Assay of Thyroid and Its Preparations J. L. DEUBLE AND JOHN WILKINSON, JR. Analptical Division, John Wyeth & Brother, Inc., Philadelphia, Penna.
Seberal analytical features of the U. S. P.
XI method of assay for thyroid lacked
are performed for analytical completeness but use of analytical grade reagents requires no correction to be made of the actual determination. Additional evidence of the effectiveness of the assay method is found in the stability of the end point. I n the assay of tablets of thyroid, provision is made for use of additional fusion mixture t o ensure complete combustion during fusion. While the method proposed by Johnson and Kelson ( 6 ) ) consisting of alkaline fusion without potassium nitrate and oxidation with bromine instead of hypochlorite, has been found to yield results concordant with those obtained by the authors' method, yet the time consumed is a distinct disadvantage.
a
sufficient degree of refinement to yield reproducible results. The proposed method satisfies this need and eliminates the analytical faults which attended the official method.
TT VIEW of
the fact that the U. S. Pharmacopoeia XI of assay for thyroid was widely considered to be inadequate, this laboratory instituted experimental work in 1937 to standardize the assay procedure. The method which was developed is applicable not only to powdered thyroid but also t o tablets. Several laboratories have made contributions from time to time regarding one or more phases of the analytical procedure. Thus, Hojer (4) postulated the existence of chlorates and perchlorates arising from the chlorine-oxidation procedure of the U.S. P. and attributed the gradual and constant shift of the end point observed in the titration to their presence. Burnett and lJ7arkow (2) took advantage of the comparative rates of liberation of iodine from potassium iodide by chlorate and iodate, performing the titration at a p H at which the desired react'ion by the iodate far exceeded that of the chlorate and thus removed chlorate interference. However, the present authors found t h a t the accurate control of pH was somewhat burdensome. Although the end point was improved, neverthless a gradual shift of the end point was still apparent due to the delayed influence of the chlorate. The coloration produced b y the thymol blue indicator was also objectionable. Hilty and Wilson (3) have proposed a modified procedure based on cerate oxidimetry, which necessitates fusion without an oxidizing agent and thus avoids the difficulties attending complete removal of chlorine and nitrite. The authors' experience with this method failed to yield complete recovery of iodine. Khile the authors' method was the result of a n attempt, to improve the entire U. S. P. XI assay method, they cannot overemphasize the improvement contributed by the elimination of nitrite from the solution. Although the A. 0. A. C. ( 1 ) uses sodium bisulfite for removal of nitrites, this reagent is troublesome in that' i t bleaches the indicator (methyl orange). The authors aroid nitrite int,erference by the simple expedient of wchling urea:
1 method
CO(SH2)S
+ 2HSOZ
+
CO2
TABLEI. EXPERIMENTAL RESULTSON POWDERED THYROID
u. 3. P. X I Proposed
Johnson a n d Nelson (5) Burnett a n d Warkow ( 8 ) Fusion without K F O s
+ HZO + 2HC1
CO,
cc .
0.212 0.206 0.198 0.198 0.199
E n d point indefinite
0.198 0.199 0.194
0.00
0.194
0,4 0.4
0.202 0.900
0.00
0.122
-
0.143
Modified Method
+
+ 2x2 + 3HZO
+
Blank
%
REAGENTS.Sodium Hypochlorite Solution. Dissolve 30 grams of chlorine in a cooled solution consisting of 44 grams sodium hydroxide in 750 cc. of distilled water. Dilute to 1 liter. Reagent grade hydrochloric acid, urea, sodium carbonate, potassium nitrate, and potassium carbonate. U. S. P. grade phosphoric acid and potassium iodide. PROCEDURE FOR THYROID. Weigh accurately a sample of about 1gram of thyroid and transfer to a nickel crucible (125 cc.), Add 15 grams of U. S. P. XI fusion mixture, and mix thoroughly with the sample. Superimpose a layer of about 5 rams of fusion mixture. Heat the crucible and contents to a dub red heat in a muffle furnace and then maintain this temperature until the carbonaceous matter is completely oxidized (about 20 minutes). Remove crucible and allow to cool. Transfer to a beaker (1 liter) containing about 300 cc. of hot distilled water. After the flux is completely disintegrated and dissolved, remove crucible, rinsing with distilled water. Cool the solution to 15" C. and add, slowly with constant stirring, hydrochloric acid (1 4) in amount sufficient to neutralize about 95 per cent of the alkaline constituents of the fusion mixture (ca. 20 cc. of hydrochloric acid and 80 cc. of water). Immediately add 50 cc. of the hypochlorite reagent and, while maintaining the solution at 15" C., add 60 cc. 2). Heat to boiling and boil to reof phosphoric acid (1 move excess chlorine, using moistened starch-iodide paper as indicator. During the boiling, maintain the volume at not less than 300 cc. by addition of distilled water, if necessary. Cool to room temperature and transfer the solution with the aid of distilled water to an Erlenmeyer flask. Add 10 grams urea, heat to boiling, and boil for 30 minutes, maintaining the volume a t not less than 300 cc. Cool to 20" C., add 10 cc. of 1 per cent,potassium iodide solution, stopper, and allow to stand for 5 minutes. Titrate the liberated iodine with 0.005 N thiosulfate, using starch as indicator. Each cubic centimeter of 0.005 N thiosulfate is equivalent to 0.0001058 gram of iodine. PROCEDURE FOR TABLETS. Keigh a counted number of not less than 20 tablets and reduce the weighed tablets to a h e powder. Weigh accurately a sample equivalent to 1 gram of thyroid and transfer t o a nickel crucible. For each gram of
Titrite is produced from nitrate during the initial fusion and, although unstable in acid medium, its complete removal by boiling is difficult and uncertain. The excess of urea is dissipated during the subsequent boiling: CO(KH,),
Iodine Found
iMethod
+
+ 2XH4C1
A study of the kinetics showed that the liberation of iodine from a potassium iodide solution by nitrite was a t a rate proportional t o the square of the concentration of hydrogen ion. Other notable improvements developed for the assay include standardization of the fusion, partial neutralization of alkaline fusion with hydrochloric acid t o yield more rapid elimination of chlorine by the salt-effect (sodium chloride), and standardization of sodium hypochlorite reagent. Blanks 463
INDUSTRIAL AND ENGINEERING CHEMISTRY
464
sample add 15 grams of U. S. P. X I fusion mixture and mix thoroughly. Superimpose a layer of fusion mixture, using 5 grams for each ram of sample. Continue with method as given above. d h e n neutralizing the alkaline solution of the flux with hydrochloric acid (1 4), use about 20 cc. of hydrochloric acid and 80 cc. of water for each 20 grams of fusion mixture employed in the assay. From the amount of iodine found, calculate to per cent of labeled amount on basis of U. S. P. mean (0.200 per cent) iodine content of thyroid.
+
Vol. 14, No. 6
Literature Cited (1) Assoc. Official Agr. Chem., Official and Tentative Methods of Analysis, 5th ed., XXVII, 57, p. 368, 1940. (2) Burnett, R. S., and Warkow, R. F., IND.ENQ.CHEM.,ANAL.ED., 12,734 (1940). (3) Hilty, W. W., and Wilson, D. T., Ibid., 11, 637 (1939). (4) Hojer, J. W., Biochem. Z.,205,273 (1929). (5) Johnson, F.F., and Nelson, H. A., Am. Drug. Mfrs. Assoc., Proc. 30th Annual Meeting, p. 186 (1941).
Determination of Salts in Crude Oil CLARENCE A. NEILSON, J. STEWART HUME, AND BERT H. LINCOLN Continental Oil Company, Ponca City, Okla.
T
HE presence of salts in crude oil is one of the most seri-
ous problems confronting the petroleum refiner. These salts may contribute to (1) the mechanical clogging of furnace tubes, condensers, and lines by deposition; (2) the corrosion of equipment by the hydrolysis of salts producing hydrogen chloride; and (3) high ash content of still residues. In predetermining the amount of salt in crude oil for the purpose of plant control, a quick and easy analytical method is of prime importance. A brief review of some of the many analytical methods now in use has been given by Blair (1). The more popular methods are essentially alike, in that they involve an intimate mixing of the oil with water, followed by the addition of a n emulsion breaker. Frequently a n organic solvent for hydrocarbons is added to reduce the viscosity of the oil. Essential differences lie in the manner of mixing the oil and water and in the demulsifying agents employed ( 3 , 4 ) . The aim of all extraction procedures is the quantitative removal of all the inorganic salts contained in the crude. This extraction is often difficult because the salts may be contained in brine-in-oil emulsions, rendered very stable by naturally occurring emulsifiers in the crude, or they may be present in the form of wax- or asphalt-protected crystals. Analysis of the extract may involve the quantitative determination of sodium, calcium, magnesium, sulfate, and chloride; but for most plant control work, only the chloride is determined. The chlorides are of first importance because of the hydrochloric acid released by the hydrolysis of the calcium and magnesium chlorides by crude oil distillation. The following procedure for extracting salt from crudes and for determining chloride in the extract has been in use for some time in Continental Oil Company laboratories and has proved to be reasonably accurate. It has the advantage of being rapid, since only one extraction is required.
Procedure Thoroughly mix the oil sample under test; then transfer exactly 125 ml. of the homogeneous sample to a 1-liter separatory funnel, If the oil is viscous, add 125 ml. of hot xylene, benzene, or toluene and shake until thoroughly mixed. Gasoline has always proved to be a very unsatisfactory diluent. The solvent is unnecessary with most crude oils lighter than 34 A. P. I. gravity, though it may always be used, particularly if an especially clear water extract is desired. For very light crudes, where the addition of a hot diluent may be hazardous, it is preferable to mix the oil and solvent at room temperature, then heat carefully to 60' C. (140" F.). Add 200 ml. of boiling distilled water and shake gently for 3 minutes, frequently relieving the pressure. (Invert the sephratory funnel and release pressure through stopcock.) Add 20 ml. of phenol in 30 ml. of boiling distilled water and shake gently for 5 minutes. Better extraction and a cleaner break of the emulsion have been found to result from this stepwise addition of the water and phenol. Allow the mixture t o stand until more than 100 ml. of clear water have separated out. Filter off exactly 100 ml. of the water
into a graduated cylinder through two sheets of heavy, qualitative filter paper. If negligible quantities of hydrogen sulfide and mercaptans are present, transfer the 100 ml. of filtered solution to an Erlenmeyer flask, adjust to a pH of 6.5 to 7.0, using 1 ml. of Continental indicator, add 0.5 ml. of a saturated solution of potassium chromate, and readjust the pH to 6.5 to 7.0. The advanta e at this point of the use of a universal indicator of the nature of Continental indicator is its clear yellow color at this particular pH range. Experience has shown that a careful pH control a t this stage is very important for a satisfactory chromate end point. Exact pH control for this titration has been found necessary in the presence of ammonium salts ( 8 ) ,and seems desirable in the presence of phenol. N o interference by the pH indicator with the chromate titration has been observed. Cool the solution to 29.4" C. (85" F:) and titrate withO.0;38 N silver nitrate to a light orange end point. Vigorous agitation of the solution during titration is absolutely necessary to ensure complete precipitation of the silver halide before the appearance of the silver chromate end point (2). A blank titration may be carried out on 92 ml. of water plus 8 ml. of phenol, the whole having been neutralized to a pH of 6.5 to 7.0 before addition of the chromate indicator. If either hydrogen sulfide or mercaptans or both are present, they must be removed before the halides can be determined. The method of acidifying and boiling is not applicable, since loss of hydrogen chloride may result. Transfer the 100 ml. of filtered extract to a 250-ml. beaker, add 1 ml. of Continental indicator, and adjust the pH to 6.5 to 7.0 (light yellow). Precipitate the sulfides and mercaptans with an excess of cadmium nitrate. Allow to stand 1 hour, then transfer to a centrifuge tube. Centrifuge until the solution is clear and the precipitate is a dense firm mass. Decant the clear solution into an Erlenmeyer flask, filtering if necessary. Rinse the beaker with 10 ml. of distilled water into the centrifuge tube. Thoroughly wash the precipitate by shaking, and recentrifuge. Decant the clear wash water into the solution in the Erlenmeyer flask and adjust to 6.5 pH (yellow). Titrate for halides. If a centrifuge is not available, filtration through Whatman No. 44 filter paper may be substituted. An alternate method for chlorides in the presence of sulfides and mercaptans is the Volhard procedure.
CALCULATIOK. (Ml. of AgNOs - ml. of blank) X 10 = grams of NaCl per barrel This relationship is true only if the exact volumes and concentrations specified are used. If the determination is desired in other units, the silver nitrate solution may be adjusted accordingly. If an appreciable amount of water is present initially in the crude, results will be too low unless corrected as follows: 250 1.25 X % water NaCl X 250
+
where the per cent of water is determined by A. S. T. M. method D-95-40.
Discussion of Procedure The use of phenol promotes the extraction of asphalt- or wax-protected crystals from the crude; hence phenol is essen-
