29976
PATRICK A. MCCWKERA N D S. I f . SCHOLASTIC.\ KENN.\RD
A reasonable structure for the monoaquotridioxatiate of tetrachloroferric(II1) acid is not readily arrived at unless one assunies that hydrogen chloride c:m replace water in the coinpleu. 0 1 1 this (HCI) I 1 basis the forniulation [FeC12(C41180?)L(tT?O) [Cl]- C4HS02is a plausible one. IVhile the conductivity data in nitrobenzene intlicates ionic character for the anhydrous dioxanate of iron(II1) chloride atid for the diaquotritlioxanatc, the electric moments of these two complexes in dioxane indicate a low polarity in solution in dioxane. The value previously reported? for the anhydrous compound is 1.27 Debye units. This suggests that the solute species present in dioxane solution and the solute species present in nitrobenzene are markedly different. Apparently ionic character does not persist in dioxane solution and a covalently bonded species is present. The electric moment found for thc tliaquodioxanate of iroti(TI1) chloride in diox: t n c solution is 3 33 R h y c units. This indicates
[ CONTRIBUTIOS
Vol. S I
that again a non-ionic species is present in solution in dioxane. The increase in moment from the anhydrous diouanate to the cliaquo-complex indicates that the dinquo-solute species has a much greater polarity than the anhydrous coiiiplex. This woultl he expected as a result of the interaction between iron and water in the complex. The greater polarity of the quo-complex in dioxane solution is reflected in a marked difference in the solubilities of the anhydrous and atjuo-conipleues in dioxane. Data obtained a t 5 . O 0 show a soltibility, in moles of complex per 1000 g. of dioxane, of 0.OS-41 for the tliaquo-complex compared to 0,9521 for the anhydrous complex. Qualitative observations on the solubility of the monoacluodidioxanate of iron (111) chloride show that its solubility in diomne is intermediate between that of the anhydrous tlioxanate :mtl tlie tliaciuoditlioxnIlnte. YOIRP I)AMC, IUIIIAYA
FROM r € I E C l I E M I C A L L A D O R A . r O R I E S O F T H E UNIVIIRSITT: OF NOTRE D A M E ]
A Spectrophotometric Study of Anhydrous Iron(II1) Chloride and Tetrachloroferric (111)
Acid in Dioxane and Other Ethers',' BY PATRICK A. ATCCUSKER AND S. M. SCIIOLASTICA KENNARD, C.S.C. RECEIVED D E C E M B E R 27, 19.j8 T h e ultraviolet absorption spcctra of anhydrous iron(II1) chloride in several alkyl ethers are essentially t h e same, b u t the spcctra in various cyclic ethers differ from those in alkyl ethers and from oiic another. The assumption t h a t a single abt o lo-' molar, is sorbing solute species is present it1 ether solutii~nsof iron( 111) chloride, in t h e concentration range, supported by the fact t h a t such solutions obey Beer's Law a t the several wave lengths of maximum absorption. Spectra in mixtures of dioxane anti n-butyl ether show t h a t dioxane is a t least four times more basic than n-butyl ether toward iron(II1) chloride. Spectral data on anhydrous iron( 111) c1ilc)ritle-h?.tlrogeli chloride i n dioxane and other ethers is consistent with the assumption t h a t only two absorbing species, solvated iron( 111)chloride and tetrachloroferric( 111)acid, are present. Tetracliloroferric(II1) acid, in contrast t o iron( 111) chloride, is spcctri,photometric3lly identical in all the ethers studied. I n dioxane-aqueous hydrogen chloride solutions the cniicentration of tetrachloroferric(111) acid is tieternlined prirnarily by the Iiydrogen chloride-water concentration ratio. Dissociation constants of tetr3chloroferric( 111) acid in dioxane and several other ethers reveal t h a t the chloro-coniplex is ~ n u c hinnre staMc in ethers t h a n in aqueous solution and t h a t its stability varies from ether t o ether.
The formation of a very stable crystalline complex of iroti(II1) chloride and dioxane, as reported in the first paper of this series,' demonstrates that strong interaction occurs between ethers and iron(II1) chloride, a t least in saturated solutions. On the other hand dipole moment data, previously r e p ~ r t e ddid , ~ not provide evidence of strong polar interaction. Although only one solid compound, the 1 : 1 complex of dioxane and iron(II1) chloride is formed, the possibility remained that other solute species, in which more than one molecule of dioxane is associated with one molecule of iron(II1) chloride, might be present in dilute solution. It was, therefore, considered of interest to study the ultraviolet absorption spectrum of iron(II1) chloride in dioxane for evidence of other solute species in dilute solution. The ultraviolet spectrum of iron(II1) ( 1 ) Paper I1 of a series on the interaction of iron(II1) compounds with ethers. Paper I , THISJOTJRNAI., 81, 2974 (1939). From the ductoral dissertation of Sister hl. Scholastica Kennard, C.S.C., 1967. (2) Contribution from t h e Radiation Project operated by tlie University of Notre D a m e and sunported in part under Atomic Energy Commission Cnntract A T - ( 1 1-1)-38. ~ 1'. 1 A . hlrCllqker ~ and I3 C. Ciirran, THISI o i I w A T , ( 3 ) 1'. J. 1 6 4 , 2 K G (1912)
chloride in an anhydrous ether has been determined only in isopropyl ether,4although the visible spectrum in ethyl ether5has been reported. For this reason our spectral studies were extended t o include a number of ethers. Numerous studies have been carried out on the aqueous iron (111) chloride-hydrogen chloride syste11P-~ and on ether extracts of this system4J~"'~11 using spectrophotometric and other means. From these studies, the absorbing species in strong Iiydrochloric acid and in the ether extracts has been identified as tetrachloroferric(II1) acid. There appears, however, to have been no work reported on coinpletely anhydrous systems of iron(II1) chloridehydrogen chloride in ethers. We have investigated (4) N. H. Nachtrieb and J. E . Conway, ibid., T O , 3517 (1948). (5) H. I,. Friedman. i b i d , 74, 5 (1952). (G) V . S. Kiss, J. Abraham and I. Ilegedus, Z.anorg. allgem. Chem.. 244, 98 ( l % i O ) , (7) G . A. Gamelin and D. 0. Jordan, J . Chenz. SOL.,1453 (19:3). ( 8 ) I,. X . Mulay and P. W. Selaood, THISJ O I J R N A L , 77, 2693 (1975). (9) K. J. Myers and 11. IS. Rletzler, i b i d . , T 2 , 3772, 3776 (1950). ( I O ) J. Axelrod and E . €r. Swift, ibid., 62, 33 (1940). ( 1 1 ) I< J , M y e r s , I) I < hlvtzler and E. 1%. Swift, i b i d . , 72, :37ti:l (
I !I.-dl)
the system : anhydrous iron(II1) chloride-hydrogen chloride in dioxane and in several other ethers, and in dioxane-water mixtures, by means of ultraviolet absorption spectra. Experimental Materials.-Anhydrous iroxi(II1) chloride was prcpared and stored and Commercial Solvents 1,ii-dioxane was purified as previously d e ~ c r i b e d . ~Mathcson-Coleman pure grade tetrahydrofuran, Matheson, Coleman and Bell practical grade tetrahydropyran and Carbide and Carbon practical grade 1,3-dioxolane were dried by refluxing over solid sodium hydroxide and were distilled from sodium amalgam through a 2-ft. glass-helix packed column into oven-dried, metal-foil covered, glass stoppered receivers. The alkyl ethers were fractionated through a similar column and collected and stored under the same conditions. Absolute ethyl ether was distilled from sodium amalgam, Matheson, Coleman and Bell practical grade isopropyl ether was dried over phosphoric anhydride and distilled from sodium amalgam, and Eastman practical grade di-n-butyl ether was dried over phosphoric anhydride, refluxed over sodiuni and distilled. 411 ether fractions, taken for spcctrophotometric work, dist~lledwithin a 0.1 degree range; they were stored in their receivers in a dry box and used within 24 hours of distillation. Anhydrous hydrogen chloride was prepared by a slight niodification of a standard procedure12 and was subjected to further drying by passage through phosphoric anhydride. Preparation of Solutions.-All solutions were prepared and stored and dilutions made in a dry box, having open containers of phosphoric anhydride, and with dry nitrogen passing through the box. Solutions were protected from light by storing in metal-foil covered flasks. Storage period for solutions never exceeded 24 hr. Master solutions, 10-2 to 10-l 111, were prepared by weighing anhydrous iron( 111) chloride into glass stoppered flasks, adding solvent ether and reweighing. The contents of the flasks were cooled during addition of solvent since solution was exothermic and uncooled solutions discolored. Dioxane, tetrahydrofuran and alkyl ether solutions, 5 X M, were stable when piepared in this way, but more dilute master solutions were used with tetrahydropyran and dioxolane. Dioxolane solutions of iron(II1) chloride polymerized to a white waxy mass on long standing. Solutions for spectral measurements were prepared by dilution of weighed quantities of master solutions with weighed quantities of solvent. The molarity of iron(II1) chloride in the ether solutions was calculatcd directly from the weight and density data, assuming the detisity of the diluted solutions to be equal to the density of the pure solvents. Solutions of iron(II1) chloride in dioxane-n-butyl ether mixtures were prepared by adding n-butyl ether to dioxanc master solutions; dioxane was added to ethyl ethcr master solutions in the study of the spectra of iron(II1) chloride in mixtures of dioxane and ethyl ethcr. Master solutions of iron( 111) chloride-hydrogen chloride were prepared by passing hydrogen chloride gas over the surface of cold iron(II1) chloride-ether solutions of known molarity, in a system protected frorn atmospheric moisture, until the golden yellow color turned t o green. T h c hydrogen chloride content of the ether solutions was determined by dissolving or suspending the solution in water and titrating with 0.03 N base from a 10-ml. microburet, using a mixed indicator of phenolphthalein and brom thymol blue. Weighed quantities of solutions of iron(II1) chloridehydrogen chloride in ether were added to weighed portions of master solutions of iron(II1) chloride in ether in a dry box. Molarities of iron(II1) chloride and hydrogen chloride were calculated from the weight data, and the experimentally determined densities of liydrogcn chloride-ether solutions, listed in Table I. For the determinations of dissociation constants, the total iron molarity in the solutions containing iroti(II1) chloride arid tetrdchloroferric( 111) acid was determined from optical densities at isosbestic points and molar absorption coeficierits determined independently. Spectrophotometric Measurements.-Ultraviolet absorption spectra were obtained with a manually operated (12) R . h‘. Maxon, “Inorganic SyntheFes,” Vol. I, AIcGrdwv-Hill Book Co., 1939, New York, N. Y , p. 147.
Beckman model D.U. quartz spcctrophotometer, using a hydrogen discharge lamp as light source. Cell chamber temperature was 30 & 2 ” . Repeated observations demonstrated t h a t the temperature variation did not introduce errors exceeding the reproducibility of the instrument. When solutions exhibited instability on standing or on exposure to radiation, wave lengths of maximum absorption were established in a preliminary survcy with a second Beckman spcctrophotonieter having a Warren Spectrachord attachment. TABLE
DESSITIES
01.’
HYDROGENCIILORIDE-ETIIERSOLUTIONS
Isopropyl ether HCI, IM d2j
I$?,
0.000 ,029 .0:19 ,059 ,112 .2G8
0.000 0.7G37 ,034 .7641 ,067 .7lj50 ,100 ,7GO ,200 .7G70 .4GG ,7713
0,7188
,7195 ,7107 ,7201
,7210 ,7242
1
?&Butylether
d25
Tetrahydrofuran HCI. AI dz5
0.882-1 ,023 ,8828 ,051 ,8832 ,8830 ,088 ,112 ,8830 ,176 ,884ti
0.000
Dioxane HCI,
M
d25
0.000 ,075 ,371 . ,490 ,826
1.0279 1.0282 1.0202 1.0297 1,0309
1.082
1,0339
T h c optical dcnsities of solutions werc measured in imtclicd 10-tiim. quartz cells; rectilincar quartz spacers, 7 X 9 mm. in thickness, were used with the cells and permitted a choice of 1, 3 and 10-mni. light paths. Matched 10-cm ., glass-stoppered cylindrical cells were used with very dilute solutions. The path lengths of the cells, with and without spacers in place, werc checked with alkaline chromate ~ o l u t i o n . ’ ~Reference solvents were taken from the same fraction of distillate used in the preparation of the iron( 111) chloride solutions. For mixed ethers reference solutions with corresponding compositioiis were prepared from thc pure components.
Results and Discussion lVliile iron(II1) chloride has been shown to have a layer structure in the solid state14and to be dimeric in the vapor’j and in such solvents as carbon disulfide, chloroform and benzene,16 it is monomeric in ethers.17 I t is clear that the solvent ethers break up the layer structure of iron(1II) chloride and that the solute species present must involve some type of association between solvent and solute. Furthermore all the iron(II1) chloride must bc present in solution as a solvated complex; no unsolvated iron(II1) chloride will exist in solution. Data obtained from spectral measurements of to molar solutions of anhydrous iron(II1) chloride in various ethers are summarized in Table 11. In this table are listcd the ethers used as solTABLE I1 MOLAR ABbORPrION COEFFICIENTS FOR ANHYDROUS IRON(111) CHLORIDE IN ETHER SOLUTIONS Ether
Amax
Ethyl Isopropyl n-Butyl Dioxanc Tetrahydrrpy-rail TetrdhydrOfUrdn 1,3-Dioxoldne
220 220 225 235 278 233 236
€
A,,,*
7070 250 7200 250 5400 250 7585 272 7400 255-205 8570 255-265 5770 260-265
e
kmax
8410 342 8615 342 7025 342 6085 340 6400 340 6830 340 4809 340
f
8646 7120 7200 8115 7940 8825 5880
(13) G. W. Haupt, J . Research S n l l . B w . S t a n d a r d s , 48, 414 (1962). (14) N. w.Gregory, THISJ O U R N A L , 73, 472 (1951). (15) C . Friedel and J. hl. Crafts, Compr. rend., 107, 301 (18881. (16) A. F. Wells, “Structural Inorganic Chemistry,” 2nd E d . , Clarendon Press, Oxford, England, 1950, p. 110. (17) E. Beckmann. 2. bhysik. Chem., 46, 853 (10031.
vciits, the ,A,, for each ether-iroii(II1) chloride solution and the molar absorption coefficients at thesc inaxinia. Figure I shows the absorption spectra of iron(II1) chloride in dioxane, tetrahydropy-ran and n-butyl ether. These three curves are representative of the types of absorption curves found for a11 the ether solutions.
0 )D i o r o n e (21 T e t r o h y d r o p y r a n
iiip region. In the case ~f the dioxane solutions, this appears as a sharp peak at 572 m p . The reinaining cyclic ether solutions have an absorption b:md in the s ~ i i i cgciicral region, but it is much broader and occurs at somewhat shorter wave lengths. I n the 340 nip region a very small but definite difference evists in the position of the wave length of niaximuni absorption for the alkyl and cyclic cther solutions. AUlthe alkyl ether solutions of iron(II1) chloride have a,,A, a t 342 nip while the cyclic etlier solutions havc Am,, a t 340 nip. The iiiterniediate absorption band in the 260-2i5 iiip region appears to be characteristic of the cyclic ethers. I t is possible, however, that the intermetliate h m d does riot show up in the spectra of the alliyl ether solutions, for the reason that it is masked by the strong nearby band at 2 X m p . Kuinerous deterininations of cqitical densities, nt t lic several diN erelit wave leligtlis of iiiaxiiiiuiii absorption, on solutions of iron(II1) chloride in dios; m e and n-butyl ether, in the coilcentration range IU-d to I O F i niolu, showed that, within our experiiiiciital errcir. Beer's Law was followed. Data for the dioxane solutions arc given in Fig. 2. There is
1.1
--
12
e
X 7 4
11.8
0 -1
The molar absorption coelXciciit, xt 8.10 nip iii tlii)saiie. is considered to be within lCj, of the truc value. The 233 and 272 nip values in cliosane and the two values of niolar absorptions for the 12-butyl ether solutions are believec! t o be within Sc)A of the true values. The molar absorption coeEicients for the remaining ether solutions were obtained from relatively few nieasuretnents and represent order of iiingnitude only. Troii(II1) cliloride solutions in the three alkyl no evidence supplied by the absorption spectra of ethers have absorption niasitiin a t exactly the sanie iron(II1) chloride solutions in ethers that more wave lengths. Differences in the molar absorption than one solute species is present in such solucoefiicients are not considered significant because oi tions. I t niay then be concluded that the absorpthe rather large esperiinental errors in the deter- tion bands in these solutions arise from different tiiiiiatioii of the concentrations in solutions of the transitions in electronically identical iron(II1) volatile ethyl :md isopropyl ethers. Froin the atoms. identical wave lengths of maximum absorption, it Since the absurptioii curves for solutions of ironiiiay 1~ coiiclutlecl that the effect of solvent asso- (111) chloride in dioxane and in n-butyl ether arc, ciatioil 011 the clectroiiic. eiiergy levels in iroii(II1) is except for the 340 inp region, cluite different, the the S:IIIIC with all the alkyl ethers. 5pectra ol ironiI 11) chloride ill mixtures of dio.aiie ;tiid n-butyl ether c;tii lie used to deteririine the (lisWith the cyclic ethers. hiwevcr, tlie ~~I~si)r~)tii!ii i i i a s i i i i a : ~ wave t lengths below ;j00 iiiy are clifferelit tributioii of iron(1II) as dioxanate or di-n-butyl irotii those i i i the alkyl etlier solutions ant1 show tlif- ctlierntc iii mixtures of the two ethers. This disfroiri one anothcr. Only the cyclic ether tribution can gil-c information on the relative basicisolutioiis ~ h m van ahsorption h a n d in the 2(X) - 2 i . i tics i d the t\.ic, cthcrs toward iron(II1) chloride and
possibly indicate a relationship between the relative basicities and the characteristic absorption spectra in the two ethers. The spectra of 1 X IO-' molar iron(II1) chloride in 5 i . 3 mole yo dioxane-. n-butyl ether and in 32.9 mole yo dioxane-n-butyl ether were measured and are shown in Fig. 3. It will be observed that, even in the 32.9 mole dioxane solution, which has a large excess of n-butyl ether, the characteristic shape of the spectra of a dioxane solution still persists. The intermediate band in each of these solvent mixtures is shifted toward a shorter wave length and becomes less prominent, as the mole fraction of iz-butyl ether increases. Theorctical curves were then constructed for the absorption spectra of iron(II1) chloride in mixed dioxaneeiz-butyl ether solutions containing 70, i 3 . 3 and hO mole dioxane. These curvcs were constructed on the assumption that the ethers were of equal basicity, that one rnoleculc of cther was associated with one molecule of iron(II1) chloride and that the distribution of iron(II1) chloride as dioxanate or di-iz-butyl etheratc would, therefore, be proportional to the mole fractions of the ethers in the solvent mixture. Data for the theoretical curves were obtained using the relation
x,
t o l d = WrX
+
eHw0
(1 -
x)
whcre C D is ~ the molar absorption coefficient of iron(II1) chloride in dioxane, E R ~ ~isO the niolar absorption coefficient of iron(II1) chloride in nbutyl ether, and X is the mole fraction of iron(II1) chloride present as dioxanate. I n the calculation it is assumed that Beer's Law is followed in the mixed ether solvctit and t h a t the iron(I1I) chloride is all present either as monodioxanate or mono-u-butyl etherate. An examination of Fig. 3 reveals that the ititermediate band, characteristic of the iron(II1) chloride dioxanate, is more prominent in the experimental 32.0 mole dioxane curve than in the thcoretical curve for i 3 . 3 mole dioxane. I t is also clear from Fig. 3 t h a t the intermediate band is less prominent in the experimental 32.9 mole yo s o h tion than it is in the theoretical curve for SO mole yo dioxaric. This permits the conclusion that, in the 32.0 mole "/o dioxane solution, more than i3.3y0 and less than SOYc of the iron(II1) chloride is present in solution as dioxanate. Using the equilibrium expression
(z
cx
[FcCla.B~nOl[Dx! = [FcCla.I)x]IBu,O]
for the reactinn FeCla.Dx
+ Bu.0
+ L3x
Fc.C1:i,Bu20
in which n-butyl ether may be considered to rcplace part of the dioxane in the association complex. upper and lower limits for the value K may be calculated. In this calculation the ratio of the concentrations of the two ethers is taken as equal to the original solvent conipositioii, all the iroti( 111) cliloride is assumcd to be present as coniplex, either dioxanate or pz-butyl etherate, and one molecule of ether is assunied to be complexed with iron(I11) chloride. X lower limit of 0.12 and an upper liniit of 0.18 were obtaiiiecl for K , using 73 mole 5%as the
220
250
:%lU
280
;MI
h inp.
Fig. 3. -i%l,surptioii curveb for IzeC13 i i i w b u t y l etlicrdioxaiie iriistures : 1, experimental curve 57 mole dioxaiie; 2 , experimental curvc, 32.9 mole dioxane ; 3, theoretical curves, 80 mole yo dioxane , 73.5 inole 70dioxane - _ _ ,- 70 mole c/o dioxane - - -.
(x,
cz
-1
lower limit of the concentration of diouanate and SO mole % as the upper limit. Use of these values of K in the equilibriuni expression gave results showing that, in an equimolar mixture of dioxane and n-butyl ether, the ratio of the concentrations of iron(II1) chloride dioxanate to iron(II1) chloride nbutyl etheratc lies between 3.5 and S.l. If i t is assumed that 2 molecules of n-butyl ether are associated with one molecule of iron(II1) chloride, a similar calculation gives a dioxanate-n-butyl etherate ratio of 7.7 rather than 5.3. Thus, even if the assuniption of the presence of mono-n-butyl etherate is not correct, the qualitative result is not changed. Since, however, the calculation on which the theoretical curve is based involves the use of molar absorption coetlicients which may be in error by as much as 3y0,the precision of the above relative basicity values is not high. T o obtain some additional data on the relative basicity of dioxane and di-n-butyl ether toward iron(II1) chloride, other values for K were calculated from optical density data a t 250 mp, a t which wave length there is a maximum difference between the molar absorption coefficients of iron(II1) ehloride in dioxane and in n-butyl ether. Values for K ranging from 0.13 to 0.20 were thus obtained. These values are close to those obtained from a consideration of the shapes of the experimental and theoretical curves shown in Fig. 3. I t appears then t o be well established from the above considerations that dioxane is at least four times more basic than fz-butyl ether towards iron(II1) chloride. An indication that a second alkyl ether was also inuch less basic than dioxane toward iroii(II1) c1ilor;de was given by obscrvatioiis of the spectrum of iron(II1) chloride in a diouaiie-ethyl ether mixture containing 57 mole (& dioxane. In this solvent mixture the three hands characteristic of iron-
(I1 I ) chloride diosaiiate were very promineiit, while the 230 mp barid characteristic of alkyl ether solutions of iron(II1) chloride was absent. I t thus seems reasonable to conclude that the stronger solvent-solute interaction that occurs in dioxane solutions of iron(II1) chloride, in comparison to alkyl ether solutions, is responsible for the difference in the electronic energy levels of iron (111) in the different solute species. 111 Fig. 4 is shown a plot of molar absorption coefficients against wave lengths for dioxane and clioxane--hydrogen chloride solutions of iron(I1I)
7 i ~ i o l a r it , ~ shows up in thc dioxane solutioiis a t a much lower hydrogen chloride conccntration. I n Fig. 5 are shown the results of iiieasureinents at the wave lengths of maximum absorption for solutions of iron(III1 chloride in dioxane-hydrogen chloride at varying hydrogen chloride conccntratioris. These results show that, a t hydrogen chloride concentrations which are above 3.0 molar, there is substantially complete conversion of iron(II1) chloride to tetrachloroferric( 111) acid and
0
2
4
G
8
IO
HCI, ,!I.
220
250
3x0
810
340
: