j16
T H E .70l’RYAL OF I-VD17STRIAL A X D E S G I N E E R I S G C H E M I S T R Y
YO].8 , SO.8
periods sen-ed t o effect a constant weight. Slight differences are undoubtedly produced by the size: weight ant? thicliness of the crucible.
11-Soap Powders proper, a mixture consisting of 1-arious proportions of sodium carbonate and soap: the soaps used are commonly prepared from cottonseed TABLEV-TIX~ R ~ Q U I R Em D HEATINGALUMINA TO CcmsThsr KEIGHT soap stock and h a w as a filler sodium carbonate, Time No. No. No. No HEATEDI N 2 0 - 1 1 1 N . P E R I O D S which itself is a water softener of great value. hlin 1 2 3 4 20 min. 40 min. 60 min. 111- scouring Powders! which contain an abrasive 10 36.9620 29.6670 21.0587 29.6647 29.6538 29.6530 29.652; 20 36.9615 29.6653 21.0570 29.6638 33.7635 33.7627 3 3 . 7 6 2 5 and soap either with or without the addition of sodium 30 36.9605 29.6643 21.0569 29 6638 33,7760 3 3 , 7 7 6 0 . . . . . . . carbonate. 40 36 9609 29 6643 . . . . . 36.9621 36.9602 36.9602 29.6730 29.6730 29.6730 I n this article are described various experiments One precipitate was given a special series of blast- which h a r e been conducted on the soap powders proper. ings t o determine whether or not t h e alumina was The basic scheme used is t h a t given by Leflmann in hygroscopic after t h e blasting. The results (Table Allen’s “Commercial Organic -4nalysis.” During our T I ) indicate t h a t this change in Tyeight was very largely studies. Dr. Leeds published a scheme for the analTABLE vI--sHOD’ISG HYGROSCOPIC PROPERTY OP 241,1-MISA ysis of soap. which is quite similar t o the one used b y Weights us for the analysis of soap powders. Additions and Constant weight obtained (cruc. 1 .p p t . ) . . 21.0598 L e f t 6 days in t h e desiccator. . . . . . . . . . . . . 21.0636 modifications t o the older schemes were made as the Blasted 10 m i n . . . . . . . . . . . . . . . . . . . . . . . . . 21.061 1 After 24 hrs. blasted 20 m i n . . . . . . . . . . . . . . . . 21.060:. work progressed. and it is t o aid t h e soap analyst After 24 hrs. blasted 20 m i n . . . . . . . . . . . . . . . . 21.0604 t h a t important details are described. After 24 hrs. blasted 20 m i n . . . . . . . . . . . . . . . . 21 .(:602 Blasted a second 70 m i n , . . . . . . . . . . . . . . . 21.0598 Left covered in t h e balance case 24 h r s , . . . . 2 1 ,0624 Our purpose in presenting this paper is t o offer o u r experience t o those interested in the analysis of soap due l o t h e precipitate, since a platinum crucible treated we do so because t h e particular problems powders. and in much t h e same \yay did not change appreciably we have found are not discussed in t h e form t h a t I s in weight. desirable for those who have similar problems t o solve, c o K’CL u SI 0 N s I--Boiling for I min. completely precipitates al! aluminum present, and longer boiling may lead t o a re-solution of part of it. 11--Sluminurn hydroxide need not be washed free from ammonium chloride before ignition. 111-The excess of ammonia present when precipitation is made should be as.smal1 as possible and never more t h a n I or z cc. in z j o cc. of solution. IT-Aluminum hydroxide freshly precipitated is soluble t o a slight extent in water and t o about t h e same extent in dilute ammonium nitrate solution. V-The precipitated hydroxide, when large, must be blasted 40 min, t o insure its being reduced t o a constant weight. ’C’I---The ignited alumina is strongly hygroscopic. SoTE-Since this paper was written a paper has been published by IT. H. D a u d t . THIS J O U X S A L , 7 (191 j): 847, confirming what we ha\Te found t o be true regarding t h e presence of ammonium chloride a t the time of ignition. CHEMICALLABORATORY. I:.I.IVERSITY MIXXEAPOLIS
OF
MINNESOTA
THE ANALYSIS OF SOAP POWDERS By LOUISROSENBERG - 4 3 0 VICTORLENHER Receiwd February 7, 1916
During recent years great progress has been made in those industries which supply cleansing materials. The advent of t h e modern mashing powder has brought into the household a coni-enient, economical cleanser. The powders which to-day are in common use divide themselves into three general classes: I-So-called Xashing Pom-ders. composed of trisodium phosphate, borax, or a mixture of various proportions of sodium carbonate and bicarbonate: it is obvious from the composition of these powders t h a t their main efficiency lies in their ability t o soften t h e water used.
3.1 01S T U R E
The ordinary methods for determining n-ater in soap are not satisfactory when applied t o washing powders. Experiments have repeatedly demonstrated t h a t heat,ing a Io-g. sample a t I o j ” C. until the weight becomes constant ‘is worthless, for the odor of decomposing soap is usually observed before this temperature is reached. When heated for I hr. t o 105’. most of t h e soap powders give off a strong odor of decomposed soap. hence the determination is of little value. Further, the soap powder upon heating t o 1 0 5 ’ frequently forms lumps which prevent complete drying: even if a weighed glass rod is used t o break u p the lumps. it is difficult t o keep the sample from caking. i n powders iyhich contain as high as 30-40 per cent water, spattering frequently occurs. Leff mann’s suggestion for soap--that the drying be done on a sand bath--is useless with many powders, inasmuch as even with the greatest care. decomposition takes place. The methods in use for the determination of moisture in a higligrade stearic acid soap are unsatisfactory, as they frequently fail entirely when applied t o soap powders containing cottonseed soap stock or lonT-grade tallow soap. T h e method of %ahrion. in which the sample is heated with three times its weight of oleic acid until a clear solution is obtained. indicating the removal of the water, was found t o be inapplicable, as difficulty was experienced in determining when t h e solution became clear. The recommendation of Fahrion, t o heat the sample (to which has been added oleic acid) in a platinum dish over a free flame, ‘Tiras also tried with the soap powders, b u t even with t h e greatest care decomposition of the soap took place. On a sand bath the powder soon decomposed. Sample I , containing actually 37.8 per cent of water, was used in experimenting mTith oleic acid in order t o test this method for moisture. Six samples of 2 g. each were
T H E JOCR,VAL O F I N D 1 7 S T R I A L A X D E,VGISEERING C H E M I S T R Y
Aug., 1916
mixed with 6 g. of anhydrous oleic acid. On heating, the odor of decomposing soap was noticed from all. The two samples showing t h e clearest solution and t h e least odor indicated j 6 . j per cent and 60.3 per cent water, respectively, or a n error of j o t o 60 per cent more water t h a n they actually contained. The method recommended is one in which t h e heating can be done a t low temperature. A vacuum drying oven is admirably suited for this purpose. The various powders examined in our studies were heated between 60 and 6 j 0 a t a pressure of 60 mm. until t h e weight became constant. I n the 2 - 8 . samples used, constant weight was attained only after I O hrs.' heating, and in some cases 2 0 hrs. of heating were required, No odor was observed and the samples did not cake. The results (percentages) were as follou~s: S o . I , 3 7 . 8 j and 37.82; N O . 2 , 23.7j and 23.99. The recommendations commonly given t o heat for 1 2 hrs. are insufficient. S o . 3 ) for example, required 2 0 hrs.' heating in the vacuum oven before constant weight was attained. P E T R 0 L E U &I E T H E R E X T R.A C T I 0 X
The usual procedure of extraction Ivith petroleum ether, with a Soxhlet extractor, was used in order t o determine t h e unsaponified fat, unsaponifiable matter from t h e f a t and free f a t t y acids. A Io-g. sample of t h e soap powders frequently required as much as 1 5 hrs. of extraction with petroleum ether. The solvent was subsequently evaporated and the residue weighed. ,411 of the powders examined were found t o contain less t h a n I per cent of unsaponified matter. F A T T Y ACID D E T E R M IKAT1 0 li
The determination of t h e f a t t y acids in the soap powders is, of course, of the greatest importance, inasmuch as it means t h e soap content of t h e powder. Of t h e various methods which have been suggested €or t h e determination of t h e f a t t y acid content in soaps, the most important are: ( I ) Direct weighing; ( 2 ) petroleum ether extraction; (3) potassium soap; (4) volumetric. Fendler and Frank' have shown the inaccuracies of all of these methods. I n t h e direct weighing method, a piece of weighed paraffin or stearic acid is added t o t h e liberated f a t t y acids, and the cooled solidified mass is subsequently weighed. The results are frequently very high. Extraction of t h e f a t t y acids with petroleum ether and evaporation of the solvent commonly gives low results. The variation in t h e results depends largely on length of drying of t h e f a t t y acids obtained. I n one instance the percentage of f a t t y acids was reduced nearly 60 per cent b y long drying. Fendler and Frank preferred for accuracy t h e potassium soap method, b u t inasmuch as the potassium soap must be dried for 2 0 t o 28 hrs. o n a water b a t h , t h e method is rather too time-consuming for ordinary commercial analysis. 1 7 0 1 ~ metric methods can be said to give only fair results. The cake method is a simple, rapid method; b u t unless due precautions are taken, discordant results ~
Z angezo. Chem., 22 (1909), 252-61.
717
are obtained. The inaccuracies of the cake method, according t o Noerdlinger, are attributed t o the stearic acid used t o collect the f a t t y acid. If the stearic acid is previously heated t o 160' the inaccuracies are avoided. I n our experiments t h e cake method was adopted as t h e most satisfactory procedure, and was carried out as follows: ,4 2 - 8 . sample of the powder is placed in a tall, narrow, 200-cc. beaker 4 in. high, and the sample dissolved in hot water. An excess of ,'OZ nitric acid is then added t o separate t h e f a t t y acids. The solution, stirred continuously, is heated for '/2 hr. in boiling water, and a weighed portion of 4 g. of stearic acid, previously heated t o 160°, added. The solution' is stirred repeatedly and warmed for another hr., after which the beaker and contents are allowed t o cool. The stirring rod, which should be about I in. longer t h a n t h e beaker, is allowed t o remain in t h e beaker. The f a t t y acid and the stearic acid on cooling solidify. and can be easily lifted out of the beaker b y means of t h e stirring rod. The liquid in the beaker can be filtered in order t o collect any solidified fat which has not adhered t o t h e cake, b u t usually this amount is small. The circle left b y the cake around the beaker can be readily scraped off and added t o t h e cake. The cake is then wiped with filter paper and placed upon a weighed cover glass, allowed t o dry for several hours, after which it is weighed. The weighed cake less t h e weight of t h e stearic acid added gives t h e amount of f a t t y acid in the sample. If t h e melted fat is allowed t o cool rather slowly, a more uniform cake is formed, and t h e cake can be more easily removed from t h e beaker; whereas if the cake be cooled quickly, t h e rapid solidification causes moisture t o be enclosed in the cake. If cavities are formed, these are likely t o enclose t h e solution. A zoo cc. beaker of t h e tall, narrow type is a convenient size, inasmuch as a cake suitably fitted t o t h e balance pan is formed. Furthermore, t h e addition of mineral acid t o a washing powder commonly causes much effervescence, due t o Iarge percentages of sodium carbonate: the use of a tall beaker lessens t h e danger of overflow during this sudden efferve-ccence. TOTAL A L K A L I D E T E R M I N A T I O N
Should t h e percentage of total alkali be required, t h e excess of standard acid required t o liberate the f a t t y acid can be titrated back with N / 2 alkali. I n washing powders, the total alkali reported does not have t h e significance t h a t it does,in soap. The presence of sodium carbonate marks the amount of alkali combined as soap. A total alkali determination cannot be used as a means for determining the relative merits of different powders, since t h e powder with the highest per cent of sodium oxide may be the one which contains t h e lowest per cent of soap. A total alkali determination is unnecessary, unless it is desirable t o check the individual alkali determinations. S ODIU Y C H L 0 R I D E D E T E R MI N A T I 0 K
I n our hands t h e gravimetric method is preferable over t h e volumetric method for t h e determination of chlorides in soap powders.
T H E J O I ' R N A L O F IA1DlTSTRI.4L i l S D EA'GIXEERIiYG C H E M I S T R Y
;I8
I
A LC 0 I3 0 L E X T R A CT I 0 K
TABLEI-SCHEME
EXTRA cT Contains Uncombined Fat. Evaporate t h e ether, dry a t llOc C. and weigh.
8
D E T E R 111S A T I 0 S 0 F S 0 D I U RI C A R B 0 S A T E A S D S I L I C A T E
I n order t o separate the soap content of the powders from such fillers as sodium carbonate, sodium silicate, borax, etc., it is necessary t o make an extraction with alcohol. Due t o t h e fact t h a t soap pomTders contain a high percentage of water, it is preferable t o use absolute alcohol. If, however, the analysis is t o be made on a sample which has been dried in t h e vacuum oven, 95 per cent alcohol is satisfactory. A 2-8. sample of t h e powder can be extracted in a hr. T h e 250-cc. beaker with IOO cc. of alcohol for alcohol becomes yellow colored and a white residue of the inorganic salts, such as sodium carbonate, sodium silicate, and borax, separates out and can be filtered off and washed with alcohol. I n this extract, phenolphthalein will show t h e presence of caustic alkali. No free alkali was found in any of t h e powders examined. On t h e other hand, all of the soap powders showed a trace of free f a t t y acid. The alkali combined as soap is found b y adding a large excess of water t o the alcoholic solution and boiling off the alcohol, decomposing t h e soap with an excess of N/2 nitric acid and titrating back with iV/z alkali. The alkali found is reported as NasO. N o distinction need orWeigh out a 2-gram sample.
1-01.8 . l o
D r y in vacuum oven a t 65' C.
RESIDUEconsists of Soap a n d Mineral Constituents. EXTRACT consists of Soap and free Alkali. (1) Add 2 or 3 drops of phenolphthalein, and titrate with N/10 acid; the amount used corresponds t o free caustic alkali: calculate as NanO. ( 2 ) Kext add a large excess of water, boil off t h e alcohol, decompose with excess of N HKO8 and boil: add 4 g. stearic acid, boil again, and cool. SOLUTIONcontains Soda CAKE contains F a t t y combined as Soap. TiAcid. Remove cake, dry and weigh trate excess of acid back with X/2 N a O H . K acid used corresponds to IianO combined as Soap.
dinarily be made between the sodium and the potassium in soap powders. The alcoholic extract should be diluted t o a t least 2 j o cc. before titrating in order t o get a satisfactory end-point. A rapid method for analysis of soap powders is based on t h e alcoholic extraction procedure: A 2-g. sample is extracted with a.bsolute alcohol, the solution filtered and the filtrate tested with phenolphthalein for free alkali. If t h e solution contains free alkali it is titrated with N 2 nitric acid: the alcohol is then evaporated off, leaving a residue of dry soap. which, if it has not been necessary t o add nitric acid. can be weighed and reported a s . percentage of total soap. T o determine the alkali combined as soap, the dry residue of soap is dissolved in water and the solution 2 nitric acid, using methyl orange as the indicator. The residue from the alcoholic extraction, which consists of sodium carbonate. sodium silicate, borax, etc., can be tested by t h e methods given belom-. The percentage of total soap plus sodium carbonate, etc.. deducted from I O O per cent will gix-e the approximate amount of water. or, an accurate water determination can be made in the vacuum drying oven.
The residue from t h e extraction with alcohol is dissolved in water and diluted t o exactly 2 5 0 cc. in a volumetric flask. The residue from all of the samples analyzed was completely soluble in hot water. In one or two powders a slight residue remained upon t h e filter paper, b u t t h e amount was too small t o warrant further examination. Aliquot portions of j o cc. are taken for t h e silica determination. The solution is acidulated with hydrochloric acid and evaporated t o dryness, as in the ordinary method for the determination of silica. The acidification and evaporation are repeated. The soluble matter is extracted with hot water and t h e residue when dried and ignited is weighed as SiO,. The percentage of silica was calculated i o sodium silicate ( KanSi4Os). I n t h e soap powders analyzed the amount of silica found was so small t h a t it was probably present as a n impurity rather t h a n as a real constituent of the powder. Borax was not found t o be present in any of the powders analyzed. I n t h e absence of borates and n-ith less t h a n I per cent of sodium silicate, a fair determination of t h e sodium carbonate can be made b y direct titration with M , I ~ hydrochloric acid, using methyl orange as the indicator. OF ANALYSIS Loss in weight = Moisture.
Extract with petroleum ether
Extract with 95 per cent alcohol. RESIDUEconsists of NanC03, NaC1, NanSiiOs, NalS01, N a r B l O i , Starch and Insoluble Matter. Dissolve in hot water, dilute t o 250 cc and filter: divide the filtrate into five equal parts. I-XanCOJ: Titrate with L / zHC1. RESIDUE 2-NaCl: Precipitate with AgN03 and weigh as Consists of Starch and AgC1. Insoluble Matter. 3-NagSir08: Add HC1 and evaporate t o dryness; repeat. Add water, filter, dry, ignite a n d weigh as SiOn. Calculate t o SanSirOs. 4-SapSO1: Precipitate with BaClz and weigh as BaSO;. Calculate to iYanSO1. If present, determine by glycerol or 5-Borates: mannitol method.
Should sodium bicarbonate be present in the powder, the treatment of the residue insoluble in petroleum ether and in alcohol. would necessarily need t o be materially modified. SCHEME OF A S A L Q S I S
The condensed scheme of analysis outlined in Table I is suggested for soap powders. I t is based on t h e solubility of soap in alcohol and the insolubility of t h e inorganic constituents. I t possesses the advantage t h a t but one sample need be taken for ana.lysis, and the soap itself is separated a t once from the other constituents of the powder. I n the simple scheme, as given. such substances as fluorides. rosins, glycerin. sugar, and sodium bicarbonate are not provjded for. Our experience with the powders examined showed them t o be of simple composition. UISCUSSIO?;
OF RESULTS
The analyses of a number of representative soap powders are given in Table 11. Their composition in a number of cases differs materially from t h a t given for the same powder by Braggl. This difference is doubtless due t o t h e introduction of the continuous 1
Chemical Ewgineer, 18 (1913), 73.
Xug
,
1916
T H E JOCR.VAL O F I . V D 1 7 S T R I d L A N D E Y G I S E E R I N G CH E M I S T R Y
drying apparatus which has changed t h e formula used by a number of the larger manufacturers. T h e continuous drying apparatus allowed the manufacturer t o TSABLE 11-TABULATION O F RESULTS (PERCENTAGES) Sample No. 1 2 3 4 5 6
7
Fatty Mois- Anhy- NazO ture dride a s Soap 3 7 . 8 2 18.39 2 . 6 3 22.06 19.44 3 . 5 8 23.75 21.63 2 . 7 1 20.86 18.93 2 . 6 5 18.52 20.34 2 . 6 8 43.38 16.74 2 . 2 3 30.40 21.46 2 . 8 6
UnsaTotal ponified Soap h'a2C03 NazSLOe NaCl M a t t e r 0.45 21.02 39.37 0.55 0.69 23.02 53.53 0.99 0.75 0 . 2 4 24.34 50.35 0.72 0.35 0.85 2 1 . 5 8 55.38 0.90 0.86 0.45 23.02 56.45 0.36 1.32 0.74 18.97 3 6 . 3 0 0.72 0.34 0.87 24.32 4 0 . 0 0 0.99 0 72 0 . 7 9
Total Per cent 99.90 100.59 100.36 100.03 100.41 100.58 97.22
increase the water content of t h e product a n d to lessen the amount of soap. Previously this addition,of water resulted in an inferior caked powder. The determinations made show t h a t the soap powders consist essentially of sodium carbonate and soap. This should be the case from t h e method of manufacture employed, i. e., t h e soap stock, caustic alkali and soda are incorporated and the product prepared similar t o t h a t of soap. After hardening it is ground into the powder form in which it appears on the market. UNIVERSITY OF WISCONSI.I-. MADISON
THE DETERMINATION OF CHROMIUM AND VANADIUM IN STEEL BY ELECTROMETRIC TITRATION By GEORGELESLIE KELLEYA N D JAMES BRYANTC O N A N T Received January 12, 1916
The method t o be described includes t h e solution of the steel in acid, t h e oxidation of t h e elements in question under suitable conditions, and t h e titration of t h e solution with ferrous sulfate using the electrometric method of determining the end-point. This method of titrating chromates has been described recently by Hildebrand' a n d by Forbes and Bartlett,2 and has been applied b y t h e authors t o t h e determination of ~ a n a d i u m . ~T h e apparatus is in all respects identical with t h a t described elsewhere for the determination of vanadium and therefore will not be dealt with here. DETERMINATION
OB C H R O M I U M Ii% S T E E I
I t was hoped t h a t b y t h e application of t h e electrometric method t o t h e determination of chromium, a n accurate and rapid method could be devised for the determination of this element. T o this end t h e oxidation of chromium salts t o chromates was undertaken after t h e methods suggested b y W a l t e r ~ Rich6 ,~ a n d Tusker.6 These methods were selected because of their simplicity and because they involve no filtering. I n these methods ammonium persulfate is used to oxidize chromium and manganese t o chromate and permanganate, sometimes with t h e addition of silver nitrate. Hydrochloric acid, added in small amounts t o the hot solution, serves t o reduce t h e permanganic acid, b u t is without action on t h e chromates. On trying the different modifications of this method which have been published, extremely variable and unsatisfactory results were obtained. This experience 1 2
3
L 5 6
J . Am. Chem. SOC.,35 (1913), 869 I b i d . , 35 (1913), 1527. I b i d . , 38 (19161, 341. M e f . Chem. Eng., 12 (1914). 310. Ibid., 13 (1915), 239. C h e m . - Z f g . ,39 (1915), 122.
719
led us t o undertake a more careful study of t h e conditions necessary t o make t h e method work with certainty. I t seemed possible t o us t h a t t h e following 'errors might interfere with the successful working of the method: ( a ) Incomplete decomposition of ammonium persulfate, ( b ) incomplete oxidation of chromium, (c) reduction of part of t h e chromate b y hydrochloric acid during t h e breaking u p of permanganic acid, and ( d ) incomplete reduction b y the hydrochloric acid of manganese in the higher states of oxidation. ( a ) a n d ( d ) would t e n d t o make the results too high, while ( b ) and (c) would have the opposite effect; thus it might be possible for some of t h e errors t o exist undetected because of counterbalancing. We therefore planned t o test each point separately. D E C 0 RIP 0 S I T 1 0 N 0 F A M 410 K I U M P E R S U L F AT E
Eight solutions of 300 cc. volume, each containing g. of chromium as chromate and 65 cc. of sulfuric acid (sp. gr. I . zo), were heated t o boiling and j g. of ammonium persulfate added. The solutions were then boiled in groups of two for j, I O , I j and 3 0 min., after which they were cooled and titrated with ferrous sulfate, using t h e electrometric end-point. The potassium dichromate solution had been standardized against ferrous sulfate electrometrically and this in t u r n had been compared with standardized potassium permanganate solution. From t h e experiments shown in Table I , it i s evident t h a t j min'. boiling of such a solution does not 0.0102
TABLE I-THE
DECOMPOSITION O F AMXONIUM PERSULFATS 0.0102 G . Chromium Present in Each Case 10 15 30 Time of Boiling (min.), , , , , , , , . . , , . , , 5 0 , 0 1 5 9 0,0101 0.0095 0.0096 Chromium Found (g.).. , . . . . . . . . . . . . 0 . 0 1 9 6 0.0117 0 . 0 0 9 4 0.0082
{
completely remove t h e ammonium persulfate and t h e oxidizing agents into which it decomposes. Even I O min. is insufficient in some instances. Boiling for a much longer time has t h e apparent effect of reducing some of t h e chromate, for t h e values are uniformly lower t h a n the expected values. I t is not improbable t h a t this reduction goes on from t h e first, b u t its effect upon t h e titration is masked b y t h e fact t h a t other oxidizing agents have not been removed. This may be explained as due t o t h e breaking down of t h e ammonium persulfate in acid solution t o give Caro's acid and hydrogen peroxide, both of which reduce chromium. We believe t h a t any method depending upon ammonium persulfate alone for its oxidizing action can give correct results only b y a balancing of errors. This opinion is supported by other experiments in which t h e sulfuric acid mentioned above was changed t o higher and lower concentrations, a n d in which it was partly and completely replaced b y nitric acid. DECOMPOSITION
OF
AMDIONIURI
PRESENCE
PERSULFATE
IK
THE
OF SILVER KITRATE
h4arshal11 suggested t h e use of ammonium persulfate and silver nitrate in dilute nitric acid as a n oxidizing agent for manganese. Experiments parallel 1
Chem. S e w s , 83, 16.
