`
`Vizio EX1021 Page 0001
`
`Vizio EX1021 Page 0001
`
`
`
`Philips Journal of Research
`
`Philips _Journal of Research, published by the Philips Research Laboratories,
`Eindhoven, The Netherlands, is a bimonthly publication containing papers on
`research carried out in the various Philips laboratories. Volumes 1 to 32 have
`appeared under the name of Philips Research Reports.
`
`Annual subscription price for Volume 36 is Dfl. 70, payable in advance.
`Payments should be made only after receipt of an invoice. Correspondence
`should be addressed to: Philips Journal of Research, Philips Research
`Laboratories, Building WBp, Room No. 42, Eindhoven, The Netherlands.
`
`Editor-in-chief: J. de Groot
`
`Cover design based on a visual representation of the sound-wave associated with the spoken
`word "Philips".
`
`© N.V . Philips' Gloeilampenfabrieken, Eindhoven, The Netherlands, 1981. Articles or illustra(cid:173)
`tions reproduced in whole or in part must be accompanied by a full acknowledgement of the
`source: Philips Journal of Research.
`
`Vizio EX1021 Page 0002
`
`
`
`Philips J. Res. 36, 15-30, 1981
`
`R 1031
`
`COLOURSHIFT OF THE Ce3+ EMISSION IN
`MONOCRYSTALLINE EPITAXIALL Y GROWN
`GARNET LAYERS
`
`by J. M. :ROBERTSON, M. W. VAN TOL, W. H: SMITS
`and J . P. H. HEYNEN
`
`Abslracl
`Starting with Ce-doped Y,AI 50 12 grown by liquid-phase epitaxy, the colour
`of the cerium emission was shifted by making various substitutions in the
`garnet lattice. The substitutions were made with the purpose of changing
`the crystal field acting on the Ce3+ ion in the garnet lattice. The light
`outputs and the colour points of the resulting phos phors have been
`measured and the influence of flltering away part of the emission has been
`calculated. It proves to be possible to obtain green and red phosphors
`with compatible light output and the right colour points for colour
`TV purposes.
`
`1. Introduction
`
`Recently it has been demonstrated that liquid-phase epitaxy (LPE) could be
`used to grow monocrystalline layers of garnets which could be used as screens
`2
`in cathode ray tubes 1
`). In this manner it is possible to make miniature
`3
`'
`'
`display tubes in which high excitation densities can be used (ca. 10 W /cm2
`)
`without thermal quenching or "burning in" of the phosphor layer. The
`epitaxially grown Ce3+: Y3Al50 12 (Y AG) gives a cathode ray efficiency of
`ca. 5% 3
`) and has a high quenching temperature (310 °C), however it has a
`broad band emission with maximum intensity at ca. 555 nm resulting in a
`colour which is too yellow to be used as a green component in a colour TV
`system. From the literature 4
`) it is known that the emission band of the Ce3+:
`YAG can be shifted towards shorter wavelengths by substituting other ions
`into the garnet structure. In the present paper we shall show that it is possible
`to shift the Ce3+ emission in LPE grown YAG so that it ha:; the desired colour
`coordinates for a green TV tube. We also investigated the possibility of
`making shifts in the emission so that blue and red tubes could be made by
`similar use of molecular engineering.
`
`2. General method
`
`The luminescent properties of Ce3+ doped phosphors have been studied in
`detail in ref. 5. In the ground state the Ce3+ ion has one electron in the 4f state.
`
`PhilipsJourn•l of Research Vol. 36 No. I 1981
`
`15
`
`Vizio EX1021 Page 0003
`
`
`
`J. M. Robertson, M. W van To!, W H. Smits and J. P. H. Heynen
`
`The ground state is split up in a doublet eF7 12 and 2F5 12) with an energy dif(cid:173)
`ference of about 2200 cm- 1
`• At low temperatures this doublet results in two
`peaks in the emission band 6
`); at room temperature it is one of the main causes
`of the width (at half maximum) of about 105 nm of the emission band. In the
`·first excited state the electron occupies a 5d state. The interaction between the
`. wave function of the electron in the 5d state and the crystal field results in a
`broadening of this state and in a splitting into five energy levels. The emission
`of the UV- or cathode ray excited Ce3+ ion is due to the transfer of an electron
`from the lowest 5d level to the 2F ground state; transitions from higher 5d
`states have a relatively low probability at room temperature owing to the
`strong interaction between the 5d states. A change in crystal field will give a
`larger or smaller splitting of the 5d state, which results in a shift of the
`emission spectrum to longer or shorter wavelengths. In most crystal structures
`the crystal field splitting is small and the Ce3+ ion shows an emission in the
`blue or UV, but as a consequence of the extraordinary large crystal field
`splitting in YAG, the Ce3+ ion emits in the green-yellow in this host lattice with
`a maximum intensity at 555 nm .
`Holloway and Kestigian 6
`) showed that if Lu3+ was substituted for Y3+
`in Y AG powders, the emission of the Ce3+ ion shifts to shorter wave(cid:173)
`lengths. Other workers 7
`) substituted Gd3+ or La3+ for Y3+ and shifted the
`8
`9
`'
`'
`emission to longer wavelengths, while the substitution of Ga3+, ln3+ or Sc3+
`for Al3+ on an octahedral site shifted the Ce3+ emission to shorter wave(cid:173)
`lengths 6
`7
`9
`•
`•
`).
`These results form the basis for the empirical rule, used in the present work:
`increasing the diameter of the ion on the dodecahedral (Y3+) site increases the .
`crystal field splitting, while an increasing diameter on the octahedral (Al3+)
`site has the reverse effect. In this study double substitutions on both the
`dodecahedral and octahedral sites were investigated, as well as the substitu(cid:173)
`tion of a smaller ion on the tetrahedral (Al3+) site.
`
`3. Experimental techniques
`
`The liquid-phase epitaxial growth technique, the cathodoluminescence
`efficiency measurement in a demountable cathode-ray system and the
`measurement of the spectra have been described in detail elsewhere 3
`) . The
`substrates on which the epitaxy was performed were Y3Al50 12 (12.001 A),
`YaGasOI2 (12.280 A) and Y3Al3 •5Ga1 .50 12 (12.119 A) 10), usually 25 mm in
`diameter, 700 ~ thick and cut and polished in [Ill] orientation.
`The chemical analysis of the layers was performed with the help of an
`electron beam microprobe, calibrated against a set of carefully seiected
`standards.
`
`16
`
`Philips Journal of Research Vol. 36 No. _t
`
`1981
`
`Vizio EX1021 Page 0004
`
`
`
`Co/ow·shift of the Ce3+ emission
`
`4• Results and discussion
`4.a. Green shift
`In order to have the right hue for the green component the emission must
`be shifted to the. short wavelength side of the spectrum, which means that Lu
`and/or Ga must be substituted. If such a phosphor is to be grown epitaxially
`on a commercially available substrate (such as YAG), this substitution must
`result in a lattice parameter which differs less than 0.018 A from that of the
`substrate. This can be ensured by a simultaneous substitution of Lu3+ for ya+
`and Ga3+ for Al3+ in the correct ratio. Since the lattice parameters of YAG,
`LuAG and YGG are 12.001, 11 .908 and 12.280 A respectively, a simultaneous
`substitution of Lu and Ga gives
`d (Y3-yLUyA15-xGax012) = 12.001 - 0.031 y + 0.056 X.
`For no lattice mismatch between film and substrate (YAG), i.e. for good
`epitaxy, Lu3+ and Ga3+ must build in a ratio of 9 : 5.
`The first experiment was set up to check this ratio and to establish whether
`the shift in the Ce emission is a linear function of the Lu and Ga substitutions
`and to investigate whether the effects of the substitutions are independent
`or not.
`We started with a melt composition
`PbO: B20a: Y20a: Al20a: Ce02 = 450.0: 11.65:2.57:5 .87:0.5 (g)
`and a growth temperature of ca. 1030 °C. Lutetium oxide and Ga2 0 3 were
`added in consecutive steps so that the lattice parameter of the film was almost
`constant. Table I shows the results of light output (L), colour point and chem(cid:173)
`ical analysis of the layers . The emission spectrum of sample 8 (doped with the
`
`100
`
`~
`
`~
`:;
`.::;.
`
`" 0 :c
`
`~ 50
`
`__. >-..(nm)
`Fig. I. Emission spectra of Ce3+:Y3 AI5 0 12 and Ce3+:Y3 AI5 0 12 + Lu + Ga (sample 8 in table I).
`
`l'hlllps Journal of Research Vol. 36 No . I 198 1
`
`17
`
`Vizio EX1021 Page 0005
`
`
`
`J. M. Robertson, M. W. van To!, W. H. Smits and J. P. H. Heynen
`
`TABLE I
`Results -of adding Lu20 3 and Ga20 3 to a Ce3+:YaA15012 melt
`
`total added
`
`sample
`
`Ga20a Lu20 3
`(g)
`(g)
`
`L
`(J.!W/sr)
`
`A max
`(nm)
`
`colour
`X
`
`point
`y
`
`Leq
`(lm/W)
`
`Ce3+:YaA150I2
`1
`2
`3
`4
`5
`6
`7
`8
`
`0.9
`1.9
`2.9
`2.9
`4.9
`4.9
`4.9
`5.9
`
`1.03
`1.03
`1.03
`1.53
`1.53
`2.03
`2.53
`2.53
`
`185
`141
`171
`166
`162
`156
`153
`143
`115
`
`555
`551
`550
`5.47
`546
`543
`542
`539
`538
`
`0.414
`0.394
`0.387
`0.387
`0.377
`0.365
`0.361
`0.350
`0.349
`
`0.552
`0.557
`0.558
`0.559
`0.561
`0.558
`0.558
`0.558
`0.557
`
`466
`466
`465
`467
`469
`458
`460
`457
`457
`
`Formula of the layers after analysis
`
`sample
`
`Ce
`
`y
`
`Lu
`
`Al
`
`Ga
`
`0
`
`I
`2
`3
`4
`5
`6
`7
`8
`
`0.039
`0.008
`0.008
`0.008
`0.007
`0.007
`0.009
`0.008
`
`2.410
`2.432
`2.422
`2.203
`2.209
`2.084
`1.887
`1.903
`
`0.546
`0.569
`0.563
`0.783
`0.777
`0.910
`1.103
`1.080
`
`4.734
`4.787
`4.681
`4.671
`4.460
`4.442
`4.426
`4.339
`
`0.189
`0.203
`0.325
`0.333
`0.545
`0.555
`0.573
`0.668
`
`12.0
`12.0
`12.0
`12.0
`12.0
`12.0
`12.0
`12.0
`
`highest quantity of Lu and Ga) is compared with that of Ce3+ in pure YAG in
`fig. 1. It is clear that the emission (Amax) shifts from 555 nm to 538 run.
`The samples 3, 5, 6 and 8 were grown with exactly zero mismatch and have
`a ratio between Lu and Ga of about 8 : 5 which is very near to the calculated
`value of 9: 5. It is also seen that the light output decreases from about
`185 J.!W/sr for Ce3+:YAG .to 115 J.!W/sr for sample 8. This is in agreement
`with previous work in which the gallium garnets were found to be less effi-
`
`18
`
`Philips Journal oiRcscorch Vol. 36 No. I 1981
`
`Vizio EX1021 Page 0006
`
`
`
`Colow"Shijt of the Ce 3+ emission
`
`555 ~-------:~ Lu203 - addition
`A.,.,oK
`
`lnm) I 550
`
`545
`
`540
`
`535
`0~~0~.1--702~~0~.3~0~4--705~706~~0~.7~
`{Ga203/
`
`Fig. 2. The shift of Amax of the Ce3+ emission as a function of the mole fraction of Ga 20 3
`in the melt.
`
`cient host lattices under cathode-ray excitation than the aluminium garnets 3
`) .
`In figure 2 the maxima of the Ce3+ emission spectra are plotted as a function
`of the mole fraction of Ga 2 0 3 in the melt. It is clear that the addition of Ga
`gives a linear shift of Amax to shorter wavelengths. Between certain samples
`only Lu 20 3 was added and this results in a vertical translation in the graph .
`From this graph it follows that there is a 30 nm shift per mole fraction
`of Ga20 3 in the melt. Chemical analysis of the layers gives the shift per mole
`fraction of Ga 20 3 in the layer which amounts to 14 nm per substituted
`Ga atom. This results corresponds exactly with a Amax at 541 nm for
`Ce3+:Y3Ga 1Al 40 12 given in ref. 9. A comparison between the mole fraction
`of Ga 20 3 in the melt and the relative concentration in the layer 11
`) gives an
`increasing segregation coefficient for Ga3+ of 0.27 to 0.38 with increasing
`Ga concentration.
`The same experimental results can be plotted as a function of the mole
`fraction of Lu 20 3 in the melt (fig. 3). This time the vertical shifts are due to
`addition of Ga2 0 3 to the melt. The shift in Amax is 13 nm per mole fraction of
`Lu20 3 • The shift per substituted ion in the layer follows again from chemical
`analysis. It amounts to 10 nm per Lu ion, which is only half the value reported
`in ref. 6 for Lu substitution in YAG. This indicates that the shifts due to Lu3+
`and Ga3+ are not completely independent and the shift in emission due to the
`Lu3+ is larger in pure YAG than in gallium substituted YAG. The segregation
`coefficient of 1.03 agrees very well with the results obtained for Lu in
`iron garnets.
`In figure 4 the colour points of the samples are given in a chromaticity
`diagram. The dotted curve in this figure represents the locus of colour points
`
`Philips Journal of Resenrch Vol. 36 No . J
`
`1981
`
`19
`
`Vizio EX1021 Page 0007
`
`
`
`J. M. Robertson, M. W. van To!, W. H. Smits and J. P. H. Heynen
`
`r----:::::::::;:::;;:;;P, Ga2 03 - addi tion
`555
`Amax
`,..
`(nm)
`550
`
`• !
`~
`~
`
`I
`
`545
`
`540
`
`535
`0~--~0~1 --~0-2 --0~.3---0~.4---0~5--~0~6--~0 -7 -
`/ Lu201 1
`
`Fig . 3. The shift of A max of the Ce3+ emission as a function of the mole fraction of Lu20 3 in
`the melt.
`
`0.70
`
`y
`
`550nm
`
`1 Ama, = 555nm
`2 Ama<= 549nm
`3 Ama<= 538nm
`
`- - x
`Fig . 4. Colour shift of Ce:YAG with Lu and Ga substitutions; ----calculated shift, o-o-o
`measured shift, ---- after filtering away the red part of the Ce emission.
`
`calculated as if the emission spectrum of Ce3+ in YAG was shifted with respect
`to the abcissa in energy units, i.e. as if the crystal field splitting of the 5d state
`is changed while its width is kept constant.
`Since the desired colour point x = 0.29, y = 0.60 cannot be obtained by
`shifting Amax alone, the red part of the emission must be filtered out.
`The effect of an ideal filter, ·i.e. a filter that transmits all radiation with a
`wavelength below a certain cut-off value and blocks all radiation above that
`
`20
`
`Philips Journal of Research Vol . 36 No.1 1981
`
`Vizio EX1021 Page 0008
`
`
`
`Colow·shift of the Ce 3+ emission
`
`value, is represented by the dash-dotted curves. Obviously the remaining
`emission has a-lower radiant intensity and a higher luminous equivalent than
`the original unfiltered radiation.
`The luminous intensity, obtained by multiplying the radiant intensity and
`the luminous equivalent, decreases only slightly because the blocked (red)
`radiation has a low lumen content.
`The colour points on the drawn line have the same hue as the desired colour
`but a different colour saturation; according to this figure, i.e . assuming
`constant energy efficiency of the phosphors, minimum filtering and just
`acceptable colour saturation is obtained with an original spectrum with Amax at
`538 nm . However, since the energy efficiency decreases with increasing Ga
`concentration (table 1), the maximum luminous intensity is obtained for a
`phosphor with Am ax = 549 nm, resulting in a colour which is slightly more
`saturated than necessary . The . corresponding ideal filter should block all
`radiation above 582 nm, which results in 65 mlm/ sr at 0.1 W cathode-ray
`excitation (5 .507o energy efficiency) .
`
`4.b. Blue shift
`It was tried to find out how far the Ce3+ emission could be pushed to shorter
`wavelengths in the garnet host-lattice in order to make a blue tube. We started
`with Ce3+: Lu 3Al 50 12 grown onto YAG and not Ce3+: Lu 3Ga50 12 since the
`gallium garnets are known to be inefficient host lattices for cathode lumi(cid:173)
`) . In the Ce3+: Lu 3Al50 12 some of the Al3+ ions on the octahedral
`nescence 3
`site were substituted by the large ions Ga3+, ln3+ or Sc3+. Since no substrates
`were available to match the small lattice parameters of these films we used
`YAG as the substrate. This means that cracked layers were produced and the
`measured light output was enhanced by multiple internal reflections and scat(cid:173)
`tering. Thus the values given in the table are only a qualitative indication of
`the efficiency of the layers .
`Starting with a melt composition of
`
`PbO: B20a : Lu20a: Al20 3 : Ce02 = 450 .0 : 11.65:3.09:3 .91 :0.5 (g)
`
`and a growth temperature of 1050 ~C Ga 20 3 and ln 20 3 were added to the
`melt. Table II gives a summary of the results . In pure Ce3+: Lu 3Al50 12 the
`emission has a Am ax of 529 nm in agreement with ref. 4. The Ga 20 3 shifted
`this Amax to 519 nm (sample 8 with 8. 1 (g) Ga20 3 added) . At this point a
`second emission band at 405 nm appears . Addition of 1.5 g ln20 3 gives a
`further shift to 513 nm and a higher intensity of emission at 405 nm.
`The emission spectra of the samples 8 and 11, i.e. those with maximum Ga
`and In content, are presented in fig . 5 and compared with emission of Ce3+
`
`Philips Journnl of Research Vol. 36 No. I 1981
`
`21
`
`Vizio EX1021 Page 0009
`
`
`
`J. M. Robertson, M . W van To!, W H. Sm its and J. P. H. Heynen
`
`TABLE II
`Results fur Ga and In substituted Ce 3+:Lu 3 Al5 0 12 films
`
`total added
`
`sample Ga20a
`(g)
`
`In20 3
`(g)
`
`L
`(~-tW/sr)
`
`A max
`(nm)
`
`colour
`X
`
`point
`y
`
`Leq
`(lm/W)
`
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`
`3.6
`3.6
`4.1
`5.1
`7.1
`7.1
`8.1
`8.1
`8.1
`8.1
`
`-
`-
`-
`-
`-
`-
`-
`0.5
`1.0
`1.5
`
`-
`-
`-
`157
`175
`79*
`168
`45
`29
`22
`
`527
`524
`524
`523
`522
`521
`519+405
`519+405
`517 +405
`513 +405
`
`0.295
`0.288
`0.286
`0.278
`0.269
`0.266
`0.263
`0.255
`0. 247
`0.239
`
`0.536
`0.533
`0.529
`0.529
`0.512
`0.500
`0.495
`0.472
`0.434
`0.395
`
`427
`420
`418
`421
`402
`396
`387
`362
`323
`286
`
`* broken sample
`
`Formula after analysis
`
`sample
`
`Ce
`
`Lu
`
`Al
`
`Ga
`
`In
`
`0
`
`3
`5
`8
`11
`
`0.007
`0.008
`0.005
`0.006
`
`2.995
`3.001
`3.030
`2.967
`
`4.463
`4.302
`3.963
`3.972
`
`0.533
`0.687
`0.999
`1.025
`
`0.0
`0.0
`0.0
`0.029
`
`12.0
`12.0
`12.0
`12.0
`
`in pure YAG. The samples with intermediate In content show intermediate
`shifts of the main band and intermediate intensities of the 405 nm peak. The
`latter peak does not shift as a function of Ga or In concentration; therefore we
`suppose that this emission does not originate in the Ce3+ ion.
`The values of Amax of the Ce3+ emission are plotted as a function of the
`mole fraction of Ga20 3 in the melt in fig. 6. The addition of Ga20 3 to the
`Ce3+: Lu 3Al 50 12 melt gives a linear shift of Amax to shorter wavelengths of
`145 nm per mole fraction Ga20 3 • This shift is again about half the shift caused
`by Ga substitution in Y3Al 50 12 which again stresses the point that the shifts
`
`22
`
`Philips Journal of Research Vol. 36 No. 1 1981
`
`Vizio EX1021 Page 0010
`
`
`
`Colow·shift of the Ce3+ emission
`
`700
`
`0
`
`400
`
`550
`
`600
`______. \ I nm!
`
`Fig. 5. The emission spectra of Ce3+: YAG, Ce3t Lu 3Al 50 12 + Ga and Ce3+: Lu 3(AIGa) 50 12 +ln.
`
`530
`
`Arnot
`lnm)
`
`515
`
`0
`
`0.1
`
`0.2
`
`0.3
`
`0.4
`
`06 0.7
`0.5
`/Ga20 1 }
`
`Fig. 6. Shift of Amax of the Ce3+ emission as a function of the mole fraction of Ga 20 3 in the
`Ce:Lu 3Al 50 12 melt.
`
`due to addition of lutetium and gallium are not independent and that the
`crystal field splitting shows a more complex behaviour than expected. The
`segregation coefficient of Ga3+ is the same as before, changing from 0.33 to
`0.38 with increasing Ga concentrations.
`In order to measure the influence of indium separately, indium was added
`to a gallium-free lutetium aluminium garnet melt with composition
`
`PbO: B20a: Lu20 3 : Al20 3 : Ce02 = 450.0: 11.65: 3.09: 6.00:0.5 (g)
`
`The results are given in table III. In contradistinction to the Ga3+ substitution
`the In3+ substitution hardly influences the shift of the main Ce3+ peak. Micro(cid:173)
`probe analysis showed that at the highest In 20 3 concentration 1.50Jo of the
`Lua+ ions are substituted by ln3+ while the Al3+ concentration is not affected
`
`l'hllips Journal of Research Vol . 36 No. I 1981
`
`23
`
`Vizio EX1021 Page 0011
`
`
`
`J. M. Robertson, M. W van To/, W H. Smits and J. P. H. Heynen
`
`TABLE III
`Results for In substituted Ce3+: Lu 3Al5 0 12 films
`
`sample
`
`ln 20 3
`(g)
`
`L
`(J..LW/sr)
`
`A max
`(nm)
`
`2
`3
`4
`6
`
`0.5
`1
`2
`2.5
`
`57.5
`35
`17.5
`12
`
`533
`532+405
`531 +405
`532+405
`
`colour
`X
`
`0.324
`0.316
`0.310
`0.305
`
`point
`y
`
`0.550
`0.541
`0.517
`0.493
`
`Leq
`(lm/W)
`
`441
`427
`359
`374
`
`so that the In builds into the dodecahedral sites instead of the octahedral sites
`in the garnet.
`With Sc 20 3 substitution a similar thing happens. Analysis shows that 400Jo
`of Sc3+ replaces Lu 3+ on the dodecahedral sites and 60% replaces Al3+ on the
`octahedral sites. The substitution has little effect on the shift of the Ce 3+ emis(cid:173)
`sion. This suggests that either the effects of Sc3+ substitution on both sites can(cid:173)
`cel each other or it is not the diameter of the substituting ion which influences
`13
`the crystal field, but some other property such as its polarizability 12
`'
`).
`The colours of the emission by Ga, In and Sc substituted garnets are
`presented in the chromaticity diagram of fig. 7. Owing to the presence of the
`peak at 405 nm, the colour points of the In and Sc containing samples deviate
`
`0.9
`
`0.1
`
`0oL-~~~~~~0.4~~0~5~~0.6~~0~7-
`--- x
`Fig. 7. The colour shift of Ce3+: Lu 3Al 50 12 due to Ga3+ and ln3+ substitutions; x--- X
`e :Lu 3Al 50 12 + Ga, o---o Ce:Lu 3(AIGa)s0 12 +In,
`Ce;Y3Al 50 12 + Lu,Ga, o---o
`o- o Ce:Lu 3Al 50 12 +In and o- o Ce:Lu 3 AI 50 12 + Sc.
`
`2C
`
`24
`
`Philips Journal or Research Vol. 36 No. I 1981
`
`Vizio EX1021 Page 0012
`
`
`
`Co/ow·shijt of the Ce 3+ emission
`
`from the dashed line, which again represents the "theoretically shifted"
`y AG: Ce emission.
`Another point to notice is that In3+ and Sc3+ have a very strong negative
`influence on the cathode-ray efficiency even stronger than gallium.
`
`4.c. Red shift
`The motive for producing a cerium doped red phosphor is to circumvent
`the problem of saturation of the relatively slow Eu 3+ -doped phosphors at
`high energy densities 14
`). It seems possible to produce a red shift in the
`cea+: Y3AJ50 12 emission by reversed substitutions compared to those which
`produced the blue/ green shift, namely, a larger ion on the dodecahedral site
`and/or a smaller ion on the octahedral site in the garnet lattice.
`We again started with Ce: YAG and substituted Gd3+ for y a+. This produced
`a rapid increase in lattice parameter of the films and resulted again in facetted
`growth so that the light output is only a qualitative measure for the efficiency .
`[twas impossible to double-substitute with a small trivalent ion on the tetrahedral
`or octahedral sites in this case in order to maintain a constant lattice parameter.
`In table IV the results are summarized for a Ce3+: Y3Al 50 12 melt to which
`Gd 20 3 had been added in small steps . The starting melt composition was
`PbO: B20a: Y20a: Al20a: Ce02 = 450.0: 11.65:3.80:6.52:0.5 (g).
`For 6 g Gd 20 3 added to the melt (sample 24) the shift of Amax was from 555 nm
`to 574 run. In fig. 8 the Amax are plotted as a function of the mole fraction of
`
`A max
`{nm}
`
`580
`
`575
`
`570
`
`560
`
`0
`
`0.1
`
`06
`0.5
`{C'(}20Jl
`[Gd2 0J]+[Y;>OJ}
`~ig. 8. The shift of Amax .of the Ce3+ emission as a function of the mole fraction of Gd 20 3
`m the melt.
`
`0.2
`
`0.3
`
`0.4
`
`0.7
`
`PhlllpsJournalofRcsenrch Vol.36 No. I 1981
`
`25
`
`Vizio EX1021 Page 0013
`
`
`
`J. M. Robertson, M. W. van To/, W. H. Smits and J. P. H. Heynen
`
`TABLE IV
`Results for Gd substituted Ce 3+: Y3Al 5012 melt
`
`total added
`
`sample
`
`Gd203
`(g)
`
`Al203
`(g)
`
`L
`(J.l.W/ sr)
`
`A max
`(nm)
`
`colour
`X
`
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
`0
`0.25
`0.5
`1
`1.5
`2
`3
`4
`4
`6
`
`6.52
`6.52
`6.52
`6.52
`6.52
`6.52
`6.52
`7.52
`8.52
`8.52
`
`139
`117
`144
`177
`179
`177
`213
`167
`172
`158
`
`555
`552
`554
`558
`560
`564
`567
`569
`573
`574
`
`0.418
`0.412
`0.420
`0.427
`0.438
`0.446
`0.453
`0.462
`0.469
`0.474
`
`point
`y
`
`0.545
`0.547
`0.544
`0.539
`0.534
`0.530
`0.525
`0.519
`0 .514
`0.511
`
`Leq
`(lm/ W)
`
`454
`450
`446
`436
`436
`429
`418
`411
`405
`400
`
`Formula after analysis
`
`sample
`
`Ce
`
`18
`21
`24
`
`0.009
`0.007
`0.010
`
`y
`
`2.589
`2.025
`1.587
`
`Gd
`
`Al
`
`0.403
`1.002
`1.395
`
`4.997
`4.964
`5.006
`
`0
`
`12.0
`12.0
`12.0
`
`Gd 20 3 in the melt. It is clear that Amax shifts linearly to longer wavelengths
`with 54 nm per mole fraction of Gd20 3 in the melt. These Amax shifts are in
`perfect agreement with a Amax of 575 nm for Ce3+: Gd~,5 Y~,5Al5 0 12 as
`mentioned by Blasse and Bril for powders 9
`). In this case the segregation
`coefficient of Gd is 0.94 which agrees very well with the value obtained with
`iron garnets.
`In fig. 9 the colour points of this gadolinium substituted series of layers are
`plotted in a chromaticity diagram. The colour shift follows a path parallel to
`the "theoretically shifted" emission spectrum of Ce3+: Y3Al 50 12 .
`A new melt was made from which pure Ce3+: Gd 3Al 50 12 films could be
`grown onto Y3Ga1.5Al1.50 12 substrates. The Amax of the Ce3+ emission was
`
`26
`
`Philips Journnl of Research Vol. 36 No. 1 1981
`
`Vizio EX1021 Page 0014
`
`
`
`Colourshift of the Ce3+ emission
`
`570nm
`
`Ce,Y:JAI5012
`
`-~---,
`
`',, 580nm
`Ce Gd3AI5012 o , ,
`
`' ~,
`
`0.60
`
`y
`
`1 0.55
`
`050
`
`Ol5
`
`040
`
`045
`
`050
`-
`
`x
`
`055
`
`Fig. 9. Colour shift of Ce.: YAG to the red; ---- calculated shift, •-• Ce: Y3AI 60" + Gd,
`x-x Ce:Y9AI 50 12 +Mg,St, 0 Ce:Gd 3AI 50 12 Umax = 587nm)and D Ce:Gd3(AI,MgMAI,St)sO"
`(Amax = 614 nm).
`
`100
`
`-
`A(nml
`Fig. 10. The emission spectra of Ce3+ in Y3AI 50 12 , (Y,Gd) 3AI 50 12 and Gd3AI 5 0 12 •
`
`found to be 587 nm which is equal to the value published by Benderskii et al. 8
`).
`The light output of these phosphor layers was 1821-J.W/ sr, but the layers were still
`facetted owing to the difference in lattice parameters of the film and substrate.
`In fig . 10 the Ce3+ emission is shown in YAG, YAG + Gd and in Gd 3Al 50 12 •
`Another large ion that might enter dodecahedral sites in the garnets is La3+.
`This was also tried, but it was found to have a limited effect on the Ce3+ emis(cid:173)
`sion in the YAG and GdAG layers .
`
`Philips Journal of Research VoL 36 No. 1 1981
`
`27
`
`Vizio EX1021 Page 0015
`
`
`
`J. M. Robertson, M . W. van Tot, W. H. Smits and J. P. H. Heynen
`
`According to the empirical substitution rule a shift of the Ce3+ emission to
`longer wavelengths might a1so be achieved by the substitution of a smaller ion
`than Al3+ in the tetrahedral site. Since no trivalent ion is available we tried to
`simultaneously substitute Mg2+ on an octahedral site and Si4+ on a tetrahedral
`site. In this way the .ionic radius of the tetrahedral site ion decreases. Addition
`of 2.75 g Si0 2 and 1 g MgO to the standard Ce3+: YAG melt shifts the
`maximum of the Ce3+ emission from 555 nm to 579 nm resulting in a colour
`point of x = 0.468, y = 0.500.
`The addition of Mg and Si to the Ce3+: Gd 3Ga5 0 12 melt gives a similar shift
`in emission of 587 run to 614 nm. The emission spectra are plotted in fig . 11.
`The broadening of the emission spectra is more than can be explained by equal
`energy dispersion and is probably due to inhomogeneities in the distribution
`of substituted ions on a microscopic scale.
`Unfortunately, the cathodoluminescence efficiency drops dramatically after
`addition of even small amounts of MgO and Si0 2 to either melt. The sharp
`decrease in light output was accompanied by the appearance of a second
`(solid) phase in the melt consisting of small crystals with a cube growth
`habit. X-ray diffraction analysis showed that these crystals were of the type
`RE 2Al 40 9 (refs 15 and 16), while analysis by the electron beam microprobe
`indicated that the cerium is highly concentrated in these crystals and the con(cid:173)
`centration in the layers was almost zero. As a result, the low efficiency seems
`to be caused by a lack of cerium activator ions in the epitaxial layer . Another
`combination of substitutions, CaO + Si0 2 , to the melt showed a similar
`decrease in efficiency as well as the appearance of some "cubic-like" crystals.
`Since it appears not to be simple to shift the Ce3+ emission to the red and
`maintain a reasonable output it is again of interest to calculate the luminous
`
`100
`
`>-lnm)
`-
`Fig. II. The effect on the Ce3+ eniission when Mg2+ an d Si<+ are substituted into Y3AI 50 12 and
`Gd3Al•O" .
`
`28
`
`Philips Journal of l!escarch Vol. 36 No. I 1981
`
`Vizio EX1021 Page 0016
`
`
`
`Colourshift of the Ce3+ emission
`
`intensities using "ideal" filters to obtain the desired chromaticity for the red
`component of a colour display. If this is done in the same way as for the green
`component, it turns out that Ce3+: Gd 3Al 50 12
`is a good choice with a
`luminous intensity of 22 Lm/sr at 0.1 W cathode-ray excitation, if the light
`output efficiency of the phosphor is equal to that of Ce: YAG (ca. 185 ~J.W/sr).
`This intensity is only 20o/o lower than the ideal combination of filter and shift,
`and is compatible with 68 mlm/sr obtained from the green tube at the same
`excitation.
`
`s. Conclusions
`The results of our studies show that the model of crystal field splitting of the
`5d level of the Ce3+ ion is valid for the epitaxially grown garnets and that it can
`be used to shift the emission over a broad part of the spectrum by adequate
`substitutions. The garnet lattice is not ideal for a blue Ce3+ phosphor since the
`natural large crystal field has to be artificially decreased to a large extent.
`However, there are plenty of other host lattices giving good Ce3+ blue
`phosphors such as Y2Si0 5 •
`It is not only the size of the ionic radius of the substituting ion on the
`octahedral site, but also some other property that causes the shift to shorter
`wavelengths. Additions of large ions such as Ga3+, and especially ln3+ and Sc3+
`have a drastic effect on the cathodoluminescent efficiency of these phosphor
`layers.
`Additions of double substitutions such as Mg2+ + Si4+ causes a shift to the
`red but it appears to be very difficult to incorporate enough Ce3+ in the Mg : Si(cid:173)
`substituted layers to maintain a high cathode-ray efficiency. This is caused by
`the formation of a second phase in the melt.
`Using the molecular engineering described above, it is possible to make a
`green and a red phosphor that are compatible in light output and useful for
`colour television displays if the colour point is corrected by some additional
`filtering.
`
`6. Acknowledgement
`
`P. A. Paans and M . Klerk 'are thanked for their work with the electron
`microprobe and J. Daams is thanked for his X-ray analysis of the crystals.
`
`Philips Research Laboratories
`
`Eindhoven, September 1980
`
`l'hlllpsJournal of Research Vol. 36 No. J 1981
`
`29
`
`Vizio EX1021 Page 0017
`
`
`
`J. M. Robertson, M. W van Tot, W H. Smits and J. P. H. Heynen
`
`REFERENCES
`1) J . M. Robertson and M. W. van Tol, Epitaxially grown monocrystalline garnet cathode
`ray-tube phosphor screens, Appl. Phys. Letts. 37 (5), 471, 1980.
`2) M. W. van To! and J. van Esdonk , A high luminescence, high resolution cathode ray tnbe
`for special purposes; IEEE Trans. ED to be published .
`3) J. M. Robertson, M. W. van To!, J.P. H. Heynen, W. H . Smits and T. de Boer, Thin
`single crystalline phosphor layers grown by liquid phase epitaxy, Philips J. Research to be
`Pl!blished .
`4) W. W. Holloway and M. Kestigian, On the fluorescence of cerium activated garnet
`crystals. Phys. Letters 2SA, 614, 1967 .
`5) G. Blasse and A. Bril, A new phosphor for flying spot cathode-ray tubes for yellow
`emitting Y3 AI 50 12 :Ce3+, Appl. Phys . Letts. 11, 53, 1967.
`6) W. W. Holloway and M. Kestigian, Optical properties of Ce-activated garnet crystals,
`J . Opt. Soc. Am. 59, 60, 1969.
`7) T. Y. Tien, E. F. Gibbons, R . G. Delosh, P. J. Zacmanides, D. E . Smith and H. L.
`Stadler, Ce3+ activated Y3Al 5 0 12 and some of its solid solutions, Electrochem. Soc. 120,
`278, 1973.
`8) L. A. Benderskii, A. F. Vedekhin, G. M. Topchiev, Z. P . Klimina, A. S.
`Plomod'Yalov and A. I. Panchenko, Synthesis and luminescence properties of solid
`solutions in the system YAG: Ce GdAG: Ce, Bull. Acad. Sciences USSR Phys. 38, 72, 1974.
`9) G. Blasse and A. Bril, Investigation of some Ce3+ -activated phosphors, J . Chern. Phys. 47,
`5139, 1967.
`10) Grown by J. Pistorius of this laboratory .
`") S. L. Blank and J. W. Nielsen, The growth of magnetic garnets by liquid phase epitaxy,
`J. Cryst. Growth 17, 302, 1972.
`12) G. Blasse, The absorption of the Cr3+ ion in host lattices containing pentavalent cations,
`J. Inorg . Nucl. Chern. 29, 1817, 1967.
`13) C. J. J~rgensen, Differences between the four halide ligands and discussion remarks on
`trigonal-bipyramidal complexes, on oxidation states, and diagonal elements of one electron
`energy, Coord. Chern. Rev. 1, 164, 1966.
`14 ) W. F. va n der Weg and M. W. van To!, Saturation effects of cathodoluminescence in rare
`earth activated epitaxial Y3AI 50 12 layers, Appl. Phys . Letts., to be published.
`15) C. D. Brandle and H. Steinfink, The crystal structure of Eu 4Al 20 9 , lnorg. Chern. 8,
`1969, 1320.
`16) 1. W. Reed and A. B. Chase, The unit cell and space group of 2Y20 3 : Al 20 3 , Acta Cryst.
`15, 312, 1962.
`
`30
`
`Philips Journnl of Resenreh Vol. 36 No. I 1981
`
`Vizio EX1021 Page 0018
`
`