Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The investigation of indium halides and graphite intercalation compounds using time-differential perturbed… Dong, Sunny Ronald 1988

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-UBC_1988_A7 D66.pdf [ 3.97MB ]
Metadata
JSON: 831-1.0302355.json
JSON-LD: 831-1.0302355-ld.json
RDF/XML (Pretty): 831-1.0302355-rdf.xml
RDF/JSON: 831-1.0302355-rdf.json
Turtle: 831-1.0302355-turtle.txt
N-Triples: 831-1.0302355-rdf-ntriples.txt
Original Record: 831-1.0302355-source.json
Full Text
831-1.0302355-fulltext.txt
Citation
831-1.0302355.ris

Full Text

THE INVESTIGATION OF INDIUM HALIDES AND GRAPHITE INTERCALATION COMPOUNDS USING TIME-DIFFERENTIAL PERTURBED ANGULAR CORRELATION GAMMA-RAY SPECTROSCOPY By Sunny R o n a l d Dong B A p S c , t h e U n i v e r s i t y o f B r i t i s h C o l u m b i a , 1981 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF APPLIED SCIENCE i n THE FACULTY OF GRADUATE STUDIES DEPARTMENT OF PHYSICS We a c c e p t t h i s t h e s i s as c o n f o r m i n g t o t h e r e q u i r e d s t a n d a r d THE UNIVERSITY OF BRITISH O c t o b e r 1988 © Sunny R o n a l d Dong, COLUMBIA 1988 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of P j-j / S The University of British Columbia Vancouver, Canada Date DE-6 (2/88) ABSTRACT T h i s t h e s i s d i s c u s s e s i n g e n e r a l terms the t h e o r y and a p p l i c a t i o n o f t i m e - d i f f e r e n t i a l p e r t u r b e d a n g u l a r c o r r e l a t i o n gamma-ray s p e c t r o s c o p y (TDPAC) to the s t u d y o f s o l i d s t a t e p h y s i c s . The t e c h n i q u e y i e l d s v a l u a b l e i n f o r m a t i o n on the e l e c t r i c f i e l d g r a d i e n t s e x p e r i e n c e d by r a d i o n u c l i d e s which have been s u b s t i t u t e d f o r n o n r a d i o a c t i v e i s o t o p e s or i n s e r t e d as i m p u r i t i e s i n t o v a r i o u s i n o r g a n i c compounds. The i nd ium h a l i d e s a r e examined i n a s e r i e s o f e x p e r i m e n t s . The f i r s t a p p l i c a t i o n s of t h i s t e c n h i q u e t o t he s t u d y of g r a p h i t e i n t e r c a l a t i o n compounds a r e d i s c u s s e d . i i i TABLE OF CONTENTS Abstract i i Table of Contents i i i Acknowledgements iv Introduction 1 The Nuclear Quadrupole Interaction 8 Foreword to the Publications 26 List of Publications 27 Quadrupole Interaction Studies of Graphite Indium Chlorides Using Time-Differential Perturbed Angular Correlations 28 Quadrupole Interactions in Chlorides of Indium . . . . 48 Quadrupole Interactions in Mixed-Valency States of Indium Chlorides 67 Quadrupole interactions in Graphite-Hafnium Chloride . . 80 Aftereffect investigations in Mixed-Valence Indium Chlorides 103 Concluding Remarks 123 References 126 iv ACKNOWLEDGEMENTS I would l i k e to e x p r e s s my g r a t i t u d e to Dr . P e t e r W. M a r t i n and the l a t e D r . J . G . H o o l e y , w i t h o u t whose f r i e n d s h i p and e x p e r t i s e t h i s work c o u l d not have been a c c o m p l i s h e d . Dr . L l e w e l l y n W i l l i a m s d e s e r v e s c r e d i t f o r h i s c a r e f u l r e v i e w of t h i s t h e s i s and h i s h e l p f u l e d i t o r i a l comments. F i n a l l y , I would l i k e to thank NSERC f o r t h e i r s u p p o r t o f t h i s e n d e a v o u r . - 1 -I. Introduction Nuclear and so l id state research may, on f i r s t consideration, be looked upon as two widely separated f i e ld s of s c i e n t i f i c endeavour. The energies encountered in nuclear reactions are several orders of magnitude greater than those found in chemical binding. Consequently, experiments in the nuclear regime deal primari ly with " f r ee " or unbound atoms, whereas so l id state experiments are performed for the purpose of measuring interatomic interact ions. In addit ion, few nuclear properties, aside from the mass and the ground state moments, are of interest to a so l id state phys ic i s t . However, there exist certa in experimental techniques which draw upon aspects of both d i s c i p l i ne s . Pr inc ipa l among these are nuclear magnetic resonance spectroscopy (NMR), nuclear quadrupole resonance spectroscopy (NQR), Mossbauer spectroscopy, and perturbed angular correlat ions (PAC). The l a t te r two techniques provide information on the interactions of a radioactive probe atom with i t s e lectronic environment. The interpretat ion of the experimental results requires an extensive knowledge of the nuclear properties of spec i f i c excited states of a given radionucl ide. In the former two resonance techniques, the probe atoms are i n i t i a l l y in their nuclear ground states. An r f f i e l d is applied which interacts , respect ively, with the nuclear magnetic moment ar i s ing from the nuclear spin, I, and the nuclear quadrupole moment ar i s ing from a nonspherical nuclear charge d i s t r ibut ion . The response to this applied f i e l d is modified by the e lectronic environment. - 2 -This thesis deals with the last technique mentioned (PAC). However, in order to give the reader a f u l l e r appreciation of the usefulness of this part icu lar form of spectroscopy, a l l four techniques w i l l be examined with a view towards pointing out s im i l a r i t i e s and di f ferences, advantages and disadvantages. The centra l theme of this thesis i s the appl icat ion of time-d i f f e r e n t i a l perturbed angular correlat ions (TDPAC) to research in so l id state physics (1,2). The usefulness of this technique is best i l l u s t r a ted by c i t i ng some examples. An area of so l id state physics in which TDPAC has been used extensively i s the study of hyperfine interactions in metals. The technique i s not sensit ive to e l e c t r i c monopole interactions so that only nuclear quadrupole interact ions need be considered. The nuclear quadrupole moment of the probe couples to the l oca l e l ec t r i c f i e l d gradient and a TDPAC measurement w i l l y ie ld information about the e l ec t r i c f i e l d gradient tensor at the probe s i t e . From Laplace's equation, the efg i s only nonzero at metal s ites which do not possess cubic symmetry. A systematic study of the noncubic metals has appeared in the l i t e ra ture ( 3 ) i n recent years. The data support the hypothesis that the efg arises from two strongly correlated sources, the l a t t i ce ions and the conduction electrons. Also, the efg was found to vary with temperature as T 3 / 2 . A typ ica l example of an experiment performed on a noncubic metal i s that which was done using l l l - C d m as a probe in Ga. NMR measurements (4) on single c rys ta l samples of Ga revealed an asymmetric efg with components that varied with temperature. TDPAC - 3 -measurements on single crysta ls of Ga (5,6,7) also showed th is temperature dependence.' Since l a t t i c e vibrations are believed to produce the decrease in efg with increasing temperature, the data points to an anisotropy in the phonon spectrum of Ga. A second appl icat ion of the TDPAC technique has been in the study of jump processes in sol ids (8). For the noncubic metals, the efg that i s measured i s t ime-invariant. However, in jump processes, the efg is time-dependent, with either i t s or ientat ion and/or strength varying randomly in time. While i t i s no longer possible to obtain information about the efg, relaxation phenomena can s t i l l be studied. A case in point i s that of (NH4) 3HfF ? doped with 181-Ta. One of the F atoms i s thought to be capable of jumping among 8 equivalent avai lable s i tes in the l a t t i c e . This high mobil ity of the F atoms generates a time-dependent efg at the Ta/Hf s i t e . Below -50°C a s ta t i c efg was found to preva i l , but at greater temperatures a smooth t rans i t ion to a f luctuat ing efg was observed. A jumping rate for the F atom was deduced which agreed quite well with NMR results obtained with 19-F in the isomorphous Zr compound. The last example i s that of TaS 2 and i t s in terca lat ion compounds (9,10). TaS2 i s one of the greater than 40 t rans i t ion metal dichalcogenides which have a layered structure. Each Ta layer is sandwiched between two S layers. Only van der Waals bonding occurs between the S-Ta-S sandwiches. The stacking of S-Ta-S layers in these - 4 -sandwiches determines whether the symmetry of the Ta s i te is octahedral (1T-) or tr igonal prismatic (2H-). The nuclear probe used in the TDPAC experiments was 181-Ta. Measurements on lT-TaS 2 demonstrated the existence of a charge density wave which undergoes a commensurate-incommensurate phase t rans i t ion as the temperature of the sample is raised from less than 180°K to 350°K (11). A charge density wave d i s torts the crys ta l l a t t i c e and i f the wave vector of the charge density wave is a multiple of a rec iproca l l a t t i ce vector, i t i s said to be commensurate. Otherwise, i t i s incommensurate. The commensurate wave corresponds to an enlarged l a t t i c e unit c e l l . In the case of lT-TaS 2 the redefined unit c e l l contains two nonequivalent Ta atoms, whereas for the incommensurate wave the t rans lat iona l symmetry of the l a t t i ce is lost and the Ta atoms are subjected to efgs of varying strengths and or ientat ions. Thus, at temperatures below 180°K two s tat ic efgs were observed for the commensurate phase, and above 350°K a single inhomogeneously broadened efg was found for the incommensurate phase. The weak van der Waals bonding between the S-Ta-S sandwiches permits the insert ion or interca lat ion of various chemical species between the sandwiches. A further requirement for in terca la t ion to occur is charge transfer to or from the host material. This apparently occurs most readi ly with 2H-TaS 2(as opposed to the lT-phase), whose e lectronic properties are metal - l ike. The most s t r ik ing trend observed - 5 -upon interca lat ion of this substance with donor type interca lants , such as NH3 and Na, i s the marked decrease in the strength of the efg from that found for the pr i s t ine 2H-TaS2- This decrease is believed to be due to the transfer of a valence electron from the intercalant to the intra layer d-band of the 2H-TaS2. However, the amount of charge transferred has not been determined. The experiments described in this thesis consist of the f i r s t applications of TDPAC to graphite interca lat ion compounds and to the systematic study of the indium chlor ides. Previous PAC work on the indium chlorides has been of the time integra l form. Graphite and some of the t rans i t ion metal dischalcogenides have layered structures and weak inter layer bonding which permits the interca lat ion of various elements and compounds. The graphite interca lat ion compounds (GICs) are distinguished by their high degree of s tructura l ordering. This i s ref lected in a phenomenom cal led staging. It i s present in a l l G ICs , even at very low concentrations of interca lant. The staging phenomenom is characterized by the periodic spacing of the intercalant layers. The number of carbon layers separating each intercalant layer i s denoted by the stage index, n. The stage index i s generally quoted when describing a part icu lar GIC. Mixed stage G ICs have been prepared, but even so-cal led "pure" stage compounds may not be ent i re ly homogeneous. In other word, the structura l ordering i s extensive but not pervasive. The graphite has a free car r ie r concentration of approximately 10 - 1 * free carriers/atom at room temperature. Upon in terca la t ion, - 6 -charge transfer occurs between the intercalant layer and i t s adjacent carbon layers with the result that the free car r ie r concentration in these carbon layers i n greatly enhanced. Since the f i n a l composition of a GIC can be control led, i t i s possible to vary the e l e c t r i c a l , thermal, and magnetic properties of a given GIC in a systematic way. GIC's have a number of prac t i ca l appl icat ions. They are used as battery materials, and as substrates for chemical ca ta lys i s . Certain GIC's such as graphite-AsF 5 , are incorporated into composite wires for e l e c t r i c a l conduction. GIC's prepared from carbon f ibres have been reported to exhibit high tens i le strength as well as unique e l e c t r i c a l properties. However, there have been problems with the reproduc ib i l i ty of these resu l t s . F i na l l y , graphite may form the basis of various storage and recovery systems. Pract i ca l applications aside, the pr inc ipa l sources of the currently strong interest in GIC's are of a purely s c i e n t i f i c nature. Though progress has been made towards a deeper understanding of the mechanism of in terca la t ion , a sat i s factory explanation of the staging phenomenom i s s t i l l forthcoming. The high anisotropoy of GIC's presents the so l id state physic ist with a unique opportunity to study two-dimensional phase t rans i t ions. In addit ion, the intercalant has been observed to undergo a var iety of chemical processes such as polymerization and disproportionation. Research techniques for the study of hyperfine interactions are wel l -suited to the invest igat ion of these matters. Since there are now known to be close to 200 reagents - 7 -which w i l l intercalate into graphite, there is no shortage of interest ing experiments to be performed. While a detai led discussion of G ICs i s quite beyond the scope of this thesis, the 200 page review a r t i c l e by Dresselhaus and Dresselhaus provides an excellent introduction to the subject (12). TDPAC provides information on the f ie lds at the nuclear probe s i t e . Its main role in GIC research w i l l probably l i e in uncovering some of the f ine deta i l s of the structure of the intercalant layer and the processes occuring therein. - 8 -2. The Nuclear Quadrupole Interaction Whereas the nuclear magnetic dipole moment (NMDM) interacts with a magnetic f i e l d which may arise to varying degrees from interna l and external sources, the nuclear quadrupole moment (NQM) interacts with the e l ec t r i c f i e l d gradient (efg) ar i s ing only from interna l sources. No d irect means currently exists by which an external ly applied f i e l d of su f f i c ient strength can be produced to couple with the NQM. Since an external magnetic f i e l d may be c losely monitored and control led, nuclear magnetic dipole moments have been measured to reasonable accuracy. However, the efg at a c ry s ta l l i ne s i te depends sens i t ive ly on the charge d i s t r ibut ion at the s i te and the calculat ion of the efg has proven to be a very d i f f i c u l t problem. A number of methods exist whereby the strength of the nuclear quadrupole interact ion can be obtained, but the theoret ica l d i f f i c u l t i e s make i t hard to separate the nuclear from the c ry s ta l l i ne f i e l d contributions. Fortunately, i t has been possible in certa in instances to measure the NQM i t s e l f , although the accuracy of these determinations does not match those of the NMDM. Thus, in spite of the sens i t i v i t y of the nuclear quadrupole interact ion to c rys ta l l ine f i e l d s , i t i s sometimes hard to interpret the measurements with the precis ion that a physicist might prefer. Nevertheless, valuable information can s t i l l be obtained. There are four main approached for studying the nuclear quadrupole interact ion. These are the NMR and NQR spectroscopies, and the Mossbauer and PAC spectrocopies. A concise discussion of the theory of - 9 -the nuclear quadrupole interaction w i l l be followed by a comparison of these four techniques. The detailed derivations of the formulae can be found i n the references (13,14). 2A Theory (13,14) The distribution of the e l e c t r i c charge i n a nucleus of atomic number Z i s given by i t s nuclear charge density, p(x). If V(x) i s the electrostatic potential at the nucleus arising from external sourses then the electrostatic interaction energy i s given by: where the integral i s over the nuclear volume. This integral can be simplified by expanding the elect r o s t a t i c potential i n a Taylor series about the center of mass of the nucleus. The series converges rapidly since V(x) varies only s l i g h t l y over the nuclear volume. Using (1) H = / p (x) V(x)d 3x Cartesian coordinates (x,=x, x2=y, x3=»z) the expansion of the integral i s : (2) H - J d 3x p(x) (V0+Z ( g - ) x + i I ( The quantities designated by the subscript 0 are evaluated at the nuclear center of mass which i s at x = 0. These are constants which - 10 -can be pulled outside the integra l sign so that (2) becomes: (3) H = Z e V 0 + l?i ( | 1 . ) 0 + I I Qjk ( S 2 V ) + . . . J 3 J » k J k Where: Z g = j d xp(x) = to ta l nuclear charge = J d J xp(x )X j = nuclear e l ec t r i c dipole moment Q^k= J d xp(x )X jX^ = nuclear quadrupole moment tensor Equation (3) i s va l id in both the c l a s s i c a l and the quantum mechanical regimes. However, in the l a t te r case, a l l the quantit ies must be replaced by the appropriate operators. Only a single state of the nucleus needs to be considered, since nuclear states are widely separated in energy and the hybrid izat ion of nuclear states does not have to be taken into account when considering the ground state or the f i r s t few excited states. Some terms may be eliminated from (3) altogether. The f i r s t term is the e lec t ros ta t i c energy of a point nucleus and is invariant with respect to the s ize, shape, and or ientat ion (as characterized by the nuclear spin, I) of the nucleus. Since i t i s the energy differences - 11 -which are measured, this constant term is dropped. The e l e c t r i c dipole term vanished because of symmetry conditions placed on the nuclear charge density. The wave function describing the nucler state has a def in i te par i ty, which means that the nuclear charge density has the property that i t i s invariant under the inversion of the coordinates, This makes the integrand in (5) an odd function and P must vanish. The interact ion energy reduces to: (7) H = £ Q .,V ., + correction terms 1 j ,k J K J K where o 2V 5 E k > (8) V.. - ° v . e l e c t r i c j k dx^ox^ & X j The correction terms are of the order of 10~8 of the quadrupole terms and are usually undetectable. Both Q ^ and are symmetric second rank tensors (3x3 matrices) by de f in i t i on . The quadrupole moment arises from a nonspherical charge d i s t r ibut ion in the nucleus. In quantum mechanical terms, this translates into the condition that the nuclear spin, I must have a value of at least 1. The V., i s the negative of the efg tensor but i t i s often referred to as the efg tensor. At f i r s t glance, there appear to be 12 independent quantit ies that must be evaluated. Fortunately, this number can be reduced to 5 - 12 -by a judicious choice of coordinate systems and the appl icat ion of some symmetry pr inc ip les . Since the quadrupole tensor is symmetric, i t has at most 6 independent terms. If the Q ^ i s redefined to have a zero trace, the number of independent terms can be reduced to 5. This i s eas i ly done. e The subst itut ion of (9) into (7) gives: The second term does not depend on the nuclear orientation (I) and can be ingnored. If the "hat" notation is also dispensed with, (10) becomes: ( I D H - z l Q f l V 6 ^ j k jk A further reduction of the number of terms required to specify the Q . can be achieved by re lat ing the 5 terms of (11) to just one. The jk rationale behind this can be developed i n tu i t i ve l y from a semiclass ical picture of the nucleus. The charges in the nucleus are imagined to be precessing rapidly about the d irect ion of the nuclear spin. Thus, the external charges see a nuclear charge d i s t r ibut ion that i s e f fec t i ve ly - 13 -symmetrical about the z axis or the spin d i rect ion. Consequently, the time-averaged nuclear quadrupole moment tensor contains zero o f f -diagonal elements, and Qn=Q]_2 by symmetry. However, from the zero trace condition i t is apparent that there i s only one independent term, and that: Now consider the efg tensor, V(x) i s the potent ia l due to charges external to the nucleus. Therefore Laplace's equation must hold for This means that V^. has a zero trace and 5 independent terms. However, a symmetric matrix can always be diagonalized by transforming to a special coordinate system. The orthogonal axes of this special system are cal led the pr inc ipa l axes and w i l l be denoted by "bars" overhead. In the pr inc ipa l axes system only 2 parameters are needed to specify the efg tensor. These are defined as follows: (12) Q l l = Q22 =<~ 1 / 2)Q33-V(x) at the nucleus. (13) 7 2V = 0 (14) oz z zz yy where e = proton charge - c^io. y _ i > i > i* XX - 14 -(15) TI = (V__ - V__)/V__, 0<T!<1 xx yy zz T) i s ca l led the asymmetry parameter and i t measures the departure of the efg from c y l i n d r i c a l symmetry. From (12) i t i s clear that the efg vanishes for e l e c t r i c f i e ld s with spherical or cubic symmetry. The other 3 terms in the o r i g ina l tensor (as specif ied by (8)) have not vanished but the information contained in them i s now included in the angles needed to specify the or ientat ion of the pr inc ipa l axes with respect to the crys ta l axes. In po lycrysta l l ine powder samples, only ri, and eq are measured. The angles may be obtained in experiments performed on large single crys ta l s . It should be emphasized, at this point, that the nuclear quadrupole interact ion i s composed of a nuclear contribution, Q , and J k an environmental contr ibut ion, V j^» which are separable. The contains the interest ing chemical and physical information. However, which factors determine the form of the V.. ? The two pr inc ipa l sources of the efg are the ions in the nearest and next-nearest neighbour she l l s , and the electrons in the p a r t i a l l y -f i l l e d atomic shel ls of the atom i t s e l f . The e lectros tat ic contribution of the distant ions is neg l ig ib le, but their d i s tort ion of the e lectronic charge d i s t r ibut ion in the incomplete outer electrons shel ls strongly affects the gradient seen at the nucleus. This effect i s ca l led Sternheimer shie ld ing. When i t enhances the gradient, i t i s ca l led ant i sh ie ld ing, and the contribution of the distant ions can be - 15 -increased from 10 to 100 times. When i t decreases the gradient, i t i s ca l led shielding, though the size of this effect is less than in the previous case. The Sternheimer ef fect i s quantif ied by the ant ishielding factor, y. The efg at the nucleus due to the distant charges and their ef fect on the outer valence electrons can be written as: The outer s electrons do not contribute to the efg because of the spherical symmetry of their charge d i s t r ibut ions: p,d, and f electron do. Furthermore, d and f wavefunctions are ca lculable, and i t is sometimes possible to derive the efg d i rect ly from them. It then becomes possible to t e l l which e lectronic configuration may have produced the observed f i e l d gradient. Shielding and ant ishielding effects are also produced by the electrons on each other, and these must be taken into account in the calculat ions. At the las t , the eigenvalues of the interact ion Hamiltonian must be given since i t i s these quantit ies which relate d i rec t l y to any physical measurement of a spectroscopic nature. The mathematical derivation can be found in the a r t i c l e by Steffen and Fraunfelder (18). Only the f i n a l result for the simple case of 1=3/2 is quoted here: (16) V zz zz ions (17) e 2qQ ~ 4I(2J-1) [3m^ - 1(1+1)] (l+n2/3) 1/2 eQ = Q 3 3 - 16 -Note that the eigenvalue has the m^  index, designating the sublevel of the nuclear spin, I, appearing only in the second power. Thus, states having m^  which d i f f e r only in sign are degenerate. The dependence of on T) i s quite complicated for I > 3/2 and the results for the 1=5/2 intermediate state of In-115 w i l l be given l a ter . Nevertheless, i t remains true that m^  only appears in the second power. 2B Methods of Measurement The four main spectroscopic techniques used in studying hyperfine interact ions, which consist of the nuclear e l ec t r i c quadrupole and nuclear magnetic dipole interact ions, can be placed into two groups. The f i r s t group is comprised of the NMR and NQR spectroscopies. Nonradioactive isotopes are used as the probes and radio-frequency ( r f ) f i e l d are applied to detect energy resonances. The second group contains the Mossbauer and PAC spectroscopies. Here, radioactive isotopes are used and the interest ing information is obtained by measuring the properties of the emitted y-rays. Each of these techniques has i t s advantages and disadvantages. The techniques w i l l be examined with regard to the physics exploited by each in making mesurements, and the type of information that can be obtained from them. Where possible, the discussion w i l l confine i t s e l f to those aspects most relevant to the quadrupole interact ion. - 17 -Group 1 Techniques (15,16) A nucleus must have I>l/2 to have a magnetic moment; i t must have I>1 to have an e l ec t r i c quadrupole moment. The magnetic moment of the nucleus is given by: (18) Uj- (gjHjj/101 = Y X I where g^ is the nuclear g factor, = he/(2m^) is the nuclear magneton, nip is the proton mass, and is the nuclear gyromagnetic r a t i o . When an external magnetic f i e l d , H Q , is applied to a c rys ta l containing a probe with I>1, and under the condition tht the re la t ive strengths of the magnetic and quadrupole interactions d i f f e r by several orders of magnitude (so that one of the interactions can be treated as a perturbation on the other in deriving the eigenvalues of the tota l interact ion Hamiltonian), the eigenvalues of the interact ion Hamiltonian are: ( 1 9 ) E = -hoa QIHJ + EQ , con 3 YJHQ = Larmor frequency H Q || V Z Z There are two cases for which 19 i s v a l i d . In the f i r s t case, the nuclear quadrupole interact ion energy is only a small perturbation on the Zeeman or nuclear magnetic dipole interact ion energy and the - 18 -relevant measurement technique is NMR spectroscopy. Resonances are observed in the rf range. However, in the second case, the quadrupole interact ion energy, even without the appl icat ion of an external magnetic f i e l d , i s large enough to generate rf resonances. Here, measurements are performed with the NQR technique. The frequencies of the nuclear quadrupole interact ion l i e in the range extending from tens of k i lohertz to approximately 1 gigahertz. The frequency of 2 megahertz is usually taken to be the point at which the " i nd i rec t " method of NMR i s replaced by the "d i rec t " method of NQR. There are two approaches to making mesurements with the group 1 techniques. They are the "stat ionary" and "pulsed" methods. In NMR the former method is implemented by applying a strong magnetic f i e l d to the sample and then subjecting i t to r f radiat ion and measuring the resonance absorption of energy. E ither the magnetic f i e l d is held constant while the frequency of the rf radiat ion is varied or vice versa. In NQR, a weak magnetic f i e l d is applied and only the rf radiat ion i s varied to scan the absorption spectum. The r f radiat ion induces transit ions between the energy levels given by (19) and resonance occurs whenever the rf matches a frequency corresponding to one of the l eve l sp l i t t i ng s . Various physical processes can broaden the spectral l ines so that the absorption occurs over a range of frequencies d istr ibuted about the idea l value. Information i s obtained by measuring the peak positions and their l ine shapes. In the pulsed method, the rf f i e l d i s applied as a series of short, powerful square pulses which rotate the tota l magnetic moment of - 19 -the sample away from the d irect ion of the s ta t i c applied magnetic f i e l d . The to ta l magnetic moment then proceeds to real ign i t s e l f with the s ta t i c f i e l d . This i t does via various relaxation processes which are characterized p r inc ipa l l y by two relaxation times; T^, the sp in - l a t t i ce relaxation time, and T 2 , the spin-spin-relaxat ion time, T-^  i s the l i fet ime for the reappearance of the magnetization along the axis of the s ta t i c external f i e l d . If the nuclear quadrupole interact ion i s strong, T^ is enlarged since a strong coupling of the nuclear spins to the interna l l a t t i c e f ie lds res i s ts the alignment of the magnetic moments by the external f i e l d , T 2 i s the l i fet ime for the decay of any transverse magnetization induced by the r f pulses. It has an inhomogeneous component due to the spread in the Larmor frequency ar i s ing from inhomogeneities in the applied f i e l d , and a homogeneous component due the dipole-dipole coupling between neighbouring spins. The main advantage that the pulsed method has over the stationary method is that i t s sens i t i v i t y depends mainly on the qual ity of the sample rather than the apparatus. Returning once again to (19), i t i s easy to dist inguish between the absorption spectra of the NMR and NQR techniques. In NMR the Zeeman l ines are shi f ted up and down by the quadrupole in teract ion. In NQR the Zeeman interact ion removes the degeneracy of the quadrupole l ines . This forms the basis of the standard technique of measuring n. What are some of the problems associated with the group 1 techniques? The inherent sens i t i v i t y of both techniques i s low because the net absorption of energy depends upon the excess population of the - 20 -low energy l e v e l . Under typ ica l conditions this excess i s small. Another disadvantage is that in powder or po lycrysta l l ine samples most of the intens i ty is concentrated in the central peak with the result that the quadrupole s a te l l i t e s are hard to detect. Most measurements on sol ids must be performed at 77°K or l i qu id nitrogen temperatue to reduce adverse thermal e f fect s . Furthermore, c rys ta l imperfections can broaden the peaks to the extent that the quadrupole spectra are undetectable in many so l ids . In materials having high e l e c t r i c a l conduct iv i t ies, the shallow penetration of the rf f ie lds only permits the exmination of surface phenomena. F ina l l y , there exist certain technical problems unique to NOR spectrometers which l imit their range of operation. The main advantage of group 1 techniques is the large number of isotopes suitable for these appl icat ions. The precis ion of the measurements, when possible, is quite high. Since stable isotopes are used, once a sample i s prepared i t can be used indef in i te ly (providing that i t i s chemically stable). Group 2 Techniques (17,18) The group 2 techniques are distinguished by a rel iance on radionuclides as probes. Nuclear quadrupole effects are studied by measuring the changes in the properties of spec i f i c radiations emitted by the probes. These changes are produced by the interact ion of the corresponding unstable nuclear state of the probe with the c rys ta l l ine - 21 -f i e l d s . In Mossbauer spectroscopy, the attention is focused on the y-ray produced by a part icu lar nuclear t rans i t ion . However, the s i tuat ion for PAC spectroscopy is much more complicated. Two radiations are detected which are the i n t i t i a l and f i n a l radiations given off In the course of a 2 or 3 step cascade of nuclear transit ions and/or transmutations. Only the two-step gamma cascades w i l l be considered here. Mossbauer spectroscopy is based on the emission of y~rays by nuclei undergoing a t rans i t ion from an excited state to the ground state, and the subsequent absorption of these y-rays by nucle i of the same type which consequently undergo the reverse process. However, the emission and absorption energies are minutely perturbed by the f i e ld s in the sol ids into which the nuclei have been incorporated. These small changes can be measured by taking advantage of the Mossbauer e f fec t . Deductions can then be made about the environment of the nuc le i . A typ ica l Mossbauer spectrometer consists of a y-ray source, a resonant absorber, and a detector. The absorber i s kept stationary while a wel l -contro l led backwards and forwards motion i s Imparted to the source. The emitted y-rays a r e thus systematically Doppler-shifted and the d i f f e r e n t i a l absorption of these y-rays by the absorber generates the Mossbauer spectum. In cases where very thin samples must be used, a resonant back-scattering arrangement i s preferred. The sh i f t in the energy leve l can be factored into three parts; the isomer sh i f t , the quadrupole interact ion, and the magnetic dipole - 22 -in teract ion. The isomer sh i f t , which is the order of 10~H eV, arises from interact ion of the nucleus with that part of the electron density which penetrates into the nuclear volume. The quadrupole interact ion, which is of the order of 10~7 eV, pa r t i a l l y reduces the degeneracy of the nuclear states when I>1. The magnetic dipole interact ion, also of the order of 10 - 7 eV, causes further s p l i t t i n g of the nuclear spin sublevels and removes a l l degeneracies. The linewidths are also of the order of 10~7 eV. However, the energy difference between the excited and ground states of the nucleus is of the order of 10 keV. Measurements of the quadrupole interact ion (on a single crysta l ) w i l l y ie ld i t s strength as well as the orientation of the crys ta l and efg axes with respect to the d irect ion of the y-ray beam. The energy resolution is usually not good enough to permit very good estimates of the asymmetry parameter, n, since the l ine broadening generally obl i terates the quadrupole s a t e l l i t e s . One of the main disadvantages of Mossbauer spectroscopy is the small number of useful radionucl ides. Most of the experimental work has been done using only two radionuclides, Fe-57, land Sn-119. About a dozen more have seen l imited use, and another two dozen are possible candidates for Mossbauer work. Only the so-cal led " r e c o i l - f r e e " f ract ion of the nuclei contribute to the Mossbauer spectrum. If this f ract ion i s small, large samples must be used. However, the low energy y-rays are s i gn i f i cant l y attenuated by photoelectric processes in a large sample. F i na l l y , while the other three techniques discussed here - 23 -require a s ingle sample preparation, a source and and absorber must be prepared for Mossbauer spectroscopy. The pr inc ipa l advantage of this technique i s that the experiments, due to the narrowness of the nuclear energy leve l s , are completely spec i f i c to the isotope being used. There is no cross- interference from other isotopes or elements. Furthermore, the sens i t i v i t y Is very high. It i s possible to obtain measurements from samples which contain as few as 10 1 2 probe atoms. Last ly , the properties of the host c rys ta l do not af fect the resonance absorption, which i s purely a nuclear process. This was not true for the NMR and NQR spectroscopies. The second group 2 technique i s PAC spectroscopy. Consider two y-rays which are emitted in cascade by a given nucleus. In the absence of extranuclear f i e l d s , the theory of multipole radiat ion predicts that the second y-ray w i l l have an angular d i s t r ibut ion with respect to the f i r s t which is determined by the nuclear parameters characterizing each t rans i t ion in the cascade. This i s referred to as the angular correlat ion of the two radiat ions. Suppose now that an efg is present and that the intermediate state of the decaying nucleus i s su f f i c i en t l y long- l ived for the quadrupole interact ion to be established. The efg a l ters the re la t ive populations of the nuclear spin sublevels of the intermediate state. As a resu l t , the re lat ive probabi l i tes of emission are also changed and the angular corre lat ion of the y-rays is perturbed. - 24 -There are two ways to perform a PAC measurement. The f i r s t ways i s to do a t ime-integral mesurement. However, this w i l l only t e l l the experimenter whether or not the efgs in the so l id are predominatly time-dependent or s t a t i c . It w i l l not reveal the number of interactions present or the parameters specifying each. Fortunately, this information can be obtained from a t ime-d i f ferent ia l measurement. Time-integral mesurements are usual ly only performed as f i r s t experimental checks, when the sources are extremely shor t - l i ved, or when the quadrupole interact ion energy is very high. A typ ica l PAC apparatus consists of two or more detectors at f ixed angles with respect to each other, some complex electronics for pulse sorting and noise reject ion (which w i l l be discussed l a te r ) , and a multi-channel analyzer for recording the time spectum. Noise reject ion Is achieved using the fast-slow coincidence method, which has the advantage of being a well-seasoned nuclear physics technique. A disadvantage of the TDPAC technique is i t s poor inherent frequency resolut ion. The intermediate state typ ica l l y has a l i fe t ime of the order of 100 nanoseconds. This translates into a natural linewidth of about 1 MHz, as compared to a natural linewidth of a few KHz in sol ids for NMR. This can be improved on in pract ice. There are about two dozen radionuclides with y-y cascades suitable for TDPAC work. However, many of these are short l ived and long chemical syntheses are precluded. Furthermore, the y-y cascades are in some cases preceded by a nuclear transmutation which can adversely affect the outcome of the experiment. - 25 -The main advantage of TDPAC spectroscopy is i t s high sen s i t i v i t y . Experiments can be performed with probe concentrations as low as 10~ 1 2M. Also, the technique looks only at hyperfine effects and i s applicable to interactions with frequencies ranging from 1 MHz to 1 GHz. The y-rays generally have energies of the order of 100 keV and are less subject to attenuation by the sample than those encountered in Mossbauer work. While i t is d i f f i c u l t to establ ish the absolute superior i ty of one of these four techniques with respect to the others, i t i s nevertheless possible to make a thoughtful choice as to which one i s the best suited for a spec i f i c appl icat ion. - 26 -Foreword to the Publications I have included in this thesis the f ive publications which I have found to be the most representative of the work done from 1982 to 1986 when I was most active as an experimental phys ic i s t . These publications de ta i l what was l i k e l y the f i r s t appl icat ion of TDPAC to the study of GICs, some interest ing work on the indium hal ides, and a cooperative ef fort with a group from Argentina. In reading these publ ications, a few important points should be kept in mind. The f i r s t of these is that the radioactive decays that produce the measured gamma-ray cascades generate an impurity at the s i te of interest . For example, Indium-Ill decays to Cadmium-Ill. This can have a s igni f icant impact on the measured efgs as i l l u s t r a ted in the paper by Kaplan et a l (20). Secondly, motions of the probe atom within the host l a t t i ce generate time-dependent efgs which are d i f f i c u l t to dist inguish from those due to electron capture processes. This i s par t i cu lar ly true of the work involving I n - I l l . - 27--LIST OF PUBLICATIONS 1. Quadrupole Interaction Studies of Graphite Indium Chloride Using T ime-Di f ferent ia l Perturbed Angular Correlat ions, S.R. Dong, P.W.Martin, J .G. Hooley, Carbon Vol.22, No. 4/5, pp.453-458,1984. 2. Quadrupole Interactions in Chlorides of Indium, P.W. Martin, S.R. Dong, J .G. Hooley, J . Chem. Phys. 80(4),15, pp.1677-1680,15 February 1984. 3. Quadrupole Interactions in Mixed-Valency States of Indium Chloride, P.W. Martin, S.R. Dong, J .G. Hooley, Chem. Phys. Letters , Vol.105,No.3,pp.343-346,16 March 1984. 4. Quadrupole Interactions in Graphite-Hafnuim Chloride, P.W. Martin, S.R. Dong, J .G. Hooley, Phys.Rev.B, Vol.33,No.6,pp.4227-4232,15 March, 1986. 5. Af teref fect Investigations in Mixed-Valence Indium Chlorides, C.P. Massolo, J . Desimoni, A.G. B i b i l o n i , A. Mowdoza-Zelis, L.C. Damonte, A.R. Lopez-Garcia, P.W. Martin, S.R. Dong, J .G. Hooley, Phys. Rev. B, Vol.34, No. 12, pp. 8857-8862, 15 December, 1986. - 28 -QUADRUPOLE INTERACTION STUDIES OF GRAPHITE INDIUM CHLORIDE USING TIME-DIFFERENTIAL PERTURBED ANGULAR CORRELATIONS S.R. Dong +, J .G. Hooley* and P.W. Martin +x Physics Department University of B r i t i sh Columbia Vancouver, B.C. Canada. V6T 2A6 * Chemistry Department University of B r i t i sh Columbia Vancouver, B.C. Canada V6T 1W5. x To whom correspondence should be addressed - 29 -ABSTRACT The technique of t ime-d i f ferent ia l perturbed angular correlat ions has been applied to the study of quadrupole interactions in graphite intercalated with indium chlor ide. Indeed, these experiments represent the f i r s t appl icat ion of this part icu lar spectroscopic technique to graphite in terca la t ion systems. With l l l I n incorporated as a probe, y-y angular corre lat ion measurements were performed on the 173-247 keV cascade in the daughter product 1 1 1 C d . Information on quadrupole interactions was obtained from Fourier analysis of the data and by comparison with theoret ica l f i t s based on standard angular corre lat ion theory. Three s i tes are found for intercalated InCl^: two of these correspond to s tat ic e l e c t r i c f i e l d gradient interact ions, which for a Gra fo i l sample yielded mean values of v ( 1 ) - 57.8 ± 0.8 MHz and v ( 2 ) - 267 ± 5 MHz; i t i s suggested that the q 9 t h i r d , time-dependent component represents a s i te to which electron transfer i s inh ib i ted. The po s s i b i l i t y that signals may have been observed from adsorbed as opposed to intercalated InCl^ was eliminated by comparison with measurements on Spheron 6 carbon black. - 30 -Introduction The interca lat ion of metal chlorides in graphite has received a great deal of attention from many authors using a variety of techniques [1,2]. In part icu lar the c l a s s i c a l hyperfine techniques of nuclear magnetic resonance (NMR), nuclear quadrupole resonance (NQR) and Mossbauer spectroscopy have played important roles in establ ishing the nature of these compounds. In a br ie f communication [3] we have recently drawn attention to the use of another hyperfine technique, that of t ime-d i f ferent ia l perturbed angular correlat ions of y-ray6 (TDPAC) for the study of graphite layer systems. Here we report further deta i l s of TDPAC studies on graphite indium chlor ide. When a radioactive nucleus de-excites by emitting two v-rays in cascade, the delayed coincidence counting rate can be shown to be [A] " t / T N W(6,t) - £ — I A P (cos 9) \ > / TJJ W V where 9 is the angle between the direct ions of propagation of the two Y~ r a y s , i s the l i fet ime of the intermediate state, P^  are Legendre polynomials and A ^ are parameters which depend on the nuclear leve l spins and mul t ipo lar i t ies of the t rans i t ions. If, however, during the time elapsing between emission of the f i r s t and second y-rays the nucleus is exposed to an electromagnetic f i e l d , the population of the magnetic sublevels of the intermediate state can be a l tered, thereby af fect ing W(9,t). In this case the perturbed angular corre lat ion i s written as - t / x N W(8.t) - ^ I A y v G v v ( t ) P v(cos 9) N v where G v v ( t ) i s the time-dependent perturbation factor [4]. Since the form of G v v ( t ) depends on the nature of the perturbing interact ion (magnetic J i po le , e l ec t r i c quadrupole e t c . ) , the TDPAC technique exploits this fact to - 31 -gain information about the external f i e l d causing the perturbation. In the s i tuat ion where a s ta t i c e l e c t r i c f i e l d gradient (efg) interacts with the quadrupole moment Q of an intermediate state with spin 1 • 5/2, the form of G 2 2 ( t ) for a po lycrysta l l ine sample i s given to second order by 3 . G 2 2 ( t ) - S 2 Q + I S 2 n (n ) exp (~ 1 ^ ° 2 t 2 ) cos [ f ^ n ) u t ] (1) n — l 2 where u (TI) • f (n)u) are the quadrupole frequencies, n is the asymmetry n n o parameter for an efg with width 6 and the S 2 n coef f i c ients depend on the nuclear decay scheme parameters. Since the f i n i t e resolving time, t „ , of the coincidence c i r cu i t r y can 1 9 lead to damping of high frequency components, a factor of the form exp(-^ o> * t_ 2 ) i s also incorporated in eqn( l ) . By f i t t i n g the data to an equation of form (1) i t i s therefore possible to extract the parameters describing the quadrupole interact ion. The strength of the quadrupole interact ion is often expressed in terms of the coupling constant e 2qQ, with eq • V » °r In frequency units as v - eQV /h . q zz 2. Experimental In the investigations described here the radioactive probe was 1 1 * In , obtained commercially as ca r r ie r - f ree 1 1 1 I n Cl^ in HC1. A suitable ac t i v i t y of indium was f i r s t electroplated on natural In (99.9992). A weighed amount of this and of graphite were then placed at opposite ends of a Pyrex tube. The tube was then sealed at the In end and evacuated to 1pm through the other end* Dry C l 2 was then admitted to 1 atmosphere before gently heating the In to convert i t completely to the t r i ch l o r i de . The tube was f i n a l l y evacuated to 1 um, f i l l e d with C l 2 to 300 mm of Hg and sealed. It -was then placed in a gradient tube furnace for times ~ 36 h with T g r - 460"C and the opposite end at a lower temperature to minimize condensation of the chloride [5]. The tube was removed and cooled quickly but in such a - 32 -fashion that the temperature of the graphite end was above that of the other end u n t i l room temperature was reached - about 1 m. The composition was calculated from the increase in the weight of the graphite. For Madagascar graphite jet mil led to 2 u.m across,, i t was C^In C 1 3 f o r 4 o 0 ° / 4 5 0 0 c « Other compositions were obtained for various temperature differences and for other graphites in agreement with previous work [5]. In order to estimate the behavior of InCl^ adsorbed on carbon, the sample preparation was repeated using Spheron 6 2700 s C. This i s a Cabot Carbon product with an area of 100 m2 gm - 1 which does not form an interca la t ion compound. In addit ion, since no information was available on quadrupole interactions in the bulk interca lant , measurements were performed on the chloride powders of indium (InCl, InCl^ InCl^ g, InClj and InCl^). Detai ls of these experiments w i l l be reported in a separate publ icat ion. The TDPAC measurements were made on the 173 keV - 247 keV cascade in ^ C d , the daughter product following electron-capture decay in ^ I n (Fig. 1). As the r e c o i l energy following the electron-capture decay in 1 1 1 In i s small, the ***Cd probe nuclei are expected to occupy regular In l a t t i c e s i te s . A system of four Na I (T£ ) detectors placed symmetrically about the sample was used to record the W(n/ 2, t) and W(u,t) angular correlat ions. Detai ls of the electronic c i r c u i t r y have been published elsewhere [6]. The advantages of a multiple detector system include higher data acquis it ion rates, el imination to f i r s t order of detector ef f ic iency rat ios and compensation of errors in source centering [7]. 3. Data Reduction and Analysis The experimental values for <>22(t) were computed d i rect ly from the expression V 3<*,t) W (Tt . t ) 1 / 2 G „ ( t ) - — [ - ] -1 (2) / U 3 A 2 2 W u ( n /2 , t ) V 2 3 ( * / 2 , t ) where W . (9,t) were the correlat ions recorded by detectors i , j with random coincidences subtracted, and A ^ incorporated corrections for the f i n i t e so l id angles subtended by the detectors. - 33 -In the second approach the G 2 2 ( t ) data w ere f i t t ed with a theoret ica l To extract information on the parameters of the quadrupole interact ions two approaches were adopted. In the f i r s t case the frequency information was obtained d i rec t l y by computing the discrete Fourier transform (DFT) of the autocorrelated G 2 2 ^ data. W i t h the advantage that no model i s assumed, the true frequencies are concentrated in r e l a t i v i l y narrow regions of the power spectrum whereas the noise is d istr ibuted over the entire bandwidth. Since n i s known as a function of the frequency rat ios v 2 / v and v ^ / v 1 [8], i t was readi ly determined from the observed frequencies. S imi lar ly , using the known relat ionship of V q / V j a s a function of n, v^ was determined. From the relat ionship v - eQV /h, n - ( V - V )/V and V q zz xx yy zz xx + + " 0, the components of the efg were determined using the measured value of Q for the quadrupole moment of the intermediate state [9]. In the second ap expression of the form G 2 2 ( t ) = c Q exp ( - C /T) + C l G ^ } ( t ) + c 2 G ^ } ( t ) (3). Here the G 2 2 ^ correspond to s ta t i c quadrupole interactions of type (1) and the c i represent the f ract ion of nuclei exposed to each interact ion, with I c 1 - 1. In order to achieve an acceptable f i t i t was also necessary to introduce the time-dependent f ract ion c n with time-constant -t. 4. Results and Discussion Character is t ic data for an intercalated sample are presented in F i g . 2, while a typ ica l Fourier power spectrum i s shown in F ig . 3. Quadrupole interact ion frequencies were Ident i f ied i f se l f -consistent rat ios of v2^v\ and ^2^1 w e r e obtained for the determination of n. The unresolved signal close to 40 MHz, which appeared to varying degree in a l l the samples, could not be ident i f ied as one of such a t r i p l e t . The so l id l ine in F i g . 2 represents a theoret ica l f i t of the form indicated in eqn(3). Since the sample c r y s t a l l i t e s may not have been t ru ly randomly oriented, but assumed preferred or ientat ions, the S 2 n coef f ic ients in eqn(l) were allowed to vary. - 34 -Detai ls of the f i t s and DFT analysis are given in Tables 1, 2 and 3. The two methods of analysis give good agreement for the derived quadrupole interact ion parameters. The theoret ica l f i t s are consistent with the existence of three s ites for the intercalated InCl, . There are two s ta t i c efg interact ions and (2) G^„ , and a th i rd f ract ion c , represented by the time-dependent term in o eqn. (3) with time constant close to 30 ns. In the case of C^^lnCl^, which was prepared from Madagascar natural f lake jet -mi l led to about 2um, this component constituted 15.6Z of the sample, r i s i ng to 29.5% for C^InCl^ prepared from Union Carbide G ra fo i l . For the g ra fo i l sample the weighted means for the quadrupole frequencies were v • 57.8 ± 0.8 MHz and (2) q v v ' - 267 ± 5 MHz. TDPAC results on bulk chlorides of indium showed that q no quadrupole interact ion frequencies were detected corresponding to those observed in the intercalated samples. However, there i s s t i l l the question of whether the time-dependent component C q comes from intercalated or from surface adsorbed InCl^ or both. From geometric considerations alone, less than 5% of the to ta l attenuation factor can originate from surface adsorbed material. Thus, from the observed cross dimension of 2p.m and the estimated thickness of 0.1 u-m for the Madagascar f lakes, i t can be calculated that i f the C Q component of C^InCl-j were ent i re ly adsorbed on the flake surfaces, the area avai lable is too small by a factor of 20. The results with Spheron 6 2700° support this conclusion. Thus, the weight of InCl^ that added to this carbon uses only 40% of the surface area avai lable so that no bulk material can be present. This figure comes from a surface area of 100 m2 gm - 1 and a mean radius of 0.25 nm for the InCl^ molecule [10]. The 293°K data in F i g . 4 for this surface adsorbed InC l 3 manifests a rapid destruction of the angular corre lat ion, quite di f ferent in character from that in F i g . 2. This could be explained in one of the following ways: (a) a time-dependent perturbation produces a fast destruction of the corre la t ion; - 35 -(b) a s ta t i c quadrupole interact ion i s present, but with a wide frequency d i s t r ibut ion as a result of a highly non-unifonn e l e c t r i c f i e l d ; (c) a s ta t i c quadrupole perturbation i s present with an Interaction frequency beyond the resolution of the apparatus (> 500 MHz). Case (a) i s preferred here as the t ime-integral attenuation factor, ^ 2 2 ^ ™ ^*140, * s w e H b e l ° w the hard-core value of 0.2, implying that the adsorbed InCl^ experiences a time-dependent quadrupole interact ion, possibly caused by a combination of two mechanisms. In the f i r s t process, the electron-capture in 1 1 1 I n which precedes the y~Y cascade under observation can leave the daughter 1 1 * C d nucleus in a number of possible highly ionized states. In good conductors e lectronic equil ibrium i s achieved rapidly by charge transfer, leaving the subsequent decay unaffected, whereas in insulators slow recovery times (> 10~ 8s) w i l l result in rapid destruction of the angular correlat ion with an unspecified time-dependence [11]. An adsorbed InCl^ molecule, being adjacent to a single carbon layer, has access to only half as many electrons as an intercalated InCl^ molecule, hence poorer charge transfer can be expected. The second mechanism which could produce a randomly f luctuat ing efg is d i f fu s ion . At low temperatures, however, the thermal motion is diminished, resul t ing in pa r t i a l recovery of the angular corre lat ion. The data in F i g . 5 for a carbon black sample at l i qu id nitrogen temperature shows precisely this behaviour, with G 2 2 ( » ) r i s i n g to 0.170. In this case G 2 2 ( t ) f a l l s to a minimum, then r i ses to a low maximum at time T • 30 ns, after which i t essent ia l ly remains constant. An estimate of the quadrupole frequency (assuming ax ia l symmetry), using the relat ionship e2qQ - hv^ • 20h/3T [ 4 ] , yie lds v • 220 MHz for this in teract ion, well removed from frequencies of 59 q MHz and 267 MHz. The evidence therefore suggests that no contributions from adsorbed InCl^ appear in the time spectra of Intercalated samples. As far as the intercalated samples are concerned, a number of conclusions can be drawn. The f i r s t of these i s that, within experimental error , the same three quadrupole interactions occur in a l l the C^InCl^ compounds. Also, for a given variety of graphite, the do not change with x for stage index n > 2. As far as the time-dependent C q component i s concerned the same remarks apply here as before regarding the mechanisms for - 36 -this contr ibut ion. Electron transfer to this s i te is inhib i ted and d i f fus ion processes may also be a factor . Because of broadening caused by y-ray attenuation in the l i qu id nitrogen, however, i t was not possible to estimate accurately the re lat ive contribution from d i f fus ion processes. Plans are underway to perform such measurements as a function of temperature with much reduced y-ray absorption. With regard to the s ta t i c interact ions, the values obtained for the quadrupole frequencies are the same within error l im i t s , regardless of the source of graphite or i t s stoichiometry. The last result i s not unexpected i f the assumption i s made that only nearest and next-nearest neighbours contribute to the efg. From this standpoint intercalant layers separated by two carbon layers would not affect the efg, nor would changes appear in the quadrupole frequencies as the stage index is increased. A pure stage 2 compound would have had a composition of C ^ I n C l 3 [ 1 2 ] . Thus the sample with lowest stage index, Cj^InCl^, was probably a mixture of stages 2 and 3 and would not be expected to experience a d i f ferent efg. We have no knowledge from other work or from our own about the structure of InCl^ between the layers. In pr inc ip le information about the structure could be obtained by comparison with theoret ica l calculations of the efg. We are unaware, however, of any r e a l i s t i c efg calculations for such a complex structure. Our measurements on bulk anhydrous powders of indium chloride indicate that the efg at In in intercalated samples is not the same as in pure so l id InCl^, InC^, or other mixed valency states. In the intercalated samples there must be charge transfer between the In nucleus and the adjacent carbon layers, presumably also involving the chlor ine, which a l ters the efg. Our results thus indicate that there are three d i s t inct degrees of electron transfer in graphite-InCl^, one of which, the CQ component, behaves l i ke an insu lator. Our data gives no measure of the extent to which the distorted In Cl^ polymeric layer structure approaches that of the carbon structure. Again, i f theoret ica l calculat ions of the efg were ava i lab le, the re la t ive contribution to the efg from surrounding indium and adjacent carbon atoms would give an indicat ion of the degree to which - 37 -InCl^ i s commensurate with the graphite l a t t i c e . From a purely geometric standpoint i t i s possible to envisage models with domains of intercalated InCl^ containing both commensurate and incommensurate species. In contrast t X-ray and neutron scattering techniques, TDPAC can readi ly observe heterogeneous intercalate d i s t r ibut ions within a single layer. It is not poss ible, however, in the present case to determine whether the intercalate layer i s s ingle phase or multiphase v l t h respect to the observed quadrupole interact ions. The existence of multiphase systems has been observed in other halide interca lants . In the case of ferrous chlor ide, Mossbauer measurements show that the presence of trapped Cl^ modifies the efg at certa in s ites [13]. Electron and x-ray d i f f r ac t i on studies have indicated that graphite molybdenum pentachloride exhibits at least two phases [14]; one of these corresponds to a dimer molecular structure, while the others are characterized by hexagonal and disordered cation d i s t r ibut ions . In the case of graphite indium chloride several hypotheses can be postulated to explain the existence of the three s i tes observed in the TDPAC measurements. For example the presence of anionic complexes such as [InCl,] in the form [graphite]"*" [InCl^] -.xInCl.j has been suggested by Rudorff and Landel[15]. Presumably p o s s i b i l i t i e s exist for the presence of several ion i species in thermodynamic equi l ibr ium. Theoretical calculations of the efg would be desirable to test such hypotheses and plans are underway to investigate these aspects. These measurements show that TDPAC can be a useful tool in characterizing s i tes in graphite interca lat ion compounds. It would be interest ing to see i f staging effects at low orders (n-1,2) can be detected. Suitable candidates for such studies, including probes in which electron capture does not occur, are currently being investigated. - 38 -References 1. M.S. Dresselhaus and G. Dresselhaus, Adv. Phys. 30, 139 (1981)'. 2. H. Sel ig and L.B. Ebert, Adv. Inorg. Chem. Radlochem, 23_, 281 (1980). 3. S.R. Dong, S. El-Kateb, J .G. Hooley and P.U. Martin, Sol id State Comm. 45, 791 (1983). Note errata In this paper for the e l e c t r i c f i e l d gradient components owing to an incorrect value for Q. Using Q • 0.83(16) from re f . 9, the values reported in this paper should be mult ip l ied by the constant factor 3.024. 4. H. Frauenfelder and R.M. Steffen, i n Alpha-, Beta and Gamma-Ray Spectroscopy, ed, by K. Siegbahn, North Holland, Amsterdam, 1965, Vol . 2, p.997. 5. J.G. Hooley, Carbon 11_, 225 (1973). 6. P.W. Martin, S. El-Kasteb and U. Kuhnlein, J . Chem. Phys. 86_, 3819 (1982). 7. A.R. Arends, C. Hohenemser, F. P l e i t e r , H. DeWard, L. Chow and R.M. Suter, Hyperfine Interactions _8, 191 (1980). 8. E. Gerdau, J . Wolf, H. Winkler and J . Braunsfurth, P r o c Roy. Soc. A, 311, 197 (1969). 9. 0. Echt, H. Haas, E. Ivanov, W. Rechnagel, E. Schlodder and B. Spellmeyer, Hyperfine Interactions _2, 230 (1976). 10. K. Wade and A . J . Bannister, The Chemistry of Aluminium, Gallium, Indium and Thallium, Pergamon Texts in Inorganic Chemistry, Vo l . 12, 1975, Pergamon Press. 11. H. Haas and D.A. Shir ley, J . Chem. Phys. 58, 3339 (1972). 12. E. Stumpp, Materials Science and Engineering 31_, 53 (1977). 13..J.G. Hooley, J.R. Sams and B.V. Liengme, CARBON 8_, 467 (1970). 14. A.W. Syme Johnson, Acta Cryst. 23_, 770 (1967). 15. Von W. Rudorff and A. Landel, Z. anorg. a l l g . Chem. 279, 182 (1955). - 39 -Figure Captions F i g . 1. Decay scheme of 1 1 1 I n . F i g . 2. The perturbation factor at room temperature for an intercalated Gra fo i l sample with composition Cg^ In C l j . The so l id l ine is the theoret ica l f i t using eqn.(3). F i g . 3. The Fourier power spectrum of an intercalated sample of Madagascar jet -mi l led graphite ( C 2 3 In C l 3 ) . F i g . 4. The perturbation factor at room temperature for a Spheron 6 carbon black sample with adsorbed 1 1 1 I n CI3. F i g . 5. The perturbation factor at 77K for a Spheron 6 carbon black sample with adsorbed m I n CI3. TABLE 1 Fit coefficients for Madagascar and Rraphotl samples. MADAGASCAR ( C 2 3 InClj) CRAF01L ( C 9 4 In CIj) c t(T) t(ne) S 2 0 S2, S 2 2 c,(Z) t(nn) S 2 0 Sjj S 2 2 15.6 30.4 - - -60.2 - 0.165 0.362 0.313 24.2 - 0.441 0.563 0.058 29.5 26.9 - -47.1 - 0.167 0.340 0.255 23.4 - 0.325 0.571 0.108 TABIX 2. Static quadrupole Interaction parameters from f i t s to eqn (2) MADAGASCAR CRAPHITE ( C 2 J In Cl,) (2|io) GRAFOIL ( C g 4 In 3Cl ) c$s> v q (MHt) n 6 v l 8 O 0 1 7 V/cm?) 56*1 0.43*0.04 0.20*0.02 2.8*0.5 270*6 0.12*0.03 0.06*0.01 13.5*2.5 vq(MHz) n 6 V„(10 1 7 V/cm2) 58*1 0.43*0.04 0.24*0.02 2.9*0.5 26B*6 0.17*0.02 0.07*0.01 13.4*2.5 TABLE 3 Static quadrupole Interaction parameters from Fourier analyses. MADAGASCAR ( C 2 3 In CI]) GRAFOTL ( C 9 4 In Clj) MHc vq(MHz) Hllz vq(MHz) (1) v l 10.2~ 10.0~ (1) v2 16.7 — 0.44 • 0.05 ( a ) 57 t 2 16.9 — 0.41 • 0.05 ( a> 57 • 2 (1) v3 26.6_ 26.2_ (2) v l (2) v2 42.8~ 84.8_ — 0.08 i 0.16 +2 283 -in 41.7~ 77.9_ — 0.24 • 0.13 262 t 13 (2) v3 - -(a) Weighted nean based on v j / v 2 a"d v l ' v 3 ratios. In ENERGY (keV) 4 2 0 2 4 7 0 Figure 1. 1 = 7/2 v ( 1 7 3 keV) •1=5/2 (T ) / 2 = 85 ns) v ( 2 4 7 keV) 1= 1/2 C d T„ = 2 . 8 d 1/2 i *» u> l - 44 -- 48 -THE UNIVERSITY OF BRITISH COLUMBIA Department of Physics QUADRUPOLE INTERACTIONS IN CHLORIDES OF INDIUM P.W. Martin*, S.R. Dong* and J.G. Hooley** * Department of Physics University of B r i t i sh Columbia Vancouver, B.C. V6T 1W5. * * Department of Chemistry University of B r i t i sh Columbia Vancouver, B.C. V6T 1W5. - 4 9 -ABSTRACT With 1 1 1 In incorporated as a probe into anhydrous powders of InCl, InCl^ and InCl^t t ime-d i f ferent ia l perturbed angular corre lat ion measurements were performed on the 173-247 keV y-y cascade in u l C d . From Fourier analysis of the data information on s ta t i c e l e c t r i c quadrupole interact ions was obtained, comprising interact ion frequencies, e l e c t r i c f i e l d gradient components and asymmetry parameters. For the t r i ch lo r ide at room temperature the angular corre lat ion i s heavily damped and structureless; at T • 203°C InCl^ reveals two s tat ic interactions with frequencies " 38.8 ± 0.7 MHz and (2) v - 95.7 ± 2.0 MHz. For InCl- a single interact ion i s detected at room q 2 temperature, with » 135.0 ± 1.1 MHz. The a and 8 phases of InCl are readi ly dist inguished; for the former v^ - 284.1 ± 2 . 6 MHz while for the l a t te r at T • 167°C the angular corre lat ion i s unperturbed. - 50 -1. INTRODUCTION In studies of quadrupole interactions of intercalated halides in graphite, previous workers have demonstrated that i t Is important to have information on the hyperfine interactions In the bulk Intercalant. For example Hooley et a l . 1 were able to dist inguish s i tes i n graphite iron chloride by comparing quadrupolar sp l i t t ings In bulk F e C l 2 with their Mossbauer measurements on the intercalated graphite system. Information on nuclear quadrupole interactions and resonances in indium halides i s sparse 2 , presumably because NMR and NQR spectroscopy on 1 1 5 I n (95.722 abundant) and U 3 I n (4.282) i s i n t r i n s i c a l l y d i f f i c u l t for I - 9/2 spins. An attempt by Ludwig et a l . 3 to detect quadrupole interact ion frequencies in InCl^ was unsuccessful. Measurements hitherto reported on 6 o l i d InCl^ using the technique of perturbed angular correlat ions of y~rays have been of the t ime-integral variety. 1* Following recent measurements in this laboratory by t ime-d i f ferent ia l perturbed angular correlat ions (TDPAC) of y-rays on graphite indium ch l o r i de 5 , we report here results of TDPAC studies on anhydrous powders of InCl, I nC l 2 (actual ly In^In m c i 4 ] ) and I nC l 3 . 2. EXPERIMENTAL PROCEDURE The TDPAC measurements were performed on the 173-247 keV Y - Y cascade in 1 1 * C d , the daughter product produced by the electron-capture decay of ^^In. Since the r e c o i l energy in the electron-capture decay of 1 1 1 I n i s very small, the probe ^ C d nuclei are expected to remain located at regular In l a t t i c e s i te s . The indium a c t i v i t y , obtained commercially as car r ie r - f ree ^ I n C l ^ In hydrochloric ac id, was electroplated onto f o i l 6 of high-purity - 51 -indium (99.999%) for subsequent preparation of the chloride samples by standard methods 6 7 . For the preparation of InCl^, the radioactive f o i l was placed in a pyrex tube (12 mm o.d.) which was then evacuated. Dry chlorine gas was admitted at a pressure of 1 atmosphere, following which the f o i l was heated u n t i l the reaction occurred, producing white c r y s t a l l i t e s of InCl^. To prepare the d ich lor ide, two Indium f o i l s with masses in the ra t io 1:2 were placed at opposite ends of the pyrex tube. In this case only the larger f o i l had been electroplated with 1 1 1 I n a c t i v i t y . The l a t te r was then chlorinated as before to form InCl^. The tube was evacuated, sealed and then furnaced for several hours at 450 8C. White c r y s t a l l i t e s of the dichlor ide were obtained when the tube was cooled to room temperature. The preparation of the monochloride was accomplished in a s imi lar manner except that in this case the roles of the f o i l s were interchanged and the system allowed to cool more slowly. This procedure yielded the yellow phase of InCl which i s stable below l l O ' C . 8 The red phase of InCl was obtained by heating the sample to 167°C i n argon gas. For the TDPAC measurements the powders were transferred in an inert atmosphere of argon gas to smaller sample tubes (o.d. 2mm). Details of the e lectron ic c i r c u i t r y and detector system, comprising four symmetrically placed Nal(Tl) spectrometers, have been published elsewhere. 9 Data were recorded at room temperature for a l l the samples. In addition runs were performed at 203°C on InCl- and at 167°C on InCl. - 52 -3. Data Reduction and Analysis A comprehensive treatment of perturbed angular correlat ions is given in the review a r t i c l e by Frauenfelder and S t e f f e n . 1 0 For the case of the y-y cascade in 1 1 * C d (1-5/2) the angular corre lat ion can be written to second order as W(9,t) - 1+ A 2 2 G 2 2 ( t ) P 2 (cos 9) where A 2 2 depends on the multipole nature of the t rans i t ions , P 2(cos 9) i s the Legendre polynomial and G 2 2 ( t ) i s the perturbation factor . The l a t t e r , which contains the information of interest , was obtained by combining the measured angular correlat ions in the form 2 W 1 3 ( * , t ) W 2 4 ( « , t ) 1 / 2 G 2 2 ( t ) ' 3 1 ^ " ( w u ( u / 2 , t ) W 2 3 (*/2 f t>) " 1 where W^(9,t) refers to the angular corre lat ion recorded by detectors i , j with A 2 2 corrected for f i n i t e so l id angle e f fec t s . This method of combining the affords the advantage that detector e f f i c ienc ie s cancel to f i r s t order and compensates for errors in source centering. By performing a discrete Fourier transform (DFT) on these data the power spectrum was extracted to y i e ld the quadrupole interact ion frequencies: V l " ( E ± 3 / 2 ~ E ± l / 2 ) / h ' V 2 " ( E ± 5 / 2 " E ± 3 / 2 ) / h a n d V 3 " ( E ± 5 / 2 _ E ± l / 2 ) / h ' This method of analysis i s model-independent, extracting frequencies - 53 -d i r e c t l y , with the addit ional advantage that the true frequencies are concentrated in re l a t i ve l y narrow regions of the frequency spectrum while the noise i s d istr ibuted over the entire bandwidth. The asymmetry parameter r\ for a s ta t i c e l e c t r i c f i e l d gradient (efg), defined in terms of the efg components by V - V ^ v zz was then determined from the frequency rat ios v 2 ^ v ^ a n < * v 3 ^ v i » w h ich are known as functions of n . 1 1 Once n was determined i t was then e2Qq straightforward to calculate the quadrupole interact ion frequency » — j p -and the components of the efg. 4. RESULTS AND DISCUSSION A summary of the results obtained for the three chlorides is shown in Table 1, indicat ing the frequencies v ident i f ied (by sel f -consistent rat ios n for v 2 / v i a n c * v 3 / ' v i ^ * n t n e Fourier power spectra, while Table 2 l i s t s the calculated values of n, v and V . The l a t ter were calculated using the q zz value Q - 0.83(16)b for the quadrupole moment of the 247 keV state of m C d (I « 5 / 2 ) 1 2 . I nC l 3 It i s known from v ibrat ional spect ra 1 3 that the metal atoms in so l id indium t r i ch lo r ide are 6-coordinate in a polymeric layer l a t t i c e . The TDPAC data for the t r i ch lo r ide at T - 20 4C and T - 203°C are shown in F igs. 1 and 2 respect ively. From F ig . 1 i t can be seen that the correlat ion i s rapidly destroyed, result ing in non-periodic behaviour for G22(t)« This - 54 -suggests the presence of a time-dependent interact ion, presumably as a result of the slow recovery (>10~8s) of the ionized state of the electronic she l l i n 1 J , 1 C d following electron-capture in the parent ^ I n . A l ternat ive ly such an effect could have been produced by a s ta t i c quadrupole Interaction with either a very wide frequency d i s t r ibu t ion , or with high frequency components (> 500 MHz) which could not be resolved by the apparatus. The measurement at T - 203° C was performed to elucidate these questions. It was not desirable to go to higher temperatures in order to avoid phase changes in the crys ta l structure, while at the same time re la t i ve variations in the l a t t i c e constants should be small. The higher temperature data reveal o sc i l l a tory behaviour in G 2 2 ( t ) , as shown in the DFT power spectrum of F ig . 3, which indicates the presence of two s ta t i c quadrupole interactions with n ( 1 ) - 0.48 ± 0.03, T / 2 ) - 0.18 ± 0.04 and frequencies v ( 1 ) - 38.8 ± 0.7 MHz (2) and ' • 95.7 ± 2.0 MHz. In the unl ikely event that over the temperature range involved a high frequency interact ion could sh i f t so d ra s t i ca l l y , these results would appear to exclude the l a t t e r . Integral measurements of the perturbation factor confirm the presence of a time-dependent perturbation at room temperature, with < G 2 2 ( « ) > - 0.14 ± 0.01, whereas at T - 203°C the hard-core value of 0.2 i s exceeded with < G 2 2 ( » ) > - 0.22 ± 0.02. At the higher temperature the e lectronic interact ion between atoms in the crys ta l i s stronger, enhancing the probabi l i ty of charge transfer and thereby reducing the recovery time for e lectronic equil ibrium following electron-capture decay. - 55 -I nC l 2 Indium dichlor ide has a mixed valency structure 1"* in which indium is present in both the +1 and +3 oxidation states i . e . In*[In***Cl^]. F i g . 4 shows the TDPAC data for the dichlor ide at room temperature, with the corresponding DFT in F i g . 5* In this case one s ta t i c quadrupole interact ion can be ident i f ied with frequency v^ - 135.0 ± 1.1 MHz. The frequency rat ios for this interact ion are consistent with ax ia l symmetry (n - 0) for the efg. In addit ion, since G 2 2 ^ " ^ * * 7 ± n , 0 * » there appears to be a time-dependent perturbation. InCl The crys ta l structures of the yellow(a) and red (8) phases of indium monochloride have been studied extensively 8 , 1 5 , 1 6 . The a-phase is found to be a deformed NaCl structure with space group P2i3. The perturbation factor and corresponding DFT spectrum are shown in f i g s . 6 and 7 respect ively. In this case only one s ta t i c quadrupole interact ion is i den t i f i ed , with n - 0.35 ± 0.01 and frequency - 284.1 ± 2.6 MHz. Van der Vorst et a l . 1 5 conclude that the most l i k e l y structure for B - In CI i s orthorhombic, with space group Cmcm. The TDPAC data in this case (F ig. 6) at T • 167°C indicate that the angular correlat ion is v i r t u a l l y unperturbed. In the absence of rapid rotat ional motion of the InCl molecules at this temperature, which could average to zero any perturbations, the results indicate an efg close to zero. - 56 -REFERENCES 1. J.G. Hooley, J.R. Sams, and B.V. Liengme, Carbon 8^ , 467 (1970). 2. S.L. Segel and R.G. Barnes, "Catalog of Nuclear Quadrupole Interactions and Frequencies In So l ids " , Part 1 U.S.A.E.C. Report IS-520 (1968); G.K. Semln, T.A. Babushkina, G.G. Yakobson, i n Nuclear Quadrupole Resonance in Chemistry (John Wiley and Sons, New York, 1975, p. 213). 3. G.W. Ludwig, J . Chem. Phys. 25_, 159 (1956). 4. R.M. Steffen, Phys. Rev. 103, 116 (1956). 5. S.R. Dong, S. El-Kateb, J .G. Hooley, and P.W. Martin, Solid State Commun. 45, 791 (1983). Note errata in this paper for the quoted efg components owing to an incorrect value of Q. Using Q - 0.83(16) from ref . 10, the values reported here should be mult ip l ied by the constant factor 3.024. 6. A.W. Atkinson, J.R. Chadwick and E. K inse l l a , J . Inorg. Nucl. Chem. 30, 401 (1968). 7. K. Wade and A . J . Banister, The Chemistry of Aluminium, Gallium, Indium Thallium, Pergamon Texts in Inorganic Chemistry, Vo l . 12, 1975, Pergamon Press. 8. J . M. van den Berg, Acta Cryst. 20, 905 (1966). 9. P.W. Martin, S. El-Kateb AND U. Kuhnlein, J . Chem. Phys. 76, 3819 (1982). 10. H. Frauenfelder and R.M. Steffen, in Alpha-, Beta- and Gamma-Ray Spectroscopy, edited by K. Siegbahn (North Holland, Amsterdam, 1965, Vo l . 2, p.997). 11. E. Gerdau, J . Wolf, H. Winkler,and J . Braunsforth, P r o c Roy. Soc. A311.197 (1969). 12. 0. Echt, H. Haas, E. Ivanov, E. Recknagel, E. Schlodder and B.Spellmeyer, Hyperfine Interactions 2t 230 (1976). 13. N.N. Greenwood, D.J. Prince and B.P. Straughan, J . Chem. Soc (A) 1694 (1968). 14. A.W. Atkinson and B.0. F i e l d , J . Inorg. Nucl. Chem. 32_, 3757 (1970). 15. C.P.J.M. van der Vorst, G.C. Verschoor and W.J. Maaskant, Acta Cryst. B34, 3333 (1978). 16. C.P.J.M. van der Vorst and W.J.A. Maaskant, J . Solid State Chem. 34, 301 (1980). - 57 -Figure Captions F i g . 1. The perturbation factor at room temperature for anhydrous In CI2 powder. F i g . 2. The perturbation factor at T - 203°C for anhydrous In C l^ powder. F i g . 3. The Fourier power spectrum of G22CO in F i g . 2. F i g . 4. The perturbation factor at room temperature for anhydrous In C l 2 powder. F i g . 5. The Fourier power spectrum of G22(t) in F i g . 4. F i g . 6. The perturbation factor at room temperature for anhydrous a-InCl ( f u l l c i r c l e s ) and from B-InCl at 167°C (open c i r c l e s ) . F ig . 7. The Fourier power spectrum of G22(t) in F ig . 6. TABLE 1 Frequencies (MHz) Identified In Fourier power spectra. In C l 3 In C l 2 a-In CI v* 1* - 6.6 ± 0.2 vj>l) - 11.9 t 0.3 v* 1* - 18.3 ± 0.3 v j 2 * - 14.9 t 0.2 v* 2* - 28.5 t 0.4 (2) v t - 20.2 t 0.3 v 2 - 40.2 ± 0.3 v 3 - 60.4 * 0.5 v t - 48.1 • 0.3 v 2 - 83.3 + 0.7 v 3 - 131.2 + 0.8 TABLE 2 t Asymmetry parameter, quadrupole Interaction frequency and prlnctpal component of electric field gradient calculated from data of Table 1. V Is expressed in units of 10 1 7 V/cra2, v In MHg. In c i 3 In c i 2 o-In Cl - 0.48 0.03 n m 0.07 • 0.05 TI - 0.35 • 0.01 v o > q - 38.8 • 0.7 V q m 135.0 • l . l v - 284.1 t q 2.6 zz m 1.9 • 0.4 V zz m 6.7 1.3 V - 14.2 ± zz 2.7 ,<2> m 0.18 • 0.04 v(2) <t •• 95.7 + 2.0 v(2) zz •• 4.8 0.9 - 61 -01 CL Figure 4. TIME (ns) - 64 -- 65 -- 66 -- 67 -QUADRUPOLE INTERACTIONS IN MIXED VALENCY STATES OF INDIUM CHLORIDE P.W. Martin and S.R. Dong Physics Department Univers ity of B r i t i s h Columbia Vancouver B r i t i s h Columbia Canada. V6T 2A6 and J.G. Hooley Chemistry Department University of B r i t i sh Columbia Vancouver B r i t i s h Columbia Canada. V6T 1Y6 - 68 -ABSTRACT Quadrupole interact ion parameters have been determined for InCl^ ^, InCl. „ and InCl- by using 1 1 1 l n as a probe for perturbed y-y angular corre la t ion measurements. For InCl. . a broadened spectrum i s obtained with 1 • -) estimated quadrupole interact ion frequency • 110 MHz. Three interact ions, comprising one s ta t i c and two time-dependent components, are required to f i t the data for InCl. a and InCl . . In each case the interact ions are i dent i ca l , but with d i f ferent population coe f f i c i en t s . The s ta t i c interact ion has a narrow width ( 6 - 4 2 ) and frequency v - 135.0±1.1 MHZ - 69 -1 . INTRODUCTION L i t t l e information i s avai lable on nuclear quadrupole interact ions in indium halides [1-3]. Such information can be valuable in e lucidat ing deta i l s of e lectron ic structure and in interpret ing the character i s t ics of composite systems containing these compounds. For example, from Mossbauer measurements Hooley et a l . [A] were able to characterize s i tes in graphite iron chloride by comparing quadrupole sp l i t t ing s in bulk F e C l 2 with those obtained in intercalated samples. While the Mossbauer technique i s res t r i c ted to a small number of nuc le i , NMR and NQR spectroscopy in 1 1 5 I n (95.72% abundant) and 1 1 3 I n (4.28%)is d i f f i c u l t for I - 9/2 spins. In a recent paper [5] we have reported measurements of quadrupole interactions on InCl, InC l 2 and InC l 3 using the technique of t ime-d i f ferent ia l perturbed angular correlat ions (TDPAC) of y-rays. In this communication further information is reported on the mixed valency compounds InCl , , InCl 0 and InCl^* The structure of these compounds has been previously studied by X-ray d i f f r a c t i on and in f ra - red absorption techniques [6-8]. For a nuclear spin I • 5/2, the perturbation factor for a s t a t i c quadrupole interact ion can be written to second order as [9] 3 . G 2 ( t ) - S 2 Q + I S 2 n (n) exp(- i w n 2 6 2 t 2 ) cos[u n(i , ) t ] (1) n*l where u> (TJ) • f (n) u are the quadrupole frequencies, n i s the asymmetry n n o parameter for an e l e c t r i c f i e l d gradient (efg) with width 6 and the S 2 n coef f ic ients depend on the nuclear decay scheme parameters. In an angular corre la t ion measurement G,(t) i s determined d i rec t l y [9], thus providing - 70 -information on the quadrupole interact ion parameters. The strength of the quadrupole interact ion i s usually expressed in frequency units by -e 2 Qq/h. 2. METHOD The radioactive probe for the TDPAC measurements was 1 1 3 - In, obtained commercially as ca r r i e r - f ree 1 1 1 l n C l 3 in HC1. Pr ior to preparation of the samples the required a c t i v i t y was electroplated onto indium metal f o i l (99.999% pur i t y ) . The indium chlorides were then prepared by standard techniques [1,6-8] in which the required proportions of indium metal and anhydrous indium t r i ch lo r ide were heated for several hours in an evacuated pyrex tube. Angular corre lat ion measurements were performed on the 173-247 keV Y-Y cascade in ^ C d , the daughter product following electron capture decay in l l l I n . Using a system of four Nal(TJl) detectors, deta i l s of which have been published elsewhere[10], the perturbation factors were determined from the measured correlat ions by 1'2 2 W 1 3( it,t) W 2 1,(*,t) J 2 ( t ) " 3 1 ^ " I [w l u (*/2,t) W 2 3 U / 2 , t ) ] where W ^ O . t ) refers to the angular correlat ion recorded by detectors i , j with A 2 2 corrected for f i n i t e so l id angle e f fec t s . - 71 -3. RESULTS AND DISCUSSION . I , T H I . . , In 3 [In C l 6 ] The TDPAC data for InCl, , in f i g . 1 reveals a heavily damped curve for G^Ct) which r i ses to a broad secondary maximum at about t"60 ns. For these measurements three separately prepared samples were invest igated, each y ie ld ing results for G^Ct) agreeing within experimental e r ror . An estimate of the quadrupole interact ion frequency (assuming ax ia l symmetry), using the re lat ionship e 2qQ » h v q - 20h/3t[9], y ie lds v q - 110 MHz. The t ime-integral attenuation factor in this case exceeded the hard core value of 0.2, with G2 ( » ) - 0.235 ± 0.002, indicat ing the absence of time-dependent interact ions. The re l a t i ve l y featureless character of the spectrum implies that the width parameter 6 is large, possibly a result of numerous coordination s ites avai lable to the metal ion. Since the l ike l ihood is then that G^(t) represents the superposition of several s ta t i c interact ions, no theoret ica l f i t was attempted. In* [ I n 1 1 * C l 9 ] and In 1 [ I n 1 X 1 C l 4 ] For InCl, „ and InCl. the perturbation factors were s imi lar in character but quite d i s t i nc t from InCl. F i g . 2 shows the data for InCl Q together 1 . -> l . o with a theoret ica l f i t of the form G 2 ( t ) t h - C Q G 2 ( t ) + cl exp (-t/Xy) + c 2 exp ( - t / t j ) (2 ) 2 where G,(t) i s given by eqn.(l) and J c. • 1 . To take account of the - 72 -f i n i t e resolving time of the coincidence c i r c u i t , a factor of the form exp(- 4 w 2 t 2 ) was also incorporated in the expression for G ( t ) . A I n R s s imi lar f i t was derived for the I nC l 2 data, the parameters for which are shown together with those for InCl. „ in Table 1. It i s also possible to derive v and r\ by extracting the quadrupole frequencies d i rec t l y from the Fourier transform of the autocorrelated G^Ct) data as shown in F i g . 3. The asymmetry parameter n is f i r s t obtained from the measured rat ios of v^/v^ and ^ / v ^ , following which v^ i s readi ly calculated [11]. It can be seen from Table 1, that the parameters thus obtained are in good agreement with those generated by the theoret ica l f i t . Comparison of the f i t parameters for InCl. 0 and InCl. shows that within experimental error the same interactions are observed for both cases, but with d i f ferent population coef f ic ients c . . In CI, „ ( i . e . In, ..CI.) contains approximately 11% more In atoms than a given InC l 2 l a t t i c e . Our data is consistent with a s i tuat ion in which a l l the addit ional In atoms are bound exclusively to CQ C 2 s i tes as the t rans i t ion is made from an InCl^ to an InCl Q l a t t i c e . On this model the value of c. for InCl, would be expected to decrease from 0 . 7 4 4 to 0 . 6 7 0 as the trans i t ion i s made to InCl. Q . This compares well with the value of 0 . 6 6 0 ± 0 . 0 0 6 quoted in Table 1. Several samples of the chlorides were prepared with reproducible re su l t s . In each case the s ta t i c efg interact ion, represented by the C Q term, i s well defined, with small width 6 ~ 42. This must correspond to a - 73 -s i te where e l e c t r i c a l conductivity i s high enough to ensure that e lectronic equi l ibr ium is re-establ ished after the electron capture decay, pr ior to emission of the y-rays. On the other hand, the time-dependent terms c^ and &2 may correspond to s i tes of poor conductivity in which Auger ef fects leave the e lectron ic she l l structure in undetermined charge states for the subsequent y-y cascade. A l ternat ive ly , these s i tes may be affected by d i f fus ion processes, par t i cu la r l y i f the binding energies are small. Measurements as a function of temperature should be useful in determining how s i gn i f i cant such d i f fus ion processes are. Theoretical calculat ions of the efg are being investigated with a view to elucidating s t ructura l aspects of these compounds. - 74 -REFERENCES [1] K. Wade and A . J . Bannister, "The Chemistry of Aluminium, Call ium, Indium and Thal l ium", Pergamon Texts in Inorganic Chemistry, Vo l . 12, 1975, Pergamon Press. [2] S.L. Segel and R.G. Barnes, "Catalog of Nuclear Quadrupole Interactions and Frequencies in So l ids" , Part I, U.S.A.E.C. Report IS-520 (1968). [3] G.K. Semin, T.A. Babushkina and G.G. Yakobson, in "Nuclear Quadrupole Resonance in Chemistry" (John Wiley and Sons, New York, 1975, p.213). [4] J .G. Hooley, J.R. Sams and B.V. Liengme, Carbon £ , 467 (1970). [5] P.W. Martin, S.R. Dong and J .G. Hooley, to be published in J . Chem. Phys. [6] J.R. Chadwick, A.W. Atikinson and B.G. Huckstepp, J . Inorg. Nucl. Chem. 28, 1021 (1966). [7] A.W. Atkinson, J.R. Chadwick and E. K inse l l a , J . Inorg. Nucl Chem. _3°_» 401 (1968). [8] A.W. Atkinson and B.O. F i e l d , J . Inorg. Nucl. Chem. _3°_. 3177 (1968). [9] H. Frauenfelder and R.M. Steffen, in Alpha-, Beta- and Gamma-Ray Spectroscopy, edited by K. Siegbahn (North Holland, Amsterdam, 1965, Vo l . 2, p.997). [10] P.W. Martin, S. El-Kateb and U. Kuhnlein, J . Chem. Phys. 76, 3819 (1982). [11] E. Gerdau, J . Wolf, H. Winkler and J . Braunsforth, Proc. Roy. Soc. A311, 197 (1969). TABU- 1 Fit parameters for the InCtjg and InCt 2 samples V q(Hllr) tl(ns) T2(ns) 1 6 133.210.5 7.0040.05 900450 0.1240.04 0.0440.02 l n C 11.8 135.0+1.1* 0.0540.06* c0-0.254±0.005 c,-0.66040.006 c2-0.08640.005 InCl 2 132.440.5 135.041.1* cn-0.18740.005 9.0040.05 c,-0.74440.006 810450 c2-0.06940.005 0.1340.04 0.0740.05* *Based on ratios of frequencies obtained from Fourier analysis - 76 -FIGURES F i g . 1 The perturbation factor at room temperature for InCl^ ^. F i g . 2 The perturbation factor at room temperature for I nC l^g . The so l i d l i ne represents a theoret ica l f i t based on equation ( 2 ) . F i g . 3 The Fourier power spectrum for the autocorrelated data of Figure 2 . - 80 -QUADRUPOLE INTERACTIONS IN GRAPHITE HAFNIUM CHLORIDE by P.W. Mart in,* J.G. Hooley* and S.R. Dong* •Department of Physics University of B r i t i s h Columbia Vancouver, B.C. Canada V6T 2A6 •Department of Chemistry University of B r i t i s h Columbia Vancouver, B.C. Canada V6T 1Y6 - 81 -"ABSTRACT 7The in terca la t ion of graphite by hafnium chlor ide has been inves-tigated by t ime-d i f fe rent i a l perturbed angular cor re la t ion measurements of 181 the 133-482 keV 7-7 cascade in Ta. E l e c t r i c quadrupole interact ion parameters were determined for samples prepared both by the two-zone vapour transport technique and from solut ion i n SOCl^. In the former case the presence of chlorine i s shown to be necessary for in terca la t ion to take place and the data are f i t t e d with the combination of a s t a t i c e l e c t r i c quadrupole plus time-dependent in terac t ion . When solvent i s used for the preparation a more complex spectrum i s obtained, requir ing an addit ional s t a t i c interact ion to obtain a reasonable f i t . - 82 -I. INTRODUCTION Although graphite-metal halides comprise the most numerous group of interca lat ion compounds, l i t t l e information is avai lable on Group IVb systems. 1 ' 2 In the case of HfCL^- graphite, the early work of Croft 3 reported that in the absence of chlorine approximately only 22 of HfCl H reacted with the graphite. The importance of the role of chlorine in the preparation of most graphite-metal chlorides has been established by several authors, while in other cases where chlorine is apparently not required i t is supplied ind i rec t l y by decomposition of the intercalants during the reaction process. 2 The review a r t i c l e by Stumpp1 reports the unpublished data of Frey in which stage 3 compounds of HfCL^-graphite were prepared both by the vapour transport technique in an atmosphere of chlorine and from solution in SOCL^ . In the former mode of preparation, however, a chlorine pressure of 2 atm.*4 was required to achieve the composition C^ g 7 HfCl^ 7 7 , const itut ing a stage 3 system with a repeat distance along the c-axis for the intercalant of 15.87 Angstroms. The technique of t ime-d i f fe rent ia l perturbed angular correlations (TDPAC) of y-rays was f i r s t applied to the study of graphite interca lat ion compounds (G1C) to characterize s ites in InCl j -graphite. 5 Here we report TDPAC measurements on HfC l^- graphite prepared both by vapour transport and solvent techniques. II. THEORETICAL BACKGROUND The theory of perturbed angular correlations of y-nys is well established and a comprehensive review can be found in the a r t i c l e by Frauenfelder and Stef fen. 6 The 42.5 day 1 8 1 H f nucleus decays predominantly - 83 -by p- emission to the 615 keV leve l i n 1 8 1 T a (I-1/2+). The re l a t i ve l y long l i fet ime of this state (17.6us) ensures that e lectronic equil ibrium can be established pr ior to the subsequent 133-482 keV Y - Y cascade in 1 8 1 T a via the 5/2+ intermediate state with h a l f - l i f e 10.8 ns. It is the interact ion of the quadrupole moment of this state (Q-2.5 b) with the loca l e l e c t r i c f i e l d gradient produced by the atomic environment that results in the perturbation of the angular corre lat ion between the 133 keV and 482 keV y-rays. The la t ter is given to second order approximation by V ( 9 , t ) « e ~ t / T { l • A 2 G 2 2 ( t )P 2 (cose)} where T is the intermediate state l i fet ime, Aj is determined by the spins and mult ipo lar i t ies of the trans i t ions and G^Ct) , the perturbation factor, contains the information of interest re lat ing to the quadrupole interact ions. For a randomly oriented ensemble of nuclei subjected to a s tat ic quadrupole interact ion, this can be written G J J U ) - S q (TI ) s n (n)cosu n t (1) where <*>n(il) a r e t n e quadrupole frequencies, ^m^xx~^yy^ ^ z z * s t h e a s y " " n s t r y parameter and the 8 n ( T l ) a r e calculable coe f f i c ien t s . For axia l symmetry (n • 0) in the e l e c t r i c f i e l d gradient (efg) the quadrupole sp l i t t i ng of the I • 5/2 state gives r i se to a simple harmonic series for u , with u - nu>„ n n o (n • 1, 2, 3), where is the lowest observable interact ion frequency; for the case of non-axial symmetry (n * 0) exp l i c i t expressions for u (TI) have n been given by GERDAU et a l . 7 in terms of v • eQV /h, while the method of ^ zx PRESTWICH et a l . 8 was used to determine the s (n) coef fec ients . - 84 -III. EXPERIMENTAL PROCEDURE A. Sample Preparation Graphite was exposed to intercalant in a device that could i t s e l f be transferred to the spectrometer for measurement and later returned to the intercalant system for further treatment. During a measurement, i t was of course essent ia l that non-intercalated vapor or solution be removed from the v i c i n i t y of the in terca la t ion compound so that i t s ac t i v i t y would not be detected by the spectrometer. A l l operations were performed in a vacuum system to ensure the absence of a i r and moisture. The following experimental deta i l s apply to the optimum sample shape of a cyl inder of diameter and height about 6mm. The various graphites used are described in section IV and the Hf f o i l contained 1 8 1 H f produced by neutron i r rad ia t ion at the McMaster University Nuclear Reactor F a c i l i t y . (1) Preparation from HfCL^ vapor. On the l e f t side of F i g . 1, the " jog " tube containing the Hf f o i l was f i r s t sealed to the vacuum system above i t and then to the 7 mm o.d. tube below i t containing the weighed graphite. The unit was then evacuated to 1 urn and flamed to remove adsorbed water vapor before dry Cl^ was admitted to a pressure of 200 mm. The Hf was gently heated to convert i t to HfCl^ which stays in the " jog" rather than subliming onto the graphite. When conversion was complete, the Clj pressure was adjusted to 450 mm or, in one run, to zero. The tube was then sealed off to a to ta l length of 30 cm and placed in a v e r t i c a l gradient furnace which maintained the graphite end at 300° C and the top end at a lower temperature - usually 290° C. After 24 hours the tube was removed, top end f i r s t and slowly enough so that the HfCl H so l id that forms on cooling w i l l condense only in or above the " jog" rather than on the - 85 -graphite compound that had formed at the bottom end. The tube can now be placed in the PAC detection un i t . However, the upper portion of the tube containing active HfCl^ so l id must be shielded from the detection unit or completely removed by sealing i f off with a torch about 6 cm above the sample. F ina l l y , when a l l measurements have been completed, the tube is cracked open in order to weigh the compound. From this weight and the known weight of graphite a formula C^HfCl^ is ca lculated. (2) Preparation from HfCL^ solut ion in SOCl^ . On the far r ight of F i g . 1, the Hf f o i l has been sealed into a 12 mm o.d. tube. This was attached to the upper system at B in place of the f r i t t e d f i l t e r shown in F i g . 1. It was evacuated to 1 before 200 mm of Cl^ was admitted. Gentle heat produced HfCl^ . The excess Clj was removed to a l iqu id nitrogen trap at lum and SOClj l iqu id was admitted from the dispenser at the top of the system to give a saturated solution at 20°C. The tap to the tube containing this solution was then closed so that the tube could be removed and used in place of the SOCL^ dispenser at the top of the system. A weighed graphite sample was next placed on the f r i t t e d glass f i l t e r . This unit was attached at B and the system from A to C was evacuated to 1 um. The tap below the f i l t e r was then closed and HfC l H solution was admitted to coyer the graphite for a known time (usually overnight). The tap below the f i l t e r was then opened to allow the solution to flow into the f i l t r a t e tube* The system from A to C was again evacuated to lum and the taps above and below the graphite were closed so that this section alone could be removed and placed in the spectrometer for the corre lat ion - 86 -measurements. It could obviously be returned to the dispensing system and, by appropriate evacuations and tap operations, washed with pure solvent or treated with more so lut ion. The taps and joints were l i gh t l y lubricated with f luorinated grease. The weight increase of an 0.080 gm sample of graphite was about 0.02 gm. If this was a l l HfCl^ then C^ggHfCl,, was the product. However, some of the 0.02 gm would be SOCl^ but the amount could not be determined. More important was the fact that further washing with solvent and evacuation did not change the PAC spectrum. B. TDPAC Measurements. Using a detector system comprising four Nal(Tl) spectrometers, 9 the angular corre lat ion of the 133-482 keV y-y cascade in the daughter nucleus 1 8 1 Ta was recorded to give information on the quadrupole interact ion between the intermediate state and the loca l efg. Using a 6 0 Co source with energy windows set to select the 133 keV and 482 keV y-rays in the Compton continuum, the time resolut ion (FWHM) was found to be 2.1 ns. The four time spectra, corresponding to two sets of 90° and 180° angular corre lat ions, were routed to quadrants of a multichannel analyser and the perturbation factor determined d i rec t l y from the data by c - 2 ( r ^(n.o ^ ^ i t , t ) 11/2 % 2 ( t ) 3X7 1 [ g i /2 , t ) W ^ U ^ O J " l ) where W „ ( 9 , t ) refers to the angular correlat ion recorded by detectors i , j with corrected for f i n i t e so l id angle e f fec t s . - 87 -IV. RESULTS F i g . 2 shows the perturbation factor obtained at room temperature for a sample of graphite-HfCl^ prepared from Madagascar f lakes jet -mi l led to 2um across. The composition of this sample, determined by weighing, was Cj jHfC l^. The so l id l ine is a theoret ica l f i t of the form G 2 2 ( t ) - S j G ^ t ) + C l exp(-t / -c) (2) where CQ + C j - 1 and G g ( t ) i s a s ta t i c interact ion of the form in Eq . ( l ) corrected for f i n i t e time resolut ion and for a f i n i t e width 6 in the quadrupole in teract ion. In the event that the graphite flakes may not have been t ru ly randomly oriented, the Sjn coef f ic ients were treated as free parameters in performing the f i t s . A reasonable f i t could only be achieved by incorporating also the time-dependent term with time constant t. The parameters obtained from the f i t are l i s ted in Table 1. F i g . 3, which indicates the Fourier power spectrum of the data in F i g . 2, allows d i rect determination of the quadrupole frequencies and from their rat ios another method of determining the asymmetry parameter n . 7 . By making an appropriate choice of the apodization f u n c t i o n , 1 3 the true frequencies are concentrated In re l a t i ve l y narrow regions of the frequency spectrum while the noise i s d i s t r ibuted over the entire bandwidth. From Table 1 i t can be seen that consistent agreement i s obtained for both methods of analys is. Runs were also performed at l i qu id nitrogen temperature on this sample in an attempt to elucidate the or ig in of the time dependent component. These data are also indicated in Table 1. An addit ional sample was prepared with tae 2 (am flakes which had a composition of C^HfCL^. A preparation using larger Madagascar f lakes (120 - 140 mesh) yielded C 8 Q HfC l^ . Both of - 88 -these samples generated time spectra at T • 293°K which are ind i s t ingu i sh -able from that of the Cj^HfCl^ sample. In order to ver i fy that the s ignal observed in F i g . 2 was in fact from intercalated HfCl^, TDPAC measurements were also performed on a sample prepared using Spheron 6 2700°C, a Cabot Carbon product with an area of 100 nr'gm"1 which does not form an interca la t ion compound. F i g . A shows that the data obtained from this sample with RfCl^ adsorbed on carbon black i s quite d i f ferent in character, being heavily damped with no periodic s ignal apparent. Furthermore, measurements on free Hf metal and anhydrous HfCl^ revealed d i s t i n c t l y d i f ferent frequencies in the time spectra. Detai ls on the l a t ter w i l l appear in a separate publ icat ion. An attempt was made to intercalate HfCl^ without the presence of chlorine in the react ion vesse l . The results of this are shown in F i g . 5, which indicates c lea r l y that no in terca la t ion has taken place. The results for HfCl^ intercalated by using the solvent S0C12 are presented in F i g . 6 together with the theoret ica l f i t . In this case the f i t required two s ta t i c interact ions of the type (1), plus a time-dependent interact ion, with G2 2 (t) - C l G s ( l ) ( t ) + e 2 G g < 2 ) ( t ) + C3 exp(-t/t ) (3). F i g . 7 shows the corresponding Fourier power spectrum, while Table 2 contains a summary of the results for this sample. V. DISCUSSION. In a l l cases of samples prepared by the vapour transport technique and maintained at T - 293*K, the data could be f i t t ed according to Eq.(2) with a combination of a s ingle s ta t i c (782) and time-dependent (222) in teract ion. Regardless of the type or size of graphite flakes used, the - 89 -f i t t e d parameters for a l l measurements at the same temperature were v i r t u a l l y indist inguishable. This result i s in contrast to that obtained for graphite-InCLj . 1 0 Such a s i tuat ion could result from incomplete in terca la t ion of the graphite host, with the formation of islands of intercalated material in which the efg i s wel l produced. The quadrupole frequency for the s ta t i c component was determined to be • 811112 MHz with essent ia l ly no inhomogeneous broadening (6-0). Previous experimental work, as reviewed by Sel ig and Eber t 2 , suggests that metal halides are s t ruc tura l l y undistorted upon interca lat ion from the vapor phase. Since the structure of HfCl^ i s not commensurate with that of graphite, i f s te r i c factors imposed by the graphite host l a t t i c e were important in determining the conformation of intercalated HfCl^ , one would expect to observe a d i s t r ibu t ion of quadrupole frequencies instead of a narrow l i n e . In accordance with the general character i s t ics of graphite metal hal ides, the s t a t i c s i te most probably corresponds to the + 4 oxidation state of the hafnium, since the +3 state can only be obtained under extraordinary conditions and i n a non-stoichiometric mixed phase. The question arises as to whether c, represents a true time-dependent component, or whether, for example, i t could result from a broad d i s t r ibut ion of s t a t i c s i tes with frequencies beyond the system reso lut ion. Certainly the measurements at l i qu id nitrogen temperature are consistent with the presence of a mobile species. A time-dependent component could correspond to a randomly f luctuat ing efg caused possibly by d i f fus ion or rapid rotat ion of the probe. From Table 1 the results at 77 K indicate an increase in the number of s i tes occupied by this species (from 21% to 48%) whereas i t s associated time constant become appreciably smaller » 0.4 ± 0.1 ns). At the lower temperature a decrease in charge transfer to HfCl^ molecules would presumably result in a lower fract ion of these in s tat ic s i tes . - 90 -The nature of such a time-dependent species is open to conjecture. It has been established that in the case of some metal chlor ides, chlorine not only enhances the in terca la t ion but is incorporated into the graphite with the in te rca l an t . 2 If we ascribe the time-dependent species to the formation of a HfCl^-chlor ine complex, then using Frey ' s composition 1 0 at room temperature, we obtain ( l - c x ) HfCl^ + CjHfC^ .2C12+ H f C l ^ 7 7 giving Cj^-0.19, in reasonable agreement with our f i t t ed value of 0.22±0.01 using Eq.(2). The data for the samples intercalated by dissolv ing in SOCl^ show a more complex and less well defined perturbation s igna l . The efg in this case i s modified considerably by the presence of SOClj or other complexes, result ing in two s i tes for s t a t i c interactions plus a th ird time-dependent species. The c^ component of Eq.(3) i s well defined, accounting for 19% of the s ignal , with v ^ - 3 9 3 ± 5 MHz. On the other hand the quadrupole 9 frequency for the second s t a t i c interact ion with C j -O.A l i s high enough ( v ^ 2 ^»1121±20MHz) that only the lowest frequency of the t r i p l e t is resolved in the Fourier power spectrum of F ig .7 . Using the value Q"2.51(15)b for the quadrupole moment of the intermediate s t a t e , 1 2 the efg components were calculated as shown in Table 2 for each of the samples. A theoret ica l ca lculat ion of the efg would be invaluable in e lucidat ing s t ructura l detai l s of these systems, but to our Knowledge none i s avai lable to date. The ro le of chlorine i n the preparation of these metal chloride in terca la t ion compounds i s of great interest and such studies are currently in progress. - 91 -REFERENCES 1. E. Stumpp, Mater. Sc i . Eng. 31_, 53 (1977). 2. H. Se l ig and L.B. Ebert, Adv. Inorg. Chem. Radiochem. 23_, 281 (1980). 3. R.C. Crof t , Aust. J . Chem. 9* 1 8 4 (1956). 4. E. Stumpp, private communication (JGH, SRD). 5. S.R. Dong, S. El-Kateb, J .G. Hooley, and P.W. Martin, Sol id State Communications 4_5, 791 (1983). Note errata in this paper for the quoted efg components owing to an incorrect value of Q. Using Q-0.83(16) from Ref.10, the values reported here should be mult ipl ied by the constant factor 3.024. 6. H. Frauenfelder and R.M. Steffen, in Alpha-, Beta- and Gamma-Ray Spectroscopy, edited by K. Siegbahn (North-Holland, Amsterdam, 1965), Vo l . 2, p. 997. 7. E. Gerdau, J . Wolf, H. Winkler, and J . Braunsforth, Proc. R. Soc. London Ser. A311, 197 (1969). 8. W.V. Prestwich, I.A. Cunningham, and T . J . Kennett, Hyperfine Interaction 12, 329 (1982). 9. P.W. Martin, S. El-Kateb, and U. Kuhnlein, J . Chem. Phys. 76, 3819 (1982). ~ ~ 10. S.R. Dong, J .G. Hooley, and P.W. Martin, CARBON 22, 453 (1984). 11. G. Frey, Diplomarbeit, Tech. Univ. Clausthal, 1975 (Result quoted in Ref. 2). 12. G. Netz and E. Bodenstedt, Nucl. Phys. A208, 503 (1973). 13. F. J . Harr i s , Proc. IEEE, 66, 51(1978). - 92 -FIGURES F i g . 1 Reactors for graphite with HfCl^ + Clj vapor on the l e f t and with HfCL^ dissolved in SOClj on the r ight . F i g . 2 The perturbation factor at room temperature for a sample of graphite-HfCl^ prepared by the vapour transport technique in the presence of ch lo r ine . The so l id l ine represents the theoret ica l f i t based on Eq.(2). F i g . 3 The Fourier power spectrum of G j j f t ) in F i g . 1. The frequency rat io v 2/v^ was used to determine v and n in Table 2 [square brackets]. F i g . 4 The perturbation factor at room temperature for a Spheron 6 carbon black sample with adsorbed 1 8 1 H f C l H . F i g . 5 The perturbation factor at room temperature for a sample prepared by the vapour transport technique, but in the absence of ch lor ine. This should be compared with F i g . 2. F i g . 6 The perturbation factor at room temperature for a sample of graphite-HfCl^ prepared from solution in S0C1 2. In the early part of the spectrum the error bars are contained within the size of the data points. The so l id l ine represents the theoret ica l f i t based on Eq. (3). F i g . 7 The Fourier poyer spectrum of ^ ( t ) in F i g . 6. The frequency r a t i o s v j 1 ' /v^ ' and v^^/v^ were used to determine v and n in Table 2 [square brackets]. Of the second interact ion G^ ^ in Eq. (3) only Vj was resolved. TABLE 1 Quadrupole interaction parameters at room and liq u i d nitrogen temperatures for grsphtte-HfCL, prepared by the vapour transport method* The interactions are defined In Eq. (2). Where possible v. and t)( square brackets] were also determined from the frequency ratios In the Fourier spectrum of Fig. 2. INTERACTION X W«l (Mlij) 1 & (1018V|cm2) <"•) y o 293K * exp(-t/t) 78(2) 22(1) 813(10) {818(10)| 0.21(2) [0.20(3)1 0.00(2) 1.20(8) 0.9(1) y o 77K * exp(-t/t) 52(2) 813(10) 0.22(2) 0.00(2) 1.20(8) 48(2) — — — — 0.4(1) TABLE 2 Quadrupole Interaction parameters at room temperature for graph!te-HfCt^ prepared from solution In SOCl^ . The interactions are defined in Eq. (3). Where posslhle and n. (square brackets] were also determined from the . frequency ratios In the Fourier spectrum In FIR. 7. INTERACTION X v1 (Mil,) n 6 (lO^Vlcm 2) (ns) C 8< l>(t) 19(1) 393(5)) 0.74(5) 0.04(1) 0.58(3) (393(5)] 10.75(6)1 C„<2>(t) 41(2) 1128(20) 0.04(3) 0.10(2) 1.67(10) — exp(-t/t) 40(2) — — — — 0.93(3) FIGURE CAPTIONS - 95 -F i g . 1 Reactors for graphite with HfCl^ • C l 2 vapour on the l e f t and with HfCl^ dissolved in S0C1 2 on the r i gh t . F i g . 2 The perturbation factor at room temperature for a sample of graphite-HfCl^ prepared by the vapour transport technique i n the presence of ch lor ine. The so l i d l i ne represents the theoret ica l f i t based on Eq.(2). F i g . 3 The Fourier power spectrum of G 2 2 ( t ) i n F i g . 2. The frequency r a t i o *> (p was used to determine v and V i n Table 2 [ square brackets ] . 2 1 q F i g . 4 The perturbation factor at room temperature for a Spheron 6 carbon 181 black sample with adsorbed HfCl^. F i g . 5 The perturbation factor at room temperature for a sample prepared by the vapour transport technique, but in the absence of ch lor ine. This should be compared with F i g . 2. F i g . 6 The perturbation factor at room temperature for a sample of graphite-HfCl^ prepared from solut ion in S0C1 2. In the ear ly part of the spectrum the error bars are contained within the s ize of the data points. The so l i d l i n e represents the theoret ica l f i t based on Eq.(3). F i g . 7 The Fourier power spectrum of G 2 2 ( t ) in F i g . 6. The frequency rat ios v J"^^/v ^ and v , ^  were used to determine v and n in Table 2 (2) q (2) I square b r a c k e t s ] . Of the second interact ion G g i n Eq.(3) only v j was resolved. - 96 -V A C , C e 2 . AIR \ / TOTAL VOLUME of 8 c m 3 30 cm Hf.1 GRAPHITE DRYING TUBE ft S0C£. A/ / Hf*C£, in soce 2 2 si ), VAC. "AIR \ / M FRITTED FILTER VAC. FILTRATE TUBE Figure 1. 0 I I Hf.*F0IL T>3~ TEFLON TAP 3 -4 ,mm l 0 / | 9 JOINT - 97 -- 99 -- 100 -- 101 -- 103 -AFTER-EFFECTS INVESTIGATIONS IN MIXED-VALENCE INDIUM CHLORIDES * . . . C P . Massolo , J . Desimoni, A.G. B i b i l o n i , L.A. Mendoza-Zelis L.C. Damonte, A.R. Lopez-Garcia Departamento de F i s i c a , Facultad de Ciencias Exactas Universidad Nacional de La P lata, C.C.67, 1900 La P lata, Argentina and P.W. Martin, S.R. Dong and J .G. Hooley Department of Physics, University of B r i t i sh Columbia Vancouver, B.C. V6T 2A6, Canada We present a new analysis of a series of measurements of the s tat ic and time-dependent interactions in the angular corre lat ion of y-ra/s from l i : C d fol lowing electron capture in 1 1 3 , I n , in In C ^ , In C ^ . In g, In CI , and In C£ compounds. The time-dependent hyperfine interact ion ar i s ing from the e lectronic rearrangements which follow the nuclear e lec -tron! capture is analysed employing an adequate parametrization. In this way, a f ter ef fects 3nd the loca l e lectronic properties of Che compounds are re lated. Results indicat ing that the loca l electron a v a i l a b i l i t y at the impurity s i tes increases with the In concent are discussed. The measured s ta t i c quadrupole interactions are tentat ively assigned to d i f ferent indium oxidation states or c ry s ta l l i ne s i tes . - 1 0 4 -1. INTRODUCTION In Che lasc few years the Time D i f f e ren t i a l Perturbed Angular Corre lat ion (TDPAC) technique has been increasingly applied to study semiconductors and insulators. One of the challenging aspects of these investigations concerns the so-cal led a f teref fects (AE). (By e lec t ron-capture a f teref fects one usually designates a l l the e lectron ic rearrangement processes which follow a nuclear decay by electron capture). In a recent study of the In203 system 1 * 2 a new picture of AE processes was presented where only the holes trapped in levels above the valence band can or ig inate the time-dependent perturbation of the angular corre lat ion associated with AE. In the proposed model, the complete relaxat ion p r o c e s s would be as follows : the primary hole in the inner-she l l moves to outer shel ls creating new holes in the outermost shel ls by Auger emission. A l l the holes f i n a l l y located in the valence band w i l l d i f fuse away. Even f o r low hole mobil ity this d i f fus ing process occurs very rapidly and so the f luctuat ing f i e l d due to the relaxation of these holes w i l l disappear i n times of the order of 10 - 1 1* - 10~ 1 5 sec. Hence, no perturbation of t h e angular corre lat ion should arise from holes i n i t i a l l y located inside che valence band. In the part icu lar case in which the nuclear electron c a p t u r e decay creates an impurity in the compound, impurity levels w i l l a p p e a r i n the band gap. If one or more holes are trapped in the impurity center, their d i f fus ion in the valence band is no longer possible and so chey may l i ve (before recombining with electrons) long enough to perturb the a n g u l a r corre la t ion. If an appropriate perturbation factor is used to analyse the angular corre lat ion - i n part icu lar a function taking account of the on-off c h a r a c t e r of the f luctuat ing interact ion and re lat ing che interact ion parameters with the l i fet ime of the holes trapped in the impurity center- the measu-rement of AE provides information about the local electron a v a i l a b i l i t y ( l e a ) at the probe's s i t e . When the electron supply to the impurity level i s mainly originated in the thermal excitacion of valence electrons, che posit ion of the impurity leve l can be obtained through AE measurements as a function of temperature. - 105 -A l l che above considerations suggest that the interest of these s t u d i e is increased by the po s s i b i l i t y of developing a new appl icat ion of che TUI'A technique : extremely loca l ized e l e c t r i c properties of semiconductors and insulators can be studied through AE measurements. ( I t is worth mentioning that in a recent study of AE in s i l i c o n , M. Deicher et a l . 3 seem co be sharing our approach). In this framework, the indium chlorides constitute a very interescing system : only In*** i s present in InCJt.3, while there is only In* in InCJL and, through the mixed valency compounds, d i f ferent proportions of both can be found. Using 1 J , 1 In for the TDPAC measurements, the dependence of die probe's e lectronic configuration on AE can be studied. F i r s t measurements of the quadrupole interact ion in indium chlorides by the TDPAC technique using 1 1 J I n have been reported previous ly 1 * ' 5. In che analyses of these experiments, no attempt was made Co account spec i f i c a l l y f o r a f te re f fec t s . For the data on InCJ&2 and I n C £ i . s , however, i t was found that the inc lus ion of time-dependent cerms of che form e x p(-c/i) greacly improved che f i t s 5 . We present in this paper a re-analys is of the complece sec of daca (InCC . 3 , InC£2, I n C £ i . a , I n C £ i . 5 and InC£) in cerms of che model proposed i n Ref. 2 and discussed above, using a paramecrizacion which r e l a c e s che cime-dependenc hyperfine inceraction wich local e lecc r i c properties of the compound. We found chat the AE perturbation of the angular corre lat ion decreases wich decreasing amount of In*** in the compound. The mean l i f e t i m e of a hole trapped in the cadmium impurity leve l decreases from 50 ns i n I n C i j to 3 ns in InCZi . 5 . The loca l electron ava i l ab i l i t y dependence wich I n ^ 1 * abundance is discussed. The scat ic component of the hyperfine interact ion is also analysed and the measured frequencies are tentat ively assigned to the differenc indium oxidation states or c rys ta l l i ne s i tes in each compound. - 106 -EXPERIMENTAL The samples were prepared and che TDPAC measurements performed ac che Physics Departmenc of the University of Bricish Columbia. The l x l I n accivi commercially obcained as carrier-free l l l I n C & 3 in hydrochloric acid, was eleccroplaced onto f o i l s of high purity indium metal (99.999 Z) which were then, placed in evacuated Pyrex tubes for chlorination Co form InCJ£3. The InCi, InC£2, InCii.a and I n C i i . 5 were prepared by standard techniques. In each case, required proportions of indium metal and anhydrous indium t r i c h l o r i d e were creaced for several hours in evacuated Pyrex tubes. More details of the samples preparation can be found in references A and 5. The TDPAC measurements were made using the well known 173-247 keV y-y cascade in 1 1 1 C d , created by electron capture of X 1 1 l n wich a 2.8 d h a l f - l i f e . Decails of the electronic set-up and detector system, cooipribin four symmetrically places NaI(T£) d e t e c t o r s have been published ei^e-where 6. The perturbation factors were determined from the measured dacj by G f ^ t ) 3 A 2 : 1 / 2 'Wi 3(TT,C)W 2„(TI,C) - 1 where W„(8,c) refers Co che angular correlation recorded by dececcors i , j wich A 2 2 correcced for f i n i c e solid angle effects. The form of the perturbation factor G 2(c) depends on che nacure and time dependence of che f i e l d s acting on the nucleus. Theoretical functions G 2 ' l e 0 r ( c ) , folded wich a gaussian prompc to account for cime resolution effects were f i t t e d to the experimental c f X p ( c ) factor. The theoretical perturbation factor Gf^ e° r(t) must Cake inco account cwo different hyperfine interactions : the s t a t i c one due to che coupling i Che nuclear quadrupole moment and che e l e c t r i c f i e l d gradient of che chargi d i s t r i b u t i o n in the l a t t i c e and the time-dependent interaction associated wichAZ. The perturbation factor corresponding to the stacic inceraction w i l l be of che form : - 107 -G!(C) 3 I i-1 3 I n=0 s . exp(- 6. n i w t) cos (OJ n. n. l l t) (2) Here f\ are the re la t i ve fract ions of nuclei that experience a given pertur-bation. The frequencies ii) are related by w • F (n)v A to the quadrupole ^ n n n y frequency v„ » eQV /h. The coef f ic ients F and s_ are known funct ions 8 ^ Q x zz n n of the ax ia l asymmetry parameter n defined by n - (V^x - v y y ) / v 2 2 > "here V.. are the pr inc ipa l components of the EFG tensor. The exponential f u m i i o n accounts for a Lorentzian frequency d i s t r ibut ion of re la t ive width •> around u . Concerning the time-dependent interact ion, use has been made of die n . . 8 perturbation factor derived by Baverstam et a l on the basis of the Abraham „ and Pound theory. The authors in ref. 8 have shown that the basic assumptions of the Abragam and Pound Theory are sa t i s f ied in AE processes and derived a modified perturbation factor by taking into account the fact that the f l u c -tuating interact ion is turned-off when the atomic neutra l i ty is reached. To do so, two simplifying assumptions were made : i) The probabi l i ty for an atom to reach i t s ground state in the time in terva l t, c + dt is given by : P (t) dt - X„ exp (-X t) dt (3) g ° S this assumption implies that the complete atomic recovery process is characterized by a single exponential recovery time X g i i ) The mean interact ion strength, characterized by the Abragam and Pound relaxation constant X r , remains constant during the time the hole (or holes) i s bound to the probe. Moreover, for atomes in the ground state, the interact ion is supposed to be much weaker. Thus, the perturbation factor can be seen as an average of one ar i s ing from atomes that reach their ground state before - 108 -the time t and those chat reach i t a f ter the time t. It can be wrin.cn then as : A A t s r G|(t) - + exp(-(A + X )t) U ; A + A X • A 8 r g r g r It i s interest ing to note that the introduction of the on-off charac-ter of the f luctuat ing interact ion strongly changes the asimptotic behaviour of the Abragam and Pound perturbation factor which leads to v a n i s h i n g G 2 (t) for t - 0 0 When the complete atomic recovery is achived, the hyperfine i n t e r a c t i o n becomes s t a t i c . Since this interact ion is weaker than the time-dependent one, . G 2 (c) can be expressed, for the combined s tat ic and time-dependent i n t e r a c -t ions, as the product G 2 h e ° r (t) - G| (t) * G.2 (t) (5) For pure s tat ic interactions we have used the perturbation faccor defined in eq. 2. In order to study the electron supply to Cd atomes located in d i f f e r e n t s i tes of the l a t t i c e , we analysed a l l che data using a theoret ical f u n c t i o n which a l ternat ive ly associates the time-dependent interact ion wich only one of the s tat ic components. The results did not show any s ign i f i cant unequivaience of the electron supply to codnium probes ar i s ing from indium ions in d i f f e -rent oxidation states. This i s why, we present in the fol lowing, the r e s u l t s of the data analysis using the theoret ical G2(t) function of eq. 5 , d e s c r i b -ed above, where the same set of time-dependent paramecers is assumed f o r both oxidation states of the indium atomes. - 109 -3. RESULTS AND DISCUSSION Figures 1-3 show the G2(t) perturbation factor measured for che di f ferent indium chlorides studied. Solid l ine represent the best theo-r e t i c a l f i t s based on eq. 5. The InCJU compound shows at room temperature a strong time-dependent interact ion which is also present when investigated at 203"C (Fig. la) and l b ) . The main di f ference between these two spectra (notice the d i f ferent times-scales) i s the pos i t ion of the second maximum. It changes from **- 100 ns at RT to around 140 ns at 203° C. This is well reproduced by a change in the parameters characteriz ing the s ta t i c part of the i n t e r a c t i o n (see Table I). The parameters X and X associated with the time-dependent interact ion are quite s imi lar at both temperatures (see Table I) and agree within errors. This means, in the assumed picture, that the promotion of electrons from the valence band to the i n i t i a l l y empty leve l do not change s i gn i f i cant l y in the investigated temperature range. The same temperature dependence of AE paramecers was found in che IruO: system between RT and 200° C 1 ' 2 . According to the reported c ry s ta l l i ne structure 9 for room temperature, there is just one c ry s ta l l i ne s i te for indium atoms in InC£.3, and t h e r e f o r e only one s ta t i c interact ion should be expected. However, the best f i t s to the data were obtained with two well defined s ta t i c quadrupole i n t e r a c t i o n s characterized by the parameters : 'Qi = 28.7 3.0 MHz "Q2 40.8 3 . 0 Hi 0.29i3 n2 1.00 To our knowledge there is no avai lable information about phase t r a n s i t i o n s in this compound. The TDPAC results at room temperature and 203° C do not show s ign i f i cant changes to conclude that such a transformation took place. In par t i cu lar , the frequency that characterizes the more populaced s i te, chat is v. , and which should be presumably assigned Co che indium s i c e Qi in I nC£ 3 remains unchanged although chere is a change in che m value (see Table I). - 110 -In the I n C £ 2 , InC£i.a and InC£i.s compounds, the indium atoms are present in boch I and III oxidation states. The ionic structure r e p o r t e d i n ' the l i t e r a t u r e , I nC£ 2 : In 1 ( I n * * * C £ „ ) , InC£i. 8 : In* ( I n 1 1 1 ^ ) and I n C £ i . 5 : I n * ( I n * * * C £ s ) suggest that at least two s ta t i c quadrupole i n ter -actions should be present in the TDPAC spectrum of these compounds, a s s o c i d t with the two oxidation states. The stoichiometric f rac t ion of In* i n c r e a s e s from 50 1 for I nC£ 2 to 75 Z in I n C £ j , . 5 , being 60 2 in I n C £ i . a 1 °> 1 1 • The In*** ions in a l l these compounds are octahedrally coordinated by 6 C£ ions. These octahedra are isolated in I n C £ i . s while in I nC£ 2 they would share e i ther two edges to form an i n f i n i t e chain or share four corners to give an i n f i n i t e sheet. F ina l l y in InC£ i.8 the complex anion consists of two such octahedra sharing a face and would have Dsh symmetry. The TDPAC spectra of the above mentioned compounds (see F ig . 2 ) , show the presence of two s ta t i c quadrupole interact ions, namely h i i and hi2, i n Table II. An addit ional low frequency is c lear l y evident in the cases of I nC£ 2 and InCHi.a- In the case of InC£i.s compound, since the f i t t e d popu-lat ions agree well with the fractions of In* and In*** ions p r e d i c t e d from che ionic structure, we assigned h i i and h i 2 to the quadrupole i n t e r -actions ac In*** and In* s i tes respect ively. The matching between che f r ac -cions f i , quoced in Table II for I nC£ 2 and InCi i . s , and chose c o r r e s p o n d i n g co In*** in che same compounds reported in the l i t e ra tu re , lead us co ascrib h i i to che same kind of ions. The change in the v and m values from \>„ s 137.5 MHz, m s 0.08 to v. =• 36.6 MHz, Hi s 1 • p r o b a b l y r e f l e c t s Qi Qi the d i f ferent d i s tor t ion and/or linkage of octahedra in the c o r r e s p o n d i n g structures. Once we assigned the interact ion corresponding to In*** ions, die remainder should be attr ibuted to In*. Hence, the addit ion of c o e f f i c i e n t s f 2 and f3 should amount to the stoichiometric f ract ion of In* in these compounds. This can be seen in Table II, where this sum is i n d i c a c e d . There is l i c c l e avai lable informacion on che crysca l l ine structure of these anhydrous chlorides of indium as to interpret the two quadrupole i n c e r a c c i o n h i 2 and h i 3 , present in I nC£ 2 and I n C £ i . s . An NMR study recorded in the molten state of I n C l 2 1 2 suggests that i t i noc, as generally supposed In*(ln***Cl,,) but ( I n I I C l 2 ) 2 . Our results support hypothesis of a mixed valence structure for I nC l 2 . - I l l -Concerning the time-dependent interact ion we can see an incremenc of che X parameter from 0.020 t O.OOA n s - 1 to 0.34 £ 0.20 ns " 1 goiny from g r InCi; to InCJLi.s. This indicates mean recovery times of 50 ± 10 ns, 9 z I and 3 r 2 ns for I n C £ 2 , I nC£ i , a and InCJLi.s respect ively. The f i t t e d values of X^ are given in Tables I and II. This parameter takes into account a l l the f luctuat ing interactions which are present during the atomic recovery. Since they may include sp i n - l a t t i ce relaxat ion and e lectronic disturbances in the probe's surroundings, i t i s expected that i t takes d i f ferent values for each compound but remains unchanged in che two measurements for InCi.3 (23 and 203° C). On the other hand, a very crude est imation 2 assuming a pure magnetic interact ion produced by a hole in the 4 d she l l of the Cd atom gives X^ » 0.2 n s " 1 , consistent with che measured values. The last spectrum to be considered corresponds to cx-InC£ and is shown in F i g . 2. The a-phase of indium monochloride is found to be a deforced NaCZ structure with space group P2i3. Two crys ta l l ine s i tes are avai lable for indium atoms, namely 12(b) and 4(a) with re la t i ve populations 3 : 1 . A least squares f i t of eq. 3 to the data revealed the presence of three s ta t i c quadrupole interact ions, one of which has the parameters characte-r i s t i c of InCij (see Table III). The existence of some InC£; in our sample is not discarded due to the preparation procedure. The other two hyperfine interactions v • 132.81 .0 MHz n. 1 5 0.9 and v - 288.5s MHz n - 0.3b} Qb D ^a . should be assigned to indium atoms in positions 12 (b) and 4 (a) respect ively. No time-dependent interact ion was f i t t ed for ct-InC£. F ie ld gradient ca lcu-lations using the point charge approximation and the pos i t ion of the atoms reported by Van der Berg in Ref. 13 gave n^*1 " 0 . 7 6 and » 0.14, not very d i f ferent from the f i t t e d ones. It i s interest ing to remark that in most of the indium chlorides s tud i -ed by TDPAC, we do not f ind a correspondence between the number of efg in te -ractions observed and the number of inequivalent s i tes reported in the crys -tal lographic l i t e ra tu re . We can f ind no obvious reason for th i s . In view of the precautions taken in the sample preparation we are inc l ined to dismiss thra p o s s i b i l i t y of contamination. However, the pos s ib i l i t y exists as to imperfect c r y s t a l l i z a t i o n of the samples. Future experiments are being considered to check these po s s i b i l i t i e s with X-ray and TDPAC techniques as a function of temperatures. - 112 -As mentioned in Sec. 2, in the model we used to analyze the TDPAC spectra, T = A - 1 provides the mean l i fet ime of a hole trapped in the cadmium impurity l eve l . The results indicate that the loca l electron a v a i l a b i l i t y at the impurity s i te increases with the content of In 1 . To our knowledge there is no avai lable information on the band struc-ture of this family of indium hal ides. It i s reasonable to suppose, in the case of InC i j , that the valence band i s predominantly constituted by orb i ta l s from che halogen, wich a 5s conduction band belonging to the indium. A d i rect consequence of our results would be to leave aside the naive picture in which the 5s electrons of each In* ion incorporate t o che conduction band leaving the band structure mostly unchanged. I f i t was so, the electron density in the conduction band would be high enough, even for InC£.2» to confer a mecal l ic characcer to the compound and t h e r u i u r c no AE would be observable. I c seems l i k e l y chat as the f ract ion of indium increases, t h e band structure is modified by the introduction of 5s metal l ic orbicals co the valence band, producing a narrowing of the band gap. The p o s i t i o n or the impurity leve l in the band gap of the semiconductor w i l l change, approaching to the top of the valence band with the increase o f che In 1 f rac t ion . Consequently, exc i tat ion of electrons from the valence band to the impurity leve l may occur at room temperature. This effect w i l l be enhanced by increasing the amount of In* and would explain che r e d u c t i o n of che hole l i fet ime when going from InCi 3 to InCA. However, further experiments, in part icu lar low temperature TDPAC m e a s u r e m e n t s , would be of much help to test these ideas. ACKNOWLEDGEMENTS The auchors are graceful Co CONICET and CICPBA, Argencina f o r p a r t i a l economic support. C.P.M., J .D. , A.G.B. L.M.Z. and A.L.G., are members of Carrera del Investigador C ien t i f i co CONICET. LD. is a fellow of CONICET. * Present address : Inst itut de Physique Nucleaire - B.P. N ° I 91A06 ORSAY - France. - 113 -REFERENCES 1 - A.G. B i b i l o n i , J . Desimoni, C P . Massolo, L. Mendoza-Zelis, A.F. Pasquevich, F.H. Sanchez and A. Lo'pez-Garcia Phys. Rev. B 29 (1984) 1109 2 - A.G. B i b i l o n i , C P . Massolo, J . Desomoni, L. Mendoza-Zelis, F.H. Sanchez, A.F. Pasquevich, L. Damonte and A. Lo'pez-Garcia Phys. Rev. B 32 (1985) 2393 3 - M. Deicher, G. Grubel, E. Recknagel, Th. Wichert and D. Forkel 11th Int. Conf. on Atomic Co l l i s ions in Sol ids, Washington, DC USA, August 4-9, 1985 4 - P.W. Martin, S.R. Dong and J .G. Hooley J . Chem. Phys. 80 (1984) 1677 5 - P.W. Martin, S.R. Dong and J .G. Hooley Chem. Phys. Let t . 105 (1984) 343 6 - P.W. Martin, S. El-Kateb and U. Kuhnlein J . Chem. Phys. 76 (1982) 3819 7 - L. Mendoza-Zelis, A.G. B i b i l o n i , M.C Caracoche, A. Lopez-Garc J.A. Martinez, R.C Mercader and A.F. Pasquevich Hyp. Incerac. 3 (1977) 315 8 - U. Baverstam, R. Othaz, N. de Sousa and B. Ringtrom Nucl. Phys. A186 (1972) 500 9 - R.W.G. Wyckoff In Crysta l Structures (Interscience, New York, 1968) - Vol. I, - 114 -10 - J.R. Chadwick, A.W. Atkinson and B.G. Huckstepp, J . Inorg. Nucl. Chem. 28 (1966) 1021 11 - A.W. Atkinson and B.O. F i e l d , J . Inorg. Nucl. Chem. 32 (1970) 3757 12 - C. Margherit is, Z. Naturforsch. 39a (1984) 1112 13 - J.M. Van Den Berg, Acta Cryst. 20 (1966) 905 - 115 -TABLE CAPTIONS Table I - F i t ted values of the parameters character iz ing the observed hyperfine interact ions in InCia compound. F i t s were performed according to eq. 5, leaving a l l the parameters free ; then, some of them were f ixed to get more precis ion in the r e l e -vant ones. No errors are quoted for the f ixed parameters. Table II - Idem Table I for I n C £ 2 , InCii.a and InCi i.s compounds. Measurements were carried out at room temperature. 1-oiii i l l - Idem Ta'blu 1 lul' u - iuCX.. liiw V u l u c s quoted uiidel t correspond to the renormalized f. (see cext). T f \) n 6 f Tm e a s . 1 Q i i 1 2 (°C) (%) (MHz) <% ) ( % ) 23 82 28.7 0.29 0 18 8 2.8 13 6 203 71 29.7 0.77 0 29 6 2.7 2 1 4 Q2 (MHz) n 6 A A 2 2 8 r (%) (ns" 1) (ns - 1 ) 40.8 1. 0 0.029 0.038 2.8 1 6 8 31.2 0.31 0 0.033 0.045 2.7 2 J -» hi i f.(%) l V (MHz) Qi n. i InCl 2 5S 1 3 7 ' 2 2 . 5 0.08 h l 2 16 3 78.5 2.5 1.0 hi 3 30 5 1 . 9 1.0 0.0 InCl 1 . 8 hi 1 46 3 138.52 5 0.09 hi 2 14 3 97.6 2.5 0.64 h £ 3 40 5 1 . 9 1.0 0.0 InCl 1.5 hi 1 25 8 36.6 2.5 1.0 h i 2 75 9 108.5 2.5 0.34 6.(Z) f + f A (ns ) A ( n s " 1 2 3 8 r 0 1 0 0.020, 0.060 2 4 6 0 46 1 1 5 0.11 0.11 3 1 1 0 54 58 10 0.34 0.09 0.28 75 2 0 2 4 - 118 -Table III f(%) v (MHz) n 6(%) f ( % ) Q i r s i te a 55 5 1325 0.9 6 6 ^ 75 s i te b 18„ 288 0.36 1, 25 2 3 1 1 19, 28.7 0.31 5 2 2.6 - 119 -FIGURE CAPTIONS Figure 1 - TDPAC spectra of I nC l 3 , measured at (a) 2 3 ° C , and (b) 2u.rc. The so l i d curves are least-squares f i t s of eq. 5 to tin.- d a t a . Insert i n Fig.1(a) i s shown the rat io G 2 E X P ( t ) ( L ) i i i L e d for the InC l 3 (23°C) spectrum. The fact that both and the f i t t e d G ^ values f a l l rapidly to cero makes the r a t i n highly dispersive above 35 ns. Nevertheless, the c h a r a c i > r i s t i < : exponential shape of the f luctuat ing component c a n c J e a r l y he seen and i s very well reproduced by the f i t t e d C^ 'u ) valm.-a (using eq. 4) shown in dashed l i ne s . The b a r s only mu< account s t i t i s t i c errors of the experimental d a t a . Figure 2 - TDPAC spectra, measured at RT, of (a) InCl , (h) and (c) InCl - . The so l id curves are l e a s t - s q u a r e s l i i s of eq. 5 to the data. Figure 3 - TDPAC spectrum of a-InCl measured at RT. The s o l i d c u r v e i s a least-squares f i t of eq. 2 to the data. - 120 -Figure 1. Figure 2 . Figure 3 . - 123 -Concluding Remarks Although the ear l i e s t publications included in this thesis date to 1984, work on the GICs began in 1982, and at this wr i t ing, i s s t i l l ongoing. It was in the summer of 1982 that I was f i r s t introduced to Dr. J .G. Hooley and this remarkable class of compounds. That f i r s t year was spent reviewing the l i te ra ture and becoming fami l iar with the synthesis of these compounds, a process of which Dr. Hooley was largely an or ig inator. By the summer of 1983, the f i r s t series of measurements on graphite-indium chloride had been completed and the results were presented at the Sixteenth Carbon Conference in San Diego. Since then measurements have been done on graphite-hafnium chloride and the various indium chlor ides. These are detai led In the enclosed publ icat ions. Also, much work has been done in ref in ing the data analysis of TDPAC spectrums, especia l ly with regards to accounting for the effects of the f i n a l time resolut ion. This issue has not been adequately dealt with in the context of today's much-improved instrumentation. In the interim, I have changed careers from experimental physics to medicine, a move which has delayed the completion of this thesis, but which has afforded me the luxury of retrospection. Therefore, I would l i ke to conclude by reviewing b r i e f l y some of the insights gained from the work on GICs and the indium halides and also comment on future direct ions in these and related invest igat ions. - 124 -In the work on GICs, compounds with stage numbers of two and higher were studied. The intercalant was consistently found to exhibit a f ixed number of efgs, with only the proportion of each varying with the stage number. The p o s s i b i l i t y that the intercalant might assume d i f ferent configurations as a result of varying degrees of charge transfer adds another l eve l of complexity to the analysis of these compounds. Further interpretat ion of data has been hampered by the lack of theoret ica l calculat ions of the efg. However, these are technica l ly challenging in view of the existence of more than one type of interact ion for a given intercalant and the lack of data about how the intercalant i s d istr ibuted in the plane of the graphite layers. Ideal ly, one should perform measurements and calculat ions on a system such as the GIC of bromine. Further investigations of GICs using TDPAC would Include the effect of temperature, a more precise de f in i t i on of the re lat ion between the stage number and the proportions of each interact ion found for a given interca lant , co- intercalated systems such as aluminum chloride and indium chlor ide, and measurements on a stage 1 compound. I ant ic ipate that rad ica l changes w i l l be observed in the efgs as one goes from a stage 2 to a stage 1 compound. In the work on the indium chlorides, there was found to be no correspondence between the number of efgs observed and the number of unequivalent s i tes reported from crystal lographic measurements. This may be due to imperfect c r y s t a l l i z a t i on of the samples. These compounds w i l l be reinvestigated with the use of x-ray crystallography as a means of qual i ty control In the sample preparation. - 125 -As I have embarked on a career in medicine, my Involvement in TDPAC experimentation w i l l be diminished. However, I hope to continue to part ic ipate in some capacity as I have acquired friendships and associations that I am unwil l ing to re l inquish and my interest in experimental science Is undiminished. - 126 -References 1. H. H. Rinneberg, Atomic Energy Review, 172_ (1979), 477. 2. J . J . Vargas, MTP International Review of Sciences, 8_ Radiochemistry, ed. A. G. Maddock, 45. 3. E. M. Kaufman, R. J . Vianden, Review of Modern Physics, 51_ (1979), 161. 4. M. Va l i c , D. LI, Will iams, Journal of Physical Chemistry and Sol ids, 30 (1969), 2337. 5. W. Keppner, W. Korner, P. Heubes, G. Schatz, Proceedings of the 5th International Conference on Hyperfine Interactions, Ber l in (1980). 6. P. Heubes, G. Hempel, H. Ingwersen, R. K e i t e l , W. Kl inger, W. Loe f f l e r , W. Witthuhn, Hyperfine Interactins Studied in Nuclear Reactions and Decay ( P roc Conf. Uppsala, 1974, E. Karlsson, R. Loe f f l e r , eds.). 7. J . Christiansen, P. Heubes, R. K e i t e l , W. Kl inger, W. Loe f f l e r , W. Sandner, W. Witthuhn, Z. Phys., B 24 (1976), 177. 8. E. Gerdau, J . Birke, H. Winkler, J . Baunsforth, M. Forker, G. Netz, Z. Phys. 263 (1973), 5. 9. J . A. Wilson, F . J . diSalvo, S. Mahajan, Adv. Phys. 24_ (1975), 117. 10. T. Butz, A. Vasquez, H. Sa i tov i tch, R. Muhlberger, A. Ler f , Physica 99B (1980), 69. 11. T. Butz, H. Sa i tov i tch, A. Lerf, H. D. Zagerka, R. Schollhorn, Phys. Le t t . , 67A (1973), 74. 12. M. S. Dresselhaus, G. Dresselhaus, Adv. Phys. 3£, (1981), 139. 13. M. H. Cohen, F. Reif, Sol id State Physics 5_ (1959), 321. 14. E. Matthias, W. Schneider, R. M. Steffen, Phys. Rev. 125 (1962), 261. 15. C. P. S l i ch ter , Sol id State Science 1, 2nd ed. (1978), Springer-Verlag "Pr inc ip les of Magnetic Resonance". 16. G. K. Semin, T. A. Babushkina, G. G. Yakobson, "Nuclear Quadrupole Resonance i n Chemistry" (1975), John Wiley & Sons, Inc. 17. R. L. Cohen, Applications of Mossbauer Spectroscopy, R. L. Cohen, ed. , Vo l . 1 (1976), 1. - 127 -18. H. Fraunfelder, R.M. Steffen, "Alpha-, Beta-, and Gamma-ray Spectroscopy", Vo l . II, K. Slegbahn, ed. (1968), 997. 19. A. H. Wapstra, "Alpha-, Beta-, and Gamma-ray Spectroscopy", K. Siegbahn, ed. , (1968), 539. 20. N. Kaplan, S.N. Sharma, D. LI. Williams, Canadian Journal of Physics, 58 (1980), 900. LIST OF PUBLICATIONS 1. Quadrupole I n t e r a c t i o n Studies of Graphite Indium Ch l o r i d e Using T i m e - D i f f e r e n t i a l Perturbed Angular C o r r e l a t i o n s , S.R. Dong, P.W.Martin, J.G. Hooley, Carbon Vol.22, No. 4/5, pp.453-458,1984. 2. Quadrupole I n t e r a c t i o n s i n Ch l o r i d e s of Indium, P.W. M a r t i n , S.R. Dong, J.G. Hooley, J . Chem. Phys. 80(4),15, pp.1677-1680,15 February 1984. 3. Quadrupole I n t e r a c t i o n s i n Mixed-Valency States of Indium C h l o r i d e , P.V. M a r t i n , S.R. Dong, J.G. Hooley, Chem. Phys. L e t t e r s , Vol.105,No.3,pp.343-346,16 March 1984. 4. Quadrupole I n t e r a c t i o n s i n Graphite-Hafnuim C h l o r i d e , P.W. Ma r t i n S.R. Dong, J.G. Hooley, Phys.Rev.B, Vol.33,No.6,pp.4227-4232,15 March, 1986. 5. A f t e r e f f e c t I n v e s t i g a t i o n s i n Mixed-Valence Indium C h l o r i d e s , C P Massolo, J . Desimoni, A.G. B i b i l o n i , A. Mowdoza-Zelis, L.C Damonte, A.R. Lopez-Garcia, P.W. M a r t i n , S.R. Dong, J.G. Hooley, Phys. Rev. B, Vol.34, No. 12, pp. 8857-8862, 15 December, 1986. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0302355/manifest

Comment

Related Items