ANALYTICAL CHEMISTRY, VOL 50, NO. 9, AUGUST 1978
Reproducibility. Many other factors such as drop shape, sample volume, and pad shape affect the reproducibility as discussed by Guilbault and Vaughan ( 7 ) . Attention should be paid to placement of the pad inside the cell and to efficient mixing of the reagents before each measurement.
pounds present, which absorb or emit a t the same wavelengths as NADH. would interfere.
ACKNOWLEDGMENT T h e authors thank L. P. Solomonson for his kind and generous supply of Chlorella uulgaris.
CONCLUSION A method for the determination of nitrate in water samples at ppb levels has been developed. T h e advantages of this method over other enzymatic and spectrophotometric procedures are: simplicity-only enzyme, NADH, buffer solution, and substrate are needed for the measurement; rapidity-only 3-5 min is required for each assay; and specificity-most diverse ions do not interfere. Therefore, this procedure can be a very useful method for monitoring of nitrate in contaminated drinking water, as well as in food and other fields. T h e main disadvantages generally associated with most fluorescence methods, Le., (1) errors from fluorescing contaminants such as aromatic organic compounds in the water sample and ( 2 ) sample matrix effects generally caused by spectral and nonspectral interferences associated with t h e matrix, are not as important in solid-surface methods such as the one described herein. Any strongly fluorescing com-
1325
LITERATURE CITED (1) J. C.W. Kuan, H. K. Lau, and G. G. Guilbautt, Clin. Chem.,( Winston-Sakm, N.C.). 21, 67 (1975). (2) H. K. Lau and G. G. Guilbautt. Clin. Chem., ( Winston-Salem, N . C . ) ,19,
1045 119731. - -, (3) L.-P.-Solomonson,K. Jetschmann, and B. Vennesland. Siochirn. Sbphys. Acta, 309, 32 (1973). (4) J. M. Maldonado, J. Herrera, A. Paneque, and M. Losada, Siochem. Siophys. Res. Commun., 51, 27 (1973). (5) "Standard Methods for the Examinationof Water and Waste Water", 14th ed.. Method 419-A, p 420: published by the American Public Health Association, American Water Works Association, and the Water Pollution Control Association, 1975. (6) L. P. Solomonson and B. Vennesiand, Planf Physiol., 50, 241 (1972). (7) G. G. Guilbault and A. Vaughan, Anal. 'Chim. Acfa, 55, 107 (1971). \
RECE~CZ~D for review January 10,1978. Accepted May 17,1978. The financial assistance of the NSF-RANN Food Technology Program (Grant No. AER-76-23271) is gratefully acknowledged.
Determination of Polypropylene Glycol Extracted from Polymers into Food-Simulating Solvents Tore Ramstad," T. J. Nestrick, and R. H. Stehl
Dow
Chemical U.S.A., Midland, Michigan 48640
A method for the determination of polypropylene glycol at sub-ppm levels in aqueous and organic media is presented. Combining the classical Zeisel alkoxyl reaction with highefficiency GLC, the method has been used to determine the amount of polypropylene glycol (P-1200) extracted from several polymers into food-simulating solvents-ethanollwater, acetic acid/water, water, and heptane. Extracted polyglycol is degraded using hydriodic acid to form 2-iodopropane as the major reaction product. The 2-iodopropane is determined either by GLC or by GC-MS. This procedure has produced 75 YO overall conversion (yield) of polyglycol to 2-iodopropane and f 2 0 YO relative standard deviation.
Polypropylene glycol has found use as an additive in a variety of polymers. Its compatibility with the matrix, low volatility, and plasticizing attributes have led to increased use of polypropylene glycols as flexibilizing modifiers. To evaluate the amount of the additive which may migrate from a plastic food package into food, in support of our concerns about product stewardship as well as those of the Food and Drug Administration (I),analytical procedures have been developed t o measure the concentrations of the additive in food-simulating media. Several studies have been published on the analysis of polyglycols at high concentrations, including percent levels. These studies include procedures utilizing nuclear magnetic resonance spectrometry (2),a modified Zeisel procedure for oxyalkylene groups, including polypropylene glycol, in which 0003-2700/78/0350-1325$01 .OO/O
evolved iodine was measured (31,thin-layer chromatography of various derivatives of polyglycols ( 4 ) , and gas chromatography of polyethylene glycols as iheir silyl ethers ( 5 , 6 ) . For low ppm determinations, a n alternative approach is required. [Since completion of this work, a sensitive technique for determining polyglycols by a derivative chronopotentiometric procedure has been published ( 7 ) ] .Alkoxy1 groups have been determined by the classical hydriodic acid decomposition of Zeisel (8). For ethers the Zeisel reaction may be represented as: ROR' 3 RI
+
R'I
4
H,O
Many modifications of the procedure exist (9, I O ) , including t h e use of GLC for determination of the alkyl iodide. The determination of alkoxyl groups in cellulosics by Cobler e t al. (11) was one of the first such applications. Crippen (12) reports a glycol determination on a 0.1-g sample using H I and propionic anhydride followed by gas chromatographic analysis on Poropak Q. Merz (13) published a method for the determination of propylene oxide polymers in which the chief reaction product of the Zeisel reaction is isopropyl iodide (68%) with 22% conversion to propionaldehyde. The volatile isopropyl iodide was collected in m-xylene in a cold trap and then determined chromatographically. More recently Pfifer (14) descrihed the use of Rohm and Haas XAD resins to sorb oxyethylene from water. The oxyethylene was then reacted with HI t o produce ethyl iodide with a final determination by GLC. We report here a microscale Zeisel/GLC procedure for the determination of polypropylene glycol of nominal molecular weight 1200 (P-1200) in four food-simulating E' 1978 American Chemical Society
1326
ANALYTICAL CHEMISTRY, VOL. 50, NO. 9, AUGUST 1978
solvents-water, 1:l (v/v) ethanol/water, 3:97 (v/v), acetic acid/water, and heptane.
Table I. ppm PPG Extracteda from GP Polystyreneb water HO Ac/H, 0 EtOH/H,O
EXPERIMENTAL The decomposition of polypropylene glycol treated with 57 70 aqueous HI was investigated. Known amounts of P-1200 were introduced into ampules of ca. 6 cm3 volume (Kontes Glass Company) along with measured amounts of acid. The sealed ampules were placed in an oven for different periods of time to study conversion efficiencies to the expected halides. Upon removing the ampules, they were cooled, scored, and opened. Using HI the major reaction product is 2-iodopropane (isopropyl iodide). It was established that at the low levels being determined in this work, a small amount of acetic anhydride added to the reaction ampule improved the conversion efficiency. Nine plates of each plastic material (8.9 cm X 5.1 cm x 0.25 cm), separated by glass rods, were placed in 250-300 mL of food-simulating solvent in 16-oz bottles. Extractions into the food-simulating solvents were conducted at 49 "C (120 "F) for periods up to 21 days. Two or three milliliters each of the heptane, ethanol/water, and acetic acid/water extracts and 5 mL of the water extract were withdrawn at designated intervals and pipetted into the reaction ampules. These solutions were then evaporated to dryness under N2 @ 55 "C. To each ampule was added, successively, 1 mL o-xylene, 100 pL HI, and 50 pL acetic anhydride. Following the addition of reagents, the ampules were sealed and placed in an oven at 150 "C for 1 to 1.6 h. Upon completion of the reaction, the ampules were cooled first to room temperature, then in dry ice. The upper layer (o-xylene) was carefully thawed and the top portion of the ampule was washed with the o-xylene by careful tipping. Quickly then, the ampule was scored, opened, and covered with a serum cap. LJpon reaching room temperature, most of the o-xylene layer (containing the isopropyl iodide) was transferred to a 2-dram vial containing 1 d of freshly prepared aqueous K,CO, (4 g in 50 mL) and K,S03 (1 g in 50 mL). These reagents neutralize remaining acid and remove any free I,. An aliquot of the o-xylene layer was injected into a gas chromatograph or into a combination gas chromatograph-mass spectrometer. Reagents. The following reagents were used as received: hydriodic acid (5770),MC&B; acetic anhydride, J. T. Raker Chemical Company; o-xylene, Burdick & Jackson; 2-iodopropane, Aldrich Chemical Company, Inc.; and P-1200, Dow Chemical U.S.A. Gas Chromatography. A Varian 1400 gas chromatograph was used in this work. Columns were packed either with 0.770 OV-275 on 80/100Carhopack C or with Carbowax 20M bonded to 150/200 Porasil E-AW (internal preparation). Operating conditions are given in the caption of Figure 1. Combined gas chromatography-mass spectrometry was performed on a LKB 9000s. The column packing was either 25% DC-200 on 80/lOO Chromosorb W-HP or Carbowax 20M bonded to 115/150 Porasil E-AW (internal preparation). Operating conditions are given in the caption of Figure 2. The concentration of isopropyl iodide in the o-xylene was determined by comparison with isopropyl iodide standards prepared by weight in o-xylene. By subtracting any contribution from blanks (ie., migration solvents alone), the amounts of polyglycols extracted and converted to isopropyl iodide were calculated. Calculations. P-1200 may be represented as: y
3
Ignoring the two end groups (OH and H), we have from the Zeisel reaction,
7% 58
---+
0.50
ppm (pg/mL) calculated for (10 cm3 solvent volume)/ (in.2 surface area) ( 1 5 ) . These are averages of replicate determinations. Addition of P-1200 at the 5% level. a
Table 11. ppm PPG Extracteda from Tyrilb HOAc/
HZ 0 0.079 0.079 0.083
0.076 0.081
av. 0.080
HZO
EtOH/ H2 0
0.12
0.12
0.12 0.10 0.10 0.10 0.11
0.13
heptane
0.14 0.12 0.14
0.054 0.053 0.054 0.053 0.053
0.13
0.053
ppm (pg/mL) calculated for ( 1 0 c m 3solvent volume)/ (in.' surface area) (15). Addition of P-1200 at the 1 % level. a
2 lodamo
0
1
2 3 4 5 T me Minuter
6
Figure 1. Gas chromatogram of EtOH/H,O extract of polystyrene. The 2iodopropane peak is equivalent to 0.5 ppm PPG in the extract. GC conditions: 2 m X 3 m m glass packed with Carbowax 20M bonded to 150/200 Porasil E-AW; helium carrier at a flow of 75 cm3/min; column temp., 7 0 OC; injector temp., 150 OC; FID detector temp., 250 OC; afs sample size, 2 PL;sensitivity, 2 x IO-'*
amount of 2-iodopropane could then he calculated from the determined yield. The percent conversion of P-1200 to isopropyl iodide was approximately 75'70 although it ranged from 6G80%. As a check on conversion reproducibility, a known amount of P-1200 was included in each set of Zeisel reactions; all data subsequently reported are corrected for the experimentallydetermined conversion yields.
RESULTS
HO-KH-CH,-Oh,H
-CH-CH,-O-
0.08 0.10
7H3 I-CH-CH3
170
Hence, for 100% conversion, 1 g of P-1200 yields -(170/58) x 1 = 2.93 g of 2-iodopropane. Since 2-iodopropane standards were available, the actual yield of the Zeisel reaction could be determined and did not require the use of an identical sample of P-1200. The concentration of P-1200 corresponding to a measured
In Table I are shown the levels of polypropylene glycol (PPG) found in extracts of General Purpose (GP) polystyrene after 21 days. Figure 1 shows the gas chromatogram of the EtOH/H,O extract of polystyrene. Table I1 shows the results of five separate determinations of PPG extracted from Tyril (Trademark of The Dow Chemical Company), a copolymer of styrene and acrylonitrile, after 10 days. The data of Table I1 show excellent reproducihility for this procedure. Single-ion chromatograms of a HOAc/H,O blank a n d a 10-day HOAc/H,O extract of Tyril are shown in Figure 2; all data are corrected for blanks. Results for high impact polystyrene
ANALYTICAL CHEMISTRY, VOL. 50, NO. 9, AUGUST 1978
1327
should also serve to minimize losses of the volatile iodide when the ampule is opened. The use of a solvent in a sealed ampule precludes lower boiling solvents, since their vapor pressure would be excessive a t the 150 "C needed for reaction. The major reaction sequence presumed to occur with HI is as follows (3):
/Extract
Q
2-IODOPROPANE
b c
FH3 HOfCH-CH,-Ot;H
2 x HI
I
y
$H3 I--CH-CHz-I
4~
+ ( x - 1 ) H,O
CH3
CH3 decornp.
I-CH-CH2-I
I
HC=CH, +
12
H
3
HC=CH,
Blank
+
+
HI
4 ~ ~ c - L - c ~ ~ I
0
1
2
3
Time, M i n u t e r
Figure 2. Single-ion mass chromatograms of HOAc/H,O blank and extract of Tyril, mass = 170. The 2-iodopropane peak is equivalent to 0.11 ppm PPG in the extract. GC-MS conditions: 2 m X 3 mm glass packed with Carbowax 20M bonded to 115/150 Porasil E-AW; helium carrier at a flow of 40 cm3/min; column temp., 135 OC; injector temp., 175 OC; separator temp., 150 O C ; ion source temp., 250 OC; sample size, 2 1 L
Table 111. ppm PPG Extracteda from HIPSb H, 0
HOAc/H,O
EtOH/H,O
0.77 0.82 0.87 0.86 0.81 0.83
1.76 1.73 1.78 1.81 1.80 1.78
0.73 0.70 0.83 0.76 0.83 av. 0.77
ppm (pg/mL) calculated for (10 cm3 solvent volume)/ Addition of P-1200 at the 5% level. a
(in.2 s u r f a c e a r e a ) ( 1 5 ) .
(HIPS) after 25 days are shown in Table 111. DISCUSSION I t was indicated t h a t solvent blanks must be included in the analysis because of possible propylene glycol or propyl iodide contamination. I t was discovered that the source of reagents is important: the blank using acetic anhydride catalyst which had been kept in a polyethylene bottle (Nalgon) was four times that of the blank using acetic anhydride kept in a glass bottle. In the latter case, the residual blank may have been due to some propyl iodide in the hydriodic acid. Nevertheless, for these solvents, the magnitude of the blank did not exceed one-tenth the amount of polyglycol originating from the polymer. Conversions of P-1200 to 2-iodopropane were approximately 75%. Essential to high conversion efficiency is the inclusion of a solvent in the reaction ampule prior to carrying out the reaction. In the early stages of our work, a solvent was added after scoring and opening of the ampule, just prior to determination by GC. It was not until a solvent was added to the ampule before reacting the contents that conversion efficiency to the expected halide exceeded 25%. (We also investigated the decomposition of P-1200 using HBr. In this case, conversion never exceeded 25% ). We believe that the effect of this solvent is twofold: to retain the propyl iodide as it is formed, promoting further reaction, and to reduce the chance of disproportionation to propane and propene (3). It
In (a) the oxypropylene units comprising the polymer are cleaved, reacting to form 1,2-diiodopropane (propylene iodide). The latter compound is unstable, being decomposed by both heat and light. Vicinal deiodination leads to propylene (b) which, in the presence of excess H I undergoes largely Markovnikoff addition to 2-iodopropane. As the conversion efficiency of P-l2(X)to 2-iodopropane was normally about 75%, side reaction(s) are probably occurring. (Increased reaction time beyond 0.5-1 h did not increase the yield.) Siggia et al. (3) postulate the reaction of 1,2-diiodoethane with HI to form ethyl iodide. The equivalent reaction for 1,2-diiodopropane would be: yH3 21-CH-CH2-I
y
3
t 2HI-*i-CH-CH3
-
CH3CH2CH,i
t 212
Under conditions where the 1-iodo- and 2-iodopropanes are resolved, there is relatively little 1-iodopropane formed. We conclude from this observation t h a t the formation of 1iodopropane does not subtract significantly from an overall yield to 2-iodopropane. In summary, the modified Zeisel/ G I L procedure developed yields a total figure for oxyalkylene units. Hence, it should be applicable to the determination of both high and low molecular weight glycols for these kinds of studies. Other possible applications include: (1)traces in waste water; (2) determination in biological materials; and (3) environmental distribution studies. LITERATURE CITED (1) Regulation 21 CRF 171.1-"Petitions", Bureau of Foods, Food and Drug Administration, Department of H e a h , Education, and Welfare, Washington, D.C. 20020. (2) A . Mathias and N. Mellor, Anal. Chem., 38, 472 (1966). (3) S. Siggia, A. C. Starke. Jr., J. J. Garis. Jr., and C. R. Stahl, Anal. Chem., 30, 115 (1958). (4) L. Favretto, L. Favretto Gabrielli, and G. Pertoldi Marletta, J. Chromtcgr., 66, 167 (1972). (5) M. K. Withers, J . Gas Chromatogr., 6, 242 (1968). (6) P. Holmqvist, Anal. Chim. Acta, 89, 315 (1977). (7) J. Tornquist, Acta Chem. Scand., 21, 2095 (1967). (8) S. Zeisel, Monatsh., 6, 989 (1885). (9) D. Grun and F. Bochisch, Ber., 41, 347'7 (1908). (10) P. W. Morgan, Ind. Eng. Chem., Anal. Ed., 18, 500 (1946). (11) J. G. Cobler, E. P. Samsel, and G. H. Beaver, M a n t a , 9 , 473 (1962). (12) R. C. Crippen, "Identification of Organic Compounds with the Aid of Gas Chromatography", McGraw Hill Book Company, New York, N.Y., 1973, pp 194-196. (13) W. Merz, Z. Anal. Chem., 62, 232 (1967). (14) L. H. Pfier, 26th Pittsburgh Conference on Analytical Chemistry and Applied Spectroscopy, Cleveland, Ohio, March 3-7, 1975. (15) "FDA Guidelines for Chemistry and Technology", Requirements of Indirect Food Additive Petitions, Bureau of Foods, Food and Drug Administration, Department of Heatlh, Education, and Welfare, Washington, D.C. 20204, March 1976.
RECEIVED for review January 16,1978. Accepted May 5,1978.
