THE NATURE OF IRON(II1) THIOCYANATE IN SOLUTION x,"
0
S. 2. LEWIN and ROSELIN SEIDER WAGNER New York University, New York, New York
A L m o u m the existence of a distinctive color reaction between iron(II1) and thiocyanate ions has been recogniwd and made use of for a t least 125 years (I),the first investigation into thenature of the productsformed in this interaction did not appear until 1931 (2). Since that time a number of additional reports have been published, but there has been no concordance in conclusions as to the uature of the species present in solutions of iron(II1) thiocyanate. As a consequence, many chemists consider that some doubt exists concerning the formula(s) to be assigned to the red color in these solutions. Thus, in the 1952 editions of three standard, widely used texts, one finds the following diverse statements reeardine iron(II1) thiocvanate: The red-colored compound has been given various formulas, Fe(CKS)s---, Fe(CNS):, Fe(CNS)zt, and Fe(CNS)++. The last formula is supported by the work of Bent and others [(S)]. The exset nature of the ferric thiacymate complex is still something of a mystery. When one considers a number of other x~ell-knownproperties of ferric iron, it might he thought that the comolex is FdCKS).---. As late as the war 1949. however. the question w a ~still consideled un~ettledby various writers in chemical literature L(4)I. More than one produrt can be formed when thiocyanate reacts with ferric ion. Depending upon the thiocyanate concentration a series of complexe~represented by Fe(CNS)n+"-", where n = 1, . 6, ran be obtained [(6)1.
..
It is the purpose of this paper to examine all the available evidence on the nature of iron(II1) thiocyanate. The reasons for the divergent conclusions of various authors are disrussed, and ambiguities and inconsistencies in the data are pointed out. These considerations will be shown to lead to the conclusiou that all the available evidence can be satisfartorily correlated 011 the basis that the interaction between Fe+++ and SCN- consists of the follorring series of equilibria: Fe+++ FeSCNt+ Fe(SCNh+ SCNFe(SCK).
+
++SSCNC N - e FeSCN++ Fe(SCN)>+ = Fe(SCN)s [or '/r FedSCN)i + SCN- = Fe(SCN)4$
(1) (2) (3) (4)
and probably also: Fe(SCN),Fe(SCN)&--
++ SCNe F~(SCN)S-SCN- e FC(SCN)~---
. (5)
(6)
Based uponaportion of the M S . thesis submitted by R. S. W.
to the Graduate School, NewYork University. 2 --.-.Prenmt,eci ~.~ a t t~h e~ -122nd Chemirnl -- - ~Mwt,inn --.. ~ - ~-~~ a of .-the ~ - American - ~ Society, Atlantic City, New Jersey, September 15, 1952.
The existing data can he classified as spectroscopic, migration, conductometric, ion exchange, and extraction. The following discussion is organized chronologically n.ithin these categories. SPECTROSCOPY
As early as 1890, quanditative measurements of the absorption of light by iron(II1) thiocyanate solutions showed that this system does not follow the Reer-Lambert-Bouguer absorption law, log (I& = rLc (6). This mas interpreted as shou-ing that iron(II1) thiocyanate, assumed to be Fe(SCS)3,must be appreciably dissociated into its component ions in solution. The first modern s~ectralstudv of the constitution of the species present in iron(II1) thiocyanate solutions was published in 1931 by Schlesinger and Van Valkenburgh (Z), who measured the light absorptiou of aqueous solutions made up from solid Fe(SCN), and NasFe(SCN)6, and of anhydrou~ether solutions of Fe(SCN)3. They concluded that the maiu features of the several spectra were sufficiently similar to establish that the same absorbing species must he responsible for the color in all cases. This conclusion was based upon the implicit assumption that the absorption spectrum of, for example, FeSCX++ n-ould he qualitatively different from that of, say, Fe(SCS)3. If, however, the spectra of these species are similar in main features and differ only in the positions and relative intensities of the absorption maxima, then the conclusions of these authors would be inadmissible-particularly in view of the fact that the instrumental technique available to them dld not permit the locatio~iof the absorptiou maxima to closer than +30 mF. The work of subsequent inrestigators, discussed below, has indeed s h o ~ wthat the absorption spectra of FeSCN++, Fe(SCN)2+,etc., are similar in main features, differing principally with respect to a progressive displacement of the absorption maximum toward longer wave lengths as the proportion of thiocyanate increases, On the basis of the assumption of identity of absorbing species in their solutions, Schlesinger and Van Valkenburgh deduced that the red color of iron(II1) thiocvanate is due to F e i S C N , -- - . , which in ether w o u l ~combine with ~ e + $ +to form Fe[Fe(SCK)6]. ~ ~weight measurements l ~ in ether ~ did seem ~ to SllppOrtthe latter in migration experiments in aqueous solution, discussed below, vere
l
JOURNAL OF CHEMICAL EDUCATION
asserted to show that the red color moves toward the positive electrode. Therefore, between 1931 and 1941 it was generally accepted that iron(II1) thiocyanate exists in aqueous solution principally as a complex ion having the formula Fe(SCN)6---. The next snectral investieation to he reported was that of von Kiss, e bra ha my and Hegediis in 1940 (7), who found that a marked shift in the wave length of the absorption maximum toward the red and an increase in extinction coefficientswere produced by the addition of an excess of KSCN. They concluded, in the light of the work of Schlesiuger and Van Valkenburgh, that with excess SCN- the principal absorbing species is Fe(SCN)6---, but that the molecule Fe(SCN)8 or Fe[Fe(SCN)B]predominates in the absence of excess SCN-.
-
= 0.011 M,ionic strength The circles nre experimental paint.; the dotted linesare oaloulated for SCX-to Fet++rstio.of V r , 1,s.and 3,reswotiuely.
[Feiitk is constant at 0.003582 M,[HCIJ
0.665, [SCN-lo is varied.
sorbing species is FeSCN++ a t low concentrations of SCN-, but as [SCN-lo grows, one or more higher complexes may he coming into the picture. This qualitative difference between solutions containing an excess of Fe+++ and those in which SCN- is in excess also is evident in the data of Frank and Oswalt (see below). Edmonds and Birnbaum tried t o fit their spectral data t o equations based upon the assnmpt,ion of a single equilibrium: h +nBeAB, The The eonoentrmtion of SCT- ir constant. while b i t + is varied. ciroles are experimental pointa: the dotted lines are ealoulated for r e * + + to SCN - ratios of %, 1. 2,and 3,respectively.
In 1941 the papers of Bent and Freuch (8) and Edmonds and Birnbaum (9) appeared, in which strong evidence was presented for the existence of FeSCN++ in iron(II1) thiocyanate solutions. The former showed that if only one equilibrium exists in a system, the slope of the plot of the logarithm of the optical density versus log [Fe+++]would give the number of ferric ions in the complex, and analogously for the thiocyanate ion. Their data are re~roducedin Figures 1and 2. The data show that wheu Fe;++ is in excess (Figure I), the only absorbing species is FeSCN++ a t all concentrations of Fe+++ (up to 0.01 M ) . On the other hand, for [Fe+++Io= 0 003682 M the principal ab-
(7)
and found the best fit for n = 1. In addition, they presented the graph reproduced in Figure 3, in which it appears that for the concentration range studied the transmission of iron(II1) thiocyanate solutions depends only on the total concentration, ( a b), and not upon which ion is in excess. Thus, according to these data, a solution in which [Fe+++Io= 0.002 M and [SCN-lo = 0.061 M has the same transmission as one in which the concentrations are reversed. Such a situation could exist only if there is a single reaction product, and it is a 1: 1 adduct; namely, FeSCN++. Although the main conclusion of Edmonds and Birnhaum that FeSCN++ predominates in their solutions is in good agreement-with the results of Bent and French, and of all the subsequent investigations described below. their result that this is the onlv sienificant constituent even a t [EN-10 = 0.064 "M i s in
+
SEPTEMBER, 1953
quantitative disagreement with the results of several of these investigations. I n the light of the data discussed below, it would appear probable that some of the data in Figure 3, perhaps only those applying to the higher concentrations of thiorvanate, are in error.
It will be noted that for a total concentration of 0.001 M, the maximum occurs a t 0.5 mole fraction Fe+++, corresponding to n = 1 in equation (8). For a total concentration of 0.01 M, n = 1.3; and when the total concentration is 0.1 M , n = 2.0. The progressive growth of n would be a consequence of the increasing proportion of one or more higher complexes with increase in concentration. The general trend of these data is substantiated by the work of Polchlopek and Smith (IS), who, also using continuous variations, found n = 1.5when [Fe+++Ia [SCN-10 = 0.1 M; n = 2.3 when [Fe+++Io [SCN-lo = 1 M; a n d n = 3.7 when [Fe+++]o [SCN-10 = 2 M. These authors state, ". . .definite evidence for higher complexes is found even a t concentrations no higher than a few hundredths molar." The same conclusion follows from the work of Frank and Oswalt (14) who, in 1947, published the results of another type of equilibrium study. They showed that for an equilibrium A B AB, if the initial concentrations a and b of the reactants are small, a plot of ab/D versus a b, where D is the optical density, should give a straight line. Figure 5 is the graph of their data, plotted in this may. The upper curve corresponds to [SCN-lo = 0.00030 JI, and increasing concentrations of Fe+++ (from 0.001 to 0.008 M); the lower curve to [Fe+++Io= 0.00030 M , and increasing concentrations of SCN-. I t is clear from the figure that at total concentrations greater than about 0.004 M, an excess of SCN- leads to a higher optical density than the same excess of Fe+++. Thus, the upper curve shows that when Fe+++ is in escess there is only one equilibrium involved; namely, the one leading to the formation of FeSCN++. When SCK- is in excess, even at as low total concentrations as 0.004 M , the presence of Fe(SCN)%+becomes erident. I n 1950 Harvey and Manning published still another
+
+
+
+ +
+
Figure 3.
Trrmsmksion of 11.0" (111)Thiooyanate Solutions docording t o Edmonds -d Birnbmvm (9)
The oiroles indicate ISCN-I and triangles [Fe+++lwhen the other ion concentration i. 0.0005 M for the upper cun.e. 0.001 M for the middle aurve, and 0.002 M for the lower curve.
One of the most generally useful methods of elucidating equilibria of the kind under consideration here is the method of continuous variations, described by Job (10). It can be shown that for an equilibrium l i e that of equation (7), a,,,, the initial concentration of A that yields a maximum concentration of AB., is related to n by the equation: n
=
M
amsx -,.a,
(8)
where M is the constant, total concentration of the system. In the case of iron(II1) thiocyanate, the optical density of the solution (proportional to AB.) is measured while [Fe+++]o (a) and [SCN-10 (16-a) are varied. Gould and Vosburgh (11), applying this technique to a system in which [Fe+++Io [SCN-lo = 0.02 M , stated that "there is no indication of the existence of any other compound [than FeSCN++] under these conditions." This is in accord with the results of Edmonds and Birnbaum, but in quantitative disagreement with the data of Babko, and Frank and Oswalt, discussed below, which indicate that Fe(SCN)%+is present in appreciable concentrations even under the conditions of Gould and Vosburgh. Babko (18) also applied continuous variations to the study of iron(II1) thiocyanate, and some of his results are reproduced in ~ i g u i 4. e
+
20 Figur. 4.
40 60 Mol per cent SCN-
continuous V.ri&on.
80
100
Study by Babko (12)
The total concentrations are: Curve A. 0.1 .M: B. 0.01 M: C.0.004 M: D,0.002 M: E,0.001M.
448
JOURNAL OF CHEMICAL EDUCATION
approach to the study of mobile equilibria (16). They pointed out that for an equilibrium mA nB = A,B., the ratio of the slopes of the optical density-concentration curves for the cases of (1) large excess of B, varying concentrations of A, and (2) large excess of A, varying concentrations of B, equals nlm. They applied this approach to iron(II1) t,hiocyanate, found the ratio of the slopes to be 1.09/1, and concluded in favor of FeSCN++. However, this procedure is based upon the assumption of a single equilihrium, and is not adaptable to a system of simultaneous equilibria. It can be shown that the existence of higher complexes than 1: 1 is not excluded by a slope ratio close to unity if the extinction coefficients of the several species are not very different a t the wave length of the measurements.
+
I .
3 -
0 X
4
k2.
I '
o
0.00s
0.001
o
rim.. F~~ the wrve,
5.
+b
~ ~ - t i ~ . t i ~ ~ of rrsnk ad
ISCN-~O ir
[Fei+%is kept at 0.00030 M .
F e i + * and
SCN-, D is optical density.
0.010 osws~t (14)
at 0.W30 M; far the lower a and b a r e initial oonoentrations of
The spectral evidence may be summarized as follows. The existence of FeSCN++ is firmly established by the work of Bent and French, Edmonds and Birnbaum, Gould and Voshurgh, Bahko, Frank and Oswalt, Polchlopek and Smith, aild Harvey and Manning. The existence of at least one higher complex is established by the work of von Kiss, Abraham, and Hegediis, Bent and French, Bahko, Frank and Osn-alt, and Polchlopek and Smith. The existence of more than one higher complex is indicated in the work of Polchlopek and Smith. Thus, taking into account only the evidence provided by absorption spectra, we may say that of the equilibria (1) to (6), (1) is definitely proved, (2) is almost as convincindv established, and (3) also is . . verv" nrobahlv . importancat large SCX' concentrations CONDUCTOMETRY
There have been three conductometric investigations of iron(II1) thiocyanate; that of Meller in 1937 (16), Ricca and Faraone in 1946 (IT), and Uri in 1947 (18). All agree in qualitative respects, providing evidence for the existence of FeSCN++ a t low concentrations of
thiocyanate, and in addition a t least one, and perhaps more than one, higher complex at larger concentrations. These conductance measurements are not, however, capable of furnishing as sensitive a qnantitative estimate of the relative stabilities of the several species as the spectrophotometric measurements discussed in the previous section. The kind of data obtained and the interpretations made may be illustrated by reference to the work of Ricca and Faraone. The conductance of iron(II1) ion is found t o he 138 ohm-', and that of the thiocyanate ion is 70 ohm-'. Upon adding one SCN- to olle Fe+++,the conductance would become 138 70 = 208, but if the ion FeSCN++ is formed, and if it is assumed that this ion has a conductance about equal to two-thirds that of the triplycharged ferric ion, the total conductance would be 92 ohm-'. Thus, the difference heheen the conductance calculated assuming no comhination, and the conductance assuming complete comhination would he approximately 116 ohm-'. Similarly, the conductmce difference for the formation of Fe(SCN)2+ would be 232, and for Fe(SCN)3it would be 318 ohm-'. Ricca and Faraone found that mith 0.004 111 FeCI3 and 0.01 M NH4SCr\' the conductance difference is about 100 ohm-', and with increasing thiocyanate concentration, the difference increases until it is approximately 200 ohm-' a t 0.1 M N H 8 C S . Hence, it-isconcluded that a t 0.01 M SCZT-, the ion FeSCS++ predominates, and with increasing SCN- concentration Fe(SCN)2+ and perhaps also Fe(SCN)s become important. Since this reasoning involves unverifiable assumptions concerning the conductance of the complex ions, the data cannot be relied upon in quantitative respects. Holi7ever, the order of magllitnde of the conductance differences would seem to establish the existence of FeSCN++ and F ~ ( S C N ) Zmith + reasonable certainty. Moreover, it is significant that the same result obtained is from absorption spectra data.
+
IONIC MIGRATION
The first migration experiments were carried out in 1931 by Schlesinger and Ban Valkenhurgh (E) who gave no details, reporting only that, ". . . electrolysis of aqueous nolutions of ferric thiocyanate causes the red color to migrate toward the anode, while ferric ions pass toward the cathode."
In 1941, Bent and Freuch (8) performed migration studies in an attempt to confirm the existence of FeSCN++ which they had deduced from spectral data. They found that with their solutions, ". . . R potential gradient of about 1 voIt/~m.w a found to give rim t o a slight movement of color across the boundary in tho direction of the cathode, hut no npprcciahle movement toward tho anode."
These statements prompted Schlesinger t o repeat and to publish the details of his migration experiments (19). I t was reported, for example, that with 0.1 N Fep(SO& or Fe(NO& and 1.0 N KSCN, under a
SEPTEMBER, 1953
potential gradient of 0.5 volt/cm., one colored boundary advanced 1.5 t o 2.0 mm. toward the positive electrode in ten minutes, and the other boundary receded 1.0 to 1.5 mm. from the negative electrode. I t will be noted that in Schlesinger's experiments the concentration of SCN- r a s very high; whereas in the work of Bent and French it was probably low. The apparent contradictiou between these results may, therefore, be explained as due to the presence of FeSCN++ and F ~ ( S C N ) Zvhen + [SCK-] is small, and of some anion such as Fe(SCiY)p-, or Fe(SCI\T)6--, or Fe(SCN)6---, when [SCN-I is large. This was, indeed, demonstrated by the migration experiments of Rabko (12), who showed that iu solutions containing 0.002 M Fe+++, the color moves toward the cathode when the SCN-concentration is in the rauge 0.01 to 0.02 M; when [SCN-] is between 0.05 and 0.1 M, the action of the colored layer is indefinite, leading to blurred houndaries; and between 0.2 and 0.4 IM [SCN-] the motion is distinctly ton-ard the anode. The migration experiments appear, therefore, to show conclusively that iron(II1) thiocyanate may exist as a cation or cations in the presence of low thiocyanate concentrations, and as an anion or anions when the thiocyanate concentration is large.
449
of a reference ion. For many ions, weights were obtained in good agreement with the formulas established by other methods. For 0.05 M Fe+++ in 2.0 N KSCN they found an ionic weight of 388. The weight of Fe (SCN)6--- would be 404. (It d l be noted that those concentrations are very close to those used in Schlesinger's migration experiments, in which one or more anionic species mere also found.) Additional proof of the existence of anionic irou(II1) thiocyanate is provided by the work of Teicher and Gordon (25), who showed that if [Fe+++]is 0.0004 to 0.0008 A t , [XH,SCN] is 1.5 M,and pH is 1, the irou (111) can he removed quantitatively from the solution by means of the anion exchaoge resin Amberlite IRA40OA. If the [SCN-] is less than 0.5 At, however, some iron passes through the column. Finally there should be mentioned the exchange experiments of Haenny and Wikler (361, who found that the exchauge between labeled Fe+++ and iron(II1) thiocyanate is rapid aud complete, showing that whatever equilibria exist between Fe+++ and SCy- are mobile and rapidly established.
The insight vhich the preceding discussion has provided into the nature of iron(II1) thiocyanate solutions may he summarized by reference to the equilibria (1) The extractability by ether and other organic sol- to (6). Equations (I) and (2) are well established from vents of the red color due to iron(II1) thiocyanate was the spectral and conductometric investigations, (3) from described as early as 1856 by Claus (20),and has served the extraction studies with organic solvenbs, (4) and as the basis of several procedures for the colorimetric perhaps (5) and (6) from the migration, ion exchange, estimation of iron (21). I t was pointed out by Durand and dialysis experiments. and Bailey (2%) in 1923 that the red color is not exSome questions still remain to be answered: tractable unless there is an excess of SCN-. This was (1) Are equilibria (5) and (6) significant a t accesconfirmed by Peters and French (2S), who showed that sible thiocyanate concentrations? with 1 part per million of Fe+++ in 0.01 N HC1, the (2) What are the equilibrium coostants of (2) to color does not begin to be extracted by ether until the ( G ) ? To date only the value of the constant, of equaratio of SCN- to Fe+++ reaches 464 t o 1. Further- t,ion (1) has been measured. more, mixing an aqueous solution of FeCI3with an ether (3) What is the structure of molecular iron(II1) solution saturated with iron(II1) thiocyanate results in thiocyanate in aqueous solution? the complete decoloration of the organic phase (17). From these facts it follom that molecular iron(II1) LITERATURE CITED thiocyanate must exist in solution, its concentration (1) BERZELIUS, J., "Lehrhuch do. Chrmie," 1826, Vol. 11, p. becoming significant, however, only in the presence of 771 a large excess of thiocyanate. (2) SCHLESIKGER, H. I., AND H. B. VAS VALKESBCRGH, J. Am. Chem. Soc., 53, 1212 (1931). Molecular weight determinations by Schlesinger and . . FOULK.C. W.. H. V. MOYER.AND W. M. MACSEVIN. VanValkenburgh (9)in anhydrous ether and benzene (3) "~u.&titrttivk Chemical ~ n i l i ~ s i a ,RlcGra\v-Hill " ~ook showed the formula to be Fe(SCN)s in these solvents. Co., Ine., New York, 1952, p. 130. This may also be the structure of molecular iron(II1) (4) PATTERSON,A., JR., AND H. C. THOMAS, ".A Textbook of thiocyanate in aqueous solution, although there is no Quantitative Analysis," Halt and Co., Kew York, 1952, p. 228. direct evidence on that score. EXTRACTION
..A.
Another type of experimentation that casts some light on the nature of iron(II1) thiocyanate is the rate of dialysis (i.e., rate of diffusion) work of Brintdnger and Ratanarat ($4). These investigators determined the effective ionic weight of a series of complex ions by comparing the rate of dialysis of the ion with the rate
(5) KOLTHOFF, I. M., A N D E. B. S.ASI)ELL, i'Te~tbookof Quantitative Inorganio Analysi~,"3rd ed., The Marmillan Co., New York, 1952,p. 635. (6) K ~ i i s s G., , AND H. MORAHT, Liebig's Ann., 260, 193 (18W); MAGNANIXI, G., Z . Physik. Chem.,AS, 1 (1891). (7) VON KISS, A., J. ABRAHAM, ASD I. IIEGED~S, 2. anovg. Allgem. Chem.,244, 98 (1940). J. Am. Chen~.Soc., 63,568 (8) BENT,H. E.,I R D C. L.FREYCH,
JOURNAL O F CHEMICAL EDUCATION
459
JOB, P., Ann. Chim., [lo] 9,113 (1928). COULD, R. K., AND W. C. VOSBURGH, J. Am. Chem. Soc., 64, 1630 (1942).
BABKO,A. K., J. Gen. Chem. USSR, 16, 1549 (1946); Compt. rend. Acad. Sn'. USSR, 52, 37 (1946). POLCHLOPEK. S. E.. AND J. H. SMITE.J. Am. Chern. Sm.. ~~,71. 3280 (1949).
'
.
FRANK, H. S., A N D R. L. OSWALT,ibid., 69, 1321 (1947). HARVEY, A. E., JR., A N D D. L. MASNING, ibid., 72, 4488 (1950).
MPILLER,M . , "Studies on Aqueous Solutions of Ferric Thiocyanate," Dana.Bogtrykerri, Copenhagen, 1937. RICCA, B., .AND G. FIRAONE, GWZ.Chim. Ital, 76,78 (1946). URI, N., J. Chem. Soc., 1947, 336.
SCHLESINGER, H. I., J. Am. Chem. Soe., 63, 1765 (1941). CLAWS, C., Liebig's Ann., 99, 50 (1856). WINSOR,H. W., Ind. Eng. Chem., Anal. Ed., 9,453 (193i); BERNHARD, A., AND I. J. DREKTER, Seime, 75, 517 (1932).
DWAND, J. F.,AND K. C. BAILEY, Bull. 8oc. Chim., 33,654 (1923).
FRENCH, C. L., A N D C. A. PETERS,Ind. Eng. Chem., Anal. Ed., 13, 604 (1941). BRINTZINGER, H., AND C. RATANARAT, Z. anorg. r~llgent. Chem., 223, 106 (1935). TEICHER, H., AND L. GORDON, Anal. Chem., 23,930 (1951). HAENNY, C., A N D E. WIKLER,Helu. Chim. Acta, 32, 2444 (1949).
