means of a fellowship grant by Eli Lilly and Co. LITERATURE CITED
(1) Bassett, L. G., Byerly,
W.,“Sodium,
Potassium, Rubidium, and Cesium,’’ in “Analytical Chemistry of the Manhattan Proiect.” ed..~. ~~~~. ited by C. J. Rodden; ~ p p’345-8, . bfcGraw-Hill, Kerv York, 1950. (2) Boltz, D. F., Mellon, M. G., Ax.4~. CHEM.19, 873-84 (1947). ~~
~~~
~
(3) Burkser, E. S., Feldman, R. V., Zavodskaya Lab. 7, 166-8 (1938). (4) Burkser, E. S., Kutschment, M. S., Mikrochemie 18, 18-21 (1935). (5) Burr, I. W.,“Engineerivg Statistics
16) (7) (8)
(9)
and Quality Control, LIcGrawHill, Yew York, 1953. Duvsl. C.. Chim. anal. 34, 209-21 (1952). Geilman, W,, Gebauhr, W.,Z. anal. Chem. 142, 241-54 (1954). Godeffroy, R., Ber. 9, 1363-8 (1876). Kahn, B., ANAL. CHEX 28, 216-18 (1956).
(10) Moser, L., Rietschel, E., ,l.ionalsh. 46, 9-22 (1925). (11) O’Leary, W. J., Papish, J., IND. ENG.CHEM.,ANAL.ED.6, 107-11 (1934). (12) Scroggie, A., J . Ana. Chem. SOC. 51, 923-5 (1929). (13) Snell, F. D., Snell, C. T., “Colorimetric hIethods of Analysis,” Vol. IV. v . 482. Van Nostrand. New York’, 1954.’ RECEIVEDfor reviex September 13, 1956. hccepted March 25, 1957.
Determination of Salicylic Acid in Aspirin CLIFTON W. STRODE, Jr.1, F. N. STEWART, H. 0. SCHOTT, and 0. J. COLEMAN F. Queeny Plant, Monsanto Chemical Co.,St. Louis, Mo.
John
b Spectrophotometric and visual techniques for the quantification of the U.S.P. limit test for salicylic acid in acetylsalicylic acid and aspirin tablets are described. By controlling the variables which may affect the ferric salicylate color and the hydrolysis of aspirin, precision and accuracy within 0.005% are illustrated.
T
ESTS LISTED I ~ Y the U. S. Pharmacopeia for salicylic acid in acetylsalicylic acid and aspirin tablets ( 7 ) do not fulfill the needs of modern industry because they are limit tests-Le., they only demonstrate that the material being examined contains more or less than an apparent 0.10 or 0.15% free salicylic acid. hfuch of the acetylsalicylic acid produced today contains less than 0.05% free salicylic acid, and i t is frequently necessary that the manufacturer know within 0.00570 or less the exact salicylic acid content of his product. Consequently, many unpublished modifications and extensions of the U.S.P. test have evolved, such as increasing sample and alcohol concentrations, using other solvents, and titrating a ferric alum blank with standard salicylic acid to match the color of the sample solution. As these changes have been developed independently, and in some cases apparently without proper consideration of the variables involved, conflicting data are frequently encountered. Edwards and coworkers (3) have made a comprehensive study of these variables, including the kinetics of aspirin hydrolysis and the effects of factors such as pH and ionic strength on the ferric salicylate test specified
Deceased.
1 184
ANALYTICAL CHEMISTRY
in the British Pharmacopoeia. They recommend conducting the test at 25” C. in a solution buffered a t p H 2.9. Colorimetric measurements are made a t timed intervals to provide an extrapolated correction for the hydrolysis of aspirin to salicylic acid. A simplified procedure is suggested involving a single measurement corrected for the previously determined hydrolysis rate a t a fixed temperature and weight of sample. Parallel work in this laboratory involving modification of the U.S.P. test, which specifies a more dilute solution of aspirin than does the B.P. test, has essentially verified the conclusions of EdJvards. The hydrolysis of aspirin must be controlled by definition of procedural details involving concentration, temperature, pH. and the time during Jvhich hydrolysis is significant. Calibration or comparison standards should be a t the same p H and essentially of the same composition as the sample solutions. il. practical test should provide a depth of solution commensurate with adequate sensitivity. Incorporating these considerations, the following spectrophotometric method has been in routine use in this laboratory for the past 12 years. The accompanying visual comparison method has been used in routine process control for the past 2 years. SPECTROPHOTOMETRIC METHOD
This method is somewhat analogous to the simplified procedure suggested by Edwards in that a previously estimated hydrolysis correction is eniployed, It is applicable for the determination of up to about 0.25% free salicylic acid in crystalline acetyl-
salicylic acid or in granular or tableted starch-aspirin formulations nhich may be white or tinted with pink or green dyes. Apparatus and Reagents. A Photovolt Model 102-E Lumetron colorimeter n-ith monochromatic filters a n d 100-mm. cylindrical cells is used for transmittance measurements. Other instruments with similar light paths t h a t transmit relatively narrow bands a t about 515 and 575 mM would probably be suitable. Acetylsalicylic acid is purified by recrystallization from isopropyl alcohol to an apparent free salicylic acid content of less than 0.005%. Ferric alum solution is prepared by dissolving 4.0 grams of ferric ammonium sulfate in 40 ml. of water and diluting to 50 ml. X 10-ml. aliquot of this solution and 10 ml. of 0.5S hydrochloric acid are diluted to 500 nil. with xater. The dilute solution is stored in a refrigerator and freshly prepared each week. Standard salicylic acid is prepared by dissolving an accurately weighed 0.1-gram portion of sublimed salicylic acid in 10 ml. of specially denatured Formula 30 alcohol (SD 30, 1 volume of methanol in 10 volumes of 190 proof ethyl alcohol), adding 0.1ml. of glacial acetic acid, and diluting to 1 liter with water. Calibration. I n oidei t o minimize t h e time during which hydrolysis takes place, the colorimeter, after a 10-minute warm-up period, is balanced t o 100% transmittance a t 515 mp with a blank. The blank is 1.9 ml. of SD 30 alcohol, 0.5 ml. of glacial acetic acid, and 2.0 ml. of dilute ferric alum solution, diluted to 100 ml. with Ivater. It should be freshly prepared. From 0.2- to 2.0-ml. aliquots of the standard salicylic acid solution, equivalent to O . O l ~ o to O.lyOsalicylic acid in 0.2-gram samples of aspirin, are
added to 100-ml. glass-stoppered cylinders. One gram of purified acetylsalicylic acid is now dissolved in 9.3 ml. of SD 30 alcohol and diluted to 100 ml. with water previously adjusted to 25" C. A timer is started as the water is added. A 20-ml. aliquot of the acetylsalicylic acid solution is added to one of the cylinders containing standard salicylic acid, followed by 2.0 ml. of ferric alum solution. The volume is made up to 100 ml. with water a t 25" C. The transmittance is measured within a maximum of 5 minutes, noting the exact time following the reading. The number of 15-second intervals required for the manipulation is multiplied by a hydrolysis constant of 0.001, and the product is added to the equivalent percentage of salicylic acid in 0.2 gram of aspirin represented by the appropriate aliquot of standard solution taken. A fresh portion of t,he purified acetylsalicylic acid is dissolved and aliquoted for treatment with each increment of salicylic acid. The corrected percent'ages are plotted against transmittances in the usual manner. This curve a t 515 nip is used in the analysis of crystalline acetylsalicylic acid, and in the analysis of white- or green-tinted starch-aspirin forniulations. For pink-tinted starch-aspirin formulations, a separate curve based on measurements a t 575 nip is constructed. Determination. I s in the calibration procedure the instrument is preliminarily warmed up and balanced a t the appropriate wave length with the reagent blank. One gram of the acetylsalicylic acid sample, or a weight of starch-aspirin formulation calculated to contain 1.0 gram of acetylsalicylic acid, is dissolved or slurried in 9.3 ml. of SD 30 alcohol in a d ~ y , 100-nil. glass-stoppered cylinder. The mixture is diluted to 100 ml. with water a t 25" C., and t'he timer is started concurrently. If starch is present the mixture is filtered inimediately by suction through a small circle of Whatman S o . 50 filter paper. A 20-ml. aliquot of the clear solution is treated JTith 2.0 ml. of dilute ferric alum solution and diluted to 100 ml. with water. The transmit,tance is measured and the time noted. ,4 hydrolysis correction, calculated as in the calibration procedure, is deducted from the apparent per cent salicylic acid corresponding t o t h e transmittance found. VISUAL METHOD
This method is limited to the estimation of free salicylic acid in essentially colorless samples. Apparatus and Standards. hlatched 100-ml. tall-form Nessler tubes are used with a suitable rack. A series of standards is prepared by diluting from 0.2- to 2.0-ml. aliquots of the standard salicylic acid solution, in increments of 0.2 nil., to about 50 ml. in separate Xessler tubes. Each solution is then treated with 1.9 ml.
Table I.
Determination of Sabcylic Acid in Acetylsalicylic Acid by Spectrophotometric Method
9cPresent
70Found
0.002 0 004 0.002 0 004 0.002 0 003 0 002 0 004 0 002 0 os0 0 027 0 025 0 027 0 031 0 027 0 027 0 052 0 052 0 052 0 043 0 052 0 056 0 052 0 048 0 075 0 077 0 077 0 076 0 102 0 102 0 102 0 104 0 SO2 0 SO2 Std. dev. of precision, 0.0024 Std. dev. of accuracy, 0.0023
Dev from Known $0 002 $0 002 $0 001 +o 002 t o 008 -0 002 +O 004 0 000 0 000 -0 009
0.002 0,002 0.003 0.008
0.002 0.027 0.027 0,028 0,054 0.049 0.060 0,051 0.076 0.074 0 102 0 102 0 098
of SD 30 alcohol, 0.5 nil. of glacial acetic acid, and 2.0 ml. of dilute ferric alum solution, then diluted to 100 ml. with water. These standards are freshly prepared every two weeks and are kept covered when not in use. Determination. The sample is treated as in the instrumental method except t h a t , if no starch is present, a 2.0-ml. aliquot of the initial undiluted alcohol solution is taken for color development: water, previously cooled t o 10" C., is used for the final dilution. By convenient prearrangement of apparatus and reagents, the time required for subsequent manipulations is held n ithin 30 seconds. Within this interval the developed sample solution is compared longitudinally n-ith the standards above a white background. From the standards most closely matching the sample, the salicyclic acid content is estimated to the nearest 0.0057,. RESULTS
Typical calibration curves for the spectrophotometric method show conformance to Beer's law. Hydrolysis curves were constructed from transmittance ineasurenients made a t timed intervals on thermostated solutions of aspirin and ferric alum. These curves indicated that, under the conditions of the spectrophotometric method, salicylic acid increased at a rate of 0.0036% per minute for the first 5 to 7 minutes a t 25' C. The rate was 0.0054% per minute a t 30" C. and 0.0028% per minute at 20" C. This emDhasizes the need for reasonably close temperature control in the spectrophotometric method. By carrying out the visual procedure within 30 to 40 seconds a t 10" C., the hydrolysis error appears to be within the experimental error. Known amounts of salicylic acid were added to specially purified acetylsali-
-to
004 -0 004 $0 002 -0 001 0 000 J-0 002 0 000
0 000 0 000 +o 001 +O 006 0 000 0 000 0 000 +o 001 +o 002 -0 003 $0 008 -0 001
+o
-0 0 0 -0
001 003 000 000 004
cylic acid by a procedure analogous to that followed in calibration for the spectrophotometric method. Results of the analysis of these mixtures (Table I) indicate that, in the range of concentrations measured, the precision and accuracy of the spectrophotometric method are within 0.005% salicylic acid a t the 95% confidence level. Comparisons of results obtained by the spectrophotometric and visual methods (Table 11) indicate reproducibility of comparable magnitude between the methods. Table II. Comparison of Spectrophotometric and Visual Methods
Sample So. 1
2 3 4 5 6 ?.
;
Salicylic Acid, 7c Spectrophotometric Visual 0 028 0 025 n nix 0 020 0.032 0.030 0.054 0 050 0.036 0.035 0.045 0.040 0 043 0 040 0 030 0 025
Std. dev. of reproducibility
0 0024
Results of the analysis of typical white, pink, and green aspirin granulations containing from 10 to 2070 starch are given in Table 111. Although the over-all precision was maintained, individual determinations on some of the pink granulations varied the most. DISCUSSION
In attempts to circumvent hydrolysis complications, some experimentation was conducted toxvard developing the ferric salicylate coIor in nonaqueous media. This work was analogous to VOL. 29, NO. 8, AUGUST 1957
1185
t h a t of Wisp and Brode (8) and Soloway a n d Wilen ( 6 ) , who described methods for carrying out qualitative ferric ‘chloride tests for phenols in organic solvents using pyridine as a n alkalinizing agent. Khile this approach appeared to have some promise as the basis of a spectrophotometric method, sufficiently stalile solutions of aspirin for permanent standards for visual comparison could not bc prepared. Standards containing only s:ilicylic acid differed in hue from solutions of aspirin, making visual comparisons subjective. I n aqueous media the ferric salicylate color is sensitive to pH and to changes in ionic strength a t the same pH (1-4). Glass electrode measurements made on solutions prepared according to the U.S.P. monograph directions (Y)showed that the pH of the aspirin solution is about 2.8, whereas the pH of the salicylic acid comparison standard is above 3.5. Figure 1 illustrates the variation in ferric salicylate transmittance when the pH of salicylic acid solutions is adjusted from 3.5 through 2.8 with acetic acid. This tends to indicomparison standcate that the U.S.P. a r d is darker than the color produced by an equivalent amount of salicylic acid in the aspirin solution and that results based thereon may be too low. I n the spectrophotometric method described, pH discrepancies and pos-
Table 111. Determination of Salicylic Acid in Starch-Aspirin Formulations by SpectroDhotometric Method
Sample NO.
Salicylic Acid Found,yo Khite Granulations
1
2 3
4 5 6 7 8 9
10
0.032 0.030 0.026 0.022 0.029 0.036 0.033 0.024 0.022 0.020
0.032 0.029 0.023 0.023 0.026 0.036
0,032 0.028 0.018 0.023
Green Granulations 1
2 3 4
5 6
0 057
0 053 0 073 0 057 0 045 0 041
0 060 0 051 0 069 0 057 0 046
0 038
Pink Granulations 1
2 3
4
5 6 Std. dev: of precision
1 186
0 0 0 0 0
043 014 054 050
059 0 061
0 0 0 0 0 0
0 0023
ANALYTICAL CHEMISTRY
038 042 059 056 053 062
PH 3.5
b
l
l
l
l
i
2.8 l
l
i
l
l
2.3
c
-
45-0.26 MG. SALICYLIC ACID I
I
0.0
4.0
0.5
ML. GLACIAL ACETIC ACID Figure 1. Effect of acetic acid on transmittance of ferric alum-treated solutions Measurements m a d e at 51 5 mg;
sible errors arising from small hgdrolgsis constant variations are minimized by incorporating equal amounts of acetylsalicylic acid in the calibration standards. I n the visual method, acetic acid is used to adjust the pH of the comparison standards approximately to the nominal value of the sample solution. The acetic acid has the further effect of preventing fading of the color and precipitation, presumably of an iron compound, which was observed to occur in stored color standards which had not been acidified. The mean ahsorptivity of a typical series of standards containing acetic acid was 10.7, while that of standards containing aspirin was 10.8. A similar series acidified to the same pH with hydrochloric acid had an absorptivity of 11.9 indicating that the effect of acetic acid is more similar to that of aspirin. Evidence has been presented that the ferric salicylate color is influenced by the concentration of organic solvent in aqueous mixtures (1, 6, 8). Visual observations substantiated by tristimulus measurements (5) on ferric salicylate solutions containing respectively 2 and 20% alcohol or acetone indicated that the color intenaity increased and the hue varied with the concentration of organic solvent, and that acetone had a greater intensifying effect than alcohol. Because the difference in intensifying effect was not large. alcohol has been retained as a solvent in the proposed methods in harmony with the U.S.P. test. The fact that the initial solution in the proposed methods, containing gram of aspirin in 9.3 ml. of alcohol diluted
100-ml. volume
t o 100 ml. with water, represents a supersaturated solution a t 25” C. might justify an increase in the alcohol concentration. I n the authors’ experience, however, crystallization during manipulation has never been a problem if the aliquoting and second dilution are performed with the rapidity specified in the procedural details. Were the alcohol concentration in the sample solution increased, it would have to be adjusted in like manner in the standards for the visual method. LITERATURE CITED
(1) Beck, P. F., Dansk. Tidsskr. Farm. 9, 289-302 (1936). (2) Broumand, H., Smith, J. H., J . Am. Chem. SOC.74,1013-16 (1952). (3) Edwards, L. J., Gore, D. S . , Rapson,
H. D. C., Taylor, 11.P., J . Pharni. and Pharniacol. 7, 892-902 (1955). (4) Illari, G., .-Inn. chim. applicata 29,490500 (1938). ( 5 ) XIellon, 31. G., “Analytical -1hsorption Spectroscopy,’’ pp. 535-7, IYiley) Sew Tork, 1950. ( 6 ) Soloxay, S . , Kilen, S. H., Ar.4~. CHEM.24, 979-83 (1952). ( 7 ) U. S. Pharmacopeia, XI‘, pp. 19-21, Mack Writing Co., Easton, Pa., 1055.
(8) Wisp, E. F., Brode, IT. R., J . Am. Chein. SOC.5 6 , 1037 (1934). RECEIVEDfor reviely October 8, 1956. Accepted ,April 10, 1957. Division of AknalYtical Chpmistry, Symposium on Analysis of Fine Chemicals and Pharmaceuticals, 130th Aleeting, AkCS, .itlantic City, K. J., September 1956.