1 0.00013 0.00017 0.0045 None
2 0.002 0.0015 0.0005 None
3 0.0013 0.0012 0.00016 None
4 0.0004 0.00007 0.00015 N one
5 0.0002 0.0003 0.00002 None
6a 0.0001S 0.00018 < 0.00001 0.00015
7 a 0.00017 0.00004 < 0.00001 0.0005
8 0.00005 0.0001S 0.000025 None
9 None 0.00005 < 0.00001 None
10 0.00008 0.0001 < 0.00001 None
11 0.00005 0.00015 < 0.00001 None
B lank None 0.00001 None N one
° A n unm olted cathode deposited from a leached solution.
Figure 1 shows a num ber of spectrogram s of the standard solutions, using 0.1 cc. for each exposure. Concentrations are given in milligrams per 5 cc. Figure 2 shows spectro
grams made from th e samples tested and the blank obtained b y electrolyzing th e filtrate from sample 9, dissolving th e de
posit in nitric acid, and treating as a sample. Some platinum was dissolved from the electrode in th e process, as m ay be seen.
The section of th e spectrum shown in the figures extends from about 2590 A. on th e left to 3200 A. on the right, and in
cludes- the most persistent lines of all of the elem ents sought.
Ac k n o w l e d g m e n t
The author wishes to th a n k for their cooperation the Cop
per and Brass Research Association and the m any copper re
finers who furnished the samples.
Li t e r a t u r e Ci t e d
(1) M oser and M axym ow icz, 7j. anal. Chem., 67, 248 (1925-26).
(2) N itchie, In d. Eng. C h e m ., Anal. E d., 1, 1 (1929).
(3) Park and Lewis, Ibid., 5, 182 (1933).
Re c e i v e d D ecem ber 9, 1933.
Determination of the Acids of Plant Tissue
III. Determination of Citric Acid
Ge o r g e W . Pu c h e r, Hu b e r t Br a d f o r d Vic k e r y, a n d Ch a r l e s S. Le a v e n w o r t h
C onnecticut A gricultural E xperim ent S tatio n , N ew H aven, C onn.
C
IT R IC acid, when treated with potassium perm anganate and p o t a s s i u m bromide under th e proper con
ditions, is converted into the insoluble substance pentabromo- a c e t o n e . T h is reaction was originally employed by Stahre (6) for th e qualitative r e c o g n i t i o n of citric acid, b u t was placed upon a q uantitative basis by the work of Kunz (4), and of H artm ann and Hillig (if); it is now widely used for the deter
mination of citric acid in plant tissues. In order to obtain tru st
worthy results w ith the method as described by H artm ann and Hillig, i t is desirable to subject a t least 50 mg. of citric acid to the oxidation; furthermore, the transfer of the precipitate
Citric acid is oxidized to pentabromoacelone by p otassiu m perm anganate in the presence of potassiu m bromide. The oxidation produ ct is extracted by petroleum ether, dehalogenated by sodium sulfide, an d the bromide ion produced is titrated with silver nitrate. Q uantities o f citric acid o f the order o f 1 to 20 mg. can be determ ined in this m anner with an accuracy o f ± 5 p e r cent.
M alic acid alone o f the common organic acids interferes in a n y w a y w ith the determ ination, a n d the error introduced b y the presence of this substance is ordin arily so sm all as to be negligible.
Inasmuch as the conversion o f citric a cid to pentabromoacelone is not quantitative, although in constant proportion under the conditions described, it is necessary to em ploy a correction factor in the calculation of the results. This factor is approxim ately 1.12.
to t h e filter, a n d t h e s u b s e
quent washing and drying, require the most careful attention to details of technic.
Although of great value for m any types of analytical studies, H artm ann and Hillig’s m ethod is seriously lim ited in usefulness for the investigation of citric acid metabolism owing to the large am ount of citric acid required for accurate results. Nevertheless the specificity of the oxidation re
action upon which it is founded makes it the m ethod of choice for the determ ination of citric acid. The authors have therefore sought to improve the pentabromoacetone method so th a t th e filtration, w ith its atte n d a n t and necessary solubility correction, m ay be avoided and trustw orthy results m ay be secured on relatively small quantities of citric acid.
I t was obvious th a t a volu
m etric m ethod to d e t e r m i n e pentabrom oacetone w o u ld b e n e c e s s a r y . Accordingly, the authors have studied s e v e r a l procedures w h e r e b y th is sub
stance can be dehalogenated and th e h a l o g e n determ ined.
K om etiani (2) has described a m ethod for th e dehalogenation of pentabrom oacetone by warm alcoholic sodium iodide in the presence of acetic acid. A care
ful investigation of this reaction showed th a t q uan titativ e results could be obtained only under the m ost r i g i d l y controlled condi
tions; in th e authors’ hands the conditions described b y K om eti
ani invariably yielded results from 30 to 40 per cent too high.
Furtherm ore it was observed th a t p e n t a b r o m o a c e t o n e is unstable in ethyl alcohol solution. The reaction was not in
vestigated in detail, b u t w ithin 3 hours a t least 15 per cent of th e pentabrom oacetone in a 0.035 per cent solution of this substance in alcohol had been converted to a product th a t was no longer dehalogenated b y sodium iodide; after 72 hours th e conversion wTas greater th a n 75 per cent. Inci
dentally it was noted th a t pentabrom oacetone is unstable in ether solution also, b u t it appears to be stable when dissolved in chloroform or in petroleum ether.
If a freshly prepared solution of pentabrom oacetone in alcohol is treated w ith silver n itrate, a precipitate of silver bromide separates in a few m inutes a t room tem perature.
A stu d y of this reaction showed th a t it was entirely unsuited
for q u an tita tiv e work: dilate (0 . 0 2 N ) alcoholic silver nitrate liberated bromine equivalent to only 32 per cent of th e sub
stance (2 atom s = 35.25 per cent) when th e solution was boiled under a reflux condenser for 45 minutes, whereas more concentrated silver n itra te solutions liberated from 60 to 70 per cent (4 atom s = 70.63 per cent) under similar conditions.
The authors have therefore turned to a reaction described by K retov, Panchenko, and Savich (8), according to whom aliphatic halogen compounds are completely dehalogenated on being heated for a short tim e w ith alcoholic sodium
Bromine w ater: saturated aqueous solution.
Ferrous sulfate: 20 grams of crystalline salt, 1 cc. of concen sulfide diluted to 100 cc.. prepared fresh every 2 to 3 days.
Hydrogen peroxide (halogen-free): 4 grams of sodium per
oxide dissolved in 50 cc. of w ater w ith careful cooling and faintly acidified to Congo red w ith 50 per cent sulfuric acid (7 to 8 cc.
required); prepared fresh every week.
Silver n itrate, 0.05207 N : 8.8462 grams of pure silver nitrate diluted to 1000 cc. 1 cc. = 2.00 mg. of citric acid or 4.714 mg. of pentabrom oacetone or 4.162 mg. of bromine.
Ammonium thiocyanate, 0.05207 N : 3.96 grams diluted to 1000 cc. an d standardized against the silver nitrate.
Ferric am m onium sulfate (ferric alum ): 100 grams dissolved boiled gently for a few m inutes to expel traces of ether, is cooled, and 3 cc. of bromine w ater are added. A fter being allowed to separatory funnel previously cleaned w ith chromic-sulfuric acid, and the beaker is carefully rinsed into the funnel with 25 cc.
of petroleum ether, used in small portions. T he funnel is shaken vigorously and the aqueous layer is drawn off. T he ether is
May 15, 1934 I N D U S T R I A L A N D E N G
transferred to a second funnel. The aqueous solution is shaken again with 20 cc. of petroleum ether an a is th en discarded. The two ether extracts are combined and washed four tim es w ith 3- to 4-cc. quantities of water to remove traces of inorganic halides.
A 5-cc. q u an tity of 4 per cent sodium sulfide is added, and the mixture is thoroughly shaken; if pentabrom oacetone is present the aqueous solution becomes red, th e intensity of th e color being very closely proportional to the am ount of citric acid originally present. (This color reaction can be used to determ ine from 0.05 to 1.0 mg. of citric acid photom etrically; th e details of th e m ethod varied less than 0.2 per cent. Oxidation a t room tem perature was therefore adopted as being more convenient.
No conditions for th e oxidation could be devised in which the yield of pentabromoacetone, calculated as citric acid, was greater than 90 per cent. T he reaction does, however, yield a constant proportion of pentabrom oacetone under rather widely varied conditions. H artm an n and Hillig likewise noted th a t th e yield was not quan titativ e and sug
gested th a t a correction factor of 1.05 be used, in addition to a weight correction for the solubility of pentabrom oacetone.
The petroleum ether extraction technic described entirely eliminates this solubility correction; this is the more de
sirable inasmuch as the solubility of pentabrom oacetone was found to be appreciably greater in solutions derived from plant extracts than in solutions obtained b y oxidation of the pure acid. In eight determ inations of the pentabrom o
acetone th a t remained in solution, after filtration of th e crystalline product obtained essentially as described by H a rt
m ann and Hillig, the q u an tity found varied from 1.1 to 2.1 controlled; under these circumstances somewhat wide varia
tions of th e solubility m ight be expected.
T he choice of solvent for th e extraction of pentabrom o
acetone from the reaction m ixture is of great im portance.
Ordinary ether gave fairly satisfactory results when pure citric acid was employed, and when th e procedure was carried through prom ptly, although it was somewhat more difficult to wash this solvent free from inorganic halides.
W hen ordinary ether was used for the extraction of oxidation mixtures obtained from tobacco leaf extracts, however, the
J E E M N G C H E M I S T R Y 191
192 A N A L Y T I C A L E D I T I O N Vol. 6, No. 3 presence of an interfering halogen compound derived from
th e oxidation of malic acid was detected in the extract.
Furtherm ore, ether solutions of pentabrom oacetone are not stable on standing. The brom inated oxidation product of malic acid is not extracted from the reaction mixture by petroleum ether, and pentabrom oacetone is stable a t room tem perature in this solvent; its use is therefore essential to the success of the present method.
T he removal of the pentabrom oacetone from the petroleum ether by aqueous sodium sulfide and its simultaneous de- halogenation are quantitative. The average recovery of pentabromoacetone, calculated as citric acid, in eight experi
m ents in which the equivalent of quantities of from 5 to 19 mg. of citric acid were taken was 99.2 per cent. The error in the determ ination of quantities as small as 1.3 mg. was not greater th a n ± 8 per cent; the variation w ith the larger usually yellow or brown in color, and silver n itrate is prom ptly reduced if added; the interfering im purities are, however, easily removed by oxidation w ith potassium perm anganate.
I t is best to add the perm anganate rapidly until a red-brown present, as otherwise th e silver bromide subsequently precipitated will be contam inated with reduced silver. The solution is frequently slightly turbid owing to the presence separate experiments. No interference was detected. Malic acid, however, was found to exercise a small effect and was therefore given especial attention.
The quantities of pentabromoacetone found by titratio n after the oxidation of from 0.6 to 1.9 mg. of citric acid corresponded to 90.3 — 4.8 per cent in six experiments. When 6.4 mg. of citric cent recovery m ay therefore be taken as the m ost probable value of the recovery of quantities of citric acid in th e range of 3 to 16 fractions obtained by extraction w ith ether from a num ber of different samples, and was of such a concentration th a t a spectroscope in titratin g sulfuric acid solutions colored by copper sulfate and by extracts of tea and stick licorice, or to th e work of M acm illan and Tingle, J. In d. En g. Ch em., 12, 274 (1920), on determ ining the acidity of red wines and fruit juices. The omission of these references was inadvertent and th e w riter wishes to call atten tio n to th e work of these authors. W . C . Sm ith
Double-Acid Method of Optical Analysis of
th a t any single inversion m ethod giving only two polarizations, the of these errors is such th a t they cannot be rightfully ignored.Another fau lt in th e generally accepted methods is th e use of various clarifiers or decolorizing agents in the inverted polarization. E ven th e best of these introduce no small dis
crepancies in th e final results.
Appreciating th e im portance of the errors involved in the current single-acid inversion methods, Paine and Balch de
veloped the double-enzyme m ethod (1, 0, 10) which perm its th e evaluation of th e three unknowns. This is adm ittedly th e m ost basic and accurate m ethod y et devised. Though th is enzyme m ethod filled a long recognized need in beet- sugar chem istry, it has n o t been accepted for general use be
cause, if properly conducted, the enzyme inversions require more tim e th a n is usually available and th e cost is prohibitive.
T he cost of th e purchased enzymes used in the work reported in th is paper was about SI per determ ination and it is doubt
ful th a t enzymes of as satisfactory q uality can be prepared m ore cheaply. Anyone who has worked w ith enzymes is also fam iliar w ith th e difficulty occasionally experienced of secur
ing satisfactory enzyme preparations and of being sure of the completeness of th e melibiase inversions. nonsugars becomes zero in a strongly acid solution and is re
stored upon neutralizing, it is possible to determ ine th e polar
izing effect of these substances. T his can be done b y m aking two acid inversions, polarizing th e one in the acid and th e other in its neutralized solution. T he nonsugar effect is zero in the acid polarization, b u t has its full norm al value in the neutralized solution. T he difference between these two in
v ert polarizations will then be th e polarization of the non
sugars, after correcting for the effect involved in the neutrali
zation itself. H ereafter th e polarizing effect of these non
sugars, designated in a previous paper (16) as Z, will be term ed th e N value. T he direct polarization m ay then be corrected for this contained N value, leaving only two unknowns, su
crose and raffinose, which can be calculated from th e two equa
tions. T his is th e basic principle of the double-acid m ethod.
T his m ethod requires very little m ore tim e or expense than th e single-acid inversion m ethods now in general use, and th e results obtained are com parable to those by th e double-A melhod previously proposed lias been further
developed fo r the determ ination o f sucrose an d raffinose in beet products which cannot be analyzed accurately by a n y single-acid inversion melhod. Advantage is taken o f the fa c t that the optically active nonsugars become inactive in strongly acid solutions, as in a Clerget in
version, by m aking two acid inversions an d polarizin g one in its neutralized solution. The polarizin g effects of this am m oniacal neutraliza
tion have been established, as well as the effects o f leading an d deleading upon the inversion con
stants. A n alyses o f molasses fro m w idely sepa
rated districts, com paring the results by the es
tablished single-acid inversion methods, the double-enzyme melhod, an d the new ly proposed melhod, show the double-acid method to be comparable to the double-enzyme inversion melhod, whereas the single-acid inversion methods in general use give results that are seriously in error. A n alyses of synthetic siru p s by the double-acid melhod account satisfactorily fo r known amounts o f sucrose added to beet molasses.
The effect o f molasses im pu rities upon inversion constants is considered in a p re lim in a ry manner.
The method has been developed a n d used at the Johnstown molasses refinery during the p a st 6 years, an d is now offered as a substitute fo r the adm ittedly basic but im practical double
enzym e melhod.
193