1496 ROBERTB.

HAGELAND LEONARD
F. DRUDING Inorganic Chemistry

complex has an (X)-asymmetric nitrogen a t the end of (~~-1,3,8,1O-~~e~trien)(T\TOz)zf and trans-C0(~~-3,8-

chelate ring. The vicinal effect from this (X)-asym- Illeztrien) (NOz)Z+ should have been determined by the
metric nitrogen has been obtained from the circular vicinal effect from the (R)-asymmetric nitrogen. This
dichroism spectra of t r a n s - C 0 ( ~ ~ - 1 , 310-Me4trien)Clz+
,8, effect is shoxm in Figure 7 also and this vicinal effect
and trans-Co(1,~-3,8-nle~trien)Clz+ and is represented in has positive Cotton effects near 20,800 and 23,500
Figure G with thatz9from (-)D-tmns-Co(N-~leen)Clzi- cm-'. From the Yamatera theory, the lower energy
and trans-Co(d-pn)zC12+. These two vicinal effects are band may be the ]A1 +- 142(D4h) component and the
very similar and have a positive Cotton effect for the higher energy band may be the ' A 1 -+ 'E(D4h) com-
81, + component and a negative Cotton effect ponent of the 'A1 'Tl(Oh) d-d transition. The
for the 'A1 -+ 'A2 component. This result indicates signs of the Cotton effects induced from the (R)-asym-
that the vicinal effect from (X)-asymmetric nitrogen metric nitrogen are different for the corresponding com-
does not depend on the positions of the two (R)-asym- ponent of the trans-CoNdClz+ and trans-CoN4(KOz)z+
metric nitrogen ( i e . , trans for trans-Co(N-Meen)2C12+ chromophores.
and cis for trcins-Co(~~-1,3,8,lO-Me~trien)Cl~+), as well
as the fact that this effect is essentially similar for Acknowledgment.-The authors thank Professors
these two tuans-CoN4Clz+ions which have (X)-asym- Kazuo Saito and Junnosuke Fujita for measurements of
metric nitrogen. circular dichroism and Kcnzo Fujii for the elemental
The difference in the circular dichroism for trans-Co- analyses.

CONTRIBUTION
FROM THE DEPARTMENTOF CHEMISTRY,
RUTGERSUNIVERSITY,NEWARK,NEW JERSEY 07102

The Triamines of Cobalt(II1). I. Geometrical

Isomers of Trinitrotriamminecobalt(III)la
BY ROBERT B. HAGELlb AZJD LEONARD F. DRUDING

Received October 27, 1969

A number of preparations from the literature for trinitrotriamminecobalt(II1) were examined for their structural arrange-
ment. Evidence for the existence of the facial and meridial isomers was established by proton magnetic resonance spectra
and thin layer chranatography and by comparison with other cobalt(II1) triamines of known geometry. Additional proper-
ties of these isomers studied included their infrared and ultraviolet spectra.

Introduction CN)3CL8 Less clear has been the situation with

In Alfred Werner's classic explanation of the theory trinitrotriamminecobalt (111), which was first prepared
of structure of coordination complexes, he submitted in 1866 by E r d m a n r ~later
; ~ preparations were reported
considerable evidence for an octahedral arrangement of by WernerlO and Jorgensen.ll KO evidence was then
groups for coordination number of 6. Most important given for the structures of these compounds or even
evidence for this structural assignment was that for whether different isomers were represented. The first
octahedral complexes containing two different ligands study of the various preparations of C O ( N H ~ ) ~ ( K O ~ ) ~
A and B in the general formula MA,Bs-, where n = was reported in 193812 by Duval, who examined five
2, 3, or 4, only two geometrical isomers are possible. methods of preparation by visible, ultraviolet, and
Where n = 2 (and the equivalent case of n = 4), infrared spectroscopy, refractive index of aqueous
many hundreds of cis-trans isomer pairs have been solutions, conductivity, and X-ray powder diffraction.
prepared and characterized. However, for n = 3, These measurements indicated that two different
or MA3B3, only five such sets of geometrical isomers2 products resulted when the Werner and Jorgensen
have been firmly identified: R u ( H ~ O ) & , ~ 'Pt-~ preparations were compared. Duval assumed that
(NH3)Br3+,jPt(NH3)313+,6 Ir(Hz0)3C13,7and Rh(CH3- these mere the two geometrical isomers of the complex
but he did not assign configurations from his data.
( 1 ) (a) Presented a t the 158th h-ational Meeting of the American Chemi-
cal Society, Sept 7-12, New York, N. Y., 1969. (b) Abstracted from the
In 1952, Tanito, Saito, and K ~ r o y a determined
'~ the
Ph.D. dissertation of R . B. Hagel, Rutgers University, 1969, crystal structure of the product of the Jorgensen prep-
(2) T h e recommended designations for the structures shown in Figure 1
are facial ( f u c ) or 1.2,3 and meridial (mev) or 1,2,4: R. F. Trimble, J . C h e m
aration to be that of the meridial isomer.
Educ., 31, 176 (1954). (8) B. D. Catsikis aud 11.L. Good, i b i d . , 8 , 1095 (1969).
(3) R.E. Connick and D. A. Fine, J . Am. Chem. Soc., 83, 3414 (1961). (9) H. Erdmann, J . Pvnkt. Chem., 97, 406 (1866).
(4) E . E. Mercer and W.A. McAllister, Inovg. Chem., 4 , 1414 (1965). (10) A. Werner, Z . Anorg. Allgem. Chem., 1 5 , 166 (1897).
(R) V. S. Oriora and I. I. Chernyaev, Z h Neovgan. K h i m . , l l , 1338 (1566). (11) S. M.Jorgensen, ibid., 17,475 (1898).
(6) V. S. Orluva and I. I. Chernyaev, ibid.. 12, 1274 (1967). (12) R. Duval, J . P w e A Q p l . Chem., 13,468 (1938).
(7) A. A. El-Awady, E. J. Bounsell, a n d C. S. Garner, Inorg. Chenz., 6 , 79 (13) Y . Tanito, Y. Saito, and H. Kuroya, Bull. Chem. SOC.JaQan, 26,
(1967). 188 (1952).
Vol. 9, No. 6, June 1970 TRINITROTRIAMMINECOBALT(III)
1497

Beattie and Tyrrell14 reported infrared spectroscopic This solution, containing CO(COa)sa- was cooled in an ice
data for a number of nitroamminecobalt (111) complexes bath and 7 g (0.13 mol) of ammonium chloride, 5 ml (0.07 mol)
of 28y0 aqueous ammonia, and 10 g (0.15 mol) of sodium nitrite
including Co(NH3)3(NO& prepared by the Jorgensen were added. This mixture was heated on a steam bath until it
method. While they could clearly distinguish between turned red-brown. After acidification with 15 ml of 6 N acetic
cis- and trans-dinitrotetraamminecobalt(II1) ions, the acid, the solution was heated for another 30 min. When the
positions of the absorption bands of C O ( N H ~ ) ~ ( N O ~solution
) ~ was allowed to stand overnight in a refrigerator, massive
a t 1320 and 1290 cm-1 lie somewhat between the ex- crystals of product were obtained with only a small amount of
cis-dinitrotetraamminecobalt(II1) ion as an impurity. The
pected values for either the facial or the meridial yield was 1.5 g (21%). Anal. Calcd for fa~-Co(NH3)a(N02)~:
isomer, although they did accept the Tanito, et al., Co, 24.1. Found: Co, 22.9.
assignment of meridial structure to their preparation. mer-Trinitrotriamminecobalt(III), Co(NHt)t(NOz)a.-This pro-
A more complete examination of the infrared spec- cedure was basically that described by Jorgensenll (a method
troscopy of these complexes was carried out by Majum- claimed to produce the fac isomer), as modified by Cooley, Liu,
and Bailar18to improve the purity of the product. A solution of
dar, Duval, and LeCompte. l5 They concluded that both
geometrical isomers did exist and that the Werner prep-
aration yielded the meridial isomer, while the Jorgensen
preparation gave the facial form! No mention was made
of the X-ray crystal analysis by Tanito, et al. l3 Postollec
and Mathieu16 confirmed the two isomers reported by
Majumdar, et al. , but recognized the previous X-ray
structural evidence. In their brief communication,
they presented the infrared spectra of the 1,2,4 (mer)
isomer prepared according to Jorgensen's method and a. MERlDlAL 8. FACIAL

the 1,2,3@c) isomer prepared by Werner's procedure

Figure 1.-Isomers of CoA3Ba.
In yet another study of the infrared absorption
spectra of a number of cobalt(II1) amines and assign- 90 g (0.38 mol) of oobalt(I1) chloride hexahydrate in 250 ml of
ment of the observed bands, Maddock and Todesco17 water was added to a solution containing 100 g (1.85 mol) of
disputed the earlier claims that the two geometrical ammonium chloride, 135 g (1.95 mol) of sodium nitrite, and 345
isomers had been prepared for C O ( N H ~ ) ~ ( N O ~ ml ) ~ of
. 28% aqueous ammonia in 900 ml of water. A rapid stream
They did prepare a nonelectrolyte whose Co :NH3 :NO2 of air was bubbled through the mixture for 4 hr. The solution
was evaporated under an air jet to half its original volume.
composition analyzed as 1:3:3 and to which they The resulting crystalline product was suction filtered, washed
assigned the meridial structure. They presented with cold water, ethanol, and ether, and air dried. The crude
infrared spectral data for the ionization isomers (cis- product was thoroughly ground in a mortar and again placed on a
CO(NH~)~(NO~)LJ(CO(NH~)~(NOZ)-~) and (trans-co- suction filter. Twelve 25-ml portions of cold water were succes-
( N H X ) ~ ( N O(Co(NH3)z(N02)4)
~)Z) and concluded that the sively poured over the crystals and suctioned off to remove am-
monium chloride. Recrystallization was carried out by pouring
facial isomer of Majumdar, Duval, and LeCompte, successive 200-ml portions of 1.5 M acetic acid a t 60" onto the
which would still give the correct chemical analysis, product, stirring, and filtering. The treatment was continued
was, in fact, one of the ionization isomers. until about 5 g of material remained. From 5 to 7 1. of acid was
In this work, a reexamination of the various methods required. The solid on the filter, consisting mostly of the ioniza-
of preparation of trinitrotriamminecobalt(II1) was tion isomer (Co(NHs),(NO,)z)(Co(NHa)z(NOz)a), was discarded.
The combined filtrates were evaporated to half their original
carried out, and their products were examined by volume under an air jet. The recrystallized complex was filtered,
chromatography and paper electrophoresis for evidence dried, and ground, then washed with water, ethanol, and ether,
of the two geometrical isomers. and air dried. The yield was 21 g (24y0). Anal. Calcd for
~zer-Co(NHa)a(NOz)a: Co, 24.6. Found: Co, 23.5.
Experimental Section Purification of fac- and mer-Co(NHa)a(NOp)a.-Both prepara-
1. Preparations.-The purity criteria used for all of the prepa- tions still contained measurable amounts of cobalt-containing
rations of cobalt complexes were their analyses by thin layer impurities so that further treatment was required for samples
chromatography and paper electrophoresis. These methods were used for spectroscopic experiments. Taking advantage of the
more sensitive t o cobalt-containing impurities than the spectro- limited solubility of these complexes in acetone compared to the
photometric methods traditionally used. insolubility of the principal contaminant, the ionization isomer,
~ ~ 500 mg of starting material was extracted with three 10-ml
fac-Trinitrotriamminecobalt(III), C O ( N H ~ ) ~ ( N O ~ ) ~ . - Tabout
method described by Shibata, et al.18 (a method claimed to pro- portions of reagent grade, dry acetone. This extraction left
duce the mer isomer), was found to give the best yields. A cold in the residue most of the ionic impurities. The extracts were
mixture of 7 ml of water with 7 g (0.03 mol) of cobalt(I1) chlo- combined and reduced in volume to about 5 ml.
ride hexahydrate and 10 ml of 30r0 hydrogen peroxide was added A small chromatographic column was constructed of 8-mm 0.d.
dropwise to an ice-cooled slurry of 23 g of potassium hydrogen Pyrex tubing packed to a depth of 10 cm with silica gel H (Brink-
carbonate in 20 ml of water, and the resulting green solution was man Instruments, Westbury, N. Y . ) , activated a t 110' for 20
quickly filtered. min, and wet with acetone after it had cooled. The acetone
extract of the complex was placed a t the top of the column,
(14) I. R. Beattie and H. J. V. Tyrrell, J. Chem Soc., 2849 (1956). about 25 ml of acetone was added, and air pressure from the
(15) A . K. Majumdar, C. Duval, and J. LeCompte, Compt. Rend., 247, 302 laboratory line was applied to the top of the column. The first
(1958).
(16) M. L. Postollec and J. P. Mathieu, ibzd , 957, 3185 (1963). 30% of the eluent contained some anionic cobalt(II1) complexes
(17) A. G. Maddock a n d A. B. J. R. Todesco, J . Inovg. Nucl. ChPm., 26,
1535 (1964). (19) W. E. Cooley, C. F. Liu, and J. C. Bailar, J . A m . Chem. Soc.. Si, 4189
(18) M. Shibata, M. Mori, and E. Kunyo, Inovg. Chem., 3, 1573 (1964). (1959).
1498 ROBERTB. HAGEL4 N D LEOn'aRU F. DRUUING Inorgunic Chemisiry

and was discarded. The remaining eluent, which required 2-5 were spotted midway between the two electrodes and the system
min t o pass through the column, mas collected and evaporated to was electrolyzed for 1 hr by a 475-V dc, a current of 4-4.5 mA.
dryness under an air stream. Any cationic impurities remained Trials a t various voltages showed that below 350 V , no movement
as a band a t the top of the column. Trinitrotriamminecobalt- of the complexes took place while above 500 V, several of the less
(111) purified in this manner produced a yield in which no other stable complexes underwent decomposition. Table I1 lists
cobalt-containing impurities could be detected by thin layer
chromatography. From the sensitivity of the analysis, an esti- TABLE I1
mate of the purity is 99.9%. ELECTROPHORESIS hIIGRA4TION RATESO F
mer-Trinitrotris(methylamine)cobalt(III),C O ( C H ~ N H ~ ) ~ - SOXE COBALT(1II) N I T R O A M I N E S
(N02)3.20-A 170-ml sample of 2570 methylamine and 50 ml of Migration Migration
dilute acetic acid were added to 30 g of sodium hexanitrocobaltate- rate, rate,
(111) dissolved in 300 ml of water. The mixture was stirred Complex mm/hr Complex mm/hr
vigorously for 1 hr after which the pH was adjusted to 5-6 by the mei~-Co(riH~)~(NOr)~ 0 Co(PjH3)g(NOda- ,Q6
mer-Co(CHsNHp)a(NOds 0 C o (NO210 3 - +143
addition of 35 ml of glacial acetic acid. After 2-3 hr, pale yellom, mer-Cofdien)(NO*)3 0 cis-Co(NHa)a(SO~)r -92
crystals were deposited, but the mixture was left overnight in jac-Co(PiHs)8 ( r i 0 1 ) 3 0 21,ans-Co(NH3)a(NOz)li -91
order to complete the crystallization. The product was suction
filtered, washed with cold water, alcohol, and air dried. The migration rates for some of the complexes. The sign indicates
yield was 8.5 g (38y0). Anal. Calcd for nzev-Co(CHLXH2)2- the polarity of the electrode toward which migration took place.
(r\;02)3: Co, 20.6. Found: Co, 19.2. 4. Spectrophotometry.-Visible and ultraviolet spectra of
mer-Trinitrodiethylenetriaminecobalt(III),Co (d ien)( S O Z ) ~ , the complexes were measured with a Cary Model 14 spectro-
was prepared by the method devised by Crayton and Mattern.21 photometer, in matched 10-mm silica cells. Solvents used were
cis- and tvans-dinitrotetraarnminecobalt(II1) nitrates (Co- water, acetone, and DMSO.
(SHs)r(?\:Oz)p)(N031,were prepared by the direction of Schles- Infrared spectra were determined with a Beckman Model IR-10
singer . 2 2 spectrophotometer using Nujol mulls, KBr disks, and disks of
2. Thin Layer Chromatography.-Thin layer chromatography pure complex. The Nujol mulls and KBr disks gave identical
was run on precoated, preactivated silica gel plates, Type Q-1, spectra for the complexes. While there is reason to doubt the
with stevch binder (Quantum Industries, Fairfield, N. J.), as reliability of the spectra obtained from the pure-complex disks,
supplied by the manufacturer. Sample sizes of 1-4 +I of either samples run in KBr pellets when the sample concentration was
acetone or dimethyl sulfoxide (DMSO) solutions of the com- varied from 1 to 60% showed no change in spectra. The use of
plexes were applied and the chromatograms were developed. the disks of pure complex greatly enhanced the spectra in the
To separate cobalt(II1) complexes of different charge, a solvent 400-600-cm-1 region.
system of CH30H-DMSO-HC104 (60 :40 :0 $1)was used. Re- 5 . Proton Magnetic Resonance Spectra.-Pmr spectra were
sults were similar to those reported earlier.23 I n this solvent measured with a Varian 4-60 instrument a t a frequency of 60
system, anionic complexes ran with the solvent front, and neu- MHz a t 25'. Sweep rates of 250 Hzlsec and sweep widths of
tral complexes ran near the solvent front, while cationic com- 250 and 500 Hz were used. Samples were dissolved in DMSO
plexes trailed, depending on their charge and geometry. (concentrations of ca. 5y0 by weight) and run against a TMS
To separate the geometrical isomers of Co(SH8)3(S02)a, ace- external standard.
tone was found to be the most effective solvent. Table I lists
Discussion and Results
TABLE I
I. Examination of the Methods of Preparation.-
THINLAYERCHROMATOGRAPHY O F The principal methods of preparation of trinitrotri-
S O M E COBALT(II1) hTITROAMISESa
Ri/Rf (inci
aminecobalt(III)lO~ll~
1 8 , 1 9 , 2 4 , * 5 were examined for the

Ri isomer)
TABLE I11
0.76 1 .00
0.85 1.15 OF VARIOUS
ANALYSIS PREPARATIVEMETHODS
FOR

0.80 1.05 Co(NHa)a(NOz)a

0.62 0.8'2 Ref Primary product Impurities
1 .00 1.30 10 ~ ~ c - C O ( N H ~ ) ~ ( N C0(XHa)z(X02)4-
O~)~
0.08 0.1 22, 24 fu~-Co(XH3)a(?jOz)a C~S-CO(NH~)~(KOZ)Q+
0.08 0.1 cO(r\T&)~(N02)4-
a Solvent, acetone; adsorbants, silica gel Q-1. 18" ~ u ~ - C O ( X H ~ ) ~ ( N cO~~s )- C
~ O ( X HN02)sQ
~)~(
CO(NH~):,(NO~)~+
the Rf values of the various complexes as well as ratios of K i / 110 I~Y-CO(NH~)~(NO tuans-Co(NH,)4(NOa)~"
~)~
Rr(wzev isomer). Rf values were found to be very sensitive to CZ~-CO(NH~)~(NOZ)~+
conditions of development while the ratios were always reproduc- CO (I&)* (NOz)4-
ible to +0.02. 19 mev-Co(i\;Ha)s(S02)8 tl.ans-Co(NHa)4(SOe).L'L
The developed plates were sprayed with a water-alcohol solu- c~s-CO( SHa).i(NO2)z"
tion of nitroso-R sodium salt and heated to 110" for 15 min. I n Co(NHa)%(N02)a-
some cases, it was necessary for the plates to stand for 2-3 days u Claimed to produce wzev isomer. Claimed to produce j h c
for certain of the complexes to develop a color. The cobalt ap- isomer.
peared as an orange spot against a pale yellow background.
3 . Electrophoresis.-Paper electrophoresis was carried out geometry of the trinitrotriammine formed and for the
on a Research Specialties chamber on 20 X 50 cm Whatman E o . 1 principal impurities by thin layer chromatography and
paper. The electrolyte was 0.01 M NHaN03. Samples (1-4 pl)
paper electrophoresis. Table I11 lists these prep-
(20) I. I. Chernyaev, A. G. Ganiev, Sh. Sh. Tarkhtaev, and Sh. Sh. arations and the identity of the products before
Rizaeva, Zh. Neovgan. Khirn., 13, 157 (1968).
(21) P. H. Crayton and J. H. Mattern, J . I v o , g . S 7 r c l . C h p v . . 13, 248 purification by the acetone extraction and chromatog-
(1960). raphy.
(22) G. G. Schlessinger, "Inorganic Laboratory Preparations," Chemical
Publishing Co., K e w York, N. Y . , 1962, p 246. (24) W. G. Palmer, "Experimental Inorganic Chemistry." Cambridge
( 2 3 ) (a) I,. F. Druding and I<. B. Hagel, A:zal. C'lrem., 38, 478 (1966); University Press, London, 1854, p 53Y.
(I,) R . B. Hagel a n d I,. F. Druding, Sefiou. Sci., 4, 89 (1969). (25) Reference 2 2 . g 262.
Vol. 9,No. 6, June 1970 TRINITROTRIAMMINECOBALT
(I I I) 1499

The nature of the impurities in each of these prep- broadening, the first pmr measurements were made
arations is consistent with its structural assignment. with trinitrotris(methylamine)cobalt(III) where both
I n all of the five methods examined, only one isomer of amine protons and the methyl protons should be
Co(NH3)3(N02)3 was produced. Preparations which observed, but, more importantly, the methyl protons
lead to the facial isomer contained only the cis form should be more easily discernible. This complex was
of the dinitrotetraammine while the two methods of first prepared by Chernyaev, et U Z . , ~ and
~ reported by
preparing the meridial isomer contained both the them to be the meridial isomer based on solubility
cis- and trans-dinitrotetraammine ions. The conditions data, refractive indices of aqueous solutions, and X-ray
of synthesis of ~ U ~ - C O ( N H ~ ) ~are ( N almost
O ~ ) ~ iden- powder diagrams which were similar to the mer-
tical with those of the cis-dinitro complex: solu- trinitrotriamminecobalt(II1) .26 Following their pro-
tions are heated above 60" after oxidation while the cedure for the methylamine complex, a pure product,
complex forms. In the preparation of the meridial containing only one geometrical isomer was obtained.
isomer, as in the synthesis of the trans-dinitro complex, The pmr spectrum of the complex dissolved in DMSO
formation of the complex is from solutions a t or near is shown in Figure 2 . Solvent bands have been
room temper a ture.
It was not possible to identify the geometry of the
tetranitrodiamminecobalt(II1) ion in these prepara-
tions. Where the anionic complex was observed, it
was assumed to be part of the ionization isomer, whose
solubility in water is as low as that of the trinitro-
triammine. Only in the Werner preparation did the
ionization isomer appear in large quantities as an
impurity. The most important criteria in the forma-
tion of the trinitrotriamminecobalt(II1) complex and
one that is most important for a reasonable yield is that
&--..-
Figure 2.-The
4.08

I
3.b2

io0 ti2
I
1.63 1.52

BO-MHz proton spectrum of nzer-Co(CHJVH2)3-

the molar ratios of ammonia and ammonium ion to WOz )a.
nitrite be equal. The only impurity then is the ioniza-
tion isomer. Indeed, Shibata, et al., l8 demonstrated eliminated from this tracing. The peaks from the
that complexes from nitropentaamminecobalt(II1) to methyl protons show up sharply as a pair of over-
hexanitrocobaltate(II1) could be prepared from tris- lapping triplets a t 6 1.52 and 1.63 ppm in the expected
(carbonato)cobaltate(III)ion by the same experimental area ratio of 1:2 . The amine protons appear as two
procedure wherein only the concentrations of ammonia bands a t 6 3.72 and 4.08 in the ratio of 1:2 , clearly
to nitrite were varied. establishing this complex as the meridial isomer.
The facial isomer is much less stable than the meridial
f ac.Co(NH 3) (NO,)
isomer, decomposing in solution or in the solid to the
ionization isomer. In DMSO solution, decomposition
was essentially complete in 3-4 hr, while in the solid,
complete decomposition required about 2-3 weeks.
A 6.70

The meridial isomer showed only a slight tendency t o

decompose in solution and was indefinitely stable in the
solid state. Majumdar, Duval, and LeComptel6 may
well have prepared both isomers, but by the time
they measured the infrared spectra the conversion to
the ionization isomer could have been completed,
thereby accounting for Maddock and Todesco's
disputation.
11. Identification of the Geometrical Isomers. A.
Proton Magnetic Resonance Spectra.-Aside from X-
ray crystallographic analysis which was not attempted
in this work, the proton magnetic resonance spectra SO Hz

provide the most definitive method of differentiating Figure 3.-The 60-MHz proton spectra of Co(NHy)s(NOz)a
between the two isomers. isomers.
The facial isomer has three equivalent ammine
groups, whose proton interactions should show up as a Figure 3 shows the pmr spectra of the facial and
single peak, while the meridial isomer contains two meridial isomers of C O ( N H ~ ) B ( N O ~ The
) ~ . initial
ammines trans to one another and a third ammine that assignment of their geometrical configurations had been
is trans to a nitro group. In this latter complex, two based on their Iif values from thin layer chromatog-
peaks whose area ratio is 2 : 1 should be observed. raphy. This assignment was confirmed by the pmr
Because quadrapole coupling of the ammine protons spectra. The facial isomer shows a single, broad peak
with the nitrogen nucleus causes considerable line (26) A. Ganiev, Z h . Neouaan. Khim., '7, 261 (1962).
1500 ROBERT F. DRUDING
B. HAGELAND LEONARD Inorganic Chemistry

a t 6.7 ppm while the meridial isomer has two bands at purity, the ionization isomer ( C O ( N H ~ ) ~ ( N O(Co- ,)~)
5.91 and 6.67 ppm whose area ratio is 1: 2. (NH3)2(NO2)J is yellow and has similar extinction
B. Thin Layer Chromatography.-Based on ob- coefficients in the visible region. Use of the d-d
servations of other cobalt(II1)-amine transition in the visible and near-ultraviolet region has
using acidic solvents on silica gel adsorbants, Rf been quite effective in differentiating between cis-
values for the complexes were observed to increase as and trans-MA4B2, and should be equally effective for
the charge on the complex decreased. Further, i t was identifying f h c - and mer-MA3B3. If one considers
possible to distinguish between cis and trans isomers of the dipoles along the three coordinate axes, the f a c
the type Co4Bzf from their relative Rfvalues. Since isomer has cubic symmetry while the mer isomer has
the cis isomer has a charge polarization that the trans rhombic ~ y m m e t r y . ~ From
’ the splitting of the energy
isomer lacks, i t is more strongly adsorbed on the silica levels of the 3d electrons of cobalt(III), two bands
gel. By this argument, the facial isomer of the neutral (AI,+ TI,, AI, +. TI,) would be expected for fac-Co-
Z ) ~ have a lower Rfthan
complex C O ( N H ~ ) ~ ( N Oshould (NH3)3(NOz)3while six would be expected for the mer
the meridial isomer because i t has a higher dipole isomer where the TI, and TZglevels are further split.
moment. It was this assumption, coupled with the In mer-Co(NH3)3(N02)3 the splitting has not been
X-ray crystallographic analysis of the Jorgensen prep- sufficient to generate separate bands and indeed the
aration, that led to the original assignment of configura- shoulders of the band a t 350 mp are only slightly more
tion of these complexes. Comparing Rf values in pronounced then the broadening of the same band
Table I, the meridial isomers of the various complexes of the fac isomer. This splitting effect could easily be
have quite similar Rfvalues. Insufficient samples of obscured by impurities such as the ionization isomer
known facial complexes were available for comparison and dinitrotetraammine or nitrotriammine complexes,
of these Rfvalues. Crayton and MatternZ1assigned all of which have absorption maxima in this region
the meridial structure to Co(dien) (NO& by its simi- The intensity of the d-d absorption band has also
larity of its uv-visible absorption spectra to that of been used to distinguish between geometrical isomers
Co(NH3)(NOz)3,a method which we found to be un- where cis-MA4B2 has a higher extinction coefficient for
reliable. Nevertheless, the thin layer chromatograms the transition than do the trans isomers. In the case
of this complex resulted in Rfvalues compatible with of Co(NH3)3(N02)3, while the less symmetrical form
other meridial complexes. I n one or two fresh prep- does have the high extinction coefficient, there are
arations of this complex, a second faint spot could be MA3B3 complexes where this is not ~ b s e r v e d . ~Re- ,~
seen at Rf0.6-0.7 which may have been the facial cently, Demas, Turner, and Crosby2*used the charge-
isomer of the diethylenetriamine complex. transfer band in the ultraviolet region to identify the
C. Electronic Absorption Spectra.-The visible and cis isomer of Ru(py)iClZ, where py is pyridine. They
ultraviolet spectra of the facial and meridial isomers proposed that there is a relationship between the energy
of Co(NH3)3(N02)3are presented in Figure 4. Both of the electronic absorption spectrum and the Kosower
Z values29for the solvents used. Compounds with a
permanent dipole moment should show shifts that are
linear with their Kosower 2 values while those without
a permanent dipole should have their charge-transfer
bands unaffected. In the case of the ruthenium com-
plexes of Demas, et al., the cis isomer has a permanent
dipole while the trans isomer does not.
This type of measurement was attempted for the
fac- and m e r - C 0 ( N H ~ ) ~ ( N 0where
~ ) 3 both complexes
have permanent dipole moments30 but the dipole mo-
ment of the former would be expected to be the greater.
Because the nature of the solvent interaction with
either the ground state or the excited state in the
charge-transfer energy change has not been established,
Figure 4.--Visible and uv spectra of Co(NH8)3(N02)3 isomers:
-, mer; - - -, fac. The solvent is HzO. the effect of solvent change on the location of the
charge-transfer band could not be predicted. Most
complexes are yellow in the solid state or in solution. (27) (a) J. Lewis a n d 12. G. Wilkins, “ R l u d e r n Coordination Chemistl-y,”
Both isomers have a similar spectrum in the visiblc Interscience Publishers, S e w I‘ork, N. Y . , 1060, p 188; (b) C. .T. XallhRusen
and C. K . Jdrgensen, Kgl. D a z s k e Videnskab. Selskab, A4at. F y s . M ~ l r d d . ,29,
region, although the nonsymmetrical appearance of the KO. 14 (1955).
absorption band a t 350 mp suggests that it might (28) J. h-.Demas, T. F. Turner, and G. A . Crosbg, I n o r g . Chem., 8, 674
represent several overlapping bands. From the simi- (1969).
(29) E. hl. Kosower, J . A m . Chein. Soc., 80, 3253 (1958).
larity of molar extinction Coefficients for these two (30) Measurement of t h e dipole moments of these comglcxcs by Lhc di-
electric constant method was not attempted These complexes were not
complexes over most of the visible and ultraviolet soluble in solvents of low dielectric strength where their differences in dipole
regions, the inadvisability of using these spectral moments would be enhanced. I n such solvents as 1)MSO and water, their
solubilities are small enough so t h a t error in measurement or t h e presence of
methods of differentiating between the two Complexes trace amounts of ionic impurities could easily exceed the expected differences
may be understood. Additionally, a common im- in dielectric constant.
Vol. 9, No. 6, June 1970 TRINITROTRIAMMINECOBALT
(111) 1501

likely, the facial isomer with its larger dipole, should TABLE IV
show a greater shift in its band location. A second INFRARED OF S O M E COBALT(II1) TRINITROAMIXES~
ABSORPTIONS
limitation to this study was the limited number of
--CO(NH~)~(NOZ)P--- mer-Co(CHa- inrr-Co(dien)-
solvents for these complexes. Two of the best solvents, fac mer NHn)a(NOz)a (NOz)
acetone and DMSO, are opaque in the region of the 1640m 1650m 1605 m
charge-transfer band. Nevertheless, the spectra of the 1565 w 1585 m
two isomers were measured in methanol and water, 1500 w
1486 vw 1470 w
whose Kosower 2 values are 83.6 and 94.6, respectively, 1426s 1432s 1421 s 1432 s
for the meridial isomer; the band a t 255 mp showed no v ( EO2 )a 1401 vw 1408 w 1399 s 1395 m
change between methanol and water, while the facial V ( NOS)a 1368m 1370vw 1368 m 1368 m
isomer band a t 258 mp in methanol appeared a t 250 a(NOn), 1330s 1332 s 1321 s 1325 s
mp in water. This seems to indicate that the dipole [U(nTOS- 11 1295 vw 1302 w
12990s 1270s 1259 w 1262 m
moment of the meridial isomer is quite small.
1223 m 1145 m
D. Infrared Absorption Spectra.-The infrared 1105 m
1073 s 1070 w
1030 w 1050 w
1009 v w 947 w
910 vw
880 w
860 w
WOds 825m 832s 826 s 838 s
732 vw 752 w
710 w
665 w 629 w
625 w 615 w
6(KOZ)I" 610m 605m 604 m
586m 565 vw 590 vw 585 w
Figure 5 , I r spectra of Co(NH~)s(N02)8isomers: --, mer; 560 w
- - -,fac. 520w 508w 535 w
v(CO-N(NH8)) 485 vw 490 vw 500 w 508 w
spectra of the two isomers are presented in Figure 5 and 472 w 450 w
summarized in Table IV. Nakagawa and Shiman- a Abbreviations: s, strong; m, medium; w, weak; vw, very
ouchi3' studied a number of cobalt (111) nitroammines, weak.
comparing calculated and experimental values of their
IV are the spectra of mer-Co(CH3NH.J3(N02)3 and
vibrational energy bands, including mer-Co(NH~)3-
mer-Co(dien) (NO2) to show the similarities of the
(N02)~. The spectrum observed in this work for the various meridial complexes. Each of these is char-
meridial isomer agrees with that of Nakagawa and
acterized by a sharp band a t 1368 cm-' which is absent
Shimanouchi to within .t7 cm-l. As the instrument
in fa~-Co(NH3)3(NO2)3. Also, the triammine and
used for these measurements is estimated to have an
tris-methylamine complexes exhibit similar nitro
accuracy of . t 5 cm-', this agreement is considered
quite good. Of particular interest is the comparison of
these results with those of Postollec and Mathieu.I6
Agreement with their reported meridial isomer is good,
although in the region 1300-1500 cm-l their results
are 20-30 cm-l lower. Even accounting for this %T
difference, agreement with their reported spectrum of
the facial isomer is not as good. In the region 400-
600 cm-', the present results show a greater number
and more distinctive peaks. -0-
600 500 400 600 500 400 Cm'
As might be expected from symmetry arguments, the
me r.Co(NH3) 3(NOZ13 fac.Co(NH3) 31NOJ
more symmetrical facial isomer has a slightly less
complex spectrum than the meridial isomer. As was 100
observed for the cobalt (111) tetraammines, an increase
in symmetry is also accompanied by a shift in absorp-
tion bands to higher wave numbers. This can be best
seen in the 6(N0& band which is a t 610 cm-l for %T 50
me~-Co(NH3)~(N0~)3 and a t 625 cm-' for the f a c
complex. The 6(N02), band shows a shift from 825
cm-' for mer to 832 cm-' for fac. The approximate
modes of vibration in Table IV are those assigned by 0
Nakagawa and Shimanouchi. Also included in Table cm"
(31) I. Nakagawa and T. Shimanouchi, Spectuochim. Acta, P8A, 2099
(1967). Figure 6.-Ir spectra of Co(II1) compounds.
1502 WILLIAME. H L i m E L DA N D EDWIN
R. JONES, JR. Inorganic Cizernistry

stretching frequencies in the 1200-1500-~m-~region tion of the OZv model for the complex. One must
and 6(N02), at 825 cm-'. therefore expect a lower order of symmetry for either
Figure 6 compares the infrared spectra of the four or both isomers of Co(NH3)3(N02)3and should not
complexes in the 400-600-~m-~region as measured on rely as heavily on infrared data for the purpose of
disks of pure complex. The three meridial isomers identification as one does for the cis-trans isomers.
show a very similar pattern of bands, with especially
strong doublets a t 600 and 500 cm-l due to 6(NO&v Acknowledgment.--This research was supported
and v (Co-N (NH3)),respectively. by a grant from the Rutgers University Research
As previously noted by Nakagawa and Shimanouchi, Council. The authors wish to thank Dr. B. 2.Senkow-
the values of the observed and calculated spectra of ski for performing carbon-hydrogen analyses and Dr.
mer-Co(NH3)3(K02)3,while in fair agreement, are not R . \Y.Kluiber for assistance in making the pmr mea-
exactly the same because of the oversimplified assump- surements.

COSTRIBUTION
FROM THE DEPARTVENT
OF CHEMISTKY,
USIVERSITY O F XORTH CAROLISA, CHAPEL HILL,NORTH CAROLIN-4 27314, AND
SAVASNAH
RIVERLABORATORY,
E. I . DU PONTDE R S CONPASY,AIKEN, SOUTH
I ~ E M O U ASD CAROLINA29801

The Magnetic Properties of Hexaamminecobalt(II1) Pentachlorocuprate(I1)

BY WILLIAM E. HATFIELD AND EDWIK R. JOSES, JR.'

Received Novembev 13, 1969

The magnetic susceptibility of hexaamminecobalt(II1) pentachlorocuprate(I1) has been measured from 2.4 t o 38'IC. The
+
susceptibility obeys the Curie-T?Teiss law, x = C/( T e ) ,in the temperature range 16-38'K, giving an effective magnetic
moment of 1.95 BM with 8 = 17'K. There is an antiferromagnetic transition near 8'K attributed to an intermolecular
cation-anion-anion-cation mechanism.

Introduction brated with a sphere of pure nickel metal. Correction was made
for the diamagnetism of the sample holder assembly.
In the first report of the magnetism of the penta-
chlorocuprate(I1) ion, Mori2 noted that susceptibility Results
data for [Co(NH&] [CuC16J obeyed the Curie-Weiss The temperature variation of the inverse magnetic
law from 4.2 to about 2"K, giving a magnetic moment susceptibility of [ C O ( N H , )[CuCl;]
~] from 2.4 to 38°K
of 1.47 BM. Our subsequent studies3 in the tem- is shown in Figure 1. The molar susceptibilities were
perature range 77-300°K yielded a magnetic moment calculated from the experimental observations after
of 1.85 BM. Because the 2,41' ground electronic correction for (a) the diamagnetism (234 X
state3s4 of the trigonal-bipyramidal CuClj3- ion is cgsu) of the constituent atoms using Pascal's constants
not expected to show a decrease in moment a t low and (b) the temperature-independent paramagnetisln
temperature, the magnetic properties of [CO("&]- of the hexaamminecobalt ion ( i d X loe6c ~ s u ) .For ~
[CuCli,] in the intervening temperature range need temperatures above 16°K the susceptibility exhibits a
to be determined. Here we report the results of Curie-X'eiss behavior with an effective moment of
measurements from 2.4 to 38°K. 1.96 B M and a paramagnetic Curie temperature of
Experimental Section 17°K. A single smooth curve fits the magnetic mo-
Preparation of I C O ( N H ~[CuCl,]
) ~ ] .-Large, well-formed, or- ments determined between 2.4 and 38°K by the vi-
ange crystals of hexaamminecobalt(II1) pentachlorocuprate(I1) brating sample magnetometer and the magnetic mo-
were deposited from an aqueous solution (200i d ) of [Co(~I%)G]- ments determined between 77 and 300°K by the Fara-
Cla ( 2 g ) and CuC12.2H20(10 g). Anal. Calcd for [Co(XH3)8]- day method.
[CuCl,]: C1,44.24; P.;, 20.86; €1,4.50. Found: C1,44.23; N, The differences noted in Figure 1 between runs 1
20.79; H, 4.47.
Magnetic Susceptibility Determinations.-The magnetic and 2 are an indication of the precision and accuracy
properties of a powdered sample of [Co(riHs)~] [CUCIS](0.1617'7 of the magnetic measurements. Thus, the magnetic
g ) in a plastic sample holder were determined with a Foner-type moment calculated from the slope of either set of
vibrating sample magnetometer5 built by Princeton Applied magnetic data is more reliable than a moment cal-
Research Corp. Measurements were made a t magnetic field culated from a line fitted to the data obtained in both
strengths of 2000 and 10,000 Oe. The magnetometer was cali-
(6) T h e susceptibility of [Co(NH?)h]Clgwas taken Lo be -87.7 X l W C t
(1) ORAU Research Participant. cgsu: G . Foex, "Constants Selectionnees Diamagnetisme e t Paramag-
(2) M. Mori, B d .Chem. Soc. J a p a n , 34, 454 (1961). netisme," Masson a n d Co., Paris, 1957, p 43. T h e diamagnetic susceptibil-
(3) W. E. Hatfield and T. S. Piper, II&O?'W. Chem., 3, 841 (1964). i t y for [Co(NHa)i]CIz calculated from Pascal's constants is - 163 X 10-6
(4) W. E. Hatfield, H . D . Bedon, and S. M. H o m e r , i b i d . , 4, 1180 (1965). cgsu, and t h e difference between the two was taken t o be TIP for Co(SHa)c8'
( 5 ) S. Foner, Rea. Sci. ImLi,,,30, 548 (1959). ion.