`
`5139
`
`and it is, therefore, unlikely that after detachment of
`the first H atom, a second H atom would be eliminated.
`Little can be said concerning the modes of formation of
`the pentenes, except that they are related to those of
`hydrogen.
`
`ACKNOWLEDGMENT
`
`The authors would like to express their gratitude to
`the referee of this paper for his comprehensive com(cid:173)
`ments on the discussion.
`
`THE JOURNAL OF CHEMICAL PHYSICS
`
`VOLUME 47, NUMBER 12
`
`15 DECEMBER 1967
`
`Investigation of Some Cea+-Activated Phosphors
`
`G. BLASSE AND A. BRIL
`Philips Research Laboratories, N. V. Philips' Gloeilampenfabrieken, Eindhoven, Netherlands
`(Received 14 August 1967)
`
`The fluorescence of a number of new Ce8+-activated phosphors is described and discussed. Only host
`lattices with a sublattice consisting of trivalent lanthanide ions are used to avoid charge compensation of the
`Ce3+ ion. Usually the Ce3+ emission is in the near-ultraviolet region. Y8Al.012-Ce and SrY20 c Ce, however,
`show emission in the visible region with a maximum in the green. Conditions for visible Ce3+ emission are
`indicated, viz., large crystal-field splittings (YaAI6Ot2-Ce) or a large Stokes shift (SrY20cCe). In a number
`of cases we were able to observe all crystal-field components of the excited 5d level of Ce3+. The cubic
`crystal-field splitting of the 5d level varies strongly with host lattice from 7000 to 14000 cm-I . The position
`of the center of the 5d levels in oxides is about 30% lower than in the free ion. For some phosphors we
`observed more than one emission band at room temperature. This is due to fluorescence from higher excited
`levels. Efficient energy transfer from Ce3+ to Cr3+ was observed in Y3Al.012.
`
`I. INTRODUCTION
`
`II. EXPERIMENTAL
`
`The fluorescence of Ce3+-activated compounds is well
`known.1,2 Usually the emission consists of a broad band
`with two peaks in the long-wavelength ultraviolet
`region. Since the Ce3+ ion has a 4j1 configuration, the
`ground state consists of a doublet (2F5/2 and 2F7/2).
`The lower excited states are the crystal-field components
`of the sa configuration. The transition 4f--+6s and
`charge-transfer transitions are at considerably higher
`energies3 and are not considered in this paper. The Ce3+
`emission is due to a Sd--+4f transition. This is an allowed
`electric dipole transition and therefore the decay time
`of the fluorescence is very short ( < 10-7 sec) .1,4
`Recently we have also reported on a Ce3+-phosphor
`which emits in the visible, viz., YsAl5012-Ce.4 This
`peculiar behavior prompted us to study the Ce3+
`fluorescence in other oxides of the trivalent lanthanides
`(La3+, Gd3+, Y3+, Sc3+). Surprisingly enough, these
`materials have not widely been investigated, although
`these host lattices offer the possibility of introducing
`Ce3+ without charge compensation.
`The present paper gives the results of this investiga(cid:173)
`tion. The complete crystal-field splitting of the sa
`configuration of the Ce3+ ion was found in a number of
`cases. Requirements for Ce3+ emission in the visible
`region are indicated.
`I A. Bril and H. A. Klasens, Philips Res. Rept. 7, 421 (1952).
`2 J. W. Gilliland and M. S. Hall, Electrochem. Tech. 4, 378
`(1966) .
`3 E. Loh, Phys. Rev. 154, 270 (1967).
`• G. Blasse and A. Bril, Appl. Phys. Letters 11, 53 (1967).
`
`Samples were prepared by firing intimate mixtures of
`high-purity oxides (or compounds, which on decompo(cid:173)
`sition yield oxides) at appropriate temperatures in
`nitrogen. The Ce3+ ions were introduced by replacing
`part of the trivalent lanthanide ions by Ce3+ ions. The
`Ce3+ concentration was 1-2 at. %. Products were
`checked by x-ray analysis. The performance of the
`optical measurements has been described previously.4-6
`
`III. REFLECTION AND EXCITATION SPECTRA
`
`A. Results
`Tables I-III contain the absorption bands of the
`Ce3+ ion in a number of host lattices in the uv region.
`These host lattices do not absorb in the optical region
`studied except for the Ga-containing oxides. The data
`were obtained from excitation and diffuse reflection
`spectra. In some cases only a continuous absorption
`was found, so that the absorption bands remain un(cid:173)
`known (Table II). Figures 1-3 show the excitation and
`diffuse reflection spectra of Y AlaB4012 : Ce, YaAhOI2: Ce,
`and ScBOa: Ceo These figures illustrate two types of
`excitation spectra, viz., those in which the quantum
`efficiency of the Ce3+ fluorescence is roughly independ(cid:173)
`ent of the exciting wavelength (YAlaB40 12:Ce, Fig. 1)
`and those in which the quantum efficiency decreases
`
`6 A. Bril and W. L. Wanmaker, J. Electrochem. Soc. Ill, 1363
`(1964) .
`6 G. Blasse and A. Bril, J. Chem. Phys. 46, 2579 (1967).
`
`LOWES 1026, Page 1
`
` Reuse of AIP Publishing content is subject to the terms: https://publishing.aip.org/authors/rights-and-permissions. Downloaded to IP: 198.65.204.101 On: Mon,
`14 Nov 2016 15:03:24
`
`
`
`5140
`
`G. BLASSE AND A. BRIL
`
`TABLE I. Efficiencies for ultraviolet and cathode ray excitation, positions of emission and excitation bands
`and Stokes shift of the emission of some Ce3+ phosphors.
`
`Composition"
`
`'1(%)b
`
`qma.(%)°
`
`Emission bands
`(10' cm-I)d
`
`Excitation bands
`(lOS cm-I)-
`
`Sc2Si2O,Ce
`ScBOrCe
`YBOrCe
`LaBOrCe
`YPO,-Ce
`LaPOcCe
`Y AlaB40 12-Ce
`Y.Al.01:rCe
`YOCI-Ce
`LaOCI-Ce
`LaOBr-Ce
`
`1.5
`2
`2
`0.2
`2.5
`2.2
`2
`3.5
`3.5
`0.4
`0.2
`
`65
`70
`50
`35
`30
`40
`40
`",70
`60
`30
`25
`
`25.4; 29.4
`26.8; ",30.5
`23.8; 25.4
`26.6; 28.4; 31.5
`28.3; 30.0
`29.8; 31.5
`27.2; 29.1
`18.2; 27.8; "'29.2
`26.3
`27.8
`22.8
`
`29.0; 33.3; ",43.5
`28.0;31.2;36.1;38.5
`27.4; (",29); 40.8; (",43.5)
`30.8; 36.9; 41.5
`32.8; (34.2); 39.6
`36.2; 38.8; 41.6
`31.0; 36.6; 39.2
`22.0;29.4; ",37; ",44
`31.6;35.8
`(",35); 35.8; 39.7
`28.4; 34.7
`
`Stokes
`shift
`(lOS cm-I) I
`
`3.6
`1.2
`2.0
`2.4
`2.8
`4.7
`1.9
`3.8
`5.3
`5.2
`5.6
`
`• Cea+ concentration is 1-2 at.%.
`b Radiant efficiency for cathode ray excitation (20 kV).
`° Quantum efficiency for excitation in the lowest excited level.
`d Position of the emission peaks.
`
`e Position of the excitation bands; shoulders between brackets .
`I Stokes shift of the emission. i.e .. difference between the positions of the
`lowest excited level and the short-wavelength component of the emission
`from this level.
`
`exciting wavelength
`drastically with decreasing
`(Y3AhOl2-Ce, Fig. 2 and ScB03-Ce, Fig. 3). All the
`excitation spectra we obtained belong to either the
`former or the latter type. In this section we discuss
`the number and position of the absorption bands and a
`possible explanation for the two types of excitation
`spectra. A comparison between the excitation spectra
`of CeH- and TbH-activated phosphors is also included.
`
`B. Discussion
`The occurrence of more than one CeH absorption
`band in the region between 25 000 and 50 000 cm-l is
`due to the crystal-field splitting of the 5d (2D) state.
`As argued before in the case of TbH we can neglect the
`effect of spin-orbit splitting, since the spin-orbit param(cid:173)
`eter S5<l is about 103 cm-l, whereas the crystal-field
`splitting amounts to 1()4 cm-l or more.7 For nearly all
`compounds mentioned in the tables the symmetry of
`the lanthanide site is known, so that it is possible to
`compare the experimentally observed number of Sd
`levels and the number expected from the site symmetry.
`Not in all cases a reasonable agreement was found.
`
`Consider, for example, YOCl-Ce and YPOcCe. The
`respective site symmetries are C4v and D2d, so that 3
`and 4 levels are expected. The experimental numbers
`(Table I) are 2 and 3, respectively. It cannot be
`excluded, however, that there are levels beyond the
`spectral limit of our apparatus ('"'-'220 m,u) or that lower
`symmetrical splittings remain hidden in the broad
`bands observed experimentally. We never observed
`more bands than expected from the site-symmetry.
`In Table IV we have tabulated those host lattices for
`which the number of levels observed experimentally
`agrees reasonably well with the number expected from
`site symmetry.
`From these data the cubic crystal-field splitting 6.
`and the center of gravity of the Sd levels can be found,
`at least approximately. In Y3AhOl2-Ce the CeH ion
`occupies a distorted cube, so that the Sd level is split
`into a lower doublet and a higher triplet by the cubic
`component of the crystal field. Noncubic components
`cause a further splitting. Unfortunately we observed
`only two components of the triplet, which causes some
`uncertainty in the value of 6. and the position of the
`center of gravity. In ScB03-Ce and SC2Si207-Ce the
`
`:;1
`
`t
`~
`~50
`~
`"" ;i
`t 0
`
`350
`
`300
`
`250
`A(mp)_
`
`100
`
`'0'
`~
`.~
`50~
`~
`t
`
`0
`200
`
`lfOO
`'8
`!
`~ 50
`j
`t 0
`
`600
`
`100
`
`i
`.s
`50j
`
`t
`
`FIG. 1. Diffuse reflection spectrum (broken line) and relative
`excitation spectrum (solid line) of the Ce3+ fluorescence of
`YO.99CeO.OlAlzB4Ot2.
`
`7 G. Blasse and A. Bril, Philips Res. Repts. 22, 481 (1967).
`
`FIG. 2. Diffuse reflection spectrum (broken line) and relative
`excitation spectrum (solid line) of the Ce3+ fluorescence of
`Y2.94Ceo.ooAI,OI2. The dotted line represents the diffuse reflection
`spectrum of unactivated Y,AI,OI2.
`
`LOWES 1026, Page 2
`
` Reuse of AIP Publishing content is subject to the terms: https://publishing.aip.org/authors/rights-and-permissions. Downloaded to IP: 198.65.204.101 On: Mon,
`14 Nov 2016 15:03:24
`
`
`
`I N V EST I GAT ION 0 F SO M E C e s+ - ACT I V ATE D PH 0 S P H 0 R S
`
`5141
`
`TABLE II. Position of emission and excitation bands and Stokes shift of the emission of some Ce3+ phosphors
`(inefficient at room temperature).'
`
`Composition
`
`Y,AJ,09-Ce
`SrY,O,-Ce
`Y 20 a-Ce
`~aY02-Ce
`NaGdO,-Ce
`LaAl03-Ce
`
`Emission bands
`(103 em-I)
`
`Excitation bands
`(l(}1 em-I)
`
`Stokes shift
`(loa em-I)
`
`(~28); 31.2
`17.4b
`19.6; (~24)
`(~21.3); 22.8
`21.3
`
`32.6; ~34; ~41.5
`25.2; ?
`?
`
`"-'4.5
`~8
`
`24. 2; 31. 2; ~40°
`
`• For explanation of columms see Table 1.
`
`C Data from reflection :3vectrum.
`
`Ce3+ ion occupies distorted octahedra, so that the triplet
`is lower. In YAlaB 40 12-Ce the Ce3+ ion is in a trigonal
`prism of 02- ions. The 2D level splits into three levels
`with symmetry species E', Ai', and E" (31000,36600,
`and 39200 cm-1, respectively). This corresponds to a
`cubic crystal-field splitting of 6500 cm- 1 (see Appendix).
`Finally Table IV contains analogous data for
`SrF2 : Ce3+ recently given by Loh.3 Although our results
`are not accurate, Table IV gives some interesting results.
`The 4f-5d distance is 51 000 cm-1 in the free Ce3+ ion.
`This value decreases to 48 000 cm-1 in fluorides and to
`roughly 35 000 cm-1 in oxides, a reduction of 6% and
`30%, respectively. This fact must be ascribed to the
`reduction of the interelectronic repulsion parameters
`by covalency effects (nephelauxetic effect8). Within
`the group of oxidic host lattices there is even a variation
`of this value. For example, ScBOa-Ce has the center of
`the 5d level at "-'32 500 em-i. On the other hand, the
`value in the case of LaPOcCe must be considerably
`higher than the 35000 cm-1 mentioned above (see
`Table I). In our paper on Tb3+-activated phosphors7
`we have already drawn attention to this fact. The
`Tb3+-activated phosphors show excitation bands due to
`
`TABLE III. Efficiency for cathode-ray excitation and position
`of the two lower absorption bands and of the emission band in
`the visible region for some Ce3+-activated garnets.
`
`Composition
`
`'1(%)
`
`Emission
`Position lower
`absorption bands bands in the
`(difference hetween
`visible
`brackets)
`region
`(l03 em-I)
`(lOS em-I).
`
`YI.5GdI .• AI50'2-Ce
`Y3AI50 12-Ce
`Y3Al4GaOI2-Ce
`YaAlaGa2012-Ce
`YaAJ,Gaa012-Ce
`YaGa.OI2-Ce
`
`... b
`3.5
`1.9
`1.7
`1.2
`
`21.5,29.6(8.1)
`22.0,29.4(7.4)
`22.5,29.1(6.6)
`23.0,28.8(5.8)
`23.3,28.6(5.3)
`23.8,28.1(4.3)
`
`17.4
`18.2
`18.5
`19.2
`19.6(18.3)
`••• c
`
`• Shoulders between parentheses.
`b Emission contains Eul + contributions due to the starting material
`Gd,O •• which contains a slight amount of europium.
`C No fluorescence.
`
`8 C. K. J¢rgensen, Absorption Spectra and Chemical BOllding
`(Pergamon Press, Inc., New York, 1962).
`
`4f--t5d transitions in the Tb3+ ion. These bands are at
`higher wavenumbers than those of Ce3+ and partly
`beyond the shorter-wavelength limit of our apparatus,
`so that it was not possible to find the center of gravity
`in the case of Tb3+.
`Table IV also shows that the cubic crystal-field
`splitting .1 varies strongly. For a cube the absolute
`value of .1 is expected to be 8/9 times the value of .1
`for an octahedron with the same distances between
`central ion and ligand.9 With this in mind all values of
`.1 in Table IV can be compared with each other. The
`.1 in YaAh012-Ce is relatively large, whereas .1 in
`YAlaB 40 12-Ce is relatively small. That .1 is strongly
`influenced by the nature of the next-nearest neighbors
`is a well-known phenomenon.lO,n We further note that
`.1 in the fluorides has the same order of magnitude as in
`the oxides. This has also been found for the transition(cid:173)
`metal ions.8
`The value of the noncubic crystal-field splitting varies
`considerably. For some host lattices it is possible to
`compare the splitting of the lower cubic crystal-field
`level of the Sri state of Ce3+ and the 4f5d state of TbH .
`The agreement is very good (Table V).
`The phenomenon that the efficiency of Ce:H- -activated
`phosphors is more or less independent of the position
`of the excitation band in some cases (Fig. 1), but
`decreases strongly for excitatiOl1 in bands with higher
`
`~100 t
`~
`g
`.~ 50
`1;
`<t
`
`t 0
`
`400
`
`100
`l
`.Ii
`50~
`d:
`t
`
`0
`200
`
`FIG. 3. Diffuse reflection spectrum (broken line) and relative
`excitation spectrum (solid line) of the Ce3+ fluorescence of
`SCO.99CeO.OlBO, .
`
`9 J. S. Griffith, The Theory of Transition-Metal Ions (Cambridge
`University Press, Cambridge, 1961).
`10 J. Ferguson, K. Knox, and D. C. Wood, J. Chern. Phys.
`35,2236 (1961);37, 193 (1962).
`11 G. masse, J. Inorg. Nuc!. Chern. 29,1817 (1967).
`
`LOWES 1026, Page 3
`
` Reuse of AIP Publishing content is subject to the terms: https://publishing.aip.org/authors/rights-and-permissions. Downloaded to IP: 198.65.204.101 On: Mon,
`14 Nov 2016 15:03:24
`
`
`
`5142
`
`G. BLASSE AND A. BRIL
`
`T ABLE IV. The sa levels and cubic crystal-field splitting (.:l) for CeB+ in several host lattices. All values in loa em-I.
`
`Composition
`
`YsAlo01.--Ce
`Y AlsB4OI.--Ce
`ScBOrCe
`
`Sc,Si,O,Ce
`
`SrF.--Ceb
`Free ionb
`
`Coordination
`of CeB+
`
`5d levels
`(observed) •
`
`Deduced
`cubic levels
`
`Cubic crystal-
`field splitting (.:l)
`
`Spherical
`5d
`
`Distorted cube (D,)
`Trigonal prism (DaA)
`Distorted octahedron
`(Dad)
`Distorted octahedron
`(C,)
`Distorted cube
`
`22.0 29.4137 44
`31.0 36.6 39.2
`28.031.2136.1 38.5 ",29.5 (I,.), 37.3 (e.)
`
`25.7 (e.), ",40 (t,.)
`
`29.033.3143.5
`
`",31 (t,.), "'43.5 (e.)
`
`33.6 48.8150.3 53.4
`
`41. 2 (eu), 52.4(tt.)
`
`",14
`",6.5·
`",8
`
`"'12.5
`
`11.2
`
`",34.5
`35.4
`",32.5
`
`",35
`
`48
`51
`
`• Vertical bar indicates separation between lower and higher cubic levels.
`b After Ref. 3.
`
`• Value of the octahedral crystal-field splitting derived from the position
`of the levels for Ce'+ in prismatic coordination using a point-charge model
`(see Appendix).
`
`wavenumbers in other cases (Figs. 2 and 3), is probably
`due to the position of the configuration coordinate
`curves relative to each other. Let us call the configura(cid:173)
`tion coordinate curves of the ground state and the two
`lowest excited states g, el, and e2, respectively (el is
`below e2)' If the curve el crosses the curve e2 within the
`configuration coordinate curve of g, excitation into el
`and e2 may both result in efficient fluorescence from el.
`If on extrapolation the curves el and e2 should cross
`outside the curve g, however, excitation into el gives
`efficient fluorescence from el, but excitation into e2
`cannot give efficient fluorescence from el, since either
`fluorescence from e2 or radiationless transition to g
`results. If the "crossings" of e2 with el and g approach
`each other, weak fluorescence from el and e2 may be
`expected upon excitation into e2. From this simple
`model it follows that a Ce3+-activated phosphor with an
`excitation spectrum in which the efficiency decreases
`with increasing excitation energy, must show fluores(cid:173)
`cence from e2 upon excitation into e2 (or higher levels) .
`This was in fact observed (see below).
`
`IV. EMISSION
`
`A. Results
`
`The positions of the emission peaks of our samples
`are collected in Tables I-III. Figures 4-11 show the
`spectral energy distribution of the emission of a number
`
`of compounds. Usually the emission is in the near-uv
`region, but not always: Ya(AI, Ga)s012-Ce and SrY20 c
`Ce (Figs. 5 and 10) show emission in the visible. The
`half-width value of the emission bands does not depend
`on their position and amounts to roughly 4000 cm-l
`for all materials. In some cases the· emission band is
`split, the difference between the two"peaks being some
`2000 cm-l (LaBOa-Ce, Fig. 4; YBOa-Ce, Fig. 9;
`YAIaB40 12-Ce and YP04-Ce, Table I). In many cases
`more than one emission band is present (LaBOa-Ce,
`Fig. 4; YaAloOI2-Ce, Fig. 5; ScBOa-Ce, Fig. 6; Sc2Si207-
`Ce, Fig. 7; and Y4AI20 9-Ce, Fig. 11).
`
`B. Discussion
`
`The fluorescence emission of the Ce3+ ion originates
`from a transition from one or more of the sd levels
`(often only the lowest level) to the 2F ground state.
`Since the ground state of the Ce3+ ion is a doublet
`(2Fs/2 and 2F7/2) with a separation of about 2000 cm-I,
`each emission band is expected to show two peaks as
`found by us in many cases (compare also Ref. 12). This
`doublet character of the emission band depends on
`temperature and Ce3+ concentration (self-absorption)
`and is not always found.la
`In the case of oxidic host lattices this emission usually
`peaks in the near uv. As a matter of fact there are two
`
`I
`
`100,-----,....------,
`t
`
`I t
`
`FIG. 4. Spectral energy distribution of the fluorescence of
`Lao .•• CeO.01BOs under 254-m~ excitation. In Figs. 5-12 the!radiant
`power per constant wavelength interval in arbitrary units (1) is
`plotted along the ordinate.
`
`FIG. 5. Spectral energy distribution of the fluorescence of
`Y2.94Ceo.OtIAI.012 (solid line) and Y2,94Ceo.OtIAI,GaaOu (broken
`line) under 254-m~ excitation.
`
`12 F. A. Kroger and J. Bakker, Physica 8, 628 (1941).
`13 Th. P. J. Botden, Philips Res. Rept. 7, 197 (1952).
`
`LOWES 1026, Page 4
`
` Reuse of AIP Publishing content is subject to the terms: https://publishing.aip.org/authors/rights-and-permissions. Downloaded to IP: 198.65.204.101 On: Mon,
`14 Nov 2016 15:03:24
`
`
`
`I N V EST I GAT ION 0 F SO M E C e a+ - ACT I V ATE D PH 0 S PH 0 R S
`
`5143
`
`TABLE V. Splitting of the lower cubic crystal-field componeI?-t
`of the Stllevel of CeI+ and the 4f75d level of Tbl + (all values In
`1()3 em-I).
`
`Host lattice
`
`YaAL.012
`YaGa6012
`YBO,
`S~Si207
`
`• From Tables I and nl.
`b From Ref. 7.
`
`7.4
`4.3
`",,1.6
`4.3
`
`7.5
`4.0
`1.4
`4.6
`
`cases in which the emission will be at lower wave num(cid:173)
`bers.
`
`(a) If the lowest 5d level lies exceptionally low.
`(b) If the Stokes shift of the emission is exceptionally
`large.
`
`We have found one example of each of these two cases.
`In YgAloOl2-Ce the lowest 5d level is exceptionally
`low, viz., at 22000 cm-I (see Table I). The YgAl60 IrCe
`is the only Ce3+-activated phosphor that has not a
`white, but a yellow body color, whereas the host lattice
`itself is white. Therefore, the emission lies also at very
`low wavenumbers. Note in Table I that the Stokes
`shift of this emission does not have a value deviating
`from what is normally found. Table IV shows, why the
`lowest 5d level lies so very low. The center of the 5d
`levels of Ce3+ in YgAl50 I2 has a normal value. In Sec. III
`we have found that the crystal-field splitting in YgAh012
`is very large. Moreover, the lower cubic level is the CD
`level, which is iLl below the center. In an octahedro?
`the t2D level is lower (only fLl below the center). ThIS
`means that the lower cubic crystal-field component is
`relatively low in YgAloOI2 (25700 cm-I, compared with
`",,30 000 cm-I for the lower t20 level of ScBOrCe and
`SC2Si20,Ce). There is yet another effect, viz., the large
`g+ in YgAloO,2,
`non cubic splitting of the CD level of Ce
`which brings the lowest level at 22 000 cm-I • Visible
`Ce3+ emission can therefore be expected, if the 5d
`crystal-field splitting and the lower symmetrical split(cid:173)
`ting are large. Cubic eight coordination is also favorable
`to obtain visible emission, because the lower cubic level
`.
`will then be at relatively low energies.
`It is obvious that long-wavelength emission will.also
`
`lO'o..--~------------,
`
`I
`
`t
`
`1001~---=,------;; ......... , - - - - - - - - -
`
`I t
`
`350
`
`4-50
`400
`--A(mpJ
`
`FIG. 7. Spectral energy distribution of the fluorescence of
`SC1.9SCeo.02Si20 7 under 254-m~ excitation .
`
`result in the case of a very large Stokes shift. This is
`the case for SrY20 C Ce. The lowest 5d level is not
`exceptionally low (25200 em-I), but the Stokes shift
`is very large (",8000 cm-I, compare Tables I and II),
`so that visible emission results (Fig. 10), Such a large
`Stokes shift can only occur if the difference between
`the equilibrium distance of the excited state and that
`of the ground state is larger than usual. EisewhereI4 we
`have argued that the increase of the equilibrium dis(cid:173)
`tance upon excitation will be large if the ions of the
`host lattice are rather weakly bounded together, i.e.,
`if large and low-charged ions are involved (this implies
`ions with relatively highly polarization or, as they are
`sometimes called, "soft" ions'S) • Such a situation exists
`in SrY20 4• We also noted weak fluorescence in the visible
`region in the case of Y20 g-Ce, Na Y02-Ce, and NaGd02-
`Ce (Table II). These are also lattices with large cations.
`The same holds for YOCI-Ce, LaOCI-Ce, and LaOBr(cid:173)
`Ce (Table I), but the oxychlorides emit in the uv. Note,
`however, that the Stokes shift of these materials is also
`large.
`Another consequence of a large increase of the
`equilibrium distance upon excitation is a low quenching
`temperature of the fluorescence. For SrY20 C Ce we
`found that the light output at 200 0 K is only 20% of
`the output at loo oK. Such a marked temperature
`dependence agrees with out model. In a number of
`other lattices with large ions (Sr2LaAI06-Ce, SrLaAI04-
`Ce, LaAlOa-Ce) the Ce3+ emission was absent or only
`very weak, even at liquid-nitrogen temperature.
`
`1001r---------~,---------------,
`
`I
`
`t
`
`50
`
`~).(mP)4-50
`
`500
`
`the fluorescence of
`FIG. 6. Spectral energy distribution of
`SCO.99CeO.OIBO. under 254-m,u excitation.
`
`FIG. 8. Spectral energy distribution of the fluorescence of
`YO.99CeO.OIOCI under 254-m,u excitation.
`- - - - -
`14 G. Blasse and A. Bril, Z. Physik.. Chern. (Frankfur~). "T).1e
`influence of crystal structure on the fluorescence of oXldlC DlO(cid:173)
`bates and related compounds" (to be published).
`15 Compare C. K. J¢rgensen, Coord. Chern. Rev. 1, 164 (1966).
`
`LOWES 1026, Page 5
`
` Reuse of AIP Publishing content is subject to the terms: https://publishing.aip.org/authors/rights-and-permissions. Downloaded to IP: 198.65.204.101 On: Mon,
`14 Nov 2016 15:03:24
`
`
`
`5144
`
`G. BLASSE AND A. BRIL
`
`I
`
`rool.---__ ~----------~
`t
`
`50
`
`FIG. 9. Spectral en(cid:173)
`ergy distribution of the
`fluorescence of
`YO.99Ceo. 01BO,
`under 254-mlL excita(cid:173)
`tion.
`
`Finally we note that also the nature of the ligands
`influences the position of the emission bands. For
`fluorides the center of the 5d level is at high energies
`resulting in emission at high energies (LaFa-Ce: 3S 700
`and 33300 cm-I, SrF2-Ce: 32200 cm-12). In the less
`electronegative chlorides
`the emission
`lies
`lower
`(LaCla-Ce: 28 500 cm-I 2). A similar phenomenon
`occurs in LaOCI-Ce and LaOBr-Ce (Table I). If the
`CI- ion of LaOCI is replaced by the less electronegative
`Be ion, the emission shifts to the visible region (due to
`the decreasing 4f-Sd energy difference).
`In many cases more than one Cea-+- -emission band is
`present. The band at higher wavenumbers is undoubt(cid:173)
`edly due to fluorescence from higher 5d levels. This
`phenomenon is not unknown, but has only been found
`at low temperature.12 ,16 In our investigations it was
`observed at room temperature in the case of LaBOa-Ce,
`Sc2Si207-Ce,
`and
`ScBOa-Ce, Ya(AI,GahOI2-Ce,
`Y4AI20 9-Ce (Tables I and II). It is noteworthy that
`the two emission bands are better separated, if the
`energy difference between the two lowest 5d levels
`increases. The YaAlsOI2-Ce shows this very markedly
`(Table I) : the lowest levels are at 22000 and 29400
`cm-I, the emission bands at 18200 and 27800 (and
`",29200) cm-I .
`The compounds with more than one emission band
`also exhibit the type of excitation spectrum in which
`the efficiency decreases on increasing excitation energy.
`The compounds with only one emission band have the
`other type of excitation spectrum. In the preceding
`section it was already shown that the former combina(cid:173)
`tion is not unexpected. At the moment our knowledge
`of the configuration coordinate curves of systems like
`those under consideration is inadequate to discuss this
`fOq----_______ --:;:,.....-'""' _______ --,
`I
`t
`
`50
`
`t~V---5~OO~--,5t,~,----OOO~h---6~~~---J~
`---A (m}')
`
`FIG. 10. Spectral energy distribution of the fluorescence of
`SrY!.9SCeO.0204 at nOK under 365-lllp excitation.
`
`16 G. R. Fonda and F. J. Studer, J. Opt. Soc. Am. 38, 1007
`(1948) •
`
`phenomenon further. As shown in Fig. 11, Y~209-Ce
`shows mainly fluorescence from e2 (31200 cm-I) and
`only weak fluorescence from el (28000 cm-I). In
`Sc2Si20,Ce both emission bands have equal intensity,
`in LaBOa-Ce, YaAloOI2-Ce, and ScB03-Ce the el
`fluorescence is more intense. The intensity ratio of the
`two emission bands may be obscured by self-absorption
`or energy transfer of the type e2+g-tg+el' The latter
`mechanism results in el fluorescence upon excitation into
`
`V. THE SYSTEM Ya(Al,Ga)SOI2-Ce
`
`The system Ya(AI, Ga)SOI2-Ce was studied in more
`detail. Results are given in Table III and Fig. 6. The
`Y3Ga;,OI2-Ce does not fluoresce under long- or short(cid:173)
`wave uv or cathode-ray excitation. The energy of all
`these types of excitation is mainly absorbed by the
`host lattice. The absence of Ce3+ fluorescence is probably
`due to inefficient energy transfer from lattice to activa(cid:173)
`tor. This fact may also explain the decrease of the
`efficiency for cathode-ray excitation with increasing Ga
`content (Table III). Figure 5 shows that with increas(cid:173)
`ing Ga content the ratio of uv to visible emission de-
`
`~0r---'~--------,
`
`I t
`
`50
`
`FIG. 11. Spectral energy
`distribution of the fluores(cid:173)
`cence of Yu6Ceo.04A120 9
`under 254-m,u excitation.
`
`creases. This again may be due to absorption of the uv
`emission of the Ce3+ activator by the Ga3+ ions of the
`host lattice.
`Table III shows that the splitting of the lower cubic
`crystal-field component decreases regularly from 7400
`cm-I for Ce3+ in YaAloOI2 to 4300 cm-I for Ce3+ in
`Y;;GasOI2. The visible emission band shifts therefore to
`the blue with increasing Ga content. That the deviation
`from cubic site symmetry is larger in YaAlsOI2 than in
`YaGasOI2 has also been found from studies on Pr3+_
`activated garnets,17 Tb3+-activated garnets,7 and from
`crystallographic investigations.ls As argued above, the
`large deviation from cubic symmetry in the case of
`YaAlsOI2 shifts the fluorescence emission into the visible
`region. Replacing y3+ by Gd3+ results in an even larger
`splitting (see Table III).
`Elsewhere we have reported that under cathode-ray
`excitation YaAloOI2-Ce shows only a very weak uv
`emission in addition to the efficient visible fluorescence. 4
`In the present model this can only be explained by
`assuming that the excitation energy of the lattice IS
`transferred directly to the lowest excited state (el)'
`17 F. N. Hooge, J. Chern. Phys. 45, 4505 (1966).
`18 F. Euler and J. A. Bruce, Acta Cryst. 19, 971 (1965).
`
`LOWES 1026, Page 6
`
` Reuse of AIP Publishing content is subject to the terms: https://publishing.aip.org/authors/rights-and-permissions. Downloaded to IP: 198.65.204.101 On: Mon,
`14 Nov 2016 15:03:24
`
`
`
`I N V EST I GAT ION 0 F SO M E C e 3 + - ACT I V ATE D PH 0 S P H 0 R S
`
`5145
`
`VI. ENERGY TRANSFER FROM Ce3+ TO
`OTHER IONS
`Energy transfer from Ce3+ to Ce3+ and to SmH, Eu3+,
`TbH, and Dy3+ in some of the host lattices mentioned
`in this paper has been described and discussed else(cid:173)
`where.19 Transfer from Ce3+ to Cr3+ in YAhB 40 12 was
`mentioned recently.20 Because of the anomalous emis(cid:173)
`sion of CeH in Y3AI.OI2 energy transfer from Ce3+ to
`Sm3+, EuH, Tb3+, Dy3+, and Cr3+ was also studied in
`this host lattice. This was done by exciting Y3AloOI2-
`Ce, X (X= Sm3+, Eu3+, Tba, Dy3+, or Cr3+) in the
`CeH-excitation band at 29400 cm-l • The absorption
`of CeH in this region is much more intense than that of
`X, so that nearly all excitation energy is absorbed by
`CeH. If X is a lanthanide ion, the emission under these
`circumstances is still mainly Ce3+ emission, which indi(cid:173)
`cates an inefficient transfer from CeH to the other four
`lanthanide ions; if X= Cr3+, however, the emission is
`mainly Cr3+ emission,21 so that the Ce3+---+Cr3+ transfer
`is efficient. These results agree with our previous con(cid:173)
`siderationsI9.2o : if the CeH emission overlaps the absorp(cid:173)
`tion lines of the 41-4f transitions of other lanthanide
`ions (which is here the case), energy transfer occurs by
`exchange interaction and is not efficient. In the case of
`CeH-Cr3+ in YaAlsOl2 the visible and uv emission of
`the Ce3+ ion overlaps the Cr3+ absorption bands con(cid:173)
`siderably: the visible mission overlaps
`the broad
`absorption band of Cr3+ between 15 000 and 20 000
`cm-l (4A~4T2), the uv emission that between 21000
`and 27 000 cm-l (4 A~4TI) .22 This gives rise to efficient
`energy transfer as in YAhB 40 12: Ce, Cr.20 In the borate,
`however, only the 4A2---+4TI absorption is overlapped by
`the Ce3+ emission.
`
`Note added in proof: We have recently found that
`at 4.2°K the visible emission band of Y sAIo0trCe con-
`18 G. Blasse and A. Bril, J. Chern. Phys. 47, 1920 (1967).
`20 G. Blasse and A. Bril, Phys. Letters, 25, A29 (1967).
`21 The emission of YaA1;OlrCr has been reported by G. Bums,
`E. A. Geiss, B. A. Jenkins, and M. 1. Nathan, Phys. Rev. 139,
`A1687 (1965).
`!2 D. L. Wood, J. Ferguson, K. Knox, and J. F. Dillon Jr., J.
`Chern. Phys. 39,890 (1963).
`
`I
`
`,
`
`~r-E"
`___ -l~ ... - -A f
`, , ,
`\""---E'
`
`FIG. 12. Trigonal-prismatic coordination of Ce3+ in Y tU,B.OI1
`(left-hand side) and splitting of the 2D level of Ce3+ under D3/o
`symmetry (right-hand side). The black circle represents Ce3+,
`open circled represents 02-.
`
`sists also of two bands located at 19 100 and 17 400
`cm-I, respectively.
`
`ACKNOWLEDGMENTS
`
`The authors are greatly indebted to Miss M. P. Bol,
`Miss D. J. M. Trumpie and Mr. C. J. Loyen, Mr. J. A.
`de Poorter, and Mr. J. de Vries for their assistance in
`the experimental part of the work.
`
`APPENDIX
`
`In a trigonal prism (D31o symmetry) the fivefold
`degenerate 2D level of the Ce3+ ion splits into three
`levels (E', AI" and E", see Fig. 12). The energy differ(cid:173)
`ence between E' and E" is
`(47r )-1/2 (0.37')'2Q1''''2+ 1. 78'Y.r0f4)
`
`and that between E' and AI'
`(47r)-1/2(0.S~20f2-1.78'Y 4Of"),
`
`whereas in a regular octahedron the splitting is
`
`Here'Y are parameters characteristic of the surroundings
`and fk= (Rn! I rk I Rn1 ), where Rnl denotes the radial
`one-electron wavefunctions. Using these formulas it is
`possible to deduce the cubic crystal-field splitting from
`the levels observed experimentally for the prismatic
`coordination.
`
`LOWES 1026, Page 7
`
` Reuse of AIP Publishing content is subject to the terms: https://publishing.aip.org/authors/rights-and-permissions. Downloaded to IP: 198.65.204.101 On: Mon,
`14 Nov 2016 15:03:24
`
`