Estimation of Oxygen Consuming Value of Coke-Plant Aqueous Wastes JOSEPH A. SHAW, Mellon Institute, Pittsburgh, P a . The present demands for substantially complete removal of phenols from plant waste water seem to imply from a practical standpoint resort to some form of oxidation. Many oxidizable substances are present in coke-plant wastes and'many of them are of unknown compositions. For plant design and plant control a reasonably rapid and accurate method of evaluating the reducing characteristics of such waste is needed. A procedure using chromic acid solution at 100"C. as the oxidant requires about 12 minutes for an individual test and less time per unit for a group of tests. No expensive or complicated instruments are required and results are reproducible within reasonable limits.
W
ITH the exception of certain special cases, the fundamental factor in stream purification is oxidation. A means of measuring oxidation requirements is a necessity; otherwise capital charges for waste disposal plants may be increased blindly. I t is difficult to exaggerate the commercial value of a quick method of analysis. The most widely used oxidation test is the so-called B.O.D. test ( d ) , which is usually of 5 days' or more duration. This test has given usable, though highly empirical, results with sanitary sewage, but its time requirements are commercially almost prohibitive, There seems to be much confusion in the minds of chemists as to the applicability of the 5-day R.O.D. test to trade wastes. The extent of the effort applied to devising a method of chemical oxidation is an indication of the need for such a method. Most of this work has centered upon the use of chromic acid salts. Adeney ( 1 ) was probably the first to use this type of reaction, and improvements upon his procedure have been proposed from time to time [Rhame ('?'), Madison ( 5 ) ,Ingols and Murray (Q), and Moore, Kroner, and Ruchhoft ( 6 ) ,who have listed considerable bibliographical data]. Certain of these proposals are based on unsound or difficult technique. Xone of them can be completed in less than 2 hours. This paper describes a procedure which can be completed in 10 to 12 minutes for a single determination and in appreciably less time per unit where a series of tests is run. These chromic acid tests are all rather similar and give empirical results. This test varies from the others principally in technique and the euceptionally small amount of time required for its performance. Although designed specifically for coke plant wastes, it can probably be applied much more widely. However, substantially all methods proposed give empirical data, and numerical data obtained by the different methods are seldom, if ever, directly comparable.
oratory the cabinet of a Shaw phenol still ( 8 ) , 150 X 100 mm. in cross section and with the substitution of a suitable lid that accommodated six tubes, was used. A larger cabinet is preferable where many tests are run. CHEMICALS AND SOLUTIONS
Mixed Acid. Equal parts by volume of 85% orthophosphoric acid and 95% sulfuric acid; this mixture is adjusted as described below. A water solution of chromic acid of approuiniately 7 grams per liter concentration, Solutions of ferrous ammonium sulfate and of potassium dichromate of 0.01 S concentration. Indicator solution is made by dissolving 0.32 gram of barium diphenylamine sulfonate in 100 ml. of a 10% sodium sulfate solution and filtering. Five drops of this solution are used per test. PROCEDURE
Into a dry 200 X 25 mni. test tube introduce 2.5 ml. of water and exactly 1.000 ml. of the chromic acid solution by means of pipets. From the dispensing buret add carefully 20 ml. of the adjusted mixed acid in such manner that the chromic acid and water previously added will be washed down the sides of the tube. Shake the tube by vigorous swirling until the solutions are thoroughly mixed, as indicated by absence of interference patterns when viewed in tran'smitted light. Place the tube in the cabinet, turnin direct steam, and start the stop clock. After exactly 5 minutes, remove the tube to a beaker of cold water. When cool, pour contents into a 300-ml. Erlenmeyer flask containing exactly 25 ml. of 0.01 *V ferrous ammonium sulfate solution washing the tube three times with water. Each water wash shouid have a volume approximately equivalent t o the volume of the sample initially in the tube (about 25 ml.). This d l ensure an acid solution of sufficient strength to give a good end point with the diphenylamine sulfonate indicator. Immediately add 5 drops of the indicator solution and titrate slowly with 0.01 N potassium dichromate to the first permanent purple end point. This constitutes a control test and from it can be calculated the amount of CrOl available t o react with organic matter in an unknown sample. This test should be run in duplicate. The unknown sample is dealt with in exactly the same manner, except that 2.5 mI of the suitably diluted sample of waste are substituted for the 2.5 ml. of water added in the control test. A sample suitably diluted for this test will have an oxygen demand fallin below 400 p.p.m. For best results samples should be fun in dujicate. A group of samples can be heated a t the same time, but it is preferable to titrate each sample immediately after pouring it into the 300-ml. Erlenmeyer flask. CALCULATION OF RESULTS
(Ml. of 0.01 i V Fe++soln. - nil. of 0.01 S KBCrPO,soln.)0.000333 = grams of CrOl found Then (grams of CrOs in blank - grams of CrOs in treated sample) 0.24 X dilution factor X 1000 = p.p.m. of oxygen consumed DISCUSSION OF METHOD
APPARATUS
Apparatus required includes a dis ensing buret (a 50-ml. buret with tip cut off may be substituted! a supply of 200 X 25 mm. test tubes, a 1-ml. capillary pipet, and a 5-ml. Mohr pipet. The steam cabinet is essentially an open-top can with a 0.125inch pipe nipple attached to the bottom as a drain and a notch a t the top to accommodate a direct steam line. The can is made preferably of copper. The open top is closed by a lid with a handle on one surface and lugs to prevent slipping on the other surface, The lid is perforated with a series of holes for the test tubes. The tubes are preferably suspended from the top by a rubber gasket, so that they extend 6 inches (15 cm.) into the cabinet. The cross section of the chamber is a matter of convenience, depending upon the volume of work customarily required. In this lab-
The mixed acid has to be adjusted to reduce impurities present. In this test the weight ratio of mixed acid to chromic acid is very great. In using the purest reagents obtainable, ahout 5% of the chromic acid added was lost in the control test. In switching from one brand of C.P. phosphoric acid to another c . p . brand, the loss jumped to 27y0. This untoward reducing action is theoretically compensated for by the control test, but satisfactorily compensated only where the amount of mived acid taken in each test is very accurately known. Because of the viscosity of the mixed acid, a volumetric measurement of sufficient accuracy cannot be readily made, and where the blank is high some form of gravi1764
V O L U M E 2 3 , NO. 1 2 , D E C E M B E R 1 9 5 1 metric measurement of the acid would be required for each test. This would be a time-consuming procedure. To meet this objection it was decided to destroy this reducing action by adding to the mixed acid a small predetermined amount of chromic acid. In this way the blank on the control test can be so reduced that a careful rough volumetric measurement of the mixed acid is satisfactory (see Table 1'). Control tests, however, must usually be run. Make up 2 liters or more of mixed acid and run two or three control tests thereon as well as a titration of the 7 grams per liter chromic acid solution. Find the amount of chromic acid destroyed by the mixed acid a t 100" C. in the 5-minute heating period and calculate the amount of chromic acid needed to be added to the whole batch. Weigh this amount of chromic acid on an analytical balance, work it carefully into solution by warming with about 50 ml. of the mixed acid, pour it into the main volume, mix thoroughly, and preferably let stand several hours. Experience in this laboratory has shown that the bulk of the added chromium, though not all, is reduced very quickly at room temperature. This acid may now be measured from a dispensing buret or a 50-ml. buret with the tip cut off. The lowest rate a t which a continuous stream of acid will flow from the latter buret will give a volume reading satisfactory t o this purpose without prolonged drainage. The manual accuracy obtainable in the test is limited now by the accuracy attainable in measuring the sample and the aqueous chromic acid solution, particularly the latter.
Table I.
Effect of Concentration of Acid and Heating Time
(1.000 ml. of CrOa solution and 2.5 ml. of 1.000 p.p m. pure phenol solution used in each test. Compensated mixed acid used.) Mixed H?O Heating Oxygen Expt. Acid, Added, Time, Consumed, % of. h 0. MI. hI1. hlin. P.P.M. Theoretical 1 20.0 None 5 219 92 (std. test) 2 20.0 Xone 5 213 90 (std. test) Xone 5 218 92 (std. test) 3 20.0 2.5 5 197 83 4 20.0 2.5 5 210 88 5 20.0 6 22.0 Xone 5 226 95 7 22.0 Kone 5 233 98 8 20.0 >-one 10 219 92 9 20.0 None 10 216 91 10 20 0 Xone 10 212 89 11 20.0 Xone 10 215 90 Expts. 1 t o 7 indicate action on IJhenol varies inversely with ratio of water t o acid in test. Expts. 8 t o 11 indicate t h a t heating time is not critical.
,45-ml. Mohr pipet was used for measuring the sample and a 1ml. capillary pipet for the chromic acid solution. The 1-ml, pipet, which had the 1.0-ml. mark about an inch from the tip, was treated by rubbing the tip on a medium fine silicon carbide block to a frustum of a cone, forming an angle of about 30' to the linear axis of the pipet and approaching very close t o the opening in the tip. This ground surface was then rubbed with a very little bit of silicone stopcock grease. This prevented the final drop of solution from running up the side of the tip when touched to the inner wall of the test tube. In handling, the pipet was filled above the zero mark, any adhering solution wiped off with a downward motion of clean dry fingers, the zero adjustment made, and the tip of the pipet touched at a wide angle to a small beaker. The solution was then run into the test tube and the 1.000-ml. adjustment made by touching the wall of the tube in the same manner. The actual volume used is relatively of minor importance, as the amount of available chromic acid present will be determined chemically in the control test; but it is critically necessary that exactly the same volume be introduced into each test. Satisfactory alternative technique may be suggested by microchemists. It was convenient in making the standard solutions to prepare 0.1 N solutions of ferrous ammonium sulfate and potassium dichromate and then dilute to 0.01 N as needed. An exactly 0.1 N solution of potassium dichromate was prepared from a known weight of the salt, 4.9037 grams per liter. The 0.01 N iron solution was standardized daily against the 0.01 N dichromate solution. When the cooled acid solution is wwhed into the ferrous iron
1765
solution in the Erlenmeyer flask for the titration, the solution will be noticeably warm. It was a t first thought that the acid must be diluted and cooled before addition of the ferrous salt solution. This cooling was cumbersome and time-consuming and titration in warm solution was tried. No significant difference was found in the results when the titration vias made a t once. Standardization of the ferrous solution in the cold and warm state gave results varying by only 2 parts per thousand. The time required for a single test by the described procedure is 12 minutes. For a group of tests, the time per test is much less. In working with an unknown sample of waste, arriving at a desirable dilution of this sample for testing is a cut and try process. It is frequently convenient to make a trial dilution of the sample, put 2.5 ml. in a tube with roughly measured reagents, and then note the amount of visual change of color in the tube. An observable reduction of yellow color should take place rather quickly, but yellow color must persist. A little experience will help with this trial dilution. A steam cabinet with a drain is specified for heating the samples. During the early part of this work a beaker of boiling water proved satisfactory. At no time did any of the tubes break in the water bath, but, from a standpoint of safety, it was deemed unwise to use hot water. The acid mixture is not nearly so bad as 95% sulfuric acid in this respect, as some heat has been dissipated in mixing, but splashing would probably occur if a tube did break in the water bath. Heating by direct steam is also quicker, where steam is available. The 5-minute heating time appears to be not highly critical, as doubling this time made no appreciable difference in the results (see experinients 8 to 11 in Table I ) This test would casually not appear to be well suited for use on samples containing volatile constituents. However, the testing of alcohol (experiments 13 and 14 in Table 11)showed 32 and 33% of theoretical found, which agrees well with the findings of lloore, Kroner, and Ruchhoft ( 6 ) , who used a refluxing method. After the first sudden heating of the air in the tube, no substantial amount of vapor leaves the tube. Under this condition there would be little tendency for any but very volatile constituents to leave the system in quantity. This method differs from other published chromic acid procedures, in that all other methods use a boiling procedure. Incidentally, coke-plant still waste is by definition substantially free of very volatile constituents, Certain other coke-plant 1% astes contain volatile constituents, but these are principally substances like hydrogen sulfide, hydrogen cyanide, and benzene, Tvhich more or less immediately decompose cold chromic acid. Table 11. Tests on Miscellaneous Materials Oxygen Consumed E r p t . Calcd., Found, Found, Description of Sample n o . p.p.m. p.p.m. % 2,4-Lutidine solution 1 259 47 19 2 259 04 2A ~~. .. _. Oleic acid solution 3 199 103 4 199 105 5 199 117 59 6 199 115 57 Wheat starch solution 7 177 148 84 8 177 146 83 Glycerol solution 9 193 180 93
432
10
Ammonium sulfate
193
175
91
Remark&
15-min. heating
Highly pure sample Nearly solid a t 26O C.
141 0.0 0.0 141 0.0 0.0 224 72 32" 224 73 335 Acetic acid 1660 26 1.6 1660 9 0.54 I n a few instances where direct comparison can be made with figures shown b y RZoore Kroner a n d Ruchhoft who refluxed for 120 minutes a t 145-150° C.. figdres for d- a n d m-cresol'are 8 and 5% lower t h a n their figures for 33% HpSO4 and for ethyl alcohol are practically identical. Attack on acetic acid b y this method appears less t h a n t h e higher temperature method a n d attack upon chlorides is practically negligible.
E t h y l alcohol
11 12 13 14 15 16
If calculated t o acetic acid end point, findings would show 96 a n d 98% of completion.
ANALYTICAL CHEMISTRY
1766
The phenolic bodies in still waste are roughly in the proportion of 70% phenol and a 30% mixture of cresols and xylenols. Together with thiosulfate and thiocyanate, they represent a very large proportion of the oxygen-consuming requirements of roughly clarified raw still waste. The latter two constituents vary considerably among plants and even from time to time in any given plant, but are seldom present in concentrations less than 0.5 gram per liter each. I t therefore seems conservative to
Tahle 111. Action of Oxygen on Phenol and Its Derivatives (Except as noted calculated oxygen consumed is based on complete oxidation' t o Con and water. Percentage column calculated t o nearest unit)
Description of Sample Pure phenol
Expt. So.
Oxygen Consumed Calcd., Found, p.p.m. p.p.m.
% of Theoretical
1 2 3 4" 5 6 7 8 9 10
238 238 238 119 179 179 143 286 286 286
219 213 218 113 160 166 131 270 264 262
92 90 92 95 89 93 92 94 92 92
Pure o-cresol
11 12 13 14 15 16
75 75 151 151 304 304
58 61 115 113 244 239
77 81 76 75 80 79
Pure m-cresol
17 18 19 20 21 22
75 75 150 150 354 354
62 60 118 115 282 284
83 80 79 77 79 80
Pure p-cresol
23 24
75 75 150 150 363 363
63 57 109 105 284 283
84 76 73 71 78 78
25
26 27 28
Table V.
Difference in Blank
[Using two different brands of C.P. phosphoric acid (A and B). 1.000 ml. of CrOa solution, 20.0 ml. of mixed acid, a n d 2.5 ml. of water in each test] Oxygen Heating ConExpt. Mixed Time, sumed, SO. Acid hlin. P.P.M. Remarks5 1 5 90 Variation high t o low 1% 2 ; 1: 116 Variation high t o low 2.9% a 174 Variation high t o low 1.0% 3 B 4 B(co1npensated) 11 Variation high t o low 0.4% 5 Average of 3 1 tests. Tests show oxygen-consuming power of impurities in two brands of phosphoric acid. I n all tests except 4 oxygen-consumed value shown (blank) is too high for good analytical procedure and in 3 it is far too high. Expts. 3 and 4 show value of compensation procedure. Remarks further indicate reproducibility of results expected from described technique.
say, judging from the figures in Tables I11 and IV,that this procedure measures about 90% of the oxygen-consuming requirement Pure 2,4-xyleno 30 29 378 266 269 70 71 represented by this major portion of the oxygen-consuming constituents of the waste. The minor portion of these oxygen-conPure 2,5-xylenol 31 303 224 74 32 303 222 73 suming substances is of almost entirely unknown composition and little can be said about their quantitative reaction in the procePure 3,4-xylenol 33 302 208 69 34 302 208 69 dure. As their source is coal tar, the components are probably Pure 3,5-xylenol 35 292 217 74 large cyclic molecules in solution and apparently easily oxidized, 36 292 220 75 except for nitrogen-containing bodies, as they break down noticeNutrient soln. ably upon exposure to air, forming insoluble solids. I t is, therefrom Hodge 37 476(?) 451 95 fore, reasonable to suppose that toward them the procedure would About 200 p.p.m. phenol(?) 38 476(?) 445 93 show a high degree of activity. This would seem to be all that can safely be said about the actual accuracy of the method. A s a 28-minute heating using 5 minute control. A 5-day B.O.D. test'on a pure phenol solution, made by George Tallon of for reproducibility of results, the data in the tables speak for Multiple Fellowship on Effluent Treatments a t Mellon Institute, showed themselves. 78 and 79% of calculated oxygen demand. B y the permanganate method the test showed 96 and 98%. The results obtained on pure compounds are in general considerably higher than those Tahle IV. Test on Still Wastes and on Impurities Likely to Be Found in Coke-Plant by the 5-day B,O,D. test, Still Waste though slightly lower than Oxygen Consumed those of the p e r m a n g a n a t e Expt. Calcd., Found, Found, Description of Sample No. p.p.m. P.P.~. % Remarks test. The reproducibility of the test appears good. Of the Still waste (untreated) 1 .. 2099 , . About 500 p.m. phenols 2 . . 2157 , . 5-day B.O.6. 1390 p.p.m. tests in Tables 11 and 111 show3 .. 2182 .. 2134 .. ing significant action, only one 4 5 .. 2177 .. varied more than 5% and only Still waste (dephenolized) 6 .. 1282 .. About 20 p.p.m. phenols three more than 3% with re7 .. 1314 , . 8 .. 2986 .. spect' to reproducibility. 9 .. 2934 .. The procedure d e s c r i b e d Still waste (treated) 10 .. 505 .. Oxidation treatment, 3.4 p.p.m. phenols h e r e a p p a r e n t l y does not 11 .. 329 .. Oxidation treatment, 1.3 p.p.m. phenols measure organic nitrogen bases, Still waste (dephenolized) 12 .. 3752 .. Dephenolized b y absorfition, 10 p.p.m* ammonia, and paraffinic acids. phenols 13 .. 3728 .. These substances exist cusPlant sewer, 1st sample 14 .. 149 .. tomarily in c o k e - p l a n t s t i l l 15 .. 149 .. waste in only minor amounts. Plant sewer, 2nd sample' 16 .. 128 .. BY KMnO4 126 p.p.m.a When the oxygen demand of 17 .. 134 .. 5-day B.O.D. 58 p.p.m.a a sample is much below 100 Pure NazSzOa solution 18 99 99.5 100.5 p.p.m. it is suggested that the 19 99 99.5 100.5 procedure be varied by adding Pure KSCN solution 20 127 116 91 Ox gen consumed calcd. according t o 5 ml. of sample, 35 ml. of 21 127 116 91 fiappold and Key (3) 22 127 119 93 10 min. in steam cabinet instead of 5 mixed acid, and the usual 23 127 118 93 min. 1 ml. of chromic acid. a Permanganate and 5-day B.O.D. tests run b y George Tallon. The tubes placed in the steam cabinet may be conveniently
.
~
V O L U M E 23, NO. 1 2 , D E C E M B E R 1 9 5 1 marked by rubbing a series of dots on the glass with a Carborundum hone and then impregnating the dots with achinamarking pencil. Because of the heat and handling, marking with a wax pencil is not very satisfactory. The method, because of its reasonable reproducibility of results and the very small amount of time required, has been helpful in recent research on the treatment of these wastes. LITERATURE CITED (1)
Adeney, “Principles and Practice of the Dilution Method of Sewage Disposal,” Cambridge, England, Cambridge University Prees.
1767 American Public Health Association, “Standard Methods for Examination of Water and Sewage,” 9th ed., 1946. (3) Happold and Key, Biochem. J . , 31,1323 (1937). (4) Ingols and Murray, Water and Sewage Works, 95, No. 3, 113-17 (2)
(1948).
( 5 ) Madison, Mellon Institute, 37th Annual Report, p. 20. (6) Moore, Kroner, and Ruchhoft, ANAL.CHEM., 21,953-7 (1949). (7) Rhame, Water and Sewuge Works, 94, No. 5, 192-4 (1947). (8)
Shaw, IND.ENQ.CHEM., ANAL.ED.,1,118
(1929).
RECEIVEDOctober 4, 1950. Contribution from the Fellowship on Gas Purification sustained a t Mellon Institute b y Koppers Co., Ino.
Polarographic Determination of Tartrates in Wines A. P. RIATHERS, J. E. BECK, AND R. L. SCHOENEMAN Alcohol Tux Unit Laboratory, Wushington, D . C . For the past 100 years there has been a divergence of opinion as to the occurrence of tartaric acid in fruits and berries other than grapes. Various governmental agencies have held that the presence of tartaric acid in certain food and beverage products is an indication of adulteration or mislabeling. There is no suitable method in the literature for determining small amounts of tartaric acid in the presence of fairly large amounts of other frui t acids and solids. Tartaric acid is quantitatively precipitated and interfering substances are removed from wine by the
N
0 SUITABLE method is described in the literature for the
quantitative determination of small amounts of tartaric acid in the presence of relatively large quantities of other fruit acids and solids in wines-e.g., 20 mg. of tartaric acid, 200 mg. of other acids, and 10 grams of solids per 100 ml. of wine. The Association of Official Agricultural Chemists method ( 1 ) is satisfactory for determination of tartrates in grape wine or wine containing more than 20% grape wine. Slathers (6) reported a colorimetric test for tartrates which has sufficient sensitivity to detect traces of tartrates in mine, but the method has not been standardized for quantitative work. Tartaric, malic, citric, and oxalic acids have been used extensively as supporting electrolytes and complexing agents in polarographic studies of metallic ions. Rleites (7-9) has dealt with the polarographic characteristics of copper(I1) in tartrate, citrate, and oxalate media. Lingane ( 4 ) has discussed the polarographic behavior of arsenic, antimony, bismuth, tin, lead, cadmium, zinc, and copper in acidic, neutral, and alkaline tartrate media. Lingane ( 5 ) reports a polarographic investigation of oxalate, citrate, and tartrate complexes of ferric and ferrous iron and shows the influence of pH on the polarographic waves. Furness, Crawshaw, and Davies ( 2 ) describe an indirect polarographic determination of ethylenediaminetetraacetic acid by forming the copper complex, precipitating the excess cupric ion with magnesium oxide, and measuring the height of the polarographic wave of the complex. As the tartrate ion does not give a polarographic wave, a number of tartrate complexes were investigated to find one suitable for an indirect polarographic determination. Kolthoff and Lingane (3)note that antimonyl tartrate produces a fairly well-defined wave with Eli2 = - 1.07 volts vs. the saturated calomel electrode with the main wave preceded by a smaller wave a t -0.5 volt. In the preeence of sodium methyl red, three waves are obtained which upon being made more alkaline give a single well-defined wave a t a potential of -0.94 volt us. S.C.E. Various metallic tartrate complexes give well-defined polaro-
procedure presented. A tartaric acid-antimony complex may be polarographed and a characteristic reduction wave obtained for use in quantitative work. Tartaric acid was found only in wines derived from grapes. The scope of the polarographic method is broadened by showing measurement of ions which in themselves are not reducible under conditions ordinarily employed in polarographic work. Most a-hydroxy acids can be determined from suitable complexes. There may be application to certain acids produced in metabolic processes.
graphic waves, but in most cases metallic complexes of isocitrate, citrate, malate, or all three give waves that cannot be readily separated from that of the tartrate complex. The antimony tartrate complex, however, is adaptable to the problem at hand. The present work describes a method for quantitatively precipitating tartrates from wine and preparing the antimony tartrate complex for polarographic analysis. APPARATUS AND REdGENTS
Polarograph. A Heyrovskg polarograph, Model XII, manufactured by E. H. Sargent and Co., was used for polarographic measurements. The drop time for the capillary was t = 2.5 seconds at E d e = 0 volt cs. S.C.E. in base solution. The temperature of the cell was maintained at 25” C. by means of a water bath. A saturated calomel electrode was constructed from a 50ml. Erlenmeyer flask by attaching a right-angle side arm of 10mm. glass tubing near the top of the flask, to serve as an agar bridge. The electrolytic cell consisted of a 50-ml. beaker with a Transite cover containing the necessary openings for dropping electrode, saturated calomel electrode, and nitrogen tube. Reagents. Bismuth nitrate solution, 20 grams of bismuth nitrate pentohydrate dissolved in 10 ml. of concentrated nitric acid and diluted to 250 ml. with distilled water. Sodium hydroxide solution, 10%. Antimony chloride solution, 3 grams dissolved in anhydrous ethyl alcohol and made to 100 ml. with anhydrous ethyl alcohol. hlethyl red indicator solution. Gelatin solution, 0.3y0in water. Buffer solution, 17 grams of sodium formate, 5 grams of glycine, formic acid, and Rater sufficient to bring pH to 3.4 and volume to 500 ml. ANALYTICAL PROCEDURE
Treatment of Wine Sample. Dilute 10 ml. of wine in a Babcock bottle with 25 ml. of water, add 5 ml. of bismuth nitrate solution, centrifuge, and discard the supernatant liquid. To the precipitate add 20 ml. of water and 2 ml. of bismuth nitrate solution, shake, and place in boiling water for 15 minutes. Cool, centrifuge, and discard the supernatant liquid. To the bottle add 20 ml. of cool water and 1ml. of bismuth nitrate solution, shake, centrifuge, and discard the supernatant liquid. Dissolve the precipitate in
