pH. A 2.00-ml aliquot was worked up as described above, yielding a deep blue 25-ml sample that gave an absorption number of 575, a little high for best use of the method. When the work-up was repeated with a 1.00-ml aliquot, the absorption number was 280, representing 0.154 mg of silicon. The unknown, therefore, had a total of 7.7 mg of silicon or 7.4% based on specimen weight. I t is not practical to take more than a 5.0-ml aliquot of the 50-ml digested unknown solution. If this amount yields too little absorption, it may be necessary to start again with a larger piece of fabric. A reliable assay for silicon can be obtained by this method for as little as 0.2% and a sure qualitative indication for 0.1%. Interference. This investigation did not include any study of interference, since it was applied only to fabric samples of known chemical composition, free of extraneous finishes. Phosphate has frequently been mentioned in the literature (7, I O ) as a possible complication, but Foulger concluded that phosphomolybdate does not give a blue reduction product when treated with sodium sulfite in the presence of acetic acid (6). Kahler (8)confirmed that in the p H range of 2.4 to 2.7 there is no significant interference by phosphate or iron. Nevertheless, Feigl ( I O ) recommends removing phosphoric and arsenic acids before carrying out a spot test for silicon involving molybdate and benzidine. Preparation of Silicon Impregnated Fabrics. This method was developed to monitor a laboratory program in which new routes to permanent silicone and silane attachments to various textiles were being explored. In a typical preparation, a doubleknit polyester specimen, approximately 15 X 20 cm, was extracted with toluene, dried, and weighed a t 8.72 g. I t was immersed for one minute in a solution of 150 g of vinyl triethoxysilane (Union Carbide Corp., catalog No. A-151) in 500 ml of toluene a t l l O o . The reagent solution was boiling and had been under reflux for a t least four hours prior to the fabric dip. The sample was dried in air, then washed twice with detergent in a standard home laundry cycle. When dried, it weighed 9.51 g. The silane weight add-on was 9.1%, corresponding to 1.5% silicon if the reagent remained intact. Colorimetric analysis gave 1.93% and gravimetric 2.30% (see Table I, sample 6) suggesting that some ethoxy groups in the silane were lost
Table 11. Dudicate Ashines - on Laundered Fabrics, % Silicon Cotton (1) (2) Average (3) g r a v i m e t r i c Polyester (1)
(2) Average
Unwashed
Z x LVashed
l o x Washed
3.72 4.15 3.93
3.83 3.75 3.79
3.75 3.62 3.69 3.62
2.42 2.42 2.42
2.42 2.75 2.59
2.74 2.60 2.67
en route. The different results of the two methods could be entirely due to non-uniform impregnation of the sample. I t was of interest to note that although the silane polymer could be gradually extracted with hot organic solvents (and was therefore not “grafted” in a true chemical sense), it was almost indefinitely stable to aqueous laundering. The sensitivity of the method was such that a slight increase in silicon content was detected on an unhemmed spun polyester sample that was washed ten times with detergent (below). I t is believed that non-silane coated fibers were preferentially lost in laundering. Reproducibility of Method. Sets of duplicate wet ashings and colorimetric determinations were carried out on samples of cotton and spun Dacron (100% polyester) that had been impregnated with vinyl triethoxysilane. Specimens for analysis were taken before the first wash and after the second and tenth launderings with household detergent. The largest discrepancy between any of the six pairs of results was 12% (Table 11). When averaged, they indicate a gradual decline in the silicon content of the cotton with washing and a similarly gradual increase in the polyester assay. Besides this evidence of internal consistency, there is a match with the gravimetric analysis: 3.69% us. 3.6296 for the 1OX washed cotton (also in Table I, sample 10).
RECEIVEDfor review May 7,1974. Accepted July 22,1974.
Determination of Stearic Anhydride in the Presence of lsopropenyl Stearate Michael F. Kozempel and James C. Craig, Jr. Eastern Regional Research Center, Agricultural Research Service, U S . Department of Agriculture, 600 E. Mermaid Lane, Philadelphia, Pa. 19118
Work is in progress in this laboratory to develop a commercially feasible process to make isopropenyl stearate (IPS) from stearic acid. Stearic anhydride is an undesirable reactor by-product. One constraint imposed by the process is that the stearic anhydride concentration be equal to, or less than, 2% by weight in the catalyst free product stream. The current method of determination for stearic anhydride, IR, is neither sufficiently accurate nor precise in this concentration range.
Johnson and Funk ( I ) reported a potentially suitable method for anhydride determination. They reacted morpholine with the anhydride and then back titrated the unreacted morpholine with HCl. They stated that ketene, diketene, and acid chlorides could interfere. The chemistry of isopropenyl stearate as an acylating agent is sufficiently similar t o these compounds that it could be expected to in( 1 ) J. 6.Johnson and G. L.
Funk, Anal. Chem., 27, 1464-5 (1955).
ANALYTICAL CHEMISTRY, VOL. 46, NO. 13, NOVEMBER 1974
2063
Table 111. Experimental Design 2 and ResultsMethanol Based Reagents
Table I. Determination of Exploratory Samples of StzOa in IPSband/or Stearic Acid
Design 2, grams
it'eighed, g IPS
Stearic acid
...
5.01 5.00 5.03 5.02 5.02 5.06
... ...
... ...
0.50 0.51
. ..
...
5.02 a
Determined, g
T i m e on steam bath,
St20
St20
min
0.21 0.59 0.05
0.21 0.62 0.10 0.01 0.33 0.39 0.00 0.02
12 12 12 12 15 15 15 15
...
0.28 0.28
... ...
A = IPS" B = Stearic acid
c = St20b
-1
5.0 0.5 0.3
0 0 0
Results
1
A B C AB AC
Stearic anhydride. b Isopropenyl stearate.
Table 11. Experimental Design I and ResultsButanol Based Reagents
B C ~.
Design 1, grams *1
+1
ABC
Response, m l HCl
Variance
46.75, 46.85 46.85 47.05 46.65 46.05 45.80 45.65 45.45
0.09 0.07 2.49' 0.015 0.0003 0.07 0.025
-1
E r r o r mean s q u a r e =
A
= IPS" B = Stearic acid c = St20b
5.0 0.5 0.3
CS2 (ABC.AB.
0 0 0
AC. BC. 2 b l a n k d = o,026
5 Isopropenyl stearate. Stearic anh>dride. p 5 0 01.
Results Response, m l HC1
Variance
Table IV. Data Used to Establish Confidence Limits 1
A B C AB AC BC ABC
47.4, 47.5 46.6 47.5 46.5 46.2 45.8 46.4 45.0
IVeighed, a
2.26d 0.195' 2.05d 0.165' 0.0003 0.0378 0.0078
IPS
E r r o r mean s q u a r e = CS2 (ABC:AC. BC, 2 blanks) = o,093
4 ~Isopropenylstearate. h Stearic anhydride. c p 5 0.05. d p
5
0.01.
terfere. Therefore, a study was made to determine whether this method could be used or adapted to accurately determine low concentrations of stearic anhydride in isopropenyi stearate. Several key questions had to be answered. Does the method give a n accurate determination for stearic anhydride in the concentration range of 0-10% by weight? Does isopropenyl stearate interfere? Will st,earic acid interfere?
EXPERIMENTAL Simulated samples for determination were prepared so that the concentration range would be equal to that found in process efflux. A nominal sample size of 6 grams was chosen, yielding mixtures of one or more of the following: IPS (5 grams), stearic acid (0-0.5 gram),and stearic anhydride (0-0.6 gram). The Johnson and Funk method was modified for use on stearic anhydride in IPS products. Fifty ml of 0.5N morpholine reagent were added to a sample of approximately 6 grams. A condenser was attached and the sample heated for 15 minutes on a steam bath. One hundred ml of methanol was added and titrations were made potentiometrically to a predetermined end point of pH 3.0. 2064
5.052 6.318 5.946 5.474 5.624 5.740 6.065 5.539 6.158 6.833 11.464 18.632 0.0 0.0 0.0
Frror,
Detern-ned, g
St2@h
St$
'
0.01.1
0.0
-
0.0 0.0 0.0 0.088 0.142 0.188 0.251 0.294 0.146 0.170 0.144 0.156 0.144 0.155 0.0 0.0 0.0
- 0.011
0.006 0.0 0.146 0.169 0.194 0.256 0.315 0.135 0.163 0.152 0.149 0.118 0.130 0.0 0.007 - 0.016 0.0 0.0 0.007 0.0 0.0 - 0.0014 - 0.016 0.0 0.0 0.047 0.0 0.0 - 0,004 0.0 0.0 - 0.024 0.0 0.0 Isopropenyl stearate. Stearic anhydride.
q
(Y,elahed - deterrrined)
0.011 0.011 0.006 0.0 - 0.058 - 0.027 - 0.006 - 0.005 - 0.021 0.011 0.007 - 0.008 0.007 0.026 0.025 -
0.007 0.016
- 0.007
0.0014 0.016 - 0.047 0.004 0.024
RESULTS AND DISCUSSION In initial trials, it was found that morpholine will react with stearic anhydride when Johnson and Funk's procedure is used. However, the reaction was very slow and unreacted anhydride remained after 24 hours a t room temperature. In an attempt t o accelerate the reaction, a group
ANALYTICAL CHEMISTRY, VOL. 46, NO. 13, NOVEMBER 1974
of samples was heated on a steam bath. Their composition and determinations, shown in Table I, indicate that the morpholine reaction is complete in 15 minutes at an elevated temperature. It was noted that condensers are required to prevent loss of the morpholine during heating. The reaction temperature on the steam bath is limited by the boiling point of the methanolic reagent (61 "C). Therefore, an attempt was made to further increase reaction rate by substituting butanol for methanol in each solution. This increased the solution boiling point to 78 "C. A 2:' factorial experimental design was made with the variabies stearic anhydride, IPS, and stearic acid concentration. Titrations were made colorimetrically. Details of the design are shown in Table 11. The results show a highly significant ( p 50.01) interference due to IPS, significant ( p 50.05) interference due to stearic acid, and significant interference attributable to the IPS-stearic acid interaction. As this approach was unpromising, use of methanol was re-introduced and the above design was repeated. Actual samples from the process are colored, and prevent colorimetric titrations. Therefore, this design was analyzed by potentiometric titration. In order to facilitate the titration, ca. 100 cm3 of methanol is added after heating. The end point was taken as the inflection point on the curve. The details and results of the design are listed in Table 111. The results indicate no significant interference. To establish confidence limits for the determination, 23 data points were statistically determined. These data are
listed in Table IV. They include 3 samples containing only stearic anhydride, 4 samples with IPS and no stearic anhydride, 8 samples with both IPS and stearic anhydride, and 8 samples containing no IPS or stearic anhydride (blanks). The variances were pooled after it was determined that the variances of the 4 sets were not significantly different by pairs ( F test, p 50.05). The confidence limits were determined as f0.044 gram. For a nominal 6-gram sample of stearic anhydride in IPS this f0.044 gram stearic anhydride is equivalent to f0.73%. The mean of the pooled error data was not significantly different from 0.0 (by t test). CONCLUSIONS The analytical method of Johnson and Funk, with the modifications described, can be used to measure stearic anhydride in isopropenyl stearate. The analysis, as modified, is not affected by IPS or stearic acid in the concentration range of 0-10% stearic anhydride. Analytical confidence limits of f0.73% were found, which is considered acceptable precision. ACKNOWLEDGMENT The authors thank Leonard S. Silbert for his advice, and Howard I. Sinnamon and Nicholas C. Aceto for their guidance.
RECEIVED for review May 31, 1974. Accepted July 31, 1974.
AIDS FOR ANALYTICAL CHEMISTS Simplified Remote Pipetting System E. M. Fortsch' and M. A. Wade Allied Chemical Corporation, Idaho Chemical Programs-Operations
The analysis of highly radioactive samples, as in the processing of irradiated nuclear fuels, requires that accurate remote pipettings be made. Several types of remote pipets have been designed for use in shielded facilities ( I , 2). Usually, the design is influenced by limitations imposed by the facility in which the pipets are used and, as a result, a wide variety of pipetting devices are being used throughout the industry with a high cost of custom design and fabrication. Thus, there would be many advantages for a low cost, easily operated, relatively accurate, commercially available remote pipet. In recent years, hand operated semi-automatic pipetting devices such as the Eppendorf, Oxford, and Sherwood pipets have become commercially available. These cover the range of 5- to 1000-microliter volumes and have many features which make them attractive for remote use. These features are: low cost. no direct contact of sample with pipet mechanism, reasonable accuracy and precision, non-
'
Present address. Fluor Engineers & Constructors, 5559 Ferguson Drive. Lo' Angeles. Calif, 90022. (1) D. C. Stewart and H. A. Elion, "Progress in Nuclear Energy," Series IX, Vol 10, Pergamon Press. New York, N.Y., 1970, p 59. ( 2 ) F W. Dykes. >. P. Morgan, and W. G.Pieder, "The Remote Analytical Facility Model E Pipetter ' At. Energy Comm. Rep.. 100-14456, 1958.
Office, 550 Second Street, Idaho Falls, Idaho 8340 1
wettable disposable tips, simple construction, and ease of operation. In the normal operation of these pipets, two spring tensions are used to control the distance a positive displacement plunger travels. One spring is used to draw a measured volume of sample into the disposable pipet tip and for expelling the bulk of the sample. The second spring is used as a "blow-out" to completely expel any liquid which remains in the tip following the depression of the first spring. Orsello, Pozzi, and Tedini ( 3 ) have adapted the Eppendorf pipet to remote use by a mechanical device which uses a lever-screw system to operate the pipet. The lever which drives the pipet button is operated by manipulators from outside the shielded cell. This paper describes an air-controlled device for operating the Eppendorf pipet remotely. The Eppendorf pipet was chosen because its critical parts (plunger and cylinder) are made of plastic and glass and will resist the corrosive atmosphere of a hot cell. The system uses controlled air pressure acting on a small air cylinder to perform the pipetting functions. The system is small, inexpensive, and requires very few remote operations for either its use or maintenance. (3) S.Orsello, F. Pozzi, and M. Tedini, "Una Nuova Telepipetta Rernotizzata Per Microprelievi: Di Soluzioni: Altarnente Radioactive," Comitato Nazionale Energla Nucleare, preliminary unpublished report, 1972.
ANALYTICAL CHEMISTRY, VOL. 46,
NO. 13,
NOVEMBER 1974
2065
