Light-emitting electrochemical cell

Abstract

A light-emitting electrochemical cell comprising a first electrode, a second electrode, and at least one light-emitting active material comprising a combination of an electrolyte, a first constituent comprising a host compound and a second constituent comprising a guest compound. A quotient between a difference in LUMO energy level between the first and second constituent, E.sub.trap.sup.n, and a difference in HOMO energy level between the second and first constituent, E.sub.trap.sup.p, is 1/10 to 10, a quotient between an electron mobility and a hole mobility on the first constituent is 1/100 to 100, a quotient between a number of ions of the electrolyte and a number of molecules or repeat units of the second constituent is to 5, and a LUMO energy level of the electrolyte is higher than the LUMO energy level of the first constituent and a HOMO energy level of the electrolyte is lower than the HOMO energy level of the first constituent.

Claims

1. A light-emitting electrochemical cell comprising: a first electrode; a second electrode; and at least one light-emitting active material separating the first and second electrodes, wherein the light-emitting active material comprises a combination of an electrolyte, a first constituent comprising a host compound and a second constituent comprising a guest compound, wherein: a quotient between a difference in LUMO energy level between the first and second constituent, E.sub.trap.sup.n, and a difference in HOMO energy level between the second and first constituent, E.sub.trap.sup.p, is 1/10 to 10; a quotient between an electron mobility and a hole mobility on the first constituent is 1/100 to 100; a quotient between a number of ions of the electrolyte and a number of molecules or repeat units of the second constituent is to 5; and a LUMO energy level of the electrolyte is higher than the LUMO energy level of the first constituent and a HOMO energy level of the electrolyte is lower than the HOMO energy level of the first constituent.

2. The light-emitting electrochemical cell of claim 1, wherein the quotient between the difference in LUMO energy between the first and second constituent and the difference in HOMO energy between the second and first constituent of the light-emitting active material is 1/9 to 9.

3. The light-emitting electrochemical cell of claim 1, wherein the quotient between the electron mobility and the hole mobility on the first constituent in the light-emitting active material is 1/50 to 50.

4. The light-emitting electrochemical cell of claim 1, wherein the quotient between the number of ions of the electrolyte and the number of molecules or repeat units of the second constituent is 0.5 to 4.

5. The light-emitting electrochemical cell of claim 1, wherein a difference in energy gap, i.e. a difference between the LUMO energy level and the HOMO energy level, between the first and second constituent is 0.1 eV-1.5 eV.

6. The light-emitting electrochemical cell of claim 1, wherein a quotient between the number of molecules or repeat units of the first constituent and the number of molecules or repeat units of the second constituent is in a range between 300 and 1.

7. The light-emitting electrochemical cell of claim 1, wherein a thickness of the light-emitting active material layer is 50-2000 nm.

8. The light-emitting electrochemical cell of claim 1, wherein the first constituent of the light-emitting active material comprises at least one solution-processable semiconductor.

9. The light-emitting electrochemical cell of claim 8, wherein the at least one solution-processable semiconductor is selected from a group comprising polymers, oligomers, small molecules, and neutral and ionic transition metal complexes.

10. The light-emitting electrochemical cell of claim 1, wherein the second constituent of the light-emitting active material comprises at least one emissive solution-processable semiconductor.

11. The light-emitting electrochemical cell of claim 10, wherein the at least one emissive solution-processable semiconductor features triplet emission and is selected from a group comprising neutral and ionic transition metal complexes, quantum dots, polymers, oligomers, and small molecules.

12. The light-emitting electrochemical cell of claim 1, wherein the electrolyte is selected from a group comprising ionic liquids, salts dissolved in an ion-solvating material, and ionic transition metal complexes.

13. The light-emitting electrochemical cell of claim 1, wherein one or both of the first and second electrodes is transparent or semitransparent.

14. The light-emitting electrochemical cell of claim 13, wherein one or both of the first and second electrodes is coated with one or more layers of a material or materials selected from a group comprising poly(3,4-etylenedioxythiophene)-poly(styrene sulfonate) (PEDOT:PSS), ZnO and graphene.

15. The light-emitting electrochemical cell of claim 1, wherein the first constituent consists of one or two host compounds.

16. The light-emitting electrochemical cell of claim 1, wherein the second constituent consists of one guest compound.

17. A method of operating a light-emitting electrochemical cell of claim 1 by providing a voltage over the electrodes of the light-emitting electrochemical cell.

18. A method of operating a light-emitting electrochemical cell of claim 17 by providing the voltage at a value sufficient to achieve a current density of 0.1 to 5000 A/m.sup.2.

19. The method of claim 17, wherein an external quantum efficiency of the light-emitting electrochemical cell, void of an outcoupling film or structure, is larger than 2%.

20. The method of claim 17, wherein more than 50% of photons emitted from the light-emitting electrochemical cell are emitted from the second constituent.

21. A method for producing a light-emitting electrochemical cell, comprising: providing a substrate; providing a first electrode on the substrate; providing at least one light-emitting active material layer on the first electrode; and providing a second electrode on the active material layer, wherein: the light-emitting active material layer comprises a combination of an electrolyte, a first constituent and a second constituent, wherein a quotient between a difference in LUMO energy level between the first and second constituent, E.sub.trap.sup.n, and a difference in HOMO energy level between the second and first constituent, E.sub.trap.sup.p, is 1/10 to 10: a quotient between electron mobility and hole mobility on the first constituent is 1/100 to 100; a quotient between the number of ions of the electrolyte and the number of molecules or repeat units of the second constituent is to 5; and a LUMO energy level of the electrolyte is higher than the LUMO energy level of the first constituent and a HOMO energy level of the electrolyte is below the HOMO energy level of the first constituent.

22. The method of claim 21, wherein the light-emitting active material layer is provided on the first electrode by solution processing under an ambient air pressure of at least about 1 kPa.

23. The method of claim 21, wherein the light-emitting active material layer is provided on the first electrode by spray-coating.

24. The method of claim 21, further comprising a step of providing one or more layers between an electrode and the light-emitting active material layer in the light-emitting electrochemical cell, wherein the material or materials of the at least one layer is selected from a group comprising poly(3,4-etylenedioxythiophene)-poly(styrene sulfonate), ZnO and graphene.

25. The method of claim 21, further comprising a step of encapsulating the light-emitting electrochemical cell so that oxygen and water penetration into the active material layer is reduced.

26. The method of claim 21, further comprising a step of introducing an outcoupling film or structure to increase light output from the light-emitting electrochemical cell.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) FIG. 1a shows a schematic of a theoretical target host-guest LEC at open circuit.

(2) FIG. 1b shows a schematic of a theoretical target host-guest LEC during steady-state operation.

(3) FIG. 1c shows symbols used in FIGS. 1a and 1b (and FIGS. 1d-1f and FIGS. 9c-d).

(4) FIG. 1d shows a simulated steady-state concentration and voltage profiles for a symmetric host-guest LEC. The left y-axis represents the density (nm.sup.3) of relevant charge carriers (anions, cations, free electrons, free holes, trapped electrons, and trapped holes) and the recombination rate (10.sup.4nm.sup.3s.sup.1) of electrons and holes into excitons.

(5) FIG. 1e shows a similar device as FIG. 1d but with a deepened electron-trap level. The left y-axis represents the density (nm.sup.3) of relevant charge carriers (anions, cations, free electrons, free holes, trapped electrons, and trapped holes) and the recombination rate (10.sup.4nm.sup.3s.sup.1) of electrons and holes into excitons.

(6) FIG. 1f shows a similar device as FIG. 1d but with a decreased electron mobility in the host. The left y-axis represents the density (nm.sup.3) of relevant charge carriers (anions, cations, free electrons, free holes, trapped electrons, and trapped holes) and the recombination rate (10.sup.4nm.sup.3s.sup.1) of electrons and holes into excitons.

(7) FIGS. 2a-2d show the CV traces recorded on thin films of different host and guest compounds, as specified in the insets.

(8) FIG. 2e shows the CV traces recorded for THABF.sub.4 as electrolyte (solid line) and a TMPE-OH:LiCF.sub.3SO.sub.3 as electrolyte (dashed line).

(9) FIG. 2f shows the electron-energy levels of the host and guest materials, as extracted from the CV data.

(10) FIG. 2g shows the triplet energies of the host and guest compounds.

(11) FIG. 2h shows the absorption spectrum of the guest compound (solid black line) and the PL spectra of the three host materials.

(12) FIG. 3 shows the CV traces of a PVK-OXD-7:Ir(R-ppy).sub.3 film, PVK-OXD-7 film, and a Ir(R-ppy).sub.3 film.

(13) FIG. 4a shows the electron-energy levels of some of the host compounds and electrolytes.

(14) FIG. 4b shows the CV trace in the cathodic regime for an OXD-7 film using THABF.sub.4 as electrolyte. (The scale of the y-axis ranging from 610.sup.5 to 0 A.)

(15) FIG. 4c shows the CV trace in the cathodic regime for an OXD-7 film using TMPE-OH:LiCF.sub.3SO.sub.4 as electrolyte. (The scale of the y-axis ranging from 210.sup.3 to 0 A.)

(16) FIG. 4d shows the CV trace in the anodic regime for a PVK film using THABF.sub.4 as electrolyte. (The scale of the y-axis ranging from 110.sup.3 to 310.sup.3 A.)

(17) FIG. 4e shows the CV trace in the anodic regime for a PVK film using TMPE-OH:LiCF.sub.3SO.sub.4 as electrolyte. (The scale of the y-axis ranging from 510.sup.4 to 110.sup.3 A.)

(18) FIGS. 5a-5d show the temporal optoelectronic response of ITO/PEDOT:PSS/Host:Ir(R-ppy).sub.3:Electrolyte/Al LECs with the host and electrolyte selection defined in the insets.

(19) FIG. 6a shows the electroluminescence spectra of ITO/PEDOT:PSS/Host:Ir(R-ppy).sub.3:electrolyte/Al LECs, with the host and electrolyte selection identified in the inset.

(20) FIG. 6b shows the electroluminescence spectra of ITO/PEDOT:PSS/PVK:OXD-7:Ir(R-ppy).sub.3:THABF.sub.4/AI LECs at different guest concentrations, as specified in the inset.

(21) FIG. 7 shows a table summarizing the performance of selected host-guest LECs.

(22) FIGS. 8a-c show the temporal optoelectronic response of ITO/PEDOT:PSS/Host:guest:THABF.sub.4/Al LECs with the host and guest selection defined in the inset.

(23) FIG. 9a shows the measured current efficacy (solid squares) and power conversion efficacy (open circles) as a function of the current density for the host-guest LEC, with a guest concentration of 29% and with the PVK:OXD-7 blend host.

(24) FIG. 9b shows the current efficacy (left) and the external quantum efficiency (right) as a function of guest concentration for the ITO/PEDOT:PSS/PVK:OXD-7:Ir(R-ppy).sub.3:THABF.sub.4/AI LEC (solid squares) and a ITO/PEDOT:PSS/PVK:OXD-7:Ir(R-ppy).sub.3/Ca/Al OLED (open circles).

(25) FIG. 9c shows the simulated steady-state concentration profiles for a host-guest OLED, with realistic injection barriers for electrons (0.2 eV) and holes (0.5 eV). (The left y-axis represents density (nm.sup.3) and recombination (10.sup.4nm.sup.3s.sup.1).)

(26) FIG. 9d shows the simulated steady-state concentration profiles for a host-guest LEC with an injection barrier of 1 eV for both electron and hole injection. (The left y-axis represents density (nm.sup.3) and recombination (10.sup.4nm.sup.3s.sup.1).)

(27) FIG. 9e shows the initial evolution of the luminance (left axis) and the current efficacy (right axis) of pristine ITO/PEDOT:PSS/PVK:OXD-7:Ir(R-ppy).sub.3:THABF.sub.4/Al LECs featuring a thin film comprising a hexagonal array of hemispherical microlenses (MLA-LEC, open circles) or a glass half sphere (GHS-LEC, solid squares) as the outcoupling structure.

(28) FIG. 9f shows the uniform green light-emission from a 45 cm.sup.2 large-area host-guest LEC fabricated by spray-sintering under ambient air. Inset: photograph depicting the airbrush performing spray-sintering.

(29) FIG. 10a shows the long-term operation of a GHS-LEC driven by j=0.39 mA/cm.sup.2.

(30) FIG. 10b shows the initial temporal evolution of the luminance and the current efficacy of the GHS-LEC (open circles) and the MLA-LEC (closed squares).

THEORY

(31) To increase the efficiency of a LEC device at high luminance, tailored traps may be introduced in the light-emitting active material layer sandwiched between and contacting the two electrodes of the LEC, blocking exciton diffusion that precedes quenching in conventional LECs. To circumvent the undesired side effect of a reduced mobility of the mobile charges in the p- and n-doped regions due to the very same traps, the trap and ion concentrations may be balanced in such a way that in the p- and n-type regions the traps are filled by the ion-induced doping, leading to a nearly trap-free transport.

(32) FIG. 1a shows a schematic of a target host-guest LEC active material, with the guest energy levels (dashed lines) symmetrically positioned within the energy gap of the host (solid lines).

(33) FIG. 1b displays its steady-state operation when the electrolyte ions have assisted with the doping (filling) of essentially all of the guest traps, as well as some of the host sites, in distinct doping regions next to the electrodes to facilitate for a large current. The central p-n junction is left doping-free to allow for efficient emission from immobile triplet excitons localized on guest traps. (Symbols used in FIGS. 1a and 1b are identified in the panel in FIG. 1c.)

(34) To investigate whether the target LEC operation is attainable, and, if so, to establish rational device design guidelines, numerical simulations were performed on a variety of different host-guest LEC architectures.

(35) Simulations

(36) The simulated steady-state concentration and voltage profiles for the best performing LEC device having a symmetric host-guest active material is shown in FIG. 1d (symbols identified in FIG. 1c). All ions have here performed electrochemical doping, as evidenced by that the ion concentration (squares) is equal to the sum of the free polarons on the host (circles) and the trapped polarons on the guest (diamonds) throughout the active material. The only exception is next to the electrode interfaces, where electric-double layers have formed to compensate for the significant energy difference (1 eV) between the Fermi level of the injecting electrode and the accepting energy level of the primary transport material, i.e. the hostthis characteristic feature of LECs that allows the use of air-stable electrodes is not affected by the introduction of traps.

(37) The coupled ion/polaron concentration profile defines a distinct doping structure for a 130-nm thick device, with 45-nm-thick p-type doping and n-type doping regions bridging the p-n junction, where the total polaron concentration remains very low. It was found that essentially all guest traps (c.sub.trap=0.03 nm.sup.3) were filled in the doping regions, while the emission was observed to originate solely from immobile excitons (stars) positioned on dispersed guest molecules in the p-n junction. The device thus accomplished the targeted doping and emission profiles, as schematically indicated in FIG. 1(b), and the in silico functionality of the concept was evidenced by the high simulated electron-to-photon quantum efficiency of =51% (assuming 100% photoluminescence quantum efficiency of the guest emitter and perfect outcoupling) at a large current density of j=4.5 mA/cm.sup.2.

(38) A second outcome of the extensive simulation study was that the performance of the host-guest LEC active material was highly sensitive to the selection and symmetry of a number of controllable device parameters. FIGS. 1e and 1f present the concentration and voltage profiles for devices with asymmetric hole and electron trap depths and asymmetric hole and electron mobilities in the host, respectively, as specified in the insets. When the electron-trap depth is increased (FIG. 1e), or the electron mobility is lowered (FIG. 1f), the exciton profile is strongly shifted towards the n-type doping region, with the consequence being a notably increased overlap between the electron-polaron and exciton populations. The concomitant increase in exciton-polaron quenching is manifested in that the efficiency dropped to =22% (at j=2.8 mA/cm.sup.2) for the asymmetric trap-depth device and to =40% (at j=1.8 mA/cm.sup.2) for the asymmetric mobility device.

(39) For the numerical simulations, a 2D drift-diffusion model was used as described in van Reenen et. al., Fundamental Tradeoff between Emission Intensity and Efficiency in Light-Emitting Electrochemical Cells. Adv. Funct. Mater. 2015, 25, 3066-3073. In brief, the model solves the coupled continuity equations for electronic and ionic charges and Poison's equation on a 2D grid by forward integration in time until steady-state has been reached. Motion of all charged species is described by the drift-diffusion equation assuming that the Einstein relation holds. Trap levels are implemented as discrete energy levels whose steady-state occupation is determined by Fermi-Dirac statistics. In the absence of traps, electron-hole recombination is described as a Langevin process; with traps present, recombination is described as a Shockley-Read-Hall process. In both cases, the recombination rate constant is R=q.sub.R/.sub.0.sub.r where the recombination mobility .sub.R equals the (sum of the) mobility (mobilities) of the mobile carrier(s). In the absence of traps, excitons can diffuse and exciton-polaron quenching is in that case described by a rate constant as:

(40) $k_1\left(t\right)=4{\mathrm{NR}}_{\mathrm{eff}}D\left[1+\frac{R_{\mathrm{eff}}}{{\left(D_{\mathrm{ex}}t\right)}^{1/2}}\right]$
where N is Avogadro's number divided by 1000, D.sub.ex the exciton diffusion constant and

(41) $R_{\mathrm{eff}}=0.676{\left(\frac{R_0^6}{{}_D^{0}D}\right)}^{1/4}$
with R.sub.0 the Frster critical radius and .sub.D.sup.0 the excited state lifetime of the donor in the absence of transfer. The competing (desired) process of radiative emission occurs with a rate k.sub.rad=1/.sub.D.sup.0. In the presence of traps, exciton diffusion is zero and the ratio of the radiative quantum yields in the presence and absence of quenching is given by

(42) $\frac{{}_D}{{}_D^0}=1-\sqrt{}\exp \left({}^2\right)\left[1-\mathrm{erf}\left(\right)\right]$$\mathrm{where}=\frac{\sqrt{}}{2}{C}_A\frac{4}{3}{R}_0^3$
and C.sub.A is the concentration of acceptors expressed in number of molecules per .sup.3. As in the trap-free case, the radiative emission in absence of quenching .sub.D.sup.0 is calculated as .sub.D.sup.0=n.sub.ex/.sub.D.sup.0 with n.sub.ex the exciton concentration. The ratio of the radiative quantum yields was used to determine the rate of polaron quenching:

(43) $k_q=k_{\mathrm{rad}}\left(\frac{\sqrt{}\exp \left({}^2\right)\left[1-\mathrm{erf}\left(\right)\right]}{1-\sqrt{}\exp \left({}^2\right)\left[1-\mathrm{erf}\left(\right)\right]}\right)$
where k.sub.rad=1/.sub.D.sup.0.

(44) Below, the parameters used in the simulations are given. The corresponding experimental values are given in the second column when available. Typically somewhat lower concentrations for ions, traps etc. are used to reflect the fact that in the real device not all salt molecules will dissociate due to the ionic binding, nor that all guest molecules will cause a trap due to e.g. aggregation.

(45) TABLE-US-00001 Simulation Experiment/comment c.sub.ion = 2 c.sub.salt = 0.06 (nm).sup.3 0.18 (nm).sup.3 c.sub.trap = 0.03 (nm).sup.3 0.02.fwdarw.0.18 (nm).sup.3; 0.15 (nm).sup.3 (best experiment) Ion/trap ratio = 2 1.2 c.sub.host = 0.3 (nm).sup.3 1.5 (nm).sup.3 (best experiment) Trap/DOS ratio = 10% 10% (best experiment) E.sub.trap.sup.n = E.sub.trap.sup.p = 0.10 eV 0.3-0.4 eV (from CV). Simulation value set lower to reflect disorder- assisted escape of the electronic charge from its counter ion. Active-material thickness = 130 nm 130 nm LUMO/HOMO = 2.7/5.6 eV LUMO (OXD-7): 2.7 eV, HOMO (PVK): 5.6 eV E.sub.F,contact (LEC) = 4.2/5.0 eV Al cathode: 4.1-4.2 eV, PEDOT:PSS anode: 5.0 Ev E.sub.F,contact (OLED) = 2.9/5.0 eV Ca cathode: 2.9 eV, PEDOT:PSS anode: 5.0 eV .sub.r = 3.6 Free n/p mobility (host, LEC) = 1 10.sup.12 m.sup.2/Vs Free n/p mobility (host, OLED) = 1 10.sup.10 m.sup.2/Vs Ion mobility = 1 10.sup.13 m.sup.2/Vs Irrelevant for steady-state solution; one order slower than electronic mobility selected to be able to observe different time scales in transients Exciton radiative decay rate: 1.0 s.sup.1 D.sub.ex = 4.6 10.sup.11 m.sup.2 s.sup.1 (trap-free) Corresponds to a diffusion length of 6.8 nm Frster critical radius R.sub.0 = 1.5 nm
Identification of Appropriate LEC Materials

(46) With the simulation results at hand, identification of appropriate LEC materials was started.

(47) Materials

(48) The following host materials were elected for the first evaluation: PVK, OXD-7 and PVK:OXD-7. The chemical structure of the host materials poly(9-vinycarbazole) (PVK, Sigma-Aldrich) and 1,3-bis[2-(4-tert-butylphenyl)-1,3,4-oxadiazo-5-yl]benzene (OXD-7, Lumtec) are presented in the inset of FIGS. 2a-b. Other investigated host materials include commercially available TH123 and TH105 (Merck KGaA).

(49) A range of commercially available guest compounds were investigated, including tris[2-(5-substituent-phenyl)-pyridinato]iridium(III) (Ir(R-ppy).sub.3, Merck, see inset in FIG. 2d), tris[2-(p-tolyl)pyridine]iridium(III) (Ir(mppy).sub.3, Lumtec), and tris[2-(4-n-hexylphenyl)quinoline]iridium(III) (Hex-Ir(piq).sub.3, Lumtec); see insets in FIG. 8.

(50) The investigated electrolytes were tetrahexylammonium tetrafluoroborate (THABF.sub.4, Sigma-Aldrich, see inset in FIG. 2e) and LiCF.sub.3SO.sub.3 (Sigma-Aldrich) dissolved in hydroxyl-capped trimethylolpropane ethoxylate (TMPE-OH, M.sub.w=450 g/mol, Sigma-Aldrich). All of the materials were used as received.

(51) Master solutions were prepared by dissolving the constituent material in chlorobenzene at a concentration of 15 mg/ml (PVK), 30 mg/ml (OXD-7), 20 mg/ml (PVK:OXD-7), 20 mg/ml (TH123:TH105), 10 mg/ml (THABF.sub.4), and 10 mg/ml (TMPE-OH:LiCF.sub.3SO.sub.3). The master solutions were stirred on a magnetic hot plate at 343 K for at least 5 h before further processing.

(52) Methods

(53) Cyclic voltammetry (CV) was carried out with an Autolab PGSTAT302 potentiostat driven by the GPES software. The working electrode comprised the material-under-study coated on an Au-covered glass substrate, a Pt rod was the counter electrode, an Ag wire was the quasi-reference electrode, and 0.1 M tetrahexylammonium tetrafluoroborate (THABF.sub.4, Sigma-Aldrich) in anhydrous CH.sub.3CN was the electrolyte. Directly after each CV scan, a calibration scan was run with a small amount of ferrocene/ferrocenium ion (Fc/Fc.sup.+) added to the electrolyte. All CV potentials were reported vs. Fc/Fc.sup.+. The reduction/oxidation onset potentials were defined as the intersection of the baseline with the tangent of the current at the half-peak-height. The energy structure (i.e., the HOMO and LUMO levels) of the material-under-study was derived using the equation E.sub.VL=e.Math.(4.8 V+V.sub.Fc/Fc+). The CV sample preparation and characterization were executed in a N.sub.2-filled glove box ([O.sub.2]<1 ppm, [H.sub.2O]<0.5 ppm).

(54) Absorption (UV-3100 spectrophotometer, Shimadzu) and photoluminescence (PL; FP-6500 spectrofluorometer, JASCO) measurements were carried out on spin-coated thin films on carefully cleaned quartz substrates.

(55) The active-material inks were prepared by blending the host and electrolyte master solutions in a desired mass ratio, and thereafter adding an appropriate amount of the guest compound. Small-area LECs and OLEDs were fabricated by sequentially spin-coating a poly(3,4-ethylenedioxythiophene):poly(styrene sulfonate) (PEDOT:PSS, Clevios P VP Al 4083, Heraeus) ink at 4000 rpm for 60 s and the active-material ink at 2000 rpm for 60 s onto carefully cleaned indium-tin-oxide (ITO) coated glass substrates (20 /square, Thin Film Devices, US). The dry thickness of the PEDOT-PSS and the active material was 40 nm and 130 nm, respectively. For the LEC (OLED), a set of four Al (Ca) cathodes was deposited on top of the active material by thermal evaporation at p<510.sup.4 Pa through a shadow mask. The light-emission area, as defined by the size of one cathode, was 8.51.5 mm.sup.2. All of the above procedures, except for the deposition of the PEDOT:PSS layer, were carried out in two interconnected N.sub.2-filled glove boxes ([O.sub.2]<1 ppm, [H.sub.2O]<0.5 ppm). The LEC and OLED devices were characterized using a computer-controlled source-measure unit (Agilent U2722A) and a calibrated photodiode, equipped with an eye-response filter (Hamamatsu Photonics), connected to a data acquisition card (National Instruments USB-6009) via a current-to-voltage amplifier. The EL spectra were recorded using a calibrated fiber-optic spectrometer (USB2000+, Ocean Optics).

(56) The planar light out-coupling structure comprised hemispherical lenses in a hexagonal pattern on the surface of a 250 m thick poly(methyl methacrylate) (PMMA) film (Microsharp). The radius and the height of each microlens were 35 and 24.5 m, respectively. A UV-curable and single-component acrylic adhesive was used for laminating the outcoupling film onto the LEC device. The hemispherical out-coupling structure comprised a half-sphere lens (d=18 mm, Thorlabs), which was mounted onto the LEC with a specialty oil (Olympus). Both light out-coupling structures featured a refractive index of n=1.5, which matched the glass substrate onto which they were mounted. The light-emission area of the out-coupled LECs was 1.51.5 mm.sup.2, and was defined by etching of the ITO anode.

(57) Large-area LECs were fabricated by spray-sintering under ambient air. The original active-material ink was diluted with 80% THF, and spray-sintered onto a pre-patterned ITO-coated glass substrate (Thin Film Devices, US) maintained at 70 C. by a hotplate. The spray-sintering deposition was executed using an in-house developed, computer-controlled spray box (LunaLEC AB, Sweden), equipped with an internal-mix spray nozzle. The N.sub.2 gas pressure was set to 45010.sup.3 Pa, and the ink-feeding rate was 1 ml/min. The spray nozzle was programmed to move back-and-forth over a 1010 cm.sup.2 area in a raster-like motion, at a height of 6 cm above the substrate, and to stop after 8 completed sweeps (t=190 s). The resulting dry active-material film thickness was 350 nm. The Al top electrode was deposited by thermal evaporation through a shadow mask, defining the 6767 mm.sup.2 emission area. The luminance was measured with a luminance meter (LS-110, Konica), and the presented luminance is the average from >6 measurements on different spatial locations over the device area. The photograph was recorded by a digital camera (Canon EOS 300D) at an exposure time of 1/120 s and an aperture of f/2.2.

(58) Materials Characterization

(59) First, it is fundamental that the target LEC comprises host and guest compounds in the active material that can be electrochemically doped, while the electrolyte should be electrochemically inert within the voltage range spanned by the electrochemical doping reactions of the host.

(60) In order to investigate whether the host:guest:electrolyte active material fulfills these requirements, a systematic cyclic voltammetry (CV) study was carried out.

(61) FIGS. 2a-e present systematic cyclic voltammetry (CV) interrogations of a number of commercially available candidate compounds, with their corresponding chemical structure displayed in the inset. FIGS. 2a-2d show CV traces recorded on thin films of different host compounds (PVK, OXD-7 and PVK:OXD-7) and a guest compound (Ir(R-ppy).sub.3). In FIG. 2e a CV for the electrolytes THABF.sub.4 (solid line) and TMPE-OH:LiCF.sub.3SO.sub.3 (dashed line) is shown.

(62) It was found that the PVK host only features p-type doping capacity, as implied by the lack of a reversible reduction event; this conclusion is in agreement with direct optical probing of planar PVK-based LECs. The OXD-7 host and the PVK:OXD-7 blend host can, in contrast, be both electrochemically p-type and n-type doped, and are thus qualified as appropriate LEC host materials. A comparison of the top three CV traces reveals that for the blend host it is PVK that is (preferentially) p-type doped and OXD-7 that is n-type doped.

(63) FIG. 2d shows that the Ir(R-ppy).sub.3 guest compound displays significant p- and n-type doping capacity, and that both its doping reactions are thermodynamically preferable in comparison to all three hosts.

(64) FIG. 3 shows that the CV trace of a Ir(R-ppy).sub.3 thin film is similar to that of a PVK:OXD-7:Ir(R-ppy).sub.3 film, but distinctly different to that of the same film void of the Ir(R-ppy).sub.3 compound. The conclusion is thus that Ir(R-ppy).sub.3 compound can be both p- and n-type doped in the PVK:OXD-7:Ir(R-ppy).sub.3 active-material film.

(65) FIG. 3 also reveals that it is primarily the Ir(R-ppy).sub.3 guest that is p- and n-type doped in a PVK:OXD-7:Ir(R-ppy).sub.3 blend film.

(66) It was further found that the THABF.sub.4 ionic liquid (FIG. 2e, solid line) is fit for the task of electrolyte, since it is electrochemically silent over the entire voltage range spanned by the p- and n-type doping potentials of the host and guest compounds. This is in contrast to, e.g., TMPE-OH:LiCF.sub.3SO.sub.3, the electrolyte of choice in recent state-of-the art CP-LECs, which electrochemical stability is insufficient (FIG. 2e, dashed line). This finding is confirmed in FIGS. 4a-e, which demonstrates that it is possible to n-type dope OXD-7 and p-type dope PVK with the THABF.sub.4 electrolyte, but not with the TMPE-OH:LiCF.sub.3SO.sub.3 electrolyte.

(67) FIG. 4a displays the electron-energy diagram for the host and electrolyte compounds. It is clear that the THABF.sub.4 ionic liquid displays a broad electrochemical stability window that encompasses the p- and n-type doping potentials of the PVK and OXD-7 host compounds. In contrast, the electrochemical stability window of the TMPE-OH:LiCF.sub.3SO.sub.3 electrolyte is much more narrow, and not encompassing the doping potentials of the host compounds. These findings are confirmed by running CV experiments on the host compounds using either THABF.sub.4 or TMPE-OH:LiCF.sub.3SO.sub.3 as the supporting electrolyte. With THABF.sub.4 as the doping electrolyte, it is possible to electrochemically n-type dope OXD-7 (FIG. 4b) and p-type dope PVK (FIG. 4d), while the lower onset potential and the absence of reversibility when using TMPE-OH:LiCF.sub.3SO.sub.3 as the electrolyte (see FIGS. 4c and 4e) is taken as evidence for that it is the electrolyte and not the host compound that is partaking in the electrochemical reactions.

(68) FIG. 2f presents the electron-energy diagram for the host and guest compounds, as extracted from the CV data (with the exception of the LUMO of PVK that was derived from a combined absorption/CV measurement). It clearly illustrates that the exciton, as well as both the hole and electron, can be trapped on the Ir(R-ppy).sub.3 guest when positioned in any of the three host matrices, i.e. PVK, OXD-7 and the PVK:OXD-7 blend. The triplet diagram in FIG. 2g shows that host-to-guest Dexter transfer is feasible for all host systems, while the significant spectral overlap of the host photoluminescence (PL) and the guest absorption displayed in FIG. 2h implies that host-to-guest Frster resonance transfer can be efficient for all three host-guest systems. The conclusion is thus that the exciton can be efficiently funneled to the dispersed guest molecules using a combination of Frster transfer, Dexter transfer and charge trapping. After arriving at a guest molecule the exciton will not move any further, provided the guest concentration is sufficiently low to suppress exciton diffusion between different guest molecules.

(69) Device Characterization

(70) The characterized materials were used for the fabrication of LEC devices, comprising an indium-tin-oxide (ITO)/poly(3,4-ethylenedioxythiophene):poly(styrene sulfonate) (PEDOT:PSS) anode and an Al cathode.

(71) FIGS. 5a-5d show temporal optoelectronic response of ITO/PEDOT:PSS/Host:Ir(R-ppy).sub.3:Electrolyte/Al LECs with the host and electrolyte selection defined in the insets with the same active-material mass stoichiometry (host:guest:electrolyte=64.6:29.0:6.4) and thickness (d=130 nm) and which were driven by the same current density, j=7.7 mA/cm.sup.2. The distinguishing characteristic was the selection of the host and electrolyte, as identified in the figure insets.

(72) FIG. 6a shows that all four devices emitted green light (.sub.peak=520 nm) from the guest compound, indicating complete energy transfer from host to guest (see FIG. 2f-h).

(73) The PVK:OXD-7 blend-host device in FIG. 5a exhibited a record-high performance for an LEC, delivering a strong luminance of 3,250 cd/m.sup.2 at a current efficacy of 42.2 cd/A; the latter corresponds to an external quantum efficiency (EQE)=11.7%. (Efficiency values as presented herein are reported at maximum luminance.) The pristine blend-host device turned on fast and reached 100 cd/m.sup.2 in less than 2 s and 3,000 cd/m.sup.2 within 25 s, and maintained a high luminance of above 1,000 cd/m.sup.2 for more than 3 h of continuous operation.

(74) A summary of relevant device and material data is available in the table in FIG. 7.

(75) The PVK-host device in FIG. 5b delivered a peak brightness of 575 cd/m.sup.2 at an efficiency of 7.5 cd/A, and the luminance began to drop already after 10 s of operation. Its comparatively poor performance may be attributed to that the PVK host only can support p-type doping (FIG. 2a), and that electrochemical side reactions accordingly will take place at the cathodic interface during device operation. The even lower performance of the blend-host device with TMPE-OH:LiCF.sub.3SO.sub.3 as the electrolyte instead of THABF.sub.4 (FIG. 5d) is attributed to the insufficient electrochemical stability window of the electrolyte, which will effectively prohibit both p- and n-type doping of the host compounds (FIG. 2e and FIGS. 4a-e). The absence of conductivity-enhancing electrochemical doping reactions of the host compounds is further manifested in that the luminance decreases monotonically with increasing operational time.

(76) The explanation for the one order of magnitude lower efficiency for the OXD-7 device in FIG. 5c in comparison to the blend-host device in FIG. 5a is at first glance less obvious. The OXD-7 host compound can be both p- and n-type doped (FIG. 2b) and the employed THABF.sub.4 electrolyte is electrochemically stable (FIG. 2e). However, with the results of the simulation study at hand (FIGS. 1c-f), the attention was directed to other material properties, in particular trap depths and charge carrier mobilities.

(77) The nominal electron-trap depth (E.sub.trap.sup.n) is equal to the difference between the (lowest) LUMO level of the host(s) and the LUMO of the guest, and the nominal hole-trap depth (E.sub.trap.sup.p) is given by the difference between the (highest) HOMO of the host(s) and the HOMO of the guest (see FIG. 1a). Using the electron-energy diagram in FIG. 2f, it was found that for the blend-host LEC E.sub.trap.sup.n=0.32 eV and E.sub.trap.sup.p=0.38 eV; for the single-host OXD-7 LEC we obtain E.sub.trap.sup.n=0.32 eV and E.sub.trap.sup.p=1.06 eV. In other words, while the trap depths for electron and hole transport in the blend-host LEC are rather symmetric, they are notably asymmetric in the OXD-7 LEC. Moreover, the PVK:OXD-7 blend-host features a balanced electron and hole mobility, while the single-host OXD-7 material is an n-type material with a distinctly higher electron mobility. The simulation results demonstrated that an asymmetry in the trap depth (FIG. 1e), as well as an asymmetry in the electron and hole mobility (FIG. 1f), will result in an increase in detrimental exciton-polaron interactions, with a concomitant significant drop in the efficiency. Thus, the conclusion from both experiments and simulations is that it is paramount to pay attention to the symmetry of the trap-depth and mobility parameters when designing a high-performance host-guest LEC.

(78) A range of different host-guest combinations have been investigated, and three additional host-guest LECs have been identified which feature high efficiency at strong luminance; see FIGS. 8a-c. With a green-emitting tris[2-(p-tolyl)pyridine]iridium (III) (Ir(mppy).sub.3) guest, and the same PVK:OXD-7 blend host and THABF.sub.4 electrolyte, a current efficacy of 24.6 cd/A (EQE=6.9%) at 1,890 cd/m.sup.2 was recorded, and with a red-emitting tris[2-(4-n-hexylphenyl)quinoline]iridium (III) (hex-Ir(piq).sub.3) guest an efficiency of 3.9 cd/A (EQE=4.9%) at 302 cd/m.sup.2 was measured. With a different TH123:TH105 host blend, and the Ir(R-ppy).sub.3 guest, green emission with an efficiency of 27.4 cd/A (EQE=7.6%) at 2,100 cd/m.sup.2 was detected. It is interesting that these three high-performing host-guest LECs also feature bipolar doping, balanced mobility, and reasonably balanced trap depths. A comparison does however reveal that the PVK:OXD-7:Ir(R-ppy).sub.3 LEC in FIG. 5a displayed the best performance (and the most symmetric trap depths), and in the following we have therefore opted to stay with this system. A summary of relevant device and material data is available in the table in FIG. 7.

(79) A further demonstration of the suppressed exciton-polaron quenching in a well-designed host-guest LEC is provided by the efficiency-current plot in FIG. 9a. If such detrimental interactions are prominent, these will be manifested in a significant drop of the current efficacy with increasing current. Here, it was found that the current efficacy dropped marginally from 50.4 cd/A (EQE=14.0%) to 42.2 cd/A (EQE=11.7%) when the current density increased by a factor of twenty from 0.38 to 7.7 mA/cm.sup.2, which implies that polaron-exciton quenching indeed is minor. Note that the highest current efficacy of 50.4 cd/A was recorded at reasonable luminance of 193 cd/m.sup.2. It was further observed that the power conversion efficacy (PCE) features a stronger dependence on the drive current than the current efficacy, which is a reflection of a lowering of the drive voltage with decreasing current.

(80) FIG. 9b presents the efficiency as a function of guest concentration for the host-guest LEC and a nominally similar OLED, void of the electrolyte and with low-work function Ca instead of Al as the cathode. The schematic LEC and OLED device structures are displayed in the insets of FIGS. 9a and 9b, respectively. Both devices display green emission from the guest compound for all concentrations (FIG. 6b), in agreement with the demonstrated efficient host-guest energy transfer. Interestingly, the LEC outperforms the OLED over the entire guest-concentration range, with the difference being largest at the higher guest concentrations. The peak efficiency of the LEC is 42.2 cd/A (11.7%), whereas the OLED peaked at 21.7 cd/A (6.0%) at a three times lower guest concentration.

(81) To investigate the cause for the differences in performance, modeling results for a host-guest OLED with device-realistic injection barriers for electrons and holes of 0.2 eV and 0.5 eV, respectively, are presented (FIG. 9c), and for a host-guest LEC with a larger injection barrier of 1.0 eV for both electron and hole injection (FIG. 9d). Importantly, the opportunity for operating an LEC with large injection barriers is very attractive since it allows for the use of air-stabile electrodes and a single-layer active material. This clearly distinguishes it from the OLED, which is dependent on either an air-unstable cathode (here Ca) or a multilayer structure in order to function efficiently. Both the OLED and LEC featured symmetrical trap depths and mobilities, but in order to make the current density similar the absolute mobility was set higher for the OLED, which is reasonable in view of the absence of the electrolyte in the OLED. The absence of mobile ions in the OLED is manifested in much lower electron and hole concentrations (since there is no ion-induced screening of the electronic space charge) and a lack of electric double-layer formation at the electrode interfaces. Two more indirect consequences are that the excitons are evenly distributed, and the voltage gradient relatively constant, throughout the entire OLED active material, whereas they are well confined to the central p-n junction in the LEC device. While the simulated OLED in FIG. 9c delivers a modest electron-to-photon efficiency of =8%, the LEC in FIG. 9d features a much higher =51% at approximately the same current density (j.sub.OLED=3.4 mA/cm.sup.2, j.sub.LEC=4.5 mA/cm.sup.2).

(82) A significant fraction of the photons are trapped within the planar device structure by total internal reflection, and in order to access these photons two different outcoupling structures were attached onto the transparent substrate: (i) a flexible thin film comprising a hexagonal array of hemispherical microlenses as the surface structure (MLA-LEC), and (ii) an index-matched glass half-sphere (GHS-LEC). The transient optoelectronic response is presented in FIG. 9e, and with the MLA-LEC a current efficacy of 66.8 cd/A at 1,285 cd/m.sup.2 was measured and with the GHS-LEC, 99.2 cd/A at 1,910 cd/m.sup.2 was reached. The latter corresponds to an impressive external quantum efficiency of 27.5%. For the same device a stress lifetime of 320 h above 100 cd/m.sup.2 was measured in FIG. 10a. The corresponding data at a higher current density of j=7.7 mA/cm.sup.2 are presented in FIG. 10b, and the GHS-LEC delivered a strong luminance of 7,200 cd/m.sup.2 at a retained high efficiency of 92.8 cd/A.

(83) All devices up to this stage have featured a small light-emission area of <0.2 cm.sup.2 as fabricated by spin-coating, but also large emission-area LECs were fabricated by more scalable spray-coating (or more specifically spray-sintering) using an in-house developed automated spray-coating apparatus. The photograph in FIG. 9f shows the uniform green light-emission from such a 45 cm.sup.2 large-area host-guest LEC, with its 350 nm thick active material deposited by spray-sintering under ambient air, as displayed in the inset photograph. The large-area LEC device was driven with a constant current density of j=1.1 mA/cm.sup.2, and delivered a luminance of 150 cd/m.sup.2 at a current efficacy of 13 cd/A, despite being void of an outcoupling structure. This result indicates that the introduced approach to high-efficiency, high-luminance LEC operation is relevant also for practical low-cost and/or large-area device configurations, and as such could pave the way for a wide range of novel and important applications in, for instance, home health care and portable signage where high-brightness operation at low energy consumption and low cost can be a fundamental criterion.