T H E J O U R N A L OF I N D U S T R I A L A N D E N G I N E E R I N G C H E M I S T R Y
426
It has been found t h a t certain grades of asbestos are affected by alkali and i t is necessary t h a t t h e asbestos, before use, should be freed from alkali-soluble constituents. For analytical purposes a very good sample of anthraquinone was obtained from a so-called chemically pure commercial sample by recrystallizing several times from hot toluene. This commercial sample was analyzed by the above method and gave the following result: Anthraquinone taken. . . . . . . . . 0.2000 g. Anthraquinone recovered.. . . . . 0.1992 g. Loss.. . . . . . . . . . . . . . . . . . . . . . . 0.008 g. 0.40per cent After purification t h e sample was analyzed as follows:
-
TABLEI-ANALYSIS
OF
PURE ANTHRAQUINONE
SAMPLE Zinc Dust Sodium Hydroxide Recovered Alcohol Gram Gram Gram Cc. of 5% S o h 0.2000 0.4 15 To 0.2003 0.3002 0.6 20 wet 0.2997 0 2003 sample 0.2007 0.4 15
Per cent
100.15 99.83 100.19
This method of purification and estimation seems t o be especially adapted t o supplant the “Hochst Test” in the estimation of anthracene because of its greater speed and accuracy. It may also be used with excellent results in estimating t h e purity of anthraquinone which is contaminated with anthracene and less t h a n I O per cent of phenanthraquinone. The examples given in Table I1 illustrate the degree of accuracy t o be expected when no special precautions are taken. Some of these analyses were completed in less t h a n z hours. TABLE11-ANALYSES OF MIXTURES COMPOSITION O F MIXTURES Anthraquinone AnthraAnthra- Phenanthraquinone Recovered quinone cene Gram Gram Gram Gram 0.1803 0.0297 0.0000 0.1805 0.2000 0.0000 0.1975 0.1982 0.1782 0.0000 0.0228 0.1805 ~
~
0.2018 0.2005 0.2004
O.Oi47
0.1342 0.0000
0.0000
0.2020
0.0098 0.0100
0.2000 0.1999
Per cent
100.09 99.65
101.28 - . ~ ~~
100.10 99.75 99.75
A single analysis may be easily completed in 1’/2 hrs., exclusive of the drying of t h e final product t o constant weight. If analyses must be completed in a short time the drying of the sample may be hastened by washing with alcohol and ether. The sacrifice in accuracy may be as little as I per cent. A modification of this method is being worked out t o increase the accuracy in the presence of large amounts of phenanthraquinone. \
COLORINVESTIGATION LABORATORY BUREAUO F CHEMISTRY WASHINGTON, D. C.
CRITICAL ELABORATION OF QUANTITATIVE PRECIPITATION METHODS EXEMPLIFIED BY A METHOD FOR THE DETERMINATION OF PHOSPHORIC ACID By H. HEIDENHAIN Received December 12, 1917
Of the numerous quantitative precipitation methods comparatively few have had the benefit of a thorough critical investigation. With most of them their authors have been satisfied when “good results” had been obtained. Such, however, are by no means proof of the correctness of a method. A compensation of errors must always be considered a possibility. A s
Vol. xo, No.
6
long as there exists any doubt in this respect, the scientific analyst will not be satisfied. On t h e contrary, he will desire t o learn how a method will work under various conditions, i. e., what influence t h e quantity of substance employed, concentration of solution, temperature, presence of certain substances, etc., may have on t h e result. T h e task of examining a method as t o its reliability has been p u t u p t o me repeatedly. While a t first such work was lined out by me just for the particular case on hand, I later found t h a t certain methods employed in my researches were applicable in a great number of cases. I might say t h a t in a measure I have found a scheme for this class of work. I t is t h e purpose of this article t o develop this scheme. However, before taking u p my subject proper, I think i t advisable t o show how in one case of my experience the scheme has been successfully applied, as by so doing i t will be easier for me t o make myself clear later on. A PRACTICAL CASE
The method of determining phosphoric acid by precipitating t h e same b y molybdic acid solution and transforming the molybdic precipitate into magnesium ammonium phosphate is generally known. This transformation was necessary as long as we did not understand how t o produce the molybdic precipitate in constant form. This, however, has finally been accomplished. Several articles on this subject have been published, but it was t h e thorough researches by Hundeshagen, Zeikchrift fgr analytische Chemie, 1889, which chiefly aroused my interest. Hundeshagen proved in convincing manner t h a t t h e precipitate contains, for every 3 equivalents of phosphoric acid, 2 4 equivalents of molybdic acid and 3 equivalents of ammonium, if produced under certain conditions, and t h a t t h e precipitate could be determined by titration with standard alkali solution, using phenol,phthalein a s indicator. Testing Hundeshagen’s method I could confirm his findings, but I noticed t h a t t h e end reaction a t titration was lacking in sharpness. Hundeshagen used a solution of ammonium nitrate as wash liquor, a n appreciable amount of which remains in t h e filter and precipitate. This, as well as the ammonium in chemical combination with t h e phosphoric and molybdic acids, evidently is t o be blamed for the uncertainty a t titration, as ammonium salts cannot be titrated with exactness with phenolphthalein as indicator. On the other hand, this indicator seems indispensable t o bring phosphoric acid t o a definite stage of neutralization. There was, however, a way out of this dilemma. After the precipitate had been washed with ammonium nitrate solution, this salt could be removed b y washing with alcohol a n d the ammonium in the precipitate could be gotten rid of by supersaturation and evaporation with t h e standard alkali solution, and determination of t h e excess of alkali b y boiling with a n excess of standard acid solution and titrating back with standard alkali solution. Thus t h e end-reaction was made sufficiently sharp and the results obtained were very satisfactory, b u t t h e method had become rather cumbersome.
June, 1918
T H E JOURNAL OF I N D U S T R I A L A N D ENGINEERING CHEMISTRY
As stated, t h e trouble was caused by the presence of ammonium. If in its stead a fixed alkali could be used, all difficulties, i t seemed, were overcome. The thought suggested itself, t o precipitate the phosphoric acid in the form of potassium phosphomolybdate, which likewise is a compound of very little solubility. Hundeshagen mentions t h e potassium compound in his article and states t h a t its composition is analogous t o t h a t of the ammonium compound, but t h a t it is more soluble t h a n the latter. M y own experiments confirmed this. At the same time, the results I obtained were so good t h a t I thought it worth while t o follow the matter up, i. e., t o subject the contemplated method t o my scheme of criticism. One of t h e first observations made was t h a t the washing of the precipitate could not be declared finished b y any test employed, be i t a test for acidity or for molybdic acid. The wash liquor used a t first was a solution of I O per cent potassium nitrate in water. The filtrate remained acid even after prolonged washing. When t h e washing finally was stopped after such large quantities of wash liquor had been used t h a t undoubtedly all free acid had been removed, the results were too low. Obviously the solubility of t h e precipitate was the cause of this loss. Following fai’nous precedents I might have declared the washing finished after a “certain” amount of wash liquor had been used. Such arbitrary methods, however, I never approved of. I tried washing until constant acidity of t h e filtrate was attained, assuming that the acidity of t h e filtrate must gradually diminish until all free acid had been removed, and t h a t when the acidity of the filtrate was caused only b y t h e solubility of the precipitate, constant acidity would prevail. This principle was a good one, only I found t h a t when constant acidity was attained, already losses had been sustained which could not very well be disregarded. It then occurred t o me t o give the wash liquor from the beginning a certain acidity b y addition of nitric acid, thus rendering the precipitate less soluble. This move was successful. I found t h a t in a I O per cent solution of potassium nitrate with so much free nitric acid as corresponds t o a I / I O O normal acid, t h e solubility of the precipitate was b u t 1 / 2 0 of t h a t in the neutral solution. The losses were now so small t h a t the results were not affected any more t o any serious extent. The small quantity of free acid remaining in the filter and precipitate could be determined with satisfactory accuracy and a corresponding correction applied. While in my first experiments t h e results were too low, I now obtained figures a trifle too high, presumably caused b y some molybdic acid which had been carried down b y the precipitate. From these preliminary experiments I had learned t o handle the method. However, in order t o become clear concerning all phases of the method I had t o proceed systematically. I made a series of experiments. I-As t o t h e solubility of t h e precipitate in a neutral solution of potassium nitrate. 11-As t o the solubility of the precipitate in a n acidified solution of potassium nitrate.
427
111-As to the completeness of the precipitation of phosphoric acid by the solution of potassium molybdate in nitric acid by standing without stirring. IV-As t o the effect of stirring on the completeness of precipitation. V-As t o the influence of nitrate of potash on the precipitation. VI-As t o the influence of nitric acid on the precipitation. VII-As t o the influence of different substances as chlorides, sulfates and tartrates on the precipitation. The results of these investigations are given in the following tables. Table VI11 shows a few determinations of phosphoric acid in a chemically pure monopotassium phosphate. TABLEI-SOLUBILITY OF POTASSIUM PHOSPHOMOLYBDATE IN SOLUTIONS O F POTASSIUM NITRATE After Digestion with Potassium Phosphomolybdate Corre10 cc. Solution sponding KNOa in Neutralized by to PzOa EXPT. 100 cc. N / 5 0 KOH Mg. No. Grams cc. 2.1 0.130 1........... 10 2 ........... 15 2.1 0.130 0,130 3 . . ......... 20 2.1
TABLE 11-SOLUBILITY OF POTASSIUM PHOSPHOMOLYBDATE IN SOLUTIONS OF POTASSIUM NITRATEACIDIFIED BY NITRIC ACID
Solution Contained KNOsin HNOs EXPT.100 cc. Corresponding No. Grams toa 1.. 10 1/100 N s o h . 2.. 10 1/125 N soln. 3.. 10 1/250 N s o h . 4 . . ... 10 1/500 N soln.
... ... ...
TABLE 111-PRECIPITATION
After Digestion with 10 cc. Neu- Potassium Phosphomolybdate tralized 10 cc. Solution Correby Neutralized by sponding to N / J O KOH N / 5 0 KOH PZOS cc. cc. ME. 5 5.1 0.606 4 4.2 0.012 2 0.018 2.3 I 1.8 0.049
ACID BY POTASSIUM MOLYB16 HRS. STANDING WITHOUT STIRRING Volume = 100 cc. Pz01 employed = 0.6 174 mg. Constant KNOI = 15 g. Potassium molybdate solution’-varying Solution of Titration Potassium of Precipitate PZOl by N / 5 0 KOH EXPT. Molybdate Lost No. cc. cc. Mg. 0.0 1..................... 0.5 0.6174 2 ..................... 1.0 0.0 0.6174 3 ..................... 2.5 7.15 0.176 4 ..................... 5.0 7.60 0.148 5 ..................... 10.0 0.145 7.65 5 ................. 20.0 7.35 0.164 1 This solution was prepared analogous t o the usual ammonium molybdate solution. It contained 5 per cent MoOl. OF PHOSPHORIC
DATE SOLUTION AFTER
I-
....
TABLEIV-PRECIPITATION OF PHOSPHORIC ACID B Y POTASSIUM MOLYBDATE SOLUTION AFTER 5 MIN. S T I R R I N G AND 45 MIN. STANDING See Table I11 for description of solution Solution Potassium Molybdate No. cc. 1 ..................... 2.5 2 ..................... 5.0 3 . . . .. . . . . . . . . . . . . . . . . 10.0 4 ..................... 20.0 5 . . . . . . . . . . . . . . . . . . . . .40.0
EXPT.
Titration of Precipitate by N / 5 0 KOH cc 0.3 9.65 9.9 9.45 8.10
.
PrOr Lost Me. 0.599 0.014 0.006 0.034 0.117
-
TABLEV-INFLUENCE OF POTASSIUM NITRATEON THB PRECIPITATION OF PHOSPHORIC ACID AS POTASSIUM PHOSPHOMOLYBDATE Volume = 1OOcc. 2 PzOs employed = 0.6174 mg. Constant Potassium molybdate s o h . = 10 cc. Potassium nitrate-varying Titration of Precipitate Pa05 Exa. Lost KNOi by N / 5 0 KOH No. Grams cc. Me. 1 ..................... 5 0.0 0.6174 2 ..................... 10 4.2 0.358 3 ..................... 15 7.55 0.151 4 ..................... 20 9.3 0.043 5 ( a ) . . . . . . . . . . . . . . . . . . 20 9.2 0,049 ( a ) This solution was stirred for 5 min. and allowed to stand 21/a hrs.; the others were also stirred for 5 min. but allowed t o stand only 1 hour,
1
I
T H E JOURMAL OF I N D U S T R I A L A N D ENGINEERING C H E M I S T R Y
428 TABLEVI-INFLUENCE
OF FREENITRIC ACID O N THE PRECIPITATION OP ACID AS POTASSIUM PHOSPHOMOLYBDATE Volume = 100 cc. Pzos employed = 0.6174 mg. Constant Potassium nitrate = 20 g. Potassium molybdate s o h . = 10 cc. Nitric acid, sp. gr. = 1.2-varying Titration of Precipitate P206 EXPT. ”0s by N/50 KOH Lost No. ( a ) cc. cc. Mg. 1..................... 0 9.3 0.043 2 10 ’ 8.4 0.099 3 20 0.0 0.6174 ( a ) These solutions were stirred for 5 min. and allowed t o stand 1 hr.
Vol.
IO,
No. 6
error, however, may be avoided if, instead of large quantities, small quantities are employed, for in small quantities the absolute amount of the impurities is very small and may, therefore, be disregarded. I n experiments t o study the solubility of precipitates we must keep apart (a)the influence of the solution from which the precipitate is separated, ( b ) the influ..................... ence of the wash liquor. ..................... (a) For the determination of the former the method TABLEVII-INFLUENCE OF DIFFERENTSUBSTANCES ON THE PRECIPITA- is the same as with a n ordinary determination, only TION OF PHOSPHORIC ACID A S POTASSIUM PHOSPHOMOLYBDATE the quantity employed is t o be taken so small t h a t Volume = 100 cc. while a considerable portion remains in solution, still PzOs employed = 0.6174 mg. Constant Potassium nitrate = 20 g. a decided precipitation is taking place so t h a t the effect Potassium molvbdate s o h . = 10 cc. Titration of of supersaturation has not t o be taken into account. Precipitate P~OS Additions by N/50 KOH Lost EXPT. According t o circumstances, either the part remaining No. I g. of Each cc. Mg. in solution or the precipitated part is determined. 1............. KC1 8.95 0.065 2.. . . . . . . . . . . . 8.60 0.086 Sometimes both parts may successfully be determined 3 . . . . . . . . . . . . . KHC4H4Oe 0.0 0.6174 4............. CaCOs dissolved in “Os 9.25 0.046 which, of course, will give the most satisfaction. I t is hardly necessary t o mention t h a t solutions employed for these tests ought t o contain all substances in the KHzPOa PzOs PzOs Time of EXPT. Employed Found Calculated Difference Standing same proportion as in a practical analysis. No. Gram Per cent Per cent Per cent Hours For the purpose of determining the influence of the 52.14 52.20 -0.06 1..............0 . 1 52.20 0.0 52.20 2 .............. 0 . 1 single substances of the solution, i t is commendable t o 52.14 52.20 -4.06 3 .............. 0 . 1 52.23 52.20 +-0.03 4 .............. 0 . 1 make series of tests in which, a t a time, one substance 52.17 52.20 -0.03 5 .............. 0.05 52.35 52.20 + O . 15 6 .............. 0.05 is vzried while the others remain constant and t o pre52.17 52.20 -0.03 7 .............. 0.1 sent the results in the form of tables. T h e study of 1 The sample of KHzPOa and 7 g. KNOs was dissolved in 30 cc. water. Then 40 cc. potassium molybdate solution was added. T h e precipitate the influence of temperature and concentration does was washed with a 1/100 normal nitric acid solution containing 10 g. KNOs in 100 cc. until 10 cc. of the filtrate were neutralized by 5.1 cc. N/50 K O H not need any explanation. solstion. (Compare Table 11, Expt. 1.) ( b ) The determination of the solubility of a preReturning t o my subject proper-it seems t o me cipitate in the wash liquor seldom presents any difficulty t h a t the accuracy of a precipitation method depends if we have plain water or another volatile liquid. All upon two circumstances-first, upon the degree of t h a t need be done is t o digest a small quantity of the solubility of the precipitate; second, upon its purity. washed precipitate for a sufficiently long time t o ensure If we fail t o find satisfactory conditions in these two saturation and t o determine the dissolved part b y respects we consider the method as worthless or em- weighing after evaporation or by titration, etc. If ploy it only conditionally. However, we must always solutions of salts are used as wash liquors the determinakeep in mind t h a t ideal conditions can never be ful- tion by evaporation and weighing the residue is out filled. All precipitation methods, therefore, are more of the question. There remain, however, all other methods of determination. or less defective. I t is hardly necessary t o mention t h a t the methods If we wish t o t r y a method as t o its accuracy, it is not sufficient t o submit a certain quantity of a substance for the determination of such small quantities as we t o a prescribed process and accept the result a s final have t o do with in this class of investigations must criticism, because the two sources of error mentioned be adapted t o the conditions, Filters, funnels, evapomay influence the result in such a manner t h a t one rating dishes, etc., must be taken proportionately small compensates the other. I n order t o avoid the danger and standard solutions proportionately dilute, or t h e of such deception, either source of error must be in- unavoidable errors from weighing and measuring may vestigated separately and the amount of either error invalidate the results. must be determined, Not until these are known will INVESTIGATIOKS AS T O T H E P U R I T Y O F P R E C I P I T A T E S i t be possible t o pass judgment upon the worth of a The purity of a precipitate may be impaired by two method. different causes-first, b y incorrect stoichiometric INVESTIGATIONS AS T O T H E S O L U B I L I T Y O F P R E C I P I T A T E S composition; second, by foreign substances. I n contradistinction t o the experiments on soluAs mentioned above, precipitates are afflicted with the evil of impurity. If, in order t o determine the bility for which small quantities are employed, the solubility of a precipitate, we would digest a large examination as t o purity requires large quantities, quantity of it with a liquid or a solution, i t is not a t because impurities or deviations from the theoretical all impossible t h a t the parts of the main substance composition are permissible only t o a small extent in and the impurities going into solution stand t o each precipitates which are adapted for quantitative deother in a ratio different from t h a t in which they were terminations. A general scheme of procedure for these researches originally present. In such a case, no matter whether t h e part going into solution or t h a t remaining undis- cannot very well be given, as the variety of cases is solved be determined, the results are misleading. This too great. All t h a t may be said in a general way is PHOSPHORIC
>-
1
June, 1918
T H E J O U R N A L OF I N D U S T R I A L A N D E N G I N E E R I N G C H E M I S T R Y
t h a t the precipitates must be subjected to analysis just as if they were original substances for an analysis. If it is found t h a t t h e composition of a precipitate comes up t o expectation fairly well, it might be sufficient t o establish a factor by which the results are t o be multiplied. However, if this is not the case we must endeavor t o improve the purity of the precipitate. I n general, precipitates are t h e purer, t h e more dilute t h e solutions from which they have separated. Still, too great hopes in this respect should not be entertained, for the ratio of the precipitate t o the substances in solution remains the same on dilution, and it is t h e influence of these substances which, in many cases, prevent the pure separation of the precipitates. The purpose is better accomplished if precipitation is repeated after t h e bulk of t h e first solution has been removed. Especially is this t o be recommended if the precipitation had to take place in highly concentrated solution. If removal of t h e impurities is impossible, nothing else remains to be done b u t t o make determinations with varying quantities in order t o arrive a t an estimation of the amount of impurities, taking into account any losses caused by the solubility of the precipitate. If a t the same time the possibility of incorrect stoichiometric composition prevails, such results, of course, allow more than one explanation. From the aforesaid it is clear t h a n two radically different methods of procedure are necessary t o criticize precipitation methods. The completeness of precipitation and t h e losses on account of solubility must be studied on small quantities. As a matter of fact, a few milligrams of the substance are sufficient in most cases, Simultaneously we learn from this part of the critical work on a method the limit of its applicability for t h e determination of small quantities. I n order t o characterize these conditions I would like t o introduce t h e term “micro-analytical.” On the other hand, examinations as t o purity and correct composition, likewise the final tests with the view t o practical application, must be made with large quantities. For these reasons I would like t o call this part of t h e criticism of t h e method “macro-analytical.” For another reason also i t would be well t o make such distinction. The micro-analytical errors are always absolute losses. Accordingly their corrections will consist in additions. But t h e macro-analytical errors have the form of p r o p o r t i o m and must, therefore, find their corrections in multiplications or divisions. The general algebraic formula for the correction of a result will be thus:
Q X M 1 X h/12
+ m1 + m2
in which Q represents t h e actual quantity of a precipitate, M1 and M2 the factors for the macro-analytical corrections and *a1 and m2 the losses caused by solubility. While methods requiring corrections have not been considered the best, there is no reason why one should hesitate t o use a method after i t has undergone a thorough critical treatment; in other words, after the errors have been scientifically determined. On the
429
contrary, results corrected on a scientific basis deserve more confidence t h a n such as are obtained by methods which are believed t o be reliable, but which never have been criticized in a methodical manner. 25 OGDENSTREET HAMMOND, INDIANA
IMPROVED METHODS FOR THE ESTIMATION OF SODIUM AND POTASSIUM B y S. N. RHUE Received September 19, 1917
A method for the estimation of sodium, involving considerable modification of the procedure of the Association of Official Agricultural Chemists, was published from this department b y Forbes, Beegle and Mensching in Bulletilz 255 of this Institution, under the date of January 1913. Since t h e time of this publication we have made extensive use of this improved method and have devised further improvements, which i t is our purpose to record. The general principles of the method as now used are the same as stated in the earlier publication referred t o above, but changes of detail have been devised which shorten the process and remove certain sources of possible error, a t the same time calling for much less use of platinum. Incidentally, improvement has been effected in the method for the estimation of potassium. For the quantitative test of the new procedures a salt solution was prepared in such manner as t o contain the same kinds and proportionate amounts of the mineral elements as are present in wheat bran. Nitrogen, also, was added to this solution, in the form of ammonium sulfate. The elements and t h e cornpounds in which they were present were as follows:
............................. ............................. ............................... ............................ ............................... .............................. .......................... ..............................
Sodium.. Potassium Calcium Magnesium Sulfur.. Chlorine Phosphorus.. Nitrogen
CzI-IiONa CzHsOK CaHPOdHzO Mgs(POr)t.4HzO HzSOa and (NH4)zSOd HC1 Salts of Ca and Mg (NHn)zSO4
The sodium and potassium ethylates were prepared from the pure metals by dissolving in absolute alcohol and standardizing by titration against benzoic acid. Calculated from the weights of the metals (in air), I O cc. of the solution should have conta’ined 0.0104j g. Na and 0.03926 g. K. The titration against benzoic acid indicated the presence of 0.0103 7 g. Na and 0.037j7 g. K (0.03202 g. sodium sulfate and 0.08373 g. potassium sulfate) in the same volume of solution. The latter weights were used as the basis for judgment as t o the correctness of analytical methods. MODIFICATION
OF THE
METHOD
FOR
SODIUM
I n the use of the method for sodium, as published i n Ohio Agricultural Experiment Station, Bu1leti.n 2 5 5 , we have found much advantage in t h e principle of the second of the optional methods of ashing. The first method proposed for destroying the organic matter, by nitric-sulfuric acid digestion, necessitates the subsequent burning off of much sulfuric acid, in which process there is great likelihood of loss through spattering and overheating. In our later work,
,
