Polyoxyethylene alkyl ether nonionic surfactants: physicochemical properties and use for cholesterol determination in food

Talanta 55 (2001) 69 – 83 www.elsevier.com/locate/talanta

Polyoxyethylene alkyl ether nonionic surfactants: physicochemical properties and use for cholesterol determination in food Alain Berthod a, Seema Tomer b, John G. Dorsey c,* a

Laboratoire des Sciences Analytiques, Uni6ersite´ de Lyon 1, CNRS 5619, Bataille 308 -D, 69622 Villeurbanne, France b Department of Chemistry, Uni6ersity of Cincinnati, Cincinnati, OH 45221, USA c Department of Chemistry, Florida State Uni6ersity, Tallahassee, FL 32306 -4390, USA Received 10 January 2001; received in revised form 8 March 2001; accepted 8 March 2001

Abstract Polyoxyethylene alkyl ethers, Cn Em, are nonionic surfactants made of an alkyl chain with n methylene groups and a hydrophilic part with m oxyethylene units. Cn Em nonionic surfactants are very useful in chemical analysis. The commercially available products are often a mixture of several Cn Em molecules with different m values. Pure Cn Em surfactants are now available. The physicochemical parameters: critical micelle concentration (c.m.c.), molar volume, density, cloud-point temperature and hydrophile– lipophile balance value for pure Cn Em surfactants were collected from the literature. Regression analyses were carried out on the data. They showed that strong correlations existed between the structure of the molecule (n and m values) and its physicochemical properties. General equations linking the c.m.c., molar volume, density and cloud-point temperature of the Cn Em surfactants and their structure (n and m values) are proposed and discussed. The use of these surfactants in chemical analysis is illustrated by the determination of cholesterol in egg yolk. Cholesterol was separated from the bulk yolk by cloud-point extraction using the C12E10 surfactant. It was quantitated using micellar liquid chromatography. The C12E23 surfactant was used to prepare the micellar mobile phase that allowed the separation of cholesterol and the use of an enzymatic detector. © 2001 Elsevier Science B.V. All rights reserved. Keywords: Nonionic surfactants; Critical micellar concentration; Molar volume; Cloud-point temperature; Micellar liquid chromatography; Egg yolk; Cholesterol

1. Introduction

* Corresponding author. Tel.: + 1-850-6444496; fax: + 1850-6455644. E-mail addresses: [email protected] (A. Berthod), [email protected] (J.G. Dorsey).

Most of the nonionic surfactants are ethylene oxide adducts. The polyoxyethylene moiety is the hydrophilic part of the surfactant molecule. The lipophilic part of the molecule can be a variety of apolar moieties including alkyl chains (from alco-

0039-9140/01/$ - see front matter © 2001 Elsevier Science B.V. All rights reserved. PII: S 0 0 3 9 - 9 1 4 0 ( 0 1 ) 0 0 3 9 5 - 2

70

A. Berthod et al. / Talanta 55 (2001) 69–83

hols, fatty acids or amides), alkyl benzenes and their fluorinated counterparts, silicone derivatives or polyoxypropylene chains [1]. Nonionic polyoxyethylene alkyl ethers are ethylene oxide adducts of linear alcohols. They were first synthesized in the early 1930s [2]. Since then, their detergency, wetting and foaming properties, and general usefulness as nonionic surfactants makes them widely used in industrial and household products. The 1999 US production of ethylene oxide reached 4.4 million metric tons (2.5 million in 1989) [3]. The 1999 US retail sales of hair care products containing nonionic surfactants (shampoos, conditioners, creme rinses and hairstyling gels) reached $9 billion [4]. Although these figures cover the polyoxyethylene glycol alkyl ethers and all the other nonionic surfactants containing polyoxyethylene moieties, they give an idea of the importance of these surfactants in the economy. Polyoxyethylene alkyl ethers have the general formula Cn H2n + 1(OCH2CH2)m OH. They will be referred as Cn Em with n indicating the number of carbons in the alkyl chain and m being the number of ethylene oxide units in the hydrophilic moiety. The problem with Cn Em nonionic surfactant is that their synthesis involves ethylene oxide polymerization. The final Cn Em product obtained is actually a mixture of molecules with n carbons in the alkyl chain, if the starting alcohol was pure, and a number of ethylene oxide units ranging from m −a to m + a, with a being a constant depending on the process [5]. A Poisson distribution is obtained. For example, the analysis of a batch of polyoxyethylene-10-dodecanol, C12E10, gave the following result (mass percentage): C12E3, 0.022%; C12E4, 0.45%; C12E5, 2%; C12E6, 5%; C12E7, 8.6%; C12E8, 11.7%; C12E9, 18.3%; C12E10, 18.2%; C12E11, not listed; C12E12, 9.57%; C12E13 and C12E14, not listed; C12E15, 3.55%; and C12E20, 0.3% [1]. Of this particular C12E10 sample, 81.8% was actually not the C12E10 molecule. In most applications of nonionic surfactants, this polydispersity of the number of ethylene oxide units is not a problem. It is a problem when precise physicochemical properties are needed. It explains the variation observed in the literature data when critical micellar concentrations (c.m.c.), molar volumes or cloud-point temperatures are listed [1,5 –8].

Nonionic surfactants can be very useful in analytical chemistry. The micelle –analytical chemistry interface was pointed out as early as 1984 [9]. Polyoxyethylene alkyl ethers are often selected as surfactants for analytical separations because they do not absorb UV light, they are little sensitive to ions and they offer a wide range of differing polarity. Many aqueous solutions of nonionic surfactants separate in two phases on heating above a particular temperature: the clouding temperature. The homogeneous solution becomes turbid (cloudy) and separates into a water-rich phase and an organic surfactant-rich phase. A solute dissolved in the homogeneous solution at low temperature may partition preferentially in one phase or the other. Cloud-point extraction is commonly used to extract and/or to concentrate apolar solutes in the surfactant-rich phase obtained at elevated temperature [6,10 –12]. The cloud-point temperature and the molar volume of the nonionic surfactant used are needed. Micellar liquid chromatography (MLC) is another analytical technique that can exploit the properties of nonionic surfactants [9,13]. The partitioning of solutes between the micelles, the aqueous phase and the stationary phase allows one to separate them according to their affinity for both the micellar phase and the stationary phase. As analytical chemists, we had often problem finding the physicochemical parameters of polyoxyethylene alkyl ethers. The c.m.c. and the molar volume of the surfactant used are needed in MLC to perform retention calculations. The cloud-point temperature is obviously the parameter to know to perform any cloud-point extraction. Since more and more different and purer polyoxyethylene alkyl ethers are commercially available, a collection of their c.m.c., molar volume and cloud-point temperature is presented in the present paper. Using correlation analysis and/or basic physicochemical justification, semi-empirical equations were derived to obtain the physicochemical parameters of pure Cn Em surfactants either available or not yet commercialized. The equations can also be used to study the trend in the changes observed on the physicochemical parameters investigated. They are compared with previously published similar equations and software [14 –16].

A. Berthod et al. / Talanta 55 (2001) 69–83

Next, the use of these nonionic surfactants in chemical analysis is exemplified by an application presented and discussed: the quantitation of cholesterol in fresh egg yolk. Cholesterol is concentrated by cloud-point extraction using a polyoxyethylene alkyl ether. It is quantitated by MLC using another polyoxyethylene alkyl ether to prepare a micellar mobile phase that is able to separate cholesterol from other extracted products and that is compatible with a sensitive enzymatic detection. 2. Material and methods

2.1. Reagents The polyoxyethylene alkyl ethers were C12E10, polyoxyethylene-10-lauryl ether, C32H66O11 (molecular weight (m.w.), 626), purchased from Fluka (Sigma Aldrich Fluka group, St. Louis, MO), and C12E23, or Brij® 35, C58H118O24 (m.w., 1200), obtained from Mallinkrodt (Paris, KY). Cholesterol (C27H46O; m.w., 387), glutaraldhehyde (C5H8O2; m.w., 100; 25% w/w solution), g-aminopropyl triethoxysilane, the enzyme cholesterol oxidase (from Pseudomonas, 40 U mg − 1) and the controlled porous glass beads (90–120 mm particle diameter, 50 nm pore size) were purchased from Sigma. Fresh eggs were purchased from a local grocery store. Water was obtained from a Barnstead Nanopore II water purification system (Boston, MA) and was filtered through a 0.45 mm Nylon membrane filter.

2.2. Chromatographic system The system consisted of a dual piston SP 8800 pump (Spectra Physics, San Jose, CA), a 6035 Rheodyne valve with a 20 ml loop (Rheodyne, Cotaty, CA), and a Zorbax CN column, 15 cm long (4.6 mm i.d.). The outlet of the high-performance liquid chromatography (HPLC) column was connected to a 10 cm column packed with immobilized enzyme cholesterol oxidase glass beads. The enzymatic detector was connected to a diode array detector model 1000 S (Applied Biosystems, Ramsey, NJ) with a 10 ml flow cell and a SP 4270 integrator.

71

A 0.03 M Brij® 35, 10% v/v propanol, 50 mM potassium phosphate buffer (pH 7.2) solution was the mobile phase used for HPLC measurement of cholesterol. The HPLC column was maintained at 25°C by a water jacket. The enzyme reactor was maintained at 35°C. The flow rate was 1 ml min − 1.

2.3. Procedure 2.3.1. Sample preparation Egg yolk (0.5 g) was accurately weighed in a test tube. Two grams of C12E10 (20% w/w) and 45 mg sodium sulfate (0.04 M) were added with 10 ml water. The mixture was sonicated for 10 min and incubated in a water bath at 90°C for another 10 min. After centrifugation and filtration, the solution was placed in a water bath at 95°C for phase separation to occur in less than 10 min. Then the lower surfactant organic layer was separated from the aqueous upper layer without centrifugation. The viscous surfactant layer (volume2 ml) was diluted to a volume of 10 ml by methanol. Reference samples were prepared in the same way, spiking known amounts of pure cholesterol in the surfactant solution. The methanolic solution of the extract was injected in the chromatographic system with the Brij® 35 micellar mobile phase for quantitation. 2.3.2. Cholesterol detection Cholesterol is very difficult to detect with a UV detector due to a lack of UV chromophore in its molecule. The enzyme cholesterol oxidase is able to oxidize a sterol in a conjugated enone that strongly absorbs the 241 nm UV light. An enzyme reactor was made by bonding cholesterol oxidase on controlled porous glass beads through a glutaraldehyde linkage and filling the beads in a 10 cm (4.6 mm i.d.) stainless steel column. The enzymatic oxidation reaction was fast enough to be competed in the 45 s residence time of the mobile phase at 1 ml min − 1. Enzymatic reactors are very fragile and sensitive to organic solvents. When not in use, it was stored at 4°C in the micellar Brij® 35 mobile phase.

A. Berthod et al. / Talanta 55 (2001) 69–83

72

Table 1 Physicochemical parameters of polyoxyethylene alkyl ethers taken from the literature [5–7,10–12]a Nonionic surfactant

m.w. (g mol−1)

c.m.c.

Molar volume

M

g l−1

l mol−1

ml g−1

Cloud point (oC)

C6E3 C7E3 C8E3 C10E3 C12E3 C14E3 C16E3

234 248 262 290 318 346 374

0.09 0.02 0.0075 0.00073 0.0001 10.0×10−6 1.2×10−6

21 5 2 0.21 0.0318 0.00346 0.00045

0.245 0.27 0.23 0.32 – –

0.988 1.03 0.793 1.006 – –

45 27.6 10.6 B0 B0 – –

C6E5 C8E5 C10E5 C12E5 C14E5

322 350 378 406 434

0.09 0.009 0.00084 0.000062 0.00001

29 3.15 0.318 0.025 0.00434

– 0.32 0.39 0.41 0.47

– 0.914 1.03 1.01 1.08

75 60.4 44 31.7 20.5

C8E8 C10E8 C11E8 C12E8 C13E8 C14E8 C15E8 C16E8

482 510 524 538 552 566 580 594

0.01 0.001 0.0003 0.0001 0.00003 0.000009 3.5×10−6 5×10−7

4.8 0.51 0.158 0.0538 0.0166 0.0051 0.002 0.0003

– 0.51 0.52 0.54 – – – –

– 1 0.99 1.004 – – – –

96 85 82 78 72 71 66 66

C8E1 C8E3 C8E4 C8E5 C8E6 C8E8 C8E9

174 262 306 350 394 482 526

0.005 0.0065 0.008 0.009 0.0098 0.01 0.013

0.87 1.7 2.45 3.15 3.86 4.82 6.84

– 0.27 0.31 0.32 0.39 – 0.53

– 1.03 1.013 0.914 0.99 – 1.008

10.6 40 60.4 74 96 100

C12E4 C12E5 C12E6 C12E7 C12E8 C12E9 C12E10 C12E12 C12E13 C12E23 C12E25

362 406 450 494 538 582 626 714 758 1198 1286

0.000065 0.000062 0.000067 0.000073 0.0001 0.00008 0.00009 0.0001 0.0002 0.00006 0.0001

0.0235 0.0252 0.0302 0.0361 0.0538 0.0466 0.0563 0.0714 0.152 0.072 0.129

0.36 0.41 0.45 0.49 0.54 0.58 0.62 0.71 – 1.12 –

0.994 1.01 1.00 0.992 1.004 0.997 0.990 0.994 – 0.935 –

5 31.7 52 63.4 78 80 90 \100 \100 \100 \100

C16E6 C16E10 C16E18 C16E30

506 682 1034 1562

0.000001 6×10−7 3×10−7 1×10−7

0.00051 0.00041 0.00031 0.00016

– – – 1.5

– – – 0.960

32 66 \100 \100

a

Cloud-point temperature for a 10 g l−1 surfactant solution (1% w/w). –, No data found.

A. Berthod et al. / Talanta 55 (2001) 69–83

73

Fig. 1. Plots of log(c.m.c.) versus the number of ethylene oxide units in Cn Em nonionic surfactants. The triangles correspond to literature values (Table 1). The lines were obtained using Eq. (9).

3. Results and discussion Table 1 presents the physicochemical parameters of CnEm nonionic surfactants that were found in the literature [1,6,7,10,17] and in a recent compilation [13]. A dash indicates the corresponding parameter was not found in the references listed.

where mE is the number of ethylene oxide in the Cn Em surfactant molecule, n is the number of data in the regression analysis set and r 2 is the regression coefficient.For the dodecyl ether surfactant: log c.m.c.=0.0046mE − 4.12 n=9, r 2 = 0.906

(2)

For the hexadecyl surfactant:

3.1. Critical micelle concentration

log c.m.c.= − 0.041mE − 5.78

3.1.1. Study of the experimental data The effect on the surfactant c.m.c. of the number of ethylene oxide units and methylene groups in the alkyl chain were studied. Fig. 1 shows the experimental values of c.m.c. (Table 1) plotted as log(c.m.c.) versus the number of ethylene oxide units in the Cn Em molecule. There is a clear linear relationship. The regression lines were as follows. For the octyl ether surfactants: log c.m.c.=0.0469mE −2.31 n=7, r 2 = 0.941

(1)

n=4, r 2 = 0.997

(3)

The same analysis was carried out for the number of methylene groups for Cn Em surfactant molecules with three, five and eight ethylene oxide units. The regression lines were as follows. For the Cn E3 surfactants: log c.m.c.= − 0.487nC + 1.828 n= 7, r 2 = 0.997

(4)

where nC is the number of methylene groups in the alkyl chain.For the Cn E5 surfactants:

A. Berthod et al. / Talanta 55 (2001) 69–83

74

log c.m.c.= − 0.503nC +1.960 2

n= 5, r = 0.998

(5)

For the Cn E8 surfactants: log c.m.c.= −0.521nC +2.213 n= 8, r 2 = 0.996

(6)

The c.m.c. of Cn Em surfactants decreases exponentially with the number of methylene groups in the alkyl chain. It changes much more slowly, but also exponentially with the number of ethylene oxide in the hydrophilic chain. The c.m.c. increases are observed with short alkyl chains, and c.m.c. decreases are observed with long alkyl chains. The polyoxyethylene dodecyl ether surfactant, C12Em, seems to have a c.m.c. value close to 10 − 4 mol l − 1 whatever the ethylene oxide number is (Fig. 1, slope of 0.0046).

3.1.2. Search for a general equation The analysis of the slopes and intercepts obtained in Eqs. (1)– (6) indicated that a strong correlation exist between the slopes and intercepts of Eqs. (1)– (3) and nC, the number of methylene groups in the alkyl chain. equations 1−3 slope = −0.011nC +0.135 n=3, r 2 = 0.999

(7)

equations 1−3 intercept = −0.433nC +1.117 n =3, r 2 = 0.999

(8)

Combining Eqs. (1)– (3) and Eqs. (7) and (8), we obtain: log c.m.c.= −0.011nCmE +0.135mE −0.433nC +1.117

(9)

Similarly, there is a strong correlation between the slopes and intercepts of Eqs. (4)– (6) and mE, the number of ethylene oxide units in the molecule. Using Eq. (9), we can form: equations 4−6 slope = − (0.433 +0.011mE) (10) and equations 4− 6 intercept =0.135mE +1.117 (11)

Eq. (10) gives the slopes − 0.466, − 0.488 and − 0.521 for m=3, 5 and 8, respectively. These values correspond to the slopes of Eqs. (4)–(6) with an error of 4.3, 3 and 0%, respectively. Since there is a similarly good agreement between the intercept given by Eq. (11) and those obtained in Eqs. (4)–(6), Eq. (9) can be regarded as a general equation giving the c.m.c. of any nonionic surfactant of the Cn Em type within the ranges 5B nB 16 and 3B mB 30. Eq. (9) gives the c.m.c. in moles per litter. The molecular weight of a Cn Em surfactant is simply expressed by: m.w. = 14nC + 44mE + 18

(12)

then, combining Eqs. (9) and (12), the general equation giving the c.m.c. of Cn Em surfactants (g l − 1) is: log c.m.c. (g l − 1) = log(14nC + 44mE + 18)− 0.011nCmE + 0.135mE − 0.433nC + 1.117

(13)

Table 2 presents the theoretical c.m.c. values (mol l − 1 and g l − 1) obtained using Eqs. (9) and (13) for any Cn Em surfactant in the 6B nB16 and 3BmB 30 range. Eqs. (9) and (13) make it mathematically possible to calculate a c.m.c. of 3 mol l − 1 for the C12E4 surfactant. This value would correspond to a 1.8 kg l − 1 solution (m.w., 594). This is not possible. It means that this surfactant is either not soluble or does not form micelles. The data presented in Table 2 were limited to 20% w/w solutions. It is pointed out that this work was carried out by analytical chemists for analytical chemists. The c.m.c. values obtained using Eq. (9) (M) or 13 (g l − 1) are not better than the c.m.c. values obtained using the Degiorgio equation [15]. Eqs. (9) and (13) are more convenient, they just use mE and nC, the number of ethylene oxide and methylene units, respectively. These equations do not pretend to give the exact value of the Cn Em molecule, but rather a good idea of the concentration range of the surfactant micellization.

Table 2 Critical micelle concentration of the Cn Em nonionic surfactants calculated by Eq. (9) (mol l−1) and Eq. (13) (g l−1)a mE

nC

3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 a

6

8

10

12

14

16

M

g l−1

M

g l−1

M

g l−1

M

g l−1

M

g l−1

M

g l−1

M

g l−1

0.46 0.56 0.69 a a a a a a a a a a a a a

90 136 198 a a a a a a a a a a a a a

0.053 0.062 0.073 0.086 0.100 0.118 0.138 0.162 0.190 0.222 0.261 0.305 a a a a

11.8 16.6 22.7 30.3 40.0 52.1 67.1 85.8 108.9 137.4 172.5 215.7 a a a a

6.22×10−3 6.93×10−3 7.73×10−3 8.61×10−3 9.59×10−3 1.07×10−2 1.19×10−2 1.33×10−2 1.48×10−2 1.65×10−2 1.84×10−2 2.05×10−2 2.28×10−2 2.54×10−2 2.83×10−2 3.16×10−2

1.53 2.01 2.58 3.25 4.05 4.98 6.08 7.35 8.85 10.58 12.60 14.94 17.65 20.79 24.41 28.58

7.28×10−4 7.71×10−4 8.17×10−4 8.65×10−4 9.16×10−4 9.71×10−4 1.03×10−3 1.09×10−3 1.15×10−3 1.22×10−3 1.29×10−3 1.37×10−3 1.45×10−3 1.54×10−3 1.63×10−3 1.73×10−3

0.197 0.242 0.292 0.348 0.409 0.476 0.549 0.629 0.717 0.814 0.919 1.034 1.159 1.295 1.444 1.605

8.51×10−5 8.57×10−5 8.63×10−5 8.69×10−5 8.75×10−5 8.81×10−5 8.87×10−5 8.93×10−5 8.99×10−5 9.06×10−5 9.12×10−5 9.18×10−5 9.25×10−5 9.31×10−5 9.38×10−5 9.44×10−5

0.025 0.029 0.033 0.037 0.041 0.045 0.050 0.054 0.058 0.063 0.067 0.071 0.076 0.081 0.085 0.090

9.95×10−6 9.53×10−6 9.12×10−6 8.73×10−6 8.36×10−6 8×10−6 7.66×10−6 7.33×10−6 7.01×10−6 6.71×10−6 6.43×10−6 6.15×10−6 5.89×10−6 5.64×10−6 5.4×10−6 5.16×10−6

0.0032 0.0034 0.0037 0.0039 0.0041 0.0043 0.0045 0.0046 0.0047 0.0048 0.0049 0.0049 0.0050 0.0050 0.0050 0.0051

1.16×10−6 1.06×10−6 9.6×10−7 8.77×10−7 7.98×10−7 7.26×10−7 6.61×10−7 6.01×10−7 5.47×10−7 4.98×10−7 4.53×10−7 4.12×10−7 3.75×10−7 3.41×10−7 3.1×10−7 2.82×10−7

0.000398 0.000409 0.000414 0.000416 0.000413 0.000408 0.0004 0.000391 0.00038 0.000367 0.000354 0.00034 0.000326 0.000312 0.000297 0.000283

A. Berthod et al. / Talanta 55 (2001) 69–83

4

a, Too high a concentration, no micelles or not soluble.

75

76

A. Berthod et al. / Talanta 55 (2001) 69–83

3.2. Molar 6olume The molar volume of a compound is the specific volume occupied by one mole of the compound. Molar volumes are listed in litters per mole or millilitres per gram. The molar volume of a micelle-forming surfactant is very important data because it is used to estimate the volume of the micellar phase. This parameter is difficult to find in the literature.

3.2.1. Study of the literature data Table 1 presents the molar volumes for Cn Em nonionic surfactants found in Refs. [1,5– 7,11– 13]. Here also, principal component analysis shows a strong correlation between the molecular weight and the molar volume expressed in litters per mole. It makes sense that a compound with a higher molecular weight occupies more space. When the molar volume is expressed in millilitres per gram, it corresponds to the inverse of the density of the surfactant. Table 1 shows that all literature values are close to 1 and it is difficult to see any trend in these data. 3.2.2. Search for a general equation The Cn Em nonionic surfactants are made of the alkyl chain and the hydrophilic polyoxyethylene moiety. Looking at the linear alkane density (hexane, 0.661 g ml − 1; heptane, 0.682 g ml − 1; octane, 0.701 g ml − 1, decane, 0.729 g ml − 1; dodecane, 0.748 g ml − 1; hexadecane, 0.770 g ml − 1), it can be estimated that the molar volume of hexadecane (m.w., 226), 0.2935 l mol − 1, corresponds to 16 times the incremental molar volume due to a methylene group, VCH2. This assumption gives VCH2 =0.0183 l mol − 1. Looking in a similar manner at the density of polyethylene glycol (PEG) with a differing degree of polymerization higher than 100, they all have a density of 1.125 g ml − 1. Assuming that the molar volume of PEG 1000 (m.w. 44 020), 39 130 l mol − 1, corresponds to 1000 times the incremental molar volume due to an oxyethylene unit, we obtain VCH2CH2O = 0.03913 or 39.13 l mol − 1. Combining these two partial molar volumes gives: VCnEm =0.0183n +0.03913m [l mol

−1

]

(14)

The density of the micellar phase (g ml − 1), which is also the inverse of the molar volume (ml g − 1), can be expressed by: d= =

1 VCnEm 14nC + 44mE + 18 [g ml − 1] 1000(0.0183nC + 0.03913mE)

(15)

Eqs. (14) and (15) were used to calculate the data presented in Table 3. Fig. 2 shows the evolution of density versus the number of ethylene oxide units (top) or the alkyl chain length (bottom). Table 3 and Fig. 2 show that the density of the Cn Em nonionic surfactant micellar phase is close to unity. This is a good point when these surfactants are used for aqueous emulsion preparation. The surfactant will not contribute to the break of the emulsion by sedimentation. It is a bad point when cloud-point extraction is performed. The density difference between the micellar phase and the aqueous phase may be so low that it is difficult to separate them. It was noted that the separation of the C12E8 micellar phase (theoretical density, 1.010 g ml − 1) at 93°C, 15°C above the clouding temperature, was almost impossible: after strong centrifugation, the two phases co-mingled almost immediately [17]. Once again, Eq. (15) should be used with caution. If the density obtained is well below or above unity, the result can be taken as an indication of the micellar phase position (lighter or denser phase, respectively). If the result obtained is close to unity (within 0.98 and 1.02), no safe conclusion can be drawn: temperature is an important parameter acting on density and, simultaneously, on phase separation.

3.3. Cloud point 3.3.1. Study of the literature data Table 1 shows that the clouding temperature decreases when the alkyl chain length increases. On the contrary, it increases with the number of ethylene oxide units. The regression equation for cloud-point temperature, CP, of the Cn E5 surfactants is: CP = − 6.88nC + 115

n= 5, r 2 = 0.994 (16)

Table 3 Molar volume (l mol−1) and micellar density (g ml−1) of the Cn Em nonionic surfactants calculated by Eqs. (14) and (15), respectivelya mE

nC 4

a

8

10

12

14

16

l mol−1

g ml−1

l mol−1

g ml−1

l mol−1

g ml−1

l mol−1

g ml−1

l mol−1

g ml−1

l mol−1

g ml−1

l mol−1

g ml−1

0.191 0.230 0.267 0.308 0.347 0.386 0.425 0.464 0.503 0.543 0.582 0.621 0.660 0.699 0.738 0.777

1.081 1.088 1.094 1.098 1.100 1.103 1.105 1.107 1.108 1.109 1.110 1.111 1.112 1.113 1.113 1.114

0.227 0.266 0.305 0.345 0.384 0.423 0.462 0.501 0.540 0.579 0.618 0.658 0.697 0.736 0.775 0.814

1.030 1.044 1.054 1.062 1.068 1.074 1.078 1.082 1.085 1.087 1.090 1.092 1.094 1.095 1.097 1.098

0.264 0.303 0.342 0.381 0.420 0.459 0.499 0.538 0.577 0.616 0.655 0.694 0.733 0.772 0.812 0.815

0.993 1.010 1.023 1.034 1.042 1.049 1.055 1.060 1.064 1.068 1.072 1.075 1.077 1.080 1.082 1.084

0.300 0.340 0.379 0.418 0.457 0.496 0.535 0.574 0.613 0.653 0.692 0.731 0.770 0.809 0.848 0.887

0.965 0.983 0.998 1.010 1.020 1.028 1.035 1.041 1.047 1.051 1.055 1.059 1.062 1.065 1.068 1.071

0.337 0.376 0.415 0.454 0.533 0.572 0.611 0.650 0.689 0.728 0.767 0.807 0.846 0.885 0.924

0.944 0.962 0.978 0.990 1.001 1.010 1.018 1.025 1.031 1.036 1.041 1.045 1.049 1.052 1.056 1.059

0.374 0.413 0.452 0.491 0.530 0.569 0.608 0.648 0.687 0.726 0.765 0.804 0.843 0.882 0.921 0.961

0.926 0.945 0.961 0.974 0.985 0.994 1.003 1.010 1.017 1.022 1.028 1.032 1.037 1.041 1.044 1.047

0.410 0.449 0.488 0.528 0.567 0.606 0.645 0.684 0.723 0.762 0.801 0.841 0.880 0.919 0.958 0.997

0.912 0.930 0.946 0.959 0.970 0.980 0.989 0.997 1.003 1.010 1.016 1.021 1.025 1.030 1.033 1.037

A. Berthod et al. / Talanta 55 (2001) 69–83

3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

6

There is no theoretical objection to using Eqs. (14) and (15) for any number of either ethylene oxide units (mE) or methylene groups (nC).

77

78

A. Berthod et al. / Talanta 55 (2001) 69–83

Fig. 2. Plots of the density (g ml − 1) of the Cn Em polyoxyethylene alkyl ether surfactants versus the number of oxyethylene units (top) and versus the alkyl chain length (bottom). Temperature, 20°C.

A. Berthod et al. / Talanta 55 (2001) 69–83

For the Cn E8 surfactant, the corresponding regression equation is: CP = −3.83nC +124

(17) The clouding temperature of a nonionic Cn Em surfactant seems to decrease linearly with the alkyl chain length. There is no linear relationship between the cloud-point temperature and the number of ethylene oxide units at constant alkyl chain length. A second-order polynomial regression is adapted to fit the octyl surfactant. This fitting was poor with the dodecyl surfactant. It was found that the square of the CP temperature was linearly related to mE, the number of ethylene oxide units in the molecule. For the octyl surfactant, the regression line was: CP2 =1843mE −5577

n =5, r 2 =0.999 (18)

For the dodecyl surfactant, the regression equation was: CP2 =1371mE −5551

dure, the best equation fitting the Table 1 cloudpoint temperatures was: CP = 11 (24 −nC)mE + 0.014n 2C − 48

n =8, r 2 =0.970

n =7, r 2 =0.988 (19)

It should be recalled that the cloud-point temperature is very dependent on the surfactant solution, concentration, added salt or solute. The temperatures presented in Table 1 correspond to 1% w/w (10 g l − 1) surfactant solutions.

3.3.2. Search for a general equation Considering the sensitivity of the clouding temperature to small changes in the surfactant solution, it is intended to find a general equation giving an idea of the clouding temperature of a Cn Em surfactant. The very exact value should be within a few degrees of the computed value. More specifically, the equation should say whether the considered nonionic surfactant will have a clouding temperature. For example, the popular Brij® 35 surfactant, C12E23, cannot be use for cloudpoint extractions because its clouding temperature is much higher than the boiling point of water. Considering Eqs. (16)– (19), the general equation should be related to nC and to the square root of mE. Using a four-parameter fitting proce-

79

(20)

Table 4 presents the calculated cloud-point temperatures for the Cn Em surfactants. Fig. 3 (top) shows the three-dimensional plot obtained with Eq. (20) within the ranges 4B nC B 16 and 3B mE B 18. Table 4 and Fig. 3 also show how Eq. (20) should be used. If the calculated CP temperature is higher than 100°C, it means that the surfactant does not present the clouding phenomena in 1% w/w aqueous solution at atmospheric pressure. A classical micellar solution is obtained in the 0–100°C range at atmospheric pressure. If the calculated value under the square root of Eq. (20) becomes negative, it means that the surfactant is either not soluble at 1% w/w or does not present the clouding phenomena. Eq. (20) does not show a direct linear relationship between CP, the clouding temperature, and nC, the alkyl chain length. However, Fig. 3 (bottom) shows that the CP versus nC lines are rather linear except for surfactants with a low number of oxyethylene units. Comparing Fig. 3 and the literature data presented in Table 1, it can be seen that the cloudpoint temperature given by Eq. (20) can be off by several degrees, especially for the surfactants with a long alkyl chain (squares in Fig. 3 (bottom)). The estimated cloud-point values obtained using Eq. (20) compare favourably with those given by the equation proposed by Huibers et al. [14]. The difference is the convenience of Eq. (20) that, once again, uses only nC and mE, whereas the Huibers’ equation uses three topological descriptors that are not obtained obviously looking at the surfactant molecule, but they should be found in physical chemistry tables. However, the Huibers’ equation works for a wide variety of nonionic surfactants (with branched or cyclic alkyl tails, or branched or linear alkyl phenol ethoxylates) [14]. Eq. (20) and Fig. 3 clearly show that cloudpoint temperatures and hydrophile–lipophile balance (HLB) of the Cn Em surfactant are linked. The HLB value of a nonionic surfactant is related to both its alkyl chain and polyoxyethylene unit number. Several equations were proposed, such as [6]:

80

A. Berthod et al. / Talanta 55 (2001) 69–83

Fig. 3. Plot of the cloud-point temperature of the Cn Em polyoxyethylene alkyl ether surfactants. Top: Three-dimensional plot leveled at 100°C; bottom, CP temperature versus the alkyl chain length. Surfactant concentration, 1% w/w.

A. Berthod et al. / Talanta 55 (2001) 69–83

HLB =7 +11.7 log(m.w.E/m.w.C)

(21)

in which m.w.E and m.w.C are, respectively, the molecular weight of the polyoxyethylene (hydrophilic) chain and the alkyl (lipophilic) chain. Eq. (21) was used to calculate the HLB values presented in Table 4. Hydrophilic surfactants have HLB values higher than 15. They also have a high clouding temperature or do not present the clouding phenomena because their clouding temperature is higher than 100°C. Lipophilic surfactants have HLB values lower than 5. They are slightly or not soluble in water or have a clouding temperature below 0°C. Eq. (20) gives an idea of the clouding temperature of Cn Em surfactants with intermediate HLB values.

3.4. Application to cholesterol determination in egg yolk Cholesterol is a lipidic compound of nutritional significance that was related to increased risk of cardiovascular disease. Its determination in foods is becoming a routine analysis [18]. The first step in the analysis of cholesterol is the reproducible

81

extraction from the food sample selected [19]. Fat digestion followed by liquid– liquid extraction involving various solvents is the most commonly used method [20]. Solid-phase extraction was also described [21]. These multistep processes lead to variabilities in the reported cholesterol concentration. Cloud-point extraction can be used to extract cholesterol from food sample.

3.4.1. Cloud-point extraction of cholesterol from egg yolk Cholesterol is a lipophilic molecule, so a surfactant with a relatively low HLB will be able to extract it correctly. Polyoxyethylene dodecyl, tetradecyl and hexadecyl ether (C12Em, C14Em and C16Em surfactants) are possible candidates with HLB values lower than 13 (Table 4). However, the density of the C16Em surfactants and of most of the C14Em surfactants is close to that of water (Table 3). This may render the phase separation difficult as signalled in the literature [17]. So the polyoxyethylene-10-dodecyl ether, C12E10, was selected. Table 4 presents its HLB and clouding temperature as 12 and 95°C, respectively, with a micellar density of 1.025 (Table 3).

Table 4 Cloud-point temperature (°C) and HLB value of the Cn Em nonionic surfactants calculated by Eqs. (20) and (21), respectivelya mE

nC 4

3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

6

8

10

12

14

16

°C

HLB

°C

HLB

°C

HLB

°C

HLB

°C

HLB

°C

HLB

°C

HLB

38.5 62.4 79.5 93.5 B100.0 B100.0 B100.0 B100.0 B100.0 B100.0 B100.0 B100.0 B100.0 B100.0 B100.0.0 B100.0

11.9 13.2 14.2 15.1 15.8 16.5 17.1 17.6 18.0 18.5 18.9 19.2 19.6 19.9 20.2 20.5

28.0 54.5 71.7 85.6 97.5 B100 B100 B100 B100 B100 B100 B100 B100 B100 B100 B100

9.9 11.2 12.2 13.1 13.8 14.5 15.0 15.6 16.0 16.4 16.8 17.2 17.5 17.9 18.2 18.5

10.4 45.2 63.1 76.9 88.6 98.9 B100 B100 B100 B100 B100 B100 B100 B100 B100 B100

8.4 9.7 10.8 11.6 12.4 13.0 13.6 14.1 14.6 15.0 15.4 15.8 16.1 16.4 16.7 17.0

0.0 33.7 53.2 67.3 78.9 89.0 98.0 B100.0 B100.0 B100.0 B100.0 B100.0 B100.0 B100.0 B100.0 B100.0

7.2 8.6 9.6 10.5 11.2 11.9 12.5 13.0 13.4 13.9 14.3 14.6 15.0 15.3 15.6 15.9

0.0 15.6 41.2 56.1 67.8 77.8 86.6 94.6 B100.0 B100.0 B100.0 B100.0 B100.0 B100.0 B100.0 B100.0

6.4 7.7 8.7 9.6 10.3 11.0 11.5 12.0 12.5 13.0 13.3 13.7 14.0 14.4 14.7 15.0

0.0 0.0 24.0 42.2 54.7 64.8 73.6 81.4 88.5 95.1 B100.0 B100.0 B100.0 B100.0 B100.0 B100.0

5.6 6.9 7.9 8.8 9.5 10.2 10.8 11.3 11.7 12.2 12.6 12.9 13.3 13.6 13.9 14.2

0.0 0.0 0.0 20.8 37.4 48.7 57.8 65.6 72.6 79.0 84.9 90.4 95.6 B100 B100 B100

4.9 6.2 7.3 8.1 8.9 9.5 10.1 10.6 11.1 11.5 11.9 12.3 12.6 12.9 13.2 13.5

a When Eq. (20) gave a value higher than 100, B100 was posted; when the equation was in error (negative value under the square root), a zero value was posted.

82

A. Berthod et al. / Talanta 55 (2001) 69–83

The cloud-point temperature of the batch of C12E10 received was exactly 95°C for a 1% w/w solution. The cloud point temperatures of a 20% and 40% w/w solution were, respectively, 96 and 98°C. The addition of sodium sulfate to the 20% w/w solution decreased the clouding temperature. With 0.02, 0.04 and 0.1 M Na2SO4 added to the 20% C12E10 solution, the cloud point temperatures were 85, 81 and 75°C, respectively. With a 0.04 M salt concentration, the cloud-point extraction is easily carried out at 95°C. Then, 2.4 mg (6.2 mmol) pure cholesterol were introduced in a mixture of 1 g C12E10 surfactant ( 1 ml) and 10 ml of 0.04 M sodium sulfate solution. Cholesterol was titrated in the two phases. Ninety-two percent of the cholesterol was found in the surfactant phase and 8% remained in the aqueous phase. This corresponds to a partition coefficient of 92/(8/10) =115. The same amount of pure cholesterol was introduced in a mixture of 2 g C12E10 surfactant ( 2 ml) and 10 ml of 0.04 M sodium sulfate solution. Then, after cloud-point extraction, the cholesterol found in the surfactant phase corresponded to 96% of the introduced amount. Four percent remained in the aqueous phase, giving a partition coefficient of (96/2)/(4/10) =120. This partition coefficient is high enough to allow a good extraction of cholesterol from aqueous-C12E10 surfactant solutions.

3.4.2. Micellar chromatography of cholesterol The interest of MLC is to use water-rich mobile phases. The surfactant selected should be relatively hydrophilic (high HLB) since the elution strength of the micellar phases is adjusted varying the surfactant concentration. C12E23 or Brij® 35 was selected due to the high HLB value (16.2; Eq. (21)) and to the wealth of information found in the MLC literature [13]. The cyano-bonded column and a 0.03 M (36 g/L) Brij® 35 solution with 10% v/v propanol separated cholesterol in 9.6 min (retention factor, k = 4.2). The peak efficiency was only 300 plates, enough to measure accurately the peak area that was used for cholesterol quantitation. The nonionic micellar mobile phase was very gentle, with the enzymatic reactor preserving its activity for weeks. Methanol- or acetonitrile-con-

taining mobile phases made the activity of a similar enzymatic reactor irreversibly disappear within hours [22].

3.4.3. Analytical figures of merit The dynamic range of the method was found to be two orders of magnitude, from 0.1 to 20 mg injected (solutions ranging from 5 to 100 mg l − 1, or from 0.013 to 0.26 mM). The limit of detection was 75 ng injected (or a 3.75 mg l − 1 solution). All the analytical figures of merit were obtained using pure cholesterol. 3.4.4. Quantitation of cholesterol in egg yolk Egg yolk (0.5 g) was submitted to the cloudpoint extraction procedure. The C12E10 extracted phase diluted by methanol was injected in the chromatographic system using the C12E23 micellar mobile phase. The added sodium sulfate favoured the precipitation of proteinaceous material contained in the sample. The cholesterol peak was easily identified, although several other peaks were obtained in the chromatogram. A remarkable constant concentration of 1.22% cholesterol was found in the yolk of four different eggs from the same origin. This value compares well with the FDA value of 1240 mg cholesterol per 100 g egg yolk. It is clear that nonionic polyoxyethylene alkyl ether surfactants can replace organic solvents in several liquid–liquid extraction processes and chromatographic separations. Their low toxicity compared with classical organic solvents is a great advantage. They will be easier and safer to handle in chemical processes, producing less toxic wastes and saving the environment. To optimize and/or rationalize the use of these surfactants, it is important to know their physicochemical properties such as the critical micelle concentration, the molar volume and micellar phase density, and the clouding temperature.

References [1] J. Cross, Nonionic Surfactants, Chemical Analysis, Surfactant Science Series, vol. 19, Marcel Dekker, New York, 1987.

A. Berthod et al. / Talanta 55 (2001) 69–83 [2] C. Scho¨ ller, M. Wittwer, German Patents P. 605,973, P. 667,744, P. 694,178 to BASF, 30 November 1930. [3] Facts & Figures, Chem. Eng. News, 26 June 2000, p. 50. [4] Hair Care Products, Chem. Eng. News, 20 March 2000, p. 15. [5] N. Scho¨ nfeldt, Surface Active Ethylene Oxide Adducts, Pergamon Press, New York, 1969. [6] M. Schick, Nonionic Surfactants, Physical Chemistry, Surfactant Science Series, vol. 23, Marcel Dekker, New York, 1987. [7] V.M. Nace, Nonionic Surfactants, Polyalkylene Block Copolymers, Surfactant Science Series, vol. 60, Marcel Dekker, New York, 1996. [8] N. van Os, Nonionic Surfactants, Organic Chemistry, Surfactant Science Series, vol. 72, Marcel Dekker, New York, 1997. [9] L.J. Cline Love, J.G. Habarta, J.G. Dorsey, Anal. Chem. 56 (1984) 1132A. [10] W.L. Hinze, E. Pramauro, CRC Crit. Rev. Anal. Chem. 24 (1993) 133. [11] H. Tani, T. Kamidate, H. Watanabe, J. Chromatogr. A 786 (1997) 229.

.

83

[12] F.H. Quina, W.L. Hinze, Ind. Eng. Chem. Res. 38 (1999) 4150. [13] A. Berthod, M. Garcia-Alvarez-Coque, Micellar Liquid Chromatography, Chromatographic Science Series, vol. 83, Marcel Dekker, New York, 2000. [14] P.D.T. Huibers, D.O. Shah, A.R. Katrizky, J. Colloid Interface Sci. 193 (1997) 132. [15] V. Degiorgio, in: V. Degiorgio, M. Conti (Eds.), Physics of Amphiphiles: Micelles, Vesicles and Microemulsions, North Holland, Amsterdam, 1985, pp. 303 – 335. [16] N.J. Zoeller, A. Shiloach, D. Blankschtein, ChemTech, March 1996, p. 24. [17] R.P. Frankewich, W.L. Hinze, Anal. Chem. 66 (1994) 944. [18] L.L. Smith, Cholesterol and Auto-oxidation Products, Plenum Press, New York, 1981. [19] C. Seillan, K. Spahis, J.E. Pie, J. Agric. Food Chem. 38 (1990) 973. [20] H.E. Indyk, Analyst 115 (1990) 1525. [21] S.M. Lai, J.I. Gray, J. Agric. Food Chem. 43 (1995) 1122. [22] S. Tomer, J.G. Dorsey, A. Berthod, J. Chromatogr. A, in press, 2001.