Determination of Traces of Boron in Nickel

(5) Glock, E., Tobacco Chemists' Re- search Conference, New Haven,. Conn., October 1957. i6) Gutsche, C. D., in“Organic Re- actions,” Vol. VIII, R...
4 downloads 0 Views 260KB Size
11. E., ANAL. CHEM. 29, 105 (1957).

( 5 ) Glock, E., Tobacco Chemists’ Re-

search Conference, Sen- Haven, Conn., October l?57;‘ 16) Gutsche, C. D., 111 Organic Reactions,” Vol. VIII, R. .4danis, ed.. n 364. \Tilev. ” , Sew Toik. 1954.-

1,i)Insull,

R.,James, A. T., Divisions of Analytical and Petroleum Chemistry, Symposium on Advances in

Gas Chromatography, 132nd Meeting, ACS, New York, PT. Y., September 1957. (8) Jacobs, T. I,., in “Heterocyclic Compounds,” R. C. Elderfield, ed., p. 72, Wiley, New Tork, 1957. (9) James, A. T., Martin, A. J. p.,

Bzocheiii J 50. 670 (19521 - (10) Ibid., 6 3 , 144 (1956). i l l ) Keulemans, A. I. >I., Iiwantes, A., Rijnders, G. W, A., Anal. Chim. Acta 16, 29 (1057). -

-

I

~

~

\

3 -

(12) Kosak, -1. I , Ezpeizentza 10, I39 (1954). (1:3) Resnik, F. E., Lee, L. -1, P o n d , JV. A , L4NA1J. CHEN 27, 028 11955). R ~ C E I V Cfor D rrview March 10, 1958. Accrnteti M a~ r . .12 1958 Southeastern RegiGnal Meeting, ACS, -Durham, N. C., Sovember 1957. Work supported in part by grants from The Damon Runyon %lemo;,ialFund. ~

I

Determination of Traces of Boron in Nickel C. L. LUKE Bell Telephone Laboratories, Inc., Murray Hill, N. J.

,Traces of boron in nickel can b e determined b y dissolving the sample in a small amount of a hydrochloric acid-platinic chloride mixture, isolating the boron by methanol distillation, and determining it by the photometric curcumin method.

s IVGESIOUS iiietliod for the tletermination of traces of boron in pure nickel has been proposed by Chirnside, Cluley, and Proffitt ( 1 ) . However, its applicability is someivliat limited by the fact that the sample to lic analyzed must be in sheet or rod form. Moreover, the method is ob~ i o u s l ynot designed to be used for the snalysis of all types of nickel, because 110 attempt has been made to remove 4 c o n and certain other metals which are known to interfere, when present in more than trace amounts, in thc photometric curcumin method for boron 13). To provide for nider applicability in the analysis of nickel a iiiethod has been developed in nhich the -miiple is dissolved under it reflux ihondeiiser in dilute hydrochloric acid plus a little platinic chloride; the boron IS isolated by mrthanol distillation and tlcterniinrd bv the photometric curI uinin iiiethod (3, 4). EXPERIMENTAL

T h e appaiatus (Figuie 1) consists of a 100-nil. standard-taper quartz conical flask n i t h a loiv-boron glass (Corning KO. 7280) air condenser. Apparatus.

Preparation of Calibration Curve.

Picpaie a calibration curve as directed ror drterniiiiation of boron in silicon ( i ) . hut add 2 drops of 0.5% rather than 2 ml.of 57’ sodium hydroxide solution to the aliquots of standard boron solution. Dissolve the alkali precipitate in 0.5 ml. of hydrochloric acid-platinic chloride solution [mix 2.5 nil. of I yGpla-

tinic chloride solutioii nitli 500 nil. of hydrochloric acid (1 I ) ] plus 0.5 ml. of n ater. Wash the solution into a 100nil. quartz flask nitli the aid of 25 ml. of redistilled inethaiiol. Add 1 drop of 0.1% methyl orange solution and ammonium hydroxide (specific gravity 0.90) dropwise until the pink color just disappears. Add hydrochloric acid (1 1) dropwise until the pink color reappears, and then add 3 drops in excess, Proceed to the distillation and photoniptric determination as dirccted (.$).

+

+

no higher tliaii ail inch ahol e the top of the flask. \Then solution i. complcte, cool the flask to room temperature, n ~ s h don 11 the condensed acid n ithin the air condewer into the flask n-itli 25 inl. of lon--boron methanol, and reiiiove the condenser. Add 1 drop of iiiethyl orange iiitlicator solution. Proceed as directed for preparation of calihrc‘1 t’ion curve. Carry a reagent blank through the entire analysis. With t h r aid of the calibration curvr determine the \I cight of boron in the \ample and rmgpiitq. DISCUSSION AND DATA

4

5c c u 9 CO’vDENSiR,

---

-

U CRO BdRNE9

-I

The m e t l i d r i m be used to dctcxrmine acid-soluble boron in iiiost nictals 1s that are soluble in Iiydrochloric. acid. To determine n lictlicr the boron i n iiickcl is likely to lie :itlid to check the accuracy of nicthod, a 0.03% hoioiinickcl alloy n a s prepared. Boron nictal was dissolved in molten nickel under a blanket of helium. A 0.1-gram portion of this sample n a s dissolved in 1 nil. of hydrochloric acidplatinic cliloridc solution. The solution n a s dilutcd to 250 ml. in a. dry volunictric flask n ith Ion -boron nictha n d . A 5-nil. aliquot of this solution, 20 nil. of loll-boron nipthanol, 0.5 nil. of h\-tlrochloric acid-platinic caliloride d u t i o i i , 0.5 nil. of nater. and 1 drop of

7p-Y Figure 1. denser

Flask with air con-

Table I.

Determination of Boron in Nickel

Salllplc T r a i i s f t ~0.100 gram of the subdivided saniple t o a d r y 100-ml. quartz conical standard taper flask i Figure 1). Add 1 nil. of hydrochloric acid-platinic chloridc solution, cap with a lowboron g1:i.s air condenser whose iiiale joint has been wetted with 1 drop of the acid mixture, and heat with a iiiicroburner so that the acid condenses Procedure.

Sivac Xivac 999 999

-1-30 -4-30

BTL-C.1-447 BTL-C.\-447

223 ‘225

Boron Fouritl, l’.P.lI. 0 0 0 0 0 0 2 2

5 6

VOL. 30, NO. 8, AUGUST 1958

2 3 6 6 7 8 0 4 6 0

1405

mctliyl orange indicator solution were transferred to a 100-ml. quartz conical flask. The boron was then determined by the proposed method. The value of 0.029% of boron, obtained on both of duplicate samples, indicates that the method is satisfactory and that the boron in electronic nickel will probably be acid-soluble. Boron was determined in duplicate in

several types of nickel by the proposed method -(Table I), The method has also been used to measure the contamination of the surface of vacuum tube cathodes by boron during sealing of the glass envelopes. By dissolving successive layers of the nickel cathode it has been possible to establish the extent of diffusion of the contaminating boron into the nickel cathodes of tubes that have been on "life test."

LITERATURE CITED

(1) Chirnside, R.

C., Cluley, H. J.,

Proffitt, P. M. C., -4nalyst 82, 18

,-"-.

114571 ,.

(2) Luke, C. L., . ~ N A L . CHEM.27, 1150 (1955). (3) Ibid., 29, 1227 (1957). (4) Luke, C. L,,Flnschen, s. s., AINAL, CHEM.30, 1406 (1958). RECEIVED for reVieTT Xovember 29, 1957. Accepted March 21, 1958.

Photometric Determination of Traces of Boron in Silicon After Separation by a Hydrothermal Refining Technique C.

L. LUKE

and S. S. FLASCHEN

Bell Telephone laboratories, Inc., Murray Hill, N. 1.

,The sensitivity of the photometric curcumin method for the determination of boron in silicon has been increased b y using a hydrothermal refining technique for the separation of the boron from the silicon. It is now possible to increase the sample size tenfold and to determine as little as 0.02 p.p.m. of boron, Because variables are controlled more closely during color development, the accuracy of the photometric determination itself i s improved. The hydrothermal refining process appears to be a powerful new analytical tool that should prove useful in other analyses.

S

methods have recently been proposed for the determination of traces of boron in silicon (1-7). The applicability of the photometric curcumin method (4) is limited because of its relatively low sensitivity. T o obtain additional sensitivity the method has been modified to permit the analysis of larger samples. Because the separation of borate from silicate by precipitation with methanol is unattractive with large samples, a new method of separation has been developed. The crystal chemistry of silicon and its oxides is such that maximum segregation of impurities would be expected if the silicon could be converted, in the presence of water, into crystalline quartz instead of sodium silicate. It seemed that this might be accomplished by oxidizing the silicon with water at high temperature and pressure to silica EVERAL

Si

+ 2H20

-P

Si02

+ 2He

and ultimately converting, in situ, 1406

ANALYTICAL CHEMISTRY

to crystalline quartz through hydrothermal recrystallization. Experimental investigation of the proposed hydrothermal method showed that the oxidation of silicon b y water alone was prohibitively slow, but was satisfactory in the presence of a small amount of sodium hydroxide. The beneficial action of the alkali is probably due to the tendency to form soluble silicate Si

+ 40H-

-.+

SiO44-

OCCE

ThlCK

L '"ER 0 03 THICK

+ 2H?

thus preventing the formation of a passive oxide film on the surface of the silicon. As the reaction proceeds to completion

+ 2H20

Si044-

+

SiO,

+ 40H-

the hydroxyl ions are regenerated. This regenerative action makes possible rapid and complete oxidation of the silicon with much less sodium hydroxide than is required to convert the silicon to sodium silicate (-4). One gram of 60-mesh silicon can be converted into crystalline quartz by heating in 14 ml. of 0.57, of sodium hydroxide solution in a platinum lined autoclave for about 5 hours at 350" C. and about 5000 p.s,i. The boron, as borate, plus a little silicate remains dissolved in the mother liquor. Because the hydrothermal recrystallization of the quartz is rapid and the solid solubility of boric oxide in the quartz lattice is negligible. the impurity distribution coefficient between the quartz and the aqueous solution is extremely favorable. Quantitative separation of the boron from the bulk of the silicon is assured. The boron in the mother liquor can

Figure 1.

Steel autoclave

be isolated by a methanol distillation and then determined by the photometric curcumin method. The hydrothermal refining separation makes it possible to analyze 1-gram samples of silicon with no increase in the size of the reagent blank. Thus the limit of sensitivity of the photometric boron determination can be extended to 0.02 p.p,m. n-ithout decrease in accuracy. APPARATUS

Steel Autoclave. T h e autoclave (Figure 1) was made in these laboratories from a 3-inch rod of Timken 17-228 bolting steel (Timken Roller Bearing Co., Canton, Ohio). T h e cavity and adjacent shoulder were lined with platinum and capped, during t h e hydrothermal operations, b y a 13/&nch disk of 0.006-inch silver. The