Determination of fatty alcohol ethoxylates and alkylether sulfates by anionic exchange separation, derivatization with a cyclic anhydride and liquid chromatography

Journal of Chromatography A, 1218 (2011) 8511–8518

Contents lists available at SciVerse ScienceDirect

Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma

Determination of fatty alcohol ethoxylates and alkylether sulfates by anionic exchange separation, derivatization with a cyclic anhydride and liquid chromatography M. Beneito-Cambra, L. Ripoll-Seguer, J.M. Herrero-Martínez, E.F. Simó-Alfonso, G. Ramis-Ramos ∗ Departament de Química Analítica, Facultat de Química, Universitat de València, Dr. Moliner 50, 46100 Burjassot, Spain

a r t i c l e

i n f o

Article history: Received 14 June 2011 Received in revised form 21 September 2011 Accepted 22 September 2011 Available online 29 September 2011 Keywords: Alkylether sulfates Diphenic anhydride Esterification Fatty alcohol ethoxylates Phthalic anhydride Transesterification

a b s t r a c t A method for the separation, characterization and determination of fatty alcohol ethoxylates (FAE) and alkylether sulfates (AES) in industrial and environmental samples is described. Separation of the two surfactant classes was achieved in a 50:50 methanol–water medium by retaining AES on a strong anionic exchanger (SAX) whereas most FAE were eluted. After washing the SAX cartridges to remove cations, the residual hydrophobic FAE were eluted by increasing methanol to 80%. Finally, AES were eluted using 80:20 and 95:5 methanol–concentrated aqueous HCl mixtures. Methanol and water were removed from the FAE and AES fractions, and the residues were dissolved in 1,4-dioxane. In this medium, esterification of FAE and transesterification of AES with a cyclic anhydride was performed. Phthalic and diphenic anhydrides were used to derivatizate the surfactants in industrial samples and seawater extracts, respectively. Separation of the derivatized oligomers was achieved by gradient elution on a C8 column with acetonitrile/water in the presence of 0.1% acetic acid. Good resolution between both the hydrocarbon series and the successive oligomers within the series was achieved. Cross-contamination of FAE with AES and vice versa was not observed. Using dodecyl alcohol as calibration standard, and correction of the peak areas of the derivatized oligomers by their respective UV–vis response factors, both FAE and AES were evaluated. After solid-phase extraction on C18, the proposed method was successfully applied to the characterization and determination of the two surfactant classes in industrial samples and in seawater. © 2011 Elsevier B.V. All rights reserved.

1. Introduction

attached to an EO chain, but AES oligomers end with a sulfate group in substitution of the –OH group of FAE oligomers [4–6]:

Fatty alcohol ethoxylates (FAE) and alkylether sulfates (AES) are two important surfactant classes, widely used in cleaners and body care products [1]. FAE are industrially obtained as complex mixtures of oligomers with the following structure (shortened below as CnEm):

Na+ CH3 (CH2 )n−1 (OCH2 CH2 )m OSO− 3

CH3 (CH2 )n−1 (OCH2 CH2 )m OH where n is the number of carbon atoms in the alkyl moiety of the molecule, and m is the number of ethylene oxide groups (EO). FAE mixtures obtained from vegetal oils contain linear hydrocarbon chains with even values of n, whereas both linear and branched chains, with even and odd values of n, can be found in FAE obtained from mineral oils [2,3]. On the other hand, AES are obtained by esterification of FAE with either sulfur trioxide or chlorosulfonic acid. Then, FAE and AES have essentially the same molecular structure, with a hydrocarbon chain

∗ Corresponding author. Tel.: +34 963543003; fax: +34 963544436. E-mail address: [email protected] (G. Ramis-Ramos). 0021-9673/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.chroma.2011.09.059

Accordingly, AES oligomers are shortened below as CnEmS. Important characteristics of both FAE and AES, including viscosity of their aqueous solutions, detergency, foam formation and skin compatibility, as well as their environmental impact, depend on the variable distributions of both the alkyl and EO chains [3,6–13]. Both the hydrocarbon cut (range of n for the predominant hydro¯ (average number of EO units) are important in carbon series) and m industrial quality control. Thus, methods for their characterization and determination are required; however, owing to the complexity of the sample matrices, lack of chromophore, and wide ranges of polarity and volatility of the oligomers, the analysis of FAE and AES, which are usually found in complex mixtures with other surfactant classes, is not an easy task. In addition, the unavailability of commercial standards constitutes an added difficulty of AES analysis. In determination of FAE, the low volatility and thermal instability of long EO chains limit the use of GC to the oligomers with m < 4 [8,14], and owing to the high volatility of the oligomers with short EO chains, the employ of HPLC with evaporative light scattering

8512

M. Beneito-Cambra et al. / J. Chromatogr. A 1218 (2011) 8511–8518

detection (ELSD) provides biased distributions [15,16]. In addition, the positive-ion MS response factors for underivatized FAE oligomers decrease ca. two orders of magnitude when m decreases from 4 to 1 [2,8,14,17]. Further, non-ethoxylated alcohols (m = 0) are not detected in a mass spectrometer [2,8,14,16–18]. Underivatized FAE can be also characterized and determined using isocratic elution and a refractive index detector [19–26], but selectivity is poor and the limits of detection are large. Derivatization procedures designed to increase volatility of FAE oligomers previous to GC analysis have been proposed [3,27,28]. Several derivatization procedures for FAE, addressed to add a chromophore or a charge to the oligomers, followed by HPLC [3,17,29–33] or CE [2,16,17,32–35], have been also described. FAE derivatives have been separated by either NP- or RP-HPLC using UV–vis [3,27,31,36–39] or MS detection [14,16,40–42]. On the other hand, nonspecific determination of AES and other anionic surfactants can be jointly performed by the methylene blue active substance method [43]. Non-ethoxylated alkyl sulfates have been studied using ion pair chromatography with indirect UV detection [44,45], ion chromatography with conductimetric detection [46,47] and HPLC with post-column ion-pair formation followed by membrane phase separation and fluorimetry [48]. Also, conversion of AES to the corresponding alkyl bromides followed by GC-FID [49], as well as HPLC coupled to ESI-MS [50–52], have been applied to the determination of AES in waters, sewage sludge and marine sediments. In former work, we have developed procedures for the RPHPLC–UV determination of FAE previous derivatization with a cyclic anhydride [37–39,53]. Afterwards, we observed that cyclic anhydrides also derivatize AES to yield exactly the same derivatives as FAE, i.e. the hemiesters of the alkyl-alcohol or alkyl-ethoxyalcohol residues. Thus, peaks corresponding to the sum of both FAE and AES oligomers were obtained on the chromatograms if the two surfactant classes were present in the samples. Therefore, in this work, a procedure for the separation of these two surfactant classes, followed by the independent derivatization of each class with a cyclic aromatic anhydride, and RP-HPLC–UV determination of the derivatized oligomers, was developed. Separation of the two surfactant classes was achieved by solid phase extraction (SPE) on a strong anionic exchanger (SAX). Then, using a cyclic aromatic anhydride, FAE and AES were esterified and transesterified, respectively. Phthalic anhydride was used for industrial samples, but diphenic anhydride which gives lower limits of detection [39] was preferred for environmental samples. Separation of the derivatized oligomers was achieved by RP-HPLC–UV, also using MS detection for environmental samples. The proposed method was applied to the analysis of FAE and AES in industrial cleaners and seawater extracts.

2. Materials and methods 2.1. Instruments, reagents and samples A liquid chromatograph (HP 1100, Agilent, Waldbronn, Germany) constituted by a quaternary pump, an on-line degasser, a thermostated column compartment, an automatic sampler and a UV–vis variable multiwavelength detector, was used. When required, peak identification was confirmed by coupling the chromatograph to the ESI source of an 1100 VL ion-trap MS system (Agilent). The column was a C8 fused-core particle type (AscentisExpress, 2.7 ␮m, 15 cm × 4.6 mm ID, Supelco, Bellefonte, PA, USA). Analytical grade reagents were: methanol (MeOH), acetic acid, acetonitrile (ACN), ammonia, dimethyl sulfoxide (DMSO) (Panreac, Barcelona, Spain), phthalic anhydride (≥99%), urea (99.5%) (Fluka, Buchs, Suiza), hydrochloric acid (37%), diphenic anhydride (98%) and 1,4-dioxane (99.8%, Sigma–Aldrich, Steinheim, Germany). The

following FAE and AES oligomers were used for peak identification or as calibration standards: C8E0, C10E0, C12E0, C14E0, C12E1, C12E2, C14E1, C14E2 (Sigma–Aldrich), C8E0S and C12E0S (Fluka); C9E0 (Sigma–Aldrich) and C8E0S were also used as internal standards. To measure response factors using MS detection, in addition to the standards of above, the following standards were also used: from C12E3 to C12E7, C14E5 and C14E6 (Sigma–Aldrich). The industrial mixtures Dehydol LT-7 (FAE with 12 ≤ n ≤ 18 and ¯ = 7, Cognis, Monheim, Germany) and LES average EO number m (lauryl ether sulfate, in fact a mixture of AES oligomers with ¯ = 3, kindly supplied by Químicas Oro, San Anto12 ≤ n ≤ 18 and m nio de Benageber, Valencia, Spain) were also used. To study the possible interferences with other surfactant classes, linear sodium alkyl benzenesulfonates (LAS, industrial mixture with 10 ≤ n ≤ 13) and sodium secondary alkane sulfonates (SAS, industrial mixture with 14 ≤ n ≤ 16) (supplied by Químicas Oro) were used. Deionized water (Barnstead deionizer, Sybron, Boston, MA, USA) was also used. The following SPE cartridges (Phenomenex, CA, USA) ˚ and were employed: Strata C18-E (500 mg/6 mL, 55 ␮m, 140 A) ˚ Industrial liquid cleanStrata SAX (1000 mg/6 mL, 55 ␮m, 70 A). ers were kindly supplied by Químicas Oro and Industria Jabonera Lina (Torras de Cotillas, Murcia, Spain). Two seawater samples were collected at a beach near the urban area of La Pobla de Farnals (Valencia, Spain). 2.2. HPLC separation of the derivatives and detection procedures Gradient elution was accomplished by mixing two solutions containing 50% ACN in water (A) and 100% ACN (B) both in the presence of 0.1% acetic acid. Unless otherwise stated, a linear gradient from A to B in 50 min was used for phtalates, and from 60 to 90% ACN in 50 min was used for diphenates. The flow rate was 1 mL min−1 . All the injected solutions (20 ␮L) were previously passed through a 0.45 ␮m pore-size nylon filter (Albet, Barcelona). Detection was performed at 230 ± 10 nm using 360 ± 60 nm as reference. Peak areas were measured with the ChemStation for LC v.10.02 software (Agilent). Mass spectra were scanned within the m/z 100–900 range; the capillary voltage was 4 kV, and 6 V was applied to skimmer 2; the voltage of skimmer 1 was automatically fixed as a function of the target mass. The target mass was set at m/z 400. Maximum loading of the ion trap was 3 × 104 counts. Nitrogen was used as the nebulizing (35 psi) and drying gas (7 L min−1 at 300 ◦ C). The Agilent LC/MSD v. 4.2 software was used for MS data analysis. 2.3. SAX separation of FAE and AES classes and derivatization procedures The optimized SAX separation procedure of the two surfactant classes is outlined in the scheme of Fig. 1. First, the SAX cartridge is conditioned with 50:50 MeOH:H2 O. Then, the sample containing FAE and AES in a 50:50 MeOH:H2 O medium is passed through the cartridge; AES are retained, while most FAE are eluted (fraction 1). The cartridge is washed with the same 50:50 MeOH:H2 O medium to remove cations (fraction 2). A small part of FAE, enriched in hydrophilic oligomers, is also eluted with fraction 2. The cartridge is then washed with 80:20 MeOH:H2 O to elute fraction 3, constituted by a FAE mixture enriched in hydrophobic oligomers. The same 15mL screw-cap tube is used to collect the combined FAE fractions 1 + 2 + 3. Then, AES are eluted using mixtures containing MeOH and concentrated HCl in water; these mixtures are next indicated as MeOH:HCl. Hydrophilic AES (fraction 4) are eluted with 80:20 MeOH:HCl (chloride concentration in the mixture, 2.4 M), followed by elution of hydrophobic AES (fraction 5) with 95:5 MeOH:HCl (chloride concentration in the mixture, 0.6 M). The combined AES fractions 4 + 5, collected in a second 15-mL screw-cap tube, are

M. Beneito-Cambra et al. / J. Chromatogr. A 1218 (2011) 8511–8518

FAE + AES in 50:50 MeOH:H2O Fraction 1 (hydrophilic FAEs) Fraction 2

SAX

Fraction 3 Fraction 4 Fraction 5

50:50 MeOH:H2O (cation removal) 80:20 MeOH:H2O (hydrophobic FAE) 80:20 MeOH:HCl (hydrophilic AES) 95:5 MeOH:HCl (hydrophobic AES) Solvent evaporation

Eluate of FAE: combined fractions 1 + 2 + 3

Neutralization Solvent evaporation

Eluate of AES: combined fractions 4 + 5

Fig. 1. Scheme of separation of FAE and AES using a SAX cartridge.

dropwise neutralized with concentrated aqueous NH3 in the presence of a drop of 1% phenolphthalein in MeOH, until colour change. Before derivatization, MeOH and water in the tubes containing the FAE and AES fractions are evaporated with a nitrogen stream, followed by removal of the last 1–2 mL of solvents in a centrifugal vacuum evaporator (MiVac, Genevac, Ipswich, UK). Procedures reported in the literature to derivatize FAE with phthalic [38] and diphenic [39] anhydrides were used. In this work, application of the same procedures to the transesterification of AES according to the reaction schemes of Fig. 2 was studied. Briefly, after solvent evaporation, derivatization of FAE and AES was separately performed in their respective tubes by adding 0.25 g finely grinded urea, 2 mL 1,4-dioxane and either 1 g phthalic anhydride or 0.5 g diphenic anhydride for the derivatization of industrial samples or extracts of environmental samples, respectively. The tubes were shaken and introduced in a silicone-oil thermostatic bath at 105 ◦ C for 90 min. After cooling, the residue was dissolved by adding 10 mL of a 2:1 MeOH:H2 O mixture containing 0.1 M NH3 ; however, only 5 mL of this mixture was added to dissolve the derivatized seawater extracts. Further, for these later, the volume of the resulting solution was reduced under a nitrogen stream to 1 mL. The solutions were injected immediately or stored in a freezer at −20 ◦ C. 2.4. Preparation of reagents and samples For calibration studies, stock solutions of C8E0 and C12E0 (2 mg mL−1 ) in 1,4-dioxane, and C8E0S and C12E0S (2 mg mL−1 ) in 1,4-dioxane containing 7% DMSO, were prepared. DMSO provided

the necessary polarity to solubilize AES. In addition, to optimize SAX separation of the surfactant classes, stock solutions of C14E0 and C12E0S (5 mg mL−1 ) were prepared in MeOH. These stock solutions were used to prepare mixtures containing 0.5 mg mL−1 of each standard in media containing different MeOH/H2 O ratios (20:80, 30:70, 50:50 and 80:20, all as v/v). Also for SAX separation optimization, stock solutions of Dehydol LT-7 and LES containing ca. 40 mg mL−1 , as well as mixtures of them (ca. 20 mg mL−1 of each surfactant class) were prepared in 50:50 MeOH:H2 O; 1mL aliquots of these stock solutions were diluted to 5 mL with the adequate amounts of MeOH and water to prepare mixtures in 20:80, 30:70, 50:50 and 80:20 MeOH/H2 O. All these mixtures were passed through the SAX cartridges, and the eluates were derivatized and analyzed by HPLC as indicated. An internal standard stock solution, containing 5 mg mL−1 of each C9E0 and C8E0S in 50:50 MeOH/H2 O, was also prepared. To estimate relative sensitivities of the phthalates and diphenates of the FAE oligomers using MS detection, a stock solution containing ca. 30–50 mg of each standard in 100 mL 1,4-dioxane (0.5 mg mL−1 each) was prepared. Aliquots of 2-mL of this solution were derivatized. After derivatization, 10 mL (phthalates) or 25 mL (diphenates) of the 2:1 MeOH:H2 O mixture containing 0.1 M NH3 was added, and aliquots were injected. For a few oligomers, MS relative sensitivities were estimated by injecting solutions of derivatized Dehydol LT-7. In this case, the concentrations of the oligomers were established from the UV chromatogram using previously reported UV relative sensitivities [38,39]. To analyze industrial raw materials and mixtures of them according to the optimized procedure, a 0.1–1 g portion was diluted to 25 mL with 50:50 MeOH:H2 O, 0.2 mL of the internal standard stock solution was added, and the mixture was passed through a SAX cartridge as described above. The FAE and AES fractions were independently derivatized and injected. For liquid cleaners, a clean-up step previous to the SAX separation was necessary. For this purpose, the samples were diluted up to ca. 25 mL with water and passed sequentially through two SPE C18 cartridges. More water (ca. 10 mL) was used for washing, and the two cartridges were then successively eluted twice with 2.5-mL aliquots of each 50:50 and 80:20 MeOH:H2 O mixtures. As indicated in the optimized procedure, water was added to the eluates to reduce the MeOH concentration to 50% before SAX separation. Concerning to seawater, two 5-L samples were collected in polyethylene containers, transported to the laboratory and allowed to settle for a few hours. Two volumes of 1.5 L each, taken from the upper layer of the containers, were passed through two SPE C18 cartridges at a rate of ca. 5 mL min−1 . Elution of each cartridge was performed twice with 2-mL portions of 50:50 MeOH:H2 O, followed by two additional 2mL portions of 80:20 MeOH:H2 O. Then water was added to the combined eluates to reduce MeOH concentration to 50% before SAX separation of the surfactant classes. Finally, the FAE and AES fractions were independently derivatized, and the two chromatograms were obtained.

O

R R

OH or

O

or

+ -

SO3

COO-

O

O

O

O

8513

+

H2O or SO3

O

COOFig. 2. Reaction schemes for the esterification of FAE and transesterification of AES with either phthalic or diphenic anhydrides.

8514

M. Beneito-Cambra et al. / J. Chromatogr. A 1218 (2011) 8511–8518

the series, which is indicated in the figure for the n = 12 series, was established by injecting standards of a number of oligomers, and was also confirmed by HPLC–MS using extracted ion chromatograms (EICs, not shown). Except for m = 0 and 1, the oligomers within the series eluted by following the order of decreasing m. The consecutive pairs of oligomers were also fairly well resolved; however, the peaks of the oligomers with m = 1 and 0 overlapped with other oligomers within their respective hydrocarbon series. Reversion of the elution order for m = 1 and 0 within the hydrocarbon series has been explained as due to the rigidity of the short hydrophilic moiety of these oligomers, which hinders intramolecular solvation, thus making their hydrophobicity to decrease with respect to the oligomers with m ≥ 2 [37–39,53].

100 Formation of diphenates Relative sum of peak areas

80

60 Formation of phthalates

40

20

0 0

40

3.2. Optimization of the SAX separation of the surfactant classes

80

Reaction time (min) Fig. 3. Relative sum of the chromatographic peak areas of the derivatives of all the oligomers given by an industrial AES sample (LES) against reaction time using phthalic (rhombus and dashed line) and diphenic anhydrides (squares and dotted line). Each point represents an independent derivatization performed at 105 ◦ C in 1,4-dioxane.

3. Results and discussion 3.1. Derivatization and HPLC separation of the derivatives As shown in Fig. 3, an industrial mixture of AES (LES) was quantitatively derivatized at 105 ◦ C in about 70 and 50 min using phthalic and diphenic anhydrides, respectively. Similar results were reported for the esterification of FAE [38,39]. Thus, to assure derivatization of the two surfactant classes, 90 min was selected. Chromatograms of Dehydol LT-7 and LES, obtained by previous derivatization with phthalic anhydride, are shown in Fig. 4. Both chromatograms showed the successive hydrocarbon series at increasing values of n. Hydrocarbon series having exclusively even values of n were observed with Dehydol LT-7, but significant amounts of the n = 13 and 15 odd series were also present in the LES chromatograms. The large differences between the peak profiles of the same hydrocarbon series for Dehydol LT-7 and LES were ¯ = 7 for Dehydol due to the different average EO number, namely m ¯ = 3 for LES. The elution order of the oligomers within LT-7 and m

Absorbance at 230 nm (mAU)

300 200

C12E0 C12E1 + + C12E5 C12E4 C12E6 C12E7 C12E8 C12E3 C12E9 C12E2

n = 12

A

n = 14

n = 16

n = 18

100 0 300 200 100

n = 12 C12E0S + C12E5S C12E6S C12E7S C12E8S C12E9S

B C12E1S + C12E4S

n = 14

C12E3S

n = 13

C12E2S

n = 15 n = 16

n = 18

0 10

20

Time (min)

30

40

Fig. 4. Chromatograms obtained after derivatization of Dehydol LT-7 (A) and LES (B) with phthalic anhydride. In both cases, ca. 40 mg were derivatized and final volume before injection was 12 mL. Elution with a linear gradient from 50 to 100% ACN in 50 min at 25 ◦ C. The insets show peak identifications for the n = 12 series.

Since derivatization of both FAE and AES leads to the same derivatives, a procedure to previously separate the two surfactant classes was developed. For this purpose, the method proposed by Fendinger et al. for non-ethoxylated alkylsulfates in water using SAX cartridges was modified [54]. The procedure was optimized to achieve quantitative isolation of the two surfactant classes, including both hydrophilic (low n and high m values) and hydrophobic (high n and low m values) oligomers. Conditioning of both sample and SAX cartridge with 80:20 MeOH/H2 O led to coelution of part of the AES jointly with the FAE. Thus, according to the optimized scheme of Fig. 1 and Section 2.3, conditioning was performed with 50:50 MeOH/H2 O. In this medium, partial elution of FAE, but without coelution of AES oligomers, was achieved. Solubilization of the most hydrophobic oligomers, together with disruption of FAE–AES mixed micelles, was also procured with 50% MeOH. At this point of the procedure, an increase of the MeOH concentration to 80% to complete FAE elution led to the partial coelution of AES. Coelution was avoided by first washing the cartridge with additional portions of 50% MeOH. In this way, residual cations from the sample (mostly, Na+ and K+ ) were washed away, thus quantitatively fixing AES on the SAX cartridge before increasing the hydrophobicity of the medium. Experiments performed with LES in 50% MeOH showed the absence of AES oligomers in the eluates obtained by increasing MeOH concentration to 80% after washing the cartridge first with more 50% MeOH (chromatograms not given). As shown in Fig. 5, application of the optimized procedure to a Dehydol LT-7 solution led to the elution of an 82% FAE with 50% MeOH (Fig. 5A, combined fractions 1 + 2 of Fig. 1), the residual 18% FAE being eluted with 80% MeOH (Fig. 5B, fraction 3 of Fig. 1). Further, in comparison to the expected oligomer distribution for Dehydol LT-7, the 80% MeOH fraction (Fig. 5B) showed a higher proportion of the hydrophobic oligomers, including both oligomers with large values of n and oligomers with low values of m within the series, than the 50% MeOH fraction (Fig. 5A). A residual amount of 0.4% of the total FAE present in Dehydol LT-7 was obtained after washing the cartridges with two additional 4-mL volumes of 80:20 MeOH/H2 O (chromatogram not shown). Next, AES were eluted using HCl solutions in MeOH as a means to achieve both high concentrations of Cl− and a hydrophobic medium. According to the chromatograms shown in Fig. 6, which were obtained using LES, a 62% of the AES oligomers was eluted with 80:20 MeOH:HCl (Fig. 6A, fraction 4 of Fig. 1). An additional 35%, containing a higher proportion of hydrophobic oligomers than the former fraction, was eluted by increasing MeOH to 95% (Fig. 6B, fraction 5 of Fig. 1). Total elution of LES oligomers was checked by washing the cartridge with two additional 2-mL volumes of both 80:20 and 95:5 MeOH:HCl; in this way, a residual 3% AES was recovered (chromatogram not shown). As indicated in the optimized procedure (Fig. 1 and Section 2.3), the combined 4 + 5 fractions containing the AES were neutralized with concentrated ammonia

M. Beneito-Cambra et al. / J. Chromatogr. A 1218 (2011) 8511–8518

300

n = 12

A

Absoorbance at 23 30 nm (mAU U)

200 n = 14

100

n = 16

n = 18

n = 16

n = 18

0 n = 14

40

B

n = 12

in the combined fractions 1 + 2 + 3. Similarly, using LES, the chromatogram obtained for the combined fractions 4 + 5 was closely similar to that observed in Fig. 4B. Also using LES, the chromatogram obtained with the combined fractions 1 + 2 + 3 showed a series of small peaks which followed the same pattern as that observed in Fig. 4B, but with much smaller peak areas. This later chromatogram was attributed to the FAE impurities which are always present in industrial AES, due to the non-quantitative sulfatation of FAE during AES manufacture. The total peak area of the FAE chromatogram of LES (combined fractions 1 + 2 + 3), divided by the sum of the peak areas of the FAE and AES chromatograms, indicated a molar percentage of ca. 1% FAE in the LES sample. 3.3. Calibration studies

20

0 10

20

30

40

Time (min) Fig. 5. Chromatograms of Dehydol LT-7 derivatized with phthalic anhydride: (A) FAE oligomers eluted with 50:50 MeOH:H2 O (combined fractions 1 and 2 of Fig. 1); (B) FAE oligomers eluted with 80:20 MeOH:H2 O (fraction 3 of Fig. 1). Chromatographic conditions as in Fig. 4; other details as indicated in Section 2.2.

to prevent the release of acid fumes during solvent evaporation. Finally, the isolated FAE and AES fractions were derivatized and injected. Application of this procedure to mixtures of Dehydol LT-7 and LES (ca. 20 mg each) according to the scheme of Fig. 1, gave rise to chromatograms of the FAE and AES fractions closely resembling the chromatograms given in Fig. 4, which were obtained by directly derivatizing and injecting aliquots of Dehydol LT-7 and LES solutions, without passing the solutions by the SAX cartridges. In addition, the optimized procedure, including the SAX separation into two fractions, was also independently applied to aliquots of Dehydol LT-7 and LES solutions. For Dehydol LT-7 (ca. 40 mg), the chromatogram obtained from the combined fractions 1 + 2 + 3 (see Fig. 1) was closely similar to that observed in Fig. 4A, whereas the chromatogram obtained with the combined fractions 4 + 5 showed no significant peaks. Therefore, FAE were quantitatively retained

n = 12

A

100

Absoorbance at 23 30 nm (mAU U)

8515

50

n = 13

n = 14

n = 15

0 n = 14

B 20 n = 12

n = 13

n = 15 n = 16

n = 18

0 20

30

40

Time (min) Fig. 6. Chromatograms of LES derivatized with phthalic anhydride: (A) AES oligomers eluted with 80:20 MeOH:HCl (fraction 4 of Fig. 1); (B) AES oligomers eluted with 95:5 MeOH:HCl (fraction 5 of Fig. 1). Chromatographic conditions as in Fig. 4; other details as indicated in Section 2.2.

Standard solutions of C8E0, C12E0, C8E0S and C12E0S were independently used to construct calibration curves for FAE and AES. For each standard, the derivatization procedure was applied to series of solutions containing increasing concentrations of the oligomers, ranging from 0.06 up to 1.7 mM in the injected solutions (six solutions per oligomer). Plotting the areas, an excellent linearity was obtained in all cases (r2 > 0.999). Further, the slopes of the calibration plots obtained with the four oligomers did not differ significantly from each other, i.e. the maximal slope difference was 1%. The LODs, estimated for a S/N = 3, were all close to 0.5 ␮M (corresponding to LOQs of ca. 1.7 ␮M), which agrees with reported values for the phthalates of non-ethoxylated alcohols [39]. Therefore, sensitivity was the same for non-ethoxylated alcohols and alkylsulfates, independently from the length of the alkyl chain. The similarity of the slopes indicated both a high degree of derivatization, presumably close to 100%, for the two surfactant classes. For FAE oligomers, sensitivity depends on m, and also increases slightly with n when m ≥ 1 [38,39]. However, since AES oligomers are quantitatively converted to the same derivatives as those obtained with FAE oligomers, then, the UV–vis response factors of the FAE oligomers, which were established in previous work for the derivatives obtained with the phthalic [38] and diphenic anhydrides [39], should be also valid for the corresponding derivatized AES oligomers. This is of practical interest, because standards of non-ethoxylated fatty alcohols are widely available, whereas non-ethoxylated alkylsulfates are rare and expensive. Further, FAE oligomers with m > 0 are commercially available, which does not occur with the corresponding AES oligomers. Thus, in Sections 3.4 and 3.5, the unexpensive and widely available dodecyl alcohol (C12E0) was exclusively used as a standard to evaluate without bias total surfactant class concentrations, hydrocarbon series distribu¯ of the complex mixtures of FAE tion and average EO numbers (m) and AES oligomers found in industrial and environmental samples. For this purpose, the C12E0 peak areas were used to construct the calibration curve. Then, the proposed procedure was applied to the samples of Table 1. A pair of chromatograms per sample, corresponding to the FAE and AES fractions, was obtained in duplicate, and all the peak areas were measured. Overlapping of the m = 0 and 1 peaks with the peaks of other oligomers within their respective hydrocarbon series made necessary to establish an indirect way of estimating the individual areas of the overlapped peaks. In certain elution conditions, to estimate the area of the m = 5 and m = 4 peaks by interpolation is straightforward and sufficiently accurate for most applications. Thus, interpolation is possible and accurate with the phthalates at 25 ◦ C, due to the perfect overlapping of the peaks by pairs of oligomers (m = 0 with m = 5, and m = 1 with m = 4) [38]. As shown in Fig. 7, the peak areas of the m = 5 and 4 oligomers were estimated by non-linear interpolation of the peak areas of the 2 ≤ m ≤ 3 and m ≥ 6 oligomers of their respective hydrocarbon series. The peak areas of the Dehydol LT-7 and LES fitted well to the cubic equation and to exponential equations of the form A = be−am ,

8516

M. Beneito-Cambra et al. / J. Chromatogr. A 1218 (2011) 8511–8518

Table 1 Declared and found composition for industrial samples. Sample

Surfactant class

Declared, %

Found,a %

Generic LES

AES FAE AES FAE AES FAE

68–72 Impurity 4 Unknown 5 Unknown

69 0.81 3.9 0.10 4.8 0.20

Cleaner 1 Cleaner 2

a b

± ± ± ± ± ±

Found distribution of hydrocarbon series,b %

3 0.04 0.2 0.02 0.2 0.02

n = 12

n = 14

n = 16

77 (2.0) – 74 (2.1) – 73 (2.1) –

22 (1.9) – 22 (2.2) – 26 (2.6) –

0.87 (2.1) – 4.1 (2.4) – 3.3 (1.8) –

Confidence limits for p = 0.05 (n = 5). ¯ found for each AES series is indicated between parentheses. In mass percentages; the average EO number, m,

respectively. Then, the peak areas of the m = 0 and 1 oligomers were obtained by subtracting the m = 5 and 4 interpolated areas from the total area of the corresponding peaks, respectively. All the peak areas up to m = 12 were next divided by the tabulated response factors of the respective oligomers [38,39], and the molar concentrations of the oligomers were obtained by dividing the corrected peak areas by the slope of the C12E0 calibration curve. ¯ was calculated from the molar Then, the average EO number, m, distribution of the oligomers. Total mass concentrations of the surfactant classes, and the mass distribution of hydrocarbon series, were finally calculated by taking into account the molar masses of the oligomers and the dilution factor of the sample in the injected solutions. 3.4. Study of interferences and application to industrial samples The common anionic surfactants secondary alkane sulfonates (SAS) and linear alkyl benzene sulfonates (LAS) were studied as potential interferences. These anionic surfactants should be retained together with the AES in the SAX cartridge; however, application of the procedure to a SAS sample gave a chromatogram with no additional peaks with respect to that of the reagent blank. Also, a 1:1 mixture of LES and SAS (ca. 20 mg each) gave rise to a chromatogram which was undistinguishable from that of LES. Then, SAS showed no reaction with the anhydrides and did not cause interference either. On the other hand, LAS are aromatic surfactants which absorb in UV–vis. The HPLC separation of a LAS sample in the conditions used to separate FAE and AES phthalates lead to a series of large and wide queuing bands, close to the dead volume,

on the chromatogram of the AES fraction. Further, this pattern was not modified by application of the derivatization procedure to a LAS sample; then, LAS did not react either with phthalic anhydride. Application of the procedure to a mixture of LAS and LES (ca. 20 mg each) gave rise to a chromatogram showing the bands of LAS at low retention times, followed by the expected peak pattern of the derivatives of the LES oligomers. The chromatogram obtained for the AES fraction of a commercial laundry liquid cleaner (nominal components: 4.3% LAS, 4.0% FAE, 2.0% AES and others), is shown in Fig. 8. The bands due to the homologues and isomers of LAS were present at low retention times, but they did not cause interference in the evaluation of the AES series of industrial interest (n ≥ 10). The determination of residual FAE in industrial AES is important to control both the sulfation process and the quality of the final product. As indicated in Table 1, application of the proposed procedure to an industrial AES concentrate showed an AES content which agreed with the declared value, plus a residual 0.81% of nonsulfated FAE. According to the data given in Table 1, the proposed method is also useful for the characterization and determination of both FAE and AES in industrial cleaners containing LAS and other unknown components. The results of Table 1 were obtained without the need of applying the internal standard correction; however, the presence of the peaks of the nonyl and octyl phthalates on the FAE and AES chromatograms, respectively, indicated that no significant losses of any surfactant class were produced during the separation and derivatization steps. 3.5. Seawater analysis Application of the proposed procedure using diphenic anhydride to seawater extracts led to chromatograms as that shown in Fig. 9. For the diphenates, the order of the lighter oligomers within

1

100

Absorbance at 230 nm (mAU)

Relative area R

FAE, m = 7 n = 12

AES, m = 3 n = 12

n = 14 n = 14 0

n = 12 0+5 1+4

50

32

n = 14 0+5 1+4 2

LAS 6

6

3

n = 16

0 2

4

6

8

10

12

Number of EOs Fig. 7. Estimation of the uncorrected peak areas of the m = 4 and 5 oligomers of the n = 12 and 14 series in Dehydol LT-7 by cubic interpolation (empty symbols, solid lines) and LES by exponential interpolation (full symbols, dashed lines). Experimental (rhombus) and interpolated peak areas (squares).

10

20

30

Time (min) Fig. 8. Chromatogram of the AES fraction of a commercial laundry liquid cleaner obtained as indicated in the optimized procedure of Section 2.2, followed by derivatization with phthalic anhydride. The numbers at the peaks are m values. The bands of LAS are indicated. Other conditions as in Fig. 4.

Absorrbance (mAU U)

M. Beneito-Cambra et al. / J. Chromatogr. A 1218 (2011) 8511–8518

12 8

A

02

FAE n = 12

1 40

AES n = 12

3 20

4

4

0

B

0

AES n = 14

20

3

0

0

C 1+2

4

8517

4 0

0

4

1+2

3

0

Intensity × 105

4 3

6

2

4

4

2 0

1

5

2

3

3

2 1

5

22

24 26 Time (min) (min)

2 1

0

0

0

4

22

32 34 Time (min)

30

24 26 Time (min)

Fig. 9. Chromatograms of a seawater extract obtained by derivatization with diphenic anhydride and using UV–vis (upper parts, at 230 nm) and MS detection (EICs, lower parts): (A) FAE, n = 12 series; (B) AES, n = 12 series; (C) AES, n = 14 series. The numbers at the peaks are m values. Elution conditions: linear gradient from 60 to 90% ACN in 50 min at 25 ◦ C.

the series differs from that of the phthalates. Thus, in agreement with a previous study [39], retention increased for all the oligomers at decreasing values of m, with the exception of the m = 0 oligomer of each series which appeared between the corresponding m = 2 and 3 oligomers. Using both UV–vis and MS detection, the seawater samples showed measurable amounts of several FAE oligomers of the n = 12 series, and several AES oligomers of the n = 12 and n = 14 series. According to the C12E0 calibration curve (constructed using the diphenate), a LOD of 0.1 ␮M (LOQ of ca. 0.3 ␮M) was estimated for the UV detection of standard solutions (S/N = 3); however, from the chromatograms of seawater extracts (which have a noisier baseline), a LOD of 3 ␮M (LOQ of ca. 10 ␮M) was calculated. This LOD corresponds to ca. 0.7 ␮g L−1 C12E0S in seawater (calculated as 1500 times less than in the injected solutions). From the peak areas, ca. 2.3 and 0.99 ␮g L−1 were estimated in seawater for C12E0S and C14E0S, respectively. Using MS detection, smaller concentrations of other oligomers with m > 0 were also observed in the n = 12 and 14 series of both FAE and AES. Small peaks and shoulders in the UV and MS traces could be due to isomers, which could be present at low concentrations [38]. Also for MS detection, an abnormally large double peak (main peak and shoulder) was observed at the location expected for the m = 4 oligomers of both FAE and AES series. These peaks, which were small at the same retention times of the UV chromatogram could be due to the presence in the seawater extracts of non-absorbing compounds, isobaric with the diphenates of the m = 4 oligomers. However, unmatching of the UV and MS profiles in Fig. 9 can be also due to sensitivity differences between the two detection techniques. Using UV detection, relative sensitivities of the diphenates of the oligomers of the n = 12 series increased slightly from 1 to 1.1 when m increased from 0 to 1, and decreased at higher values of m [39]. Thus, UV relative sensitivities ranged between 0.68 and 0.62 for 3 ≥ m ≥ 8. This depressed UV peak areas of the m > 2 oligomers with respect to those of the m = 0 and 1 oligomers. On the other hand, relative sensitivities of the diphenates of the oligomers of the n = 12 and 14 series up to m = 7, established using HPLC with MS detection, and taking C12E0 as reference, are given in Table 2.

As observed, relative sensitivities of the diphenates decreased for m = 1 and increased when m ≥ 2, showing a pronounced maximum for m = 4 and finally decreasing steadily at higher values of m. In addition, sensitivity of the diphenates increased with n. Thus, in relation to the diphenates of the m < 4 oligomers, MS peak areas are fairly enhanced for the diphenates of the m = 4–6 oligomers. This, together with the depression of the UV response for the diphenates of the oligomers with m > 2, could explain most of the differences observed between the UV and MS profiles of Fig. 9. MS relative sensitivities of the phthalates showed a similar behaviour (see also Table 2), with the difference that the maximal sensitivity at increasing values of m was produced for the m = 6 oligomer instead of the m = 4 oligomer. For the phthalates, relative sensitivity also increased with n. Using either UV or MS detection, the proposed method was capable of distinguishing the oligomers of the two surfactant classes in the seawater, AES being present at higher concentrations than FAE. This could be due to the faster biodegradation of FAE in comparison to AES. This also indicates that previously reported data for FAE in environmental samples, obtained by methods based on derivatization with anhydrides, should be actually seen as the sum of FAE and AES concentrations. Due to the labitity of the ester bond of AES, the concentrations obtained by other analytical methods also based on FAE derivatization could

Table 2 Relative sensitivities for the phthalates and diphenates of the FAE oligomers using MS detection.a Derivatives

n

Phthalates

12 14 12 14

Diphenates

m 0

1

2

3

4

5

6

7

1.0 2.0 1.0 1.6

0.5 0.9 0.5 1.5

0.4 0.8 1.7 3.7

0.7 2.0* 1.5 2.1*

0.9 2.0* 3.1 5.2*

1.4 4.2 3.0 4.8

2.0 4.5 2.7 3.8

2.1 4.4* 2.3 2.2*

a Values with an asterisk were estimated using Dehydol LT-7 and the other values with standards of the individual oligomers.

8518

M. Beneito-Cambra et al. / J. Chromatogr. A 1218 (2011) 8511–8518

be biased by the presence of AES. Therefore, the selectivity of these other methods with respect to AES should be also revised. 4. Conclusions In previous work, procedures for FAE characterization and determination based on derivatization with a cyclic anhydride, followed by RP-HPLC with UV or MS detection, were developed [37–39,53]. However, when FAE are esterified, AES are also transesterified, leading to the same derivatives as FAE. Thus, if the two surfactant classes are present in the samples, the resulting chromatograms show the sum of both FAE and AES oligomers. In this work, a procedure for the SAX separation of both surfactant classes, followed by their independent derivatization and RP-HPLC–UV, was developed. Quantitative isolation of the two surfactant classes, including both hydrophilic and hydrophobic oligomers, was achieved by eluting the FAE and AES fractions in three and two steps, respectively. As previously shown for FAE, quantitative derivatization of AES with either phthalic or diphenic anhydrides has been demonstrated. All the steps involved in the separation of the two surfactant classes and independent derivatization previous to HPLC are simple and can be easily automatized. Separation of the derivatized oligomers, with good resolution between both the hydrocarbon series and the successive oligomers within the series, was achieved by RP-HPLC using gradient elution with ACN/water. It has been also shown that FAE oligomers can be used as calibration standards for AES, which is an advantage because standards of alkylsulfates are rare and expensive, and standards for AES with m > 0 are not commercially available, whereas standards of FAE, including ethoxylated oligomers at many n and m values, are widely available. On the other hand, anionic surfactant classes commonly used in cleaners, including SAS and LAS, did not interfere. The proposed method was successfully applied to the characterization and determination of FAE and AES in industrial liquid cleaners and seawater extracts. The proposed procedure is also useful to control the quality of industrial AES, where FAE are always present as an impurity, due to the non-quantitative sulfatation of FAE during AES manufacture. Acknowledgements Work supported by Project CTQ2010-15335/BQU (MEC of Spain and FEDER funds). M.B.-C. thanks the Universitat de València and Químicas Oro (San Antonio de Benagéber, Spain) for a Cinc SeglesEmpresa grant for PhD studies. References [1] H.P. Fiedler, Lexicon der Hilfsstoffe fur Pharmazie, Kosmetic und angrenzende Gebiete, Cantor, Aulendorf, 1989. [2] C.J. Sparham, I.D. Bromilow, J.R. Dean, J. Chromatogr. A 1062 (2005) 39. [3] A. Marcomini, M. Zanette, J. Chromatogr. A 733 (1996) 193. [4] B. Strain, L. Theoharous, D.D. Whyte, Ind. Eng. Ind. 51 (1959) 13. [5] C.M. Suter, Organic Chemistry of Sulfur, Wiley, New York, 1944. [6] Environmental and Human Safety of Major Surfactants, in: Final Report to the Soap and Detergent Association, vol. 1, Arthur D. Little Co., Cambridge, MA, 1991, Part 2.

[7] T.F. Tadros, Applied Surfactants: Principles and Applications, Wiley-VCH, Weinheim, 2005. [8] P. Rudewicz, B. Munson, Anal. Chem. 58 (1986) 674. [9] C.V. Eadsforth, A.J. Sherren, M.A. Selby, R. Toy, W.S. Eckhoff, D.C. McAvoy, E. Matthijs, Ecotoxicol. Environ. Saf. 64 (2006) 14. [10] S.E. Belanger, P.B. Dorn, R. Toy, G. Boeije, S.J. Marshall, T. Wind, R. van Compernolle, D. Zeller, Ecotoxicol. Environ. Saf. 64 (2006) 85. [11] R. van Compernolle, D.C. McAvoy, A. Sherren, T. Wind, M.L. Cano, S.E. Belanger, P.B. Dorn, K.M. Kerr, Ecotoxicol. Environ. Saf. 64 (2006) 61. [12] I. Ribosa, J. Sanchez-Leal, A. Marsal, M.T. Garcia, Afinidad 64 (2007) 528. [13] E. Jurado, M. Fernandez-Serrano, J. Nunez-Olea, M. Lechuga, J. Surfactants Deterg. 10 (2007) 145. [14] C. Crescenzi, A. Di Corcia, R. Samperi, A. Marcomini, Anal. Chem. 67 (1995) 1797. [15] W. Miszkiewicz, J. Szymanowski, Crit. Rev. Anal. Chem. 25 (1996) 203. [16] V. Bernabé-Zafón, E.F. Simó-Alfonso, G. Ramis-Ramos, J. Chromatogr. A 1118 (2006) 188. [17] J.C. Dunphy, D.G. Pessler, S.W. Morrall, Environ. Sci. Technol. 35 (2001) 1223. [18] K.B. Sherrard, P.J. Marriott, M.J. McCormick, R. Colton, G. Smith, Anal. Chem. 66 (1994) 3394. [19] Y. Mengerink, H.C.J. De Man, S. Van der Wal, J. Chromatogr. 552 (1991) 593. [20] B. Trathnigg, D. Thamer, X. Yan, B. Maier, H.R. Holzbauer, H. Much, J. Chromatogr. A 657 (1993) 365. [21] B. Trathnigg, D. Thamer, X. Yan, B. Maier, H.R. Holzbauer, H. Much, J. Chromatogr. A 665 (1994) 47. [22] B. Trathnigg, A. Gorbunov, J. Chromatogr. A 910 (2001) 207. [23] B. Trathnigg, C. Rappel, J. Chromatogr. A 952 (2002) 149. [24] D. Cho, J. Hong, S. Park, T. Chang, J. Chromatogr. A 986 (2003) 199. [25] B. Trathnigg, S. Fraydl, M. Veronik, J. Chromatogr. A 1038 (2004) 43. [26] B. Trathnigg, C. Rappel, S. Fraydl, A. Gorbunov, J. Chromatogr. A 1085 (2005) 253. [27] A.T. Kiewiet, P. deVoogt, J. Chromatogr. A 733 (1996) 185. [28] J. Hübner, R. Taheri, D. Melchior, H.W. Kling, S. Gäb, O.J. Schmitz, Anal. Bioanal. Chem. 388 (2007) 1755. [29] B.J. Hoffman, L.T. Taylor, S. Rumbelow, L. Goff, J.D. Pinkston, J. Chromatogr. A 1043 (2004) 285. [30] C. Sun, M. Baird, H.A. Anderson, D.L. Brydon, J. Chromatogr. A 771 (1997) 145. [31] M. Zanette, A. Marcomini, E. Marchiori, R. Samperi, J. Chromatogr. A 756 (1996) 159. ˇ cik, J. Hlaváˇc, J. Chromatogr. A 1021 (2003) 19. [32] K. Lemr, J. Sevˇ [33] A.M. Desbène, L. Geulin, C.J. Morin, P.L. Desbène, J. Chromatogr. A 1068 (2005) 159. [34] R.A. Wallingford, Anal. Chem. 68 (1996) 2541. [35] K. Heinig, C. Vogt, G. Werner, Anal. Chem. 70 (1998) 1885. [36] H. Bachus, H.J. Stan, Tens. Surf. Det. 40 (2003) 10. [37] A. Micó-Tormos, C. Collado-Soriano, J.R. Torres-Lapasió, E.F. Simó-Alfonso, G. Ramis-Ramos, J. Chromatogr. A 1180 (2008) 32. [38] A. Micó-Tormos, E.F. Simó-Alfonso, G. Ramis-Ramos, J. Chromatogr. A 1203 (2008) 47. [39] A. Micó-Tormos, F. Bianchi, E.F. Simó-Alfonso, G. Ramis-Ramos, J. Chromatogr. A 1216 (2009) 3023. [40] L.H. Levine, J.L. Garland, J.V. Johnson, J. Chromatogr. A 1062 (2005) 217. [41] P. Jandera, M. Holˇcapek, G. Theodoridis, J. Chromatogr. A 813 (1998) 299. [42] K.A. Krogh, K.V. Vejrup, B.B. Mogensen, B. Halling-Sørensen, J. Chromatogr. A 957 (2002) 45. [43] R.A. Llenado, T.A. Neubecker, Anal. Chem. 55 (1983) 93R. [44] J.A. Boiani, Anal. Chem. 59 (1987) 2583. [45] S.A. Shamsi, N.D. Danielson, Anal. Chem. 67 (1995) 4210. [46] N. Pan, D.J. Pietrzyk, J. Chromatogr. A 706 (1995) 327. [47] L.M. Nair, R. Saari-Nordhaus, J. Chromatogr. A 804 (1998) 233. [48] F. Smedes, J.C. Kraak, C.F. Werkhoven-Goewie, U.A.Th. Brinkman, R.W. Frei, J. Chromatogr. 247 (1982) 123. [49] T.A. Neubecker, Environ. Sci. Technol. 19 (1985) 1232. [50] D.D. Popenoe, S.J. Morris, P.S. Horn, K.T. Norwood, Anal. Chem. 66 (1994) 1620. [51] F. Bruno, R. Curini, A. Di Corcia, I. Fochi, M. Nazzari, R. Samperi, Environ. Sci. Technol. 36 (2002) 4156. [52] P.A. Lara-Martín, A. Gómez-Parra, E. González-Mazo, Environ. Toxicol. Chem. 24 (2005) 2196. [53] A. Micó-Tormos, E.F. Simó-Alfonso, G. Ramis-Ramos, J. Sep. Sci. 33 (2010) 1398. [54] N.J. Fendinger, W.M. Begley, D.C. McAvoy, W.S. Eckhoff, Environ. Sci. Technol. 26 (1992) 2493.