase (14-16). Therefore, these chromatographic systems might be applicable to the separation of homologous series of other surfactants. Determination of Alkyl Chain Distribution of LAS. The separations of the phenyl isomers of LAS by HPLC were incomplete compared with those by capillary-GC analysis of desulfonated LAS. However, the results agreed very closely with those by the capillary-GC analysis as shown in Table I. The detector response of HPLC and GC is a very important factor for the determination of the alkyl chain distribution and the average molecular weight of LAS. By HPLC with
HPLC 8.1
c,
2'3 1.9 3.9 38.6
c,,
5.6 8.5 12.5 32.0
I
10
sample A
6-phenyl 5-phenyl 4-phenyl 3-phenyl 2-phenyl
..
6-phenyl 5-phenyl 4-phenyl 3-phenyl 2-phenyl CI, 6- to 7-phenyl 5-phenyl 4-phenyl 3-phenyl 2-phenyl
c,4
6- to 7-phenyl 5-phenyl 4-phenyl 3-phenyl 2-phenyl
av mol wt as sodium alkylbenzenesulfonate
GCb
sample B HPLC
GCb
8.4 11.1 1.2 1.4 1.9 0.4 3.9 2.7 40.0 31.2
10.5 3.2 2.5 2.3 2.5 32.9 4.5
7.2 1.3 7.1 8.2 5.5 5.7 13.8 5.3 6.0 31.5 28.7 29.2 4.6 7.4 10.0 5.0 15.0 7.7 5.5 4.8 5.8 5.5 5.9 7.2 4.0 4.5 10.6 9.9 3.9 4.1 21.3 20.1 20.0 19.5 4.1 3.9 9.4 6.8 3.9 3.2 2.6 4.5 3.3 3.3 3.2 3.2 3.8 3.6 2.4 2.6 6.2 6.1 2.4 2.4 0 0 9.0 7.9 3.9 3.5 2.0 1.7 1.1 1.1 0.9 0.9 1.1 0.7 343.8 343.3 346.3 345.9
Presented as peak area %. benzene was determined.
Desulfonated alkyl-
UV detector, the peak area ratio of LAS corresponds to the molar ratio of each component of LAS (131,because the molar absorptivity of each component of LAS is identical (our experiment, Cl0-, Cll-, C12-, C13-, C14-, and C15-LAS in methanol, ,A, 225 nm, E,, 1.25 X lo4L.mol-l.crn-l, respectively) (17). On the other hand, GC results using a FID reported earlier was the only comparison of the original alkylbenzene and the volatile derivatives of its sulfonate (2,4). In the case of the GC analysis of long chain fatty acid methyl ester, Ettre and Kabot found a very close correction between weight percent concentration and peak area percent for methyl ester (18)and Sheppard et al. indicated that the correction factors were needed to obtain accurate results (19). As shown in the above examples, the FID response is unclear and those of the raw alkylbenzene of LAS have not been reported. Therefore, it is necessary to correct the FID response in order to determine the average molecular weight of LAS. However, it is recognizable that the peak area ratio obtained by the capillary-GC analysis of desulfonated alkylbenzene corresponds to the molar ratio of each component of LAS, because good agreement between the HPLC and the capillary-GC results was obtained as shown in Table I. Application. Branched alkylbenzenesulfonates (ABS), which have until recently almost never been used in Japan, were separated by this chromatographic conditions. The chromatogram obtained is very complicated as shown in Figure 3 and the elution peaks have not been identified due to the existence of numerous isomers. However, LAS and ABS are easily distinguished by the comparison of their chromatograms. The proposed HPLC method was applied to detergent analysis. LAS in several commercial detergents, as well as the standard LAS, was separated without any interferences
ANALYTICAL CHEMISTRY, VOL. 53, NO. 12, OCTOBER 1981
1821
Table 11. Determination of Alkyl Chain Distribution of Alkylbenzenesulfonates in Several Commercial Detergents a alkyl chain length and isomer distribution
c
10
4- to 5-phenyl 3-phenyl 2-phenyl Cll 5- to 6-phenyl 4-phenyl 3-phenyl 2-phenyl
Cl,
5- to 6-phenyl 4-phenyl 3-phenyl 2-phenyl
c,,
6- to 7-phenyl 5-phenyl 4-phenyl 3-phenyl 2-phenyl av mol wt as sodium a1kylbenzenesulfonate
A
12.8 9.8 3.0 32.6 20.0
B
C
D
17.5 11.1 7.6 6.1 8.3 0.6 4.8 3.6 6.6 2.8 24.0 50.6 34.5 12.4 23.4 22.8 1.0 0.9 1.7 5.9 10.4 4.7 6.3 5.7 15.1 6.0 5.4 3Q.7 54.6 31.9 34.3 15.6 22.3 12.0 18.1 6.1 10.2 6.0 6.6 4.4 9.5 4.8 5.9 12.6 4.6 8.0 4.8 23.9 9.6 0 20.1 3.0 13.0 11.9 1.3 2.4 0.8 1.5 3.8 3.2 2.9 1.3 2.4 2.9 1.4 1.8
343.7 Presented as peak area 7%.
11.8
343.2
336.5
343.4
Table 111. Determination of Alkyl Chain Distribution of Alkylbenzenesulfonates in a Commercial Detergent a alkyl chain length and isomer distribution CO 4- to 5-phenyl
3-phenyl 2-phenyl
c,,
5- to 6-phenyl 4-phenyl 3-phenyl 2-phenyl
c,,
5- to 6-phenyl 4-phenyl 3-phenyl 2-phenyl
CI,
6- to 7-phenyl 5-phenyl 4-phenyl 3-phenyl 2-phenyl
Cl4
6- to 7-phenyl 5-phenyl 4-phenyl 3-phenyl 2-phenyl av mol wt as sodium alkylbenzenesulfonate
run 1
run 2
run 3
mean
12.5 7.7 0.5 4.3 35.2 18.5 0.7 6.8 9.2 27.1 11.4 4.9 4.6 6.2 16.7 7.5 0.6 2.6 2.5 3.5 8.5 3.0
12.7 12.6 12.6 7.8 7.8 7.8 0.5 0.5 0.5 4.4 4.3 4.3 34.6 34.7 34.8 18.1 18.1 18.2 0.7 0.7 0.7 6.5 6.6 6.6 9.3 9.3 9.3 27.5 27.2 27.3 11.4 11.4 11.4 5.1 5.0 5.0 5.0 4.7 4.8 6.0 6.1 6.1 16.3 16.7 16.6 7.4 7.6 7.5 0.6 0.5 0.6 2.4 2.6 2.5 2.4 2.5 2.5 3.5 3.5 3.5 8.9 8.8 8.7 3.0 3.0 3.0 1.6 1.6 1.4 1.5 1.2 1.4 1.1 1.2 1.2 1.3 1.3 1.3 1.5 1.6 2.0 1.7 344.8 344.9 344.9 344.9
Presented as peak area 7%. since other surfactants used do not absorb light at 225 nm. The results obtained are shown in Table 11. The average molecular weight of LAS was also determined precisely as shown in Table 111.
I
I
0
5
I
10 min
Flgure 3. Chromatogram of branched alkylbenzenesunonates. ditions are as In Figure i.
Con-
On the other hand, the selective quantitative analysis of
LAS in detergents, of course, could be carried out by the proposed HPLC method by means of making the calibration curve for LAS and also the composition of LAS in the polluted environment could be determined in detail.
ACKNOWLEDGMENT The authors wish to acknowledge the technical assistance of Hideko Ueno. LITERATURE CITED (1) Knight, J. D.; House, Rr. J. Am. 011 Chem. Sm. 1859, 36, 195-200. (2) Setzkorn, E. A.; Carel, A. B. J. Am. Oil Chem. Soc. 1083, 40, 57-59. (3) Lee, S.; Puttnam, N. A. J. Am. OIIChem. Soc. 1887, 44, 158-159. (4) Lew. H. Y. J. Am. 011Chem. Soc. 1972, 49, 665-669. ( 5 ) Nishl, S. Bunsekl Kagaku 1885, 74, 917-920. (6) Watanaba. S.; Nuklyama, M.; Takagl, F.; Ilda, K.; Kalse. T.; Wada. Y. Shokuhln Elseigaku Zasshl1875, 16. 212-217. (7) Imalda, M.; Sumlmoto, T.; Yada, M.; Yoshlde, M.; Koyama, K.; Kunita, N. Shokuhln Elselgaku Zasshll875, 76, 218-224. ( 8 ) Matsutanl, S.; Shige, T.; Nagal, T. Yukagaku 1878, 28, 847-851. (9) Llddlcoet, T. H.; Smlthson, L. H. J. Am. 011 Chem. Soc. 1885, 42, 1097-1 102. (IO) Lew, H. L. J . Am. 011Chem. Soc. 1887, 44, 359-368. (1 1) Denig, R. Tenside Oeterg. 1873. 70, 59-63. (12) Nakae, A.; Tsujl, K.; Yamanaka, M. Anal. Chem. 1980, 52, 2275-2277.. (13) Nakae, A.; Muto, G. J . Chromatogr. 1876, 720, 47-54. (14) Nakae. A.; Kunlhlro, K. J. Chromatogr. 1878, 752, 137-144. (15) Nakae, A.; Kunlhko, K.; Muto, G. J. Chromatogr. 1977, 734, 459-466. (16) Nakae, A.; Kunhlro, K. J. Chromatcgr. 1878, 756, 167-172. (17) Weber, W. J.; Morris, J. C.; Stumm, W. Anal. Chem. 1982, 34, 1844-1845. (18) Erne, L. S.; Kabot, F. J. J. Chfomafogr. 1883, 1 7 , 114-116. (19) Shepprd, A. J.; Meeks, S. A.; Ellbtt, L. W. J. Gas Ckomatogr. 180, 6 , 28-38.
RECEIVED for review November 19,1980. Accepted June 19, 1981.