244
Anal. Chem. 1984, 56,244-247
same series of experiments ranged from 17% to 20%. Sample Analysis. The feasibility of performing multielement analysis of natural water samples by coupling standard additions with liquid chromatographic separation of PAR chelates is demonstrated in Figure 4. Figure 4A depicts the chromatogram of a river water sample which was collected immediately downstream from the effluent of the local wastewater treatment plant. The sample was made 1.4 X M PAR prior to injection into the liquid chromatograph. Figure 4B shows the chromatogram of the same sample after the addition of metal ion standards. The added metal ion concentrations ranged from 0.11 the 0.61 ppm. Future efforts will concern extending the capability of this approach by lowering detection limits, optimizing the detection modes, using different bonded-phase materials as stationary phase, employing gradient mobile phase delivery, and investigating the separation of PAR-chelates of numerous additional metal ions.
(8) (9) (10) (11) (12) (13) (14) (15) (16) (17) (18) (19) (20)
LITERATURE CITED ( I ) Schwedt, G. Chromatographla 1079, 72, 613-619. (2) Fritz, J. S. Pure Appl. Chem. 1077, 49, 1547-1554. (3) Timerbaev, A. R.; Petrukhin, 0. M.; Zolotov, Y. A. J. Anal. Chem. USSR (Engl. Trans/.) 1081, 36, 811-830.
Willeford, B. R.; Veening, H. J. Chromatogr. 1083, 257,61-88. Cassidy, R. M. "Trace Analysis"; Lawrence, J. F., Ed.; Academic Press: New York, 1961; p 121. Snyder, L. R.; Kirkland, J. J. "Introductlon to Modern Liquid Chromatography", 2nd ed.; Wiley: New York, 1979; Chapter 7. Snyder, L. R.; Kirkland, J. J. Introductlon to Modern Liquid Chromatography", 2nd ed.; Wiley: New York, 1979; Chapter 4. Bond, A. M.; Wallace, G. G. Anal. Chem. 1982, 5 4 , 1706-1712. Bond, A. M.; Wallace, G. G. Anal. Chem. 1983, 55, 718-723. Shih, Y. T.; Carr, P. W. Anal. Chlm. Acta 1982, 742, 55-62. Schwedt, 0. Fresenius' 2.Anal. Chem. 1077, 288, 50-55. Berthod, A.; Kolosky, M.; Rocca, J. L., Vittori, 0. Analusls 1070, 7 , 395-400. Uden, P. C.; Parees, D. M.; Walters, F. M. Anal. Lett. 1075, 8 , 795-805. Gaetanl, E.; Laureri, C. F.; Mangia, A.; Parolari, G. Anal. Chem. 1976, 48. 1725-1727. Guha,R. C.;Carr, P. W. J. Chromatogr. Sci. 1982, 20, 461-465. Shibata, S. "Chelates in Analytical Chemistry"; Flaschka, H. A,, Barnard, A. J., Ed.; Marcel Dekker: New York, 1972; p 1. Anderson, R. G.; Nickless, G. Analyst (London) 1067, 92, 207-238. Lapkowskl, M.; Zak, J.; Strojek, J. J. Electroanal. Chem. 1983, 745, 173-1 80. Roston, D. A.; Kissinger, P. T. Anal. Chem. 1981, 53, 1695-1699. Peters, D. G.; Hayes, J. M.; Hieftje, G. M. "Chemical Separations and Measurements", 1st ed.; W. B. Saunders Co.: Philadelphia, PA, 1974; Chapter 2.
RECEIVEDfor review August 22, 1983. Accepted October 26, 1983.
Determination of One-Carbon to Three-Carbon Alcohols and Water in Gasoline/Alcohol Blends by Liquid Chromatography Mikio Zinbo Ford Motor Company, Research Staff, Dearborn, Michigan 48121
A dlrect llquld chromatographlc method for determlnlng onecarbon to three-carbon (Cl-C,) alcohols and water slmultaneously In gasollne/alcohol blends Is descrlbed. The method employs a mlxed mode of llquld chromatography-Le., slzeexcluslon and adsorption (or afflnlty) chromatography. The separatlon Is performed on elther one or two mlcropartlculate sire-exclusion columns wlth toluene as a moblle phase. The quantlflcatlonof alcohols and water In the effluent Is achleved by a dlfferentlal refractometer at 30 OC. Analytlcal data for a gasollne/ethanol/methanol blend are presented and the proposed method Is applled for determlnatlon of water mlsclblllty of a gasollne/ethanol blend. The lower llmlts of detection for C1-C, alcohols and water are 0.005 and 0.0025 vol %, respectlvely, wlth a maxlmum Injection volume of 200 pL of neat gasohol. The relatlve standard devlatlon Is less than 1% for both alcohols and water In thelr volume ranges of 0.04 to 0.2 pL per analysls.
World-wide trends toward lead-free gasoline in recent years have generated renewed interest in nonlead octane quality improvers, especially oxygenated organic compounds such as alcohols and ethers. Methanol (MeOH), ethanol (EtOH), and/or tert-butyl alcohol (t-BuOH) have been blended in gasoline as fuel extenders in addition to being octane improvers. Hence it is important to check on the fuel property changes, which alter the driveability, emissions, and reliability of existing automobiles. MeOH, EtOH, and t-BuOH in gasoline have been determined by gas-liquid chromatography (GLC) following aqueous 0003-2700/84/0356-0244$01.50/0
extraction of the gasoline/alcohol blends (1). It has been reported, however, that the aqueous extraction-GLC method is satisfactory and quantitative for EtOH, but not for MeOH, 2-propanol (2-PrOH), 1-butanol (1-BuOH), and higher molecular weight alcohols (2). Two direct GLC methods for determination of C1-C4 alcohols in gasoline/alcohol blends have been reported recently: one using a 7 m X 3 mm glycerol column and 1-pentanolas an internal standard (3) and another using a valveless multidimensional switching system with two packed columns (4). Two infrared methods for the determination of EtOH in gasohol have been also reported recently, using the wavelength regions of 6000-7000 cm-' (2) and 800-1200 cm-l (5). The Karl Fischer titration is still the most popular method for the determination of low levels of water in petroleum products. Water in gasoline has been also determined by reaction gas chromatography with a calcium carbide reactor column (6). A reversed-phase high-performance liquid chromatographic procedure has been utilized for determination of water in both liquid and solid samples, following a water/phenyl isocyanate reaction in dried N,N-dimethylformamide (7). In size-exclusionchromatography (SEC) for the separation of polar, low molecular weight substances, the existence of adsorptive solute-substrate and associative solute-solvent interactions has been recognized previously by several workers (8-1 I). Applications of high-performance size-exclusion chromatographic (HPSEC) columns for partition or interactive chromatographic mode by employing mixed solvent of varying polarities have been reported for analysis of a series of aromatic compounds, food additives, and pesticides (12, 13). 0 1984 Amerlcan Chemical Society
ANALYTICAL CHEMISTRY, VOL. 56, NO. 2, FEBRUARY 1984
Table I. Calibration Dataa slope, A RI response/ corr standard fiL coeff
I
0
I
I
I
12
24
36
I
ELUTION VOLUME (mL1
Flgure 1. Chromatograms of alcohol (-) and water (---) standard solutions (procedure 2): (a) t-BuOH, 0.04 pcLL; (b) 2-PrOH, 0.08 pLL; (c) EtOH, 0.04 pL; (d) MeOH, 0.08 pL; (e) H20, 0.0125 pL.
The present paper describes a direct liquid chromatographic method for simultaneously separating CI-Co alcohols and water from gasoline components on HPSEC column sets with toluene as a mobile phase and sample solvent. The proposed method has been successfully applied to a series of gasohol/water mixtures for determining the water miscibility of a gasoline/EtOH blend. EXPERIMENTAL SECTION Materials. Alcohols used as standards and the solvents were of analytical reagent grade and were used without further purification unless otherwise stated. Unleaded gasolines A and B were obtained from two different commercial sources. Apparatus. A Waters Associates (Milford, MA) 150C ALC/GPC liquid chromatograph equipped with an automatic injector (1to 500 pL in microliter increments) and a differential refractometer (ARI) was employed. Two Ultrastyragel column (30 cm X 7.8 mm i.d.) sets, either (1 X 100 %.)or (1 X 500 A) + (1 x 100 A), were used. The mobile phase was toluene (J.T. Baker Chemical Co., Phillipsburg,NJ). The flow rates were 1.0 mL/min for the single-column set (procedure 1) and 1.2 mL/min for the two-column set (procedure 2). The column oven and detector cell temperatures were controlled at 30 f 0.1 "C. The relative sensitivity of the refractometer used was at 32 throughout this work unless otherwise stated. A Waters 730 data module was used for recording chromatograms and for integration of peaks. Standard Solutions. An alcohol standard solution containing MeOH, EtOH, 2-PrOH, and t-BuOH was prepared by placing the measured amounts (50 or 100 pL each) of the alcohols into a 25-mL volumetric flask and filling to the mark with the chromatographic solvent (toluene). The solution was chromatographed by procedures 1 and 2. Typically, the injection sizes of the alcohol standard solution were 20,25, 30,35, and 40 pL for the column set calibration. Standard solutions for water were prepared by placing a series of the measured amounts of distilled water, i.e., 10,25,50,and 75 pL, into 10-mL volumetric flasks, adding 5 mL of "distilled-in-glass" grade acetone (Burdick & Jackson Laboratories Inc., Muskegon, MI), and filling to the mark with toluene. Twenty microliters each of the prepared solutions was chromatographed for calibration. Precision and Accuracy. A sample solution containing MeOH (100 pL), EtOH (50 pL), and gasoline A (2000 pL) in toluene (total volume 25.0 mL) was prepared and analyzed by procedure 2 to obtain the measures of precision and accuracy of the present method. Water Miscibility. A series of gasohol/water mixtures (seven samples) were prepared by placing 1 to 20 pL of distilled water into dried vials (15 X 45 mm) and adding 2.0 mL each of gasoline B. The sample vials were tightly capped with Teflon-lined screw caps, shaken vigorously, and left to stand overnight at room temperature. Then, 500 pL of the mixtures, or the gasoline phases in the case of phase separation, was mixed well with 2.0 mL of toluene in the 150C ALC/GPC sample vials, from which 25 pL
245
lower detectn limits, fiL/ analysis
std error of estimate
2.16 0.005 0.9999 3805 t-BuOH 4.75 0.005 0.9999 3984 2-PrOH 1.55 0.005 0.9999 3773 EtOH 2.28 0.005 0.9999 5062 MeOH 26.38 0.0025 0.9993 9140 H,O a By procedure 2 ; based on ten analyses with the concentration range of 0.04 to 0.16 fiL per analysis ( n = 2 at each concentration), Relative refractometer sensitivity at the setting of 32.
-~
Table 11. Retention Times of Alcohols and Water retention time, min
I _ _ _
t-BuOH 2-PrOH EtOH procedure l a procedure 2 b
11.07 19.05
11.66 19.98
12.46 21.38
MeOH
H,O
13.45 23.06
16.32 29.01
a Single Ultrastyragel column, 100 A , with a flow rate of Two Ultrastyragel columns, 500 A i100 1.0 mL/min. A, with a flow rate of 1.2 mL/min. __-p______I
each of the sample solutions was analyzed. RESULTS AND DISCUSSION HPSEC separations of a mixed alcohol and a water standard solution for calibration are shown in Figure 1, in which two modes of liquid chromatographic separation ranges, Le., SEC and SEC plus adsorption or affinity (AC), are indicated based upon the deviation of elution volumes from a molar volume calibration curve (11). Linear regression analyses of the HPSEC results of the standards by procedure 2 are listed in Table I. The resulting four alcohol calibration curves possess high correlation coefficients and relatively low standard errors of estimate. A slightly larger standard error of estimate for the water calibration curve is probably caused by a result of small background corrections made for unknown ARI responses at the elution volume of water (see Figure 1). Included in Table I are the lower detection limits under the present chromatographic conditions. Retention times of four alcohols and water obtained by procedures 1 and 2 are compared in Table 11. Unleaded gasolines A and B were characterized by both HPSEC and GLC in order to be used as base gasolines in the preparation of synthetic gasohol and gasohol/water mixtures. As shown in Figure 2, gasoline B contained 10.4 vol % EtOH, which was determined by five replicate HPSEC analyses with a relative standard deviation (RSD) of 0.72%. GLC analyses of both gasolines A and B with a packed Silar 1OC column (180 cm X 4 mm i.d.) demonstrated very similar chromatographic peak distribution profiles, indicating the similarity in gasoline components except EtOH in gasoline B. A standard mixture of EtOH and MeOH with gasoline A was prepared for determination of the measures of precision and accuracy of the proposed method for gasohol analysis. The prepared gasohol contained 2.33 vol % EtOH and 4.65 vol % MeOH, which was diluted to 0.086 times with toluene for the HPSEC analysis by procedure 2. The chromatogram of a synthetic gasohol sample is shown in Figure 3 together with that of a corresponding concentration of base gasoline A for comparison. Results obtained for the synthetic gasohol are summarized in Table 111. The percentage relative accuracy of the mean for EtOH is 0.601 and that for MeOH is 0.968.
246
ANALYTICAL CHEMISTRY, VOL. 56, NO. 2, FEBRUARY 1984 W
3
4 k\
c
-/
1
IO
P
w
3
,*---a
5
'L
I 0
1
12
0.0
36
24 ELUTION VOLUME ( m L )
0.2
0.4
0.6
I .o
0.8
WATER V O L . % ADDED TO GASOLINE B
Flgure 2. Chromatograms of unleaded gasolines A and B (procedure 2). Peak identity: gasoline components, elution volumes from 19 to 24 mL; EtOH, elution volume at 25.7 mL.
Flgure 4. Determination of maximum water miscibility in gasoline B and EtOH vol % changes in the gasoline phases of its water mixtures.
I
H
a
Q, v) W
z
2
w v)
a
a
0
c 0 W
+ 0 W
GASOLINE A
--
1
0 0
36
24
12
ELUTION VOLUME (mL)
Flgure 3. Comparative chromatograms of (i) base gasoline A and (ii) synthetic gasohol (procedure 2): (injection size) 20 pL contalning (i) 1.6 pL of gasoline A and 18.4 pL of toluene, and (ii) 0.04 pL of EtOH, 0.08 pL of MeOH, 1.6 pL of gasoline A, and 18.28 pL of toluene; (peak identity) (1) aliphatic hydrocarbons, (2) monomeric aromatic hydrocarbons except toluene, (3) EtOH, and (4) MeOH. I
IO
15
Flgure 5. Comparative estimation of monomeric aromatic hydrocarbons in gasolines A and B (procedure 1): (1) aliphatic hydrocarbons, (2) monomeric aromatic hydrocarbons except toluene, and (3) EtOH.
HPSEC results are summarized in Figure 4. The plot between water vol % added to gasoline B and water vol % found in the gasoline phases clearly demonstrates a maximum water miscibility of 0.3 vol % at room temperature. The procedure can be used to determine the water tolerance of gasoline/alcohol fuel blends a t various temperatures. Furthermore, the proposed HPSEC method can be utilized for a comparative estimation of benzene and alkylbenzenes except toluene in gasoline fuels, which, for example, is shown in Figure 5 . The analysis indicates that gasoline A contains ca. 23 % more monomeric aromatic hydrocarbons except toluene than those in gasoline B (excluding 10.4 vol 5% EtOH), based on the fact that both gasoline A and gasoline B possess very similar detector response ratios among benzene, ethylbenzene, and xylenes by GLC analyses. A single-column set (e.g., 1 x 100 hi) is appropriate for the analysis of gasoline/MeOH blends. However, a three-column set (e.g., (1 x 500 hi) (2 x 100 A)) will be needed for the analysis of gasolinelt-BuOH blends to minimize interference from monomeric aromatic hydrocarbons in the base gasoline. Obviously the recently available microparticulate size-exclusion column materials ( < I O pm) do possess significantly more adsorptive forces between polar solute and substrate in nonpolar solvents such as in toluene than the larger particulate size-exclusion column materials (>30l m ) previously investigated (8-11).
-~___-_--II
Table 111. Percent Alcohols in Synthetic Gasohol injection volume,
&La
alcohol vol % in
I _
sample no.
as toluene soln
1 2 3 4 5
20 25 30 35 40
as gasohol 1.72 2.15 2.58 3.01
3.44
mean RSD a
5
ELUTION V O L U M E (mL)
gasohol ---EtOH
MeOH
2.328 2.329 2.304 2.314 2.303
4.625 4.594 4.583 4.614 4.611
2.316 0.542
4.605 0.364
n = 2 at each injection volume.
-_I_____
A series of seven mixtures of water with gasoline B was prepared in the 0.05-1.0 vol70 range. It was observed for the mixtures that phase separation would first occur when the . amount of added water was between 0.25 and 0.40 ~ 0 1 % The concentration changes of both EtOH and water in the gasoline phases were determined by the proposed method. The
+
Anal. Chem. 1084, 56,247-250
In conclusion, the proposed method is very accurate and precise for determination of C1-CB dcohols in gasoline/alcohol fuel blends. The estimated lower detection limits for C1-C3 alcohols and water are 50 and 25 ppm, respectively, injecting 200 pL (max. 100 p L per column) of neat gasohol samples. It requires minimum sample preparation and yields the simultaneous determination of C1-C3 alcohols and water in gasoline/alcohol blends.
ACKNOWLEDGMENT The author thanks R. E. Baker and R. K. Jensen of the Fuels and Lubricants Department for their helpful discussions with this work. Registry No. Ethanol, 64-17-5; methanol, 67-56-1;2-propanol, 67-63-0; tert-butyl alcohol, 75-65-0; water, 7732-18-5.
LITERATURE CITED (1) Pauls, R. E.; McCoy, R. W. J. Chromatogr. Scl. 1981, 79, 558-561. (2) Wong, J. L.; Jaselskis, B. Analyst (London) 1982, 707, 1282-1285.
247
IDurand, J. P.; Petroff, N. Rev. Inst. Fr. Pet. 1982, 37, 575-578.
Chem. Abstr. 1982, 97, 112187a. Sevcik, J. HRC CC, J . H/gh Resolut. Chromatogr. Chromatogr. Commun. 1980, 3 , 166-168. Battiste, D. R.; Fry, S. E.; White, F. T.; Scoggins, M. W.; McWilliams, T. B. Anal. Chem. 1981, 53, 1096-1099. Konopiynski, A.; Siedlecki, A. J. Chem. Anal. (Warsaw) 1980, 25, 777-781. Chem. Abstr. 1980, 95, 45612d. Bjorkqvist, B.; Toivonen, H. J. Chromatogr. 1979, 178, 271-276. Hendrickson, J. G.; Moor, J. C. J. folym. Sci., Part A 1968, 4 , 167-1 88. Smith, W. B.; Kollmansberger, A. J. f h y s . Chem. 1965, 69, 4157-41 61. Hendiickson, J. 0. Anal. Chem. 1968, 40, 49-53. C a m , J.; Gaskill, D. R. Sixth GPC International Seminar Preprlnts, Miami, FL, Oct 1968, 147-157. McComas, D.; Benson, J. Presented at the Pittsburgh Conference on Analytical Chemistry and Applied Spectroscopy, Oleveland, OH, 1979; paper no. 3. McKay, V.; Stevenson, R. Presented at the Pittsburgh Conference on Analytical Chemistry and Applied Spectroscopy, Celeveland, OH, 1979; paper no. 98.
RECEIVED for review July 22,1983. Accepted October 3,1983.
Determination of Total Phthalate in Urine by Gas Chromatography Phillip W. Albro,* Satldra Jordan, Jean T. Corbett, and Joanna L. Schroeder National Institute of Environmental Health Sciences, National Institutes of Health, U.S. Public Health Service, P.O. Box 12233, Research Triangle Park, North Carolina 27709
Whlle urlne rarely contalns slgnUlcant quantnles of phthalate diesters, It does contaln a variety of metabolites (phthalate monoesters) Including conjugates. I n the absence of baicterlal action followlng excretlon, the metabolltes retain an Intact phthalate rlng. A procedure for the hydrolysls 6f phthalate esters and metabolltes to free phthallc acld, recovery and esterlflcatlon of the acld, and gas chromatographic quantlflcatlon of the product ester all relatlve to an Internal standard of 4-chlorophthalate has been developed. The measurement limit Is 0.5 nmol of total phthalate/mL of urlne, and the relatlve standard devlatlon Is approximately 1.8% for four or more replicates. The assay Is llnear between 0.5 and 50 nmol/mL urlne, which spans the range of phthalate levels found thus far In human urlne samples. The procedure can also be used to determlne levels of lsophthalate and terephthalate slmultaneously wlth phthalate.
Phthalate esters are used primarily as plasticizers for vinyl and related plastics, with an annual production exceeding a billion pounds (1). Dimethyl phthalate is used topically as an insect repellent (21,bis(Zethylhexy1) phthalate is present in vinyl blood storage bags from which it migrates into stored blood (3-5),and several phthalates are found in medical grade vinyl tubing used for infusions and dialysis therapy (6,7). Two of the major sources of phthalates in the home are vinyl floor tiles and electrical insulation, from which the plasticizer slowly volatizes as the products age (2). Although most phthalates have a very low order of acute toxicity, reports that high-level chronic exposure to bis(2ethylhexyl) phthalate causes liver cancer in rats and mice (8) while several phthalates including bis(2-ethylhexyl) and diThis artlcle not subject to
n-butyl (the two most widely used phthalate plasticizers) cause testicular atrophy in rats (9)have stimulated a great deal of concern about the potential hazard of these compounds to humans. Phthalate diesters have been found in a variety of animal tissues (10,II)and food (12,13)but are very rapidly cleared from human blood after exposure (14). In general, intact diesters of the more common plasticizers are not found in urine (except dimethyl phthalate (15)). In most cases only hydrolyzed and oxidized metabolites occur in urine of mammalian species (16);however, mammals appear to be unable to metabolize the phthalate ring system, which is intact in all of the identified metabolites ( 17 ) . Metabolites of phthalate diesters are eliminated from the body in both feces and urine (18),as a mixture of free and glucuronide-conjugated acidic forms in the case of all animals studied extept rats (If?),which do not form conjugates of bis(ethylhexy1) phthalate metabolites. Therefore analysis of clinical samples for phthalate diesters will not effectively reflect exposure. Since a large number of different metabolites are excreted (17),it would be most useful to have a single meaburemknt of total phthalate as an indirect indicator of level of exposure. The ratio of urinary to fecal excretion of phthalate metabolites appears to be reasonably constant for a given plasticizer in a given species but may differ for different plasticizers (15). Therefore, determination of total urinary phthalate can only serve as a relative indicator of exposure and not as an absolute measure. The classical procedure for determination of phthalate in urine (19) involved hydrolysis to phthalic acid, oxidation of interferences with nitric acid, and precipitation of phthalic acid as the lead salt for gravimetric determination. This procedure has obvious limitations in terms of specificity and sensitivity. The method described in the present paper was
U S . Copyright. Publlshed 1984 by the American Chemical Society
