A NEW FUNDAMENTAL HYDROGEN DEFECT IN ALKALI HALIDES

A NEW FUNDAMENTAL HYDROGEN DEFECT IN ALKALI HALIDES

Spero Penha Morato and Fritz Liity

PUBLICAÇÃO IEA 507

CPRD - AMD 36 ABRIL/1978

CTMO-AM0

m

A NEW FUNDAMENTAL HYDROGEN DEFECT IN ALKALI HALIOE8

SfNm 9mtm Mora» md f r t n LOty

curmo oi

MMTI1VTO M f N f M ' A ATÔMICA

•AO PAULO - MAWL

AfKOVADO PAHA PMKCAÇAO I M AINIL/lff't

CONKLHO DCLIKHATIVO

Klau* RaWwch - Roterto OUtr» V v tMcio UedMo d» Con*

MRTKIMNTCS

F l«*lo Gori

•UKIIIWTENDfNTI

INSTITUTO O I I N f ROM ATÔMICA Cftat P0MM 1I.0M IFhUwlfMl CMM* lMt»»r«il«fl» "ArnNntfs dt SUM* O i M M "

•AO PAULO - •RAtlL

A NEW FUNDAMENTAL HYDROGEN DEFECT IN ALKALI HALIDES

Nnht Mora»' and Frttt Uty1

AKTRACT

Atom h»d»o—i In i w n H (M°) «nd nteaih* ( M l (brm on utatnvitonM and tnwnlilW tonin ttMt f t * * *m to <«a rimiwnl m M iMacn to e»aH*«tM>» «J. U,. U,. U , n m m l . nine» he» beat

• " » * * > MM«le*rad in It» pot «Ann «tudyina ih» photcMXKompotitKm of OH'dcfarn. • K M conipfufnlm Of atomic dtvead liyd.ugwi M dicwml. «Aid) can t» pradund in Ian* • W H I M tn In* crnW and h espewitry not «numwd to my o»n»t knpunty Thtt nan» hvdrogan dafact doa* not tho* any pronounce*

akKtrome ataorptlon. but dnpleyi a ain»l* tftarp toed mid» band (ai I I M c m *1 tn KOI wtth a pattacl > / l H-^J nus»» «n.ti The d^aet canVa pnducad bv variou* UV or X f»v lachniquai tn cfytnn dopad with OH".

I H " Of M" dXartt A dtfBlad itudy ol ill Iwrwioo kintuci « low tampacaiuna «nom thai II It prinwitt

*om<ail by !*• raactwn o» a n v * * Cl) tnxHon IH<(ntw) «r»i hyii'ngan

1 - (NTROOUCTION

impuntiM alk*)i hatid* cryftal» m* wall itudiari antititi in ttwif configuration and p»opa<iiat. So 1m tnara. tra only lour known httic tormt of hydrogm centers in i n n * lalt«cn Occupying ona anion uta w* can tilhar hav* an hydrogen ion (H~ or U center « moilt commonly k n o w n " • " ) or a hydrogen atom | H ° oi U , center'¹'). In intentmal potitiont w t can ha««

•gam an hydrogen n n IH~ or U , center¹⁴*) or an fiydtogen atom (H°⁽ or U> center¹⁵'). All these form»

of hydrogen center* ate very eccnuble to a lull ttudy with optical and magnetic technique» that detect their electronic or vitxational transitions or even a powblt paramagnetic character M shown by the U j center tor example. H y d r o g e n interiiitialt acated by "optical radiation damage" through photodiuoo«t<on of substitulional OH or H", impurities have played and t M I play a model role in the attempt to underitand both ihe primary proem and secondary reactions of the formed interstitials. In the case of t^e H" defect, §n "ex t r i m * th«ngert Frenkel pair" consisting of anion vacancy and H*

(miteed of Ihe Cl⁽)is formed The H,"defect gives rite 10 a very broad itructuteles*electronic absorption tn the UV and to a local mode absorption in the IR (around 12M for KCO. This latter displays at lowest temperatures, a fine structure camming of several groups of sharp lines that correspond to H ' w H h various special corrtletions end interactions to the anion vacancy, or "correlated extrinsic Fienfcel pairs"

Ithe M,' local modes a>e split bv the interactions with the anion vacancy). The thermal stability of Ihe H"

center shows the seme behaviour of Ihe intrinsic chlorine Interstitials: thermal annealing curvet with character*!* steps are observed. A first sharp annealing step corresponds to the extintion of ttM speccially correlated extrinsic Frenkel pairs. A higher temperatura annealing step corresponds to Ihe extinction of the free uocorrelated H^dalect

Several "perturbed hydrogen centers" with the hydrogen located dose to other impurities have feaan studied recently, Specially correlated U H center pairs ware found to display a broadening and

•tufting of the Ü center UV absorption and • renu? j l of the degeneracy of its IR local mode absorption'*1. Pairs of mutually perturbed U centers IH H " peirsl were also the subject of recant

( t l t u a p v M «v a mt «re* # DMM - M O n t S AOt

IW M M m e d by FAPtfT erann # M / Í 5 I and 7 V P » and CNEN Now «I rtw Initiiuto rt» fnwOM 4iAmir«. C.P tfOM - ««o*aMlo m

111 **yM» DepartnWM umv*»lty " ' Mian, lull L*»> City. Uiih M M ? U S A

. Thaw < 1 1 0 > oriantad H ' H ' p a i r t produoa several Knat in the IR local moda region

«tat could be interpreted with coupled otdlujtor modal». Tha presence ol additive manalic NnpurrtMS o n aho produce othar forms ol parturbad hydrogen a» the UA I N . * | center»'. For example, a metallic impurity Hk» N» m KCI titiwtad in one of the six nearest neighbor site* of an H * cantar (than a UA

cantar) will cairn rha «putting o l tha triply dcganatata locai moda vibration of tha M" cantar. Thia splitting it dua to tna reduction o l tha O^h lymmatry of tha perfect lank» into C^{4 V} symmetry.

t - EXPERIMENTAL PROCEDURES

Tha KCI «ingla cryitat praparition and tha experimental optical techniques haw already been described i» datail in th* preceding work"¹. To ariditnely color our umplat wa employ** the method described by Van O o r n "^{0 1} with toma modilicatiom. I n thi« method th» F canter concentration wa*

obtained by a controlled vapor prntur* o l t h * metal «apor uamg an m e a n pressure o l nitrogen ojn. Wa i m d a low lived preuur* of nitrogen gat and varied th» liquid potassium temperature, thut varying i n vapor preuure. Uting an inttrmadiat* calibration curve Iror» Van Oorn and Rogrnar't method»^{1 1}", which oiv*t the F center concentration at a function of t h * potassium vapor pressure, we ware able I » obtatr. quit* rrproductibl* F center concentraiiont on the order ol 1 01* to 1 01* c m "1. HydrogenatiOfi o l an additively colored tempi» conrnti in the controlled diffusion o l hydrogen into t h * cryttal under high temperature and hydrogen pressure. By heating the crystal M 650°C undar 20 atmotphere» of hydrogen gat, hydrogen molecules ditlu»* into th* cryttal and react with lhe F center», producing

hydrogen ions (U centers).

To irradiate our tamplei with X Rtyt at low ttmparaturat the cryostad head was provided with a fifth window made ol * 1 mm thick aluminium toil at 45* bisweting the perpendicular optical paths.

Th» X Ray source was a Horeico with a W target. T h * untilteved aait window placed at approximately 6 cm of th» sample. The X Ray exposure W M always *1 30 KV and 30 mA.

To inveitigate all the trampareni ring* of th* crystal with U V , V is and IR tpectroscopy - atpecially in the hydrogen center local mode region w* «at up two experiment* with the two parallaj and perpendicular geometries using crystals of suitable dimension». The aim W M to prrduce tha «am*

photochemical reactions in both samptet in partlltt to we could carefully detaet and correlate any IR effect» to th* already known specual behavior in the UV region.

9 - RESULTS ANO DISCUSSIONS

As it was shown in the previous work'*'. • Sfquenct of U V irradiations at I N T in KCI O H "

system tvgiestad that a missing Cl^ crpwdion W M being trapped at some unknown form o l hydrogen d*t»ci undetectable by UV or visible sptctroscopy. «Vith th» changes in geometry previously described wa sat up an experiment that would show all the affect* of this U V irradiation tequenee at LNT by monitoring *H th# changes in th» optical absorption of I ' M KCI band gap.

H i - PhowprodiKtion» of • New UiMogaw Cewte»

A t shown in figure 1A we started with tha photodlttocinion of tha OH'delect by inadiating monochrcmttk UV light at 304 nm IOH^-ebsorptton bind). As a result wa o t M r v M in the UV range th»

OH*-> H*| cpnvanton m prevtoualy desaibad. In the IR rang» no ügniNcant spectral change» can b»

observed M thl» flat», M npactad. T n * OH" itindiing vibration absorption at 2.7*» a) too «Mtk to be detetubla for t h * low O H * doping used, and t h * neutral H*, trrterstltfal doe» not giva f I M to an optical active local mod*. In tna nmt Map «•• photo-destroyed H*| centers, rvproducing our formar U V spactra) result». I n (nu Intufikx» turn, new pronouncad afftctt ifculopwi m tha IR range Beside» th»

appoarane» of th» H" local moda trantitioni - M a«p»ct»ii from th» U V rMult - w» ohs^ved th»

growth of a new single band to far not reported in tha literature. This b*.vt *r. «ii mo.m.im n

(ai LNTI. at a 30% higher energy compared to the H~canter local mud» Hamilton. Wa eat) lhe center responsible for tht» n*w hydiogtn local mod» absorption tentatively H^a. with the " • "

Indicaling it» unknown structure and sit*, and th» minus sign indicating that this new o n t o should ba charged by being IR active.

Proceeding with the series of spectral irradiation» showed in Figure 1. w * illuminated in tha third stag» with U band light, producing t h * watt known H - H" conversion. This step basically did not chang* the optical density of this naw center. A» ntpectad wa observed in both ranges OR and UV) tha overrate of the H" canter ( U band) and basidn the trcreaw of lhe Hf center, th* U band destruction i»

aho irtnonsible for some H*/F ctntar pain formation. The new H* deter* apparently does not participa in all these transformation». A further U , band irradiation was performed in C M » B. mostly to ftllow in the IR the behavior of the new H * band. Th» maior effect» after this irradiation war* th» increase of the H ' center, th» incr-ai* of the H ; band and a dacraat* c l th» H ; center». A deems* of the H'center could partially b» ««plained by the fact that U , band irradiation aho reaches the U, electronic absorption band which is very broad and amends from 280 to 230 nm in the far UV. We should point out here that m a similar «ft of »*partm*nts with a ICCIOH"* Na* crystal, tha monochromatic U , band

•nartiahon only brouojht up H^A cantor» a» mentioned bafora with no traces of H~ cmttr formation wrtatsoever in t h * IR region.

All the abov» described affects can happen simultaneously if w * irradiate t h * crystal with undisposed U V light. This we» indeed observed by irradiating two KCI sampla» of different O H "

concentrations. For t h * low OH'concentration sample wa observed a parallel ris» of t h * H~, H'and H * center transitions (Figura 2A). Attar a prolonged irradiation thaw transition bands reached some sort of dynamical equilibrium and saturated. I n several attempts wa disturbed tht» dynamical equilibrium by riummation with monochromatic light into th» U band for enampt*. producing the H"-»H^-»O conversion already described in Figura 28. Conversely w* would invert (hit p r o m t by ihimminating into rh* u , band. While th» relativo amount» of H " and H * local mode absorptions changed reversion/ in the»* « i t » , th* hvight of th» H~ transition band remained approximately constant (as it did in t h t eap*>iment) ehown in Figure I B ) . At lh» and w * wcrt aM* to return to tht dynamical equilibrium relativt height» of the»» thro» transition» as shown in Figura I A fust by repeating for • few hours tha fuH Xenon lamp irradiation. T h * IR local mod* transition» and the corresponding UV spectral absorption» after t t n noun of «mditperstd light irradiation» am «hown respectively in Figuras 2A and 3 8 . From tha U V spectrum wa can confirm that to naw band» war* found in i h * U V region This to comment with th* fact that tha H * . apparently «ring no significant eleclionic transition in the transparent crystal rang», will ba tha most piomintnt product of t h * OH'photodtcomposition at I N T . as observed from this saturation «pariment.

We repeated this fun Xenon lew*» irradiation aaporimem for a K O O H " » Nt* sample (see Figure 28) and observed t h * tame saturation behavior a» the one obtained with a ICCIOH" sample (Figure 2A). Sine* our irradiation alto contains t h * wav»i*ngth whrch to abscind by the HA center», thto kind of undiepartad irradiation proven» their formation. (W* confirmed the otiutodojctrwclion of HA

cantor» in another eiowimem «hare we Matched the MA band by irr»diaiing monoehiorrietic light that

• absorbed by t h * H^A e*nt*r>. A t t remit of fM» •eruretian Mpailmom w * atam ended up with H ; bradujiion condition», w * W H O *bt* to tea M tome mort d m H « w Initial boHd-w» of the*» thra*

VarMnon band» (Figura*). The M f M formation ret* of t h * M ; osnMr» to tow, but Ultra—» markedly under prolonged Irradiation. Thto mean» that undar undiapaned Irradiation oondHkra t h * H ; trirarnon to formed * t * h * «iparae of anoffajr notllon product « M d * mutt b* mttialiy creetad by t h * U V

Mion oomponent.

^ f»_ »vaeai vjsaw a a w w e^w ivweaw • * * • • * vt_ V^P^IPI^W» w H H W H ^ P ^ W*W^«*> B

A ctoaw look at «to H * «antHlom band r o M o M «an/ kmmHt% roaturaj. It did not «how any

«ructvrt or aatininf whan cooled to torn Mmperatur*», dhvlevmi • akigla band with maiimum at

1 1 1 4 c m¹ and a halfwKtth ol 7 . 5 c m '¹ at 6K. The variation of the band shape at a funaion of temperatura can b* teen in Figure 5 Our gtnetat ob»i*atiom of th» H~ canter to lar strongly indicata th» funrnna of a hydrogen. But how would ona confirm tha presence of a hydrogtn in this naw defect? A rather straiqhforwwd test was to repeat th» full Xmon lamp irradiation for a lyitem such as KCI-00 and March for isotope shifts in tha obtarvad transitions. This experiment it described in F*jurt6. Wt indntl obsrtved the displacement of tha H* transition horn 1112 to 77» e m "1, an almost peifeet v ' 2 isotope thilt (H - 0 ) . Parallel to this we also observed th» isotope shift for tha local moda of tht '-T center fiom 845 cm"' to 60b cm" *. Finally, in a briaf attempt to demonstrate tha generality of t r * H^a difrct ar*l m photoproduction from OH"defects, we took K B f O H ' a n d RbBrOH'samples and leprated th» iull Xennn lamp irradiation experiment. We indeed found in these two systems the formation of H~ anrl H" centers as in the KCIOM system. For KBrOH" tht peak positions of H~ and H « t » «t 1095 c m '1 ami 790 cm"1 inprctively. For HbBrOH" they were at 10?1 cm"1* end 7«5cm ' rr»t>*ct'*eiy Th«e results indicate the general nature of the H~ center excluding tha ptmih'lity if Hi hrii«) foimni only in K C I O H ' systems.

Up to thu point we know or anticipate that hydrogen and mobila' Cl" crowdionf are tha basic cind.diiri fof the formation ol tfc* H( hand. However we must nmembaithat so far we only dealt with o'ir oi the two i»:«1iict> of the OH'photodtuoaalion. namely the H ° center. In otner words, tht O" crntrr .at no if lal^n into consideration although it was always prevent and in the same quantities at the imr.jl H " cvnifr. In the full irradiation experiments, the presence of the 0*center is even mora cj.t.c al uncr here we are alto irradiating in the O" band. This could result in photochemical reaction of new o!e<:t ewreqattont Since the precise optical detection of tha 0"center and this contribution of tha vanout drlrtt rrjrtiom was experimentally imprarticel, its eaclution from lhe formation of the H*

center . j n only be pioven by indirect experiments.

3.3 - Th» Formation of H^a Canters in KCIH"

1o answer «II the penrjing questions mat remained open up to this point we X irradiated • t C M tyttrm at 77. The reasons to this are. a) X-rays are well known to produce Cl° crowdiona in K O ,rt low temperatures" } >: b) a KCI H" system would offer hydrogen traps for the mobile crowdions; c) • KCi H" tyiirm would exclude the presence of th» 0 ' centers. We X irradiated • KCI sample containing 5 6 x 10"' H~. Out to the high concentration of H ' defects, its local mod» transition was monitored by the ant i Siohm phonon srde band, since the mam local mode transition is completely off scale (by • facior of «OhKfirr than the phonon tide band at L N T( l iY After a prolonged X irradiation tea gbsrrve the formation of structure* correspond to transition* of M(" centers that have different spatial correlations to their ami centers - the enion vacancies - as explained by F r i u '⁴' . Tha Xirradiation also produces H" centers as we htd antecipafed (Figura 7AI.

In Figure8A w» display th» growth and decay curves of the centers involved in Figure 7A. We observa that X irradiation produces lha following effects: a) N * centers undergo a gradual reduction, reflecteJ by th* jeoeete of its phorron sideband; b) H(" centers measured by the main peak from lha spatially uncorrtfefed ones are formed with approximately constant rata; c) N^ canters are formed Initially with a small rat» wtiich increases under further Xlrradiatron. From C wa recogniia that M~

centers are formed more efficiently with the help of soma reaction product that is obtained during th»

initial X irradiation exposure. This observation makes tha H* «amar tha strongest candidata for a trapping M e loi the C(° crowdions.

tn order to further verify tha participatim. of H* defects m the H^ production process, we sat up another experiment <r. which prexposed a KCIJT tampla to U V «raiiation. As a result of this UV irradiation we obtained, very efficiently, tha H'-« H*conversion process without any trace of H* cantar formation (sea Fiuur* 7BI Under X-frrariietion of this previously UV exposed sample, wa observed an initial daciaaa* of H⁽' centers and an increase of H^ centers with m Initial formation rata a factor of -r 16 larger tnen the initial Hf formation rate obtained by a direct comparison of Figura J0A with r - » « » 7 0 e Onder further X-ray »«posore fha system lend» tv reach some dynamical equilibrium. The

observed lacti conlirm again strongly tha pronounoad participation of tht H~ oamar at a trapping sit*

for t h * CI*J crowdion. Wa cannot dacida from these exp*rimantt if H~. defects play any rol* at tha Cl?

crowdion trap*. Wa can howavei. concluda that if thay participa in tha formation proem, thn trapping probability for C l * aowdion it contidarably lower than that of H" defects.

A doubt that «merges aftar doing these X-ray aitpirlmantt may ba expressed by tna lollowing question: Could tha KC!:H~ sample contain unwanted OH" impuritiat which would have produced the H * effects under X irradiation/ Due to the high concentration of the H"center, their UV absorption would eomphmty mask any tmall OH" band. Two conclusive arguments can ba given agaimt tuch a pouibility:

a) The tame UV light exposure which formed no H" eentert in the experiment of Figure 7B produced, in expvriment shown in Figure 1. A an optical density of 0.15 in the H~ local mode absorption If our KCI H" sample of experiment Figure 7B would hive a factor of six lower concentration of OH" impurity than experiment Figure 1A. we should have bean able to produce 0 0? 0 0 . of H* local mode absorption ( 0 0 2 0 . 0 . wai the lower limit of our detection system ability at the vgnel to noise ratio used). We can therefore say that if OH" was present in the KCI H" sample it should be in concentrations less than 4 x 10"*;

bl In another experiment wa exposed a sample of low OH"concentration (2.5x 1 0 " ' I t o Xtiyt at LNT during 16 hours and were unable to find any trace of H~. H~ or H * centers This shows that X irradiation at LNT does not decompose at all the OH" defect into any of I f » hydrogen reaction products H~. H~ and H~ as observed under U V

•nictation of OH". The participation of any unwanted OH" additions in the experiments of Figures 19 and 70 is therefore definitely excluded, and to is the possible participation of any oxygen defects in the H" formation.

For a further confirmation of t h * rton participation of lha O ' center in the H" formation w * did another experiment in which wa used a KCISH" crystal and repaeted tha full UV irradiation treatment at we d 4 with the KCI OH crystal. Wa know from previous w o r k " * '1* ' that th* SH"center can be also decomposed into S~ and H ° defects. Proceeding with this full irradiation treatment wa observed exactly the same hT band in tha KCISH* crystal. Tha experiment» described in this section, betides confirming, th* participation of Ht' and CI*J centers into the H" center formation, ruled out completely and possibility of the participation of the O* cantar In tha H* center production.

1 4 - Pivlimewy ConctwtM» About Ha formation from LNT Eapartarant»

Up to this point tha various experiments at LNT allow us t o draw several conclusions about t h * nature and formation process of th* new H* center: a) under full UV light irradiation, H* centers *r*

lha most prominent hydrogen reaction products of tha O H * photodecomposit'ion; b) under stepwiM monochromatic OH" photodecomposilion, H" defects from in the secondary stag* by pholcexcitetion of H*j centers. At this process creates mobile Cl° crowdions, t h * latter are very likely candidates for t h * H~

formation; c) T h * IR absorption of the H * centers shows, by the H -» D Isotope shift, that it is caused by a perfectly localwd vibration of a charged hydrogen ion; it* single band structure indicates • sit* of high symmetry for th* hydrogen, which does not split its focal mod*; d) M~ centers c m b* formed by X irradiation of K O . H ' ayttals at LNT. This excludes any contribution of th* oxygen in th* H* formation process, and confirms th* idea thai H~ canters are formed by the reaction of mobila Cl° crowdions with hydrogen defecu Th* ineraat* of th* initial H ; formation rat* In this experiment by a previas» M"-» H"

conversion makes the interstitial H ' o * f * d t h * moat likely candidate for t h * trepping of t h * C«^ crowdion forming the H^g canter.

All the preceding Mp*rim*ntf war* don* f t LNT, wher* C l * crowdions «r* mobile right after the» creation and thus form th* H* defects instant*n*ou«ly. If this picture It correct, w»

should be *t>l* to prnduc* H* canters In oontiotted rlept at tower temperatures wher» Cl^

rrnwrfions êt% thrim»My statile. F.vnarimann In fhit temperature rang* should therefore provide a

conclusive test un the fotmation process end • definitiva identification of the hydrnqen delect trapping the Cl° crnwriion.

3.6 - Controlled Production of H ' Center at U t o T

We initially made the photodissociation of the O H ' canter at 77K (fo« experimental convenience! by monochromatic 204 nm light (Step 0 in Figure 9 A ) . We then proceeded with the further photochemical reactioni at 6 K. After photodecompoting H*j centers (Step 1, Figure 9 A | at BK we oburvad thai the M" center local mode wai built up at expected from previous experiments. In contrail to the ciwreiponding experiment at LNT ( i n Figure I I , we observe at this ttep no trace of H~

land H I forn-jiion Thit confirmt our previous assumption that the photorlettroyed H ° centers am quantitatively converted into H ' centers and Cl° crowdiont. with the latter ttabibfed as H center» m the lattice. It further condrmi our assumption that only mobile Cl° crowdiont. reacting with hydrogen detr-cit. are able to form the H( center» The "telf trapping" of the Cl^ crowrliont at H center»

thrrrfore prevent* the H" formation. If our line of arquing it right, the M~ centers should be formed if we m.ike the puxtuced Cl° crowdiont mnbilr in tome way. One way t o achieve this it the optical e x ' U t i o r i m the electronic trennt.on (H band! of lhe Cl" crowdion. which leads to an optically stimulated mot'oo of the defect. If wt shine monochromatic light into the if band (Step 2 in Fit)o'»9A), we irN<r>«ii ofovrrve the appearance of the H^ local mode band. Simultaneously we tee the rrilti.tt>on of thp H local mode band. A t no H ' delecit have been present during this reaction we mutt conclude thtt in this step KT drfrctt have been formed by the reaction of "optically mobilized" a*J cruwl'ons with H'rlvlrcts. A further irradiation into the U j band (Step3 in Figure9B) proceeds with the M'| • H » Cl™ formation at in Step I without further H~ formation A subsequent monochromatic rrratl Mtion m |h» (J band biouqht up mainly the local mode oi the H 7 D close pairs (Step 4. FiguraSB) w<th lhe crwrespontlmt decrease of the H" absorption center. Now that we have two kinds of hydrogen centers cnmpptinq at traiiping sites, we again optically bleach t h * H band and observe further increase of the Ha local mode band Simultaneously we see a considerable reduction in the local mode absorption of the H_'( 1 close pairs, and a very small reduction in lhe H"local mode. Apparently when both H ' a n d H 7 ! } port o f f n n l are much more effective for the H~ formation. This it exactly what we observed in the X ray e«periment with the KCI H~crystal! of I N T .

The second potubiliiy to mrjhiii/e the C l " crowdions after their optical creation at LHeT is a ttvrmtl ann»a'inrj pretest into the temperature range of their thermal instability (T > 5 5 ° ) . This proem was p<f formnl at Step 6 in Figure 9 0 . It leads to • further considerable increas* of lhe H ' centers and • s<mu)tiineou« d'ltiuriion of the H V H ciosa pairi At the latter are clearly thermally stable in the used trmprriture rarv|» (they are thermally annealed only at T > M K ' * ' ) , we conclude that lhe H~ increase was achieved at the expense of lha H⁽7 1 ! close pani. (Tha H"centers remained approximately constant during the iheimal annealing Step 6 ) . Tfms again we can conclude (hat thermally mobiliied Cl"J crowdiont form H ' centers by /eacting with H⁽7Ociosa pairs. Tha invohrement of the Cl? in tha H * formation is conclusively demonstrated in Figure 10, in which we monitor lha ebsorptici during tht thermal annealing process Tha correspondence of lha thermal destruction of tha Cl" centers Wound S5 Kl to the increase of tha Hf absorption is clearly demonstrated (lha change m the H-band absorption at 10K is caused by thermal reorientatton of tha Cl° crowdions).

A mon gvntwal experiment confirmed tha two above described results. Wa employed wnditperiad broad band UV irradiation et 6K «s we previously did at 77K. Tha results of this aapenmeni ara displayed in Figures I I A and 11B. From Figure 11 A , Step 1, w» sea that after an enfioture of undiifwriari U V light wa observe immediate formation of all three hydrogen defects H^a. H^t"

lcorre!*tad and uncorrelatndl and H* centers Since our irradiation contains tha wavtlenghis of lha O H ' band, of tha riifWm hydrogen centers involved IH . H(, H0}, and of lhe H band absorption of tha Cl*J (•«•f»r. the simuitarwims appearances of lhe Hâ H' ên<l H'canters is expected « in lhe LNT full

Alin ihf. 1,.i\ (JV r«(«ni'i» w* I I I O N I M I I Ih* H hanrl ill.ny ttr«|itr»<iA< Injhl In;' MtKlxng iny

< I'J I ' - ;1- ; l i f ( i o l t l O n n w i > > * 't:' • " ! I ' i ' i l v « » S « f p I , \ H i t , , o y\ t,\ A » | » r v n ' r n l i w r v d . »v»

again H I the raise of the H~ at the expense of H⁽" centers. T h * following ttap was to practically repeat previous UV treatment to re establish the dynamical equilibrium among the three defect! (Step 3, Figure 11 A| at previously done i t LNT. As a last step, we peiformed again the thermal annealing to 77K procedure to mike the crowdmns thermally unstable After this treatment we again observed an increase of KT centers, a decrease of H~ centers Iboth correlated and uncorrelaled) and a corresponding decrease of H centers. This experiment, although being different in terms of irradiation procedures and intermediate products obtained, confirms the result! of the previous monochromatic experiments

I f * - T h * Thermal Destruction of the H * Centart

To estimate the relative oKillator strength cl the H" center, we report to the previous work (Part I) where we learned that at LNT approximately 50% Irs: H'centers are formed when compared to the LHeT experiment We atsume that the other 50% of "missing" H'centers are being consumed to form H^< centers Usuxj the relative strengths of the integrated ahsorptiont in Figure I B . we obtain ' H " ' H " 0 7. Considering the value 0 5 lor the oscillator strength of the H~ Center. After we have studied the rlrtaiir*d kinetics of the H~ center creation and concluded that the H~ center was the final and optically stable products of the Hg thermal destruction. When heating a sample containing H~

centers and following their local mode absorption at LNT after pulse annealing to various higher ternjwatu'es we found that the H~ centers decay thermally in the temperature range 1B0210K (Fiqtire 12) This annealing behavior is very close to the H -* H l h e r m a l decay process of H" interstitial*

drifi l>ed by F r i t /^{1 4 1} We imlepcj observed, toqr-ther with the extinction of the H ' centers, this thermal reaction H • ( ] • H" by the H decrease and H increase, as shown in Figure 12A. A simple comparative/

anjiyvi of the strengths of absorption changes in the three IR bands involved in this process shows that the putmrMvi of H ' center» does not create Hcenters Taking the loss in absorption strength in the M"

local mode (A OD 0 8) and converting it into the corresponding gain in absorption strength for the H * center, we woukl expect an optical density increase of 1.7 for the H'absorption. We instead observed a mere 0 3 increase in O D . of the H local mode, a fact that by itself excludes the possibility that t h * destruction of H" centers will form H" center*. The small increase of the H" local mode is fully accounted for by the thermal destruction of tha H ^ Ü extrinsic Frenkel pain, at teen by the destruction of the H ' local mode band No new local mod* absorption in the IR rang* and no new electronic absorption m the U V / V n range it observed to develop after the thermal destruction of the H~ defects.

1 7 - Final Conclusion* on the H~ Formation Procea* and Structural Model

The large variety of experiments described in tha previous sections offered a consistent pic.jre about the two components which participate in th* dynamic H~ center formation process We proved in different wavt that Cl° erowdions are participating in the H~ center formation: a) in the LN7 rang*, where U , band irradiation create* th* K deficit, we ihowed thai competing Na* defect* can capture and ttabili/* Cl° crowdiont and tupprtst th* H~ formation, b) creating Cl° crowdions by X-rays lit LNT m a KCI H tyttem leads to the formation of H ' (excluding the contribution of oxygen in anv form in thu process), c) at LH*T we wer* al>l* to create H< center* itepwise in a controlled way by first creating stable Cl° crowdions. Only when these crowdiont were mad* mobile by optical or thermal excitation, th* H~ defects appear.

T h * • xperiment* under various temperature, irradiation, end defect condition* showed continently the participation of hydrogen defect* In th* HJ formation process: a) in a crystal containing both HVi;] pair* and H~ defect», t h * formation of H~ cent*» is accompanied mostly by a reduction in close H(/ O pair» and vary little, if any, b y * H ' defect reduction; bl If only H" defect* are present, the MM formation proceeds at the expense of th* H" defect*. We therefor* conclude that t h * M^ defov» it fo'mud by the reaction of mobile Cl^ erowdion» with either clot* H ' / U pairs (preferred proeetr'/, or with M" defect* As the H(7 f l Frenkel pair» it - In term» of it» net «tructural components eq»n».)lcnt to U<* substitution!) H defect, both ttu>te hydnxpn tr*(H for the C l ' crnwdion can l<-a<1 to tl<e w m

~«l prrHlijr.t From trie H" local mode »1rr>rir|th, siMtrtrjl ih«|ie and i»ot"|>e shift wr- f(lncliclift tli.it trie

defect mutt consist of a charged tocaKied hydrogen delect in a Mia ol high symmetry which doe* no»

split tha local mode. Tht h q * frequency of the r i ; local mode indicate* a Wong» vibration»! potential of the H~ compared to the H" defect. From tha thermal destruction of the r T cantar we know that it n not converted back into H^ center*, but disappear! into W M unknown opticaHv maccanibh structure.

A structural modal must be able to integrate and satisfy aM t h e » static and dynamic experimental features. We tee only one possibility to achieve this in a single model: A li»itwain ion H~

in a bodyxentered intentitial petition, with a trapped hole shared lyimaMj Haiti by the foe*

surrounding nearest neighbor onion* (see Figure 131. This model first satisfies tht observed static features, a localized charged hydrogrn dtfrct in a site of high symmetry with a single unsplit thric-loM degenerate iurji moite transition. A side from the trapped hole, it is similar to the H~ defect (saa Figure 131 In frV Utter one. the charged interstitial H~defect will have strong electrostatic interaction*

with the fuuouraling ions, repelling and shifting outwards, the enions. and attracting and pulling inward»

the cation» The addition of a hole (positive charge) to tha shell of nearest neighbor aniont in tht H*

centrr will mtiice the H < > Q~ repulsion and thus produce a closer distance between the hydrogen and the thionne wni As the potential for the localired vibration of the H~ is mainly produced by tha Born M.ivf rrputnon interaction with tht (large sin) anions. we will expect a higher local mode Iroiu.Mii v ol the Hp compared to the H~ defect. This is in agreement with the frequency shift to higher

tounrl experimentally.

For ih» dynamic formation process, we have to regard the two cases: a) reaction of a Cl^

aovwi.on (an interstitial C l ' ion with a bound hole) approaches a H 7 O Frenkel pair, the most natural proem to Kiurnt it the recombination of the Cl~ interstitial with lhe empty tcancy of the Frenkel pair At i remit we are left with the H* interstitial and the hole which was earned previously by the CI*J o o w t w i II th'i hole iiisi reoonitxnes with the interstitial H~ we would restore the original neutral inff.iii'.jl H " tenter. Apparently, however, the polarization of the surrounding ions around the H("

deflect (Fiçure 13) make it possible to «rap tht hole at the CT ions surrounding the H," defect in a (table way. r» reaction of a O ° with a H" defect m order to form the tame H* defect in this reaction, the approaching Cl interstitial crowdion will exchange the lattice place with the H" defect, so that an intern.T.JI H" ion is formed. The hole, carried by the Cl° crowdion then gets bound to the H~ defect in the ume way ts above. In terms of its net components (after recombination of the hole with the Hj"

deffctl. the H_ it a neutral interstitial hydrogen atom. Thus when it gets thermally unstable and recrxntnrvt v> *h another H ' delect, it would form a neutral interstitial H , molecule. This delect is known ,o he prewnt in alkali hehdet and spectrally invisible in both UV and IR range. The "spectral d'ttpoatrance' ol the H^ defect alter thermal annealing is therefore « i l l understandable with our model

vVe should point out thai the proposed model is constructed strictly from en extended static and dynamic experimental material, which is conclusive and d o n not leave any alternative choice for • different model, consistent with the experiments. A theoretical justification for tha proposed interstitial H" structure with a stabilised hole snared by the surrounding tniorw is beyond tht scope of this work.

Tht theoretical understanding and (unification for this peculiar "inverted and self trapped axciton" at ÊI\

irrterHitnl hydrogen defect appears to be a challenging and interesting problem for further theoretical studies

RUUMQ

Arorm <*• hdroeinw n«a»"wi IN* I OH rteiniM IH I quanas m> poatçòei imwMiclata ow «abMkucien»)* afc or**m a mHnot rmrtMf bwn caranwinrJot rws halalo» atoallno* Icemro* U. U i . Uj. Uj) O M twn *MD exi»r*i«*n«r<w «(furiviat no pMMdo EMuaanl»w a foiovjeernposstaa oe dtfCHot OH'ume cpndewtçío now a>

h>iuie*n» «Amkn arrcaarta 'ol entontreoa Cda tonne tfe finliualiila pea* *er pradvMa tm arandea ouaniklafaa no crtttal * nfc ex» mremtfnenw vMcviate com nwihvme owrr* impiireaa. law dtfetio * Mo>aaink> nto Wttena newhwns atxwçio *\m<l>n*a pronunoeae mm apreasma «m úmeo modo loctJ dt nwitme em 1114 em"1 em KO e asm um IMWIO rwiApvo pmltno Ó» iji (M • 01 i « * tfefeno pod* wr nortuiMO per veriet leenica* UV ov 4* rekx N *m amtn dnuak>i enm mvunim OH. SH ow M~ um eriudo «nalha. « d* sua cinetica d» totm*Mo i betxet

Muiai mrxix nut t>» nefwio * primartameni* formato oa nxKB» o> «m "crowrjlon" Clj IcenSfO III earn

KCI:*i.lX|C5 O H " Photo-reactions at 77K 5nmirr

.4a tu

204nmirradi 235nmirrad

4

- 0 h (OH") | / (U

)

235nmirrad.

(U

) -

A Electronic- TransitiortB

B Loca! Mode Transitions

0.2 -

20 4 0 Irradiation tima [hrs]

f'tfur» 1 A Electronic ibvxpiiott («rxrt si < fureiion of Irriviitllon tim* for • taquenc* of ic irradiatfvit m » KCI OH crvit»l «t t N T IthukoM» - 0 5 m m l . mod* rr«rif>tinnt

fl 0 mm)

!R t*m« utrjutnc* o( *« in A (thichnati

10

Formotion of Hydrogon Locol Modos ot 85 K

10 20 30 40 Irrodiotíon timt [hit]

Plot of th« local modi Irintition of H , M**nd H~ ctnltri •• • function H irtidution undtf full X*fion lamp «rf«di«tion tl KK Imwii/Rj *t 77K)

A - KCIOH'lYfitm

• - KCI OH * N#'

11

KCI:OH" T-77K

lOhrs full X-lampirrod. 85K

Local Mode Transitions

(§) Electronic Transitions

. 1 . — r — » — I . . * T I >L

5 0 0 4 7 !

Q2 -

200 220 240 260 280 Wove length(nm)

600

F l f i » 3 A Hydfog«n ctntw» locil mwl» «»nntioni iftw 10 houri ol lull X»non l»mp irridiition H 85K (tTMMur«d «I 77K)

8 - Th« miMciivt UV

12

i.o- KCI • I.I full X-lamp

irrad. 85 K

10 20 30 Irradiation time [hrt.]

- Plot of fh» local mod* tramitioM of H". H("and M* cantar» • * • function of irradiation timt undar full Xtnon lamp IrradKlion fur • KCI OH"< itlêi wrth highar OH'ooncantration.

13

0.0.

08

06

04

02

Holf

6K

, KChOH" HJCenters

J

50 ^{100 150}

T«mp«ratuni[K]

77K

H

I46K

A

I76K

1125 1100 1125 1100 1125 1100 Wave number [art

]

1125 1100

F*ur*8 thtmnimict of fh» *nd of lh« H loral mode

14

K C h O D * Loca» Mode Spectra 77K I8hr«.ful X-lamp irrod 85K

0.6

aa

0.4

0.2

°f

Ill2cnr 845 779 606

I I

«(HP

-(Dp' ^1.394

17^*'

1200 1000 800 600

Wave number fcnf

]

»Nfi «»»ieti of MN H ; and H,'ln • KCIiOO"erv«it • * * M l Xwon temp Irrwftotlon

•i MK Irrmuni it 77K|.

16

KCI

H" hydrogen local modes 77K

O.D.

0.6 04

6.2

1 Bocltyound 2 7hr« X-trod 3 2lhn • •

4 42hr» • •

0 6

0.4

02

1 Pocfcyouoo 2 90min UV-inad.

3 5 min X * vrad.

4 Thm •

5 21 ha

1125 0 7 5 ^ 0 0 85(3 800 600 560 Wow numbtr cm

A - Formation of H]| and H(" lot di(!«r#m tpttUI oorralttlonil Mnttrt In • KCI * 6.0 x 10"' rt'umpM unriar X-ftyi Irradiation at 77K.

• - Tha Mma a* abovt but aftar a ortvtout UV Irrarftation, producing H * H, convwiton

16

0.6 0 4

02

0.0.

08

0 6

KCI*5.6XK) ⁹ H"77K

HJ (phonon

—' sideband) K) 2 0 3 0 4 0

X-Irradiation time [hrsl

Fkjwvl A - InoMW of H* tnd M(" luncorrtlMMl ontt) M • lunction of X irradiation Itmt. Tnt contumpiion of H'ctnttri it monitortd by th« phonon «ktelMnd of the H' local moot.

• - Tnt H I M M tbovt but tfftr Iht prtvtom UV irradblion. All data art laktn from tht

•paetral manuramtnti In Figurt 19

17

0.D 0.05

0.05

0.05

005

local mod» ct 6K 0 l6hn204nm(0H*)irrad.

1 Ihr. 235nm(Uj) • 2 ?.3fw.340nm(H) •

. 0.1.2

3 I hr. 235nm(U,)inqd 4 I hr20nm(U)

Li.

S 1.2 I w 340n<n<H)trrad.

6 Thcmt. Anmol to 80K

" (isolated)

90o Á

nwnter [cn']

00- (H-)

02

H0.2

- Q 2

- Q 2

Local m o * U»«h»ofW of M", H," WK» H ; un*K ÍÍK dlffw.nl «*Mqu<mt mooochromtHc 6K <«•»» O - 51, »rrf • fin«l thwmrf «nítMlIng to 80K Irtup fl)

18

-0.25

5 10 20 30 40

Tcmperofur» [K]

60

10 - M ctnttr decay and H'n build up curvtt v». itmptrtlurt. Thii mM*ui4m«nt ccvr«tpond« to Ih« tnpmimêtn thown In Ftgurt 7 1 0

IB

0.0.

0 4

02

0 4

0 2

K C I + 2 . 5 X IÓ

OH"Hydrogen

Modn 6K

0 KCI OH'background 1 6 hrs. full UV irrod 2 2 hrsH bond "

3 2.5hrs.full UV "

- ^ 4 Thermal Annealing 6 - * - 7 7 K

1125 IÍ75 700 850 r ,800 500 475 Wove number [cm¹

11 - Production of hydrogtn omwri in K C i O H ' « 6K undw undiipffMd light (broMl tend lMt>(«ftnM filter • Xtnon lamp).

A - S u p 1. Formation of H*. H ; (dilfarant oorralattont) and H , ctntari undtr full UV Irradiation. Stap 2. H^> formclion and H~ dt»tructk)o unrifi H band (Cl° oowdion) liQhi. Step 3. Rapatiiion o< Slap 1.

• - H ' production *W a tharmal annulling proem to 77K. K'and H"cantar»diamrn altar tnii traatmani.

O.D.

06

04

0 2

77K

1 40hrs. full X lamp irrod. 85 K 2 annealing to 2 3 0 K 1 2

O.D.

0 8

04

02

1150 1100 900 850 _,, Wove Number cm]

I

_,, 800 525 475

KCI*l.lx|C)

0H' 77K

HJj centers thermal decay

80 100 120 140 160 180 200 220 Temperoture [K]

12 A - H ' , H⁽" and H" C«nt«r« local mod* irtniltionj b«for« «id t f i t r pulM «nnwling to 240K.

• - H" c»nt*r tnnMling curv* t l I N T iikan under • pulM fnr»allng procedurf.

21

• Hydrogen Ion O Fotassiom Ion O Chloride Ion

Hole

19 - Structural mod* of Iht M; c m n (A) and tht MJ" cam* IB) with Indication of tht ihlftt In pofxion of th» «grounding «ilont ind enloni duf to «iNlroitatic intartction with tht Intaritiiial hydrogan Ion.

22

REFERENCES

1. R. W. PoM. Proc Phyi. Soc 49. 3 (1937).

2. R. Hihch and R. W. PoM. Tram. Faraday Soc 34. 883 (1938).

3. W. HayM and J. W Hodby. Proc Roy. Soc A294. 360 (1986).

4. G. G. K. Whrta, Exptrmrnital TmJméqum to Lorn Tmmumtum Pftyrt». 2nd «d.. Oxford (I960).

5. A. Haiamtnn. Zaitt F. Phytic 182. 313 (1986).

& G. Ku/z. Phy*. Stat. Sol. 31. 93 11969).

7. M. d* Souza and F. LQiy. Phyi. Raw. 8. 5866 (1973).

8. D. N. Mirlin and I. I. Rainina. Soviat Phyi. Sol. Stata { . 116 |I988).

9. S. P. Morato and F. LOty - PuWieaçfc IEA 606. COURP - AMD 36.

10 C. Z. Van Doorn. Rav. Sd. Inrtr. 32. 756 (1961).

11. Von H. Roganar. Ann. dar Phytik 29.386 (1937).

12. W. 0 . Awl« and J. J. Marktwm, Phyi. Rav. 8J. 1043 (1962).

1 1 B Fritz. J. Phyi. Cham. Solkh. Suppl. 1 . 485 (1966).

14. F. FiwtMr and H. GrCndig. Z. Phyiik 1J4, 299 (1966).

15. J. Rorft, Appl. Phyi. Lattan ft, 66 (1965).