983

Luminescent M40. Luminescent Materials

This chapter surveys the field of solid-stateluminescent materials, beginning with a discussionof the different ways in which luminescence canbe excited. The internal energy-level structures ofluminescent ions and centres, particularly rare-earth ions, are then discussed before the effects ofthe vibrating host lattice are included. Having setthe theoretical framework in place, the chapterthen proceeds to discuss the specific excitationprocess for photo-stimulated luminescenceand thermally stimulated luminescence beforeconcluding by surveying current applications,including plasma television screens, long-termpersistent phosphors, X-ray storage phosphors,scintillators, and phosphors for white LEDs.

40.1 Luminescent Centres ............................ 98540.1.1 Rare-Earth Ions ........................ 98540.1.2 Transition-Metal Ions ................ 986

40.1.3 s2 Ions ..................................... 98740.1.4 Semiconductors ........................ 987

40.2 Interaction with the Lattice .................. 987

40.3 Thermally Stimulated Luminescence ...... 989

40.4 Optically (Photo-)StimulatedLuminescence ..................................... 990

40.5 Experimental Techniques –Photoluminescence ............................. 991

40.6 Applications ........................................ 99240.6.1 White Light-Emitting Diodes

(LEDs) ...................................... 99240.6.2 Long-Persistence Phosphors ....... 99240.6.3 X-Ray Storage Phosphors ........... 99340.6.4 Phosphors for Optical Displays .... 99440.6.5 Scintillators .............................. 994

40.7 Representative Phosphors .................... 995

References .................................................. 995

Luminescent materials are substances which convert anincident energy input into the emission of electromag-netic waves in the ultraviolet (UV), visible or infraredregions of the spectrum, over and above that due to

Table 40.1 Types of luminescence

Designation Excitation Trigger Acronym

Photoluminescence UV, visible – PL

photons

Radioluminescence X-ray, gamma – RL

rays, charged

particles

Cathodoluminescence Energetic – CL

electrons

Electroluminescence Electric field – EL

Thermoluminescence Photons, Heat TSL

charged

particles

Optically/photo-stimulated Photons, Visible/IR photons OSL, PSL

luminescence charged

particles

black-body emission. A wide range of energy sourcescan stimulate luminescence, and their diversity providesa convenient classification scheme for luminescencephenomena, which is summarised in Table 40.1. Pho-

PartD

40

984 Part D Materials for Optoelectronics and Photonics

Luminescing state

Ground state

Excited state

P12

2

1

P31NR P31R

3

P23

Fig. 40.1 Optical pumping cycle for a generic photolumi-nescent system

toluminescence, where the luminescence is stimulatedby UV or visible light, is a widely used materials sciencetechnique for characterising dopants and impurities, andfinds applications in lighting technologies such as flu-orescent lamps. Radioluminescence involves excitationby ionising radiation, and is used in scintillators for nu-clear particle detection; the special case of stimulationby energetic electrons is called cathodoluminescence,the name arising from early atomic physics experi-ments involving gas discharges. The major applicationof cathodoluminescence is in cathode ray tubes for tele-vision sets and computer monitors. Electroluminescenceinvolves collisional excitation by internal electrons ac-celerated by an applied electric field, and with a muchlower energy than in the case of cathodoluminescence.Electroluminescence finds applications in panel lightingused in liquid-crystal display (LCD) back-planes, and inlight-emitting diodes.

There are other forms of luminescence which wemention for completeness but will not discuss further:bioluminescence and chemiluminescence where the en-ergy input is from chemical or biochemical reactions,sonoluminescence (sound wave excitation), and tribolu-minescence (strain or fracture excitation).

There are several books which describe the lu-minescence of materials in more depth than ispossible in a short article, for example that edited byVij [40.1] and the monograph by Blasse and Grab-maier [40.2].

The terms phosphorescence and fluorescence are of-ten used in connection with luminescent materials. Thisclassification is based on the time-domain response ofthe luminescent system. Figure 40.1 shows a genericphotoluminescent system where incident UV radiationexcites a system from a ground state 1 with probabilityper unit time P12 into an excited state 2. The system de-cays with probability P23 to a luminescing level 3, fromwhich there is a probability P31R and P31NR of radiativeand nonradiative decay, respectively, to the ground state.Nonradiative decay generally involves phonon emission.

First suppose that the transition probabilities aresuch that P12, P21 � P31(= P31R + P31NR). If the opticalpumping (excitation) is abruptly stopped, the populationof the luminescing state 3 decays as,

N3 = N3(0) exp (−P31t) (40.1)

and the rate of luminescent energy emission is−(hνP31R/P31)dN3/dt for an energy differenceE31 = hν between states 1 and 3. Hence the lumines-cence intensity at distance r from the sample is,

I = (hν) N3(0)P31R exp (−P31t) /(4πr2) (40.2)

and the characteristic luminescence lifetime isτ = (P31)−1. Thus the lifetime is governed by both ra-diative and nonradiative processes, whilst the intensityof the luminescence depends on the relative magnitudeof P31R.

This discussion provides the basis for understand-ing the terms fluorescence and phosphorescence appliedto luminescent materials. A material is often classifiedas one or the other according to the relative magnitudeof τ = (P31)−1, with 10 ns being set in a relatively ar-bitrary way as the boundary between a fast fluorescentsystem and a slow phosphorescent one. For compar-ison, the theoretical lifetime for spontaneous emissionfor a strongly allowed hydrogen atom 2p → 1s transitionis about 0.2 ns.

However, by this definition phosphorescence canalso arise from luminescent states with short lifetimeswhich are populated through ones with long lifetimes.In Fig. 40.1, if state 2 is long-lived in the sense thatP23 � P31 then the measured lifetime for emission fromthe luminescing state will be τ = (P23)−1, and the sys-tem will be labelled phosphorescent, even though theluminescing level itself has a very short lifetime. Inconsequence, a second classification [40.2] is basedon whether or not the luminescing level is fed bya metastable state which sets the lifetime. Sometimesthe metastable state is a long-lived intermediate form

PartD

40

Luminescent Materials 40.1 Luminescent Centres 985

of energy storage which can be triggered by an ex-ternal stimulus to undergo a transition to a fluorescentlevel. Thus in Table 40.1, thermally stimulated and op-tically stimulated luminescence involve a metastablestate consisting of trapped electrons and holes, whichcan be triggered to recombine by heating or by optical

stimulation; the recombination energy is transferred toa fluorescing centre.

Overall, the fluorescence/phosphorescence clas-sification is somewhat nebulous, and it is de-bateable whether the classification is necessary ordesirable.

40.1 Luminescent Centres

A wide variety of centres give rise to luminescencein semiconductors and insulating materials, includingrare-earth ions, transition-metal ions, excitons, donor–acceptor pairs, and ions with a d10 or s2 electronicconfiguration ground state. Some luminescence spectraconsist of broad emission bands arising from the inter-action between the electronic system of the luminescentcentre and the vibrations of the atoms or ions, whichsurround it; the broad bands arise from simultaneoustransitions of both electronic and vibrational systems.For others, such as the rare earths, the spectra comprisesharp lines arising from purely electronic transitions,and the effect of the environment is felt mainly throughtheir effects on the lifetimes of the states. Thus in dis-cussing the physical background to luminescence, it issimplest to start with a discussion of rare-earth lumi-nescence, where the effect of vibrations can be initiallyignored.

40.1.1 Rare-Earth Ions

The trivalent rare-earth ions have n electrons(n = 1−14) in the 4f shell. In a free ion, the eigen-states resulting from the various atomic interactions arelabelled by the total spin S and orbital angular momentaL. Spin–orbit coupling breaks up each L, S multipletof degeneracy (2S +1)(2L +1) into sub-multiplets la-belled by the total angular momentum J = L + S, whereJ can range from L − S to L + S. The 4fn orbitals liewithin the outer 5s2and 5p6 filled shells, which partlyshield them from the effects of a crystalline environment.The effects of the latter are quantitatively described bythe crystal field [40.3], and this term in the Hamilto-nian splits the J multiplets into 2J +1 sublevels, theso-called crystal-field splitting. Some of these crystal-field levels may still be two- or threefold degenerate,depending upon the symmetry of the environment. Odd-electron systems always have at least twofold (Kramers)degeneracy. The resulting energy-level structures arecomplicated, and are summarised in the classic Dieke di-

agram [40.3]. The original has been updated by Carnallet al. [40.4] and is reproduced in many books and pa-pers on rare-earth ion spectroscopy [40.2]. In Fig. 40.2,we show a schematic version (not accurately to scale)of the diagram appropriate for Pr3+ with n = 2, whichserves for our discussion. The crystal-field splitting isusually smaller than the spin–orbit splitting, and is illus-trated schematically by the vertical extent of the bandsin Fig. 40.2. The multiplet labels follow the usual 2S+1LJsystem.

In the figure we also show a generic highest excitedband which does not belong to the 4f2 configuration.In the specific case of the rare earths this can bethe excited-state configuration resulting from promo-

4f 5d or charge transfer

1S0

Energy (104cm–1)

40

20

0

3P23P13P0

1I6

1G4

1D2

3F4 3F33F2 3H63H5

3H4

Fig. 40.2 Schematic energy-level diagram for Pr3+

PartD

40.1

986 Part D Materials for Optoelectronics and Photonics

tion of one electron to the 5d state, giving an overall4fn−15d1configuration; more generally it could be a so-called charge transfer band, which corresponds to thetransfer of one electron from the ligands to the lumi-nescent ion. The relative location and importance ofthese bands varies with the luminescent ion and the crys-talline environment, but they play an important role inthe excitation of luminescence.

Excitation and luminescence transitions within thevarious levels in Fig. 40.2 are governed by the goldenrule of quantum mechanics [40.5] for interactions withthe electromagnetic field; in summary the probability ofa transition between two states i and j is proportional tosquare of the matrix element < i|H| j >, where H is thetime-dependent perturbation Hamiltonian representingthe interaction of the electrons with the electromagneticfield. The proportionality constant contains the light in-tensity in the case of excitation. The perturbation can beexpanded in a power series involving electric and mag-netic multipoles of the electronic system, but of thesethe electric dipole term is dominant, with the magneticdipole term being much smaller by a factor of morethan five orders of magnitude. Since the electric dipoleoperator er has odd parity, the matrix element for tran-sitions rij =< i|er| j > is necessarily zero unless i andj have opposite parity. This is the most important selec-tion rule governing luminescence: transitions betweenstates of the same parity have zero transition probabilityand so are forbidden (Laporte’s rule). In the case of therare earths, all states of a single 4fn configuration havethe same parity, and so all optical transitions within theconfiguration are strictly forbidden. But this rule is re-laxed by several considerations. First, if the crystallineenvironment lacks inversion symmetry, the crystal fieldadmixes a small fraction of the excited configurations(eg 4f15d1 for Pr3+) of opposite parity into the groundconfiguration, which makes such transitions weakly al-lowed. Secondly, the selection rule for magnetic dipoletransitions is that they are allowed between states of thesame parity, although they are typically about five or-ders of magnitude weaker than for electric dipole ones.Finally, odd-parity vibrations and an electron–phononinteraction produce a similar configuration admixtureeffect to static lattice odd-parity mixing although thiseffect is more important for 3d ions than for 4f ones.

With only weak transitions possible within the 4fconfiguration, one might wonder how it would be possi-ble to optically excite any significant luminescence. Theanswer lies in the 5d or charge-transfer bands which ei-ther lie at higher energies, or overlap with the upperlevels of the 4f configuration. These give rise to strong

absorption and efficient pumping. Relaxation can occurvia the parity-allowed transitions to the upper levels ofthe 4f configuration, and from there via single or multi-ple radiative emissions back to the ground state. Becausethese intra-configurational transitions are only weaklyallowed, the lifetimes are generally quite long, of the or-der of µs–ms. Figure 40.2 shows some of the observedtransitions in the case of the Pr3+ ion. There are fur-ther constraints on possible transitions which arise froman analysis of the angular momenta of the initial and fi-nal states. For example, a transition between two statesboth of which have J = 0 is forbidden since there isno angular momentum change as required for a photon;similarly for dipole transitions we require ∆J = 0,±1.

The transition probability per second for sponta-neous emission [40.6] is given by

Pij = 64π4ν3

3hc3

∣∣rij

∣∣2

, (40.3)

where ν is the frequency of the transition, h is Planck’sconstant, c is the velocity of light, and |rij | is the ma-trix element of the electric dipole operator erij betweenthe two states i and j, and e is the electronic charge. Forabsorption, this must be multiplied by N , the mean num-ber of photons with energy hν, which thus incorporatesthe effect of the incident beam intensity. Experimentally,one measures an absorption coefficient k as a functionof energy k(E) [40.6] which is linked to Pij through,

∫

k(E)dE = Ni

(π e2h

nmc

)(n2 +2

3

)2

fij , (40.4)

where n is the refractive index of the crystal environ-ment, m is the electronic mass, Ni is the concentration ofthe luminescent centres, and f is the oscillator strengthfor the transition. For both absorption and emission, thedimensionless oscillator strength fij defined as [40.6]

fij = 8π2mν

3h e2

∣∣rij

∣∣2

(40.5)

is often quoted to compare the relative transition proba-bilities. For an electron harmonic oscillator,

∑fij = 1,

and so oscillator strengths of the order of 0.1–1 arestrongly allowed transitions.

40.1.2 Transition-Metal Ions

Transition-metal ions from the 3d series are char-acterised by a much stronger interaction with thecrystalline environment than the 4f ions since there is noequivalent of screening by the 5s, 5p outer shells. In addi-tion, the spin–orbit coupling is weaker, and so the order

PartD

40.1

Luminescent Materials 40.2 Interaction with the Lattice 987

of perturbation is reversed: the atomic L , S multipletsare split by the crystal field, with spin–orbit couplingbeing a smaller interaction. Intra-configurational tran-sitions are again strictly forbidden, but become weaklyallowed by inter-configurational mixing through odd-parity crystal fields, and by odd-parity vibrations. Theresult of this is that the strongest selection rule af-ter parity is that transitions should have ∆S = 0, sincethe electric dipole operator does not involve spin. Theother major difference, again due to the strength of thecrystal-field interaction, is that transitions which arepurely electronic, the so-called zero-phonon lines, arerarely observed. Rather what are seen are broad bandswhich correspond to the simultaneous excitation of anelectronic transition and vibrational transitions, whichoverlap to give the broad observed bands. In particular,transitions involving odd-parity vibrations have a hightransition probability through the effect of configura-tion admixing. This will be considered in a followingsection. The most commonly observed luminescent ionsare those from the d3 configuration (Cr3+, Mn4+) andfrom the d5 configuration (Mn2+).

40.1.3 s2 Ions

The 5s2 (e.g. Sn2+ and Sb3+) ions and 6s2 (e.g. Tl+,Pb2+, Bi3+) ions are of considerable importance be-cause transitions to and from the excited s1p1 states areLaporte-allowed. The interaction of the p state with the

crystalline environment can be very strong, and so broadspectra are often observed.

40.1.4 Semiconductors

Luminescence in semiconductors is dominated by near-band-gap luminescence arising from recombination ofelectrons and holes. This process is most efficient indirect-band-gap materials such as ZnS and GaP ratherthan indirect-gap materials such as Si and Ge be-cause the transition probability requires conservationof wavevector, but the photon wavevector is ≈ 0 onthe scale of the Brillouin zone. Hence creation or de-struction of a phonon is required for band-to-bandluminescence in indirect-gap materials, which is lessprobable. The near-edge emission may correspond toluminescence from a variety of shallow energy-levelstructures such as free or trapped excitons, or fromdonor–acceptor recombination. These are both exampleof electronic systems with spatially extensive wave-functions, in contrast to the atomically localised 3dand 4f wavefunctions considered earlier. However, it isalso possible to observe deep-level luminescence fromtransition-metal ions and rare earths in semiconductorsprovided that the electron affinities and band-gap en-ergies are such that the pertinent energy levels fall inthe band gap. Since semiconductors are discussed else-where in this volume, we shall not consider them furtherhere.

40.2 Interaction with the Lattice

For rare-earth ions, the interaction with the vibrationsof the crystal lattice can be ignored for most purposes;the observed luminescence spectrum consists of sets ofsharp electronic transitions. But for other luminescentions which interact strongly with the vibrating ions ofthe surrounding crystal, the incorporation of the latter iscritical to explaining the observed spectra. The simplestmodel of ion–lattice interactions is to consider only theN nearest neighbour ions and their atomic displacementsXn , Yn , Zn , (n = 1, N) in Cartesian coordinates. Thevibrational Hamiltonian involves cross terms in thesecoordinates, but may be transformed to harmonic form ifsymmetry-adapted forms of these coordinates (normalmodes) are used instead of the actual displacements.For example, the so-called breathing mode Qb, for anoctahedrally coordinated ion takes the form

Qb = (Z1 − Z2 + X3 − X4 +Y5 −Y6)/6 . (40.6)

If all the other modes have zero amplitude, the ionsmove radially towards or away from the central lumi-nescent ion. The key point in considering the influenceof the crystal lattice is that the vibrational potentialenergy is just the variation of the electronic energywith ionic displacement, or equivalently with the nor-mal modes (within the spirit of the Born–Oppenheimerapproximation). Put another way, the crystal field de-pends on the ion positions so that the electron and latticequantum-mechanical systems are linked through thiselectron–lattice coupling. We can therefore expect a dif-ference in the harmonic vibrational potential from oneelectronic state to another, so that it will in general havethe form (1/2)mω2

i (Q − Q0i)2; i. e. both the magnitudeof the potential and the position of the minima Qi0 willdepend on the electronic state i. The vibrational statesin the harmonic approximation are just the usual simpleharmonic oscillator states with energies (n +1/2)hνι,

PartD

40.2

988 Part D Materials for Optoelectronics and Photonics

Energy

Coordinate Q

ExcitationEmission

Fig. 40.3 Configuration coordinate diagram for excita-tion/emission cycle

with n an integer. Thus we arrive at a configuration co-ordinate diagram such as that shown in Fig. 40.3, withthe potential energies for a ground state g and excitedstate e being offset parabolas of different curvature. Thesquare of the vibrational wavefunction is shown for theground state and that with n = 11. We note that for a clas-sical oscillator there would be a peak in the probabilityfunction at the extreme lengths of travel, correspond-ing to the maxima for n = 11 at the outer limits of thewavefunction in Fig. 40.3.

We consider first luminescence from the excitedelectronic state/lowest vibrational state. The transitionprobability may be calculated using the Franck–Condonprinciple, that is we assume that the duration of theelectronic transition is much shorter than a vibrationalperiod, so that the Q remains constant during a tran-sition. The transition is therefore taken to be verticalon the configuration coordinate diagram. The transitionprobability is just 〈φe|φg〉2, the square of the vibra-tional wavefunction overlap, multiplied by the electronictransition probability considered earlier. From Fig. 40.3,this overlap will be a maximum for some vibrationalstate other than the ground state, unless the positionsof the minima of the two potential energy curves coin-cide accidentally. (In Fig. 40.3, this maximum wouldbe for the n = 11 ground vibrational state). Put an-other way, the maximum transition probability is not

Energy

AbsorptionEmission

Stokes shift

Fig. 40.4 Absorption and emission line shapes for strongelectron–lattice coupling

for the zero-phonon transition (no change in vibra-tional state), but corresponds to the creation of a finitenumber of phonons. A range of transitions is allowed,and the result in a semiclassical analysis, allowingfor finite line widths, is a Gaussian-shaped band. Theanalysis has to be extended to include finite temper-atures and other modes, but the overall result is thatthe emission line shape is approximately a Gaussiancentred at an energy lower than that of the differ-ence between the minima of the two potential curves.The same argument can be applied to the excitationprocess, and again an approximately Gaussian lineshape results but this time centred on an energy abovethat of the difference in potential-energy minima. Fig-ure 40.4 shows the overall result; the difference betweenthe maxima of the two curves, known as the Stokesshift, is an indicator of the degree of electron–latticecoupling.

It is clear from the diagram that the, in caseswhere there is strong electron–lattice coupling of thistype, that: (a) there will be very little intensity inthe zero-phonon line, and (b) there will be a largeStokes shift between the energies for maximum absorp-tion and maximum emission. Thus the luminescenceof transition-metal ions, colour centres, and closed-shell ground-state ions (s2, d0), which have stronginteractions with the lattice in the excited states, are typ-ically broad bands with only occasionally weak, sharpzero-phonon lines being observed. For luminescencefrom within the 4f states of the rare earths, the re-verse is true; we are in the weak-coupling regime, andzero-phonon lines are the predominant features of thespectrum.

PartD

40.2

Luminescent Materials 40.3 Thermally Stimulated Luminescence 989

40.3 Thermally Stimulated Luminescence

Thermally stimulated luminescence (TSL), or simplythermoluminescence (TL), refers to luminescence in-duced by thermally stimulated recombination of trappedelectrons and holes in materials which have been sub-ject to prior irradiation. The irradiation, which may bein the form of UV light, X-rays, gamma rays, or ener-getic electrons, creates free electrons and holes, most ofwhich promptly recombine, but some of which are lo-cally trapped at defect centres such as impurities andvacancies. If the trap binding energies are sufficientlylarge, thermal promotion of the electron or hole to theconduction band or valence band, respectively, is im-probable at the irradiation temperature, and so thesecharge carriers remain trapped after irradiation. How-ever, if the sample is then heated, thermally assistedrecombination becomes increasingly probable, and theresult is an initially increasing light output with increas-ing temperature until the traps are depleted, whereuponthe light intensity drops. The curve of light intensityversus temperature I(T ) is known as a glow curve, andmay be analysed to extract the trap depths and concen-trations. A comprehensive review of the field has beengiven by McKeever [40.7], and a shorter discussion isgiven by Vij [40.1].

The process is shown schematically for a simplesystem comprising a single electron trapping level Tand a single recombination centre R in Fig. 40.5.

Irradiation results in a trapped electron at trap T anda trapped hole at R. The trapped hole binding energyis larger than that of the trapped electron, so the latteris depopulated first, with a probability P which has theform,

P = s exp (−E/kT ) , (40.7)

where s and E are the attempt frequency and the acti-vation energy respectively. We have shown the hole trapand recombination centre as being one and the same inFig. 40.5, but it is also possible that the recombinationenergy is transferred to a separate luminescent centre.

In the measurement process, the sample is heatedat a fast and linear rate, typically 1–10 K/s, whilstthe light emission is monitored by a sensitivefilter/photomultiplier combination, or by a monochro-mator system. The resulting glow curve may be fitted toa theoretical curve to extract the trap parameters. Gener-ally the light output is quite weak, and TSL systems havesensitivity as one of their prime design factors, so cooleddetectors and photon counting are commonplace. Aboveabout 400 ◦C, thermal radiation from the sample/heater

is a problem and must be eliminated by filtering or bysubtraction of a glow curve recorded using a thermallybleached sample.

The main uses of TSL are in determining trap depthsand irradiation doses. In archaeological and geologicalapplications, a comparison is made between accumu-lated natural dose and a dose from a radioisotope; bycombining this with a measurement of the activity of thesurroundings, a date since last thermal or optical erasure

Conduction band

Valence band

T

R

Fig. 40.5 Thermally stimulated luminescence process

1.2

1.0

0.8

0.6

0.4

0.2

0500450400350300

Intensity (arb. units)

Temperature (K)

0.9 eV1.0 eV1.1 eV

Fig. 40.6 Computed glow curves for first-order kinetics forfixed escape frequency and various trap depths

PartD

40.3

990 Part D Materials for Optoelectronics and Photonics

of the object can be determined. The nature of the trapsis often poorly understood for these chemically com-plex samples. For medical dosimetry, room-temperaturestable traps with high thermoluminescent output are re-quired. At present, the material of choice is LiF dopedwith a few hundred ppm of Mg and Ti.

The mathematical form of the TSL glow curve de-pends on the physical model used for the TL process. Inthe simplest case, assuming first-order kinetics [40.1,7],the light intensity I at temperature T is given by,

I(T ) = n0s exp (−E/kBT )

× exp

⎡

⎢⎣− (s/β)

T∫

T0

exp (−E/kBT ) dT

⎤

⎥⎦ ,

(40.8)

where n0 is the number of occupied traps at time t =t0 when the temperature is T0, β is the heating ratein K/s, and kB is Boltzmann’s constant. The result isa glow curve whose peak position varies approximatelylinearly with trap depth E, as shown in Fig. 40.6 (forfixed escaped frequency s = 3.3 × 1011 s−1 and heatingrate 1 K/s).

Of course, the peak position also depends on theescape frequency s, but is less sensitive to s than toE. For second-order kinetics [40.1, 7], the glow curveequation becomes

I(T ) =(

n20s/N

)

exp (−E/kBT )

/

⎡

⎢⎣1+ (n0s/Nβ)

T∫

T0

exp (−E/kBT ) dT

⎤

⎥⎦

2

,

(40.9)

1.2

1.0

0.8

0.6

0.4

0.2

0

–0.2500300 450400350

Intensity (arb. units)

Temperature (K)

E = 1 eVs = 3.3 x 10–11 s–1

ß = 1 K/s

Second order

First order

Fig. 40.7 Comparison of first- and second-order glowcurves

where N is the number of available traps. The twoforms both give glow curves of similar shape, butthe first-order form shows an asymmetric form witha sharp fall off on the high-temperature side of thepeak, whilst the second-order form is more symmet-ric as shown in Fig. 40.7. The parameters are extractedby least-squares fitting of these expressions to the ex-perimental glow curves, which is cumbersome due tothe integral, but Kitis et al. [40.8] have given analyticalapproximations to (40.8) and (40.9). In practice, manyglow curves do not follow first- or second-order kineticsprecisely.

40.4 Optically (Photo-)Stimulated LuminescenceThermally stimulated luminescence is sometimes alsoaccompanied by optically stimulated luminescence(OSL), in which one of the trapped carriers is ex-cited by optical stimulation to a level from whichit can recombine with the conjugate carrier by tun-nelling, or completely to one of the bands so thatrecombination is achieved through what is essentiallya photoconductivity effect. For OSL to be significant,

there must be an appreciable optical transition prob-ability, so not all TSL centres are OSL active. Thestimulation energy measured in OSL generally dif-fers from that determined from TSL because of theFranck–Condon principle. The OSL effect finds practi-cal application in dosimetry, e.g. [40.9], and in X-raystorage phosphors used for medical imaging, as de-scribed later.

PartD

40.4

Luminescent Materials 40.5 Experimental Techniques – Photoluminescence 991

40.5 Experimental Techniques – Photoluminescence

A typical traditional photoluminescence measuring sys-tem involves a broad-spectrum source, either a combinedtungsten filament for the visible spectrum and deuteriumlamp for the UV, or a xenon flash lamp. The lampemission is passed through a grating monochromatorand so selectively excites the luminescence. Band-passor band-edge filters are generally required to eliminateunwanted second- and higher-order diffraction maximafrom the grating. The luminescence is efficiently gath-ered by low-f/number optics and fed to a second gratingmonochromator, also equipped with filters, to monitorand analyse the luminescence. The final detector maybe a photomultiplier, or preferably a charge-coupled de-vice or photodiode array for improved data collectionefficiency at multiple wavelengths.

This arrangement relies on good monochroma-tion/filtering to remove what is sometimes a relativelystrong component of scattered light from the beam anal-ysed by the emission monochromator. An alternativemethod of removing scattered light is to use time dis-crimination, by replacing the source by a xenon flashlamp (for example, as in the common Perkin ElmerLS55B luminescence spectrometer). The flash has a du-ration of about 10 µs, so any scattered light has decayedaway to an insignificant level when the emitted beam issampled some time (0.1–10 ms) after the flash. A timedelectronic gate is used to sample the emission imme-diately after a flash and just before the next flash; thedifference between these two sampled signals gives theshort-term luminescence whilst the second sample alonegives the long-term luminescence with long and shortbeing relative to the flash repetition period. Of course,luminescence with lifetimes shorter than the pulse width(≈ 10 µs) cannot be readily measured with this system.The luminescence intensity is normalised with respectto the excitation intensity by steering a sample of theexcitation beam to a rhodamine dye cell which hasa quantum efficiency of essentially unity for wavelengthsbelow about 630 nm. The fluorescence from the dye ismeasured with a second photomultiplier.

To minimise the effect of scattered light, a con-ventional laser with its intrinsically narrow linewidthand high intensity is a very convenient replacementfor a broad-spectrum lamp, but suffers from the dis-advantage of a fixed wavelength. Typical lasers ofinterest are nitrogen (pulsed), argon (UV lines, orfrequency-doubled visible lines), krypton, and the new

generations of GaN/GaInN blue/violet/UV laser diodes.For rare-earth spectroscopy, or other systems which arecharacterised by narrow absorption lines, it is very use-ful to have a scanning dye laser as the excitation source.This selective excitation facility enables tagging of par-ticular luminescent levels with excited states belongingto the same centre, so that a picture of the energy-levelstructure of each luminescent centre can be built up incases where several such centres contribute to the overallluminescence.

For decay kinetics on faster time scales, fluorimeterssuch as those developed initially by the Spex company(now Horiba) use a fast modulator and phase-sensitivedetection to measure the phase shift between fluores-cence and excitation; it is claimed that fluorescencedecays can be measured with a resolution of 25 ps thisway. An alternative is the time-correlated single-photon-counting technique which can measure decay constantsin the ps–ns range. In this method, the excitation comesfrom a fast laser pulse, and the light level reaching thephotomultiplier or micro-channel plate detector is re-duced to such a low level that less than one photon perexcitation pulse is detected. The time delay between thephoton detection and the time of the pulse is measured,and a histogram produced of numbers of detected pho-tons versus arrival time taken over a large number ofexcitation pulses. For efficient data collection, a highrepetition rate and fast-pulse laser are required, oftena Ti–sapphire laser. The wavelength for Ti–sapphire istoo large for stimulating many materials directly withsingle-photon excitation, but stimulation is nonethelesspossible by a two-photon excitation process, or by theuse of a nonlinear crystal acting as a frequency doublerto produce laser output at one half the wavelength of thebasic laser.

There are a number of specialist techniques, suchas hole-burning, fluorescence line-narrowing, and pho-ton echo methods associated with the use of laserswith either very narrow line widths or short pulseduration which have developed in a parallel way to tech-niques first introduced in nuclear magnetic resonance,and which are mainly used to investigate the dynam-ics and quantum mechanics of the luminescent speciesrather than the material in which they are contained.Meijerink gives a review of experimental luminescencetechniques [40.1] which includes a short discussion ofthese specialist techniques.

PartD

40.5

992 Part D Materials for Optoelectronics and Photonics

40.6 Applications

The largest market for luminescent materials has tradi-tionally been in the areas of lighting, through fluorescenttubes, and in cathode-ray-tube screen phosphors for im-age display. Both of these areas can now be regardedas mature in terms of materials development. However,new discoveries in the past decade, and the advent of newtechnologies, have rekindled interest in phosphor mater-ials. Some of the current areas of activity in applicationsare outlined below.

40.6.1 White Light-Emitting Diodes (LEDs)

The development of blue, violet and UV LEDs based onGaN, InGaN and other semiconductors and alloys hasstimulated great interest in the possibility of producinga white-light LED for use in lighting applications. LEDsare now available with emissions which peak as low as365 nm in the ultraviolet. The concept is to use the blueemission to stimulate luminescence from yellow, or redand green phosphors, so that when mixed with residualblue light from the LED the result is simulated whitelight. For UV LEDs the LED output is used to stimulateblue, red and green phosphors.

The first generation of white LEDs from companiessuch as Nichia relied on a YAG:Ce phosphor to convert

1000

100

104.00 0.5 1.0 1.5 2.0 2.5 3.0 3.5

Intensity (arb. units)

Time (h)

Fig. 40.8 Dark decay of persistent luminescence in a com-mercial lighting strip based on Nemoto SrAl2O4:Eu/Dymaterial

some of the emission from a 465-nm GaInN LED intoan orange/yellow emission centred at 550 nm; the com-bination of blue and yellow simulates white light. Morerecent versions of this scheme include Sr3SiO5:Eu2+,which emits at 570 nm, and is claimed [40.10] tobe more efficient than YAG:Ce, and a CaSiAlON:Euceramic phosphor which offers improved thermal sta-bility [40.11]. Improved colour rendition is obtained byusing more phosphors and better balance between thevarious emissions, for example by replacing the YAG:Cewith SrGa2S4:Eu (green) and SrY2S4:Eu (red). Most re-cent developments based on 375 nm UV LEDs have usedmultiple emissions to achieve even better colour balance;for example Kim et al. [40.12] use Sr3MgSi2O8:Eu2+,Mn2+, which has blue (Eu2+), yellow(Eu2+) and redemissions (Mn2+).

A second development in this area is the sub-stitution of another semiconductor for the phosphor– the so-called photon recycling technique. In thismethod, a layer of AlGaInP is used to absorb someof the blue incident radiation and down-convert itto the complementary colour. The advantage is thatthe fabrication/integration process is simpler com-pared to combining phosphor and semiconductortechnology.

1.2

1.0

0.8

0.6

0.4

0.2

0700200 600500400300

Intensity (arb. units)

Wavelength (nm)

Excitation Emission

Fig. 40.9 Excitation and emission spectra of a commerciallighting strip based on Nemoto SrAl2O4:Eu/Dy material

PartD

40.6

Luminescent Materials 40.6 Applications 993

40.6.2 Long-Persistence Phosphors

It has been discovered in the past decade that somematerials show a very long-lived but intense afterglow,arising from thermal emission of charge carriers fromdeep traps, followed by electron–hole recombinationat or near a luminescent ion, i. e. room-temperaturethermoluminescence. The persistence is for periods ofseveral hours, much longer than the well-known ZnS:Cupersistent phosphor. This new class of materials havewidespread applications in areas such as signage andpassive emergency lighting for public buildings, aircraftcabins etc. The material is activated or charged by theblue/UV content of solar radiation or fluorescent indoorlighting during normal conditions; in subsequent darkconditions the energy is released as an afterglow. Nopower supply is required, the operation is entirely pas-sive, and so a high degree of reliability is assured. Gen-erally the phosphor powder (≈ 10 µ grain size) is mixedwith a resin binder and applied as a thick (≈ 1 mm) coat-ing. The materials which have been used are rare-earth-doped strontium or calcium aluminate [CaAl2O4:Eu/Nd(blue), SrAl2O4:Eu/Dy (green), Sr4Al14O25:Eu/Dy(blue–green), and Y2O2S:Eu/Mg/Ti (orange)]. Fig-ures 40.8 and 40.9 show the decay and emis-sion/excitation characteristics of a commercial phosphorstrip of this type, based on Luminova® SrAl2O4:Eu/Dypowder sourced from the Nemoto company. The decayis clearly non–exponential, but after the initial rapid de-cay the intensity may be fitted with an exponential witha decay constant of 3.1 h.

The traps which are responsible for the long-liveddecays in these systems have not been clearly iden-tified, and it is likely that there are several involved.Jia et al. [40.13, 14] have studied persistent violet/UVand blue (≈ 400 nm, 450 nm) luminescence from rare-earth ions in CaAl2O4 and BaAl2O4 and report that thethermally stimulated luminescence from these materialscontains multiple glow peaks. They also describe howco-doping the calcium aluminate with Tb3+ results ina green persistent phosphor as a result of energy trans-fer from the cerium to the terbium ion; this mechanismcould be the basis for convenient colour control, includ-ing white persistent phosphors. Aitasalo et al. [40.15]report that the nature of the traps is affected by theparticular rare-earth ion or couple.

40.6.3 X-Ray Storage Phosphors

Optically stimulated or photo-stimulated luminescenceis the basis for an X-ray imaging technology known

commercially as computed radiography (CR). CR wasthe first of a number of imaging techniques which aresteadily replacing traditional photographic film meth-ods. Several other techniques such as those basedon amorphous selenium photoconductor/flat panelsand a-Si arrays, and scintillator CCD/complementarymetal–oxide–semiconductor (CMOS) arrays have alsoemerged in recent years, but CR is still the dominantof the new technologies on the basis of numbers ofinstalled units. The many advantages of CR over pho-tographic film, and the principles of the method aredetailed in [40.16,17], but the basic mechanism is that in-cident X-rays create electron–hole pairs in the material.Most of these promptly recombine, but some are trappedat defects and impurities, and remain trapped for periodsof hours after the X-ray source is turned off. In an X-raystorage phosphor (XRSP), one of the trapped carriers isoptically stimulable, and can be excited to the conduc-tion band or valence band, or to a level from which it canrecombine by tunnelling with the conjugate trapped car-rier. The resulting recombination energy is transferredto a luminescent ion, and the intensity of the photo-stimulated luminescence (PSL) is in direct proportion tothe incident X-ray intensity. The dominant material usedin current XRSP systems is BaFBr1−xIx:Eu2+, where theelectron traps are F-centres, the hole traps are unidenti-fied, and the luminescence is the 5d–4f transition of theEu2+ ion. One disadvantage of systems based on this(powder) phosphor is that when the image is extractedwith a raster-scanned He–Ne laser beam, light scatter-ing from the powder grains means that material outsidethe focussed laser spot is also stimulated, limiting thespatial resolution to around 200 µm, which is inade-quate for applications such as mammography. Severalways to overcome this are currently under development.RbBr:Eu is also an X-ray storage phosphor [40.18] andcan be grown by vapour deposition in a columnar form.

Fig. 40.10 PSL image of a BC549 transistor recorded ona glass-ceramic imaging plate

PartD

40.6

994 Part D Materials for Optoelectronics and Photonics

The columnar structure has a light-guiding property,restricting the scattering effect, and improving the res-olution. A second development is that of glass-ceramicstorage phosphors, where PSL active crystals are embed-ded in a glass [40.19–21]; the combination of particlesize, separation and refractive-index mismatch meansthat these composite materials are semitransparent andthe problem of scattering of read-out light is reduced.Figure 40.10 shows an image of a BC549 transistorrecorded in a glass-ceramic X-ray storage phosphor.

40.6.4 Phosphors for Optical Displays

There are several new technologies being developed toreplace cathode ray tubes for domestic televisions, in-cluding plasma display panels (PDPs). In these units,each pixel is a sealed cell containing a mixture of Xeand Ne in a dielectric-shielded electrode structure (fora review, see Boeuf [40.22]). An alternating current(AC) voltage applied between the electrodes results ina glow discharge being set up in the gas, and a Xe dimervacuum UV (VUV) emission predominantly between147 and 190 nm occurs. (In comparison, the mercurydischarge in a conventional fluorescent tube emits pri-marily at 254 nm.) The UV discharge excites red, blue,or green phosphors coated on one of the cells; eachcolour is activated by an adjacent electrode. The require-ments for efficient output from these phosphors differfrom conventional tubes since the latter were chosen onthe basis of their luminescence efficiency at a 254-nmpump wavelength, and their resistance to degradationby the UV light and chemical attack by Hg+ ions.These requirements are evidently different for the PDPtechnology; in addition the phosphors must have a sig-nificant reflection coefficient in the visible to optimisethe light output [40.23], and the surface quality is ofgreater significance due to the short penetration depthof the VUV. The phosphors which have been used sofar include BaMgAl10O17:Eu2+ (blue), Zn2SiO4:Mn2+(green), and (Y, Gd)BO3:Eu3+ and Y2O3:Eu3+ (red).The blue phosphor is prone to degradation.

The widespread introduction of Xe excimer excita-tion in PDPs can be expected to stimulate applications inother lighting technologies. In this regard, the possibil-ity of so-called quantum cutting is of much interest. Thisrecognises that the energy of a VUV photon is equivalentto two or more visible photons, so that quantum efficien-cies in excess of 100% can in principle be achieved. Thedifficulty lies in finding a luminescent ion system whoseenergy-level system provides for both efficient pump-ing and two-photon luminescence in the visible. One

example which has been reported to have a quantumefficiency of up to ≈ 145% is Pr3+ in YF3 and otherhosts [40.24, 25], where the excitation is through theallowed 4f2 → 4f15d1 or host transitions. The systemthen decays to the 1S0 excited state of the 4f2 configura-tion from which two-photon decay is possible throughsuccessive 1S0 →1I6 and 3P0 →3HJ, 3FJ transitions, asshown in Fig. 40.2. (The intermediate step from 1I6 to3P0 is provided by a nonradiative transition.) A diffi-culty is that the photon for the first transition is in theUV region of the spectrum, and so it is necessary toincorporate a second luminescent ion pumped by thistransition to convert the UV to visible output, and thevisible quantum efficiency is necessarily reduced.

40.6.5 Scintillators

Although semiconductor detectors of ionising radiationare making increasing inroads into the particle detectionmarket, traditional scintillators are still widely used andare indispensable for some applications. The operatingprinciple is that an incident gamma ray creates a largenumber of electron–hole pairs in the scintillating mater-ial directly or indirectly through the photoelectric effect,Compton scattering, or pair production, and that some ofthe energy of recombination appears as photon emissionfrom luminescent ions. Charged particles such as pro-tons produce electron–hole pairs through the Coulombinteraction with the band electrons. The scintillation ormultiphoton bursts which signal the event is detectedby a photomultiplier, and the height of the output pulsefrom the photomultiplier is proportional to the energyof the particle. A pulse-height analyser sorts the pulsesaccording to energy, and so an energy spectrum can beobtained. Key figures of merit for a scintillator materialare the numbers of photons per MeV of particle energy,the radiative lifetime of the luminescent ion (since pos-sible pulse overlap limits the maximum count rate whichcan be measured), and the weighted density ρZ4

eff, whichreflects the gamma sensitivity. (Here Zeff is the effec-tive atomic number.) A recent review of scintillators hasbeen given by van Eijk [40.26]. The most widely usedscintillator for many years has been NaI:Tl, but many dif-ferent scintillators are being investigated, driven by theneed for improved performance and lower cost for med-ical applications such as positron emission tomography(PET) and single-photon emission computed tomogra-phy (SPECT), and for large-scale elementary-particlefacilities such as those at the Centre Européen pour laRecherche Nucléaire (CERN). In the latter regard, theCrystal Clear collaborative project and other programs

PartD

40.6

Luminescent Materials References 995

have resulted in several new materials such as LaBr3,LaCl3, Lu2SiO5, Gd2SiO5, and LuAlO3, all Ce-doped,and undoped Bi4Ge3O12 and PbWO4, with typical per-formance figures of 10 000–50 000 photons/MeV andlifetimes of 10–50 ns. For gamma spectroscopy thepulse-height resolution is critical and LaBr3:Ce hastwice the resolution of NaI:Tl. The fastest scintillators

are based on core-valence luminescence where a holecreated in a core level recombines with an electron inthe valence band. For example, BaF2 shows this cross-luminescence effect with a lifetime as short as 600 ps.The effect is only shown by materials with a core-valence(CV) band energy gap less than the usual band gap,otherwise CV luminescence is absorbed.

40.7 Representative Phosphors

To conclude, we present in Table 40.2 a list of several lu-minescent materials of practical significance. The tableis intended to be representative rather than comprehen-

Table 40.2 Some luminescent materials of practical significance

Host Dopants Colour Excitation Application

Bi4Ge3O12 – Blue Ionising radiation Scintillator

ZnS Ag+ Blue Electrons Colour TV screens

Zn0.68Cd0.32S Ag+ Green Electrons Colour TV screens

Y3Al5O12 Ce3+ Yellow Blue, violet White LED

Gd2SiO5 Ce3+ UV Ionising radiation Scintillator

ZnS Cu+ Green Electrons Colour TV screens

BaFBr Eu2+ UV/blue X-rays X-ray imaging

BaMgAl10O17 Eu2+ Blue UV fluorescent lamps,

plasma displays

Sr3SiO5 Eu2+ Blue UV White LED

SrGa2S4 Eu2+ Green UV White LED

SrAl2O4 Eu2+, Dy3+ Green UV, violet Persistent phosphor

CaAl2O4 Eu2+, Nd3+ Blue UV, violet Persistent phosphor

Y2O3 Eu3+ Red Electrons, UV Colour TV screens,

fluorescent lamps

Sr2SiO4 Eu3+ Yellow UV White LED

(Y, Gd)BO3 Eu3+ Red UV Plasma displays

Y2O3 Eu3+ Red UV Plasma displays

SrY2S4 Eu3+ Red UV White LED

LiF Mg2+ and Ti4+ UV//blue Ionizing radiation TL dosimetry

ZnS Mn2+ Yellow Electric field Panel displays

Zn2SiO4 Mn2+ Green UV Plasma displays

CeMgAl11O19 Tb3+ Green UV Fluorescent lamps

sive. It is noticeable from the table that just a few ionsare responsible for a large number of applications, andprimarily as oxides.

References

40.1 D. J. Vij: Luminescence of Solids (Plenum, New York1998)

40.2 G. Blasse, B. C. Grabmeier: Luminescent Materials(Springer, Berlin, Heidelberg 1994)

40.3 G. H. Dieke: Spectra and Energy Levels of Rare EarthIons in Crystals (Interscience, New York 1968)

40.4 W. T. Carnall, G. L. Goodman, K. Rajnak, R. S. Rana:J. Chem. Phys. 90, 343 (1989)

PartD

40

996 Part D Materials for Optoelectronics and Photonics

40.5 E. Merzbacher: Quantum Mechanics (Wiley, NewYork 1970)

40.6 D. Curie: Luminescence in Crystals (Methuen, Lon-don 1962)

40.7 S. W. S. McKeever: Thermoluminescence of Solids(Cambridge Univ. Press, Cambridge 1985)

40.8 G. Kitis, J. M. Gomex-Ros, J. W. N. Tuyn: J. Phys. D.31, 2636–2641 (1998)

40.9 L. Botter-Jensen, S. W. S. McKeever, A. G. Wintle:Optically Stimulated Luminescence Dosimetry (El-sevier, Amsterdam 2003)

40.10 J. K. Park, C. H. Kim, H. D. Park, S. Y. Choi: Appl.Phys. Lett. 84, 1647–1649 (2004)

40.11 S. Ken, O. Koji, K. Naoki, O. Masakazu, T. Daiichiro,H. Naoto, Y. Yominobu, X. Rong-Jun, S. Takayuki:Opt. Lett. 29, 2001–2003 (2004)

40.12 P. L. Kim, P. E. Jeon, Y. H. Park, J. C. Choi, L. P. Park:Appl. Phys. Lett. 85, 3696–3698 (2004)

40.13 D. Jia, R. S. Meltzer, W. M. Yen: Appl. Phys. Lett. 80,1535–1537 (2002)

40.14 D. Jia, X. J. Wang, E. van der Kolk, W. M. Yen: Opt.Commun. 204, 247–251 (2002)

40.15 T. Aitasalo, A. Durygin, J. Holsa, J. Niittykoski, A. Su-chocki: J. Alloys Comp. 380, 4–8 (2004)

40.16 S. Schweizer: Phys. Status Solidi 187, 335–393(2001)

40.17 J. A. Rowlands: Phys. Med. Biol. 47, R123–R166(2002)

40.18 P. Hackenschmied, G. Schierning, A. Batentschuk,A. Winnacker: J. Appl. Phys. 93, 5109–5113(2003)

40.19 S. Schweizer, L. Hobbs, M. Secu, J.-M. Spaeth,A. Edgar, G. V. M. Williams: Appl. Phys. Lett. 83,449–451 (2003)

40.20 M. Secu, S. Schweizer, A. Edgar, G. V. M. Williams,U. Rieser: J. Phys. C: Condens. Matter 15, 1–12(2003)

40.21 A. Edgar, G. V. M. Williams, S. Schweizer, M. Secu,J.-M. Spaeth: Curr. Appl. Phys. 4, 193–196 (2004)

40.22 J. P. Boeuf: J. Phys. D 36, R53–R79 (2003)40.23 H. Bechtel, T. Juestel, H. Glaeser, D. U. Wiechert: J.

Soc. Inform. Display 10, 63–67 (2002)40.24 S. Kuck, I. Sokolska, M. Henke, M. Doring, T. Schef-

fler: J. Lumin. 102-103, 176–181 (2003)40.25 A. B. Vink, P. Dorenbos, C. W. E. Van Eijk: J. Solid

State Chem. 171, 308–312 (2003)40.26 C. W. E. Van Eijk: Nuclear Instruments and Methods

in Physics Research A 460, 1–14 (2001)

PartD

40