780
ROBERTC. BRASTED AND CHIKARA HIRAYAMA
with increasing X / P , reaching ca. 12 e.v. a t 2 volts/ cm./mm. This corresponds t o the energy of the first excited state of argon. Evidently the loss of energy by elastic collision is very inefficient and the energy of the electrons increases readily until inelastic collision is possible. On the other hand, a polyatomic molecule can undergo inelastic collisions a t much lower energies. The data show that e increases much more slowly with X / P and the increase continues to much higher values of X / P than in the case of a monoatomic gas. Consequently much higher values of X / P will be required to produce an appreciable number of electrons of the high energy required for chemical decomposition. In the argon-methane mixtures the possibility exists that either component could be the seat of the primary excitation. If it were argon it then would be necessary t o postulate a process such as A* CH, = CH3 H A to initiate reaction. The higher yields of methane decomposition in the presence of argon and the increase in yield with increas-
Vol. 63
+
ing proportion of methane tempt one t o adopt this view. However, the decision should not be made on this basis, for the electron energy distribution is certainly a function of gas composition and is undoubtedly a strong factor in determining the observed yields. Little can be said a t present about the nature of the products of these reactions other than t o note that they correspond, qualitatively, to those obtained in radiolysis reactions. It is difficult t o conceive of a radical reaction which could produce a product such as isopentane from methane a t room temperature. The possibility remains that ionic reactions may contribute some small part of the total product yield and may be primarily responsible for such a product. Further investigation to examine this and other aspects of the problem has been initiated.
(10) B. Rossi and H. Staub, "Ioniaation Chambers and Counters," Ch. 1, N. N. E. S., Div. 5, Vol. 2, MoGraw-Hill Book Co.,New York, N. Y.. 1949.
Acknowledgment.-The author wishes to acknowledge the several valuable suggestions and comments received from Prof. W. H. Hamill of the Universit,y of Notre Dame during the conduct of this work
+ +
AN EXAMINATION OF THE ABSORPTION SPECTRA OF SOME COBALT(II1)AMINE COMPLEXES. EFFECT OF LIGAND AND SOLVENTS I N ABSORPTION BY ROBERTC. BRASTED AND CHIKARA HIRAYAMA Contribution from the School of Chemistry, University of Minnesota, Minneapolis, Minn. Received October 24, 1968
There are numerous reports in the literature with regard to the absorption spectra of coordination compounds of cobalt(II1). Practically all of these reports are concerned with the absorption spectra in aqueous solutions. There are some inconsistencies in the data in regard to the absorption characteristics of some of the cobalt compounds. I n some cases the positions of maxima are not clearly defined in aqueous solutions, whereas some of these maxima become clearly defined in non-aqueous solvents. A study was initiated to coordinate some of the absorption spectra of some of the cobalt(II1) amine complexes, compare a.nd contrast reported spectra from other sources, and to make a study of the absorption spectra in alcohols as well as in aqueous solutions. Certain of the spectra are interprete,d in terms of crystal field theory. Experimental Preparation of Compounds.-The compounds were prepared by well-established methods.'-* Most of these com(1) H. F. Walton, "Inorganic Preparatione," Prentice-Hall, Inc., New York, N. Y., 1948, p. 91. (2) A. Werner, Ber., 40,4821 (1907). (3) S. M.Jorgensen, 2. a n o ~ uChem., . 14,416 (1897). (4) S. M.Jorgensen, J . prakt. Chem., 42, 211 (1890). (5) W. C. Fernelius, "Inorganic Syntheses," Vol. 11, John Wiley and Sons, Inc., New York, N. Y., 1946,p. 222. (6) W. C.Fernelius, ibid., Vol. 11, p. 222. (7) A. Werner and A. Frohlich, Ller., 40, 2228 (1907). (8)J. C.Bailar, Jr., "Inorganic Syntheses," Vol. IV, John Wiley and Sons. Inc., New York, N. Y., 1953,p. 176.
pounds were assayed by analyzing for cobalt, and the analyses in each case showed satisfactory agreement with the theoretical percentage of cobalt. The diaquo complexes were prepared only in solution, by hydrolysis of the corresponding dichloro complexes. Method of Analysis for Cobalt.-The analyses for cobalt were done by the electrolytic method.9 I n the analysis of the ethylenediamine and pro ylenediamine complexes, it was first necessary to treat t f e sample with concentrated sulfuric acid and ignite to remove t,he amine. The residue was then dissolved by treatment with nitric acid and a 30% HzOgsolution. The nitric acid subsequently was removed by evaporation to fumes after adding sulfuric acid. Determination of Absorption Spectra.-The absorption spectra in the various solvents were determined at room temperature by use of a Cary recording spectrophotometer. Fused quartz cells of two- and ten-cm. lengths were used. Certain of the com lexes isomerize as soon as they are dissolved, e.g., cis-dichyoro complexes of the tetramine, bis(ethylenediamine) and bis-( propylenediamine). Some are unstable in aqueous solutions in which they undergo rapid aquation, such as the chloro complexes. Still others are unstable to heat, such as the nitritopentammine cobalt( 111) chloride. Therefore, the solids of the more readily soluble compounds were weighed into volumetric flasks and the solutions were made immediately before determining the spectrum. In this way, it was possible to determine the spectra of these compounds within 15 to 20 minutes after the solutions were made. In those instances in which the compounds were only slightly soluble (as the nitritopentamminecobalt(II1) chloride in methanol) a solution was made by shaking a large excess of Lhe solid with the solvent followed by rapid filtration and then obtaining the spectrum. The (9) I. M.Kolthoff and E. B. Sandell, "Textbook of Quantitative Inorganic Analysis," 3rd Ed., The Macmillan Co., New York. N. Y., 1953, p. 410.
EFFECTOF LIGAND AND SOLVENTIN ABSORPTION
June, 1959
781
TABLE I OPTICALABSORPTION OF COBALT(III) AMMINESAND AMINESIN DIFFERENT SOLVENTS Compound
1,2- [Co(NH,)rClz]Cl 1,6- [Co(NH3)rCL]Cl
1,2- [Co(en)zClzIC1 1,6- [Co(en)zChIC1
Water
log
540 630 465 366 253 530 380 620
1.64 I .40 1.37 1.62
c
..
1.89 1.84 1.53
.. ..
1,2- [Co(pn)zClz]Cl
1,6- [Co(pn)zCl~]Cl
304 247 529 382 244 615
3.11 4.45 1.81 1.84 4.26 1.55
..
.. ea. 394
1.6 3.15 4.44 1.98 3.06 2.28 3.55 4.31 2.34 3.23 4.05
307 246 447 335 433 338 247.5 457.5 325 239 486 ca. 330 220 535 362 228 1,2- [Co(NHs)a(H20)Cl]Cl?
.. .. ..
..
.. *.
513 360 502 364 498 362 474 (ref. 349 (ref. 465 (ref. 349 (ref.
Wave length of maxima (md Methanol log a Ethanol
ea. 560 627.5 468 384 256 540
..
608 442 390 305 250 540
..
246 GO5 450 384 307 251 434 333.5 416 340 250 450 330 240
1.58 1.31 1.71 .. 1.93
..
540
log
e
Proganol
log
540
1.90
249 608 454 385 306 252-253
4.26 1.60 1.40 1.70 3.06 4.2
@
1.98
..
..
1.55 1.38 1.63 3.05 4.44 1.83
610 450 390 305 252 540
1.68 1.42 1.67
1.87
..
..
..
4.20 1.57 1.43 1.67 3.09 4.38 2.30 3.07 2.51 3.50 4.06 2.17 3.24 4.04
247 608 450 386 306 252
4.25 1,60 1.45 1.71 3.12 4.3G
..
.. 342 251
3.51 4.10
330 242
.. .. 223 532 364 231 530 360 240.5
10) 10) 10) 10)
concentrations of these solutions were not determined in most instances. However, since we arc chiefly interested in the positions of the maxima rather than the intensity of the bands, this procedure waR quite satisfactory. In instances in which the compounds aquate readily, such as the dichloro complexes, ths spectra were determined in two (or more) different parts. This process was accomplished by weighing two (or more) different portions of the solid in volumetric flasks, making the solutions immediately before measurement, and carrying out the measurement for any one solution over a short wave length region. In this way the spectrum was measured before it could be significantly affected by aquation. The nitro complexes are slightly soluble in alcohols, but
they are stable in both aqueous and alcoholic solutions. Therefore, these alcoholic solutions were made by shaking the solvent with an excess of the solid over a period of a halfhour t o several hours, filtering and measuring the spectrum. The concentrations of most of these solutions were estimated by evaporating a known volume of the solutioii to redness in a platinum dish and determining the weight of the residue. The nitro complexes are stable up to 80' (the t,emperature a t which the alcohol was evaporated) so that the concentration could be estimated to about &15%. Since the absorption increases greatly in the deeper ultraviolet region, it was necessary to make dilutions of ten- to a hundred-fold. This dilution was especially necessary in the region of 2500 A., where the absorption coefficient was
782
ROBERTC. BRASTED AND CHII~ARA HIRAYAMA
in the order of 10-20,000 as compared to approximately 75 in the visible. Ordinary distilled water was used to make the aqueous solution. The methanol, ethanol and 1-propanol were r& distilled after drying with sodium. The molar absorption coefficient is defined by the expression log ZO/Z = ed.
Experimental Results The positions of the maxima and log E values are shown in Table I. The detailed results are described for the individual compounds. cis-Dichloro-(ethylenediamine)-cobalt(II1) Chloride.The data of Table I show this compound to have two distinct. maxima, a t 530 and 380 mp, in aqueous solutions, but only one distinct maximum in the alcohol solutions. This compound isomerizes readily to the trans form, and the rate of isomerization is markedly affected by ultraviolet light. Linhard and WeigellO show a definite shoulder on the absorption curve at around 560 mp, with a distinct maximum at 625 and calculated bands at 392 and 312 mp. Basolo11 reports a diffuse fourth band a t 240 mp. The present work gives no indication of a shoulder in 560 mM region. A lateau at about 240 mp is indicated in aqueous solution. &he alcoholic solution, however, shows a continued increase in absorption. trans-Dichlorobis-( ethylenediamine)-cobalt( 111) Chloride.-Basololl reports four maxima for a 95-99 % methanol-water solution of this compound, these appearing at 252, 385, 450 and 625 mp. Linhard and Weigello re ort only two maxima in aqueous solution, although three otfers are calculated. I n the present investigation, three distinct maxima are found in aqueous solution, while in the alcohols there are five. The 450 and 390 mp bands are broad, permitting only & 2 mp estimate of the peaks. These two bands are overlapped by their neighboring bands in the shorter wave length region to such a degree in a ueous solution that the eak does not rise to its expecte% maximum. The b a n i a t 305 mp appears as a definite maximum with the peak slightly displaced to shorter wave length in the aqueous solution. I n dilute solutions this band does not appear to have a maximum in the alcoholic solutions, but, instead, appears as a plateau in this region. cis-Dichlorotetramminecobalt(II1) Chloride.-The spectrum of this compound was obtained only in the visible region. The compound is unstable in a ueous solution. A maximum in aqueous solution is foun8 at 540 mp. The spectrum of this compound in methanol could not be obtained due to its slight solubility and very rapid isomerization to the trans form. When an excess of the cis compound was shaken with methanol, the purple solution rapidly changed to green and subsequently exhibited a yellow turbidity. It was apparent that the compound reacted with methanol, resulting in the formation of the insoluble solid. This reaction with methanol also was noticed with the trans complex. The concentrations of the complexes whose spectra are recorded in Table I are: cis-complex in water = 8.6 X 10-4 M; trans-complex in water = 8.5 X M; and trans-complex in methanol = 3.7 X lo-' M . trans-Dichlorotetramminecobalt(TII) Chloride .-The present work shows four maxima for t h e compound. Tsuchida and Kashimoto12 report only three of these bands at 666, 475 and 312 mp. Linhard and Weigello report bands at 629 and 253 mp, with additional bands calculated at 475, 401 and 304 mp. I n the present investigation the maxima in aqueous solution appear at 630, 366,465 and 253 mp. cis-Dichlorobis-(propylenediamine)-cobalt(II1) Chloride. -The aqueous solution of this compound has three maxima at 529,382 and 244 mp (see Fig. 1 and Table I). The shape of the absorption curve of this compound suggests at least four bands similar to the analogous ethylenediamine complex, with an additional apparent band in the neighborhood of 600 mp where a slight hump appears. The f i s t band ap-gars at 540 mp in methanol! ethanol and propanol Solutions. owever, the absorption increases rapidly as the wave length diminishes, such that the band in the 380 mp region does not show a clear maximum. (10) M. Linhard and M. Weigel, 2. anow. allgem. Chem., 871, 101 (1952). (11) F. Basolo, J . A m . Chem. Soc., 7 8 , 4393 (1950). Japan. 11,785 (12) R.Tauchida and S. Kashimoto, Bull. Chem. SOC. (1936).
Vol. 63
trans-Dichlorobis-( ropy1enediamine)-cobalt(II1) Chloride.-As with the etEylenediamine malo there are five maxima in the alcohol solutions and only t k e e well-defined maxima in the aqueous solution (see Fig. 2 and Table I). The ultraviolet spectra are ractically the same as those for the ethylenediamine compEx. I n the 395 mp region in water the maximum is ill-defined. It is estimated that this maximum in the aqueous solution is located a t 394mp. Due to the overlapping of bands in the region between 450 and 380 mp the maxima in this region are determined to an accuracy no better than & 2 to 3 mp. The band a t 307 mp appears at practically the same position for all four solvents. The 250 mp band appears a t shortest wave length in water and at longest wave length in ropanol. Basolo11 reports only four bands in 95-99% metianol-water solution, these being a t 610,450, 380 and 255 mp. As was the case for the ethylenediamine analog, he does not report the 307 mp band. Nitropentamminecobalt(II1) Chloride.-The first maximum for this compound appears at a longer wave length in the aqueous solution than for a corresponding alcoholic solution. The maxima in the ultraviolet region appear at shorter wave lengths in the aqueous solution than in the alcohols. Tsuchida and Kashimotola report only two bands at 455 and 323 mp in aqueous solution, and they claim that the third band is absent. Linhard and WeigeP report a third distinct band at 238.6 mp, in agreement with our results. Nitritopentamminecobalt(II1) Chloride.-This compound was not sufficiently soluble in the alcoholic solvents to he able to obtain the absorption characteristics with the exception of the third band in a very dilute saturated solution of the complex in methanol. The band at 202 mp appears as a flat band with slight maximum at 220 and 223 mp in water and methanol, respectively. I n aqueous solutions there are bands at 485 and 330 mu. Tsuchida and Kashimoto12 re ort just two bands at 486 and 330 mp. Linhard and 8eigella report bands at 491, 361.5 and a shoulder in the region of 263 mp. There is a discrepancy in the position of the second band between Linhard and Weigel, on one hand, and Tsuchida and Kashimoto and the present investigators. Linhard and Weigel observed a distinctly peaked hand at 361.5 mp, whereas the band (this work) at 330 mp is somewhat overlapped. It is, however, definitely not in the region of 360 mp. There is also a discrepancy in the position of the third band. A much shorter wave length is therein indicated for this band. It was noticed that the nitrito and nitro complexes d;composed in aqueous solutions after being stored at 35 This decomposition takes place at a slower rate a t room temperature. A blackish-brown precipitate is formed. The decomposition appeared to be revented in acid solutions. The precipitate is cobalt(II1) Eydroxide,14 which is formed by the photodecomposition of the complex. cis-Chloroaquotetramminecobalt(I1I) Chloride.-The spectrum of this compound was obtained only in methanol solution. Three maxima were obtained at 530, 360 and 240.5 mp. Shimurals reports bands at 529 and 363 mp in aqueous solution. trans-Dinitrobis-( ethylenediamine)-cobalt (111) Nitrate.The first band in the ethanolic solution was overla ped by the second band Ruch that the maximum was ogscured. However, the first band in water appears at a longer wave length than in methanol. The second and third bands showed maxima at shorter wave lengths in water than in the alcohols. Basololl reports maxima in 95-99 yo methanolwater solution at 433,347 and 250 mp. cis-Dinitrobis-( ethylenediamine)-cobalt( III) Nitrate .This compound has only two absorption maxima between 700 and 230 mp. Both maxima appear at longer wave lengths in the aqueous solution as compared to those in methanol. Basolollreportsthree maxima for this compound in methanolwater solution at 438,325 and 240 mp.
.
Discussion Bands I and 11.-In the past five years the crystal field theory has been used very successfully in the interpretation of magnetic, spectroscopic and (13) M.Linhard and M. Weigel. 2. anorg. allgem. Chem., Z67, 113 (1961). (14) M.Linhard and M. Weigel, ibid., 866, 49 (1951). (16) Y. Shimura, Bull. Chem. SOC.Japan, 86. 49 (1952).
June, 1959
EFFECT OF LIGAND AND SOLVENT IN ABSORPTION
stereochemical data of discrete transition metal complexes.lB In the case of complexes of octahedral symmetry, the first two bands are due to electronic transitions within the central metal ion. However, in the case of a complex of tetragonal symmetry the first band is further split into two components, the magnitude of the splitting depending upon the relative magnitude of crystal field contribution of the ligands. It has been shown by Orgel" that the magnitude of the crystal field effect on the d-electrons of the central ion is in the increasing order of the spectrochemical series, i.e., the order I- > Br- > C1- > F- > HzO > (3204- > pyridine > NHI > en > NOz- > CN-, where en is ethylenediamine. The greater crystal field produced by one ligand compared to another, on a delectron of the central ion in a complex of identical Fymmetry, will manifest itself in the shifting of the first maximum to shorter wave lengths. The results of Linhard and Weigel'O show this shift for the ammine and ethylenediamine complexes of cobalt(II1). The results in Table I for the aqueous solutions agree with those of Linhard and Weigel. However, whereas some of the peaks are not observed in aqueous solutions, these peaks are actually observed in the alcoholic solutions in the present investiga tion. In the case of cobalt(II1) complexes of octahedral symmetry (hexamminecobalt(III), hexaquocobalt (111), etc.) a calc~lation'~ based on a simple perturbation theory shows that the band separation should be about 10,500 em.-'. However, it is found that the actual band separations are 8,500 and 8,000 cm. -l for hexamminecobalt(II1) and tris-(ethylenediamine)-cobalt(II1) complexes,1° respectively. The band separation between the same first-order electron configuration is related to the intermixing of molecular orbitals,'* which gives a measure of covalency. As the crystal field effect of the ligand increases (increasing order in the spectrochemical series) the band separation in the octahedral complexes decreases. Thus the separation is greater in the hexammine than it is in the tris-(ethylenediamine) complexes, and the hexacyano complex is even smaller a t 6600 em.-'.'* The decrease in band separation is, to a reasonable approximation, in the order of increasing covalency. The effect of changing the symmetry of the crystal field results in different orders of splitting of the electronic energy levels. The splitting of the first band is increased as the difference in the crystal field effect of the incoming ligand and that of the other ligand becomes greater." Thus a change from the octahedral symmetry of the hexammine to the tetragonal symmetry of the chloropentammine predicts a further eplitting of the first band. Linhard and Weigel14 have shown that the splitting of the halopentammine first band is greatest for the iodopentammine and that there is no splitting of the fluoropentammine first band. The broad f i s t band of chloropentamminecobalt(II1) is attributed to a splitting of this band. The nitro and nitritopen(16) See, for example. L. E. Orgel, J . Chem. Soo., 4756 (1963),and papers by Orgel and by Nyholm in "Proceedings of the Tenth Solvay Conference in Chemistry," R. Stoops, Ed.,Brussels. 1956. (17) L. E. Orgel, J . Chem. Soc., 4756 (1953). (181 C. K.Jorgensen, Acta Chem. Seand., 10, 500 (1956).
783
4.1 3.7
3.3 $2.9
3 2.5 2.1 1.7 1.3 240 280 320 360 400 440 480 520 560
600
(md. Fig. 1.-Absorption of cis- [ C ~ ( p n ) ~ C l ~ l C l . 2.0
-
I
I
I
I
0
1.8 1.6
4
-
6
o
-
I
I
I
I
In Water In Methanol In Ethanol In Propanol
-
' 1.4 -
-
bD
3
-,
1.2
-
-
1.0
-
-
0.8
-
I
I
I
I
I
I
I
I
_
tnmminecobalt(II1) complexes do not show evidence of splitting in the first band since these ligands lie very close t o ammonia in the spectrochemical series. Since the nitro ligand produces a greater crystal field than either the ammonia or amine ligands, the position of the first band for the latter ligands is more bathochromic than for the nitro complexes, Shimura16has determined the order of the spectrochemical series and has placed the nitrito ligand before ammonia. The earlier position of the nitrito ligand is manifested in the slightly longer wave length of the first band of nitritopentammine as compared to that of the hexammine complex. In disubstituted complexes, the crystal field theory predicts that the splitting of the longest wave length transition would be much more marked in the trans complexes than in the cis.10 As predicted by the theory the intensity of the first band of the trans-dihalo complexes is greater than that of the short wave length component. In addition, the cis-complexes show greater intensity than the trans complexee. Basolo, Ballhausen and Bjerrumz0 report similar observations. These latter (19) L. E. Orgel, J . Chem. Phys., 23, 1004 (1955).
ROBERT~C. BRASTED AND CHIKARA HIRAYAMA
784
Vol. 63
investigators also reported a splitting of the first ever, the difference ( A v a - Avb) remains quite conband of the cis isomer of [ C O ( ~ ~ ) ~ ( N O ~ )in C ~ ]stant. +, This value represents the splitting in the contrast to the trans isomer. They explained the first band of the trans-dichloro complexes. From splitting in the cis complex as being due to the much the constancy of this value, it seems that the effect .mailer contribution of the chloride, compared to of the chloro ligand on the symmetry of the field due one-half ethylenediamine and the nitro groups. The to ammonia, ethylenediamine and propylenediamultimate result is ascribed to a greater tetragonal ine is of the same magnitude. symmetry in the cis complex. The cis-chloroaquoBands I11 and 1V.-The results obtained in the tetramminecobalt(II1) chloride does not show any present investigation are in general agreement with splitting of the first band due to the small crystal those previously reported (vide infra). The only field contributions of the chloride ion and water as disagreement is in the absorption of nitropentamminecobalt(II1) chloride. The present results for compared to the chloronitrotetrammine complex. The trans-dinitrobis-(ethylenediamine)-cobaltthe positions of bands I and I11 agree with those of (111) does not show any splitting in the first band. Tsuchida and Kashimoto,12these being at 486 and Basolo" and Linhard and WeigelZ1also have shown 330 mp, respectively. Linhard and WeigeP reno splitting in the trans-dinitrotetramminecobalt- port band I11 at 361.5mp. There is also a disagree(111) Complexes. It is seen from Table I that the ment in the position of band IV between that of the first band of the trans-dinitro complex has the latter workers and that obtained by the present ingreatest hypsochromic shift. Both the cis and trans vestigators. There is, however, general agreement complexes exhibit a first band whose maxinium is in that the charge-transfer band I11 of this commore hypsochromic than the tris-(ethylenediamine)- pound is more bathochromic than that of the correcobalt(II1) complex. These relative band positions sponding nitro complex. It is thus indicated that can be explained easily on the basis of the nitro there is greater photosensitivity in the nitrito comgroups exhibiting a greater crystal field effect than plex. The greater photosensitivity of the nitro ethylenediamine. For this reason it would be ex- and nitrito complexes is shown by the comparapected that the trans-dinitro-(ethylenediamine)-co- tively rapid decomposition of these complexes in balt(II1) would have a more hypsochromic first aqueous solution in the presence of light. Effect of Solvents on Absorption.-The most band than the corresponding cis complex. The lack of splitting in the trans isomer may be attrib- prominent effect of solvent on the absorption is uted to the comparatively similar crystal field seen in the shifting of the maxima (see Table 11). contributions of the ethylenediamine and nitro There is a difference in the direction of the shift for ligands. the longer wave length bands depending on whether The replacement of a propylenediamine by two the solvent is water or an alcohol and on the type of chloro groups to form the dichlorobis-(propylenedi- complex concerned. The position of band IV, the amine)-cobalt(II1) complex results in the appear- most hypsochromic band, is always at the shortest ance of the first band a t a shorter wave length than wave length for the aqueous solution, increasing in the ethylenediaminecobalt(II1)complex. Assuming the order: water < methanol < ethanol < propanol. that the dihalo complexes follow the same order as It is to be noted that the diacid0 complexes show the pure amine complexes in their relative first band this band at longer wave lengths than the monopositions (as the trend seems to indicate from the acido complexes. The hexammine and tris-(ethylspectra for other diacidoammine complexes), the enediamine) complexes do not show either bands 111 propylenediamine ligand would be placed in a later or IV down to 200 mp.2a The cis-diacid0 complexes position than ethylenediamine in the spectrochemi- always have their respective bands at shorter wave length than those of the corresponding trans comcal series. In examining the frequency difference (YII - plexes. Via) (see Table 11) it is seen that the magnitude The most significant shift of the maxima with of this difference increases in the order of complexes solvent is observed in the visible bands. The maxof ethylenediamine < propylenediamine < am- ima of the first band of cis-dichlorobis-(ethylenedimonia. In every case the difference in the trans- amine)-cobalt(III), cis-dichlorotetramminecobaltdichloro complex is greater than that in the mono- (111)and cis-dichlorobis-(propylenediamine)-cobaltchloro and ammine complex. If the decreasing (111) complexes all appear at shorter wave lengths magnitude of (VII - VI&) is a measure of increasing in water than in the alcohols. The position of the covalency, then this order follows that of increasing band in the alcohols remains constant. On the other stability of the complexes since the stability of am- hand, cis-dinitrobis-(ethylenediamine)-cobalt(II1) mine complexes increases in the order ammonia, has the maximum for this band at longer wave propylenediamine, ethylenediamine.22 In this case, length in the aqueous solution. All of the trans it would appear that the spectrochemical series does complexes, nitropentammine, and chloropentamnot give the order of increasing stability of the lig- minecobalt(II1) chloride also have their first band and. ilt longer wave length in the aqueous solution. It is The values of ( Y I ~ YIb) for the trans-dichloro noted, on the other hand, that the magnitude of complexes show a fairly wide variation for the dif- the shift in the trans-dinitrobis-(ethylenediamine)ferent complexes in the different solvents; how- cobalt(II1) nitrate is greater than that of the corresponding cis-complex. (20) F Basolo, C . J . Ballhausen and J. Bjetiuin, A d a Chemzca The charge-transfer band, 111, in trans-dichloroScaiid., 9 , 810 (1955) (21) h l . Linhard and M. Weigel, Z . anorg. allgem. Chem., 278, 287 bis-(ethylenediamine)-cobalt (111) and the corre-
-
(1955).
(22) G. Schwarzenbach, Helu. Chim. Acta, 35, 2344 (1952.)
(23) M. Linhard, Z . Eleklrorhem., SO, 224 (1944).
P
',
EFFECT OF LIGAND AND SOLVENTIN ABSORPTION
June, 1959
785
TABLE I1 FREQUENCY DIFFERENCES IN MAXIMA (FREQUENCY IN Cia.-' X 108)' Compound
Solvent
trans- [CO(NH,)~CI,]C~
Water
trans- [Co(NH,)&lz]Cl cis- [Co(en)zCIA]Cl trans- [Co(en)zCln]CI
Methanol Water Water
trans- [Co(en)&lZIC1 trans- [Co(en)zClz]C1 cis- [Co(pn)zClzIC1 trans- [ C ~ ( p n ) ~ C l ~ ] C l trans- [Co(pn)pCI,]Cl trans- [C~(pn)~Cl,]Cl trans- [Co(pn)zClz]C1 trans- [Co(en)z(NO&]NO3 trarzs-[Co(en)z(NO~)~]NOa trans- [Co(en)z(NOr)z 1x03 trans- [Co (NHJ r C l ~C1 ]
Methanol Ethanol Water Water Methanol Ethanol Propanol Water Methanol Ethanol Water
trans- [Co (NHS)4CIJIC1 cis- [Co(en)ZClz]Cl trans- [Co(en)&lz]C1
Methanol Water Water
VI1
-
UIa
11.45 (9.o)b 10.1
...
(9. 8)b 9.11 9.2
...
YII
-
VIb
-
Avb
VI11
5.5
...
*.
3.61 3.4 7.3
5.5 5.8
(9.7)b 7.6 (7.8)b 7.2 6.9
..
...
..
..
.. .. .. 5.7 5.8 5.6
..
..
...
..
11.45 (9.0)b 10.1
5.85
5.60
4.65 7.45 (3.4)b
5.5
(9,8 y
-
9.2" (6.5)b
4.65 7.45 (3.4)b
3.9 3.70 4.0
...
UIV
5.60
9.7 9.5 9.5 9.6
...
4Avs
5.85
..
*. ..
...
8.0 7.2 7.0 ca. 7 . 0 10.8 10.6 10.6 9.2" (6.5Ib
... (9.7)b 7.6 (7.8)h 7.2 6.9
9.11 3.61 5.5 trans- [Co(en) ] C1 Methanol 9.2 3.4 5.8 trans- [Co(en)&ln]CI Ethanol .. ... 7.3 ... Water cis- [ Co (pn )zCL1Cl .. .. 8.0 9.7 Water trans- [Co(pn)tCI?IC1 7.2 3.9 5.7 Methanol 9.5 trans- [ C ~ ( p n ) ~ C l ~ ] C l 7.0 3.70 5.8 Ethanol 9,5 trans- [Co(pn)zClz]Cl ea. 7.0 4.0 5.6 9.6 Propanol trans- [Co(pn)ZClz]C1 10.8 .. .. ... Water trans- [Co(en)2(NOz)z]NO3 .. 10.6 ... .. Methanol trans- [Co(en)z(NOz)ZINOJ .. 10.6 ... .. Ethanol trans- [Co(en)z(NOZ)z]NO3 11.1 .. .. ... Water [COWHdsNOz IClz .. 11.3 ... .. Methanol [CO(NH~)~NOZIC~Z .. .. ... 11.1 Ethanol [Co(NK)sNOzIC12 .. ... .. ca. 15 Water [CO(NH: