NOTES AND QUERIES ON DR. DUDLEY'S METHOD OF

NOTES AND QUERIES ON DR. DUDLEY'S METHOD OF DETERMINING PHOSPHORUS IN STEEL. O. S. Doolittle, and Alban. Eavenson. J. Am. Chem...
0 downloads 0 Views 732KB Size
2 34

DOOLITTLE A N D E A V E N S O N

[ a ) with one per cent. nitric acitl, dry ant1 weigh, or ( h ) with one per cent. nitric acid and ,I,., per cent. potassium nitrate, tlissolve and titrate ( H a n d y ' s niethod).

T h e greater part of the analytical work of this paper has been done by illy assistant Mr. Geo. 0 . Loeffler. METHODS l~S*~ll For SfeeC a d Pig-iron.-The inethod used in a11 analyses of steel atid pig-iron, was the oiie published by the author in 1892 ( Tmns. Engimcr's Society of W e s f e m Pe.nnsyhania March, 1851.2, nriti]. A n a l . .4$$. Chmz., April, 1892). Foie Fei-roi,iai~~anc,se.-Two grams, dissolved iii nitric acid, ( I . 4 2 ) , and the solution evaporated to dryness. T h e residue, having been redissolved in hydrochloric acid, ( I . 2 0 ) the solution was diluted arid filtered. Tlie filtrate was treated with ainriioniuiii hydroxide till aiiimoniacal, then acidified with nitric acid, heated to 85' C . and precipitated with fift17 cc. of molybdate solution. Foi- /?-on and Manga7rcse Oim.-T\vo granis dissolved in aqua regia (five per cent. H Y 0 , ( 1 . 4 2 ) miti ninety-fire per ceiit. HCl ( I . 2 o ) , and the solution evaporated to dryness. Subsequent procedure was as described above for ferroniaiiganese after t h e evaporation. Phospliorus, in ' ' insoluble residues \vas separated and tleteriiiiiietl. .-lrse?zic n ~ i s added as d r y Xs,O., to the samples Ivhen \wiglied oiit. Tlie reagents used i n the analysis oxidized it to arseiiic ositlr. "

NOTES AND QUERIES ON DR. DUDLEY'S METHOD O F DETERMINING PHOSPHORUS IN STEEL. 0 . 5. ~ O O I . I T T 1 . 1 : . C H K l I I S T , A S D .lLl3.\X 1 < . \ \ - E S S O S . . 1 S S l S T A S T c l i K 3 f l S T ~ ' l i l : A I > ~ L l ' H l . XA S 0 R E A U I S Gl < . X l l , X O ~ I >C O > l I ' X S Y . X..,.Cl\i.!I 1.c """'\

T

H E object of the work outlined

1 ' ) .

Ib?l,

i i i this paper was to ascertain the accuracy of the iiiethod proposed by Dr. Dudley for the volumetric analJ-sis of phosphorus in steel, described in /. Auz. C/zc*m. Soc.! Sept., z893, p . 519, and to tlctcrniirie if a separation of phosphorus could he made in the presence of arsenic with a sufficient degree of accuracy for coiniriercial purposes

235

PHOSPHORUS. I N STEEL

by this method or any modification of it. Our first undertaking was to ascertain if the new ratio between phosphorus and niolybdi~acid in the yellow precipitate of I .go, brought forward by Dr. Dudley, was correct. The commonly accepted factor as found by Emmerton and others is I ,794, Trans. A . Z. 174. E . , 15, 93, R . Finkener, Ber. d. them. Ges., II,1638, Henry Peniberton, Chem. News, 46, 4, Von der Pfordten, Ztschr. anal. Chem., 23, 422, and very recent work by H. C. Babbitt, J , A n a l . A#$. Chem., 7, 1 6 5 , has confirmed this ratio. However, as Dr. Dudley used an entirely different method for determining this ratio, we have followed as nearly as possible his work as outlined in the American Engineer a n d RailroadJournal, Jan. r893, p. 18, but have been wholly unable to obtain his figure. Defevminatioii o j the Ratio bcfsueen Phos~horusand Molybdic Acid i?i the Yellow Precz)itak.-Four samples of steel of very different carbon composition were taken, and the phosphorus determined both gravimetrically and volumetrically in each, and from the figures obtained, the ratio between the phosphorus and the molybdic acid in the yellow precipitate, formed in the volumetric analysis under the conditions used, was computed. T h e complete analyses of the steel used in this and subsequent investigations are as follow : Carboii. p e r cent.

....0.188 0.997 KO. 344.. ..0.588 No.

27

NO.789..

*

No. 6j.. . .o. 166

Sulphur, p e r cent.

0.020

Gravimctric phosphorus, p e r cent 0. I IO

0.226

0.021

0.024

1.295

0.079

0.065

0.107

0677

0.072

0.012

0.02j

hlangaiiese, p e r cent. 0.321

Silicou. p e r cent.

0.427

0.052

For our gravimetric determination we used the combination method as described hy Dr. Dudley. T h e acetate method was followed up to the point of obtaining the basic acetate precipitate, at which stage enough bromine water was added to oxidize fully one-half a gram of iron, and the solution boiled. I t was then cooled d o v a and the acetate precipitation made exactly as described in Chemical Analysis of Iron, by A. A. Blair. This precipitate was dissolved in hydrochloric acid, and all traces of iron washed from the paper. After expelling the hydrochloric acid from the filtrate by repeated evaporations with nitric acid. seventy-five cc. of nitric acid 1.135 sp. gr. were added, the solu-

236

DOOLITTLE AS1) I.:h\:ESSOS.

tion boiled, oxidized n-ith potassium periiiaiigaiiate, aiid reduced n.ith ferrous sulphate, as described iii Dr. Dudley's routine method. T h e solution at S j' C . \vas the11 treated with severityfive cc. of iiiolybtiic acid at 27' C., the !.ello\\- precipitate sliakeii dowii, filtered ant1 lvashed free of molybdic acid. U-c test for .he presence of inolyl~cticacid as follo\ifs; All)outfour cc. uf tlic acid wash water are caught i n ;t test tube as the>- ruii through the filter, and treated ivitli a feiv drops of dilute aiiinioiiiiiiii sulpliicle. T h e least presence of inolylxlic acid will cause the liquid to darkeii slightly. T h e cloud of sulphur produced 1)y tlie ntlditioii of aiiiiiioiiiuiii sulpliitle to tlie acitl solution tiocs iiot interfere i i i an). n a y with the delicacy of tlie test. Il-e have found this test far iiiore sensitive tliaii the oiie used ijy 1)r. I)uciley, ivlio alloirs the acid \ w s h water to tlrc,p into R dilute ,solution of :iiii:iioiiiuiii suipliitle, the test thus 1)eiiig in the alkalilie solutioii. \Ye have frequently fouiid yellow precipitates which accordiiig to the alkaliiie solutioii test \vere thoroughly washed, yet with the acid test shoiv uiiiiiistakatile s i p s of niolj-btlic ncitl which required oiie or two iiiore \vmliiiigs to reiiio\'e. T h e \-ellow precipitate \vas then dissolvetl i i i ainiiioiiia and s:itur:ited while \varni with hydrogeii sulpliiilc.. : i i i ( t the red solution of ~iiolylxleiiuin sulphide slightly aci(iifiec1 \vit!i 1i:;drocliloric acid. Tlic precipitate of iiiolybderiuiii sulpliidc \vns alloivetl to scttlc. then filtered off rapidll- on :i large folded filter. avoidiilg coiitrict \\.it11 the air as much as possible, am1 \vaslicti thoroughly \\-it11strotig hydrogen sulpliide water coritaiiiiiig a little hydrochloric acid. T h e large filtrate \vas tlieii e\.apornteti, aiid 011 1)oiIiiig. it !vas i iivari ably f o uiid that n small a iiiount of mol y 11 tl c II 111i i 51111111i de separated out. This \vas filtered off :iiitl the resu1ti:ig filtrate concentrated to a bulk of ntiout four c c . The pliosphorus \vas then precipitated with from five to eight r c . of iiiag-iiesiainistiire and a siiidl aiiiount of aiiiiiioiiiuiii h!xiroside. the colt1 solutioii being 1.igorously stirred until precipitatioii began, ivlieii aiiimoiiiuiii hydroxide arriouiitiiig to one-tliirtl of tlie l ~ u l koi the solution was added, and the aiialj-sis allo\ved to stnnti lor tivel\-e hours. The precipitate was tlieii filtered off. \\aslied ivitli aiiiiiioiiiuiii iiitrate uiitil free froiii cliloriiie, aiid igiiited. Tliis precipitate was invariably coiitaiiiiiiatetl with iiiol!.lxlic acid. It

23 7

PHOSPHORUS I N STEEL.

was purified by dissolving in hot dilute hydrochloric acid, filtering, evaporating the filtrate to a small bulk, and precipitating as before. For our volumetric determination of the phosphorus in these samples we followed the preceding method up to the point of obtaining the ammoniacal solution of the yellow precipitate. This solution was acidified with sulphuric acid, passed through the reductor and titrated with permanganate, one cc. of which equals 0.003350 grams iron, and assuming as Dr. Dudley has, that the ratio between iron and molybdic acid is 90.76, the corninonly accepted factor, we obtain the results in the table lie 10w

.

No. I' ('

Permanganate used by voluniet. ric method, cubic centimeters. 27 103.1

....

789 .... 19.0 3 4 4 * . . * 59.4 65 11.2

....

Phosphorus by grsvimetric method. per cent. 0.110

Ratio between Error in ratio f o r each 0.001 percent. phosphorus and mol bdic error in gravimetric acidl determination.

I : 1.759 I : 1.826

0.016

0.021

0.06 j

I:

1.813 I: 1.806

0.027

0.012

oh87 o.rg1

A glance at the ratios obtained between phosphorus and molybdic acid in the above table will show that they are very irregular, but this will be explained by the last column of figures which point out that the gravimetric determinations of phosphorus inust be made with absolute accuracy in order to har-e these ratios agree. The only point we wish to call attention to in the above table of ratios is the fact that they are all much lower than the ratio I .go proposed by Dr. Dudley, and obtained by him in a similar manner. Xot satisfied with the above results, we turned our attention to the actual analysis of the j-ellow precipitate, which was prepared for our investigations as follows : 'I'o a solutioii of ferric nitrate in nitric acid of 1.13j sp. gr., was added a sufficient anio~intof phosphoric acid to make the solution correspond to that of a steel containing about one per cent. of phosphorus. T h e yellow precipitate \vas shaken dowi from this, exactly in accordance with Dr. Dudley's niethod of phosphorus analysis. I t n-as washed by decantation with aninioniuin sulphate, then with water, and finally dried to a constant weight at 130' C. This precipitate was found to be so hygroscopic that it n-a.: practically impossible to obtain accurate weights of it for anal\.-

238

I)OOZITTI.E

X S I) E AVE NSO S .

sis. To avoid this difficulty, after carefull). weighing i t , tlit. entire precipitate was dissolved i i i a slight excess of aniiiioiiiuiii hydroxide, filtered through a lialniicetl filter papei into a tnrccl ,700 cc. flask, the filter lieitig washed with water until tlie XIO cc. iiiark was reached. T h e sinal1 miouiit of insoluble matter o i i tlie filter \vas deteriiiined after dryiug at 100' C . . arid the weight tieducted froin the original weight of the !-ellow precipitate. 1 i - e iio\v ha1.e a kiioivn xveiglit of yellou. precipitate i ~ ia kiion.u \veiglit of solution, successive portions of which ivere ~vciglied out froin time to tinie for the deter-iiiitiatioii of phosphorus :iiitl iiiolyhdic acid. Phosjhorus ill Yellow Prrcipifatc.--This \vas tlrteriiiinetl as .iccurately weigh off about tn.ent)--five granis of s u l u tioii. which is equivalent to about one gram of yt.llo\v precipitate. x i d ten cc. of iiiagriesia mixture, stir well until precipitatioii beg-ins, then add aiiiinoniuiii hydroxide to tlie aiiiount of about oiie-third tlie bulk of the original solution, and allow to starid for several hours. P'ilter aiid wash with aiiiiiioriiuiii iiitrnte solutioii until free froill' chlorine, then igiiite lvliile moist. I'urify the magnesium pyrophosphate by dissolving. iii dilute li\-tlrochloric acid, filtering froin any insoluble residue. and precipitatiiig a s liefore by the additioii of aiiiiiioniuiii hydroxide. five cc. of aninioiiiulii chloride solution, and five cc. of inagnesiuni mixture. -Ulow to stand tlvelve hours, igiiite aiid weigh. T h e reagents used were made u p as follons : Arnnioniurri chloride.-One part anirnoniuni chloride t o eight parts of water. Magnesia mixture.--l:roni magiiesiuni chloride a s per Chemical .inalysis of I r o n , p. 58, by A. A . Blair. Arrimoniuin hydroxide.-0.90 sp. gr. Amnioniurri nitrate wash \rater.-One p a r t a m m o n i u m hydroxide 0.90 sp. gr. t o t h r e e parts water, t h e n add t w o g r a m s of ammonium nitrate for each IOU cc. of solution.

Molybdic A c i d in Yellow [email protected] method used for this deterniitiation was as follows : Weigh off accurately an amount of solution equivalent to about 0.25 grain of yellow precipitate, dilute with I jo cc. of water and bring to a boil. Xeutralize the boiling solution with acetic acid. and add about five cc. in excess of the neutral point. Add a t once twenty-five cc. of lead acetate,

PHOSPHORUS I N STEEL.

239

or one cc. for every 0.01 gram of yellow precipitate present in the solution, and allow to digest an hour and a half a t nearly a boiling temperature, in order to render the precipitate compact and granular. Filter on a Gooch crucible and wash with hot water about eight times by decantation. After the fifth or sixth washing the filtrate should give no reaction for lead when tested with ammonium sulphide. Dry to a constant weight a t 130' C. T h i s gives us the weight of P b , ( P 0 4 ) , f PbMoO,. From the previous determination of phosphorus we now compute the P b , ( P 0 4 ) , , and deduct this weight from that of the mixed precipitates. T h e difference gives us the weight of PbMoO,, which multiplied by the factor 0.39237 equals molybdic acid (Moo,). T h e reagents used were made up as follows: Lend Acetate.-Sixty g r a m s crystallized salt i n t w o liters of water, to which fifteen cc. of acetic acid were added. Aceetic Acid.--I .03 specific gravity.

Before using the above method for the deterniinatioii of molybdic acid, we carefully investigated its merits and found it to be thoroughly reliable under all ordinary conditions of precipitation, the acidity, bulk and temperature of the solution having only a slight influence. On attempting to ignite the precipitate, however, as directed in the method as originally published in the Bey. d. clzzm. Ges., 4, 280, we invariably found a loss varying from 0 . 2 per cent. to 3.0 per cent., and even more, according to the duration and intensity of ignition. But by drying on a Gooch crucible at 130' C. we obtained very constant results, which we believe to be iii every way reliable. T h e analysis of the yellow precipitate by the above methods gave us as follows: Phosphorus, 1.648 and 1.644 per cent., an average of I .646 per cent. Molybdic acid, 91.61, g ~ . j g ,91.61 and 91.57 per cent., an average of 91.595 per cent. Our ratio between phosphorus and molybdic acid as computed on these analyses is 1.797. W e deteriiiined the molybdic acid in the filtrates from the pliosphorus deterniinations and obtained 92.2 j and 92.41 per cent. These results are higher than those previously obtained in the separate determinations, but as this difference was not sufficient to change to any extent our ratio we did not investigate the cause.

We now went a step further, in order to prove that the ratio between the phosphorus and molybdic acid in the yellow precipitate actually obtained in a steel analysis, is identically tlie same as that in the yellow precipitate sliakeii down from the prepared solution of phosphorus. W e separated a considerable amount of ellon on^ precipitate froin a steel containing 0.10per cent. phosphorus, 1)). Ilr. Dudley's method of analysis, and after dr).iiig to a constant weight at I 30' C. made the determitiations of phosphorus aiitl niolybdic acicl exactly as in the case of the !>r-eviousprecipitate, obtaiiiing the followi~igresults : 1~1iosp!iorus,1.638 arid I . h 4 S per ceiit.. ail average of I .643 per cent. Xolybdic acid, 91.63 and 91.67per cent., an average of 9 1 . 6 j per cent. T h e ratio between phosphorus and molybdic acid given by these figures is 1 , 7 9 2 , thus confirming our previous results aiid the generallj. accepted figure of 1.794. From the work above outlined it would seem evideiit that tlie ratio of phosphorus to iiiolybdic acid cat1 be derived inuch riiore accurately from the aiial)-sis of the yellow precipitate tliaii by the method suggested by Dr. Dudley. T h e elaborate investigations of Hundeschage~i( Ch2cnz. N i ~ ~ i ~ ~ , Oct. I I , 18, 2 j , and SOY. I , ISS~.) have shown that ammoniuiii phosphomolybdate is far iiiore uiiiforin wlieu prepared under varying coiiditioiis than \\'as foniierly supposed. If the coiiditioiis of precipitation of the yellow precipitate are iiiaintaiiied constant, there is but little doubt as to its uiiiforiii coniposition. T h e ratio then becomes a question of tile accuracl- of the determinations ot pliospliorus and iiiolylxlic acid in tlie !.ellmy precipitate. We have accoidiiigl!. given the iiwthods used for these deterniiiiatioiis as coiiipletel). as possible, in order- to a2ord ail opportunit!- for criticisiii aiitl duplication. R a f i o 6 e f i L a - u 0 . 0 ) ~ ( 1 mi .Ilolqnbtz'ic :Icid--Tlie ratio between phospliorus and molybdic nciil lxiiig establislied, \ye then took up the ratio between iron :wd iiiolylidic acid, as the i n v e h g a C h c u ~ 7, . ~ I h j , aiitl oiir 0 ~ v i 1 tioils of hIr. Babbitt. /.>.liLa/../)Iexlxriencc had indicated that the coiiillioill\~ncccpteil ratio. 90.176, is too high where the reductor is used. tlic reductioii lieing considerably greater than with the ordinary inethod of reducing with zinc powder and shot.

241

PHOSPHORUS I N STEEL.

Our first experiments were on the standardization of a permanganate solution by the ordinary method of dissolving the steel in a flask in an atmosphere of carbon dioxide gas, as compared with the figure obtained when the steel was dissolved in an open beaker and reduced through the reductor. No appreciable difference was found, the reduction being complete in both cases. W e then niade a large solution of molybdic acid of such a strength that fifty cc. contained 0.25 gram of the salt. T h e first experiments with this solution were made after the reductor had been in use some time, and the zinc was packed closely together, so that the solution encountered considerable resistance to its progress, and the acid had some time to act on the zinc, the result being that the reductor became quite hot. Upon cleaning out the reductor and filling with entirely fresh zinc, we were surprised to find our results very appreciably lower, no matter how slowly the solution was passed through the reductor, and only after heating the solution nearly to boiling did we obtain the results in the first case. This demonstrates that care must be taken to have the solutions hot when they are reduced, in order to obtain complete reduction, but with this precaution the results are very uniforin, as will be seen irom the table following. In each of the following experiments fifty cc. of the above mentioned molybdic acid solution was used. Experiments with reductor : RESCLTS EXPRESSED

IN C U B I C C E N T I M E T E R S O F ??ERMANG.qNATE.

REDUCTOR CLOGGED F R O M

REDUCTOR CONTAINXNG F R E S H

LONG USI;.

ZINC.

Five c c . sill p huric ncid, culiic ceiitiiiieters.

Ten c c . sill pliiiric acid, cubic centimeters.

75.25 75.35 75.45

75.2 75.3 75.4 75.3

....

_-

Average, 75.35

Five cc. sulphuric acid, cubic centimeters.

Ten cc. Five cc. sulphuric sulphuric acid solution acid, heated to 95'C.. cubic cubic centimeters. centimeters

....

-

74.8 75.0 74.9 74.95

-_

74.6 74.7 74.7 74.8 -

-_

75.3

74.91

74.70

75.38

75.45 75.35 75.35

Our experience with the reductor has been that in order to secure coinplete reduction, it is necessary to have the column of zinc of considerable length, and to pass the solution through

very slowly. \[?e would not feel safe in using a column of zinc less than ten inches in length. 111reducing :I soliitivii \vliich has a bulk of 2 0 0 cc. at least two iniiiittes h110uld lic: allo\ved for passing t hro 11g11. Turning i i o ~ vto tile retluction oi iiioIy!~lic n i T i ( l l)y liratiiig with zinc ant1 su1I)huric acid, ive iriuml that very different results could be obt:iiiietl !)!. the \xririu> !:io