4670
JAMES
C. !VI. LI AND
of magnesium ortho- and dititanates from their constituent oxides are - 0.4 and - 0.1" cal. /deg. mole, as compared with - 0.7 for magnesium metatitanate. Thus, the entropies of the compounds in this system do not deviate markedly from the
N.MJ. GREGORY
Vol. '74
sums of the entropies of the oxides. Similar behavior for magnesium meta- and orthosilicates was noted by Kelley.6 ( 6 ) K K Kelley, THISJ O U K U A L , 6 6 , 339 (1943). UERKE[,EY
4, CALIF()R\rA
/ ~ O S ~ I . K I B C I . I OFROM N THE DEPARTMEU f 01' CHEMISTRY AND CHEMICAL E S G I N E E R I S G , USIVERSI'I'Y Of( 'i\7.4SIIlS(; i'OS
Heats of Solution and Formation of Some Iron Halides 1 3 JAMES ~ C. 11. LI
AKD
X. iV. GREGURV
R r r c e ~ vFEBRUARY ~~ 18, 1952 Heats of solution have been determined for E'eCL, IJeCla, FeBrr, FeBrr and FeBrClp a t concentrations of the order 10 molal. Heats of formation calculated from these data agree with quantities measured by independent methods when corrections are applied for hydrolysis and complex ion formation in the iron( 111)solutions.
The complex nature of the ionic species in aqueous iron halide solutions makes interpretation of heats of solution of these compounds diflicult. A question arises concerning the reported value for the heat of formation of iron(II1) chloride based on determination of its heat of so1ution.l.z A correction for complex ion formation and hydrolysis does not appear t o have been included in the interpretation of data and the value obtained is not in good agreement with that determined from equilibrium studies in anhydrous s y ~ t e m s . ~We have reinvestigated the heats of solution of FeClz and FeCl3 a t low concentrations and have obtained similar results for FeBr2, FeBr3 and FeBrClz. Taken in conjunction with the reported characteristics of iron(II1) solutions,2 these data lead t o heats of formation which are in good agreement with quantities observed in equilibrium studies. Experimental Part Thermal measurements were made in a simple adiabatic Calorimeter, consisting of a dewar flask equipped with stirrer andaample-holding and breaking device. 750 ml. of water, carefully saturated with and maintained under a nitrogen atmosphere to minimize oxidation of iron(II), served as the solvent. The calorimeter thermometer was constructed from 6-mm. Pyrex glass tubing in the form of a spiral of 17 turns, 2 inches in diameter and 5.5 inches long, filled with toluene and a small amount of mercury. The lower end was connected t o a capillary of 0.4 mm. diameter which turned upward and served as the thermometer stem. The sensitivity, as measured with a cathetometer read t o 0.1 mm., corresponded t o 0.0004 degree, or 0.36 calorie. The entire device was placed in an air-bath controlled to f0.01' a t 25" with only the stirring rod, breaking rod and thermometer stem extending out of the thermostat. 111 tests for temperature drift the thermometer reading was observed t o stay withiti the error of the cathetometer for periods of more than two hours during blank runs under normal operating conditions. Samples were carefully prepared, purified by vacuum sublimation where possible, analyzed, and transferred t o small Pyrex capsules in a dry-box. The capsules were evacuated and sealed and the quantity of material present determined by subsequent analysis of the solution. The calorimeter was calibrated by measuring heats of solution of LiCl, NaCl, KC1, KBr and KI.* For each compound various amounts were used to give approximately 36, 72, 108, 144 and 180 calories. Deviation of results for the various substances was within 50.3 calorie. ( I 1 F. K. Hichowsky aiid Ti. D. Rossini, "Thermochemistry of the Chemical Suhstances." Reinhold Publ. Cur* , New Yurk, X. Y., l93F. ( 2 ) F. D. Rossini, c1 ril , S a t l . Bur. S t ~ l s .Circ. 500, l i . S. G o b t. I'rioting Otficr, Washington. L). C.. l!lT,O, ( 3 ) W. Kdiigronnd E, Peterseii, Z o i i u r g . Cizc,tt , 2 6 1 , 1 : i (1!150),
The heat conductivity of toluene is much lower than that of mercury which offsets some of the advantage gained by its greater coefficient of thermal expansion. However, our experience confirms an approximate calculation of the time required (somewhat less than two minutes with efficient stirring) t o reduce the temperature differential between the center of the liquid column and the outer edge of the glass tube to 0.001 of the original value produced on dissolving the sample. Stirring appears t o be the major factor in the problem of heat transfer; in the absence of stirring in the calorimeter, more than 30 minutes was required for the system to reach uniform temperature. Thermal equilibrium in times of the order of two minutes was considered satisfactory for the purpose of the present work. it should also be noted that the compressibility of toluene is sufficiently large to require a correction in the temperature reading if barometric pressure varies appreciably. The error becomes significant for variations of the order of 2 cm. The heat of stirring a t the rate employed was of the order of 0.2 calorie per minute. Allowance for this was included in the calibration by adjusting the temperature of the airbath so that the heat leaking out of the calorimeter matched the input from stirring ( L e . , the calorimeter remained a t constant temperature with a constant rate of stirring). This steady state was maintained for a period of 15 minutes or more before breaking the sample holder. After solution of the sample the temperature would again stay constant a t its maximum value and was observed for a period of 10 minutes or more. Hence it was assumed that the heat of stirring was exactly compensated by heat leakage during the run and no correction applied for either. The change of heat of stirring while the sample was dissolving was neglected inasmuch as these compounds dissolved almost immediately in water. As an illustration of the temperature behavior during a run, the following examples are observations made with FeBr:$samples. i'athe-
'l'imc. miniites
0 to 19 Capsule brokeri
20 21 t o 28
l(1meter reading
'lime, minutes
(1 to 8 Capsule broken 95.855 9 95.865 10 t o 20
94.GC;3
Cathrlnmetrr reading
9-1. 7 2 1
9.5. 92.5
95.940
Results and Discussion Iron(I1) Chloride.--Samples were prepared by thermal decomposition of the hydrate with subsequent sublimation of anhydrous FeC12 in high vacuum a t 550". After the thermal measurements, solutions were tested for Fe+++ ion with negative results in all cases. Data are shown in Table I. An average value for the heat of solution I:e.CI:(s\ = I:e ++(aq)
+ '7C1-(aq)
has been taken as - 19.5 f 0.2 kcal./iiiole, tieglectirig the sriiall contribiition accompanying fur-
HEATSOF SOLUTION AND FORMATION OF IRON HALIDES
Sept. 20, 1952
ther dilution of the solution. Assuming that the aqueous solution of FeClz involves only Fe++ (AHo = -21.09 and C1- (AHo = -40.02), the heat of formation of FeClz(s) is found to be -81.5 =t 0.2 kcal./mole a t 25". This value is in good agreement with other results, including those based on independent m e t h ~ d s . ~ ~ ~ * * J ' Iron(I1) Bromide.-The heat of formation of FeBrz reported by Hieber and Woerner4 is not based on a direct determination of the heat of solution and the latter has not been reported. Three independently prepared samples were studied, including vacuum sublimed material and FeBrz obtained by thermal decomposition of FeBr3 at 100". Data are shown in Table I . A slight trend toward larger negative values observed with decreasing concentration ; howeve;, criechange is of the order of experimental error and in as much t similar behavior was not observed with FeCL uI.average value for the heat of solution of -20.1 f 0.4 kcal./mole has been taken rather than extrapolating the data. Using AHo = -28.932 for Br-, a heat of formation for FeBrz(s) of -58.7 i 0.4 kcal./mole is obtained. Whereas this is somewhat less than -60.02 given by the N.B.S. tablesj2 our value is consistent with results for FeBrs when compared with the difference in heats determined from thermal dissociation equilibrium data.6 TABLE I HEATSOF SOLUTION FOR IRON(II)HALIDES Molality
FeCIz Heat of solution, kcal. /mole
0.0117 .0029 ,0022 ,00167 .00139
-19.5 -19 5 -19.8 -19.5 -19.5
Molality
FeBrr Heat of solution, kcal./mole
0,00828 .00652 .00386 ,00316 ,00247
-19.7 -20.2 -20.3 -20.3 -20.4
Iron(II1) Chloride.-The heats of solution which have been reported for FeC13 were determined in rather concentrated s ~ l u t i o n . I~n~ ~the present work measurements have been made a t lower concentrations using samples prepared by resublimation of Fisher C.P. anhydrous FeCL. Solutions were carefully tested for Fe++ with negative results. The heat of formation of FeC13(s)has been estimated from these data by considering the reactions F e + + + + H 2 0 = FeOH++ + H+ (1) m(l
m(1
- a - p)
Fe+++
- a - p)
ma!
+ C1m(3 - 8)
ma
= FeCl++
(2)
mS
where a is the fraction of Fe+++ hydrolyzed and /3 that fraction complexing with chloride ion. We have used values for AFo and AH" characterizing these equilibria as given by the N.B.S. (4) H. Hieber a n d A. Woe,rner, 2. Elcktyochem., 40,287 (1934). (5) Kokiti Sano, J . Chcm. SOC.Japan, 69, 1069 (1938). (6) N. W. Gregory and B. A. Thackrey, THIS JOURNAL, 78, 3176 ( 19.50). (7) G.Lemoine, Ann. Chim. Phys., SO, 373 (1893). ( 8 ) W.Kangro and R. Flugge, Z. physik. Chcm., 8176, 187 (1935). (9) For a discussion of the hydrolysis problem see A. B. Lamb and A. G. Jacques, THISJOURNAL. 60, 1215 (1938), and others. O n chloride complexes, C. Brosset, Svcnk. Kcm. Tid., 68, 434 (1941); J. Badozlambling, Bull. SOC. Chim., 652 (19051,and others.
4671
Corresponding quantities for the second stages of these reactions indicate that they may be neglected for the purposes of this treatment. The equilibrium constants for (1) and (2) were expressed as functions of ionic strength of the solutions by the approximation given by the DebyeHuckel limiting law. log K I = -2.43 log Kj = 1.49
wherep
- 2p1/2; AHal = 12.32 kcaI. - 3p1/2; AHa2= 8.52 kcal. = m(6 - 201 - 3p).
With these three equations and the expressions for the equilibrium constants in terms of m, a and /3, one may solve for a! and p for various values of the concentration m. The heat of reaction (defined as A) FeClp(s) = Fe+++(aq)
+ 3Cl-(aq)
may be calculated from the experimental heat of solution and AH0* and AHoz. Data are summarized in Table 11. The constancy of X over a rather wide range of values of a and /3 tends to support the validity of the assumptions made in this treatment. An average value of X is -38.0 f 0.2 kcal./mole a t 25") which leads to -93.4 kcal./mole for the heat of formation of FeCt(s> (using AHo for Fe+++ as -11.42). This is 3.4 kcal. smaller than the value previously reported from heats of solution but is in good agreement with -93.5 obtained by Kangro and Petersen3 from equilibrium studies in the iron-chlorine system. TABLE I1 HEATSOF SOLUTION FOR IRON(III)HALIDES Molality
K
0.00318 .00311 .00221 .00150 .000766
0.53 .535 .615 .69
.81
B FeCla 0.05 .049 .035 .024 .012
0.00736 .00307 .00271 .00220 .00206 .00191 .00098
0.36 ,555 .58 .63 ,645 .66 .78
FeBrr 0.0127 ,005 .005 .0038 .0035 .003 .0017
Exptl. heat of solution, kcal./mole
-31.1
x,
kcal./mole
-30.8 -30.1 -29.7 -27.8
-38.1 -37.8 -38.0 -38.4 -37.9
-29.0 -29.2 -28.6 -26.4 -28.3 -27.6 -24.3
-33.5 -36.1 -35.8 -34.2 -36.3 -35 7 -33.9
The heats of solution measured by us are consistent with earlier data7V8; however, a quantitative comparison involves rather uncertain approximations concerning ionic strength and composition in extrapolating t o the higher concentrations. Iron(1J.I) Bromide.-Heats of solution of FeBra have not been reported previously. Preparation and analysis of samples have been described.6 Techniques employed were the same as with FeCL, including the corrections for hydrolysis and complex formation. The equations log K1 = -2.43 log K: = 0.59
- 2p'/s; AH', - 3p1/3; AH''
= 12.32 kcal. = 6.10 kcal.
( K ; referring t o the equilibrium involving Fe+++,
13r and FeRr4+ ? ) were used with the results shown Table 11. Four independent samples lead to an average value for X of -33.3 f 1.0 kcal./mole. This gives -62.8 f 1.0 kcal. 'mole for the heat of formation of FeRr?(s) at 22". The difference between 11i( heat\ oi fnrrnatioii of F'cl3r)(qi ,tiid FcRrl(sj ol,t,iined i i i this work 15 i n qoot'l :igrcciiicnt with that found froin equilibriuin stutlici," I L = I 2 kcal. and 4.0 =t0.;3 kcdl , respectively The somewhat greater scatter of data for FeBr, Liscompared with those for the chloride is thought to be due largely to the greater difficulty of purifyiilg the bromide. Solutions gave negative tests for I'e- -, however. Iron (111) Monobromodich1oride.-This substance was prepared by bromination of FeClZ.'O Its thermal instability makes purification difficult and varying amounts of FeT+ were found in the solutions following thermal ineasurements. The Fe+* ioii was assumed to represent FeCI2 in the original $ample aiid the heat of solution corrected for the amount present as determined by analysis. Interpretation of results was carried out as for the other iron(l1I) compounds. The fraction of iron in the inrm of FeBr + predicted in these solutions is \cry small; hence only FeOH+* and FeCl'f were considered. Results are shown in Table 111, where nz' represents the molality of FeC12 in the solution and X is again defined as A H o for Rr-iaq) the reaction FeBrCl?(s)= Fe"+"(aq) 2Cl-iaq) . i n 'iverage value for X of -;;; 1 t 0 .i III
+
10
/
\
'il' C,rrgoi\
IKI\
ILT it\
\I
73
+
1051
kcal. mole is obtained; this leads to -82.!1 f '11101~~ for the heat of formation of FeBrCl.! which agrees well with -83 kcal. obtained from equilibrium studies of the thermal dissociation of FeBrC12into FeCILand bromine. l o Values for heats of formation of substances 11iv01v111g bromine :ire hascrl O I I thc liquid Iir? ds the standard state. 0 3 kcal.
'rABI,E 111 HEATS01- SOLUTION FOR FeBrCll
\Iola!ity n1
a
@
0 00585
0 00218
noioc
0 38 M
0 10
o04.a 00197 00123
Ifolality
00037 00013
tj3
(1; 03
73
02
Exptl. heat of solution, kcal. /mole
A,
kcal. / m d e
'3
-31 8
-37
-.?o
9
-:G 2
-29 5 -28 6
-37 5 -37 8
Conclusions.-Whereas several assumptions are involved in the interpretation of measured heats of solution of the iron halides in order to estimate heats of formation, the agreement of the values obtained with those from independent methods offers some support to the validity of the treatment. -1linear relation between the heats of formation of FeC13, FeBrCl? and FeBra and the number of C1 or R r atoms per atom of iron is observed which is not apparent if one uses the higher value for the heat of formation of FeC13 previously reported. This relationship also applies to the heats of solution (A) and might be anticipated from the close similarity in structure of the mixed halide to FeC1 and FeRri as discussed in an earlier paper l o AI 11 I i \ \ ' % Y H I \ G I O \
1
, ~ O S T R l I ~ c l . l i l . \ - l ' IFROM O ~ ~ I I I E IIEPARTMEST OF CHE4fISTRl ASI) CHEMICAL ESC,INEERISG, 3TATE I'SIVERSITY OF
Spectrophotometric Study of the Phosphorus-Interhalogen Tetrachloride and Acetonitrile1
IO\!'.\
1
Complexes in Carbon
BE' .II.EXANDF,R I. POPOV , ~ X DEDW.IRD H . SCHMORR RECEIVED FEBRTARY 14, I R,i2 Spectrophotometric study of phosphorus hexachloroiodide and phosphorus hexabromoiodide reveal that in carbon tetrachloride solutions the complexes dissociate into component molecules PCIEI % PC1, IC1; PBrsI & PBra IBr Br?. ! IC1.- and PBrJ Fr! PBrlf IBr?-. PhosI n polar solvents, like acetonitrile, the dissociation is ionic: PCII;I ~ rPClif phorus pentahromide---carbontetrachloride cotriplex PBra.2CC14is prrpared by B neir method, and its decomposition point is (ietermiried.
+
Introduction Although a large number of phosphorus-halogen complexes have been reported in the literature, it seems that relatively little work has been done on the elucidation of their structure and their phvsical and chemical properties. In recent years Fialkov and Kuz'menko2 have studied the addition of iodine, iodine chloride and iodine bromide to phosphorus tri- and pentahalides. I t was found that stable solid complexes, phosphorus hexachloroiodide (PCleI) and phosphorus hexa(1) Abstracted from a thesis presented by Edward H. Schmorr to t h e Graduate College of t h e S t a t e University of Iowa in partial fulfillment u f the requirements for the degree of I f a s t e r of Scienre. Felxuary, 1!452, ( 2 ) (a) Ya. A . Fialkov and A. A. Kuz'menko, J . Gcn. Chcm. (L'..S. S . R . ) , 19, 812 (1949); (b) 19, 997 (1949); (c) 21, 433 (1961); (d) A -4 Kuz'menko xnd Y a . .4 Fialkov. ihid., 21, 47.7 (3953).
+
+
+
+
bromoiodide (PBrJ) can be obtained by the simple addition of iodine chloride and iodine bromide to the respective phosphorus pentahalides. Likewise Fialkov and Kuz'menko have verified the reaction reported by Baudrimonts in which phosphorus hexachloroiodide is formed by the addition of iodine to phosphorus pentachloride. A study was made on the electrical conductivity of these complexes in various organic solvents.?d I t was found that while acetonitrile, nitrobenzene, ethyl bromide and chloroform yield conducting solutions-benzene, toluene, carbon tetrachloride and dioxane solutions do not. These results were based on a study of saturated solutions of these complexes in the respective solvents, and i t must be noted that in the case of non-conducting solutions ( 3 ) \I
F Raudrimont, 4 i i i i i h i t n f i h y s , 141 2 , 8 fl864)
