of dissociation is given with sufficient accuracy by the first approximation : -r~= 1 -
3.0~
=
1
- 50
+ 2x02,
(4)
and f i 2 is obtained from Equation 2 using the d u e of yo. Substitution of these approyimations into Equation 1 yielded thr values of h given in Table I. The results may be considered accurate to =kO.lyi. The data of Table I may be represented by a plot of log (2i3--,i) against log concentration of dissolved oxygen (Figure 1). The following procedure should be used with the plot: estimate o ~ y g c n concentration roughly; read plot and determine 1; calculate oxygen concentration from the following equation, if concentration is greater than 0.005 p.1i.m. : 0 2 =
(0.0017
+ 8AL/A)/e,
Equivalent Conductance of Thallous Hydroxide Solutions a t 25’ C.
TlOH concn. (moles/liter)
x 5x 2 x 4 x 6X 8 x 1x 2 x 3 x 4x 5 x 1
where
p.pm
Table I.
(3)
At higher concentrations, a second approximation to y is obtained from y1
~
(6)
where AL is the conductivity increase of water passing through the column (pohms per em.), and t is the fraction of dissolved oxygen reacting, normally close to 1.0. (Xore complex calculations are required if the oxygen concentration i. less than 0.005 p.p,m.). Repeat, j f necessary for desired accuracy. This procedure converges so
6 X 7X 8X
0 10-6
10-6 10-5 10-5
10-5 10-4 10-4
10-4 10-4 10-4
9 x 10-4
Oxygen concn.“ (p.p.m.) 0
0.008 0.04
0.16 0.32 0.48 0.64 0.8 1.6 2.4 3.2 4.0 4.8 5.6 6.4 7.2
Equivalent conductance* (mho-cm.?/ TlOH concn. es.) (moles/liter ) 273 0 1 x 10-3 272 88 272 72 1 . 5 x 10-3 2 x 10-3 2 . 5 x 10-3 272 43 272 18 3 x 10-3 271 99 3 . 5 x 10-3 4 x 10-3 271 82 4 . 5 x 10-3 271 67 5 2’10-3 2’71 O i 5 . 5 x 10-3 270 59 270 13
269 269 269 268 268
80 46 1.3
83
51
6
X
7
X
6.5
x 10-3
7.5 X
x 10-3 8.5 X 9 x 10-3 9 . 5 x 10-3 1 x 10-2 8
Equivalent Oxygen conductance* c0ncn.a (mho-cm.z / (p.p.m.) es.1
..
8
12 16 20 24 28 32 36 .. 40 44 48 52 56 60 64 68
72 76 80 a Corresponding to TlOH concentration, assuming complete reaction. b All -4 values f0 . 1 yo.
rapidly that the initial estimate can be very approximate. LITERATURE CITED
( I ) FUOSS, R. M.,J . d i n . Chem. Soc. 81,
2659 (1959). (2) Fuoss, R. AI., Accascina, F., “Electrolytic Conductance,” Interscience, Sew York. 1959, (3) Hlasko. ill.. Salitowna. A , . Rocrniki Cheni. 14, 1038 (1934). (4) Ibzd., 15, 2T3 (1035). ( 5 ) Lindsay, IT.T., Jr., Pittsburgh Con-
268.31 267.09 266,03 265.07 264.19 263.37 262.59 261.86 2 6 i , ‘16 260.49
259.84 259.22 258.62 258.05 257.48 256.93 256.40 255.88 255,37
ference on Analytical Chemistry and Applied Spectroscopy, February, 1961. (6) Lindsay, W. T., Jr., J. Phys. Chem. 66, 1341 (1962). (7) Ostwald, W.,“Lehrbuch der Allegemeinen Chemie,” Engelman, Leipzig, 1891; “International Critical Tables,” 1st ed., Vol. 6, p. 243, McGraw-Hill, Kew York, 1929. (8) Wright, J. M., Lindsay, W. T., Jr., Proc. Am. Power Conf. 21,706 (1959). W.T. LIXDSAY, JR. Physical Chemistry Department Westinghouse Research Laboratories Pittsburgh 35, Pa.
Determination of Water Content in Potassium Difluoride SIR: -4 method has been developed to determine sniall amounts of n ater in KHF2 with Karl Fischer (KF) reagent. It is well known that KHF2is insoluble in common solvents at room temperature, nith the exception of a fen- organic acids, in iThich it is sparingly soluble. Hon el-er, esterification reactions of such acid solvents with KF reagent cause serious interferences, especially with iamples of low water content. The experinierital technique reported here makes it possible to separate the water due t o esterification of acetic acid solutions nith the methanol from the K F reagent. A fixed volume of a solution of sample in acetic acid containing less than 0.1% of water is added in the analysis cell to a constant volume of pyridine that has been dehydrated beforehand with a K F qolution. The titration is checked by a voltammetric method. A constant current of
about 2.5 pa. is imposed across t n o bright platinum microelectrodes, and the voltage at the tn-o ends is measured and recorded. It varies between a few millivolts in the presence of iodine t o about 350 mv. in its absence. Since the kinetics of the Karl Fiqcher reaction are approximately first order nith respect t o the water concentration, the average rate of the reaction is approximately inversely proportional t o the tranqition time, 7 , necessary to crow a prefixed voltage interral. As a convenient reference interval, the steepest part of the voltage m. iodine concentration curve (50 t o 250 mv.) is choqen. The titration is effected by immediately adding the K F reactant in small and constant quantitieq each time the voltage exceeds the upper limit (260 mv.). So, during the intervals between each addition, the iodine concentration is the .&me; the KF reaction rate depends only on the nater concentration. The asymptotic value of 7 corresponds
to the concentration of water in equilibrium with the esterification n-aterproducing reaction and the KF waterdestroying reaction. When all the original 15 ater contained in the sample has been titrated (theoretically a t time a), the latter becomes constant. The values so obtained are plotted aq a func. tion of the volume of the K-F added solution (Figure 1). A pure solvent blank, of the same volume RS the previously titrated acetic acid solution, is run folloning the same procedure. The inflection points of the 7 us. volume of KF solution curl-es are taken as reference points for both tests. An objection that can be made to this procedure is whether the time required by the titration is the same for the sample as for the blank. This condition, in the case of all-manual titration, is dependent upon operator skill. The condition is always attained using an automatic device driven by the voltammetric circuit that adds the titrant VOL 35, NO. 3, MARCH 1963
409
The water content is given by:
% Hz0 50
1,
101
0 7.15
[(G
+ G o b - d,/da.
Gan'lT
1 220
' 225
230
735
740
n (cc.) Figure 1 . Transition time, 7,as a function of volume of added KF solution
because the bigger part of the reactant is added a t the very beginning of the titration. I n other words, in the titration of two different water-containing samples, the time for reaching a point which differs from the inflection point by a definite small amount is practically the same. Let n and n' be the titrating volumes a t the inflection points corresponding to the titration of the sample and of the blank, respectively; G and G, the weight, in grams, of the salt and of the acetic acid in which it is dissolved; T,the titer of the Karl Fischer solution (expressed in mg. of HzO per cc.) ; V the volume of acetic acid and of the acetic acid solution introduced in the cell; and d, and d, the respective densities.
Table 1.
Analysis of Four KHF Samples
Water found,
10Vd.G
+
LO.
20
=
since d, = d, 0.0074b, where b is the percentage of salt in the acetic acid solution and d, = 1.051 a t 20" C. Some of the results obtained are reported in Table I. The results are independent of salt concentration, which means that the extraction of the water contained in the sample is complete. The volumes of acetic acid solution and of the blank were of the order of magnitude of 5 to 10 cc.; T was 1.5 to 2 mg. of HzO per cc. The same KHF2 samples mere analyzed by electrolytically generated iodine after dissolving in a salicylic acid solution in the Karl Fischer reagent. I n this case, account was taken of the rate of esterification by measuring the current necessary to maintain a constant voltage across the platinum voltammetric-control microelectrodes. The results are in agreement with those obtained with the acetic acid method, although the coulometric method does not permit a routine analysis, since the time required for dissolution is too long (1). The separation between the anodic and cathodic compartments of the cell is accomplished with an anionic exchange membrane (BDH-Permaplex A-20). Finally, we measured the limiting current relative to the evolution of hydrogen on a bright platinum cathode
Sample 1
2
3 4
Run 1 2 3 4 5 6 7 8 9 10 11 12 13 14
b 2.1 2.4 2.8 1.8 1 .6 ~. ~
%
0.657 0.645 0.648 0.649 0.788
1.3 1.3 1 .o 1.2
0.793
1.6 0.9
0.310 0.330 0.310 0.309
1.2 2.8
0.040 0.036
1.6
0.776
0.030
immersed in a melted bath of KHF2 a t 250" C. The values obtained from galvanostatic measurements are approximately proportional to the water content as determined by one of the two abovedescribed methods (2). LITERATURE CITED
(1) Barbi, G., Pizzini, S., High Tempera-
ture Chemistry Section Laboratories, CCR Euratom, Ispra, Italy, unublished data, 1961. (2P Pizzini, S., Barbi, G., Sternheim, G., High Temperature Chemistry Section Laboratories, CCR Euratom, Ispra, Italy, unpublished data, 1961. GIOVANNI B. BARBI SERGIOPIZZINI Chemistry Department Euratom C.C.R. Ispra, Italy
Vacuum Output Gas Chromatography SIR: The operation of the output of a capillary chromatographic column in vacuum has not yet been evaluated from the viewpoint of its analytical value. Such technique was not considered advantageous in either its experimental or theoretical aspects ( 3 ) . For example, the linear gas velocity in a column operating a t 40-p.s.i.g. input and 1p.s.i.g. output pressures increases only in a ratio of 1 to 3; while in the case of a column having the same input but vacuum (IO+ torr) output pressures, it will increase in a ratio of 1 to 10'. Recently Giddings (1) showed in his theoretical paper that in case of capillary columns using vacuum output and a carefully selected input pressure, a high speed separating system could be obtained. While-operating the output of a standard Golay column in vacuum (10-3 torr) the input of the column can be operated a t or above atmospheric 410
ANALYTICAL CHEMISTRY
pressures. Such pressure differences can be held off by the capillary column itself without using additional restrictions and valves. I n a previous paper (3) describing a vacuum type quantitative and qualitative (QQ) detector, we suggested the possibility of coupling the outlet of a capillary column directly into the input of the detector tube and operating the output of the capillary, accordingly, a t low pressures. The feasibility of such a system has now been investigated, and experiments have been made to evaluate the performance of such an arrangement. The experimental system is shown schematically in Figure 1. The carrier gas was helium. The sampling system consisted of a standard Perkin-Elmer type gas sampling valve ( S V ) ,a rubber septum injector port (SZ), and a split (8).I n some of the experiments the sampling system was connected directly to the input of the column, not utilizing the split. The capillary column (C) was a 150-foot, 0.010-inch i.d. stainless
steel tube coated on the inside wall with Dow Corning silicone high vacuum grease. A conventional high vacuum system pumped the QQ detector tube, and through this the column, continuously. Both the QQ detector tube and the vacuum system were described in a previous paper ( 3 ) . The column output pressure was monitored by the QQ detector tube by switching the ionization potential from 22 volts to 150 volts, ionizing the helium gas and measuring its pressure. The input pressure was measured by a manometer (MI. The performance of this vacuum output chromatograph system was compared to a conventional system operating a t atmospheric output pressures and using a flame ionization detector. The measurements were carried out using identical columns, identical input pressures, sample amounts, and operating temperatures. The gas mixture to be tested contained CH4, C2H6, and
