THE UNIVERSI TY OF MI CHIGAN COLLEGE OF ENGINEERING Department of Nuclear Engineering Technical Report VANADIUM SOLID-STATE CHEMISTRY BY PARAMAGNETIC RESONANCE C. Kikuchi N. Mahootian ORA Project 06029 under contract with: U.S. ARMY MATERIEL COMMAND HARRY DIAMOND LABORATORIES CONTRACT NO. DA-49-186-AMC-80(x) WASHINGTON, D.C. administered through: OFFICE OF RESEARCH ADMINISTRATION ANN ARBOR July 1964

TABLE OF CONTENTS Page ABSTRACT v LIST OF TABLES vii LIST OF FIGURES ix I. INTRODUCTION 1 II. GENERAL PRINCIPLES 2 III. VANADIUM IN TUTTON SALTS 9 IV. ANGULAR DEPENDENCE OF CaO:'VAND MgO:V2 FINE STRUCTURE 16 V. VANADIUM IN MgO POWDER 21 VI. VANADIUM IN A1203 POWDER 24 REFERENCES 25 iii

ABSTRACT A brief review of the electron spin resonance properties of vanadium in crystalline solids, such as in sapphire, Tutton salt, and magnesium oxide, is given. v

Table I. II. LIST OF TABLES Zn++ - 0- BOND DISTANCES IN TUTTON SALT COMPARISON BETWEEN THE CRYSTAL AND THE POWDER DATA. Page 12 15 vii

LIST OF FIGURES Figure Page 1. ESR spectrum of V4+ in sapphire. 3 2. ESR spectrum of V3+ in sapphire.4 3. Angular dependence of ESR spectrum of V3+ in sapphire. 5 4. ESR spectrum for V2+, in green sapphire at k-band. 7 5. Energy levels of V2+, Cr3+, and Mn4+ 8 6. ESR spectra of V02' and V2+ in Tutton salt. 11 7. Simplified schematic of the Zn++ site in Tutton salt. 10 8. Vanadyl pentahydrate.17 12 9. Vanadyl ion in substitutional site in Tutton salt. 13 10. ESR spectra of randomly oriented VO2+ (x-band). 14 11. ESR spectra of V2+ in CaO. 17 12. Energy levels of V2+ in MgO. 19 13. ESR spectra of vanadium in MgO powder samples. 22 ix

I. INTRODUCTION In an earlier report, it was suggested that the electron spin resonance (ESR) technique can be developed into an analytical tool for the study of the changes in oxidation states brought about by high energy radiations. As the ionizing radiations proceed through the crystalline lattice, some of the energy is inevitably absorbed, resulting in the release of electrons through the photoelectric and Compton processes, or even by the pair production process, if the incident gamma radiation is sufficiently energetic. The electrons thus freed will for a short period of time move about in the solid and may eventually become trapped at metastable centers, resulting in ions with rather unusual valence states. Or possibly, the incident radiation may result in the breaking of electron bonds, again leading valence states not ordinarly found. Such changes in solids, however, can be expected to be small, so that the detection of the effects may fall outside range of standard chemical techniques. The purpose of this report is to indicate how ESR can be used to identify traces of chemical species produced by high energy radiations, and to indicate the progress that has been made in our laboratory subsequent to the suggestion. Also, as indicated in our earlier report, vanadium is particularly well suited for the study of electron transfer and trapping processes in solids. The reason for this stems from the fact that vanadium is essentially isotopically pure V51 (v51, 99.7%; V50, 0.25%) whose nuclear spin is 7/2 so that the ESR spectrum will have the characteristic 8 line structure. This hfs can then be used to identify vanadium. Furthermore, as the valence changes, the electron spin of the ion will change, and the electron spin of the vanadium in question can be ascertained from the angular dependence of the fine Structure spectra. Combined with the ease of identification, there is the additional property that the valence state of vanadium can be changed easily. Thus it becomes possible to study the properties of the different valence states in the same host lattice. In undertaking the present report, we wish to thank Dr. R. Borcherts and Mr. H. Azarbayejani for technical assistance, and for the financial support provided by the Reynolds Aluminum Company through the Michigan Memorial Phoenix Project. 1

II. GENERAL PRINCIPLES We shall first discuss vanadium sapphire in order to illustrate the general principles. As indicated above, the nuclear spin of V51 is so that the ESR spectrum, regardless of the valence state will consist of 8 lines, of about equal intensity and spacings. The spacing between the components, known as the hyperfine coupling constant will vary from crystal to crystal, and from one valence state to another. In general, however, the spearations fall in the range from about 70 to 150 gauss. What distinguishes one vanadium valence state from another is the number of such groups of 8 lines and the angular dependence of the groups. The electron configuration and the associated electron spin of the vanadium paramagnetic species that have been observed in sapphire are as follows; Ion Electron Configuration Electron Spin V4+ 5d 1/2 v3' 3d2 1 v2+ 3d3 3/2 Consider first V4+. The electron spin is 1/2, so that ESR absorption spectrums should consist of a single group of 8 lines. The ESR spectrum of this vanadium state in sapphire is shown in Figure 1. The lines are very weak. The spectrum seems to be isotropic with g = 1.97 and A = B = 1.52 x 10-4 cm-1. This spectrum is observed in vanadium sapphire, prepared from a powder mixture containing 0.1 vanadium pentoxide. Consider next V3+. Vanadium in sapphire is considered to be in A.13+ substitutional sites, so that V3+ in corundum was first observed by Zverev and Prokhorov,2 and their ESR results constitute a confirmation of the deductions from magnetic susceptibility measurements by Siegert3 and by van der Handel and Siegert.4 Subsequently, further confirmatory optical investigations have been reported by Pryce and Runciman.5 Theory and experiment indicate that the V3+ ground state, whose spin is 1, consists of two levels, one for M = 0 and the other for M = 1 is studied. The ESR spectrum at X-consists of a single group of 8 lines, with lines very nearly equally spaced, with separation of about 114 gauss when the magnetic field is along the crystal C-axis. The spectrum, furthermore, is observed at 4.2~K, but not at 78~K. Also at liquid helium II temperatures, the intensities of lines are reduced, stemming from the depopulation of the magnetic M = 1 state. A typical spectrum and its angular dependence are shown in Figures 2 and 3. An interesting fact is that the hfs. components are about 110 gauss apart, or about 1.02 x 10-4 cm-1. This value is smaller than that for V4. Information on g- and B are still lacking. 2

V 90O 3000 K 4000 gauss 2980 gauss Figure 1 - ESR spectrum of V+ in sapphire.

V e=o0 T= 4. 2~K 1440 gauss -- 114g 2240 gauss -p Figure 2. ESR spectrum of V3+ in sapphire.

-8000 7000 Position Of Center Of Hyperfine Group 0-Experiment -— Approximate Theory — 6000 Ho = Ho(00) X cos 0 u cos 9 O s^^fc 0 — 5000 o 4000 — 3000 0 — 2000 0 15 30 45 60 75 90 e Figure 3. Angular dependence of ESR spectrum of V3+ in sapphire. 5

The ESR spectrum for V2+ can be produced by subjecting sapphires containing vanadium to ionizing radiations. Figure 4 gives the spectrum obtained from the so-called green sapphire, which contains a small amount of vanadium and cobalt. This spectrum was taken at K-band frequencies. There are three groups of 8 lines, indicating that the electron spin of the center is 3/2, and the characteristic 8 line structure shows that it is due to vanadium. The strong lines accompanying the vanadium lines are due to Cr3+ which is isoelectronic with V2+. The relative positions of the V2+ and Cr3+ lines indicate that the zero-field splitting, 2D, for V2+ is somewhat smaller than that of Cr3+. Our measurements indicate that 2D(V2+) = 9.8 Kmc/sec., and is to be compared to 2D(Cr3+) = 11.5 Kmc/sec. Also g = 1.98. For Cr3+, g = 1.97. A question raised by this study is the origin of the electrons responsible for the conversion of V3+ to V. A study of the growth of the intensity of V2+ lines relative to V4+ seem to preclude the possibility that an electron is transferred from one vanadium ion to another. Furthermore, the saturation intensity of the V2+ spectrum seems to vary from sample to sample. However, to date, we have not been able to determine the factors that contribute to the stability of the V2+ and/or V4+ ion. The practical reason motivating our investigations is that the stabilization of specific valence state is important in a program for the development of new materials for solid-state devices. For vanadium sapphire, V2+ is isoelectronic with Cr3+, so that a properly prepared vanadium sapphire could be a material useful for masers and lasers, particularly in the presence of ionizing radiations. Figure 5 summarizes the present known information6'7 about isoelectronic ions V2+, Cr3+ and Mn4+ in A1203. 6

Figure 4. ESR spectrum foxr'2+ in green sapphire at k-band Fige,4. ESR spectra -~

2+ v 3+ Cr 4+ Mn -I 11,600 Cm " 12.3 / / / / / / I1,679 / /!- ~~31 29 18,000 Cm/'/ / 14,400 / / / / I I 21,000 Cm I / ~I~ / I / / / I / I/.39 / / 4,800 80 Figure 5. Energy levels of V2+, Cr3+, and Mn4+

III. VANADIUM IN TUTTON SALTS After a few attempts to study vanadium chemistry in sapphire by purchasing single crystals of sapphires..at considerable cost... it became apparent that a direct attack on the problem would not yield useful results. Consequently, our attention was turned to vanadium in Tutton salts, because such crystals can be grown in our laboratory without difficulty. The vanadyl ion, O can be readily introduced into the zinc ammonium Tutton salt, and furthermore, our preliminary studies showed that the vanadyl vanadium can be converted to vanadous vanadiumo It should be noted that randomly oriented VO++ has been studied by a number of investigators. Among the firs- to make a study of VO++ in frozen aqueous solution was Kozyrev.8 Investigation of VO++ radicals on various absorbers was carried out by Faber and Rogers;9 vanadyl porphyrins by Roberts and Koski;10 vanadyl etioporphrins dissolved in benzene and high viscosity oil by O'Reilly;ll and vanadyl ions in aqueous, acetone, and ether solutions by Pake and Sands.12'13 Much of the interest in the vanadyl ion appears to have been stimulated by oil companies who are concerned with the relatively high concentration of vanadyl vanadium in crude oil. Aside from being important from the standpoint of the technology of refinery operation, there appears to be some geologic interest, because the presence of vanadium in oil casts some light on the nature of organisms involved in the production of petroleum.14 Despite this relative wealth of studies of randomly oriented VO++ in solutions, it appears that very little study has been made of oriented VO+. Brief comments on tetravalent vanadium occur in the paper by Hutchison and Singer,15 and some unpublished investigations by W. B. Gager16 exist. Consequently, we have undertaken a systematic study of VO++ radicals in the Tutton salts, to determine whether vanadium occurs as VO++ or V44 in these crystals, and to study the relation of V++ produced by ionizing radiation to that grown directly into the crystal. Our studies reveal that VO++ maintain its identity in these crystals, and that following irradiation, the vanadyl vanadium is converted to V++ that is indistinguishable from those grown directly into the crystal lattice. The basis for this remark is that the vanadous vanadium, V2+ produced by irradiation from VO++ in the Tutton salt, has within experimental error the same spin - Hamiltonian parameters and the same orientation of crystalline electric field as those of V++ Tutton salt grown from aqueous solution. While our investigations were in progress, the paper by Ballhausen and Gray17 appeared. We feel that our experimental results constitute a confirmation of the validity of their theoretical calculations. 9

The top two traces in Figure 6 give the ESR spectra of a single Tutton salt crystal containing a small concentration of VO2+ radical before and after x-irradiation. The VO2+ spectrum before irradiation is actually quite complex, due to the fact that there are two non-equivalent sites in the unit cell and to the observed three orientations of the VO2+ axis. However, if the crystal is properly oriented, the spectrum simplifies to the 8 lines shown at the top. The second trace shows the effect of x-irradiation. New groups of 8 lines occur, characteristic of vanadium, and by inspection it is seen that there are three such groups, suggesting that V2+ is produced by X-rays. the bottom trace is included to show that the V+ ions produced by X-rays are identical to those obtained from the VO2+ doped Tutton salt by electrolytic reduction. A further check is obtained by growing a VO2+ crystal using as a seed a Tutton salt crystal containing V2+. The V2+ spectrum produced by X-rays is found to coincide exactly with the V2+ spectrum present before x-irradiation. To interpret these experimental results, let us consider the structure of the site occupied by VO2+ or V2+. The crystal structure of Tutton salt is quite complex. However, for our purposes, we need to consider only the water molecules surrounding the Zn++ sites. X-ray diffraction studies indicate that each Zn++ site is surrounded by 6 ligand water molecules, with the oxygens lying closest to the Zn++ ions, as shown in Figure 7. 07 2.138 A" 90 -08 00 Zn 80 09 07 Figure 7. Simplified schematic of the Zn++ site in Tutton salt. According to the recent X-ray diffraction analysis by H. Montgomery, the bond distances and angles are as given in Table I. 10

X - IRRADIATED UNIRRADIATED ZnSO4(NH4)2S04 6H20: V++,2% Figure 6. ESR spectra of VO2+ and V2+ in Tutton salt.

TABLE I Zn++ - 0- BOND DISTANCES IN TUTTON SALT Bond Distance e i Zn-07 2.138 50.2 79.5 Zn-08 2.117 49 -71.5 Zn-09 2.066 65 4 The antle 8 is the polar angle measured from the crystal b-axis, and ~ is the azimuthal angle measured in the crystal ac-plane from the a-axis. Suppose now that a V2+ ion is substituted for Zn2+ ion, without any distortion. The V2+ ion will be subjected to the crystalline electric field due to the 6 ligand water molecules. Both the hydrogens and oxygens will contribute to this field, but clearly the major contribution will come from the oxygens, because they lie closest to the V2+ ion. And further more, of oxygens, we can expect the nearest oxygen, which is 09, to make the greatest contribution. Consequently, it is reasonable to suppose that the direction of the crystalline electric field to be more or less along 09. In fact measurements show that the angles for the V2+ crystalline electric field are (690, 0.6~), very nearly equal to the polar angles for 09 (65,40). Consider next the effects of V02+ radical substitution. Here it seems most appropriate to suppose that the structure is similar to that found in vanadyl pentahydrate, Figure 8. The vanadyl vanadium is assumed to occupy 02H2 0 OH2 1.67 A 2.3 A V4+ 2.3 A \I H2 0 {2 —- H2 Figure 8. Vanadyl pentahydrate.17 12

the Zn2+ site, and the vanadyl oxygen is assumed to lie along the directions of one of the water oxygens. This then suggests that three directions for the V02+ axis are possible, in agreement with the observed experimental re-; sults. Furthermore, it is reasonable to suppose that the V02+ axis will be in the direction of water molecule vacancy. Thus we can expect the most preferred direction to be along the most loosely bound water molecule, associated with 07. Since the bond distances to 07, 08, and 09 are 2.138, 2.117, and 2.066 A, so that the 07 water molecule most weakly bound and the 09 water molecule is most strongly bound. We can then infer that the 07 direction is the most preferred orientation for the VO2+ axis, whereas the 09 direction is the least preferred. Measurement of the relative intensities indicate that the probabilities for the three orientations are in the ratio of 20:5:1. Another evidence supporting this model is provided by the close agreement between the experimental g-values and the values calculated by Ballhausen and Gray, as shown in Figure 9. O O o 0 1 o 0\ V / V. V 10'\ 1 I \\ 1\ 0 0 2.138A0 O 0 gz = 1.9537 calc. gz = 1.9538 calc. gz = 1.9541 calc. ggz = 1.9331 Exp. gz = 1.9316 Exp. gz = 1.9299 Exp. Figure 9. Vanadyl ion in substitutional site in Tutton salt. Another fact is -the excellent agreement of the single crystal and powder measurements. Thus, for the VO++ radical, sufficiently reliable results can be obtained from powder measurements. In Figure 10-a, we give the spectrum VO++ absorbed on Amberlite IR-4B, vanadium in glass is given in Figure 10-b, and vanadium in crushed Tutton salt in Figure 10-c. The spectrum shown in Figure 10-b indicates that vanadium is present as vanadyl vanadium in glass. The spectrum of V0++ in aluminum oxide powder is quite similar to those indicated above. A partial summary of our results is presented in Table II. 13

(a) VO' ON9 AMBERLITE O.S o V__, IN (b) 0 Nt. 3 Sto _ GLASS i-I 1.0% VO IN (c) ZnNH4)_2S. 6H*H.O (crushed crystal) Figure 10. ESR spectra of randomly oriented V02+ (x-band).

TABLE II COMPARISON BETWEEN THE CRYSTAL AND THE POWDER DATA Calculated (Ballhausen and Gray) Orientation Powder gll 1.9238 + 2 gl 1.9802 ~ 3 g 1.9644 ~ 3 1.9314 ~ 2 1.9812 + 3 1.9646 2 1.935 1.940 1.980 1.983 1.965 1.969 (A) (B) 0.01824 ~ 2. 182 0.007162 ~ 5.00729 Several interesting facts stand out. One is that for the two orientations, the individual g-values differ appreciably, Yet, the average g-values are equal, within experimental error. This perhaps is to be expected if the crystalline field provides the major contribution of Ag. For the two sites, the crystalline fields have different orientations with respect to the VO++ but the trace of the g-tensor should be the same, if the ligand bonding effects do not change. 15

IV. ANGULAR DEPENDENCE OF CaO:V2+ AND MgO:V2+ FINE STRUCTURE Recently we have observed another interesting characteristic of V2 in CaO and in MgO. The spectra at four different angles for CaO:V2+ are presented in Figure 11. We note that the V++ spectra show an angular dependence very much like that of Mn4+ At 0~, where the Mn4+ fine structure splitting is maximum, that of the V++ is also the largest. This angular dependence of the V++ spectrum is very surprising, because ions with S = 3/2 in a cubic field should show only an isotropic splitting coming from the second order terms in the electron-nuclear interaction. A similar anomalous angular dependence was observed during the course of our investigation of vanadium in MgO single crystals. The small angular dependence of the vanadous vanadium ESR spectrum in MgO appears to have been first noticed by Low who attributed the effect to a small rhombohedral distortion. Ham et al1l have observed a similar angular dependence in the spectrum of Co++ in CdTe. The spin Hamiltonian for s = 5/2 ions in a cubic crystalline site is usually written as te = gP S.H + A S.I - gNPN I.H (1) Bleaney19 and Koster and Statz20 have indicated that the observed unexped angular dependence can be explained in some cases by adding to the above spin Hamiltonian a correction term AU, Eq. (2) which accounts for higher order interactions of the Zeeman term P(L.H+2S.H), the spin-orbit coupling term \L.S, and the hyperfine interaction term PL.I with the energy levels of the ion in the crystalline field. A,2 = upSx HxHy + S Hz ( - (S)[S(S+l)-l] (2) + U S3 Ix + Sy3 y + Sz3 Iz - (SI)[5S(S+l)-l where x, y, and z are the cubic axes of the crystal and u and U are constants depending on g. Using the corrected spin Hamiltonian cib + A< we find energy levels of the ion to second order in A and first order in u and U to be

' ~ I].~, I'j~ f-~ z.'i.,LJL^.|,,:,|'1 ~......... f>._...... i.......'.... _. I. -—'i ~ - ~ - > ".'~. -.?.-.-y. ^'^ rt. ~*.*. -.-'.'. -: 4;! i -. i' I H -L......... i —-.-....,!'i -'P-.. /!-'. i' 1 -i. —O.... i i I, i ~0 I" ]",;... F I..1 I.I I 2 i i a I-I I i i i II i I L..i I i I i.. i s ~ I, ~, >"i.. [, I,;,h: I i'- I i.:.. -..-. i........... I. i.._i... i i'"T'-r'-........-... i....:., 1"-" -..-..^.....-l'_" ~, I I.^~t.:4 —-..... 4. 4. 0 CJ.r4 CH O Cd t C) p-i rr rl (D.r-... - 1... I. I I. i. I > r.I I i. i,. i I. I I f.I: oo 00 I i ~ ~ u m?5 ~D3Z 17

A2 EM,m = go HM + AmM + [mM2-mM+MI(I+l)-mS(S+l)] 2gPH0 (3) + (uH+Um){M3 1 [5S(S+l)-l]M}p where A = hyperfine coupling constant, m, M = eigenvalues of I and S along the applied magnetic field H respectively, I = nuclear angular momentum of the ion, S = spin angular momentum of the ion, Ho = hv/gP, and p = 1 - 5(nl2n22+n22n32+n32nl2) with nI, n2, and n3 being the direction cosines of the magnetic field with respect to the three cubic axes of the crystal. In the case of the V2+ ion I = 7/2 and S = 5/2 and positions of the resonance lines in gauss are found from Eq. (5). For (m, M = 3/2) + (m, M = 1/2) transitions, A. A[2. 5 ( m A2m 6 Upm (4) - [H A K - 2(gB - 4 )o (g-)2HO 5 g For (m, M = 1/2) - (m, M = - 1/2) transitions, and for (m, M = - 1/2) -> (m, M = - 3/2) transitions Hy = K LHO A-2 3 _ A m 6 Upm] (6) 2(gB)2Ho Q (4) 5 g2

where k = (1 + 5/6 up/g) 1 As we observe, the effect of A)fon position of the absorption lines depend on magnitude of u and U. In our case u and U were experimentally determined to be u - 3 x 10-5 p10-6 -1 U 106 cmi we found that these values can account only for a small part of the observed angular dependence of the ESR spectrum of the V2+ ion in MgO and in CaO. In the case of Co+2 ion in CdTe, investigated by Ham18 and coworkers, the constants u and U are much larger than those in our case and hence the observed anomalous angular dependence could be explained adequately by the correction term A A theoretical estimate to the values of u and U can be made using 4th order perturbation calculations, to include interactions of the Zeeman, the spin-orbit coupling, and the hyperfine interaction term with the energy levels of the ion, Figure 12. (3) r4 4F3 3/2 6Dq e__._ (3) 12 Dq rS rP (I) Figure 12. Energy levels of V2+ in MgO. 19

The result is 120X3 ul i = E 22 E42 52 42 (7) U1 = Pui (8) where ul and U1 are a part of contributions to u and U respectively E42- (E)r4 - (E)r2, and Es2 (E)r5 - (E)r2 and P 0.0225. The values of E42 and E52 as given by Sturge7 are E42 = 19,900 cmES2 = 13,200 cm-1 and the value of k can be approximately taken equal to that of the hexahydrated vanadous ion,21 \ =_ 44 cm'1 Substituting k, Es2 and E42 with their values, we find from Eqs. (7) and (8), uL 3 x 10-6 U1 - 0.7 x 10-7 cm-1 which are a lower limit estimate to the experimental values cited earlier. 20

V. VANADIUM IN MgO POWDER Next a study of vanadium in MgO powder samples was undertaken with the hope that a recipe for the stabilization of vanadous vanadium could be worked out. The second objective to this program was to develop the necessary chemical techniques to measure the nuclear spin and magnetic moment of V49. The nuclear parameters for this nuclide, which decays to Ti4 under electron capture 330 day half-life, were first reported by Weiss, Walter, Gilliam, and Cohen.22 However, the observations were made in a xylene solution of a vanadium (IV)-Cupferron Chelate, the individual vanadium Hfs. lines were quite broad, so that there is still some doubt on the reported values. On the other hand in MgO powder, the hfs lines are sharp, so that a precision value for the nuclear magnetic moment can be expected. The MgO powder samples containing vanadium are prepared by wetting MgO powder with a solution of VOS04 so as to obtain about 0.1% for the V/Mg ratio. The samples are then dried, fired at a suitable temperature, and then irradiated with 50 Kvp X-rays. The dried samples often exhibit a spectrum suggestive of VO, but this disappears completely upon firing at 900~C or higher. If, however, the sample is irradiated subsequently with X-rays a spectrum is observed suggesting that vanadium is present as 2n and/or V+, as shown in Figure 13a. A typical spectrum obtained from a sample heated at temperatures of about 12000C and then X-rayed as shown in Figure 15b. There are the characteristic 8 lines corresponding to I = 7/2 for the nuclear spin of V51. In addition, the principal hfs lines are flanked by two satellites, suggesting that the electron spin of the paramagnetic center is 3/2. In fact, very good agreement between theory and experiment is obtained by assuming that the satellite splitting is due to second order effects in the electron-nuclear interaction term. Thus, there is every reason to believe that the spectrum presented here is due to vanadous vanadium, V2+. The presence of satellites in the spectrum of the powder samples is not readily expected because, as we have seen earlier, both the intensity and position of satellites in ESR of V2I in a single MgO crystal are angular dependent. In fact it can be shown that we can have powder samples containing vanadous vanadium with an ESR spectra showing no distinguishable satellites. In order to show this, a relatively large MgO:V2+ crystal was cut in small and average size crystals and following observations were made at x-band. (a) Each small crystal showed a weak V2+ spectrum. (b) Two small crystals, oriented in the same direction and glued 21

(a) 5891.9 Gauss 6453.8 R) ro V4 V+t (b) J a —-l~~~~~~~~~~~~~~~~~~~~~~~~~ Figure 13. ESR spectra of vanadium in MgO powder samples.

together, produced stronger ESR signal with more distinguishable satellites. (c) Ten small crystals, including the ones used in (a) and (b), were placed randomly in a quartz probe. The ESR spectrum of this sample consists of a set of 8 main lines without any satellites. The observations (a), (b), and (c) demonstrate clearly that the ESR spectra of the MgO:Vanadium powder samples which produce no satellites cannot be necessarily attributed to V4+ ion. In fact they may result from randomly oriented crystallites of MgO:V2+' Unlike the samples prepared at 900~C, the powder samples fired at 1200~C are corse and "lumpy'. Thus, it is assumed that among the factors that contribute to the appearance of the satellites in powder samples, the degree of coarseness of the powder may be playing an important part. In order to show the connection between the size of the powder grains and the observed satellites, the powder samples and a small steel rod were placed in a quartz probe. The probe was then flushed with helium gas and sealed. By shaking the probe we were able to crush the powder grains, in an oxygen-free atmosphere. We observed that as the coarse grains were gradually ground to finer powder, the satellites became smaller until they finally dissappeared. The ultra-violet excitation techniques were used as an alternative method of identify the vanadium ions in the powder samples. From our preliminary work it was found that U-V emission spectrum of the powder samples, heat-treated at 1200~C (whether they produce satellites or not) are somehow similar to that of the single MgO:V2+ crystal. More works in this area are scheduled. The direct chemical preparation of MgO:V2+ powder samples were intended to remove some of the existing ambiguity in associating the observed spectra with V2+ and V4+. It was found that samples heated at 1250~C in a stream of carbon monoxide or hydrogen for about 10 hours and cooled to room temperature under a steady stream of these gases show strong ESR spectra before xirradiation. These spectra, which remain qualitatively unaffected by a subsequent x-irradiation, are identical with the ones obtained by x-irradiation. It is known7,23 that heating at elevated temperatures in a reducing atmosphere converts higher valancies of vanadium to V2+ and V3+ states. Since the latter has no ESR spectrum at room temperature, the observed spectrum of these powder samples is attributed to V2+ ions. 23

VI. VANADIUM IN A.120 POWDER Although our investigations on A1203 powder samples are not complete yet, we shall present some of our preliminary results. The preparation of powder samples here is similar to that of MgO:Vanadium samples. We have observed that unheated unirradiated samples of aluminum oxide powder with vanadyl sulfate, VOS04, produces the characteristic spectrum of VO++ ion similar to those shown in Figure lOa, b, and c. Upon heating the sample at about 1000~C, the vanadyl spectrum disappears. The heated samples upon x-irradiation in an oxygen-free atmosphere such as C02, N2, or He2 show a spectrum somewhat different from that of VO+. It is tentatively held that this spectrum belongs to V++"

REFERENCES 1. C. Kikuchi, S. Yip, and S. H. Chen, "Application of Spin Resonance Techniques to the Study of Radiation Effects," Michigan Memorial Phoenix Project (June, 1961). 2. G. M. Zverev and Am M. Prokhorov, "Electron Paramagnetic Resonance of the V+++ Ion in Sapphire," Sov. Phys. JETP, 7, 1023 (1958). 3. A. Siegert, "Zur Dentung des Magnetischen Verhaltens der A.laune der Eisengruppe," Physica 4, 138 (1957). 4. J. Van Den Handel and A. Siegert, "Uber das Magnetischen Verhaltens vom Vanadium - Ammonium Alaum," Physica 4, 871 (1957). 5. M.H.L. Pryce and W. A. Runciman, "The Absorption Spectrum of Vanadium Corundum," Disc. Faraday, Soc. 26, 34 (1958). 6. S. Geschwind, P. Kisiluk, Mo Klein, J. P. Remeika, and D. L. Wood, "Sharp-line Fluorescence, Electron Paramagnetic Resonance, and Thermoluminescence of Mn4+ in -A1203," Phys. Rev. 126, 1684 (1962). 7. M. D. Sturge, "Optical Spectrum of Divalent Vanadium in Octahedral Coordination," Phys. Rev. 130, 639 (15 A.pril, 1963). 8. B. M. Kozyrev, Disc., Faraday Society 19, 135 (1951). 9. R. J. Faber and M. T. Rogers, "Paramagnetic Resonance Spectra of Absorbed Manganese," (IIO) Cooper (II) and Oxovanadium (AV)," J. Am. Chem. Soc. 81, 1849 (1959). 10. C. M. Roberts, W. S. Koski, and W. S. Caughey, J. Chem. Phys 29, 118 (1958). 11. D. E. O'Reilly, "Paramagnetic Resonance of Vanadyl Etioparphyrin I," J. Chem. Phys. 29, 118 (1958). 12. G. E. Pake and R. H. Sands, "Hyperfine Structure in the Paramagnetic Resonance of Vanadium Ions in Solution," Phys. Rev. 98, 266A (1955). 15.. H. Sands, "Paramagnetic Resonance Absorption on Glass," Phys. Rev. 99, 1222 (1955). 25

REFERENCES (Concluded) 14. A. J. Saraceno, D. T. Fanale, and N. D. Coggelshall, "An Electron Paramagnetic Resonance Investigation of Vanadium in Petroleum Oils," Analytical Chem. 33, 500 (1961). 15. C. A. Hutchison and L. S. Singer, "Paramagnetic Resonance Absorption in Salts of V and Mn," Phys. Rev. 89, 256 (1953). 16. R. Borcherts, G. Wepfer, and C. Kikuchi, "Paramagnetic Resonance Spectrum of Vanadyl Ammonium Sulfate," Bull. Am. Phys. Soc. 7, 118 (1962). In a private communication, Dr. W. B. Gager of the Battelle Memorial Institute has informed us that he too has made some preliminary unpublished investigations of VO++ in Tutton salts. However, a complete identification of the centers, was not made. 17. C. J. Ballhausen and H. B. Gray, "The Electronic Structure of Vanadyl Ion," J. Inorganic Chemistry 1, 111 (1962). 18. F. S. Ham, G. W. Ludwig, G. D. Watkins, and H. H. Woodbury, "Spin Hamiltonian of CO2+," Phys. Rev. Let. 5, 468 (1960). 19. B. Bleaney, Proc. Phys. Soc. (london) A73, 939 (1959). 20. F. Koster, and H. Statz, Phys. Rev. 115, 445 (1959). 21. J. Owen, "The Coulors and Magnetic Properties of Hydrated Iron Group Salts, and Evidence for Covalent Bonding," Proc. Roy. Soc. A227, 183 (1955). 22. M. M. Weiss, R. I. Walter, 0. R. Gilliam, and V. W. Cohen, "Paramagnetic Resonance Spectrum of V49," Bull. Am. Phys. Soc. 2, 31 (1957). 23. J. E. Wertz, P. Auzins, J.H.E. Griffiths, and J. W. Orton, "Electron Transfers Among Transition Elements in MgO," Faraday Soc. of London, No. 26, 66 (1958). 26

UNIVERSITY OF MICHIGAN 3 9015 03023 1495