QUADRUPOLE INTERACTION OF CO 5o IN COBALTIC COMPLEXES

BY V. SARASWATI AND R. VlJAYARAGHAVAN

(Tata Institute of Fundaraental Research, BOmbay-5, India)

Received March 27, 1967

(Communicated by Prof. B. V. Tho*ar, r.A.sc.)

ABSTRACT

Quadrupole coupling constants of Co s9 has been measured in several octahedral cobaltie complexes. The compounds are diamagnetic and polycrystalline in nature. The magnitudes of the coupling constants were evaluated assmning ,~ as zero. The temperature variation of the coupling constants are correlated with the nature of Co-ligand bond and molecular motions in the solid.

INTRODUCTION

Co 59 has a nuclear spin 7/2 a n d a high quadrupole moment Q equal to 0" 5 barn. The quadrupole interaction would be appreciable compared to Zeeman interaction with the external magnetic field and the magnetic reso- nance lŸ will be broadened beyond observation, unless the nucleus is in a near cubic environment. This condition is satisfied for Co 59 in the co- ordination compounds. Tl~e transition metal atoms, because of their unfilled d orbitals, have a tendency to form eomplexes in which the metal atom is in the octahedral surroundings of the ligands. Trivalent cobalt complexes are diamagnetic. However, large chemical shifts occur due to second order paramagnetism. 1 Quadrupole coupling constants of Co 59 in a few octahedral cobaltic complexes are reported in this paper. The results have been corre- lated with the nature of cobalt-ligand bond and molecular motions in these compounds. A brief account of some of the results have been presented earlier. ~

CALCULATION OF THE QUADRUPOLE COUPLING CONSTANT

The quadrupole coupling constant [e~qQ)/h] can be determined accurately either by performing a pure quadrupole resonance experiment a or by studying the nuclear magnetic resonance in a single crystal. 4 The former method is applied when e2qQ is large enough and the latter when the quadrupole energy

253

254 V. SARASWATI Alfil) R. VIJAYARAGHAVAN

is smaU and could be treatc'd a sa perturbation on the Zeeman energy. Co z+ in octahedral surroundings is expected to have a smaU coupling constant. Hence we shall concern ourselves only with the situation where quadrupole interaction acts a s a perturbation on the nuclear magnetic interaction. For ah axially symmetric field gradient (i .e. , ~/-= 0), with quadrupole perturbation carried to second order, the transition frequencies approp¡ to a nucleus of spin I in a single crystal specimen is 5,e

vm.~m-~ = Yo q- �89 Yo (,3/*= - - l ) ( m - - � 89

--k ~ (1 --/,2) {[102m (m -- 1) -- 18I (I q- 1) q- 39]/,2

- - [6m (m - - 1) - - 2I (I + 1) + 3]} (1)

where

3e=qQ . vo -- 21(21_ l) h ' t z = c ~ and 0

is the angle between the z-axis in the principal axis system of the field gradient tensor and the external magnetic field. 1,o is the Larmor frequency or the transition frequency in the absenr of the electric quadrupole interacfion. From equation (1) it is clear that when quadrupole perturbation is only up to first order ( i .e . , first order in yo) the central transition m = �89 = - �89 (for half integer spins) is unshifted from the Larmor frequency yo ('." m - - �89 = 0). When the second order term is also considered the central transition shifts from v0 and the shift is inversely proportional to v 0 or the external magnetic field. In a polycrystalline sample all values for q are permissible. The intensity maxima for the resonance absorption lineshape occur at/z = 0 for quadrupole perturbation up to 1st order. The absorption lineshape consists of (for half integer spins) an unshifted centre line flanked by small maxima separated by

�89 vQ (2m - 1). (2)

When seeond order perturbation is significant the averaged expression for ir possesses for the central transition (�89 ~ - �89 two maxima corresponding to q = 0 and q 5/9. These occur at

v~ (�89 --+ -- �89 = 1,o + < [I (I + 1) -- �88 (3)

~ ~ ~ ~ - ~ - ~ ~ = ~o- ; { . ~, + , ~ - ~ J . ~~~

Quadrupole Interaction o f Co sa in Cobaltic Complexes 255

Thus the absorption shape consists of two peaks asymmetrically about v0 and the peak-to-peak separation Av would be,

25 r o, A v = v~ -- v~ ----- 1-T~vo [1 (I q- 1) -- ~].

situated

(5)

Subsfituting for I, which is 7/2 for CoSa:

1 (e~---QQ) ~ Av = 75"25v£ (6)

It should be noted that in the second order only the transition �89 ~ - ~- is observed, the others, e.g., 4- ~ ~ 4- ~, 4- J ~ 4-' ], 4- ] --+ -4- �89 are un- observable by tkis method.

In the magnetic resonance experiment, one records the derivative of the resonance line. The frequency separation between the derivative maxima, referred to as �91 differs from the absorption peak separafion Av by 2cr, the dipolar width. Ir has been shown by Casabella 7 that cr a n d a more accurate (e~ could be obtained by measuring �91 at several fields. The plot of �91 vs. 1/v o will be a straight line for Av/2<r > 2, and is given by

1.02B �91 - - + % (7)

where

25 B --- f44 ~o' [I (I + 1) - k].

The slope of the curve (7) gives the coupling constant accuratcly and the intercept gives the dipolar contribufion.

When ~ is not equal to zero, the peak-to-peak separation in the absorption line shape is modified from equation (5), Av is given by, a

24 yo, [ 3 ] [ 74 ~~] A ~ = i ~ 4 ~ ~ I ( I + 0 - - 1 + i ~ 3 ~ + 4 3 . (8)

Reccntly, the detailed lineshape for the powder pattern of �89 ,-+ -- �89 transition for ~ ~ 0 has been worked out by Narita et aL a

Another method, a more accurate one, would be to measure the pure quadrupole resonancc frequencies for the d:: 7/2 -+4- 5/2, d: 5/2 --~--4- 3[2,

256 V. SARASWATI AND R. VI/AYARAGHAVAN

4-3/2---~',4-1/2 transitions. For non-axial field gradient case these occur respectively at lo

v ( 7 "-+~)=--]-4-h-3e2qQ ( 1 ~72 487 74 + ) 10 27000 " " "

v(5 __~.~)= 2e2qQ (1 17o 5020774 -y07"+27-~0 +...) (3 1) e2qQ( 1097~ v "-'2 =q~h- 1 +~6- 196093 %

7 ~ + . } 27000 "" ~, " (9)

In principle these transitions are observable in quadrupole resonance using nuclear inductionY x As will be shown below the coupling constants were small and this method could not be used.

E~~ERLMENTAL RESULTS

The compounds (listed in Tablc 1) were prcparcd in this laboratory and investigations were made only in polycrystalline compounds. Cobalt forros numerous complexes having different types o~ ligands as donors. A few compounds having the most symmetrical environment around cobalt atom have been chosen for this study. The environment in the compounds chosen is schematically illustrated in Fig. 1.

A VarŸ wide line spectrometer with variable temperature accessories has been utilised. Measurements were made at different field strengths ranging from 3 to 12 k-gauss and at varied temperatures in favourable cases. The chemical shifts in these compounds have been studied by Dharmatti et aL 1~ and the data were utŸ to fix the frequency of Co 59 resonancr Ah asymmetric line due to �89 - + - - �89 transition was observed in all the com- pounds. The linewidth (i.e., derivative peak-to-peak separation) decreased with increasing field indicating the presence of second order quadrupole interactions. The peak separation in frequency �91 was plotted against 1~yo and from this curve (e2qQ)/h and dipolar contribution to width/lave been evaluated. The values are given in Table I. The resonances were extremely weak for ' trans' [Co (NH3)4(No2)~] No, Co (NH3)3 (7N02)8, Na3 [Co (Nos)s] and Co (dipy)3 (CIO4)3. Mean of a few traces at 12 kg. was used to calculate the coupling constants (e2qQ),/h. For Co(en)3 C13.3HaO coupling constant obtained on putting �91 for Av in equation (6) compares well with the value obtained from �91 vs. 1/v o plot. This is of course understandable f of in these compounds the dipole interactions contnbute much less to the width compared to quadrupo/ar interaction. It has been assumed throughout that 7 is zero.

Quadrupole Interaction of CM' in Coba~tic Complexes 257

In trans Co (NH3) ~ (NO~)2NO3 and Co (]~,IH3) 3 (~q'O2) 3 the lineshape indŸ that ~ may not be zero. As we have not been able to determine ~7, the coupling constant has been calculated for ~ = 0.

eH3

~ r ~ ? ~ ~ / H3N eH 3 H3N - eH 3

I/, c o / \1 % , , /~j.,,,.~ - . . % /

NH 3 3"l- (i~) 'TRAES' Co(NH3)4X2 3+ ( i ) co(eH3) 6

HtC -,-~- ' -" lCHŸ ~J\/,-.~ I'~1~ -"-"---" C o - '-"-"~ N %

.2c / / \ ,c.~

(i i i) Co(en)3 3+

H3 C - C C - C H~

c O

0 C o - -~- , - - - - 0 / \ O 0

~ N H

"'~- ?l.H" N H 3

(iv) 'Ci•' Co(NH3)4X2 $'1"

~ Co - - - - - - - . . N

, / \ . (v) Co ACETYLACETONATE (vi) Co (d ipy ) ] 3+

FI6. 1. Schematic diagram of the Cobaltic Complexes.

In the hexammines the high field side oŸ the resonance spectrum shows a few extra humps indicative of non-equivalent sites for cobalt. These

A$

258 V. SARASWATI AND R. "VIJAYARAGHAVAN

TABLE I

Quadrupole coupling co/tstants and the chemical shifts o f Co s9 in coba# complexes

No. Compound Chemicat shifts at 10 K. gauss

(in liquid)

Coupling constant (e2qQ)/h in Mc/s.

(in sofids)*

1 KaCo (CN)6 . . . . . . 0 (Reference) 2.74-0.1

2 Co (dipy)3(C104)s . . . . --66"2 11"2

3 Co (NHs)8(NO2)3 . . . . --69"4(s) ~ 3.6 --71.0 (w)

4 Co (en)a CI~ . . . . . . --70.1 3.34-0.1 ~

5 " Trans' [Co (NH3)4 (NO~)~] NOs.. --71.5 5"66

6 NazCo (NO~)e . . . . --73.5 8.4

7 Co (NH~)e Cla . . . . . . --80.8 1"54-0.1'

8 Co (NH~),(NOs)z . . . . --80.8 1'84-0"1

9 Cobalt Oxinate . . . . . . --94"2 4.3

10 Cobalt acetyl acetonate . . . . --122.9 5.74-0.2" . . . . f,

(a) ~/assumed to be zero,

(b) In liquid samples two linos, one woak and the other stroAg wcro observed.

(c) At 100 ~ K. (#gQ)lh = 3.8 Mc/s.

(d) Weak line hr error is bound to be large. Lineshape indicat�9 that ~ may -ot be zero as assumed.

(e) Only the dispersion mode could be observed.

humps become prominent at low fields. The width of the centre line and separation between one pair of outer humps at 6 Mc./s. correspond to the coupling constants 1.2 and 2.2 Mc/s. respectively. As the field (or yo) is increased/X v.vo also increases instead of rernaining constant. This shows overlap and the coupling constant calculated from the width of centre line at 7.5 McŸ is found to be 1"5 Mc/s. At this frequency the first order satellites were observed. The coupling constant obtained by using equation (2) is 1" 1 Mc/s. for Co (NH3)8 Cia.

Quadrupole lnteraction of Co ~ in Cobaltic Complexes 259

In pentammines, e.g., [Co (NH3)5 CI] Cl” cobalt is in the environmcnt of tire ammine ligands and one chlorine ligand. The octahedron around cobalt would be distorted and cationic reorientation is absent as observed from proton resonances. ~ A large field gradient would be present at the cobalt site owing to the asymmetry of charge distribution. For these reasons no Co 5~ resonance cou]d be traccd in any of the pentammines. In "trans" [Co (NH3)4 (NO2)2] NO3, there is ah axial symmetry around cobalt which is lacking in the ' Cis' isomer. The cationic motion was observed in the former and it was fotmd to be absent in the latter. This explains the failure in record- ing the nuclear magnetic resonance of Co s~.

0)

FI~. 2. Lineshape of Co s9 resonance in (i) C-~(NHs)s (NOs)s and (ii) Co(r In Tra~ leo (NI-Is)~(NO2)z ] NO8 the resonance shal~ is like (i), and in the others lik�9 (ii).

As mentioned in Section 2 if one could trace the quadrupole resonance of at least two transitions v could be determined. The induction method could be used for obtaining the quadrupole resonance line. The minimum radio frequency available in the spectrometer used by us was 2 Mc/s. Hence this method could be used (vide equation 9) only for the 7/2 --~-5/2 transition of Co (dipy)3 (C10~)3. A broad line was observed at 2.38 Mc/s. which con- firms the coupling constant obtained from the powder sample. ~ could not

260 V. SARASWATI AND R. VIJAYARAGHAVAN

be determined as ir is not possible to observe other transitions. Search for the quadrupole resonance of Co 59 in pentammines was not successful.

Temperature dependence measurements were carried out in hexammines, Co(en)8 C1~.3H30, K3Co(CN)e and cobalt acetyl accetonate. In the last two the coupling constant remained practically unchanged whereas in the former ones large change was noticed and is shown in Fig. 3.

~5

A

V

4"0

:Ÿ

3-0 O

2"0

20

100 150 200 2~0 ] 0 0 Tr'~ ~00 150 200 250 ~,00 T~K

�9 ~ ,~ 4"0

_ ~ 3-5

~oo 150 ZOO 250 500 T~K ~ 100

FIo. 3. Cormlation between tho tcmperaturr dependence of sccond moment and quadrupolo coupling oonstants. (a), (b). Sccond momgnt and (e2qQ)/h against T in Co (NHs), (NO~. (e), (d). Second momcnt, aud (e~qQ)/h against T in Co (en h CI, 3HsO.

DISCUSS~ON

The faaors that appreciably affect the quadrupole coupling constant ate 3 the nature of bond associated with the nucleus concemed and molecular motions in the solid.

The electronegafivity difference relates to the covalency of cobalt-ligand bond, The bonding atoms of the groups NHs, ethylene diamine (en) and

Quadrupole lnteraction of Co ~9 in Cobaltie Complexes 261

NO~- have almost the same electronegativity, z3 We fiaad that coupling con- stants at liquid-air temperature for Co (en)3Cl~, Co (NH3)6 (NO3)3 and the room temperature value of Co (NH3)3 (NO2)3 are nearly the same, 3.8 Mc/s. For Co (NH3)6 Cls the coupling constant at 120 ~ K. is 2 Me]s. but it increases on lowering the temperature showia~g that it is not the rigid lattice value. In 'trans' [Co(NH3)4(NO2)~] ]qO3 the coupling constant of 5"6 Mc/s. is comparatively high. This could be due to the asymmetry introduced by replacing two of the NH3 grotlps of the hexammine by NOs groups. The deviation flora octahcdral symmctry could hence be more. This is a]so brought out by the fact that whereas in the former compounds ~ appears to be small from the lineshape, in the latter compound 7/appears to be large (Fig. 2). The ]ineshape is grossly different for ~ close to zero and close to one and it can, therefore, be easily distinguished. Oxygen, nitrogen and carbon ate in the decreasing order of electronegativity, z3 When these atoms belong- ing to a ligand group ate bonded to cobalt atoms the coupling constant shou]d also be in this order if the distortion e~ects are equal or sma|l. As expected in cobalt acetyl acetonate, where cobalt is linked to six oxygen atoms, the coupling constant is 5.6 :J: 0.2 Mc/s. and in KsCO(CN)n w¡ the donor atom is carbon it is 2"7 -4-0.1 M c/s. In the compounds with nitrogen, as donor, the coupling constants ate in between these two values. A correlation thus seems to exist between electronegativity difference and the value of coupling constants when all the bonding ligands ate alike and distortion effects negligible. This requires confirmation with more compounds having octahedral carbon and oxygen surroundings.

Molecular motions in Nt-I 4 is found to have a significant effect on the coupling constant of bromine in NH,Br? 4 A similar situafion occurs in the cobalfic complexes. The coupling constant Ÿ found to increase in Co (NI-I3), CI3 at lower temperatures when the cationic mofion freezes. The corresponding increase is much higher in Co (N]-I3) * (NOn)~. In Co(en)8 Cl3.3HaO the motion of water molecules above 170~ could account for the decrease in the coupling constant above this temperature. Figure 3, which is the plot of proton resonance second moment and the quadrupole coupling constant against temperature,shows evidently the correIation between protonic mofions and the coupling constant decrease. The proton reso- nance width in cobalt acetyl acetonate is independent of temperature, the coupling constant is also found to be so. Hence it could be concluded that molecular motions such as the cafionic mofion in the hexammine or the hindered rotation in water molecule reduce the coupling constant of cobalt.

As mentioned in these compounds the chemical shifis ate large (Table I) and ate directly related to the strength of the r field due to the

262 V. SARASWATI AND R. VlJAYARAGHAVAN

ligand. The quadrupole coupling constant depends on the anisotropy of the electric field around cobalt nucleus whereas the chemical shift depends on the magnitude of the crystalline fie]d. Therefore a correlation between the two is not expected.

The only work available on quadrupole coupfing constant of Co 5~ is that of Sugawara, a5 who has studied Co ('NHz)e Clz in the powder forro a n d a twin crystal of K3Co (CN)e at 4 ~ K. The coupling constant obtained in the former is 1.5 Mc/s. assuming ~/ to be zero. This compares favourably with our values. In Kz CO (C.N), , (e2qQ)/h = 7.3 �91 0.1 Mc/s. with ~7 = 0.75. Both these values are incompatible with our results. In polycrystalline K3Co (CN)8 we find that (e~qQ)/h = 2" 7 + 0.1 Mc/s. is temperature indepen- dent up to 100 ~ K. and ~ from the lineshape is < 0-5. The large value obtained by Sugawara at 4 ~ K. could then only be due to some kind of a transi- tion between 4~ and 100 ~ K.

ACKNOWLEDGEMENT

It is with pleasure we thank Mr. S. Jagannathan for the preparation and analysis of the compounds.

1. Proctor, W. G. and Yu, F .C .

2. Dharmatti, S. S., Saraswati, V. and Vijayaraghavan, R.

3. Das, T. P. and Hahn, E. L.

4. Volkoff, G.H. . .

5. Cohen, H. H. and Reif, F.

6. Jones, W. H., Graham, T. P. and Barnes, R. G.

7. Casabella, P .A. . .

10. Jeffrey, G. A. and Sakurai, T.

REFERENCES

Phys. Rev., 1951, 81, 20.

"Nuclear magnetic resonance and relaxation in soilds," Proceedings of the XIII Colloque Ampere Lenven, 1964, p. 133.

Solid State Physics Supplement, 1958, 1, 1.

Can. J. Phys., 1955, 31, 820.

Solid State Physics, 1957, 5, 321.

Phys. Rey., 1963, 132, 1898.

J. Chem. Phys., 1964, 40, 149.

8. Silver, A. H. and Bray, P.J. Ibid., 1958, 29, 984.

9. NarRa, K., Uneda, J. and Ibid., 1966, 44, 2719. Kusonoto, H.

Progress in Soild State Chemistry, 1964, 1.

Quadrupole Interaction of Co 59 in Cobaltie Complexes

11. Smith, G . W . . .

12. Dharmatti, S. S., Kanekar, C. R. and Mathur, S. C.

13. Pauling, L. . .

14. Itoh, J. and Yamagata . .

15. Sugawara, T. . .

263

Phys. Rey., 1966, 149, 346.

Proceedings of Symposium on the Chemistry of Co-ordination Compounds, Agra, 1959, p. 157.

Nature of Chemical Bond, Cornell University Press, Ithaca, 1948.

J. Chem. Phys., 1956, :24, 62J.

J. Phys. Soe. Japan, 1959, 14, 859.