Liquid chromatographic separation of alkane sulfonate and alkyl

Alkyl Sulfate Surfactants: Effect of Ionic Strength. Daan Zhou and Donald J. Pietrzyk*. University of Iowa, Chemistry Department, Iowa City, Iowa 5224...
0 downloads 0 Views 705KB Size
Anal. Chem. 1992, 64, 1003-1008

(5) Bradley, J.; Stkkleln, W.; Schmld, R. D. Roc. Confr. Qual. 1991, 1 , 157-183. (8) Tljssen, P. Laboratory Techniques in Biochemktry and Molecular Biology; Elsevler: Amsterdam, 1985; Chapter 1. (7) Chard, T. An Infroductfon to RadEolmmunoassey and Rehted Techniques; Elsevler: Amsterdam, 1987. (8) Oellerlch, M. Methods of Enzymatic Analysis, 3th ed.;Verlag Chemle: Welnheim, 1983; Chapter 2.7. (9) Hemmilae, 1. Clln. Chem. 1985, 31, 359-370. (IO) Rubenstein, K. E.; Schnelder, R. S.; Ullman, E. F. Biochem. Res. Commun. 1972. 47, 846-851. (11) Mane, M.; Cleudel, J. P.; Clret, P. Clln. Chem. 1987, 33, 209-213. (12) Collet-Cassart. D.; Limet, J. N.; Van Krleken, L.; De Hertogh, R. Clln. Chem. 1080, 35, 141-143. (13) Kooyman, R. P. H.; Lenferlnk, A. T. M.; Eenlnk, R. G.; Greve, J. Anal. Chem. 1091, 63. 83-85. (14) Lukosr, W. Fresenius’ J. Anal. C h m . 1990, 337, 24-25. (15) Boever, J.; Kohen, F.; Bouve, J.; Leyseele, D.; Vandekerckhove, D. Clln. Chem. 1900, 36, 2036-2041. (16) . . Brbht. F. V.: Betts. T. A.; Lltwller. S. Anal. Chem. 1990, 6 2 , 10&5-1069. (17) Fagerstam, L. I n Technlques In Protein Chemistry II; Vlllafranca, J. J., Ed: Academlc Press: New York, in press. (18) Janata, J.; Blackburn, G. F. Ann. N.Y. Acad. Scl. 1984, 428, 286-292. (19) Borman, S. Anal. Chem. 1987, 59. 1161-1164. (20) Davls, K. A. Anal. Chem. 1989, 61, 1227-1230. (21) Muramatsu, H.; Dlcks, J. M.; Tamiya, E.; Karube, I. Anal. Chem. 1987, 59. 2760-2763. (22) Wohltjen, H.; Dessy, R. Anal. Chem. 1979, 51, 1458-1464. (23) Wohltjen, H.; Dessy, R. Anal. Chem. 1979, 51, 1465-1474. (24) Kurosawa, S.; Tawara, E.; Kamo, N.; Ohta, F.; Hosokawa, T. Chem. hem. Bun. 1990, 38, i i i 7 - i i 2 0 .

IO03

(25) Eddowes, M. J. Anal. Roc. 1989, 26, 152-154. (26) Kooyman, R. P. H.; Kolkman, H.; Van Gent, J.; Greve, J. Anal. Chlm. Acta 1988, 213, 35-45. (27) Bergveld, P. Blosensws Bldectron. 1991, 6 , 55-72. (28) Newman, A. L. Patent WO 88/09499. 1988. (29) Stanbro, W. D. Patent WO 88/08528, 1988. (30) Newman, A. L. Patent WO 90/02792, 1990. (31) Peareon. R. K. AIChESymp. Ser. 1985. 267. 42-45. (32) Betalllard, P.; Gardles, F.; Renauld, N. J.; Martelet, C.; Colln, B.; Mandrand, B. Anal. Chem. 1988, 60, 2374-2379. (33) Gardles, F.; Martelet, C. Sens. Acfuators 1989, 17, 461-464. (34) Valdes, J.; Wall, J. G.; Chambers, J. P.; Eldefrawi, M. E. John Hqpklns APL Tech. Dig. 1988, 9 , 4-10. (35) Newman, A. L.; Hunter, K. W.; Stanbro, W. D. Roc. Int. Meet. Chem. Sens., 2nd 1988, 596-598. (36) Keli. D. B.; Davey, C. L. Bksensors, a pracfkal approach; IRL Press: Oxford University, Oxford, U.K.. 1990; Chapter 5. (37) , . Dbale. J. W.: Downie. T. C.: Gouldlnn. C. W. Chem. Rev. 1989. 69, 36g-405. (38) Vermilyea, D. A. Acta Metall. 1953, 1 , 282-294. (39) Pethlg, R.; Kell, D. B. phvs. M .Bid. 1987, 32, 933-970. (40) Sarma. V. R.; Sliverton, E. W.; Vavies. D. R.; Terry, W. D. J. Blol. Chem. 1071. 246. 3753-3759. (41) Schwan, H. P.yFehs, C. D. Rev. Sci. Inst”. 1988, 39, 481-489. (42) Gobrecht, H. Lehrbuch der €xperimenfa@ysk, 7th ed.; Walter de Gruyter: Berlin, New York, 1978; Vol. 111, Chapter 3. (43) Weetall, H. H. I n Methods of Enzymology; Mosbach. K., Ed.; Academic Press: New York. 1976; Voi 44, Chapter 10. (44) Sonnenfeld, R.; Hansma, P. K. Science 1988, 232, 211-213.

RECEIVED for review September 27,1991. Accepted January 27, 1992.

Liquid Chromatographic Separation of Alkanesulfonate and Alkyl Sulfate Surfactants: Effect of Ionic Strength Daan Zhou and Donald J. Pietrzyk* University of Iowa, Chemistry Department, Iowa City, Iowa 52242

The retention of alkanesulfonate and alkyl sulfate surfactants, which was determined on a reversed stationary phase as a function of mobile-phase ionic strength, Is consistent with a double-layer type interaction at the stationary-phase surface. Increasing the mobilephase lonlc strength not only increases retention but also Improves resolution because peak widths are dgnlficantly reduced. The type of cation provided by the ionk strength salt also enhances retention, reduces peak width, and improves rewlutlon. Lithium hydroxide is an ideal electrolyte for the separation of mutticomponent mixtures of aikanesulfonate and alkyl sulfate surfactants. When the c d m n effluent is passed through a postcolumn anion micromembrane suppressor, the conductivity due to the electrolyte is minimized and conductivity detection is sendive, yielding a detection limit of about 0.3 nmoi of injected anaiyte for a 3:l signaknoise ratio. Multicomponent alkanesulfonate and alkyl sulfate mixtures from C2to C,8 are baseline resolved by udng a moblleghase gradient whereby CH&N concentration increases and LlOH concentration decreases.

INTRODUCTION Typical anionic surfactants will contain hydrophilic (anionic) and hydrophobic centers, and both contribute to the surfactants physicochemical properties. Major classes of anionic surfactants are the long-chain alkanesulfonates (ASO,-), alkyl sulfates (AOS0,-), and the linear alkyl-

benzenesulfonates(LAS). All three are water soluble and are used widely in commercial products and processes where detergent action is required. The need to accurately determine their presence in these applications and in the environment, particularly at trace levels, is of growing concern. The development of analytical methodology for the determination of anionic surfactants has focused on the properties of either the hydrophobic or the hydrophilic portions of the surfactants. In general, classical procedures, such as precipitation, color formation, and color quenching involve a reaction between the anionic center and another reagent containing a positive center to produce a neutral product of low dis~ociation.’-~These methods, however, do not discriminate among individual compounds within homologous groups of anionic surfactants and, consequently, their main application is generally toward group determinations. Classieal separation strategies, such as planar methods, ion exchange, and solvent extraction, which often require reactions a t the anionic center, are also primarily suited to group separations. In contrast, modern, efficient separation methodologies including gas Chromatography (GC),9*4-7 high-performance liquid chromatography (HPLC),&’*and capillary electrophore~is~~ are strategies which have been successfully used to discriminate between members within the homologous or isomeric groups of AS03-, AOS03-, and LAS surfactanta. The hydrophobic center, the anionic center, or both are key surfactant structural features which influence retention and resolution. In GC anionic surfactant volatility is often a limiting factor and has been overcome through desulfonation procedures4or

0003-2700/92/0364-1003$03.00/00 1992 American Chemical Society

1004

ANALYTICAL CHEMISTRY, VOL. 64, NO. 9, MAY 1, 1992

conversion to volatile derivatives." Both strategies have been successfully applied to mixtures of AS03-, AOS03-, and LAS surfactants including environmental samples. Characterization and in some cases detection are enhanced by interfacing the GC separation to mass s p e ~ t r o m e t r y . ~ * ~ , ~ In LC of anionic surfactants detection is often a limiting factor particularly for the ASO,- and AOSO, surfactants. For LAS surfactants UV and fluorescence detection will provide favorable detection limit~.~@'lWhile AS03- and AOS03surfactants can be detected at low UV wavelength, sensitivity is modest because detection at a low wavelength is subject to interference from other low UV absorbing materials. More suitable detection limits are obtained by postcolumn color formation reactions12or by a precolumn conversion of the surfactant into a detector-active derivative.13 Suppressed cond~ctivity'~ and indirect d e t e ~ t i o n ' ~can J ~ also provide favorable detection limits. The LC separation can be interfaced to mass spectrometry and used to characterize and determine ASO, and AOSO, components in multicomponent mixtures. 17,18 The AS03- and AOS03- surfactants have been separated on reversed stationary phases because of hydrophobic interactions between the surfactant and the stationary phase,3,8,g,11-13,15,'s on anion exchangers due to anion-exchange processes,zoand on reversed stationary phases in combination with mobile phases containing ion interaction reagents of opposite ~ h a r g e . ~ J ~ For J ~ Jseparation ~J~ of multicomponent mixtures mobile-phase gradients are required. Retention of AS03- and AOS03- surfactants on reversed stationary phases increases as surfactant hydrophobicity inthat retention, selectivity, creases. It is also and peak shape for benzenesulfonate derivatives are markedly improved if a high ionic strength mobile phase is used. Retention of these sulfonates is accompanied by double-layer formation at the stationary-phase surface. The sulfonate analyte constitutes the primary layer due to an interaction between the analyte hydrophobic center and the stationaryphase surface, and the diffuse secondary layer is made up of countercations. As mobile-phase ionic strength increases, interactions favoring double-layer formation are enhanced whether using an organic polymeric or bonded silica reversed stationary phase. A detailed discussion of this phenomenon is reported elsewheren*% This report summarizes OUT studies on (a) the effect of mobile-phase ionic strength on retention of AS03- and AOS03- surfactants on reversed stationary phases and (b) the optimization of the mobile-phase conditions for the separation and conductivity detection of multicomponent mixtures of AS03- and AOS03- surfactants. Mobile-phase ionic strength is high to take advantage of ionic strength effects on enhanced retention, selectivity, and resolution of the ASOf and AOS03- surfactants. Favorable detection limits are obtained by employing a postcolumn membrane suppressor, which is often used in ion chromatography,2o to minimize the effect of mobile-phase ionic strength on conductivity detection.

EXPERIMENTAL SECTION Reagents and Instrumentation. Alkanesulfonates and alkyl sulfates were purchased from Eastman Kodak Co. and Chem Service Co. Inorganic reagents were analytical reagent grade, and organic solvents were LC grade when possible and used as received. LC water was obtained by passing laboratory distilled water through a Milli-Q-Pluswater treatment system. All stationary phases were purchased as prepacked columns. Zorbax ODS was obtained as a 6-pm, 4.6-mm X 150-mm column from Mac Mod Analytical, Inc., and PRP-1, a poly(styrene-divinylbenzene) copolymeric stationary phase, was obtained as a lO-pm, 4.1-mm x 150-mm column from Hamilton Co. The instrumentation consisted of Beckman M-11OA pumps and controller or a Spectra Physics M-8800 gradient pump, a Rheodyne 7125 injector, a

Dionex AMMS-1 anion micromembrane suppressor, a Waters M-430 conductivity detector, and a Spectra Physics M-4270 integrator controlled by Spectra Physics Autolab Software. Procedures. Standard analyte solutions and mixtures were prepared by dissolving known quantities in LC water or 2:3 CH3CNH20. Mobile-phase solutions were prepared by dilution of a known volume of standard aqueous salt, acid, or base solution and organic solvent with LC water. Solvent composition is expressed as percent by volume. All mobile phases were degassed before use. Aqueous 25 mM H2S04solution at 1.0 mL/min was used to regenerate the anion membrane suppressor. PRP-1 and Zorbax ODS column performance was verified periodically with a phenol, benzene test sample and a 9 1 CH3CN:H2Oor a 85:15 CH3CN:H20mobile phase. An average of at least four measurements were made to establish each data point on the calibration curves (peak area versus analyte amount injected). Statistical analysis was completed with Slide Write software. Sample aliquots were injected (usually 10 pL) with a Hamilton 702 25-pL syringe into a 5O-pL sample loop in analyte quantities of 0.5-6 nmol. The flow rate was 1.0 mL/min, temperature was 25 "C, inlet pressure was 500-900 psi, and column void volume was 1.0-1.2 mL depending on the column and mobile phase.

RESULTS AND DISCUSSION Initial studies demonstrated that the retention of alkanesulfonate, AS03-, and alkyl sulfate, AOSO,-, surfactants on PRP-1 and Zorbax ODs, was consistent with typical reversed-phase type interactions in that retention of the surfactants increases as the alkyl chain length increases in the AS03- and AOS03- homologous series and decreases as CH3CNin the CH3CN:H20mobile phase increases. Previous studies indicate that the bonded phase silica often provides more favorable chromatographic peak shapes than those obtained with the PRP-1 stationary phase and is often preferred unless there is a specific factor in the separation which prevents its use. For example, the separation may require a basic (pH > 10) mobile phase. In a head to head comparison of the two columns retention for a given surfactant from a CH3CN:H20mobile phase is higher on PRP-1. For example, C8S03-retention is larger by about 12% on PRP-1, and as the AS03- and AOS03- alkyl chain length increases, the percent retention increase on PRP-1 becomes larger. Resolution, however, is better on Zorbax ODS because column efficiency (an increase by about 2 or more depending on the analbeing compared) and selectivity (about 6% or greater) are more favorable on this column. While differences in particle size and amount of stationary phase per column are contributing factors, the major reason for the difference between the two columns is that maas transfer is more favorable on the ODS column. These initial studies also demonstrated two other properties of surfactant retention. First, the AOSO, surfactant is more highly retained than the AS03- surfactant of the same alkyl chain length on both PRP-1 and Zorbax ODs. Second, the separation of multicomponent mixtures of AS03- and AOSO, was inadequate (see Figure 5C) when only a CH3CN:H20mobile-phase gradient was used. For simpler surfactant mixtures resolution is still not markedly improved depending on the AS03- and AS04- mixture even though the required CH3CN:H20gradient range is less. When the mobile phase contains electrolyte, AS03- and AOS03- retention on the stationary phases increases as electrolyte concentration (ionic strength) and surfactant alkyl chain length increases. These trends are illustrated in Table I where retention data for several AS03- and AOS03- surfactants on PRP-1 are listed as a function of mobile-phase NaOAc concentration. Similar ionic strength effects were determined when NaC1, sodium benzoate, or LiOH was the mobile-phase electrolyte. Enhanced retention was the same whether NaOAc or sodium benzoate was used and differed only slightly from NaCl because of the small effect of ionization on ionic strength. Although not studied extensively,

ANALYTICAL CHEMISTRY, VOL. 64, NO. 9, MAY 1, 1992 Oe8 0.8

0.6

1005

1.4

1

1.2

1

1.0

/6c

0.8

Vk' 0.4 0.8

0.4 0.2 0.2

0.0

n, 0.0

0

10

20

30

40

10

0

20

30

40

iuonic strrngtn, M Flguo 1. Effect of mobile-phase ionic strength on retention on PRP-1 (A) and Zorbax ODS (B): a 1:9 CH,CN:H20, NaCl (A) and a 5:95 CH,CN:H20, NaCl mobile phase.

Table I. Effect of Mobile-Phase Electrolyte (NaOAc) Concentration"on Retention of Alkanesulfonate Analytes analyte

0M

CzSO< C4SO< C5SO< C6S03-

0.17 0.26 0.42 0.72 1.17 1.98

c7so3CsSO
Ba2+ > TMAB+ = Mg2+> Na+ > Li+. These trends, plus the observation that AS03- and AOS03retention increases significantly as ionic strength increases, suggest that an appreciable association is taking place between the cation and the AS03- and AOS03- analytes. The association reduces the charge at the anionic center in the surfactant and hence increases the hydrophobic interaction between the analyte and the stationary phase. When adjacent analyte peak shapes were compared, efficiency, resolution, and selectivity were also cation dependent, as shown in Table 111. The enhancement in resolution follow the order Mg2+> Ba2+ > A13+> Li+. Resolution enhancement produced by other monovalent cations (see Table 111) was less but not markedly different from Li+. The improved resolution due to the cation is the result of a decrease in peak width or improved column

1006

'

ANALYTICAL CHEMISTRY, VOL. 64, NO. 9, MAY 1, 1992 40

Table 111. Effect of the Cation on Peak Shape salt

efficiency" resolution selectivity C8S03- C l o S O ~ c1Oso~-/c~so3~ Cl0SO3-/C8S03

LiCl 4200 NaCl 2000 NH,Cl 1600 7600 MgClz (CH3)4NCl 1200 BaCl, 5100 Al(NO3)3 3700 HCl 2300

lo

1 0

2

4

6

1 I

8

1

3.86 3.84 3.95 3.84 3.92 4.60 4.00 4.03

the best detection limits, but not the largest enhancement in retention (seeFigure 2). From a practical view improving peak shape is more of an advantage and achieving this effect is a more useful goal in establishing optimum conditions for separation of multicomponent mixtures. However, since the simpler detection options for AS03- and AOS03- are limited primarily to conductance, the best detection limit is not realized because of the high background conductance due to the mobile-phase ionic strength. Using a postcolumn anion membrane suppressor will not solve the problem since a HCl background is produced by the suppressor when MgC1, is used. A Mg2+salt of a weak acid would be suitable, and Mg(OAc),, which is water soluble, was evaluated as the ionic strength salt. For separations where isocratic elution is satisfactory, favorable resolution, narrow peaks, and low detection limits are obtained. The detection limit can be lowered further by using the postcolumn anion membrane suppressor. When a mobile-phase gradient, where Mg(OAc), decreases and CH3CN increases, is used, which is required to separate a multicomponent AS03-/AOS03- mixture, a background conductance drift becomes a limiting factor in the detection limit for the Mg(OAc), mobile phase even with the anion membrane suppressor. The effect of the solvent gradient on background conductance is shown in Figure 4. When the solvent gradient contains Mg(OAc),, the suppressor converts the Mg(OAc), into HOAc, and because of the large solvent change during the gradient, dissociation of HOAc is altered. This produces a

0

Hydrated Ionic Radlus, AngrlromlCharpr

Correlation between retention and mobile-phase cation hydrated ionic radius to charge ratio.

Flgure 2.

efficiency rather than to an altered selectivity. No reversals in the selectivity were found. The improved efficiency, as listed in Table 111, follows the same cation order or Mg2+> Ba2+> A13+> Li+. When mobile-phase CH3CN concentration is increased, retention of AS03- and AOSO, decreases. This is illustrated in Figure 3A, where retention of several alkanesulfonates on PRP-1is shown as a function of the CH3CNH20ratio in the presence of Mg(OAc),. Mg(OAc),, rather than MgC12,was used as the ionic strength salt for reasons outlined later. If Mg(OAc), concentration is increased at any given CH3CNH,0 ratio, retention of the surfactants increases. Figure 3B shows the magnitude of the enhanced retention for several alkanesulfonates as Mg(OAcI2is increased in a 2080 CH3CN:H20 solvent. A MgC1, mobile phase, according to Tables 11and 111, yields the best resolution, the narrowest peaks, and consequently, 25

i

c2 c4 0 C8 C8 0 c10

I

20

z

4.83 4.68 4.69 9.11 5.04 7.69 4.80 4.33

Same conditions as Table 11; efficiency is plates per meter.

I

\

4900 2100 1400 8600 1300 5300 4500 2600

15

0

*

f'

0

a

L

5

0 0

10

20

30

40

0.1

1

10

Magnesium Acetate, m M Percent Acetonitrile Flgure 3. Effect of mobilephase CH,CN and Mg(OAc), concentration on retention of several aikanesuifonates: a CH,CN:H,O, (A) and a 1:4 CH,CN:H,O, Mg(OAc), (B) mobile phase and a PRP-1 column.

0.10 mM Mg(OAc),

ANALYTICAL CHEMISTRY, VOL. 64,NO. 9, MAY 1, 1992 Bo

8

analyte to the protonated form. Although not shown, step gradient separations of simpler mixtures of sulfonates and sulfates are possible using the LiOH, CH3CN mobile phase. For example, C3 to Ca and Ca to c16 linear chain sulfonate mixtures were readily separated. For more complex multicomponent separations, where alkyl chain length covers a wider range, a successful separation requires a gradient. Several types of gradients are possible. One is where the solvent composition is changed at constant LiOH Concentration, while the second is the reverse. Changing both, that is CH3CN concentration increases and LiOH concentration decreases, was shown to be optimum, particularly when multicomponent mixturea that vary widely in alkyl chain length were separated. Examples of this kind of gradient separation of alkanesulfonates from C2to c16 and alkyl sulfates from c6 to C16are shown in Figure 5A,B, respectively. The electrolyte enhances retention, improves the selectivity, and reduces the peak width; if the electrolyte is omitted, resolution (see Figure 5C) is reduced sharply. Decreasing the electrolyte concentration during the gradient reduces retention of the more highly retained analytes while the favorable effects of the electrolyte on peak shape are still maintained. If the mixture contains both AS03-and AOSO,, excellent resolution is still possible (see Figure 6). Resolution of the Cz to C4 analytes can be improved by increasing elution time prior to initiating the gradient. Also, if the lower alkyl chain length surfactants (