THERMODYNAMICS OF TITANIUM CHLORIDES
Mar., 1956
309
THERMODYNAMICS OF THE TITANIUM CHLORIDES. I. HEAT OF FORMATION OF TITANIUM TRICHLORIDE’ BY DAVIDG. CLIFTON~ AND GEORGE E. MACWOOD Contribuiion from the McPherson Chemical Laboratory, The Ohio State University, Columbus 10, Ohio Received August 16, 1866
The heat, of formation of titanium trichloride has been determined by measuring the heats of solution of TiCI,(I) and Tic&(s) in a solvent of HCI-FeCI3 in an ice calorimeter. The heat of formation of the trichloride is based upon the heat of formation for TiClr(l) of -192.1 f 0.6 kcal./mole. The value obtained for the heat of formation of TiC13(s) at 298°K. is -172.2 f 0.7 kcal./mole.
Introduction There are few references in the literature for the heats of formation of the lower titanium chlorides. Brewer3 gives estimated values for the trichloride and the dichloride. Also, Kubaschewski4 reported estimated values for the tri- and dichlorides. In the light of recent work by Schaffer, Bred and Pfeffer,b and by Skinner and Ruehrwein,s these previously reported estimates are low. This research provides an independent determination of the difference between the heats of formation of TiCL(1) and TiCls(s). With these results and the results on the heat of formation of TiClz(s),’ an attempt is made to give consistency to the heats of formation of the titanium chlorides. Apparatus The heats of solution measurements were made using an ice calorimeter. The calorimeter, shown in Fig. 1 , wag a modification of the calorimeters developed by the National Bureau of Standard~.e*~ The main alteration was in the method of determining the volume change in the calorimeter. A dilatometer was used instead of the weight method employed by the Bureau. The calorimeter was calibrated electrically for the purpose of comparison with the Bureau’s calibration fact,or. The factor obtained was 878.08 f 0.09% cal./cm.S This agrees within its recision with the value found by the NILtional Bureau of Ktandards.
modified by Reed. The trichloride was formed by hydrogen reduotion of titanium tetrachloride on a hot tungsten flament. After removal of any tetrachloride from the trichloride by pumping, the trichloride was further purified by sublimation: The trichloride was analyzed for total titanium using the method of Rahm.11 Analysis gave an empirical formula of TiCls.aw. Analysis of the reducing power was also made. The trichloride was dissolved in a solution of FeCh and. HISO, and the ferrous ion produced was determined by tltration with ceric sulfate solution. The ceric sulfate was standardized against AS&. An inert gas atmoAphere was found necessary for precision. ~
To i n e r t gas --TO
adotometer
cower vacuum II
Supportma slrqn
Materials Titanium Tetrachloride.-The titanium tetrachloride used in the heat of solution determinations was provided by the Inorganic Chemical Section of the National Bureau of Standards. Their analysis gave the material a purity of 99.9991’%. Samples were handled in sealed Pyrex bulbs and exposed only to dry nitrogen. The actual purity of the experimental samples was therefore not as high as the original, but any contamination was not apprecbiahle. Titanium Trichloride.-The titanium trichloride waa prepared in this Laboratory by the method of Sherfey’o as ( 1 ) Work performed under the Office of Naval Research, Contract No. Nonr-405(00). (2) Taken in part from the dissertation aiibmitted b y David G.
Clifton in partial fulfillinent of the requirements for the Ph.D. degree a t The Ohio State University, March, 1055. (3) L. Brewer, Paper 6, “National Nuclear Energy Series,’’ Vol. IV-IgB, edited by L. L. Quill, McGraw-Hill Book Co., New York. N. Y., 1950. (4) 0. Kubsscliewski. “Metalluraical Thermochemistry.” Academic Press Inc.. New Yurk, N. Y., 1951. ( 5 ) 11. Schaffer. G. Brei1 and C . Pfeffer, 2. anoro. Chem.. 276. 325 (1954). (6) 0. B. Skinner and R. A. Ruehrwein. T H l s JOURNAL, 59, 113 (1955). (7) D. G. Clifton and G. E. MacWood, THISJOURNAL,60, 311 (1956). ( 8 ) D . C. Ginninas and R. J. Corruccini, J . Research Natl. Bur. S f a n d a ~ d sS, 8 , 583 (1947). (Y) D. C. Qinnings, T. B. Douglas and A. F. Ball, ibid., 46, 2 1 (1950). (IO) J. M. Slisrfey. ihid.. 46, 299 (1951).
n o r - f r e e water Reaction well
Concentric ~ b s jars s Mercury lolled capillary
Rubber podding
--eruur Sllrrup SUFQOIl
Fig. 1.-Ice calorimeter. The reducing power analysis gave an empirical formula of TiCl3,ma. The formula obtained by the total titanium analysis was used for the calculations on the heats of solution of the trichloride. The contaminant was assumed to be TiCI,. Solvent.-The solvent used was a solution of HCI and FeCla. In preparing this solvent an HC1 solution, HCI. 8.859H20, and 370 g. of FeCla.6HnO were mixed to form 2 liters of solution. This concentration was chomn to duplicate a solvent used by Bilts and FendiusI2 as thcy have re( 1 1 ) J. A. Rahm, Anal. Chem., 84, 1832 (1052).
(12) W. Biltz and C . Fendius, 2. anoro. Chem.. 116, 40 (1028).
DAVID G. CLIFTONA N D GEORGEE. MACWOOD
310
Vol. 60
pork& for a solvent of this concentration, results which are used in submquent calculations. The denRity of the solvcnt at 27" was determined to be 1.1588 g./ml. This was used for calculations of the Ti-tosolvent weight ratios. Initially, an arbitrary weight ratio of Ti ion to solvcnt was estabkhed in thc heat of solution measurements of the TiCla. This same ratio was maintained throughout the heat of solution measurements, at 123.19 g. of solvent per g. Ti.
AHI, -21.4 f 0.1 kcal./mole, gives the heat. of formation of TiC13, AHs, in terms of the heat formation of TiC14(1)
Heat of Formation of Titanium Trichloride The heat of formation of Tic13 can be obtained if one has the heats of the reactions
a t 0 ° i s A H 6 = AH3-AH6= - 2 0 . 2 f 0 0 . 3 k o a l . / mole. With the aid of heat capacity data, this heat of reaction can be corrected to 298°K. The heat capacities used are
+ solvent = end soln. 1 + solvent = end soln. 2
TiCb(1) TiCl(s) Ti(s) FeCl,(soln. 2)
(1) (2) (3) (4)
+ 2Clz(g) = TiCh(I) + 1/zC12(g)= FeC13(soln. 1)
Adding together the heats for reactions (1) and (3) and subtracting those for reactions (2) and (4)) one obtains the heat of the reaction Ti(s)
+ 8/2Cll(g) = TiCl,(s)
(5)
The heats of reactions (1) and (2) were measured experimentally. The heat of reaction (3) is the heat of formation of TiCldl). The heat of reaction (4) was taken from the literature. Biltz and FendiusI2 made a direct determination of the heat for reaction (4)and found it to be -21.4 f 0.1 kca1.l mole. This same heat of reaction was also determined by MacWood13 in an HC1-FeC4 solution of stronger concentrations as -21.0 f 0.2 kca1.l mole. Since the solvent used here is the same as that of Biltz, his value will be used. The results of the heat of solution measurements are listed in Tables I and 11. Corrections were made on the Tic13 heats of solution for the Tic14 contamination as shown in Table 11.
.
TABLEI HEATOF SOLUTION OF TiC14IN SOLVENT Sample wt.,
Run No.
1 2 3 4 5 6 7 8
9 10 11
g.
1 2 3 4 5
TABLEI1 SOLUTION OF TiC1,
Sainplc wt., E.
TiCh a t . , R.
IN
I*cat iiieasrircd, CSl.
-
AHi, kcal./riiole
0.5775 123.3 ,5028 107.5 .5774 123.3 .5232 111.1 ,5411 114.9 ,6100 130.6 .5592 118.7 .5722 121.9 123.9 ,5792 119.8 ,5685 127.9 ,6064 Av. AH1 = -.40.4 f 0.2 kcnl./mole IIEAT O F
Run No.
Heat measured, cal.
40.5 40.6 40.5 40.3 40.3 40.6 40.3 40.4 40.6 40.0 40.0
SOLVENT Cor. for TiClr, cal.
0.4568 0.4513 1.2 115.7 .4053 .4004 1.0 103.3 1.0 .3788 ,3742 96.9 ,3787 ,3741 (36.0 1.0 .3814 .3768 1.0 96.7 Av. AH2 = -39.2 f 0.12 kcal./mole
-AfIv*
kcal./ mole
30.1 39.4 39.1 39.2 39.2
The combination of the measured heats of solution, A H , and AH,, with the literature value for (13) L'n~~~il~li~l~r~l,
A H s = 20.2
+ A H 3 f 0.3 kcal./mole
Therefore the heat for the reaction TiClr(s)
+ l/.~Cl?(g)= TiC14(1)
CpTic14(I) = 37.0 cal./deg. mole, Cpclz(s)= 8.82 0.06 X lo-, T
+ CPTiC13(,,) = 23.0 + 4.0 X
(6)
- 0.68 X
T-
lo5 T-*cal./deg. mole 1.7 X lo5 T - 2 cal./deg. mole
The heat capacity of liquid Tic11 is taken as the average of the values reported by Gmelinl4 and the National Bureau of Standards.'s The heat capacity of gaseous chlorine is that given by Kelley. l6 The heat capacity of the trichloride is not reported. The expression used above is the roundedoff value reported for vanadium trichloride by Kelley,16 and should be a fair estimate because of the similarity of the compounds. The change in heat capacity for reaction (6) is T + 2.04 X lo5 AC, = 9.59 - 4.03 X cal./deg.
mole
and, therefore AH6(2980) = -19.9 f 0.3 kcal./mole
Thus the heat of formation of titanium trichloride a t 298°K. is 19.9 f 0.3 kcal./mole more positive than the heat of formation of titanium tetrachloride. Recently, Johnson, Nelson and Prosen" have reported the heat of formation of TiCL(1) as - 192.1 f 0.6 kcal./mole. Using this value and the heat of reaction (6) a t 298°K. the heat of formation of TiCl.?(s)is obtained AHI~~= ~ , -172.2 ~ ( ~ ) f 0.7 kcal./mole
Discussion Schaffer, Brei1 and Pfeffer5 reported two independently determined values for the heat of formation of Tic&. Effectively, they measured the difference between the heats of formation of TiCld(1) and TiC!3(s), and in reporting the heat of formation of TiC13 based their calculations on the Bichowsky and RossiniI8 value of -181.4 kcal./mole for the TiCL(1). By combining their difference values with the lately reported heat of formation of TiC14(1),the heats of formation of TiCl,(s) reported by Pfeffer can be recalculated. They are - 172.3 f 0.8 and -172.8 f 0.8 kcal./mole. (14) Ginelin, "Handhuch der Anoraanisclien Chernie, Titan," WeinIieiin. 1051. (15) "Selected Values of Cheiiiical Thermodynamic Properties." Circular No. 500, National Bureau of Standards, 1952. (16) K. K. Kelley, United States Bureau of Mines, Bulletin 476. (17) W. H. Johnson, R . A . Nelson and E. J . Prosen, unpublished results, Reported to the Officeof Naval Research in National Bureau of Standards Report No. 3663 revised (1955). (18) F. R. Bichowsky and F. D. Rossini, "Thermochemistry of ('Iirinical Subshnres," Reinhold I'iilJl. Corp., New York, N . Y., 1936.
Mar., 1956
HEATO F FORMATION OF T I T A N I U M
Skinner and Ruehrweid determined the heat of formation of TiC13(s) by a direct method, reporting the value of - 170.0 f 0.8 kcal./mole. In view of the agreement between the present results and those of Pfeffer and co-workers by two independent methods, the difference between the heat of formation of Tic14 and TiCla appears t o be fixed within f0.3 kcal./mole. Therefore, unless
DICHLORIDE
31 I
there were an error in the heat of formation of TiCh, which is rather improbable, the heat of formation of TiC13 a t 298.16 is -172.2 f 0.7 kcal./ mole. Acknowledgment.-D. G.C. would like to express his appreciation to the Eastman Kodak Company for a fellowship which he held during part of this investigation.
THERMODYNAMICS OF THE TITANIUM CHLORIDES. 11. HEAT OF FORMATION OF TITANIUM DICHLORIDE1 BYDAVID G. CLIFTON~ AND GEORGE E. MACWOOD Contribution from the McPherson Chemical Laboratory, The Ohio State University, Columbus 10, Ohio Received August 16, 1966
The heat of formation of titanium dichloride has been determined by two independent methods. In the first method the heats of solution of TiC14(1)and TiClz(s) in a.solvent of HCl-FeCla were measured. In the second method the heats of solution of TiCla(s) and TiC12(s)in an HCI solution were used. The two values obtained are -123.3 f 0.7 and - 123.7 i 1.0 kcal./mole, based on - 192.1 f 0.6 kcal./mole for the heat of formation of TiClr(l).
I. Introduction Brewer3 and Kubaschewski4 report estimated values for the heat of formation of titanium dichloride. Skinner and Ruehrweins recently have reported a value for the heat of formation of the dichloride, calculated from disproportionation data. This research gives the differences between the heats of formation of Tic14 and TiClz and between TiC13 and Tic&. This information and that reported for TiC136can then be used t o give consistency t o the heats of formation of the titanium chlorides. 11. Apparatus and Materials The apparatus and some of the materials used for the measurements are described in the previous paper on Ticla.& The titanium dichloride was prepared by disproportionation of sublimed titanium trichloride.' The trichloride was placed in a nickel boat which was put in a Vycor tube in a furnace. The heating chamber was evacuated using a mercury diffusion pump backed by a mechanical pump. The trichloride was held at 485" for 8 hours and then at 470" for 10.5 hours. The titanium tetrachloride formed during the reaction was condensed out in a liquid air trap. Chloride analysis of the product gave an empirical formula of TiClz.oor. Total titanium analysis of the sample gave an empirical formula of TiC11.g9g. The sample dissolved completely in distilled water with slow evolution of hydrogen. This indicates that there was no appreciable metallic titanium impurity. Since the chloride analysis afforded the greater precision, the empirical formula used in all calculations was TiClz.oor. (1) Work performed under the Office of Naval Research, Contract No. Nonr-495(06). (2) Taken in part from the dissertation submitted by David G. Clifton in partial fulfillment of the requirements for the Ph.D. degree a t The Ohio State University, March, 1955. (3) L. Brewer, Paper 6, "National Nuclear Energy SerieR." Vol. IV-19B, edited by L. L. Quill, McGraw-Hill Book Co., New York, N. Y., 1950. (4) 0. Kubaschewski, "Metallurgical Thermochemistry," Academic Press Inc., New York, N. Y., 1951. 59, 113 (5) 0 . B. Skinner and R. A. Ruehrwein, THISJOURNAL, (1955). ( 6 ) D. G. Clifton and G. E. MacWood, T H I SJOURNAL, 60, 309 (1956). (7) W. C. Schumb and R . F. Sundstrom, J . Am. Chcm. Soc.. 55, 596 (1933).
The contaminant was assumed to be TiCls, and the necessary corrections on the measured heats were made. The same HCl-FeCls solvent as previously reported6 was used. The same weight-ratio of solvent to Ti-ion was maintained as in the previous experiments (123.19 g. of solvent per g. Ti). The water-to-HC1 ratio in this solvent was 9.69. The HC1 solution used in the heat of dilution measurements of the solvent analyzed as HCL9.60HzO. The waterto-HCI ratio in the solution is not exactly the same as that in the solvent. However, this slight difference between the two ratios does not result in an appreciable difference in the heats calculated. This same HCI solution was used to determine the heats of solution of TiCl3 and Tic12 in hydrochloric acid. The density of this solution was measured to be 1.0843 g./ml. at 27". The weight-ratio of solution to Ti-ion used for the trichloride and dichloride heats of solution in hydrochloric acid was 104.32 g. of solution per g. Ti.
111. Theory The heat of formation of titanium dichloride can be obtained by using the following set of reactions
+
TiCZ(1) solvent = end soln. 1 TiClz(s) solvent = end soln. 2 Ti(s) 2Clz(g) = TiCb(1) FeClz(end soln. 2) l/zClz(g) = FeC13(endsoln. 1 )
+
+ +
(1) (2) (3) (4)
Adding together the heat for (1) and (3) and subtracting those for (2) and (4), the heat of reaction ( 5 ) is obtained Ti@)
+ CL(g) = TiClz(s)
(5)
ie., the heat of formation of TiC12. Practically, complications arise in the above scheme. Reaction (2) involves the evolution of hydrogen. I n order t o use the above series of reactions, a correction must be made for the hydrogen evolution so that the end-solution after correction corresponds to the end solution 2. This required knowing quantitatively the amount of hydrogen evolved. If no hydrogen were evolved, one equivalent of ferrous ion would be formed for every equivalent of divalent Ti-ion introduced into the solvent. Hence, the difference between the number of equivalents of ferrous-ion in the end solution actually found by titration, and that calculated from the