3 4MSb 03bOS7D 5

CENTRAL RESEARCH LIBRAS*POCUMJSNX COLLECTION

ORNL-599

Chemistry-General

&4*

W*

ESTIMATION OF TRIBUTYL PHOSPHATE (TBP)

IN MIXTURES OF TBP-VARSOL

OAK RIDGE NATIONAL LABORATORYOPERATED BY

CARBIDE AND CARBON CHEMICALS DIVISIONUNION CARBIDE AND CARBON CORPORATION

porr office box p

OAK RIDOE, TENNESSEE

n

rs

ORNL-599

This document consists of

13 pages.

Copy T ofH^ Series A.

Contract No. W-7^05 eng. 26

Estimation of Tributyl Phosphate (TBP)

in Mixtures of TBP-Varsol.

M. T. Kelley, P. F. Thomason,#A. D. Horton,and J. L. Horton

Date Issued:FEB 2] ]950

Chemistry Division

OAK EIDGE HATIQHAL LABORATORYOperated by

CARBIDE AND CARBON CHEMICALS DIVISION

Union Carbide and Carbon CorporationPost Office Box P

Oak Ridge, Tennessee

*Instrument Department, Oak Ridge National Laboratory,

i^|AmiTJN MARIETTA ENERGY SYSTEMS LIBRARIES

3 44SL 03bD57D E

*Estimation of Tributyl Phosphate (TBP) in Mixtures of TBP-Varsol «

Abstract; Two methods for the estimation of tributyl phosphate

(TBP) in mixtures of TBP-Varsol are presented. The first

procedure involves the colorimetric analysis for phosphate following an

acid or a micro combustion bomb decomposition of the mixtures. The second;

and preferred method depends upon the measurement of the dielectric constant

of the mixtures.

Introduction:

Colorimetric Phosphate Method. The Technical Division of

(I)ORNL has studied^ ' the recovery of uranium from metal waste solutions

using TBP-Varsol mixtures. During their investigations it was frequently

necessary to determine the amount of tributyl phosphate in the solvent mix

ture. Therefore a rapid and accurate control method was desired.

Experimental; Since there is a good colorimetric method available for

phosphate, it was decided to attempt to effect complete

(2)decomposition of the mixture by the method recommended by Snell and Biffen .

This was accomplished by boiling a volume of the TBP-Varsol mixture, contain

ing approximately 50 mg of tributyl phosphate in a 150 ml beaker, with 5 ml

of concentrated HNO^, 5 ml of concentrated ^SOjp and one gram of Na^SO^ as

a catalyst. The volume was reduced by evaporation to approximately 2 ml.

The beaker and watch glass cover were rinsed down with 25 ml of distilled

water and 5 ml of concentrated HN0o« The solution was boiled for five

minutes, cooled, and made up to 100 ml with distilled water. One hundred A

and 200 A aliquots were analyzed by the familiar molybdenum blue colorimetric

methodVJ/ for orthophosphate. The percent by volume of TBP in Varsol was

*A1though Varsol is a trade name for a particular solvent, it is used ge-nerically in this report to denote solvents of this nature.

-3-

calculated from the total phosphate found. The results of several

synthetic samples are shown in Table I.

Table I

Phosphate Analyses of 15$ TBP-Varsol Solutions

Sample NOo Weight of TBP Calculated PO^" P0^~ Found Per cent(Grams) Present (Mg) Error

(Mg)

1 .<*56 16.28 16.50 +1.352 .OkQk 17.28 16.80 -2.78

3 .01+82 17.21 16.80 -2.38k .0476 16.99 16.80 -1.11

5 .01+60 16.1+0 16.50 +0.61

It was found that the sodium peroxide micro combustion

method of Elek and Hillv ' was equally effective in decomposing the organic

mixtures and was somewhat faster.

Dielectric Constant Method. It was felt that a measurement

of some physical property of the mixtures would offer a better routine

analytical method. Although there are differences in specific gravity of

the mixtures, the differences in the range used {% - 2% of TBP) are too

small to permit accurate estimation of the TBP. Optical properties such as

refractive index, and ultra-violet absorption did not disclose a good method

of estimation. As the dielectric constants of Varsol and TBP are quite

different, it was believed that a measurement of this property would give

an accurate estimation of the TBP in these mixtures.

Preliminary measurements with a "Q meter" showed that different

mixtures gave a decided range of readings. Therefore, a dielectric constant

meter was designed and built. A photograph of the instrument is shown in

Figure 1. It consists of a crystal controlled oscillator operating at 1000 kilo-

-k-

FIG. I

DIELECTRIC METER

NOT CLASSIFIEDPHOTO 5963

cycles loosely coupled to a high Q tank circuit which can be tuned to resonance

by a combination of two condensers. Resonance is indicated by a maximum read

ing on a vacuum tube voltmeter. The cell to contain the samples is a small

copper box containing a 50 mmfd. midget condenser^ which is connected in

parallel with the tuning condensers with a length of coaxial cable. This

instrument in some respects is similar to the Q-meter manufactured by the Boonton

Instrument Company# but is considerably simplified^ since it is designed to

measure capacity only. The circuit diagram of the instrument is shown in Fig. 2.

The two tuning condensers were calibrated with a precision

condenser, Type 722»D Serial No. 2539 manufactured by General Radio Company.

The calibration curves are shown in Figures 3 and k»

These calibrations are based on an arbitrary zero at 100 on the

scale, since the instrument provides no way of measuring absolute capacity of

the condensers.

In use^ the cell is filled first with a blank (Varsol) and the

circuit tuned to resonance with the 200 mmfd. condenser leaving the 100 mmfd.

condenser set at a fixed setting at the low end of the scale. The cell;, which

holds a total volume of 50 ml, is then filled with a mixture to be analyzed and

the circuit retuned with the 100 mmfd. condenser. The increase in dial reading

is an indication of the increase of capacity of the cell caused by the change

of dielectric.

Accurately measured volumes of TBP and varsol were mixed to give

a range of 5 to 25 per cent by volume of TBP in Varsol. These were measured

with the instrument and the values plotted on coordinate graph paper. The cali

bration curve is shown in Fig. 5. While this eurve is not strictly linear it is

quite reproducible. Most of the observed curvature is probably caused by

-6-

••*

NOrT: Alt f<fi*t*rs*£ w. Vhfc*. SptziVt.*

t

•

. '

/MS 423I MC

6AG75

I .s 1* i

L-

1"

;

• <-•

r r 3 q •" -;•nary ISOyu-h- IZ8*«**t^ in

A/VW1 -f-

C>*

•i - vV- »1

^T K-4 ,cok -f ,^ Zero I

paf>e.

T—G0?J*-

ctr: i

!

< > 1 \

l; • . .o • •

-v. ^ ^ r

00^f, , •; L. Cf

va (

la^ . 5V. *•<- 1

115 v

a-c . .

5">y)K

1-4 \ — >. • 2. K

SC ua

REVISIONS

*• -

LIMITS OK DIMENSIONS

UNLESS OTHERWISE SPECIFIED

FRACTIONS

DECIMALS

ANGLES

FIA5T USEO ?—n—:OESIGNSD

APPROVED

OAK RIDGE NATIONAL LABORATORYINSTRUMENT DEPARTMENT

n! i ' . r TAMTU\ l LL L . . ;/V /

METF G. 2

L/C Utvb5CALC

I Q- ? 6-/ A

o

0 r —

0

20

30

40

< 50

a

LU

<

60

70

80

^90 O^

O

iooO-

10

JO'

.OrJO'

20 30

HOT CLASSIFIEDDWG 8318

JO'

JOT

JO'

>>'.O

£r

JO'jo-

jo-JO'

JO-

FIG. 3

CALIBRATION OF SMALL LINEAR CONDENSERFOR DIELECTRIC CONSTANT METER Q946

40 50

mmfd60 70 80 90

JO

100

20 I—

Q50 ._.

UJ60 !--

1 1

o 70 r-J

<

S80 —

20 4( 60

lALIBRATION OF LARGE NON LINEAR 00N0FNSERFOR DIELECTRIC CONSTANT METER 0946

140

NOT CLASSIFIEDDWG. 8319

i80 200

0

5Q\

! * I

Lu r- f '

\r- d[J > -< •

«>

X P ^ '- -/—**O ~

cr/">

oWJL..

<£ 30 i—K :

z !

—/ 3 5 !•-—X. i

o l

<UJ 40 L-C) i

A C, '"J J 1<t

ijj

a:

_j 50 L-< |

o i

UJ 55 U-O(T 1

<i

! 60 1—

6 5 h

70 -

75

0

\\

v

X

ex

v

QX

Q\

\V

\V

w.

x

\FIG. 5

CAL IBRATION CURVES OF TRIBUTYLPHOSPHATE-VARSOL MIXTURES

SATURATED WITH WATER

10 15 20

PERCENT TRIBUTYL PHOSPHATE IN VARSOi

\

O

L

dissolved water. As the measuring condenser is not linear below 5 on the

dial, the instrument is balanced at a dial reading of 5 with the large 200

mmfd. condenser with a straight Varsol blank in the cell.

The effect of temperature was found to be very small in the

range of 20° to k0° C as shown in Table II.

Table II

Temperature Effect on Dielectric Constant of 15$ TBP-VarsolMixture.

Temperature 19.5° C 28.5° 1*0.0° C

Beading 36.5 3^ 35.5

. It was found that TBP-Varsol mixtures that contained acid had

to be stripped with three equal volume water washes before measurement. The

organic layer was finally filtered through filter paper in order to remove

the excess water. A comparison of the water washed mixtures with unsaturated

mixtures is shown in Table III.

Table III

Dielectric Constant Readings of Water Saturated TBP-Varsol Mixtures and Unsaturated Mixtures.

Dial Beading *$ TBP by Volume Water Saturated Unsaturated

%10$15*20$

♦Using materials as received from supplier.

It is seen from this table that TBP mixtures that have

been in contact with water must be estimated from a calibration curve made

with water saturated mixtures. Therefore all samples received were washed

three times with an equal volume of water before filtering into the cell.

-11-

35.0 13 0825.2 23«536.5 33-5*8.5 *6o761.5 55.5

The different mixtures were analyzed for water with Karl

Fischer reagent. The results are shown in Table IV.

Table IV

Water Content of the TBP-Gulf BT Mixtures

Sample Per cent H2O by Weight

Straight Gulf BT <0.0l$Straight Gulf BT-15$ TBP at 25° C 0.05$Water Saturated BT at 25° C 0.01$Water Saturated Gulf BT-15$ TBP at 25° C 0.20$

These results show that the TBP-Gulf BT mixtures contain more water than

the Gulf BT reference, however the difference is relatively small. It is

also important to use the correct diluent as the reference to adjust the

100 mmfd. condenser to 5*0 before measuring the unknown samples. A comparison

of Esso Varsol No. 2 and Gulf BT is shown in Table V.

Table V

Dielectric Constant Headings of Mixtures of TBP-Isso TTarsol No. 2 VersusEsso Varsol No. 2 Beference and Versus a Gulf BT Beference.

Readings

%TBP in Esso Varsol No. 2 Vs. Esso Varsol No.2 Vs. Gulf BT

5 1U.1 18.510 2-!4.5 28.515 3^.3 38.520 k6,Q 1+9.925 57-5 61.0

From these data it appears that the same diluent used for

the mixtures must be used for the reference and the calibration curve.

A number of unknown samples estimated by this method have

agreed within -0.2$ with the phosphate procedure. A comparison of the

-12-

dielectric method with the uranium saturation method'5) employed by the

Technical Division is shown in Table VI.

Table VI

Per Cent of TBP Found in Unknown Mixtures

of TBP-Varsol

By Dielectric Method By Uranium Saturation Method

9.9$ 10.2$13.1$ 13.0$15.*$ 15.1$18.1^ 17.8$

The uranium saturation method of course is another indirect

method and depends upon very accurate uranium analysis. Several known samples

analyzed by both methods showed the dielectric method correct within -0.2$

while the uranium saturation method was correct within -0.5$°

The effect of contaminants such as mono- and di-butyl phosphates

was not investigated since these materials were absent in the TBP reagent.

These compounds can be readily detected because of their acidic nature.

Summary; The dielectric constant method has been found to fulfill

the requirements of a good analytical control method for the

estimation of tributyl phosphate in TBP-Varsol mixtures in the range of 5$ -

25$ TBP by volume.

Referencess (l) Ellison, C V., Ferguson, D. I. and Bunion, T» C,Solvent Extraction of Uranium from Metal Wastes,OBNL-258.

(2) Snell, F. D. and Biffen, Frank M., Commercial Methodsof Analysis, McGraw Hill Book Co., Inc., New York, 19^*p. 1*82.

(3) Fisk, P. B., et al CN-1791»

(!»•) Elek, A., and Hill, D. W., J. Am. Chem. Soc. 5J5, 3^79»

(5) Memo to F. L. Steahly fromT. C. Bunion, 0RNL-Central Files No. U9-10-218.

-13-