April, 1958
431
DENSITIESOF HEAVYWATERLIQUID
monosalt and disalt at wave lengths a, b and c, Kz is the disalt dissociation constant (PH+2e P H + H f) and K1 the corresponding monosalt constant (PH+ P H+). I n practice this is prohibitive and one is forced to approximate values of el by use of fortuitous wave lengths and at which the ratios of microspecies may be approximated, so that the K’s and H+’s follow directly. This approximation procedure is, of course, not required for the chlorophyll chlorins, where, pending potentiometric titration, one assumes that K1 = Kz and this may be calculated directly from the data. The results of both the calculations and approximations are presented in Table I. The strength of perchloric acid in nitrobenzene is apparent from the narrow limits of the reference compounds whose aqueous dissociation constants are approximately
+
+
3.2 X and 7.9 X 10-lo. It is interesting to note that the K’s of the chlorophyll chlorins follow their “acid numbers” roughly, the inversion of pheophorbide a and purpurin 18 presumably arising from a solubility effect in EtzO. The very small values of K z imply, by comparison with p-naphtholbenzein, much weaker basicity for the second N than has generally been supposed. From these values, and by the assumption of a Hammett acidity function, one may calculate apparent, approximate aqueous over-all dissociation constants. This hardly appears warranted at the present stage. Acknowledgment.-The author wishes to express his appreciation for the excellent tracings made in the Medical Arts and Scientific Exhibits Section of the National Institutes of Health, Bethesda, Maryland. -
THE DENSITIES OF HEAVY WATER LIQUID AND SATURATED VAPOR AT ELEVATED TEMPERATURES1 BY G. M. HEBERT, H. F. MCDUFFIEAND C. H. SECOY Contribution from the Chemistry Division of the Oak Ridge National Laboratory, Oak Ridge, Tennessee Received November 8, 1967
I n connection with studies of the homogeneous catalysis of the deuterium-oxygen reaction in heavy water systems, it became desirable to know the densities for the liquid and saturated vapor of heavy water at temperatures above 250”. Previous studies had provided information concerning the density of the liquid up to 250” and information concerning the critical properties.a No data relating to the vapor densities appeared to be available. An experimental method used for determining the volume of uranyl sulfate solutions as a function of concentration, temperature and fractional filling of sealed quartz tubes4 appeared suitable for application to the determination of the fractional filling, at various temperatures, of tubes containing heavy water. Such determinations were made for heavy water and light water samples. From the relationship between fractional filling and temperature it was possible to calculate the desired densities of the liquid and saturated vapor. Experimental Method.-Mass balance considerations lead to a relationship between fractional filling at room temperature, Fo,and that, F T , a t some elevated temperature: If V = total volume of tube Po,FT (fractional filling) = vol. of liquid i V mo, mT = total mass of water in the tube d = density v = vapor 1 = liquid (1) This paper is based upon work performed a t Oak Ridge National Laboratory, which is operated by Union Carbide Nuclear Company for the Atomic Energy Commission. (2) J. R. Heiks, e t al., THISJOURNAL,58, 488 (1954). (3) “The Reactor Handbook (USAEC),” Vol. 2, (Engineering), 1955, p. 22. (4) G. M. Hebert, D. W. Sherwood and C. H. Secoy, H R P Quar. Prog. Rep. Oct. 31, 1954, ORNL 1813, p. 164.
then
+
V F O ~ I , V ( 1 - Fo)dvo (1) VFTdlT f V(I - FT)dv, (2) Since no mass is added mo = mT; hence, equating the righthand portions of (1) and (2) and cancelling V Fodlo (I - Fo)dv, = FTdlT f (1 - FT)dv, (3) Equation 3 reveals a linear relationship between FOand F T . This equation could be used for work of the highest accuracy. When, however, as in the present case, d,, = 0, equation 3 can be reduced to a simpler form VLO =
mT =
+
(4)
When the relationship between POand F T is established experimentally, the slope and intercept may be used to calculate dl, and d,, if dl, is known (as it is). Procedure.-Samples of heavy water of known weight and density were sealed in quartz tubes (approximately 14 cm. long, 1.6 mm. i.d. and 4 mm. 0.d.) which had been selected for uniform bore. The total volume of the sealed portion of the tube was determined in the following manner which automatically corrects for the distortion introduced by the two seals. 1. With reference to Fig. 1 measure the height of the liquid hl with a cathetometer and the total height (tip to tip of the inner space) of the inner volume of the tube, hz. The height of the liquid is measured to a position one third the distance from the bottom to the top of the meniscus. 2. Invert the tube and measure the height of the liquid again, as ha, and remeasure hz to obtain an average. 3. Calculate h, = h2 - hl - h) as the length of undistorted tubing in the center of the tube which is never (or always) wet by the liquid in the vertical position, depending on whether the tube is more or less than half full. 4. The total volume of the tube is, then, twice the liquid volume plus the volume ham- (this volume will be positive or negative depending on whether h2 - hl - ha is positive or negative) ( r = half the measured tube diameter). 5. The liquid volume is known from its weight and density. Six tubes were prepared as described above and placed in a circulating air-bath furnace, shown with its instrumentation in Fig. 2. The temperature was measured with a calibrated
G. M. HEBERT, H. F. MCDUFFIE AND C. H SECOY
432
I
I
I
Vol. 62 I
I
I
I
I
I
(Measured with
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
1.0
FT. Fig. 3.-Graphical determination of relative liquid and vapor densities of D20at 353.4'.
Fll Fig. 1.-Measurements C4LIBRATEO C t . l l CONTROL I.C.
volume and fractional filling a t room temperature and the change in volume on heating to elevated temperature, it was possible to calculate the liquid volume and fractional filling at each of the elevated temperatures. Of the six experimental tubes prepared, four contained heavy water and two contained light water. The tubes containing light water were present to serve as an internal check upon the agreement of the experimental data with the known properties of light water liquid and vapor. This check also established the reliability of the measured temperature and made possible the direct comparison of the data for heavy and light water in the same experiment. Data and Calculations.-The experimentally determined fractional fillings at room temperature and a t the various elevated temperatures are presented in Table I.
of liquid heights in quartz tubes.
TABLE I FRACTIONAL FILLING vs TEMPERATURE
rHERMOCOUPLE 7
Temp. ("C.)
H
GL4SS COVERED SLOTS THROUGH W I L L O F S4MPLE UOLOER
._
Fig. 2.-Air-bath furnace and instrumentation: S, sample tube; SB, sample block (6 samples); G, glass tube around sample holder; asbestos packing around top and bottom; Nic, nichronie heater wire (wound vertically); H, handle for rotating sample holder; A, asbestos cord packing; C, glass chimney. chromel-alumel thermocouple, and the temperature control precision was better than 0.1". Readings of the liquid heights in each tube were made a t various temperatures after temperature stabilization. The change in volume accompanying each change in tem erature was calculated from the measured change in the heigft of the liquid and the measured diameter of the tube. It was assumed that the change in height was occurring in a section of the tube which had not been distorted by the sealing process. It was also assumed that the change in volume of the tube, due to thermal expansion of the quartz, was negligible when compared with other experimental uncertainties. From the known liquid
Room temp." 113.2 229.1 268.5 305.6 353.4 367.0 371.5 372.6 373.8 374.8 "24'.
Frl
Ha0
Fre
Fr3
Fr4
DzO
Frl
Frl
0.261 0.348 0.198 0.260 0.346 0.447 .272 .366 .208 .274 .362 .491 .303 .409 .226 .310 .406 ,533 ,314 .430 .230 .314 .429 .567 .320 .455 ,227 ,322 .455 .613 .315 .508 .I68 ,311 .514 .753 .250 .560 .252 .601 .215 .583 .608 .207 .603 .112 .646 .074 .668
Values for d,, and dr can be derived from these data either by graphical or mathematical procedures. A graphical solution for the heavy water data a t 353.4' is presented as Fig. 3. Multiplying the intercepts by 1.1047, the density of heavy water liquid at room temperature (d& gives 0.6109 g./cc. and 0.1414 g./cc. as the values for dlT and d v T ,respectively. A mathematical solution can be obtained for each pair of experimental points by means of simultnneous equations; for four tubes a t one temperature there are six such sets of equations, and the average of all such values was determined a t each temperature. Figure 4 presents the graphical and mathematical values for the liquid and vapor densities so determined for light and heavy water, togethe: with appropriate values from the literature for comparison. Table I1 presents values for D20 taken from the smoothed curve at convenient temperature intervals.
Discussion The expected linear relationship between fractional filling a t room temperature and fractional filling a t elevated temperatures was demonstrated,
VAPOR-SOLID EQUILIBRIA IN THE IRON-CHLORINE SYSTEM
April, 1958
433
TABLE I1
LIQUID AND VAPOR DENSITIES OF DtO Tern C0C.P'
Density (g./ca.)
175 180 190 200 210 220 230 240 250 260 270
280 290 300 310 320 330 340 350 360 370 371.5 0 Critical point.
Vapor
Liquid
0.004 ,005 ,006 .007 .009 .OlO .013 .016 ,020 .024 .029 .034 .040 .048 .058
0.989 .983 .970 .957 .943 .929 .913 .898 .881 .864 .847 ,829 .so9 .787 .763 .735 .705 .668 .626 .573 .462 .363"
.070 .OS7 .lo5 .129 .163 .248 .363
0.9
0.8
0.7
-
1
0.6
U 0
\
.? 0.5
t
t
0.4
LITERATURE VALUES H-0
-
EXPERIMENTAL VALUES FOR HzO 0 GRAPHICAL SOLUTION A MATHEMATICAL SOLUTION EXPERIMENTAL VALUES FOR D20 -$ GRAPHICAL SOLUTION MATHEMATICAL SOLUTION
A --D20
0.3
and the agreement of the experimental data with this relationship was sufficiently close to permit calculation of the densities of the liquid and the saturated vapor. The values so calculated were in TEMPERATURE rG), good agreement with the available data for heavy Fig. 4.-Liquid and saturated vapor density of H20 and water and the established data for light water. The DzO. precision of the method may be estimated from the This method is so simple to apply, and the oppordeviations between accepted values for light water and those found by the present technique. Be- tunities for direct comparison with liquids of known tween 200 and 370" the average deviation in liquid properties in order to establish internal checks of densities was less than 1% of the accepted value the consistency and temperature in a particular and the average deviation in vapor densities was experiment are so attractive, that the method might less than 10% of the accepted value. Above 370" be considered for use with other liquids, such as the effects of small temperature errors are so much pure hydrocarbons, for which it may become demagnified, and below 200" the vapor densities sirable to know the vapor and liquid densities at are so low, that larger deviations occur. elevated temperatures.
VAPOR-SOLID EQUILIBRIA IN THE IRON-CHLORINE SYSTEM BY LAURENCE E. WILSON~ AND N. W. GREGORY Contribution f r o m the Department of Chemistry, University of Washington, Seattle 6 , Washington Received November 19. 3067
The vaporization and thermal decompositionequilibria of iron( 111)chloride have been investigated by gas saturation flow and diaphragm gage techniques. A brief summary and discussion of results of previous investigators is given. Based on the present study, equations for the preysures of Fed% and Clz in equil!br!um with FeC13(s)and/or FeClz(s) as a function of temperature are presented. Solid solution of FeCla and FeCl2 IS not significant below 300".
Vaporization of solid iron(II1) chloride is cornplex. Three equilibria (or alternate combinations thereof) must be considered 2FeCls(s)
FezCls(g)
FezC&(g)I _2FeCls(g) FezCldg) J _ 2FeClds) Cl&)
+
(1) National Scienoe Foundation Fellow, 1854-1957.
(1)
(2) (3)
Although a number of studies involving these equilibria have been made previously, a survey of published results reveals inconsistencies such that one cannot, with confidence, predict the equilibrium partial pressures of the various components as a function of temperature. Four independent measurement,s of the total pressure above solid iron(II1) chloride have been
