2
n C
n
-1 0 I
-I
-4
-3
-2
.I
0
I
Log (phen), Figure 3. Extraction of zinc with 8quinolinol and 1 ,I 0-phenanthroline in the presence of 0.06M sodium perchlorate a i pW 4.85
LITERATURE CITED
u. (8-quinolinol)o = 0.05M
b. (8-quinolinol)o = 0.0026M Slope of log D vs. log (phsn)o: a, 1.0;
Zn phensf2, 2c104-, which might be expected from the larger value of its conditional stability constant than that of ZnOsz. Rather, as shown in Figures 3 and 4, the addition of phenanthroline results in the formation of a mixed 8-quinolinol-phenanthroline zinc complex, ZnOx phen+, Clod- (slope log D vs. log Ox or log CIO4- should each be 1.0 for this composition). The driving force of the formation of the mixed ligand complex would seem to be the more favorable extraction characteristics of the mixed chelate complex, On the basis of these findings, we are currently investigating the possible occurrence of mixed ligand ccmplexes in a variety of chelate extraction systems.
b, 1.2
ment with the observed slopes of 2.5 and 1.8. The addition of phenanthroline to the extraction system Zn(II)-$-quinilinol-CHCls does not result in either the formation of a phenanthroline adduct of Zn(Ox)z, as our experience with pyridines might lead us to expect. Nor does it lead to the transformation to
(1) Chou, F.,Fernando, Q., Freiser, H., ANAL.CHEM.37,361 (1965). (2) Chou, F., Fresier, H., University of Arizona, Tucson, Ariz., unpublished work, 1966. (3) Qucret, L., Anal. Chim. Acta. 20, 560 (1959). (4) Gahler. A. R., ANAL,CHEM.26. 577 (1954). ’ (5) Sillen, L. G., Martell, A. E., “Stabilitx Constants of Metal-Ion Complexes, See. 11, The Chemical Society, London, 1964. (6) Tachibana, K,, Mem. Fac. Sci., Kgushu Univ. Ser. C, 4,229 (1961).
”! I
-4
-3
-2
-1
0
I
2
L o g ( Eq u i n 01 i no1 lo
Figure 4. Extraction of zinc with 0.0005M 1.1 0-phenanthroline and a varying amount of 8-quinolinol in the presence of 0.06M sodium perchlorate pH = 4.85,(Zn) = 10-6M,slope = 1.0
(7) Yemamoto, Y., Kinuwaki, S., Bull. Chern. Soc. Japan 37,434 (1964). (8) Yamamoto, Y., Kotsuji, K., Ibid., 37,785 (1964). FA-CHUNCHOU HENRY FREISER Department of Chemistry Vniversity of Arizona Tucson, Ariz. 85721
Ion Flotation Method for Analysis of Some Cationic und Anionic Surfactants below Critical Micelle Concentration SIR: The method described here was developed from a suggestion by Tomlinson and Sebba (11). The principle of the technique (9) was based on the removal of a charged dye from solution by ion flotation using the surfactant ion as collector. Their method was limited to the analysis of anionic surfactants, in particular, the oleate ion. The method has been improved, used for the determination of the laurate ion, and extended to the determination of cationic surfactants. This method depends for efficiency on the absence of micellar or colloidal aggregates and for this reason the surfactant solutions used have concentrations below the critical micelle concentration (c.m.c.). The 0.m.c. values of the surfactants used are given in Table I together with the appropriate references. The use of dyes for the determination of charged surfactants is not new, the 1926 *
ANALYTICAL CMEMISTRY
methods differing from the one to be described in the technique of renioval of the dye/surfactant complex from the solution. Few and Ottewill (6) used a method in which the surfactant/dye complex is quantitatively extracted into an organic phase. Burger (a) developed a similar method in which the surfactant to be determined was precipitated with either methylene blue or
Table 1.
C.M.C. Values c.m.c. (molarity) Ref.
Surfactant Potassium laurate Dodecylammonium chloride Dodecyl trimethyl ammonium chloride Cetyl trimethyl ammonium chloride Laur 1 pyridinium ch%ride
0.020 0.013
0.062
(9) (9) (8, 10) (8)
0.0008
(8)
0.0028 0.0044
(4)
0.030
(1)
pyrocatechol violet and dissolved in alcohol. This was then analyzed spectrophotometrically. In addition, Gregory (7) Bas recently described a method for the determination of residual anionic surface-active reagents, in which the cationic copper(I1) triethylenetetramine complex reacts with anionic surface-active agents to give an adduct that can be extracted inLo an isobutanol-cyclohexane mixture. The copper associated with the surfactant ion is determined photometrically as the colored complex using diethylammonium diethyldithiocarbamate. EXPERIMENTAL
Materials. Lauryl pyridinium chloride was obtained from Hooker Chemical Corp., Niagara Falls, N. Y., and was purified by recrystallization from 9.570 ethanol, absolute alcohol with an activated charcoal treatment, and acetone, successively.
~~
Potassium laurate was prepared from coinmercially pure lauric acid from Armour and Co. and Analar potassium hydroxide, recrystallized from absolute alcohol and washed with acetone. All other surfactants used were obtained from Armour and Co. and recrystallized from absolute alcohol. The dyestuffs used were British Drug House samples. Negative Surfactant Ion Analysis. I n a study using potassium laurate as surfactant, a positively charged dye had to be used in the analysis, the one chosen being crystal violet. I n alkaline or weak acid solution, this purple dye exists as the singly charged ion (6):
The procedure for the preparation of the calibration curve was as follows: 400 mg. of crystal violet was dissolved in a minimum volume of ethanol and the solution made up t o 1 liter with redistilled water, enough ammonium hydroxide being added to ensure an alkaline solution. This solution was then placed in a dark bottle and stored in a cupboard away from light for approximately 16 hours to ensure homogeneity. The use of redistilled water and a dark bottle is important, because crystal violet (as well as the other dye used) rapidly loses its color on exposure t o light or oxidizing agents, this decay being approximately 1% per hour in the case of the aboTe solution. For this reason, the analysis was performed in as short a time as possible. In addition, crystal violet undergoes fading in alkaline solution (3) and thus, pH values should not be too high. Trro hundred and fifty milliliters of solutions containing known amounts (0, 10, 20, 30 and 40 mg./liter) of potassium laurate were taken, to which were added 5 ml. of 0.5N sodium bicarbonate to ensure an alkaline solution (pH 8 3 , . This is important because, in addition to the dye becoming less stable, potassium laurate hydrolyzes in acid media. To these solutions 50 ml. of the crystal violet solution was added, and after an induction period of 5 minutes, the mixture was transferred to the ion flotation cell consisting of a sintered-glass Buchner funnel of approximately 14 cm. diameter and porosity KO.4. This cell was attached to a compressed air line with a manometer and Venturi flowmeter; bubbles were generated which buoyed the product between the dye and the surfactant to the surface. There it produced a small froth which broke almost at once t o produce a scum. To ensure that this scum did not reenter the solution, it was sucked off through a glass nozzle attached, through a flask, t o a water suction pump. Because the scum was removed from the surface, the speed of bubbling, which tended to agitate the surface, could be reasonably fast, provided the bubbling was continued
Analytical Data for Solutions of Surfactants A Dodecyl Cetyl Dodecyltrimet'hyl trimethyl Lauryl Potassium ammonium ammonium ammonium pyridlnium laurate chloride chloride chloride chloride
Table 11. Concn. of surfactant, mg./liter 50 40
0 : 030 0.298
0.092 0.239 0.370 0.502
until all the surfactant had been removed. This was determined by bubbling for increasing lengths of time until the absorbance of the residual solution remained constant. Fifteen minutes of bubbling at a flow rate of 250 ml. per minute and a manometer reading of 150 mm. mercury was sufficient. Too great a bubbling rate, however, may cause the funnel t o break, because such funnels are constructed to withstand downward pressure and not upward force. For this reason. a safety valve was installed initially, consisting simply of a tube from the air supply passing into a gas-jar of mercury to a depth of approximately 200 mm. Immediately after aeration, a pipet was inserted below the surface and cleared by blowing lightly through it. Twenty-five milliliters of solution was then removed and diluted t o 250 ml. The absorbance of this solution was measured in a 1-cm. cell in a Unicam spectrophotometer a t a wavelength of 590 nip where crystal violet has a pronounced spectral absorption peak. Unfortunately, crystal violet tends to adsorb on glass and hence, inconsistent results were obtained initially until sufficient dye had been so strongly adsorbed on the glass that it could not be removed with ethanol. I t is probably for this reason that the calibration curves for negative surfactant ion analysis didnot attain the high degree of precision observed in those for positive surfactant ion analysis. Other cationic dyes were tried-e.g., malachite green and methylene blue-but no advantages over crystal violet were apparent. Positive Surfactant Ion Analysis. I n this case, an anionic dye had to be used and several were investigated. These were bromphenol blue (BPB), bromcresol green, potassium indigo tetrasulfonate, and potassium indigo disulfonate, of which B P B was found t o be the best. This dye (shown below) develops an alkaline purple color over the pH range 3.0 t o 4.6 and, in fact, the absorbance of a solution of BPB decreased sharply at a pH of approximately 6.0.
0,205 0.323 0,440 0.565 0 675
0.265 0.362 0.464 0.561 0.666 0.760
0,275 0.381 0.487 0.593 0.702 0.797
As was the case with crystal violet, this dye fades in alkaline solution and should not be stored at a high pH. The surfactants investigated were dodecylamnionium chloride, dodecgl trimethyl ammonium chloride, cetyl trimethyl ammonium chloride, and lauryl pyridinium chloride. The general technique for the preparation of the calibration curve was the same as previously, different quantities, however, being used. Thus, a solution of 800 mg./liter BPB was prepared and only 25 ml. of th'is was added to the 250-ml. quantities of surfactant solution. Ten - milliliter quantities of buffer solution were added, the pH of which depended upon the pH of the solution to be determined. For example, analysis of the dodecylammonium ion, which hydrolyzes in the alkaline region, was performed a t a pH of 6.5 utilizing a sodium hydroxide/ potassium dihgdrogen phosphate buffer, whereas analysis of the potassium laurate ion which hydrolyzes in acid solution, took place at, a pH of 10.0 in a sodium bicarbonate solution. As before, 25-ml. aliquots were drawn off at the end of the float and diluted to 250 ml. for absorbance measurements. A calibration curye for the analysis of dodecglammoniuni chloride in the range 0 to 5 mg./liter was also prepared. I n this case, control of such factors as flow rate and float time was of great importance and any slight fluctuations in either appeared to affect the curve adversely. The technique and conditions were the same as before, it being found, however, that with solutions of dodecylanimonium chloride of 0 to 5 mg./liter, 25 ml. of a solution of 60 mg./liter BPB Tvas convenient. RESULTS AND DISCUSSION
Results for the determination of the various surfactants studied are shown in Table 11. By plotting surfactant concentration against absorbance, straight lines were obtained. The relative standard deviation of points on the potassium laurate curve varied from &5y0to i87& in strong contrast to the relative standard deviations for the positive surfactants which were all about 1%. The large increase in the precision VOL. 38, NO. 13, DECEMBER 1966
e
1927
Table Ill.
Dodecylammonium Chloride Analysis
Concn. of dodecylammonium chloride, mg./liter
A
0.244 0.340 0.439 0.530
0,686
0,624
of the positive surfactant determinations is considered to be due to the lack of adsorption of bromphenol blue on the sintered-glass plate, the dye being removed easily after each float by washing with ethanol. The reproducibilities above agree favorably with those obtained in the dye method of Few and Ottewill (6), who obtained values of 1 2 % to 1 5 7 , at concentrations of 10-Pll.
Gregory (7) reported a relative standard deviation of approximately 1% for the determination of oleic acid in a 4.74 p.p.m. solution. This author reports, however, that when lauric acid was investigated, much lower recoveries were obtained. The results for determining dodecylammonium chloride in the concentration range 0 to 5 mg./liter are shown in Table 111. This calibration curve tailed off a t low concentrations, which could be due to a small amount of inherent surface activity of the bromphenol blue. The calibration curves above were prepared over the concentration range 0 to 50 mg./liter (0 to f 2 X 10-41V) because this was convenient for the analysis of the unknown solutions used. However, calibration curves over greater concentration ranges can be prepared by increasing the amounts of dye added. For calibration curves over smaller concentration ranges, attention must be paid to variables such as float time and gas flow rate which become important.
LITERATURE CITED
( I ) Brown, -4.S.,et al., J . Phys. Chem. 56, 701 (1952). ( 2 ) Burger, X., 2. Anal. Chem. 196, 15 i1962). \ - - - - I -
(3) Chen, D. T. Y., Laidler, X. J., Can. J . Chem. 37, 599 (1959).
(4) Cushman, A., Brady, A. Pa,McBain, J. W., J . Colloid Sci. 3, 425 (1948). ( 5 ) Fern, A. V., Ottewill. R. H.. Ibid.. 11, 34 (1956).‘ ( 6 )7Finar, I. L., “Organic Chemistry,” 101. 1, 759, 3rd ed., Longmans, London. 1959.
(7) G&ory, G. R. E.C . , Analyst 91, 251 (1966). 3) Ralston, A. W.,Hoerr, C. W., J. Am. Chem. SOC.64, 772 (1942). 3) Sebba, F., “Ion Flotation,” 1st ed., Elsevier, Amsterdam, 1962. 10) Tamamushi, B., Tamaki, K., Trans. Faraday Soc. 5 5 , 1007 (1959). 111 Tomlinson. H. S.. Sebba., F.., Anal. Chim. Acta 27, 596 (1962). IT.&I. LOVELL FELIXSEBBA Chemistry Department Universit of the Witwatersrand Johanneszur g Financial assistance to V. RI. Love11 from the Kational Institute for Metallurgy, Johannesburg.
Direct Gas Chromatographic Determination of Isopropyl N-( S c h lorop he nyl)ca rb a mate (CIPC) SIR: An analytical method was required for determination of the herbicide, isopropyl N-(3-chlorophenyl)carbamate (CIPC). This material may be encapsulated to facilitate its controlled release into the soil. Since various experimental materials and conditions are used in preparing the encapsulated herbicides, a rapid analytical procedure to determine CIPC mas needed. Most methods for pesticide analyses deal with trace quantities of material, thus requiring extremely sensitive detecting devices-e.g., colorimetric, hydrogen flame, or electron capture. The prime consideration of this method is the determination of the encapsulated herbicide at levels of 50 to 7570 CIPC. Methods available in the literature make use of the acid hydrolysis of the phenyl carbamate with subsequent colorimetric analysis of the resultant aniline derivative (1,S). The method of Gutenmann and Lisk ( 2 ) utilizes a gas chromatographic separation with electron affinity detection of the brominated hydrolyzate of CIPC. An attempt a t the direct chromatographic separation of the CIPC was made in the hope of obtaining a onestep analysis of the extracted herbicide. Initial experiments using packed columns of 2 meters or more and temperatures exceeding 230’ C. caused degrada1928
a
ANALYTICAL CHEMISTRY
0
I
2
3
4
5
6
7
TIME ( MINUTES 1
Figure 1.
Chromatographic separation of CIPC
tion of the CIPC. The resulting chroEXPERIMENTAL matograms contained many peaks of the Apparatus. A gas chromatograph pyrolytic products of the material. equipped with a tungsten wire thermal By lojvering the operating temperature conductivity detector was constructed in the laboratory and used for the below 2000 C, and using a 4-foot by 1/4-inch with 15% ~ ~ 3 3 0 analyses. Chromatogranls were recorded on a Leeds and Worthrup on Chromosorb IT, separation of the Speedomax H recorder 0- to l-mv. CIPC was attained without apparent response, ~h~ column used was a pyrolysis. Precision and linearity of 4-foot by l/4-in& aluminum tube concentration versus detector response packed in the usual manner with 15% were good as is shown in the results. SE30 on Chromosorb X, 60/80 mesh.
