1717
HEATSOF COMBUSTION OF COMPLEX SALTSOF IRON(III)
Heats of Combustion and Bond Energies in Some Octahedral Iron(II1) Complexes with p-Diketones by David T. Farrar and Mark M. Jones Department of Chemistry, Vanderhilt Universitg, Nashville 6 , Tennessee
(Received September 11, 1963)
The heats of combustion of seven inner complex salts of iron(II1) have been determined by oxygen bomb calorimetry. From these, the heats of vaporization of the complexes, and other thermal data, the Fe-0 bond energies have been calculated for dissociation of the complexes to neutral species. The values obtained vary from 45.0 to 73.7 kcal. Values for the cleava,ge of these bonds to give the gaseous ferric ion and enolate anion can be estimated and are found to be higher by about 180 kcal. The thermodynamically preferred dissociation process for such gaseous complexes thus leads to electrically neutral species.
The present work was undertaken to obtain bond energies for the Fe-0 bond in several octahedral inner complex salts, The general procedures were similar to those used in previous work in this laboratory.’ Prior to this work there has been no experimental data available to determine how nearly constant a given metaldonor atom bond energy would be in a, series of environments which differ only slightly from each other. The data on boron compounds which has been collected by Brown and his co-workers2 indicate that coordinate bond energies tire subject to a very considerable variation which is interpreted primarily in terms of steric factors. In the complexes examined here the range in steric parameters is less than that studied by Brown for a very simple reason: moderate steric hindrance which will only weaken a bond in a tetrahedral complex of the type F&B:S R a ’will completely prevent the formation of an octahedral complex where six groups must be accommodated by the acceptor. This is seen in the absence of reaction between highly hindered 8diketones and iron(III[). Experimental Material. The acetylacetone used was Eastman White Label and was used without further purification. The propionylacetone was prepared by a Claisen reaction.* It was separat,ed as its copper complex. This was then decomposed with dilute sulfuric acid, the pdiketone was extracted into ether, and the resultant solution was dried with anhydrous sodium sulfate prior
to fractional distillation. The fraction boiling over the range 157-159’ was used. Benzoylacetaldehyde was prepared by the condensation of acetophenone with an excess of ethyl forma,te in the presence of a sodium slurry in ether.5 This diketone is unstable, and no attempt was made to purify it prior to the preparation of the iron(JI1) chelate. The benzoylacetone was Eastman White Label and was used without further purification. Propionylacetophenone was prepared by a Claisen reaction,6 as was butyrylacetophenone.’ Dibenzoylmethane was obtained by the method given in Organic Syntheses* and was recrystallized from ethanol prior to use. The method used to prepare the complexes was varied as the complex was more or less readily obtained in a crystalline form. A slight excess of ligand was used in all cases. To prepare tris(acetylacetonato)iron(III) the acetylacetone was added to an aqueous ferric chlo(1) (a) M. M. Jones, B. J. Yaw, and W. R. May, Inorg. Chem., 1, 166 (1962); (b) J. L. Wood and M. M. Jones, J . Phgs. Chem., 67, 1049 (1963). (2) H . C. Brown, D . M . McDaniel, and 0. Hafliger in “Determination of Organic Structures by Physical Methods,” E. A. Braude and F. C. Nachod, Ed., Academic Press, Inc., New York, N. Y., 1955, Chapter 14. (3) J. Marshall, J . Chem. Soc., 107, 518 (1915). (4) G. T. Morgan and H. G. Reeves, ihid., 123, 447 (1923). ( 5 ) K. Van Auwers and W. Schmidt, Ber., 58, 535 (1925). (6) K. Van Auwers and H . Jacobsen, Ann. 426, 231 (1922). (7) C. Beyer and L. Claisen, Bar., 2 0 , 2181 (1887). (8) A. Magnani and S. M . McElvain, Org. S y n . , 2 0 , 32 (1940).
Volume 68, Number 7
J u l y , 1964
DAVIDT. FARRAR AND
1718
ride solution and the pH was raised slowly to 4.5 with aqueous ammonia. The red precipitate was collected on a filter and allowed to dry a t room temperature. It was then recrystallized from carbon tetrachloride and dried a t 60". Tris(propionylacetonato)iron(III) was obtained by the addition of aqueous ferric chloride to the neutralized reaction mixture obtained in the preparation of propionylacetone. An oily red material was obtained which was allowed to evaporate and the complex obtained as a sticky solid after long standing. Tris(benzoylacetaldehydo)iron(III) was obtained by the interaction of an ethanolic solution of benzoylacetaldehyde with an aqueous ferric chloride solution. The addition of aqueous ammonia caused the complex to precipitate. It was then collected on a filter, was washed with small amounts of ethanol, and was dried a t 50". Tris(benzoylacetonato)iron(III) was obtained by a similar procedure, except that dilution with water was sufficient to precipitate the complex. I t was collected on a filter, was washed with water, was dried a t 55", was recrystallized from chloroform, and was dried again. Tris(propionylacetophenono)iron(III) was obtained similarly, but as a red oil. It was washed several times with small amounts of ethanol and dried in a vacuum desiccator a t 50". Tris(butyry1acetophenono)iron(II1) was obtained by the interaction of the sodium salt of butyrylacetophenone and ferric chloride in aqueous solution. The complex was extracted from this with chloroform, the extract washed with water, and the chloroform evaporated to obtain a very viscous red oil. This was washed several times with a waterethanol solution and was finally dried in a vacuum desiccator a t 50". Tris(dibenzoylmethano)iron(III) was obtained by the interaction of an ethanolic solution of dibenzoylmethane and aqueous ferric chloride. The completion of the precipitation was assisted by the addition of aqueous ammonia and further water. The complex was recrystallized from chloroform, was washed several times with ethanol, and finally was dried at 50". Analyses of the Complexes. A small weighed sample of the complex (0.3-0.5 g.) was put into a weighed cruciTable I : Analysis of Iron Complexes Complex
Tris( acetylacetonato)iron( 111) Tris( propionylacetonato)iron( 111) Tris( benzoylacetaldehydo )iron(111) Tris( benzoylacetonato)iron( 111) Tris( propionylacetophenono )iron(111) Tris( butyrylacetophenono )iron(111) Tris( dibenzoylmethano)iron( 111)
The Journal of Physical Chemistry
15.84 14 27 11.46 10.28 9.64 8 88 7.84
15.81 14.13 11.23 10.35 9 60 8.96 7.70
nlARK
M. JONES
Table I1 : Ligand Analyses" Theoretical %C % H
Ligand
Propionylacetone Dibenzoylmethane a
63.13 80,34
8.83 5.39
Found
%C
% H
62.72 80.08
8.87 5.39
Analyses by Micro-Tech Laboratories, Inc., Skokie, Ill.
ble and an excess of sulfuric acid added. The crucible was heated slowly to evaporate the acid and the liberated ligand, and when dry the ferric sulfate was ignited at red heat to Fe203which was the form weighed. The results of these analyses are summarized in Table I. Those ligands on which heats of combustion determinations were carried out were analyzed with the results given in Table 11. Calorimetry. The calorimetric measurements were carried out as described previously.' At least four determinations were carried out for each reported heat of combustion. Gelatin capsules were used as sample holders for the liquid diketones. Vapor Pressure Measuremants. The vapor pressures of the complexes and most of the ligands were determined over a range of temperatures by the method of Burg and T r ~ e m p e r . This ~ is an isoteniscope method and was used with a silicone oil constant temperature bath and a cathetometer which allowed the heights of the mercury columns of the isoteniscope to be read to j~0.01mm. No difficulty was encountered with the decomposition of most of the complexes and temperatures up to 130" could be used for all but tris(butyry1acetophenono)iron(ITI) which decomposed above 90". The vapor pressures were used with the Clausius-Clapeyron equation to obtain the heats of vaporization.
Results and Discussion The thermochemical data on the ligands and complexes are presented in Tables 111, IV, and V. The Fe-0 bond energies for the rupture of the complex into iron atoms and ligands from which one hydrogen atom has been removed were calculated using a thermochemical cycle of the sort shown in Scheme I. The relevant parameters and the bond energies are collected in Table VI. The bond energies all fall in the range 45.0-73.7 kcal. and are similar in magnitude to the bond energies in inner complex salts obtained previously. There seems to be no systematic indication of special interaction between the aromatic rings of the ligand and the chelate rings. though slight differences in a steric hindrance could easily off set any such trend. An aliphatic (9) E. Burg and J. W. Truemper,
J. Phys. Chem.,
64, 487 (1960).
HEATSO F
branch on the ligand will be more flexible than an aromatic substituent of the same size. These values are related to the bond energies for heterolytic cleavage (where both electrons in the coordinate bond go to th.e dcnor atom) by the equation
Table I11 : Thermodynamic Properties of the Ligands (kcal./mole) AE"
combustion estimated
AE
Compound
Acetylacetone Propionylacetone Benzoylacetaldehyde Propionylacetophenone Rutyrylacetophenone Benzoylacetone Dibenzoylmethane
AH
AH
vaporization
sublimation
6.5 5 6 10.Ob 10.Ob 10,Ob 13.3
combustion exptl.
784.2 1058,O 1364.1 1520.4 20.0' 12.0b
1719
IROS(II:[)
c O i V I B U S T I 0 ~O~F COMPLEX SALTS O F
1200.0 1764.9
a M. S. Kharasch and El. Sher, J . Phys. Chem., 29, 625 (1925). 'Estimated. ' A . Aihara, Bull. Chem. SOC.Japan, 32, 1242 (1959).
Table IV : Heats of Sublimation of the Complexes (kcal./mole)" Compound
AH8
Tris( acetylacetonatci)iron(111) Tris( propionylacetonato)iron( 111) Tris(benzoylacetaldehydo)iron(111) Tris(benzoylacetonato)iron( 111) Tris(propionylacetophenono)iron( 111) Tris( butyrylacetophenono )iron(111) Tris(dibenzoylmethano )iron(111)
5.6 10.0 13.3 10.9 6.3 7.6 7.6
Tris(propionylacetonato)iron( 111) and tris(butyry1acetophenono)iron(III) were not completely crystalline a t room temperature. They were sticky semisolids whiclh turned into viscous liquids when heated above room temperature.
where E(Fe-O)h,t is the bond energy for the splitting of the complex in such a manner that the products are the gaseous ferric ion and the gaseous anion derived from the ligand by the loss of a proton, E(Fe-O)hom is the bond energy obtained from a thermochemical cycle of the sort given in Scheme I, the I , are the ionization potentials for the loss of the ith electron from the iron atom, and E l i g a n d is the electron affinity of the ligand species from which a hydrogen atom has been abstracted. The data required for an exact calculation of the desired bond energies are not presently on hand, though it is possible to estimate these. The sum of the first three ionization potentials of iron amounts to 54.69 e.v. or 1161 kcal. The term E l i g a n d can be estimated to be approximately equal to the electron affinity of the oxygen atom for one electron, which is 1.48 e.v. or 34.1 kcal.l0 Using these values, E(Fe-O)h,t takes the value of 230 kcal. for the complex with acetylacetone. The major contribution to this can be seen to be the energy required to form the ferric ion, or alternatively, frdm the energy released when the ferric ion takes on electrons. This value is of the same order of magnitude as the few coordinate energies previously reported for complexes where charge neutralization occurs during their formation. l1 An examination of these previously reported coordinate bond energies (most of which were calculated rather than experimental values) shows that they fall roughly into two distinct groups. The first group consists of complexes in which the bond energies range from 58 to 134 kcal. ; these are complexes which possess the same charge as the cations from which they are derived. The second group shows values from 233 to 413 kcal. and consists of complexes which form with the neutralization of charge. The question then arises as to which of the bond energies calculated above is a more realistic, estimate of the strength of the Fe-0 bond. It
Table V: Combustion Data, in kcal. AHP,
Cornp o u n d
Tris( acetylacetonato )iron(I11) Tris( propionylacetonato)iron( 111) Tris( benzoylacetaldehydo )iron(111) Tris( benzoylacetonato)iron( 111) Tris(propionylacetopheiiono)iron( 111) Tris( butyrylacetophenono )iron(111) Tris( dibenzoylmethano :)iron(I11) a
Formula weight
AEo,
AEota
kcal./g.
kcal./mole
353,183 395.264 497.315 539.396 581.477 623.558 725.609
-5.315 -5.942 -6 229 -6.450 -6.978 - 7 057 -7.266
-1877 -2349 -3098 -3479 -4058 -4400 -5272
i1.8 f2.5 f3 . 3 +2.2 i2 . 1 i5 . 4 & I .9
An
-3.0 -4.5 -3.0 -4.5 -6.0 -7.5 -6.0
koal./mole
from elements
- 1879 - 2352 -3100 - 3482 - 4062 - 4404 - 5276
-347,s -361.7 -254.5 -359.4 -266.2 -411,l -179.8
Estimated uncertainties are indicated after these entries.
Volume 68, Number 7
July, 1964
DAVIDT. FARRAR AND MARKM. JONES
1720
Table VI : Data Used in the Calculation of Bond Energies, all in kcal. (Estimated Uncertainties are Indicated beneath Entries)
Tris( acetylacetonato)iron( 111)
-2226.8
Tris(propionylacetonato)iron( 111)
-2713.7
Tris( benzoylacetaldehydo)iron( 111)
- 3354 5
Tris( benzoylacetonato)iron( 111)
-3841.4
Tris( propiony1acetophenono)iron(111) Tris(butyrylacetophenono)iron(III)
-4328.2 -4815.1
Tris( dibenzoylmethano)iron( 111)
-5455.@
I
99.@ -lO1.Oc f0.4 99,8 - 120.8’ f0.7 99.8 - 61.2h 110. 99.8 - 81.2’ fl.1 99.8 - 79.2h f10. 99.8 - 84.gh 110. 99.8 - 54.6’ f2.9
-3.gd -3.9‘ -3.98 -3.9# -3.98 -3.9‘ -3.90
6.5e 5.6’ f0.2 f1.6 5.6” 10.0“ f 0 . 0 7 f0.5 1o.oi 13.36 fl0. f0.5 20.0’ 10.Q 6 320.2 f O .1 10.oi 6.3° f10. f0.5 7.66 1o.oi f10. f0.4 12.04 7.6e 1 5 . 0 f0.3
-187ge -146.8 11.8 12.7 -2352” - 94.2 f 2 . 5 f3.4 -31006 -175.7 f 3 . 3 f43. -3482e -253.0 f2.2 f4.0 -4062e -140.4 f 2 . 1 f43. -4404° -266.9 f 5 . 4 143. -5276e -132.5 1 1 . 9 f9.0
53.7 0.5 45.0
0.6
58.5
7.2
71.4
0.7
52.7
7.2
73.7 7.2 51.3
1.5
a Calculated from data given in ref. f and in the “Handbook of Chemistry and Physics,” 41st Ed., C. D. Hodgmar,, Ed., Chemical Rubber Publishing Co., Cleveland, Ohio, 1959. The oxide of iron resulting from the combustion was taken as FezOs. All the oxide D. R. Stull and G. C. Sinke, “Thermodynamic Properties of the Elements,” American residues formed in the combustions were red. Chemical Society, Washington, D. C., 1956. G. R. Kicholson, J . Chem. Soc., 2431 (1957). E. Funck and R . Mecke in “Hydrogen Bonding,” D. Hadzi and H. W. Thompson, Ed., Pergamon Press, London, 1957,pp. 433-441. This paper. Circular 500,Kational Bureau of Standards Washington, D. C., and from heats of combustion determined a t Vanderbilt University. Estimated to be the same as the value for acetylacetone. Circular 500, National Bureau of Standards, Washington, D. C., and from heats of combustion estimated by the method presented by M. S. Kharasch and B. Sher, J . Phys. Chem., 29,625 (1925). ’ Estimated from data on similar A. Aihara, Bull. Chem. SOC.Japan, 32, 1242 (1959). The standard deviations for AH,,, A H I , ~and ~ ~ AHtrana , are compounds. all 0.05kcal./mole or less.
’
Scheme I : Thermochemical Cycle for the Empirical Evaluation of Coordinate Bond Energies in Fe(C6H702)3
Fe(s)
+ 2102(g) + 15C(s) + 12Hz(g) - 1.5H2(g)-0.5Fe203(s) 3AHf,i0+ 3AH2,vap + 3AHtrana + -3E(O-H) + 6E(Fe-0) + 1.5E(H-H) AHOX
AHoX = AHl.vap
AHf.chei
AHf,chei
AH3,vap
=
is apparent that the lower value, E(Fe-O)k,,,, represents the preferred route for the disruption of the bond. Such large values as are reported for the union of gaseous ions of opposite charge may be of theoretical interest but are unrealistic as these do not represent effective measures of the stability of the coordinate bonds. They would also seem to be very misleading if they are used in the estimation of activation energies for the reactions of these metal complexes. The Journal of Physical Chemistry
+ 15c02(g)+ 2.5H20(1) + AH,
The effect of the experimental uncertainties on the bond energies of the Fe-0 bonds can be seen most readily in the spread of bond energies. This is from 45.0 to 73.7 kcal. in our computations. If the maximum (10) M. C. Day and J. Selbin, “Theoretical Inorganic Chemistry,” Reinhold Publishing Corp., New York, N. Y., 1962, p. 110. Reac(11) F. Basolo and R, G , Pearson, c~Mechanisms of tion,” John Wiley and Sons, Inc., New York, N. Y . , 1958, p. 60.
STUDY O F
1721
THE TELLUIRIDE I O N SYSTEM
uncertainty were to occur a t these extreme values, the spread of bond energies would be from 44 to 81 kcal. Thus the magnitudes of the bond energies change by from 1 to 7 kcal. These bonds are still much more easily split homolytically, and this conclusion stands even when the uncertainties in the thermochemical quantities are allowed. to run to their maximum values. Because each of the iron atoms have six coordinate bonds, the uncertainties in the bond energy of any one
bond will be only one-sixth of the uncertainty in the bonding energy for the entire complex. The method presented here can be used to obtain bond energies for the heterolytic cleavage process if the ionization potentials of the ligand anions can be determined. Acknowledgment. We wish to thank the U. S. Atomic Energy Commission for its financial support of this work.
A Study of the ‘Telluride Ion System
by Armand J. Panson Westinghouse Research LCkbOTatOTieS, Pittsburgh 86, Pennsylvania
(Received September 28,1963)
The valence with which tellurium cathodes dissolve was studied as a function of pH. It was found that the major product below pH 9 is Te-2 and above pH 12 is Te2-2. Polarographic analysis of solution composition and electrode weight loss measurements were used in the study. The potential of tellurium electrodes in contact with telluride ion solutions was measured and interpreted. An equation relating valence to pH was derived.
Introduction In the course of developing an electrolytic method for preparing metal tellurides from solution, a study of the pH dependence olf the valence oE cathodically dissolving tellurium waE; initiated. The results obtained in this study are reported here. Cathodic dissolution of tellurium in acid solution gives colorless H2Te1 while in alkaline solution red Tes-2 is obtained. No pH limits for the existence of the monotelluride and ditelluride ions are given in the literature; therefore, it was the purpose of this investigation to determine precisely the limits of stability for these two slpecies. Experimentally, this was done both by electrode weight loss rneasurements and by polarographic analysis for monotelluride and ditelluride ions in solution. Thus two independent methods were used. Measurements were made of the potential of tell~iriumelectrodes in contact with telluride ion solutions. From these data the solution
composition was calculated and found to agree with the polarographic analysis and electrode weight loss measurements. It was found that tellurium cathodes dissolve mainly as Te-2 below pH 9 and mainly as Te2-2 above pH 12. The results agree with an equation which was derived relating valence to pH.
Theory The composition of solutions obtained by cathodic dissolution of tellurium has been calculated from the following equilibria and associated equilibrium constants: the electrolytic reduction reaction
Te
+ ne- + 2(n - 1)Hf -+ 2 - n (n - 1)H2Te+ -2
Tez-2 ( 1
5
n
I 2)
(1) M. DeHlasko, Bull. Acad. Pol. Sei., Ser. A , Sci. Mat., 73 (1919).
( 2 ) E. Muller and R. Novakowsky, 2 . Elektrochem., 11, 931 (1905).
Volume 68,Number 7 July, 1964
