867
V O L U M E 2 8 , N O . 5, M A Y 1 9 5 6 amount of resin an 1 acljiisting the f l o ~rate, the rxrhange capacity might be extended. T h e reprodncildity of ttir remlt3 for any given sample is rhntvn tiy the staiid:ird c!evi:ition,
Hence, the average amount of fluoride foilrid in each expwinient was in error by no more than 8 parts in a thoiisand. The largest deviation from the true value of a n y one determination wa3 0.1 mg. of fluoride ion per 8.00 mg. of fluoride sample, or 1 p i r t i n 100, or 1.25%. ACIiNOWLEDG.\IENT
w h e r ~tl = 0 - 31, 0 is the experimental n,illigritnis of fluoride, and JI is the arithnirticnl mean. From the table it is app:rrent that trace amounts of cation impurities cause a niarkr 1 i:inc'curacy in the direct determination of the fliioritle ion m:!ceiitration, :tltho:igh a precision of the same order of magnitiic'c, is obtained. Aiisiiniitig a trrie value, 1', of 8.00 mg of fluoride ion p p i ' 3O-ni1.
aliqnot, the rrlative error in any one determination
ip
0 - T
,-~X
1
100. Xpplying this rspreesion for precision t o each c~xl)erin-.rllt, :LII :rvc'r:ige precision was determined as fo1lon.s:
The author is i:idebted to Y. A4.Long for S1,iggeStiIig .init)etditc IR-l20(H) as the specific ion exchange resin heitlg $ought for this cation removal application. LITERITURE CITE11
(1) A???.SOC. Trsti/ig .llaterials. Siarida>ds, I-'. T-11. D 1179-5lT (1952). (2) Boruff, C . S.. -1bbott. G . R . , ISD. Esr,. CHL:\I... 1 s . k 1 . . ED. 5, 2 % (1933). (3) Bumstead. IT. E., Tl-ellz. J . C , , -1s.u..( ' H E X 24, 1595 (1952). (4) Castor, ( ' . I+., Saylor, J. H., Ibid.,24, 1369 (1952). (5) Shell. H. K..r r a i g , I?. L.. Ibid., 26,990 (1954). (Gj Willard, H . FI.. H o r t o ~ iC. . d..Ibid..22, 1194 ( 1 9 5 0 ~ . ( 7 ) Willard. H. 13.. LT-intcr. 0 . B.. I s u . ESC. ( ' H L J I . , .\s.i~.. E n . 5, 7 (193.7). H V C E I ~ Efor D rei-iew April 8.
19%. .i-\rcepted Felirrior>- 9 t .innlytical Clieii~istri.,127th l l e d n g , .iCS, Cincinnati, Oi
Ion Exchange Separation of Morphine Prior to Its Determination in PdpdVer somniferum C. H. VAN ETTEN, F. R. EARLE, T. A. MCGUIRE, and F. R. SENTI Northern Utilization Research Branch,
U. 5. Department o f Agriculture, Peoria, Ill.
\ o rapid, accurate method could be found for the determination of morphine in the poppy plant o r extracts from it; a new procedure was de\ eloped for isolation and
purification of morphine bj ion exchange methods. The morphine was finallj measiired b j the color it produces with nitrous acid, by its ultrntiolet absorption, or by titration. The method when applied to pure morphine ga\e an aierage recovery of 98qo with a standard detiation of 1.9 on 18 analyses of samples that contained from .5 to 20 mg. of morphine. The analjtical \slues obtained on opium and tarious extracts from the poppj plant were lower than those obtained by a colorimetric and a solvent extraction method, but 10 to 4070 higher than those obtained by methods in which morphine +I. as isolated lij crj stallization. Colorimetric methods applied to extracts of the poppy plant without prior separation of the morphine from interferences w i l l Fire high results.
I
N FOLI,O\YISG the processing of morphine from poppy plants, a rapid and accuiate method of analysis was desired. The I:. S. Pharmacopeia (ESP) method (14) for morphine in opium involves isolation of crystalline morphine and requires about 600 mg of the compound. This method, as well as others of similar nature, was not applicable to the problem, because some samples of plant fractions and partially processed material contained no more than 5 mg. of morphine in portions of convenient size for analysis. Methods sufficiently sensitive ( 2 , 4, 11)include the solvent extraction method of Levine and Matchett, which has not been published in detail (If ), and a colorimetric method ( 2 ) based on the color of the nitroso compound formed by the reaction of nitrous arid with morphine. When these methods were used on samples which permitted comparison, they gave results much higher than
those obtained by the I-SI' method. I n the cxtruction methoci, alkaloids are extracted from aqueous solutions at p H 8 to 9 by chloroform-ethanol and recovered by evaporation of the solvent'. The residue is dissolved in sodium hydroxide solution a t p H 10 to 11, and nonmorphine material is removed by extraction with benzene, first from the alkaline solution and later aftt.1 acidification. Morphine is extracted by chloroform-2-propanol after adjusting the solution to p H 8 to 9. The morphine is recovered by evaporation of the solvent, dissolved in methanol, and then titrated with acid. The colorimetric method gave higher results in the present investigation than the Levine and 3Iatchett method when applied to purified plant extracts anti even when applied t o the material isolated by the Levine arid X a t c h e t t procedure. The ionic character of the opium alkaloids suggeeted the use of ion exchange resins for the separation of morphine from interfering materials prior t o analysis. Exchange resins have been used analytically for the separation of codeine from morphine (1, 5 ) and for the preparation of free morphine from its salts (6, 10, 19). Ion exchange has also been used as part of procedures for separating bases, including morphine, from other substances in body fluids (13, 1 7 ) . I n the method presented here, morphine was separated from interferences by the procedure outlined in Figure 1. It was finally estimated by the nitroso colorimetric method, by it? ultraviolet absorption, or by titration. REAGENTS AND EQUIPMEIVT
Ion Exchange Resins. The cation exchange resin Dowex 50 X 1, 50 to 100 mesh, and the anion exchange resin Dowex 1 X 1,
50 t o 100 mesh, were prepared as previously described (18). A supply of the cation resin mas stored in the hydrogen form and the anion resin in the chloride form. Reagents for Ion Exchange Separation. Boric acid buffer solutions of p H 8.6 and 9.4 (9). These solutions were diluted with distilled water to make them 0 . 0 2 5 in concentration with
868
ANALYTICAL CHEMISTRY
respect to the boric acid present. Ammonium hydroxide, 0.5S. Sodium hydroxide, LV. Acetic acid, 0.3N. Hydrochloric acid,
3,v.
Reagents for Colorimetric Analysis of Morphine ( 2 ) . Sulfuric acid, 1volume of concentrated acid to 4volumes of water. Sodium hydroxide, 40 grams per 100 ml. of water. Sodium nitrite, 1% aqueous solution. Equipment. Colorimeter, micro ion exchange columns (19), steam bath, and usual volumetric ware. PROCEDURE
Preparation of Samples. I n an analysis of opium, a water extract of the sample was prepared as described in the USP method (14). For analysis of different parts of the poppy plant, solutions Tvere obtained b y 3-hour aqueous extractions of 5 grams of finely ground material in an ASTM (American Society for Testing Materials) rubber extractor ( 7 ) which had the siphon removed in order to obtain percolation of the water through the sample. These water extracts were filtered free of sediment and diluted t0.a convenient volume, from which samples were taken for analysis. STARTING MATERIAL
-
;t
Recovery. Average
Determinations 4 5 4
70
Kumber of Average, P a r t of DeterminaPlant 70" tions Capsules 6 0.61 Septa 2 0.38 Nodes 6 0.30 Stems 3 0.12 Corrected for approximate moisture content.
Standard Deviation 0.01 0:01
..
LV sodium hydroxide. This was followed by a distilled water wash until the effluent had a p H of about 7. Contents of the column were stirred once or twice with a small stirring rod during the washing. After passage of t h e cation eluate through this column, three 5-ml. washes followed. Then ampholytes were eluted a i t h 50 ml. of 0.3N acetic acid. The eluate, collected in a 100-ml. evaporating dish containing 3 to 5 ml. of 3N hydrochloric acid, was evaporated to dryness on a steam bath with the aid of an air current.
Ampholytes
EFFLUENT Anions. Neutrol Substances
I
Coiions, Bores
h
n
pH 8.6 Eluted Arnpholytes
ANA CY ZE D pH 9.4 Eluted Morphine
Figure 1. Diagram of ion exchange method of separation of morphine from impurities
PH
U.b
a 5.01 5
CATION DEVELOPER COLUMN No-Form
DISCARD
h'umber of
Mg. 5
Table 11. Morphine Content of Various Parts of Poppy Plant Determined by Ion Exchange Separation and Nitroso Color hIethod
a
ANION COLUMN OH-Form
Sample,
97.2 10 97.6 1.5 98.2 20 5 98,6 a USP morphine sulfate pentahydrate vhich gave theoretical carbon and hydrogen analyses and lost moisture equivalent t o five molecules of water on drying t o constant weight.
",OH ELUATE Boses, Cotions, Ampholytes
I
CATION COLUMN H Form
Table I. Recovery of Pure Morphine' Carried through Ion Exchange Procedure and Measured by Colorimetric Nitroso Method
'I
'
NH40H 1.0 N
Y.4
1
i-
+.4.0 2
3.0
w
2.0
17.9 mg. 71.4%
W
f
a
1.0
a 0
=
o
2.5mg
4.7mg 18.7%
10.0% 400
500
600
E F F L U E N T , ml.
I n analyzing various concentrates and other samples obtained in processing morphine from the poppy plant, suitable volumes of the solution nere taken for analysis without pretreatment. T h e senior author found morphine can be quantitatively removed by Doxvex 50 X 1 from solutions containing large amounts of inorganic salts. Honever, \Then solutions of low morphine and high salt content were used, morphine leaked through the column. I n samples of this kind, morphine 1% as extracted from the aqueous solution a t p H 9 in a continuous extraction apparatus n i t h a chloroform-ethanol (4 to 1 by volume) solution. Organic solvents nere evaporated, and the salt-free organic bases n ere dissolved in rvater or dilute acid. Samples xere taken from these solutions for analysis. Separation of Morphine from Impurities (Figure 1 ) . A water suspension of Dowex 50 X 1 was poured into a micro ion exchange column to a drained height of about 5 em. Analytical samples which contained 5 to 20 mg. of morphine in a 5- to 50-ml. volume were passed through the column a t a flow rate of 2 to 4 ml. per minute, after which the column was washed with 5 t o 10 ml. of water. Samples were eluted with 50 ml. of 0.5h' ammonium hydroxide. T h e anion exchange column of Douex 1 X 1 C1 was prepared in the same manner as t h e cation exchange column, but was filled to a drained height of 6 em. This resin was converted to the hydroxide form in the column by washing with about 15 ml. of
Figure 2.
Elution from Dowex 50 X 1 Na by 0.02N buffers in final purification step of procedure Sample, water extract of poppy capsules
After the residue had been dissolved in 5 to 10 ml. of water, it was put on a 5-em. developer column of D o v e s 50 X 1 S a . T h e column was elutriated with p H 8.6 buffer until the effluent had reached a p H of about 8.6. Then 30 ml. more buffer Tvas passed through. Total volume required was 50 to 100 ml. T h e morphine-containing fraction then was eluted with 225 ml. of the p H 9.4 buffer. This eluate, after acidification with 3A7 hydrochloric acid, was concentrated on the steam bath, then diluted to 50 or 100 ml. From this volume, samples containing 0.1 to 2 mg. of morphine were taken for analysis. Nitroso Colorimetric Method of Analysis ( 2 ) . Two samples, each containing 2 mg. of morphine or less, mere pipetted into 25ml. volumetric flasks, and the total volume of each was brought to approximately 5 ml. T o each of these solutions, 3 ml. of the sulfuric acid solution mas added. Two milliliters of the nitrite solution was added to one flask, the second flask serving as a blank. After the flasks stood for IO minutes a t room temperature, 4 ml. of sodium hydroxide solution was added to each flask a i t h mixing, following which the flasks were cooled in an ice
869
V O L U M E 28, NO. 5, M A Y 1 9 5 6 pH 8.6
pH 9.4
NH40H
KNOWN
ELUATE
ELUATE
ELUATE
MIXTURE
2.
I.
I.
2.
I.
2.
I.
2.
. . .... [< ...:
'ii'i
00
a 0 0
0
Figure 3. Paired chromatograms of allraloids separated in Figure 2 with known alkaloids for comparison 1. Color developed with nitroso reagent Color developed with Munier'e reagent .II. Morphine
2.
bath for 5 minutes. The cooled solutions were diluted t o volume and the color was read in a colorimeter using a Klett No. 42 filter. The instrument was adjusted to zero against each blank before the reading was made on the sample. RESULTS
Precision and accuracy of the method when applied to pure morphine are shown in Table I. Average value for recovery of morphine n-as 98%, with a standard deviation of 1.9 Recovery of known amounts of morphine added t o opium was equally good. To determine whether excessive amounts of narcotine, papaverine, codeine, and thebaine ivould interfere in the method, additional experiments vere run. I n these mixtures of two alkaloids were analyzed: 10 mg. of morphine and 10 mg. of one of the associated alkaloids. I n these mixtures narcotine, papaveiine, and thebaine reduced the recovery of morphine to as IOK as 91%. On elution with ammonium hydroxide, the first ration exchange column turned light gray in the presence of these alkaloids. Further elution of this column with methanolic ammonium hydroxide yielded large amounts of the associated alkaloids. These observations indicated precipitation of the associated alkaloids by the aqueous ammonium hydroxide and occlusion of small amounts of morphine. Codeine did not interfere and was recovered quantitatively in the anion column effluent. Analyses of parts of the poppy plant, Papaver somniferum, are given in Table 11. Values are in the range reported by other investigators for the morphine content of different parts of the plant (15, 16). Table I11 compares values obtained by this method with values obtained by the Levine and Matchett method (11) and modified USP methods on different morphine-containing natural products. Application of paper strip chromatography (as described later) to the end product of the Levine and Matchett method showed that nitroso-color-producing substances other than morphine were present. These other substances probably account for a t least part of the positive error in the Levine and hlatchett method when applied to a poppy plant extract. Examination of Developer Cation Column Eluate. Xitrosocolor-producing materials, selectively eluted by different buffers from the developer cation column used in the analytical procedure, are shown in Figure 2 These results shoFv t h a t the mixture of ampholytes obtained from the poppy plant contained bases
both stronger and weaker than morphine which develop a color with nitrite. The material eluted with the p H 8.6 buffer gave an ultraviolet absorption peak in acid from 273 t o 277 mp. The absorption curve did not resemble that of morphine, and it varied considerably from sample t o sample. The material eluted with the p H 9.4 buffer gave an ultraviolet absorption typical of morphine. It had a peak of 285 mp in acid and 297 mp in alkali. Of the solutions eluted a t pH 9.4, morphine content for a large number of samples was measured by ultraviolet absorption. ,4 small background correction n'as required. Some of the absorption curves indicated a trace of narcotine or papaverine. Evidence for the presence of these alkaloids in the p H 9.4 eluted fraction also was obtained by paper chromatography (Figure 3). The values calculated from the ultraviolet absorption of this fraction agreed closely with those obtained by the nitroso colorimetric method of analysis and by titration on the same samples (see Table IV). Following removal of the morphine, the column was elutriated mith ammonium hydroxide, giving an additional amount of nitroso-color-producing material. The ultraviolet absorption curve of this eluate was the same shape as that of morphine, but it differed in that the maximum was a t 280 mp in acid and 294 mp in alkali. Paper strip chromatography (Figure 3) showed, however, that no morphine was present in this eluate. The composition of the eluted fractions was examined further by paper chromatography according to the procedure of Munier (12) as described by Block ( 3 ) . The alkaloids from each fraction were extracted from inorganic salts a t pH 8.9 by means of chloroform and ethanol. llfter removal of the solvent the residue n-as dissolved in aqueous ethanol. Duplicate chromatograms from each sample and a knovn mixture of alkaloids were run, after which the chromatographs were separated. One strip was developed with the nitroso reagent used in the colorimetric determination. This reagent gives no color with thebaine, codeine, narcotine, and papaverine, and it gives a yellow color x i t h most compounds containing a phenolic hydroxyl. The companion strip was developed ivith Munier's reagent nhich gives a positive test for both phenolic and nonphenolic opium alkaloids. Typical chromatograms of the alkaloids in vaiious eluates detected by the two reagents are shown in Figure 3. As shown by the recovery of pure morphine (Table I), a small loss of morphine occurs in its separation from aqueous solutions on ion exchange resins. This loss may result from some leakage of morphine with the pH 8.6 buffer. Evidence for such a loss
Table 111. Comparison of Methods of .4nalyses hlethods, % Levine Modified USP and Ion Sample hlatchett exchange method Feed solids (poppy capsules) 0.88 0.59 0.45 Concentrate from ion exchange process 2 16 1.82 1.62 Concentrate from distillation process 3.72 3.25 2.28 ESP opiumQ 12.7 11,25 10.3 a Specifications for U8P XIV opium permit a range of 10.0 t o 1 0 . 5 7 ~ morphine. Sample contained 4.7% moisture.
Table IV. Agreement of Analyses by Different Methods of Measuring Product of Ion Exchange Separation hlorphine. % Nitroso color
Ultrariolet Sample Titrationa absorption b Water extract of straw 0.003 0.059 0.062 Straw sample 0.51 0.48 0.51 Aqueous concentrate 0.38 0.38 0.39 Free base was extracted from t h e final eluate b y chloroform-ethanol a t pH 8 i o 9. After evaporation of solvent, t h e base was dissolved in methanol and titrated with 0.01N acid. b Small background correction required for these estimations.
870
ANALYTICAL CHEMISTRY
is indicated in Figure 2, in that the p H 8.6 buffer continuously elutes a small amount of nitroso-color-producing material after the bulk of the nitroso-color-producing material was I emoved with this buffer. Additional evidence for the elution of morphine with the p H 8.6 buffer is shown in Figure 3. The chromatograms In Figure 3 also show the elution of a small amount of nitrosocolor-producing impurity with the morphine at p H 9.4. These small positive and negative errors of from 2 to 5% are not of great enough magnitude to account for the difference betn-een the values by the ion exchange separation and those obtained by the modified USP method or the Levine and RIatchett method given in Table 111. Therefore, if the ion ewhange separation method gives incorrect high results because of nonmorphine color-producing substances, these unknolvns cannot be differentiated from morphine by ion exchange resins, paper chromatography, or ultraviolet spectrophotometry as applied here. The possibility of such substances being present, however, must not be ignored. Evidence from these ion exchange separations shows the presence of color-producing substances other than morphine in the poppy plant. If these other substances are not removed, they interfere in colorimetric methods and cause high results. Thiq may account in part for the high values obtained by previously reported colorimetric methods (8, 15. 1 6 )
4Ch\ OW LEDGM ER T
The authois wish to thank ,4.H . Homeyer of the Mallinckrodt Chemical Works for the first three analyses by a modified ESP
method reportrd in Table 111)anti 13ettye Wilson for the spectrophotonietric analyscs. L I T E R - ~ T U R Ecrrm
hchor, L. B., Geiling, E. A I . K . , -4x.41..CHE:Jf. 26, 1061 (1954). -4damson, D. C. AI., Handisyde. F. P., Analyst 70, 305 (1945). Block, R . J., “Paper Chromatsgraphy.” p. 137, Amdemic: Press, Kew York, 1952. Cranier. J. W.W., Voernian. J. G., Seta Pharm. Intern. 1 , 219 (1950). Grant, E. W.. Hilty, W. W., J . A m . Pharin. Assoc. 42, 140 (1953). Jindra, .4.,J . Pharm. Pharmacol. 1 , 87 (1949). Joint Rubber Insulation Committee, J . I n d . Eng. Chem. 9 , 311 (1917) * N e e , F. C., Kirch, E. R., J . A m F‘hnrn. dssoc. Sci. Ed. 42, 146 (1953). Lange, X. A., “Handbook of Cheinist.ry.” 7th ed., p. 1127, Handbook Publishers, Sandusky. Ohio, 1949. Leri, L., Farmilo, C. G., Can. J . Chem. 30, i93 (1952). Matchett, J. R., Levine, J., ISD. Esc. CHEY., ANAL.E D . 13, 2G4 (1941). Rlunier, R., Bull. soc. chim. b i d . 33, 857, 862 (1951). J . Lab. Clin. Med. 24, 318 (1938). Oberst, F. W., “Pharmacopeia of the Cnited States XIV,” 14th rev., p. 400, Alack, Easton, Pa., 1953. Poethke. W.von, .Irnold, E., Phawn. Zentralhalle 88, 1 (1949). Reith, J. F.. Indexmans, A. W. >I., Phnrnt. Weekblad 8 5 , 309 (1950). Ytolman, A , , Stewart. C. P.. .477alust 74, 538, 543 (1949). Van Etten. C. H., AKAI..CHEII.27, 954 (1955). Van Etten, C. H., Wiele, 11.. Ihid..25, 1109 (1953). RECEIYED ior review Sol-ember 11, 1955. -4ccepted February 23, 1956. Uention of firm niimes or commercial products under a proprietary name or names of their nianufarturer does not constitute a n endorsement of such firms or prodiicts hy the U. S. Department oi Agriculture.
Use of Ionic Dyes in the Analysis of Ionic Surfactants and Other Ionic Organic Compounds PASUPATI MUKERJEE Department of Chemistry, University of Southern California, Los Angeles 7, Calif.
On the basis of previous investigations on the interaction between ionic dyes and ionic surfactants of opposite charge, a partition technique for the analysis of all classes of ionic surfactants and similar organic compounds has been developed. The theory of this new technique is presented. Experimental results with cationic and anionic surfactants of various kinds support the theoretical expectations; sensitivity of the new method is high. The qualitative detection limit in favorable cases is of the order of 1 to 2 p.p.b., while quantitative estimations of concentrations of the order of 0.1 to 1 p.p.m. and amounts of the order of 0.001 to 0.01 mg. are found to be possible. Hydrolyzable and nonhydrolyzable surfactants in mixtures can be determined.
I
K THE analytical chemistry of detergents and surfactants
in general the colorimetric methods involving the use of ionic dyes of various kinds have been very popular (1, 2, 6, 7 , 10, 11). The interaction between ionic dyes and ionic surfactants which forms the basis of these methods is also involved in the popular spectral-change method for the determination of critical micelle concentrations of surfactants, The precise nature of the interaction between dyes and surfactants involved was recently studied in some detail and the findings applied to the understanding of the detailed nature of the spectral-change method
(9). I n t h r present paper further application is made to t h r development of a new analytical technique for the detection, estimat,ion, and separation of small quantities of various classes of ionic surfactants and related compounds. Of the various kinds of possible interaction between dyes and surfartaiits, the one that is of interest here is the interaction between ionic dyes and surfactants of opposite charge. Various previous investigators have explained this interaction in terms of the formation of a “complex” between the dye and the surfactant (11). The investigations in this laboratory (9) showed that the react,ion involved is a simple metathetical one between the large surfactant ion and the large dye ion of opposite charge to produce compounds which usually have only slight solubility in water. These compounds are stoichiometric simple 1-1 salts in rvhirh electroneutrality is maintained by the dye ions and the surfactant ions only (except when acid or basic salts are formed by divalent dyes). The isolation, purification, and analysis of some of these compounds have been reported ( 9 ) . Most of the existing analytiral methods for surfactants involving t,he use of dyes utilize this compound formation betwecn a dye and a n oppositely charged surfactant. I n the titration methods of Hartley and Runnicles ( 7 ) and Salton and Alexander (IO) the color change accompanying the formation of some of these compounds is utilized as an indication of the end point. This is made possible by the fact t h a t the salt formed by two oppositely charged surfactants usually has a lower solubility in water and higher solubility in organic solvents than a dye-sur-
