CHAPTER l

GENERALINTRQDUCTIQE

PART;g

l.l The Basic Phenomenon

Emission of light from materials has attracted manfrom time immemorial. The shining light flashed by fire­flies at night, apparently without any harmful heat accompany­ing it, had remained an enigma for the early man. A substancecan be made to emit visible light in a multitude of ways.These include the application of heat, pressure, chemicals,frictional force, high energy radiation etc. Electroluminesc­ence (EL) is the phenomenon of light emission from certaincrystals (phosphors) caused by the passage of an electriccurrent. Such electric field induced emission was firstobserved in SiC crystals by Lossew [1] in 1923. Since thenextensive investigations have been carried out by a host ofworkers on the different aspects of EL in a number of mate­rials [2]. The subject is of great contemporary importanceas it finds a variety of potential applications in many hightechnology areas.

Light emitting devices based on EL have been producedand they find numerous applications as indicators and displayunits in instruments and systems. In particular EL devicescan serve as the information linkage between complex elec~tronic systems and their human users (as a man-machine inter­face). Hence there is a continued attempt to improve the

2

performance and efficiency of EL devices. This has resultedin a large number of new EL materials and new configurationsfor EL devices.

Currently the most prominent EL materials are III-Vand IIQVI compound semiconductors. The former type is usuallyknown as the LED material since they are widely used in thefabrication of the light emitting diodes (LED) in which theemission is due to the recombination of the injected minoritycarriers across a p-n junction. These display elements areroutinely used in visual alpha—numeric information displaysin many hand held and desk electronic calculators, as indi­cators in a wide variety of electronic instruments, as lightmodulators etc. These cold emitters have in their emissionspectrum bands or lines unlike incandescent sources whichusually give a continuous spectrum of visual as well as heatradiation. Light emitting diodes with emission in the nearinfra red (IR) region are used in fibre optic communicationsystems [3,4].

Electronic systems are continually increasing intheir complexity and in range of applications. This hasproduced a need for electronic displays of large area,capable of showing complex digital, analogue and graphicinformation. The LED displays are not quite suitable forthe development of such systems due to their high cost andthe much higher current density required for their excitation.Even the most efficient devices require a current density

of O.25A/Cmz [5]. This will need bulky excitation sources,making battery operation quite difficult.

1- 2 lygeertance ref II-a-Y1 ..¢.QI11P_°11}1.@$ as a11EI» 1>hQs2l1.9r

In the 1950's and early 1960's several thousandpapers in the field of EL were published [6]. These dealprincipally with panels based on powder ZnS and related com­

pounds and with very few exceptions, all devices describedwere operated under AC excitation. The work during thatperiod has been reviewed and assessed in several books [6,7]and articles [8,9,lO]. These studies suggested that EL _display panels of low power requirements can be developedwith the II-VI compound semiconductors. But it is verydifficult to form p-n junction with these materials partic­ularly due to their self compensation effects [ll]. Thefirst and the most efficient EL material of this group isZnS. Electroluminescence in ZnS-tuna discovered by Destriau

[12] in 1936. Compounds belonging to this class are com­paratively less expensive and have good convertion efficiencyFor example Lehmann [13] in 1958 has prepared one optimisedZnS EL device activated with copper and chlorine (ZnS:Cu,Cl)which has an efficiency of 4.5 x lO'2 compared to the valueof 1.7 x lO'l of a 40 W fluorescent white lamp which is themost efficient visible light source available.

Most of the work related to II-VI chalcogenideshas been concentrated on the compounds which have relativelylarger band gap namely ZnS, ZnSe and CdS. Among these

4

ZnSe and CdS exhibited some typical semiconductor properties

under certain preparation conditions, However, ZnS stillstands out as the only material that possessfis the mvst desir­able feature in luminescent application, both in cathode­luminescence (CL) and in EL. It is still intrinsically themost efficient phosphor far surpassing GaAs in propertieslike quantum efficiency (QE) and brightness. Although thedefect chemistry and the physical properties of narrow andmedium band gap co-valent semiconductors are relativelywell known, this cannot be said of wide band gap semi­conductors such as ZnS which tend to host various type ofrecombination centres creating energy levels within theirband gaps. The efficiency of luminescence in such phosphorsis determined by the nature of the band structure and thevarious kinetic processes that occur at these centres. Thesecentres interact, and each type of centres plays a significantrole in determining the overall efficiency and the emissioncharacteristics of the luminescent materials. Hence for allmaterials, in particular for this IIQVI group of compounds,the procedure adopted for the preparation of the phosphorsand the nature and amount of the additives called activatorswhich create these centres are of great significance [ll].

1- 5 .1§_¢ti3ra1=Qrs-an§_@Or-act ivatgrs

The activators are certain acceptor type impuritiesintroduced into the pure host crystal lattice to produceluminescence. There are numerous activators that cause

5

luminescence in these materials as well as certain defectcentres that may act as if they were activators. Some ofthe elements commonly used as activators in II-VI phosphorsare copper, silver, aluminium, manganese and various lantha­nides. A number of these, copper and silver in particular,may occupy various sites in the crystalline lattice, therebycausing different emission colours and efficiencies. But formany activators to operate properly, it is necessary to in­corporate an additional donor impurity usually called as theco-activator. The presence of a co-activator in phosphorsmay be required to ensure physical incorporation of theactivator or to create certain energy levels within thephosphor band structure which aid the luminescence. Forexample, ZnS with Erbium activator and copper co—activator(Zn:Er,Cu) system, both of these functions are required andapparently performed by the copper co-activator. The copperaccomplishes this by performing the two separate and distinctfunctions, viz. charge compensation and sensitization. Whena tripositive ion is substituted for a zinc ion, it createsa charge imbalance that must be compensated by adding amonovalent ion, such as Cu+, to the lattice. This is knownas the charge compensation. An analogous situation occurswhen monovalent copper is present as an activator. In thiscase, a negative monovalent co-activator such as the halides,typically chlorine, bromine or iodine, is used. "Sensitizationinvolves the presence of energy levels within the crystalwhich transfer exciting energy to certain activators [ll,l4].

6

1. 4 _lj1iQs;Qh<>rs cbased c c 911, Zns

ZnS:Cu,Cl phosphor is the most extensively studiedEL phosphor. As discussed above, it is observed that theproperties of the phosphor depend very strongly on the condi­tions of preparation and on the concentrations of the activatorand co-activator present. For example it is found that thisphosphor can emit blue, green or red or its combinations[15,16]. In addition, in some cases IR emission at 1.5 and1.7 pm is observed [17]. There is also a very strong depend­ence of the emission colour on the excitation frequency andtemperature. And it is not at all certain whether these twoemission bands are due to two copper levels or due to themodification of the energy levels of the sulphur ions locatednearest to the copper ions because of the difference in ioniccharge associated with a zinc or copper ion [2]. However, inorder to explain the experimental observation G. Curie andD. Curie [18] and Riehl, Schon and Klasen [19] have suggestedtwo different models. In Curie's model (Fig.l.l) it isassumed that the ground state due to C'u+ ions plays a partin both blue and green centres in ZnS. But the green emissionoccurs from a transition arising from.the donor levels intro­duced by the oxygen or the co-activator ions. The blue emissionis either due to the transition from the conduction band orfrom an excited level of the luminescent centre which isshallower than.the donor levels involved in the green emission.But the Riehl-Schon-Klasen model (Fig.l.2) assumes that theempty low lying (blue) centres will be filled by electrons

CONDUCTION BAND-_____---~_-------- --__.__.....___.-_._ _...__ -__...__.._.._..-_...__-.--..__. -..____......._.._. ___._._ _ _.

v-.- _ T _.. -­P

4\

1‘

'”T_*"DONOR LEVEL

awe em Arssrom GREEN EMISSION\ ‘ ‘ACCEPTOR LEVEL(Cu)

VALENCE BAND

I'1¢.1.1. The model proposed by O. Curie and D. Oufor blue and green emission centres inZnS:Cu.

rie

'C ONDUCTION BAND__ __ _ i___,,_,__;A- _,_,,T[,_',, ,_~_~ ,-;’—,--...__.

H

(\ Y1* \V

LJL-'UE EMISSION ‘GREEN EMISSION

<1 'JL“I”\ l

""'i"_T"'

;1> ?F.-­

O11

._ _. -5. _ .. - .. -,VALENCE BAND

I1g.1.2. Riel-Soh6n—K1asen scheme. The transportof Gnwrgy from blue to green centres byhole migration is shown by the arrows‘QB Blfl C.

8

from filled higher (green) centres via the valence bandand the emission will be predominantly green if sufficienttime and activation energy (temperature) are available.Both these models explain some of the experimental resultsbut not all. The details of a series of studies which throwmore light on the role of Cu centre in ZnS made by the authorare presented in Chapter III of the thesis.

At present in ZnS, the activator of practical import­ance, other than copper, is manganese. ZnSnMn has an yellow­orange emission. Eventhough this is one of the earliest ELphosphor system a complete picture of the physical phenomenoninvolved in the emission process is not yet well understood.Recently Skolnick et al. [20] have made an attempt to studythe light emission mechanism with the aid of the time resolvedspectroscopic approach and obtained evidence for the hotcarrier impact excitation process in its EL Qmission. It isfound that the emission from ZnS:Cu,Mn is more efficient thanthe ZnS:Mn phosphor.

1. 5 Barbee. earth asle-eciv irate;

Rare earth (RE) dopants, as activators in lumines­cent systems (eg. in EL, CL and in photoluminescent (PL)phosphors) are of special importance because of their abilityto yield narrow band emissions. It is found that by choosingproper host and rare earth activator, emission in any regiorcovering the UV, visible and near IR regions, can be obtained.RE elements have been used as the activator in ZnS phosphors

9

by Bryant et al. [21]. Recently laser action has also beenreported by Zhong and Bryant [22] in a thin film ZnS:Cu,Nd,Clcell. One problem regarding this type of activator is thatit is quite difficult to introduce them in the ZnS crystaldue to smaller cation size of the lattice. So more sophisti­cated ion implantation technique or sometimes co-evaporationtechnique has to be adopted for this purpose [21].

1- 6 .N;e21:.rh@aPh9r. syrsrams

Experience in solid state physics in general and ELin particular has shown that cases where theory leads experi­ment are more of an exception than the rule [25]. So EL

researchers are constantly engaged in the synthesis of newand sometimes exotic materials, in addition to the attemptsto improve the existing phosphor systems. In their hunt fornew materials, they are sometimes gdided by the idea that agood CL phosphor may be a good EL phosphor as well. This isbecause the charge carriers and the luminescent centresinvolved in the two processes are the same but differs onlyin the mode of acceleration of the charge carriers.

One such EL phosphor which has gained some importance

recently is the ZnO phosphors doped with various activates[24-27]. The most significant achievements in recent timesis the conversion of the efficient alkaline earth sulphidephosphors developed by Lehmann [28] into good EL phosphors

by Vecht et al. [29]. They have developed a series of band

and line emitting EL phosphors. The notable one is CaS:Cewhich has a power efficiency comparable to the establishedZnS:Cu,Mn phosphors”

1- '7 liischanism 01131

Three distinct processes are involved in the ELphenomenon:

(1) Excitation of crystals by the applied electric fieldto an energy state atleast a few electron volts aboye theground state. The excited state can be an ordinary conduct­ion state of th% crystal, an excited state of an impurity

\.

.-.'.\

system or a state of high kinetic energy within the conduct­_ K­_.. .\. J

ion or valence band. The excitation process may inject.!'

minority charge carriers, field ionize valence electrons orimpurities, or accelerate charge carriers within a band to

»

optical energies.

(2) Transport of the excitation energy through thecrystals to a region where de-excitation can occur with large_1 | '.

probability for radiative transition. The energy transportcan be by charge carriers, exciton migration or by resonancetransfer.

(3) Radiative-de-excitation. This involves localizedstates oi impurity systems, interband transition or intra­band transitions [50].

ll

Various combination of these excitation, energytransport and emission process can occur in actual crystals[31]. Depending on the carrier generation process EL can beclassified into two types: (l) Intrinsic EL and, (2) Inject­ion EL.

1 - 8 c 111-"ins_i_9_.E§151

High field EL devices based on the II-VI semiconduct­ing compounds generally fall in this group. Curie [18] hassuggested a three step process for such an EL emission whichis given below:

(1) Transfer of electron from donor levels to theconduction band under the action of the applied electricfield and or temperature.

(2) Acceleration of these electrons in the conductionband and the subsequent carrier multiplication or an avalancheeffect.

(5) Collision of these electrons with luminescent centreswhich are thereby excited (ionized) or with the basic crystallattice itself followed by the production of electron holepairs.

However, Curie was not specific about the firstprocess and assumed that it is not the important rate deter­mining step. But Piper and Williams have independentlyconsidered field acceleration of electrons as the exciting

12

mechanism in EL. The quantitative treatment by Piper andWilliam [52] differs from that of Curie and considers thatthe initial production of carriers can influence the ELemission. They have also assumed that the phosphor possessesdeep donor levels lying perhaps uqQ.5 eV below the conductionband.

The effect of the field on the ionization of thesedonor levels can be explained on the basis of the idea putforward by Frenkel [18]. He has assumed that the field andthe phonon mutually assist in producing ionization of thelevels. If the depth of the level is s in the absence of thefield, the probability of the ionization per second is

P = S exp ( — §% )

which is very small if e is large. In the presence of thefield E the depth is reduced to

6* = e — f(E)

and the probability of the ionization becomes

where f(E) is the apparent change in trap depth due to theapplied field. So during first few cycles of the applied

13

field on an EL cell, the brightness will be very low.Gradually the electrons fall into the traps which at thebeginning were empty, and the ionization of the latter bythe same process as above soon plays the dominant role inthe supply of electrons to the conduction band. This willnaturally produce a build up in the EL emission and isactually observed [33]­

3-_1 . 9 .A..<t§ele-WP‘ 0.11 .02 Q2119 2har_a¢_. 2 arroioerrs

In order to accelerate charge carriers to opticalenergies, high fields capable of accelerating them to largekinetic energies must exist and also the carriers (holes orelectrons) must be injected into or created in this highfield region. The latter usually occurs by the field ionisa­tion of shallow donor levels as explained earlier. But hereconduction electrons experience an accelerating force as aresult of the field and a retarding force resulting frominteraction with the lattice phonon. with moderately strongapplied field, conduction electrons which experience un­usually few collisions with lattice phonon will attainsufficient energy to ionize impurities or valence electronsby inelastic collisions. A typical value of the fieldintensity is FulO5 volts/cm. At a much higher field(/\JlO6 volts/cm) average conduction electrons acquireenergy from the field faster than the rate at which energyis lost to the lattice phonons, thereby attaining sufficientenergy to ionize impurities or valence electrons by collision.

14

This may also lead to the conditions for dielectric break­down [30].

But in order to eliminate the catastrophic consequ­ence of dielectric breakdown Piper and Williams assumed thatthere occurs inhomogeneous field distribution within the ELcrystals, for example, a region where a narrow high fieldexists in series with an extensive low field region. Thiswill permit operation over a broad range of applied voltageswithout creation of an unstable breakdown condition. A cry­stal which is particularly stable against dielectric break­down is one having non linear characteristics such that thewidth of high field region as well as the magnitude of thefield increases with the applied voltage.

Piper and Williams [34,55] have also assumed thatin EL phosphors the acceleration is actually occurring atsuch a Mott-Schottky" exhaustion barrier (Fig.l.5) whicharises from the ionization of shallow donor levels. Theconcentration of deep donor is assumed to be much less thanthat of shallow donors so that the latter alone determinesthe potential distribution in the barrier region. Most ofthe applied potential will appear across the barrier andhence a moderate applied electric field in the barrier regionwill be sufficient to produce ionization of these deep donorsand subsequent acceleration and collision excitation can leadto luminescence. Zalm et al. [36] suggested that the donorsresponsible for the positive space charge barrier are the

i

r

‘r

Hti?\

4 \

K A__ *1 ."_'_i m 1

\

¢, ‘4, ‘W

J i;

; »w-i-' QFi5z1.3. Metteflchettky exhaustion layer (A) otential

F1g.1.4

configuration (B) distribution of e£ectro­static charge density.

\] 2\ ---9--...\ \ \

1 '1'

II

f-->-L 04'--1I-a, $ taI

CONDUC TION BAND5 ,_ ,-+ - -i-i0o~ one

-- ACTIVATOR cswrsns

VALENCE BAND

Schematic representation of acceleration collisionmechanism of electroluminescence. Electrons fromtraps in the localized-high field region are libe­rated by‘the action of the field (1) or injectedfrom a contact and then accelerated by the field (2)to acquire kinetic energy above the bottom of theconduction band and collide with activator centreswhereby they lose their energy and the activatecentre in ioniocd (4) or excited.

16

ionized activator centres themselves and that the initial

electrons are supplied by a surface layer of Cu2S.

In the case of Mott-Schottky type barrier an increasein applied voltage increases the breadth of exhaustion regionand field intensity increases as the square root of the appliedvoltage.

An intrinsic layer is another suitable type ofbarrier. This barrier remains constant in thickness with .applied voltage. The field which is independent of positionin the barrier varies linearly with applied voltage.

Yet another possibility is a p~n junction biased inthe reverse direction. An increase of the applied field inthis case increases the width of the depletion region andhere the field strength varies as the two third power of theapplied voltage [30].

1-10 Qgllisionexeitatign

The energy of these accelerated electrons is trans­ferred to various activator centres'myimC}lastic collisionsand they lose energy by release of further electrons fromthe centres into the conduction band (Fig.l.4). Bothcategories of electrons after impact possess some residualmotion along the accelerating field direction and will bein the same state they were before getting accelerated.

17

The other modes of energy transfer relevant to ELemission are the migration of minority charge carriers,exciton and resonance energy transfer processes [30].

1-llvfirishtness-volfiaeelgherecteristies

From the above discussion it can be seen that the

carriers which induce emission are generated in a high fieldregion formed within the crystal. It is also known that,depending on the type of barriers involved, the nature ofthe field producing the acceleration of the carriers variesconsiderably. This causes the number of carriers generatedto produce luminescence and hence the brightness to vary withthe type of barrier. In other words we can have an idea aboutthe type of potential barrier involved in the process bystudying the average brightness voltage characteristics.Destriau has obtained an empirical relation which relatesthe brightness (B) and the applied voltage (V) as

B = a exp (-b/V)

where a and b are constants. He has later modified thisrelation to account for the slight curvature of the log B

vs .% plot and wrote

B = a vn exp (-b/V)

The value of n is found to depend on the phosphor used [55].

18

A number of other empirical relationshavealso beenproposed by various workers and the most popular among them

is due to Alfrey, Taylor aha Zalm [56,37]. Here the bright­ness is given by

B = BO exp [ -( 2? )%]

where BO and V0 are parameters which depend on temperatureand frequency of the alternating voltage, phosphor type, andon details of the construction of the test cell. The squareroot term in the relation clearly shows that there involves aMott—Schottky type barrier for the acceleration of electrons.Over a limited range of voltage this equation may be approxi­mated with a power law. At very low voltages and thereforeat very low output the brightness may vary as the 9th or 10thpower of the applied voltage. At voltages approaching break­down of the layer the dependence usually approximates to V2or even to a lower power of V. In some cases it is smallerthan unity. But Thornton [38] pointed out that since thephosphor contains particles of different sizes, for a givenvoltage the field experienced by each particle will bedifferent and hence the observed result will be only theaverage.

1- 12 .T_1L_a<-2_d2;>en.91@n¢@ t - Oi EL

The obvious involvement of transport process,carrier trapping and related effects in EL emission is clear

19

from the above discussion. So it is quite natural to expecta dependence of the brightness of emission and its spectralcontents on the applied frequency of excitation and also onthe shape of the wave form.

Generally it is observed that for a sine wave ex­citation there exist two light peaks (known as brightnesswaves) per cycle of the applied voltage. The two peaks havedifferent amplitudes [55]. Various workers have observed aphase shift between the brightness peak and the voltagemaxima. Destriau [35] has attributed it to the phase shiftof the field in the dielectric medium surrounding the phosphorparticles. But Zalm [36] has attributed it to the delayedrecombination of the accelerated electrons. There occurs asecondary peak for green or blue emitting ZnS:Ou phosphors.These minor peaks are often not well resolved and occur near(and usually before) the instant the applied voltage passesthrough zero. The detailed characteristics ofthese peakswere studied by Destriau [40] Thornton [2] and Zalm [39].

Thornton [41] and later Georgobiani and Fok [42]assumed that the brightness wave is a kind of field cont;rolled glow curve since the rate determining factor is thefield controlled thermal release of trapped electrons.Thornton developed a model based on reduction of the trapdepth by the field. This mechanism facilitated return ofthe electron to the excitation region, and hence an earlierprimary peak, for high voltage and low frequency as well as

2O

for high temperature. Since it is known that trapping ofcharge carriers plays an important role in the temperaturedependence of integrated output, secondary waves, EL buildup etc., one can expect its effects on the brightnessfrequency characteristics also. This is because of somesort of matching (or mismatching) between the field frequencyand the trapping time which can affect the release of electronsfrom the traps and hence the brightness. Such a phenomenonwas not observed earlier. But the author has obtained clearevidence for the carrier trapping from the B-f characteristicof the various CaS phosphors and the same is described indetail in Chapter IV of the thesis.

1- 13 l1'1j@<2’°i°n

l.l5.l p-n Junction

Perhaps the simplest type of EL process is thatfollowing injection of minority charge carriers, either atan electrode contact or across a p-n junction. At such ajunction, in the absence of an applied voltage there is astate of dynamic equilibrium between the competing processesof thermal production and subsequent recombination ofelectron-hole pairs. Some of recombinations occur with theemission of radiation, which contributes to the normal thermal(black body) radiation of the material. When voltage isapplied in the forward direction and additional carriers areinjected, this equilibrium is upset and the rate of re­combination is increased (Fig.l.5). The resulting emission

N_ TYPE P-TYPE

4‘ A /“ n FORBIDDEN BAND

/ /nz/,VALENCE BAND A

\(§*2

n1~b° F?»“* n

\\

"é-5' """""" "' ‘:e\;P.*?'~_'E2_v9e AGE ®//HOLE ///A0’ //§, //// ///My/ INJECTION .

*-> EM/SSIQN QUANTA

H‘:B

Energy level diagram for a p-n Junction inthe absence of an applied field (A) gnd fqgcurrent flow in the forward direction vélhconsequent injection of minority vhargwcarriers (B).

22

of radiation may be called injection electroluminescence[43,44]. Lossew [1] was first to find such an EL from therectifying crystals of SiC. Since then this phenomena isobserved in a variety of compounds. Reviews of the subjectare available in the literature [5,4].

l.l3.2 Heterojunction

In order to inject only one particular type ofcarriers into a luminescent material usually a hetero junctionis made use of (Fig.l.6). In this case the source of theinjected carriers is a material having a wider band gap withrespect to the luminescent region. Then the asymmetry ofthe barriers seen by electrons and holes assures a high in­jection efficiency from the layer of the material with widerband gap. Although hetero junctions can be made betweencompletely different materials, the most successful devicesare those made from different compositions of miscible alloyshaving similar lattice constants at the temperature at which

they are fabricated, eg. All_XGaXAs with two different valuesof X on either side of the junction, commonly used in LEDsand diode lasers [45-47].

1.15.5 Schottky barrier

A Schottky barrier usually occurs at the surface ofsemiconductor in contact with a metal, such a barrier canact as p-n junction. Whenever surface states induce aninversion layer, the surface has the opposite type of

1­

‘ ‘A (AI (er‘i '_11L;_E9"7 N-TYPE "_"""“_"

--—-—---—-—-‘—--—-FL --—­it-n-1'o 0 66 6-5* *

P..TYPE 592

I1g.1.6. Diagram cf heterojunction (A) at ­equilibrium (B) with a forvard bias VJ.

(A) (B)‘|

Pi i i 1 1 i i i 1I

\ 7

I1|.1.7. Diagram of Sohottky barriers to n tyP° (A)and P-tYP@ (B) material. The arrows showthe minority carrier injection duringforward biaa.

conductivity compared with the bulk. In effect this isequivalent to a p—n junction immediately below the surface(Fig.l.7). A forward bias tends to flatten the bandstructure, allowing the injection of minority carriers intothe bulk, where they can recombine radiatively just as theywould in a p—n junction [43].

1.15.4 MIS structure

when surface charges are insufficient to inducethe desired inversion layer, one can induce them by acapacitive coupling to the surface through an insulator.In a metal-insulator-semiconductor (MIS) structure, thesurface charge on the semiconductor is determined by thepotential applied to the metal [48]. An immediate benefitof the MIS structure is that the phenomenon of band bendingcan be controlled by the applied voltage (Fig.l.8). Thusan inverse layer can be induced by one polarity, then thecarriers accumulated at the surface can be injected by

\3

reversing the polarity of the metal electrode. This methodhas been used to obtain EL in GaAs without p—n junction [49

If the insulator is made very thin electrons can also tunnelacross the insulator [50].

Once the minority carriers are injected, a varietyof recombination mechanisms are possible. Some of them donot lead to luminescence and thus contribute to the observed

low efficiency. These include processes like multiphononemission, Auger effect and non-radiative trapping by defects

"'7?

‘ ‘

N9 I 8

P13.1.8

£3! I B) !CI

Band structure at HIS Junction(A) without bias (B) metalnegatively biased fur Holegeneration (C) metal positivelybiased for reeombinatieno

26

At present the most successful electroluminescentdevice prepared from the II~VI compound semiconductor viz.ZnS:hn is the AC thin film EL device. These cells have asymmetric structure which can be thought of as having twoM18 junctions.

1-14 Q9- Powder EL-diS_P__...¥__1~‘-’~ S

The renewed interest in the EL of II-VI compoundshas resulted in improved devices and their wide applicationin different display systems. At present nearly 14 differenttype of EL devices are either available in the market or arein the different stages of development. Among them a promis­ing type of recent origin is the DC powder EL devices. Thesedevices can be operated at very low voltage (fiv5O V). Butthey are still in the experimental stage eventhough Vechtet al. [10]-have prepared a few good quality Direct CurrentPowder Electroluminescent (DCPEL) panels using ZnS:Cu,Mn and

CaS:Ce phosphors. Television display screens have also beendeveloped based on this technology [52-55]. These devicesessentially consist of the phosphor.powder coated with a

CuXS skin [56] and dispersed in a suitable insulator matrixand placed in between a pair of electrodes deposited in acrossed grill fashion. The peculiar nature of these devicesis that they have to undergo a current forming process tobecome luminescent [56]. A detailed account of the structure,the fabrication, the current forming and an analysis of thevarious characteristics of a DCPEL cell is given in Chapter VIIof the thesis;

PART—B

1.l5 Thin Film Electroluminescence - The Present StatusQilj __ii-—*jf _, 1 ’ 1 _;.:€_*+,%_ g _ — _: , _' _ ; _—-— _ i — ~~*~__ __iu_.:;, i i_1’ i W _ r i ’__i;;*_ i :1’_ e ;-~

1.15.1 Introduction

Electroluminescence in phosphors deposited in theform of thin films was first suggested by Halsted andKoller [57] in 1954. They observed a steep nonlinear B-Vcharacteristic in their devices. This property attractedThornton, a well known display technologist, to this newtype of device. Thornton [58] has made such a device butit lasted only for a few hours. Subsequently, a number ofworkers have entered into this new area. They were success­ful in developing devices of high brightness (j>lOOO ft L)and efficiency (PulO_4) [10,66]. But their poor workinglife still remained as a serious problem. Most of the earlywork on thin film EL was centred on thin film devices havingthe active phosphor layer in between two electrodes. Obviouslyone of these electrodes should be transparent. In additionto their poor maintenance, preparation of such large areadevices was difficult owing to the severe burn out and pinhole problems. Because of this technical constraints apractical and commercially viable display device of thistype has not become a reality despite the fact that theycan be operated at a relatively low voltage and can havehigh brightness levels. However, recently, Abdalla eta1. [ZND] have prepared an X4Y matrix display panel of theDC thin film type using a new preparation technique. At

28

present, such DC thin film EL (DCTFEL) cell structurewidely used for the experimental investigation of variousluminescent centres in different host materials [21].

l.l5.2 AC thin film EL devices

Russ and Kennedy [60] have attributed the failureof the DC cells to the lack of current limiting. So theyhave prepared an AC thin film EL device of double insulatinglayer structure. This structure, because of itssymmetric nature was able to minimize any slow polarization

effect and also offers the required protection by localcurrent limiting. This is possible since the current flow­ing through the active film (e.g. ZnS:Mn) would charge upthe interface with the dielectric and reduce the internalfield. The first successful application of this type ofcurrent limiting in a practical device was by Soxman andKetchpel [59]. They have employed an unsymmetrical AC

coupled structure to make segmented numeric and X-Y matrixdisplays. ZnS:Mn is the preferred active material for ACcoupled devices. But following the work of Kahng [61 ]there was also considerable interest in films of ZnS doped

with rare earth fluorides such as TbF5 which produces greenluminescence [ 62]. The studies on Nn doped ZnS filmsstarted early in the history of EL. Since 1960, Vlasenkoet al. [63-65] have made extensive investigation on ZnS:MnEL thin films and have published a number of interestingpapers. It seemed that all these workers were more

29

interested in the study of the physics of these devicesrather than giving importance for prolonging the life ofoperation. But the most significant impetus to developmentin the thin film EL for display applications came, however,with the report of Inoguchi et al. [66 ] of very high lumin­ance (;>lOOO fL), coupled with very long life (;>20,000 hrs)achieved through the use of ZnS:Mn in a symmetrical structure

similar to that of Russ and Kennedy but with a layer of YZO5which has a very high dielectric strength. A very strong

non-linearity of luminance vs voltage characteristic of thisdevice was evident and this is quite suitable for applicat­ion in display systems.

The interest among display technologist was furtherstimulated by the subsequent report [67] of hysteresisbehaviour exhibited by the luminance vs voltage variationsfor devices with the same structure but made by a slightlydifferent procedure. This report raised the possibility ofdisplays with inherent memory function. An EL display panel,making use of this memory phenomenon with 1248 characters( 120 X 160 mmg area) was reported by the Sharp Group [ 68]in 1975. From that year to the present, scores of publicat»ions have appeared describing the attempts to reproduce orimprove the results obtained by the above team. Othercommercial manufacturers presently in the market are:(1) Sigmatron Nova, U.S.A with a 38 mm X 58 mm device ofan average brightness of ‘-30 fT. [AG] and Lohfia Corporation.

30

Finland, which produces a transparent device fabricatedby,adopting a noval technique known as atomic or alternatelayer epitaxy (ALE). This technique was first demonstratedby Suntalo ['H)]. The films obtained by ALE are reportedto have better quality than those prepared by the convent­ional evaporation technique.

l- 16 leviceoeo Rbreliss

l.l6.l Non memory

The most important feature of the non-memory EL

device is the steep rise of its brightness with voltage and asaturation of the brightness at a high excitation field.(A typical characteristic curve is shown in Fig.l.9)The analysis of the characteristics involves two componentsthe equivalent circuit aspect and the internal mechanismassociated with the active layer. As shown in Fig.l.lO itcan be assumed that the device consists of a series capacit­ance made up of the two dielectric layers and an activelayer which has both capacitance and non linear conductance,i.e. at low fields the active layer acts as an insulatorbut at higher internal fields (in this case 1.5 x 106 V/cm)the breakdown of this layer occurs. This increases theelectrical conduction followed by the onset of light emiss­ion. The steep rise in brightness reflects the correspondingrise in current density as long as the voltage drop acrossthe ZnS:Mn layer continues to increase. At higher intensityall the charge carriers produced do not recombine and as a

I P3: 970+» A“ c, _L_

enronmesslml‘?§> " 55w»

.\F“ .‘ \‘ ‘’ ‘~

\7f ~ Aw, .1 ‘ ' Ll

‘ 44 \c, |2 '—f 1/ '10 ":=*~ =1: o-o.r.~o ii to _o iii: 'o so roo 250 zoo 250vouzaos (vows; g

Fis.1.i9 r1;.1.1o (A)Typical B-V-oharaoteristio The oquivaloni circuit

‘>\‘- M' \Electrode * ".‘ ‘.‘\W ._

I\

r--1 |--­

3d5OFMn 1;'777777W/~EIecrrode\ 1

r

lflccrrddt ZnS Elccfrodellr1g.1;1o (2)

Shows the rain features of the field clampingmodel(l) high field induced tunneling from statesnear the interface (2) oolliaion excitation of Mn**

_ by olocirona accelerated in the_hi§h field ())retrapping near tho opposite inter ace. ,

32

result, a surface charge will be formed at the active layerdielectric interface. This will offset the increasedexternal voltage and is known as the field clamping effect[ 71,72]­

The internal efficiency of thin film EL devicesdepends upon: the field at which current flows, the fractionof electrons at such fields with sufficient energy to pro­duce internal excitation of the activators, the cross actionand density of activators and finally on the luminescentprobability of the activator. Mn2+, for example, shows astrong concentration quenching, even in PL, at concentrationsgreater than about 0.5 wt percent in ZnS. It is well knownthat Mn2+ has a more favourable cross section for excitationthan the rare earth which is probability related to thedifference between d electrons and f electrons, but thereare efforts still underway [62] to realize the lumocenconcept of Kahng [61] whereby a complex (presumably withlarge cross section) is impact-excited and then, via.internal coupling, energy is transferred to the activator.

1.16.2 Device Physics-Memory

Yamauchi et al. [ 67] were the first to report thehysteresis behaviour in the brightness vs applied voltagecurve of the thin film EL device with double insulatinglayer structure. The typical nature of this hysteresisis depicted in Fig.l.ll. One practical consequence of this

GHTN£S$ (fL)

L

BR

7’250; 4

59

‘\

-SO‘

\

‘ k

I

ia~»%

780 790 ZOO 270APPLIED VOLTA GE

Iig.1.11 (A). Typical B-V characteristic of a hyateretic

INTERFACE ‘STATES §

omscrn/c\ H“ Q

o\

thin film EL device. The right hand curvesare obtained for increasing voltagc.

\i >

\

1

K3 $

4

0’

/c»rn

LECTRIC

§

\Iig.1.1l (B). Band diagram of an EL memory device

illuetrating echematically some of themechanisms involved ouch an tunnel inJoe­tion, impact excitation, impact ionizationand hole trapping.

34

characteristics is that in a matrix display all elementscan be excited with a steady AC sustaining voltage withinthe range of hysteresis, at a relatively low frequency.Individual cells can then be switched on or off by transi­ent wave form applied to the X-Y electrodes, and the overallresult is a more efficient display system. Additionalpractical interest in the hysteresis behaviour was generatedby the observation that the device could be switched bylight [73 1 and by electron beam [ 74,75] thus making themsuitable for applications like image storage and in highcontent storage cam tube displays [ 75,75]. Another importbant application of these devices are in the fabrication ofimage amplifiers [ 77].

Inoguchi et al. [66 ] have explained the occurrenceof this hysteresis behaviour on the basis of a charge storageand the release of trapped electrons from some deep levelsand then subsequent detrapping at the removal of the field.

They have also verified the existence of the deep traps fromthe thermally stimulated current measurement across the activelayer.

Howard believes that eventhough the charge storagephenomenon is important in memory behaviour, this by itselfcannot produce the observed bistability. So Howard and hiscolleagues [ 78-80] proposed a model to explain the observedbistability by incorporating the idea of lattice ionization,electron-hole pair production, electron heating, and theimpact excitation luminescence.

35

There are several key empirical requirements forproducing hysteresis behaviour. The active layer filmsmust have good crystallinity, must be thicker than about0.5 pm and should be activated with Mn with a concentrationgreater than 0.5 wt percent. The former two conditions canbe associated with the necessity of inducing enough electronheating to produce ionization. The latter requirement isto form deep traps in order to create and maintain spacecharge. Essentially the bistability is a consequence ofAC coupled negative resistance which arises because of thespace charge induced by the trapping of holes followingimpact ionization of the lattice [71].

1- 1"! New <%Q1’19?_P?°-$

The observed maximum efficiency of a high fieldEL device of the type discussed above is about one percent.Ferd Williams [81] has attributed this low efficiency tothe low effective temperature of the hot carriers. It isobserved that all cathodoluminescent materials willluminescence when placed in contact with hot carriers andCL efficiency can be as high as 30 percent. This promptedthe idea that if electrons can be accelerated rapidly tooptical energies in high field EL devices efficienciesapproaching this value might be obtained. To achieve this,the functions of acceleration and collision excitationprocesses may be separately performed in different layers

36

so that their independent optimization can be achievedmore conveniently. The schematic diagram of the proposedstructure is shown in Fig.l.l2.

In this type of devices, electrons are injectedinto insulating layers such as SiO or even ZnS at fieldssuch that there is considerable electron heating. The hotelectrons so produced can then be injected into a secondmaterial where they may produce impact ionization as in CLfollowed by recombination at donor acceptor activator pairssuch as Ou,Cl. One luminescent layer which has been used

is ZnF2:Mn. This effort has not yet resulted in improvedefficiency. However, the work of Hamakawa et al. [82]can be considered as the implementation of the above conceptinto a practical device.

One of the drawbacks of AC thin film devices is

their higher operating voltages. Attempts have been madeto reduce the operating voltage by using piezoelectricfilms as insulator layers. One such device has been report­

ed by Okamoto et al. [35] with PbTiO3 (eeul5O). They haveprepared devices which can be operated at around 5O V.Another problem is regarding the preparation of deviceswith reproducible hysteresis behaviour. It is found thatwhen these devices are operated at high brightness, adrifting in voltage of the hysteresis loop accompanied bya change of shape of hysteresis curve occurs [78]. The

ACCELERATOR PHOSPHOR ACCELERATORH FILM FILM FILM

QR 1* ‘vJ? £3-Aii AIR }

M;:L;**

I1g.1.12. Schematic band diagram of theproposed structure to separateacceleration and excitation.

7"’

38

exact reason for the observation of these undesirablequalities are yet to be understood.

1.18 Summary

Part A of the chapter gives a detailed descriptionof the basic phenomenon of electroluminescence together withan account of AC and DC powder EL devices. The fundamentalprocesses associated with EL emission are mentioned. Briefoutline of the theoretical concepts are given. Part Bpresents a concise review of ACTFEL deviceswhich are of

current importance. Here the latest trends and possibilitiesin this area are also discussed.

Rsfsrsnces

1. O.W. Lossew; Telegrafia i Telefonia l8(l925)6l.

2. H.F. Ivey; Advances in Electronics and Electron Physics,Suppl.1, Electroluminescence and Related Effects(Academic Press, New York, 1963).

5. A.A.Bergh and P.J. Dean;Monographs in Electrical andElectronic Engineering, Light Emitting Diodes(Clarendon Press, Oxford, 1967).

4. F.W.Wil1iams and R. Hall, International Series in theScience of Solid State Volume 13, Luminescence andthe light emitting diode (Pergamon Press, Oxford,New York, Toronto, Sydney, Paris, Frankfurt, 1978).

5. T.Mishina, W. Quan-kun and K. Takahashi; J. Appl. Phys.59(l98l) 5797.

6. H.F. Ivey; IRE Trans. ED-6 (1959) 203: J. Electro Chem.Soc. 108(196l) 590; Electrochem. Techno1.l(l963)42.

7. H.K. Henisch; Electroluminescence (Pergamon Press,Oxford, 1962).

8. A.G.Fisher; Luminescence of Inorganic Solids, Ed.P.Goldberg (Academic Press, New York, l966)54l.

9. F.F.Morhead; Physics and Chemistry of II-VI compounds,Eds. M.S. Ayen and A. Prener (North Holland,Amsterdam, 1967) 613.

10. A. Vecht; J. Cryst. Growth, 59(l982) 81.

11. Elliott Schlam; Proc. IEEE, [7161 (1975) 894.

12. G. Destriau; J. Chem. Phys. 35(l956) 587.

13. W. Lehmann; J. Electrochem. Soc. lO5(l958) 585.

14. M. Aven and J.S. Prener, Eds. Physics and Chemistry ofII-VI compounds (North Holland, Amsterdam,Netherland, 1967).

15. M.H. Aven and R.M. Potter; J. Electrochem. Soc.l05(l958) 154.

16. H.C. Froelich, J. Electrochem. Soc. lOO(l953) 280.

17. G.F.Garlick and M.J. Pambleton; Proc. Phy. Soc.London, B67(l954) 442.

l8. D. Curie, Luminescence in Crystals (Methuen and Co.Ltd, London, 1963)

19. P. Zalm, G. Diemer and H.A. Klasens; Philips ResearchReports 9(l954) 81.

20. M.S. Skolnick and Paul J. Dean; IEEE Trans. on ElectronDevices ED-28[4] (1981) 444.

21. C.C. Yu and F.J. Bryant, Solid St. Comm. 28(l978) 855;G.Z. Zhong and F.J. Bryant, J. Phy. C. Solid St.Physics [4] l5(l98O) 797.

99- G-'7.- 7.hnn.o' srnfi 1-?‘-.T1 'R'r~v:an1'.- 901-16 S13- Comm- "§Q(lQ8l) 907

Egon E. Loehner; Proc. IEEE [7] 61 (1973).

S. Bhushan M. Saleem and S. Chandra; Pramana 1O[1](1978) 1.

L.N. Tripathi, B.R. Chaubey, C.P. Mishra; Phys. StatusSolidi A 57(198o) K 157.

4. Pfahnl; J. Electrochem. Soc. 110 [5] (1963) 581.

S. Bhushan, A.N. Pandey and Balakrishna Rao Kaza;Phys. Stat. Solidi (a) 46(1978) K 123.

W. Lehmann; J. Lumin. 5(1972) 87.

A. Vecht, M. Waite, M.H. Higtan and R. Ellis;J. Lumin. 24/25 (1981) 917.

W.W. Piper and F.E. Williams; Electroluminescence,Solid State Physics Vol. 6, Eds. Frederik Seitz .and David Turnball (Academic Press 1nc., NewYork, London, 1958)

W.W. Piper and F.E. Williams, Phys. Rev. 98(1955) 1809.

L. Martin, Solid State Luminescence Ed. F.E. Williams,Advances in Electronics Vol.5 (Academic Press, NewYork, 1953) 137.

C.H. Haake, J. Appl, Phys. 28(1957) 245.

W.W. Piper and F.E. Williams; Brit, J. Appl. Phys.Suppl. 4, 8 39 (1955).

G. Destriau and H.F. Ivey; Proc. of IRE lO(1955) 1911.

P. Zalm, D.G. Diemer and H.A. Klasens; Philips ResearchReports 1O(1955) 205.

G.F. Alfrey and J.B. Taylor; Proc. Phy. Soc. LondonB 68(1955) 775. Brit. J. Appl. Phys. Suppl.[4] (1955) 44.

W.A. Thornton; J. Appl. Phys. 52(196l) 2379.

41

P. Zalm; Philips Research Reports [ll] 353 [1956] 417.

‘G. Destriau; Brit. J. Appl. Phys. Suppl. No.4,(l955) 49.

W.A. Thornton; Phys. Rev. 1o2(1956) 38; 1o3(1956) 1585.

42. A.N. Georgobiani and M.V. Fok; Optics and Spectroscopyll(l96l) 48.

43. J.I. Pankove; Electroluminescence, Topics in Appl. Phys.Vol.)l7, Ed. J.I Pankove (Springer-Verlag, Berlin,1977 .

44. D.V. Skobeftsyn; Ed. Electroluminescence, Proc. of theP.N. Lebedev Physics Institute Vol.5O (ConsultantsBureau, New York, London, 1972).

H. Kroemer; Proc. IRE 45(l957) 1555.

A.G. Fisher; Luminescence of Inorganic Solids, Ed. byP. Goldberg, Academic Press, New York (1966)

H. Kressel; J. EleCtrOn. mat. 4(1975) 1081.

48 A.S. Grove; Physics and Technology of Semiconductordevices (Wile and Sons, New York, London, Sydney,Toronto, 1969? 425.

49. C.N. Berglund; Appl. Phys. Lett. 9 (1966) 441.

J.I. Pankove and P.E. Norris; RCA Reviews 53(l972) 557.

51. W. Heinke and H.J. Queisser; Phys. Rev. Lett. 55(l974) 55.

Hiroshi Kawarada and Nobumasa Obshima; Proc. IEEE [7]61(1975). ­. M. Yoshiyama; Electron 42(l969) 114.

M. Yoshiyama, H. Kawarda and T. Sato; Proc. IEEEInt. Computer Group Conf. (Washington DC,1970) 261.

55. A. Arin, T. Yoshizawa, K. Awazu, K. Kuruhashi andS, Ibuki; Proc. IEEE Conf. Display Devices(New York- 19703 ‘-39­

A. Vecht, N.J. W r in§ R. Ellis and P.J.F. Smith;Proc. IEEE Y7? o (1973) 902.

R.E. Halsted and L.R. Koller; Phys. Reva 95(1954) 549.

W.A. Thornton; J. Appl. Phys. 35(l962) 3045.

E.J; Soxman and R.D. Ketch Pel; ElectroluminescentThin Film Research, JANAIR Report 72o9o3(1972).

M@J. Russ and D.I. Kennedy; J. Elaotrochemo Soc.7 114(1967) loos.

D. Kahng; Appl. Phys. Lett. l3(l968) 210.

J. Benoit, P. Benalloul and B. Blanzat; J. Lumin.[1.2] 25<1981> 175. ~

N.A. Vlasenko; Opt. Spect. l8(l965) 260.

N.A. Vlasenko and A.M. Yaremko; Opt. Spect. l8(l965) 263

N.A. Vlasenko, S.A. Zynio and Yu V. Kopytko; Phys.Stat. Sol. (a) 29(1975) 671.

T. Inoguchi and S. Mito; Eleotroluminescence,Topics in Appl. Phys. Vol. 17 Ed. J.I. Pankove(Springer-Verlag, Berlin, I977).

Y. Yamauchi, H. Kishishita, M. Takeda, T. Inoguchi andS. Mito; Digest Int. Elect. Dev. Mfg. IEEE(New York, 1974) p. 552.

M. Takeda, Y. Kakihara, M. Yoshida, M. Kawaguchi,H. Kishi Shite, Y. Yamauchi, T. Inoguchi ands. Mito,'75 SID Ihtorh. Symp., Sect. 7.s(1975)

M.P.R. Panicker, R.J. Bell, S.K.Tiko; SID Inter.Symp. Digest (1980) 112 and Sigmatron NovaDisplay Spec.

T. Suntola, J.J. Antson, A. Pakkala and S.Lindfors;SID Intern. Symp. Digest (1980) 108.

71. w.E. Howard; J. Lumin.24/25 (1981) 855.

8. Tornquist; J. Cryst. Growth 59(1982) 599.

7 . D.H. Smith; J. of Lumin. 23(l98l) 209.

W.E._Howard and P.M. Alt; Appl. Phys. Letters,31 (1977) 599.

O. Sahni, P.M. Alt, D.B. Dove, W.E. Howard andD.J. Mo Clure; IEEE trans. Ed-28 [6] (1981) 708.

D. Theis; J. Lumin. [1,2] 23(1981) 191.

P.K.C. Pillai and Nilonfee Shroff; National Conferenceon Instrumentation Proc. (1983) 285.

78. W.E. Howard; J. Lumin. [1,2] 23(l98l) 155.

79- P.M. Alt, W.E.Howard and O. Sahni; IEEE Trans. ED26(l979) 1850.

80. W.E. Howard, O. Sahni and P.M. Alt; J. Appl. Phys.[1] 55(1982) 859.

Ferd Williams; J. LumhL23(l98l) l.

K. Okamoto and Y. Hamakawa; Appl. Phys. Lett.,35(l978) 508­

K. Okamoto Y. Nasu and Y. Hamakawa; IEEE Trans. ED.28 [61 (1981) 705.