Potentiometric Nonaqueous Titration of Substituted Fatty Acids JACK RADELL and E. T. DONAHUE Eastern Regional Research Laboratory, Philadelphia 18, Pa.
course of a study of wool wax fatty acids, it was found Iin 95% that these mixed acids could not be titrated satisfactorily ethyl alcohol with aqueous sodium hydroxide and
by reversing the plug in the receptacle, or occasionally by adding a milliliter of methanol to the solution being titrated.
phenolphthalein indicator. The difficulty was the result of the fact that highly colored samples obscured the end point, and even with the less highly colored samples a gradual change of indicator color occurred, making the end point difficult to determine. Attempts to circumvent these difficulties by using glass and calomel electrodes with a Beckman Model H-2 p H meter, produced curves that did not have a steep enough slope a t the equivalence point to give accurate results. Consequently, the method of nonaqueous potentiometric titration was investigated. The feasibility of titrating fatty acids in nonaqueous solvents has been indicated in the literature (1, 2, 5 ) . Essentially, the method used in this paper has been described by Fritz and Lisicki ( 3 ) . Some simplifying modifications have been introduced, and the method has been extended to the titration of substituted fatty acid$.
RESULTS
N THE
0.1X-
Results of the titrations are listed in Table I. Characteristic titration curves are shown in Figure 1, in which the volume of titrant is plotted against millivoltage readings. Titration of the benzene-methanol solvent with lithium chloride added indicated no alkali was consumed by these reagents. There is reason to believe that some of the hydroxy acids present in wool wax acids may be lactonized, but a sample of y-stearolactone gave a blank titration. The titration curve for a-sulfopalmitic acid and ammonium a-sulfopalmitic acid in Figure 1 showed the presence of two breaks. The first break, unlike the second, was not pronounced
Table I.
Determination of Xeutralization Equivalents
0.1N NaOhle SaOH Found Founda Stearic 284 5 285 284 2 a-Bromopalmitic 337 336 5 335 3 a-Sulfopalmitic 171 168 2 170 7 182 6 Ammonium a-sulfopalmitic 184 176 7 9,lO-Dihydroxystearic 312 315 4 316.5 9.10-Euoxvstearic 303 298 7 298 d 12-Ketbsteario 296 294 0 298 5 Mixed aminosteario 293 299 9 296 8 Mixed wool wax 384 372 y-Stearolactone s o riaition No reaction KO reaction KO reaction No reaction No reaction +Caprolactam 6 This determination was made n i t h , O . l N sodium hydroxide, 95% ethyl alcohol solvent, and a phenolphthalein indicator.
EXPERIZlEYTAL
Acid
REGEXTSA N D APPARATI > AIethanol, ACS reagent grade. Benzene, ACS reagent grade. Solvent, benzene-methanol (3 to 1 by volume). Sodium methoxide, 95% or better. Benzoic acid, Kational Bureau of Standards, primary standard. Lithium chloride, ACS reagent grade. Beckman p H meter, ;\lode1 H-2. Electrodes, calomel and antimony. Titration flask, modified iodine flask for potentiometric titrations ( 4 ) . Buret, 10 ml., graduated to 0.05 ml. SODIUM METHOXIDE SOLUTIOS. To prepare an approximately O . 1 N solution, about 6 grams of sodium methoxide, 105 ml. of methanol, and 615 ml. of benzene are put in a 1-liter Erlenmeyer flask, the flask is stoppered, and the solution is stirred nith a magnetic stirrer for 15 minutes. A large quantity of undissolved sodium methoxide indicates that it has deteriorated excessive11 and fresh material should be used. The solution, if necessary, is filtered and stored in a glass-stoppered borosilicate glass bottle It is necessary to keep the sodium methoxide and the sodium methoxide solution in stoppered flasks. No precautions are necessary to prevent the stopper from sticking. A solution of sodium methoxide showed little change in normality over a period of 6 months. The solution need not be restandardized more often than once a week.
Theory
3
500
/I
I
5 400 E
PROCEDURE
-
I
The acid sample or the standard benzoic acid is weighed into a tared 250-ml. titration flask ( 4 ) . The sample size should be chosen so that about 7 ml. of 0.1,V sodium methoxide solution are required. About 0.2 gram of lithium chloride and 50 ml. of the benzene-methanol solvent are added. The flask is stoppered and the contents are allom-ed to go into solution (10 to 15 minutes). A glass or Teflon-encased stirrer is introduced, and the electrodes, previously soaked for 0.5 hour in the benzene-methano1 solvent, are inserted into the side arms of the flask. The antimony electrode is inserted into the electrode jack normally used for the calomel electrode, and the calomel electrode is adapted to fit the other jack by means of a terminal connector. The solution in the titration flask is stirred by the magnetic stirrer and titrated a t a reasonably uniform rate. At first, readings are taken a t every 0.5 ml., but at every 0.1 ml. near the equivalence point.
4,
Attempts to avoid plotting each titration by titrating to a fixed millivoltage reading proved impractical because of the difficulty of reproducing conditions. Erratic behavior of the meter needle, which may be caused by several factors, may be eliminated by permitting a 0.5-hour warm-up period for the pH meter,
Figure 1. Titration of Wool Wax Acids and Substituted Fatty Acids
0
>
300
’i 200
3
4
5
6
VOLUME,
7
8
9
rnl
A. Representative c u r v e for wool wax, a-bromopalmitic; 9,lO-dihydroxyrtearic; 9,10-epoxystearic, 12-ketostearic, and mixed a m i n o s t e e r i c acids
B.
590
Representative curve for a-sulfopalmitic and ammonium =-sulfop a l m i t i c acids
V O L U M E 26, NO. 3, M A R C H 1 9 5 4 enough to obtain an accurate end point. Attempts to titrate a-mercaptopalmitic acid were unsuccessful. DISCUSSIOV
Application of the potentiometric nonaqueous titration method to mixed wool wax acids has made possible the determination of neutralization equivalents, which usually were unattainable or grossly inaccurate. The relatively pure and colorless model compounds liqted in Table I demonstrate the comparable accuracy of the nonaqueous methods and the accepted method foi the determination of neutralization equivalents. The mixed wool wax acid sample whirh gave the best agreement by the two methods is given in Table I. Other mixed wool wax acid samples gave either poorer agreement between the two methods or could be titrated only by the nonaqueous method. The compounds used demonstrate the feasibility of titrating fatty acidq containing substituted bromine or an amino, epoxy, dihydrosy, a-sulfonic acid, and ammonium a-sulfonate group. This has indicated that lactone, lactam, and epoxy functional gioups do not interfere in the determination of the neutralization equivalent. This method has also been used to analyze samples derived from wool wax which have an acid number as low as 17 and are so highly colored that they could not be analyzed by the indicator method. Determination of aminostearic acid (mixed position isomers) dissolved in 95% ethyl alcohol and titrated potentionietrically
59L with 0. liV aqueous sodium hydroxide was unsuccessful unless the amino group was first allowed to react with formaldehyde. In the method here employed, the methoxide ion is capable of titrating the aminostearic acid directly. Similarly, the ammonium a-sulfopalmitic acid which resulted in appreciable precipitate and a gradual color change when titrated with 0.LV aqueous sodium hydroxide in 95% ethyl alcohol could bereadily titratedpotentiometrically with sodium methoxide in the benzene-methanol solvent. 4CKNOWLEDGMENT
The samples of a-broinopalmitic, a-sulfopalmitic, and amnionium a-sulfopalmitic acids were kindly supplied by J. K. Weil. The 9, 10-dihydroxystearic, 9,10-epoxystearic, 12-ketostearic acids; and mixed aminostearic acids were supplied by Daniel Swern and the ystearolactone by Ahner Eisner, all members of the lahoratory staff. LITERATURE CITED
Folin, O., and Flanders, F. F., J . A m . C'hern. SOC.,34, 774 (1912). Folin, O., and Wentworth, -4.H., J . Biol. Chem., 7, 421 (1910). Fritz, ,J, S., and Lisicki, N. >I., ANAL.CHEM.,23, 589 (1951). Ogg, C. L., Porter, W. L., and Willits, C. O., ISD. ENO.C H m r . , ~ A L ED., . 17, 394 (1945). (5) Rescorla, A. R., Carnahan, F. L., and Fenske, 11. R., Ibid., 9, 505 (1937).
RECEIVED for review October 7, 19.53. Accepted November 23, 1953.
Possible loss of Iron during Sodium Carbonate Fusion of Silicates and Rocks HASKIEL R. SHELL Analytical Laboratory o f the Electrotochnical Laboratory, HE first step generally employed in the analyEis of silicates T a n d rocks containing much silica is a basic fusion with .qodium carbonate to prepare the sample for acidic separation of silica. This fusion is usually done in a platinum crucible with a gas burner as a source of heat. That the loss of iron during this st.ep may be serious is known, but' in practice is frequently disregarded. Duparc ( 2 ) said that the loss of iron to the platinum crucible \\-as caused by the reducing effect of flame gases that diffuse through t,he platinum and avoided loss by fusion in a muffle. Jacob (4)discussed losses of iron to the containing platinum rrucible during sodium carbonat'e fusions, pointing out t,hat !ossep occurred even if all iron was present initially as iron(II1) oxide. Also, he suggested that some (hut not too much) potas$ium chlorate be added to the podium carbonate to prevent this loss of iron. S o mention was made of possible loss of iron as a volatile chloride by such additions. Potassium chlorate, which decomposes a t 400" C., would be largely gone before t'he sodium caarbonate fusion started; therefore any effect noted by Jacob may have been from the residual potassium chloride. Dittler ( 1 ) also noted that such losses occurred and suggested that potassium nitrate be added to iron( 11)oxide-enriched samples to prevent such loss. I n analyses in this laboratory, potassium nitrate did not prevent the loss of iron to the containing platinum crucible. Hillebrand and Lundell ( 3 )objected to the use of R muffle because of a strong attack of the crucible around its rim by the molten sodium carbonate, with subsequent introduction of platinum into the solutions. Some procedures may state that "oxidizing conditions should be maintained" but do not give :my concrete data on how to obtain them or point out the serious errors involved when oxidizing conditions are not obtained.
U. S.
Bureau o f Mines, Norris, Tenn.
Usually, however, the Puhject is ignored entirely and each analyst must discover the details for himself or falsely assume that all iron of the sample, and only the iron of the sample, is contained in his later solutions. That this false assumption is frequently made is exemplified by the analysis of a kyanite mineral by well-known and respected laboratories both in the United Stateq and abroad. On identical samples of mineral the following rewlts were obtained by the various laboratories for percentage of iron(II1) oxide: 0.57, 0.66, 0.44, 0.24, 0.75, 0.39, 0.36, 0.51, 0.73, and 0.58. Estimation of the true value for iron in this sample \ \ a
