Luminescence method o f examination of
3 d
level
structure of Mn2+ ions in glasses
STANI8LAW GrĘBALA
In stitu te of P hysics, Technical U n iversity of W roclaw, W ybrzeże W yspiańskiego 2, 50-370 W rocław, Poland.
It has been shown th at in th e glasses activacted w ith Mn2+ no absorption bands are noticed w hich would correspond to th e transitions d -d because of th e presence of th e absorption band Mn3+ am ong others. I t has been stated th a t th e application of the lum inescence excitation spectra allow s to exam ine the structure of the su b level 3d for Mn2+ ions. T he m ethod of th e lum inescence excitation separate offers a num ber of advantages: am ong others, it allows to spectra Mn2+ (4) and Mn2+ (6)* in a sim ple w ay.
1. Introduction
T h e m anganese ions (M n2+) o f electron configuration
3d5
belon g to the grou p o f v e ry efficient lum inescence activators, therefore the p rob lem o f lum inescence M n2+ is w idely discussed in papers [1 -8 ].The lum inescence properties o f M n2+ a ctivated glasses o f different com position s are the su bject o f interest o f m an y authors [ 1 -5 ]. T he lum i nescence o f M n2+ ions is solely con n ected w ith th e transition o f electrons fro m th e excited level
*Tlg
[4£ ] to th e fundam ental level6A lg
[6$ ] . The electrons transit from th e other ex cited levels to th e fundam ental level*Tlg[*G-}
w ith ou t lum inescence em ission [9 ]. Since the p osition o f the level4T lg [4G]
depends on th e coordin ation sym m etry o f the surrounding ligands (for oxid e ion glass), then the spectral properties o f lum inescence is also con dition ed b y this sym m etry. In general, it is assumed th a t M n2+ (4) gives green lum inescence, w hile M n2+(6) — th e lum inescence o f red colour. In th e presence o f b o th M n2+(6) and M n2+(4) the resulting cu rve o f th e spectral distribution o f lum inescence w ill depend on th e quantitative M n2+(4 )/M n 2+(6) ratio, w hich depends u p on its glass com position , an essential part bein g p la yed b y the qu a n tity and sort o f M e20 oxides.Thus, b y increasing th e num ber o f M n2+ ions fo r th e constant com posi tion o f b asic glass w e will increase th e num ber o f M n2+ w hich can n ot b e im bedded in the glass structures in the tetrahedric surrounding [10].
132
S. G
abalaThe exam in ation o f the structure o f th e level
3d
fo r M n2+ ions and the determ ination o f crystallic field param etersB , C, Dq
is based exclusively on thp results o f absorption m easurements. T he absorption bands as well as the corresponding w ave num bers are sought fo r the electron transitions from level ‘ ¿ „ [ ‘ A] t o the levels: % ,[* < ? ], % , № , · ! „ № 4i f s [46 ] , 4H „ [4D ]. The ions M nI+ (6) and M n2+(4) show similar splitting o f th e levels depending on param etersDq.
T h ey are presented in the diagram due to Tanaka and Sugano [6, 7 ] (fig. 1). E lectron transitionsF ig. 1. The energy diagram for ions of electron configuration d5 as a function of Dq
param eter
fo r M n2+ ions o f coordinations 4 and 6 d iffer in intensity and shift in the respective light w avelengths. The absorption bands are o f v e ry low intensity due to the fa ct th at these transitions (betw een th e levels) belon g to th e fo r
b id d en ones. T he dependence o f the sublevel splitting u p on the crystal field is different. I f the sublevels
4A lg[4G], 4E g[4G],
and4E g
[4D ] are independent o f the field strength (up toD J B fa
3) then th e rem aining levels are characterized b y the directional coefficients and th e respective ban d w idth depends on the valuedEId [Eg).
Thus, th e energy o f transition o n the levels4Tlg
and4T 2g
depends on the field strength. W ith th e increase o f the field strength the transition energy lowers, causing an absorption ban d shift tow a rd th e longer w avelength region. I n a field octahedric sym m etry th e levels4E g [4G]
and4A lg [4G]
superim pose each other. A redu c tion o f sym m etry results in splitting the a b ov e levels [6, 7].In th e exam inations o f the
d-d
ty p e transitions fo r M n2+ the m ost d ifficu lt p roblem consist in finding th e corresponding absorption bands. I f th e con cen tration o f M n2+ is low or if som e other absorption bands appear sim ultaneously, fo r instance, those attributed to M n3+, the re co v e ry o f those bands is p ractically im possible. In this paper w e shall show th a t in spite o f th e fa ct th at the spectral exam inations cannot b e carried ou t b y using th e absorption m eth od it is possible to determ ine th e structure o f3d
level fo r M n2+ ions b y applying th e m eth od o f lum inescence e x c i ta tion spectra.2. Measurement method
T h e o p tica l glasses w ith an addition o f M n2O s m elted in the ironless ceram ics w ere used fo r exam inations. The con ten t o f M n20 3 in the respective glasses is th e fo llo w in g : 0 .5 % in F K 5 , 0 .1 % in B K 7 and S K 4. T h e glass samples w ere o f th e sizes 13 x l 4 x 3 0 m m and 0.5 x l 2 x 2 0 m m . The absorption m easurem ents w ere perform ed on a Specord TTV Y IS spectrophotom eter. T he m easurem ents o f the spectral lum inescence distributions w ere m ade in th e setup consisting o f H B O -5 0 spectral lam p, H g M on 436 and S IP 436 filters o f C. Zeiss m ake, th e UM-2 m onochrom ator, the M12 PC 51 p h o to m ultiplier o f spectral ty p e, S20/SbK N aC s cathodes o f th e sensitivity range 3 2 0 —760 m m , and the G l - B l p lotter w hich, bein g cou pled w ith a m on och rom a tor was able (beside drawing the graphs) to change the w avelength. The excited lum inescence was recorded in the setup com posed o f a halogen lam p (100 W ), deuter lam p, a SPM -2 m on och rom ator w ith a quartz prism , a P K attachm ent, interference and glass filters, an M12 P C 51 ph otom u ltiplier and a G l - B l plotter, w hich being again coupled w ith a m on och rom ator changed the excitin g w avelength.
T h e lum inance tem perature o f th e halogen lam p was 3200 K , its spectral distribution I 0(A) was the same as th at o f b lack b o d y o f th e same tem perature. In the near ultraviolet Z0(A) takes v e ry low values, and w ith increasing ligh t w avelength (in accordance w ith th e P la n ck form ula)
134
S. Gabala th e intensity becom es respectively higher. The application o f the steady geom etric w idth o f th e m on och rom ator slit, as it was in onr case, results in som e increase o f th e differences in th e value o f lum inescence excitin g ligh t flu x betw een th e near ultraviolet and th e visual range. T o ob ta in th e true distributions o f th e excitation spectra th e indication o f the p lo tte r should b e d ivid ed b yI 0(X)AX.
F o r exam in ation o f the excitation spectra o f M n2+ ions lum inescence th e deuter lam p was used. This allow ed to broaden the excitation spectral range (starting w ith 200 nm ) and to m ake it closer t o the true spectra. I n order to select th e intervals o f the lum inescence spectra w ith th e help o f filters th e generally used division was assumed, accordin g t o w hich th e lum inescence w ith in the w a v e le n g th 'ra n g e o f 50 0 -6 0 0 n m com es fro m th e M n2+ ions in th e tetrahedric com plexes, w hile th a t w ithin th e 600 -7 00 n m w avelength range is em itted b y the M n2+ ions in th e octahedrie com plexes.
Tig. 2. The spectral distribution of excitin g ligh t beam from th e deuter lam p (1) and halogen lam p (2), as related to the constant differences of th e ligh t w avelength
In th e figure 2 th e dependence betw een the ligh t flu x and th e w avelength is show n fo r a deuter lam ps (1) and a halogen lam p (2) m easured w ith a va cu u m V T h 5 therm oelem ent supplied w ith a quartz w in dow . T he results w ere referred t o a constant ligh t w avelength interval.
3. Results of examinations
T h e presentation o f th e results of exam ination w ill b e started w ith showing th e dependence o f th e op tica l density u p on th e w ave-num ber fo r the sam ples 14 m m th ick m ade o f F K 5 glass containing 0.5 % o f M n2O a and o f B K 7 and S K glasses w ith 0 .1 % con ten t o f M n 20 8. I n th e glasses F K 5 and B K 5 th e w ide absorption ban d o f M n3+ m a y b e n oted . In contrast to this, bands con n ected w ith the
d-d
transitions fo r M n2+ ions are p ra cti ca lly unn oticeable (fig. 3).F ig. 3. D ependence of the optical d ensity upon the w ave num ber for glasses activated w ith m anganese (d = 14 mm).
1 - FK 5 (0.5%Mn2O3), 2 - BK 7 (0.1%Mn2O3), 3 - SK 4 (0.1%Mn2O3)
In th e figure 4 the curves o f lum inescence spectral distribution are show n fo r the sam ple o f F K 5 , B K 7 and S K 4 glasses a ctiva ted b y m angane se ions. The spectral distribution o f m anganese lum inescence indicate
136
S. G
íbalaa pproxim ately th e p roportion s o f those ions depending on th e coordination. Thus th e results ob ta in ed fo r the B K 7 glass (curve 2) speaks fo r a decisive dom inance o f lum inescence c o m ing fro m the M n2+ (4) ions. In the S K 4 glass th e quantities o f M n2+(4) and M n2+(6) are com parable (curve 3), w hile in the F K 5 glass th e predom inance o f th e com pon en t com ing fro m M n2+(6) is v e ry distinct (curve 1). F o r these glasses th e intensity o f lum i nance due to M n2+ is different. The curves in fig. 4 are p lo tte d fo r different ranges o f th e p lotter sensitivity.
In th e figures 5 -7 th e excita tion spectra o f lum inescence are shown in th e spectral range corresponding t o th e M n2+(4) an d M n2+(6) bands. A b o v e 600 n m w avelength th e lum inescence was separated b y the edge filter (A > 600 nm ). F ro m the longer w avelength side the spectral range was restricted b y the lim it o f ph otom eter sensitivity. O n th e other hand, the spectral range o f M n2+(4) lum inescence was separated b y I F 550 interfe rence filter o f th e half-w idth 5.0 n m — fo r excita tion w ith a halogen lam p,
P ig. 4. The spectral distribution of lum inescence in glasses activated w ith m anganese
and b y a glass filter (green V G , 2 m m ) fo r the denter lam p. The excitation spectra are registered in th e graphs, w hich correspond t o M n2+(4) lum i nescence (curves 1) and to th e M n2+(6) lum inescence (curves 2). The curves o f .excitation spectra induced b y th e deuter lam p (la , 2a) as w ell as b y the halogen lam p (la , 2b) are presented fo r th e fix e d w id th o f the m on och rom ator slit. W h ile th e intensities o f the excitation b a n d fo r deuter lam p are closer th e true ones, th e intensities o f th e excita tion ban ds indu ced b y th e halogen lam p (for w hich an increase o f th e light flu x occu rs w ith the increase o f th e ligh t w avelength) are enhanced.
M any bands are observed in th e excitation spectra o f M n2+ lum inescence in the F K 5 glass (fig. 5). A ban d w ith a m axim um at 220 nm appears fo r
F ig. 5. The lum inescence excitation spectra for Mn2+ ions em bedded in F K 5 glass under th e influence of deuter lam p (a) and under th e influence of halogen lam p (b) la — green glass filter (VG-9), 2a — edge filter (600 nm), lb — interference filter (550 nm), 2b — edge filter (600 nm)
lum inescence w hen m easured a t the presence o f either a green filter or a red one. T he intensity o f this ban d is tw ice less fo r lum inescence o f M n2+(4) than th a t fo r lum inescence o f M n2+(6). T here exist several bands w ith in the interval 3 00 -4 00 nm , the highest intensity being associated to the ban d w ith th e m axim um position ed betw een 350 and 360 nm (422a [4D ] . W ith in this spectral region the intensity o f the ban ds is defin itely higher fo r lum inescence due to Mn2+ (6). H ow ever, the highest intensity fo r this glass corresponds to the bands w ith m axim a fo r th e w avelength ranging
138
S. Gabala betw een 420 and 430 n m{*Eg *A
lg) . I t m ay b e n o te d th at these bands are splitted. T h e bands, corresponding to the transitions to the sublevels4T 2g
and4T lg,
are v e ry low and w ide. O nly b y increasing the light flu x (halogen lam p) th ey m a y b e m ade w ell visible. A n increase o f the light flu x w ith th e increase o f th e w avelength causes an apparent shift o f these bands tow a rd th e longer w avelength.In the B K 7 glass th e con cen tra tion o f manganese ions is considerably low er (0.1 % ) and b y the same means th e ratio M n2+ (4) /M n2+(6) is different. H ere, the greatest num ber o f ions w ith fou r ligands occurs and therefore the ban d w ith m axim um at 220 nm fo r M n2+(4) has th e highest intensity (fig. 6). O n the other hand, th e excita tion spectrum fo r lum inescence o f M n2+(6) w ithin th e interval 2 0 0 -3 0 0 nm has tw o bands w ith m axim a at 220 and 275 nm , b u t o f low er intensity. W ith in the interval 30 0 -4 0 0 nm tw o bands o f m axim a at 350 -3 60 n m and 380 n m are visible and v e ry low intensity is observed if com pared to bands fro m th e interval 2 0 0 -3 0 0 nm . The bands
4E g 4A lg
(420-440 nm ) are also o f less intensity b u t their splitting m a y b e seen v e ry well. The ratios o f these bands are different fo r M n2+ (4) and M n2+(6).In th e case o f S K 4 (fig. 7) glass th e differences in th e excitation spectra distribution are observed betw een 20 0 -3 0 0 nm . H ere, there exists one
w ide ban d exten ded betw een 230 and 240 nm . In spite o f this, th e
*Ea *Alg
(420-440 nm ) are splitted to a less degree. This is p ro b a b ly due to th e excitation o f M n2+ in tw o coordinations.
4. Discussion o f results
T he lum inescence m eth od m a y b e used to exam ine the structure o f the 3d level in M n2+ ions and determ ine th e crystal field param eters in a m uch m ore p erfect w a y than it is allow ed b y absorption m easurements. The m eth od discussed m a y b e also applied t o exam ine the samples fo r w hich th e
d-d
transitions m a y n o t b e con trolled b y absorption m easurements.T h e positions o f the absorption bands o f M n2+ fo r the silicon glasses are know n fro m th e literature data. F o r instance, the b a n d con n ected w ith th e electron transition to th e levels
*Eg
[4$ ] and4A lg [*G]
is associated w ith the w ave num bers w ithin the region 23500-23800 cm 1 [1 ], while th e transition t o the level*Eg
[4D ] is associated w ith th e w ave num bers 28000 cm - 1 . These results are absolutely consistent w ith those obtained fro m th e excita tion spectra lum inescence. The exparim ental data fo r th e lu m inescence excita tion spectra o f M n2+ (4) and M n2+ (6) as w ell as the param eters o f th e crystal field are contained in table. In this table no excita tion bands are reported w hich w ou ld b e contained betw een 200 and140
S. GębalaT a b l e . W ave numbers for Mn2+ (6) and Mn2+ (4) and tbe param eters for the crystal field N otations F K 5 [cm “ 1] B K 7 [cm “ 1] SK 4 [cm - 1 ] Mn2+ (6) Mn2+ (4) Mn2+ (6) Mn2+ (4) Mn2+ (6) Mn2+ (6) *Tlg 4P 30300 29400 % 4L> 28170 28410 28090 28250 27930 28200 b 26400 27300 26300 27300 26400 27200 % 4G 23700 23580 23700 23500 23700 23600 ‘•¿ i, * 6 23250 23000 23000 23000 23400 23000 *T2g 4G 20400 20400 20400 'T la * O 21500 21500 20200 B 638 690 627 678 604 657 Dą 870 460 900 460 770 650
300 nm . T h e latter bands o f v e r y high intensities m a y b e ascribed t o con d u ction bands. T he respective calculations have b een carried o u t on th e base o f form ulae fro m the b ook s b y Ba b t e c k i [ 6 ] and Be k s t j k e b [ 7 ] .
T he results obtained in this w ork are (w ithin th e m easurem ent a ccu racy) consistent w ith th e literature data. T h e m easurem ent accu racy o f th e applied m eth od m a y b e im proved b y th e respective changes in th e m easu ring setup and b y perfection in g th e m easurem ent technique.
The splitting o f the
*Eg
and*Alg
ban ds indicate th a t there exist consi derable deviations fro m the octahedrical sym m etry. F o r instance, in th e B K 7 glass the difference betw een the w ave num bers fo r4E g [46]
and*Alg [*0]
bands is equal to 700 cm - 1 . T h e m eth od o f th e lum inescence excita tion spectra is th e simplest so far as th e separation o f th e com pon ents com ing fro m M n2+ (4) and Mn2+ (6) is concerned. Such a separation w ou ld n o t b e possible w ith th e help o f ab sorp tion measurem ents.References
[1] Ma r g a r y a n A . A ., et al., D oklady A N SSSE 221 (1975), 665.
[2 ] Ma r g a r y a n A . A ., Ka r a p e t y a n S. S., X l- t h International Congress on Glass, Prague 1977, Section A8, p. 79-85.
[3] Go r b a c h e v a N . A ., Kab a k o v a A . J ., Zh. Prikl. Spektr. 6 (1967), 478. [4] Sz ó r e n y j T ., Szollósy L ., Sz a n k a K ., P b ys. a. Cbem. Glass. 17 (1976), 104. [5] Go l d b e r g P ., Luminescence o f Inorganic Solids, Academ ic Press, N ew York,
London 1966.
[6] Ba r t e c k i A ., Spektroskopia elektronowa związków nieorganicznych i komplekso
wych, P W N , W arszawa 1971.
[7] Berstjker I. B ., Elektronnoe stroenie i svoistva koordinacionnykh soedinenii, Izd. K him iya, Leningrad 1976.
[8] Bo k sh a O. N ., GtRu m-Gzhimailo S. V ., Iesledovanie optichesTdch spektrov Tcristallov 8 ionam i gruppy zheleza p r i Tcomnatnoi i niskikh temperaturakh, Izd. N auka, Moskva 1972.
[9] Ko n s t a n t in o v a-Sh l e z in g e b M. A ., K h im iya lampových geterodesmicheskikh ljuminoforov, Izd. N auka, Moskva 1970.
[10] Appe n A . A ., K him iya stekla, Izd. K him iya, Leningrad 1974.
Received A p ril 15, 1980 ^ш йш иссцсвпш ш MC I и д и ш ю д ш ш ш х структуры УРОВНЯ 3d ИОНОВ Мп+2 В стеклах Показано, что в активированных стеклах Мп+2 полосы спектра поглощения, соответству ющие переходам d-d, не обнаруживаются, между прочим — из-за наличия полосы спектра поглощения Мп+3. Доказано, что применение в этом случае спектров возбуждения люми несценции позволяет успешно исследовать структуру подуровня 3d ионов Мп+2. Метод спектров возбуждения люминесценции обладает рядом преимуществ, позволяя, между про чим, несложным образом разделять Мп+ 2 (4) и Мп+2 (6).