LED display device, method of controlling the same, and method of manufacturing an LED display device

Abstract

A display device comprises a light emitting diode (LED) which includes a porous semiconductor material, wherein the device comprises a pixel comprising a plurality of subpixels each having a light-emitting layer. A first subpixel has a first light-emitting layer having a first area A1, and a second subpixel has a second light-emitting layer having a second area A2 different from the first area A1. The first subpixel is configured to emit at a first peak wavelength, and the second subpixel is configured to emit at a second peak wavelength different from the first peak wavelength. A method of controlling this display device and a method of manufacturing said display device are also provided.

Claims

1. A display device comprising a light emitting diode (LED) which includes a porous semiconductor material, wherein the device comprises a pixel comprising a plurality of subpixels each having a light-emitting layer; wherein a first subpixel has a first light-emitting layer having a first area A.sub.1, and a second subpixel has a second light-emitting layer having a second area A.sub.2 different from the first area A.sub.1; wherein the first and second subpixels have identical LED epitaxial structures and identical layer compositions; wherein the first subpixel and the second subpixel are variable-wavelength subpixels, and in which the display device is configured to provide a variable drive current to the variable-wavelength subpixels, such that each variable-wavelength subpixel emits at different peak emission wavelengths in response to different drive currents; wherein the first subpixel is a variable-wavelength subpixel configured to emit at a first peak emission wavelength in response to a first drive current applied to the first subpixel, and to emit at a third peak emission wavelength in response to a third drive current applied to the first subpixel; wherein the second subpixel is a variable-wavelength subpixel configured to emit at a second peak emission wavelength different from the first peak wavelength in response to a second drive current applied to the second subpixel, and to emit at a fourth peak emission wavelength in response to a fourth drive current applied to the second subpixel.

2. The display device according to claim 1, in which the first subpixel has a first geometry or shape, and in which the second subpixel has a second geometry or shape, in which the first geometry or shape is different from the second geometry or shape.

3. The display device according to claim 2, in which the first subpixel is circular in shape, and in which the second subpixel is formed as a ring arranged concentrically around the circular first subpixel.

4. The display device according to claim 1, in which the first subpixel and the second subpixel are positioned on a shared n-type conductive layer of semiconductor material.

5. The display device according to claim 1, comprising driver circuitry configured to control the drive current provided to each subpixel in the display device.

6. The display device according to claim 1, in which the first subpixel is configured to emit at a first emission intensity at the first peak emission wavelength in response to the first drive current applied to the first subpixel, and/or in which the second subpixel is configured to emit at a second emission intensity at the second peak emission wavelength in response to the second drive current applied to the second subpixel.

7. The display device according to claim 1, in which the first subpixel is configured to emit at a first luminosity at the first peak emission wavelength in response to the first drive current applied to the first subpixel, and/or in which the second subpixel is configured to emit at a second luminosity at the second peak emission wavelength in response to the second drive current applied to the second subpixel.

8. A method of controlling the display device according to claim 1, comprising the steps of: providing a variable-magnitude drive current to the first subpixel, and providing a variable-magnitude drive current to the second subpixel; wherein providing the variable-magnitude drive current to the first subpixel comprises providing the first drive current to the first subpixel such that the first subpixel emits at the first peak emission wavelength, and providing the third drive current to the first subpixel such that the first subpixel emits at the third peak emission wavelength; wherein providing the variable-magnitude drive current to the second subpixel comprises providing the second drive current to the second subpixel such that the second subpixel emits at the second peak emission wavelength different from the first peak emission wavelength, and providing the fourth drive current to the second subpixel such that the second subpixel emits at the fourth peak emission wavelength.

9. The display device according to claim 1, in which the first area A.sub.1 of the first light-emitting layer is larger than the second area A.sub.2 of the second light-emitting area, or in which the first area A.sub.1 of the first light-emitting layer is smaller than the second area A.sub.2 of the second light-emitting area.

10. The display device according to claim 1, in which the first area A.sub.1 of the first light-emitting layer has a different shape than the second area A.sub.2 of the second light-emitting area.

11. The display device according to claim 1, in which the first area A.sub.1 and the second area A.sub.2 are footprints of the first and second light-emitting layers over the porous semiconductor material.

12. The method according to claim 8, in which the magnitude of the first drive current is the same as the magnitude of the second drive current.

13. The display device according to claim 1, comprising a first electrical contact which contacts the first subpixel over a first contact area, the first electrical contact being configured to apply a drive current to the first subpixel, and a second electrical contact which contacts the second subpixel over a second contact area, the second electrical contact being configured to apply a drive current to the second subpixel, in which the first subpixel has a first contact ratio defined by the ratio of the first contact area:first area A.sub.1, and the second subpixel has a second contact ratio defined by the ratio of the second contact area:second area A.sub.2, and in which the first contact ratio is different from the second contact ratio, so that the first and second subpixels are configured to emit at different peak emission wavelengths in response to the same drive current.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) Embodiments of the invention will now be described with reference to the figures, in which:

(2) FIG. 1 illustrates a porous template suitable for an LED according to the present invention;

(3) FIGS. 2-18 illustrate the steps of manufacturing an LED according to a preferred embodiment of the present invention;

(4) FIG. 19 is a graph of normalised electroluminescence (EL) intensity vs wavelength, for an InGaN LED over a porous region;

(5) FIG. 20 is a graph of normalised electroluminescence (EL) intensity vs wavelength at different current injections, for an InGaN LED on a non-porous substrate;

(6) FIG. 21 is a graph of normalised electroluminescence (EL) intensity vs wavelength at different current injections, for the same InGaN LED as FIG. 20 grown over a porous region;

(7) FIG. 22 is an I-V curve measured for InGaN micro-LEDs of different pixel sizes on a non-porous substrate, with the inset image showing yellow emission;

(8) FIG. 23 is an I-V curve measured for InGaN micro-LEDs of different pixel sizes on a porous substrate, with the inset image showing red emission;

(9) FIG. 24 is an I-V curve measured for InGaN micro-LEDs of different pixel sizes on a template with a sub-surface porous region;

(10) FIG. 25 is a low-current I-V curve measured for InGaN micro-LEDs of different pixel sizes on a template with a sub-surface porous region;

(11) FIG. 26 is a series of five EL images of the same MicroLED pixel being driven at different currents in constant wave mode (CW), showing five different colours of emission;

(12) FIG. 27A is an emission wavelength vs current density plot for a 25 m25 m 100100 variable-wavelength LED pixel array driven in pulsed mode with a 100 s pulse at 1% duty cycle;

(13) FIG. 27B is an emission wavelength vs current density plot for a 30 m30 m 100100 variable-wavelength LED pixel array driven in pulsed mode with a 100 s pulse at 1% duty cycle;

(14) FIG. 28 is a plot of intensity vs wavelength for a single variable-wavelength LED driven at different currents in pulsed driving mode with a 100 s pulse at 1% duty cycle;

(15) FIG. 29A is a schematic plot of an LED structure comprising a current constraining layer, according to a preferred embodiment of the present invention;

(16) FIG. 29B is a schematic plot of an LED structure comprising a current constraining layer, according to another preferred embodiment of the present invention;

(17) FIG. 30 is a schematic cross-section of an LED according to a preferred embodiment of the present invention;

(18) FIGS. 31-34 are schematic cross-sections of the LED of FIG. 30, including two v-shaped pits;

(19) FIG. 35 is a TEM image of a cross-section of an LED with a v-shaped pit, according to a preferred embodiment of the present invention;

(20) FIG. 36 is a graph of normalised intensity vs peak emission wavelength, for a variable-wavelength LED according to the present invention;

(21) FIG. 37 is a graph of peak emission wavelength vs driving current, for a variable-wavelength LED according to the present invention; and

(22) FIG. 38 is a graph of peak emission wavelength for variable-wavelength LEDs of different pixel sizes;

(23) FIG. 39 is a schematic side-on cross-section of a variable-wavelength LED structure according to a preferred embodiment of the present invention;

(24) FIG. 40 is a schematic side-on cross-section of a variable-wavelength LED structure according to another preferred embodiment of the present invention;

(25) FIG. 41 is a schematic illustrations of the contact area between a contact layer and a semiconductor layer in an LED diode;

(26) FIG. 42 is a schematic plan view of an electrical contact positioned over an LED subpixel mesa;

(27) FIG. 43A is a schematic illustration of three subpixel mesas having the same subpixel mesa area, but different contact areas and thus different contact ratios;

(28) FIG. 43B is a schematic illustration of three subpixel mesas having the same subpixel mesa area and the same contact pad size, but different contact areas and thus different contact ratios;

(29) FIG. 43C is a schematic illustration of three subpixel mesas having different subpixel mesa areas, different contact areas and different contact ratios;

(30) FIG. 44 is a schematic side-on cross-section of three subpixels on a shared semiconductor template, having different contact areas and thus different contact ratios;

(31) FIG. 45 is a schematic side-on cross-section of three subpixels on a shared semiconductor template, having different contact areas and thus different contact ratios.

(32) FIGS. 46A-G illustrate alternative embodiments of non-uniform, fragmented or discontinuous light-emitting regions of variable-wavelength LEDs according to the present invention;

(33) FIG. 47A is a TEM image of a cross-section of a conventional non-variable-wavelength LED;

(34) FIG. 47B is a TEM image of the light-emitting region of a variable-wavelength LED comprising v-shaped pits, according to an embodiment of the present invention;

(35) FIG. 47C is a TEM image of the variable-wavelength LED of FIG. 47B, showing a porous region and a light-emitting region comprising a plurality of v-shaped pits, according to a preferred embodiment of the present invention;

(36) FIG. 48A is a graph of peak emission wavelength vs driving current density for a conventional non-variable wavelength LED;

(37) FIG. 48B is a graph of peak emission wavelength vs driving current density for a variable-wavelength LED according to an embodiment of the present invention;

(38) FIG. 48C is a graph of peak emission wavelength vs driving current density for a variable-wavelength LED according to another embodiment of the present invention;

(39) FIG. 49A is a graph of peak emission wavelength vs driving current density for another variable-wavelength LED according to the present invention;

(40) FIGS. 49B-E are photographs of the variable-wavelength LED of FIG. 49A, with inset emission spectra showing the different peak emission wavelengths at different driving current densities;

(41) FIG. 50A is a schematic illustration of the spatial combination of single-colour subpixels in a conventional display pixel;

(42) FIGS. 50B-50G are schematic illustrations of LED display device pixels containing at least one variable-wavelength LED according to the present invention;

(43) FIG. 51A-C illustrate colour gamuts obtainable with prior art displays and the devices of the present invention;

(44) FIGS. 52A-D illustrate a display pixel according to an embodiment of the present invention and the colour gamuts obtainable with that pixel;

(45) FIG. 53 illustrates a method of controlling a display pixel containing variable-wavelength LEDs according to the present invention;

(46) FIGS. 54A & 54B illustrate a display pixel according to an embodiment of the present invention and the colour gamut obtainable with that pixel;

(47) FIGS. 55A-D illustrate a display pixel according to another embodiment of the present invention and the colour gamut obtainable with that pixel;

(48) FIG. 56 illustrates a display pixel according to another embodiment of the present invention and the colour gamut obtainable with that pixel;

(49) FIG. 57 illustrates a display pixel according to another embodiment of the present invention and the colour gamut obtainable with that pixel;

(50) FIG. 58 illustrates a display pixel according to another embodiment of the present invention and the colour gamut obtainable with that pixel;

(51) FIG. 59A illustrates a variable-wavelength display pixel according to an embodiment of the present invention, and FIGS. 59B-60D illustrate driving schemes for controlling the pixel of FIG. 59A;

(52) FIG. 61A illustrates a display pixel according to an embodiment of the present invention, and FIGS. 61B-61C illustrate driving schemes for controlling the pixel of FIG. 61A;

(53) FIG. 62 illustrates a display pixel according to an embodiment of the present invention, and driving schemes for controlling the pixel;

(54) FIG. 63 illustrates a display pixel according to an embodiment of the present invention, and driving schemes for controlling the pixel;

(55) FIG. 64A is a schematic illustration of driving current conditions for a variable-wavelength LED according to the present invention

(56) FIG. 64B illustrates the peak emission wavelength emitted by a variable-wavelength LED according to the present invention in response to different drive currents;

(57) FIG. 65 is a schematic illustration of driving current conditions for a variable-wavelength LED according to Embodiment 1 of the present invention;

(58) FIG. 66 is a schematic illustration of driving current conditions for a variable-wavelength LED according to Embodiment 2 of the present invention;

(59) FIG. 67 is a schematic illustration of driving current conditions for a variable-wavelength LED according to Embodiment 3 of the present invention;

(60) FIG. 68 is a schematic illustration of driving current conditions for a variable-wavelength LED according to Embodiment 4 of the present invention;

(61) FIG. 69 is a schematic illustration of driving current conditions for a variable-wavelength LED according to Embodiment 5 of the present invention;

(62) FIG. 70 illustrates the peak emission wavelength emitted by a variable-wavelength LED according to the present invention, overlaid with five different drive currents;

(63) FIG. 71 illustrates a pattern of digital pulses of five different driving currents;

(64) FIG. 72 illustrates the resulting emission spectra emitted by a variable-wavelength LED in response to the pulse pattern of FIG. 71;

(65) FIGS. 73A-73D illustrates spectral reconstruction using a plurality of digital pulses of driving current according to Embodiment A of the present invention;

(66) FIG. 74 illustrates an exemplary target emission spectrum;

(67) FIG. 75 illustrates the peak emission wavelength emitted by a variable-wavelength LED according to the present invention in response to different drive currents;

(68) FIG. 76 illustrates an exemplary analogue driving current pulse usable in Embodiment B of the present invention;

(69) FIG. 77 illustrates a perceived output spectrum produced by a variable-wavelength LED in response to the analogue current pulse of FIG. 63;

(70) FIG. 78 illustrates a large area spectrally tuneable illumination source according to a preferred Embodiment C of the present invention;

(71) FIG. 79 illustrates a spectrally correct display device comprising spectrally tuneable pixels according to a preferred Embodiment D of the present invention;

(72) FIG. 80A is a schematic plan view of a display pixel having two variable-wavelength subpixels with the same diode structure but different light-emitting areas;

(73) FIG. 80B shows an array of the red-green (RG) pixel packages of FIG. 80A, mounted on a back plane driver integrated circuit with an array of blue LED subpixels;

(74) FIG. 81 is a graph of drive current density vs emission wavelength for a variable-wavelength LED subpixel according to an aspect of the present invention;

(75) FIGS. 82A and 82B illustrate an exemplary device pixel comprising two subpixels with the same diode structure but different areas;

(76) FIG. 83 illustrates the relationship between emission efficiency and emission wavelength for an exemplary variable-wavelength LED subpixel according to an aspect of the present invention;

(77) FIG. 84 illustrates the photopic luminosity function for emission at different wavelengths for an exemplary variable-wavelength LED subpixel according to an aspect of the present invention;

(78) FIG. 85A is a schematic plan view of a two-subpixel device pixel according to a preferred embodiment of the present invention; and

(79) FIG. 85B is a schematic side-on cross-sectional view of the pixel of FIG. 85A, taken along the line A-A.

(80) FIG. 1 illustrates a porous template suitable for an LED according to the present invention.

(81) The porous template comprises a porous region of III-nitride material on a substrate, with a non-porous layer of III-nitride material arranged over the top surface of the porous region. Optionally there may be further layers of III-nitride material between the substrate and the porous region.

(82) As described in more detail below, the porous region may be provided by epitaxially growing an n-doped region of III-nitride material and then an undoped layer of III-nitride material, and porosifying the n-doped region using the porosification process as set out in international patent applications PCT/GB2017/052895 (published as WO2019/063957) and PCT/GB2019/050213 (published as WO2019/145728).

(83) As described above, this porosification leads to strain relaxation in the crystal lattice, which means that subsequent overgrowth of further semiconductor layers benefit from reduced compressive strain in their lattices.

(84) The porous region may comprise one or more layers one or more III-nitride materials, and may have a range of thicknesses, all while still providing the strain relaxation benefit that shifts the wavelength of InGaN light emitting layers overgrown above the porous region. In preferred embodiments, the porous region may for example comprise GaN and/or InGaN.

(85) A variety of LED structures may be overgrown over the template illustrated in FIG. 1.

(86) In particular, LED structures containing InGaN light emitting layers, which are known in the art for the manufacture of yellow or green LEDs, may be overgrown on the porous template using standard LED manufacturing steps. When grown on the porous template, however, a LED structure which normally emits at a first wavelength, will emit at a red-shifted longer wavelength.

(87) In this way, the use of a porous region of III-nitride material as a template or pseudo-substrate for overgrowth of known InGaN LED structures allows longer-wavelength LEDs to be manufactured in a straightforward manner.

(88) In a preferred embodiment, a LED according to the present invention comprises the following layers, and may be manufactured using the step by step process described below.

(89) The following description of the LED structure relates to a Top emission architecture being described from the bottom up, but the invention is equally applicable to a bottom emission architecture.

(90) FIG. 2Substrate & III-Nitride Layer for Porosification

(91) A compatible substrate is used as a starting surface for epitaxy growth. The substrate may be Silicon, Sapphire, SiC, -Ga2O3, GaN, glass or metal. The crystal orientation of the substrates can be polar, semi-polar or non-polar orientation. The substrate size may vary from 1 cm.sup.2, 2 inch, 4 inch, 6 inch, 8 inch, 12 inch, 16 inch diameters and beyond, and the substrate may have a thickness of greater than 1 m, for example between 1 m and 15000 m.

(92) A layer or stack of layers of III-nitride material is epitaxially grown on the substrate. The III-nitride layer may contain one or a combination of these elements: Al, Ga, In (binary, ternary or quaternary layer).

(93) The thickness T of the III-nitride stack is preferably at least 10 nm, or at least 50 nm, or at least 100 nm, for example between 10-10000 nm.

(94) The III-nitride layer comprises a doped region having an n-type doping concentration between 110.sup.17 cm.sup.3-510.sup.20 cm.sup.3. The III-nitride layer may also comprise an undoped cap layer of III-nitride material over the doped region.

(95) The doped region may terminate at the exposed upper surface of the III-nitride layer, in which case the surface of the layer will be porosified during electrochemical etching.

(96) Alternatively, the doped region of the III-nitride material may be covered by an undoped cap layer of III-nitride material, so that the doped region is sub-surface in the semiconductor structure. The sub-surface starting depth (d) of the doped region may be between 1-2000 nm for example.

(97) FIG. 3Porosification to Porous Region

(98) After it is deposited on the substrate, the III-nitride layer (or stack of layers) is porosified with a wafer scale porosification process as set out in international patent applications PCT/GB2017/052895 (published as WO2019/063957) and PCT/GB2019/050213 (published as WO2019/145728). During this process, the doped region of the III-nitride material becomes porous, while any undoped region of III-nitride material does not become porous.

(99) Following the porosification step, the structure therefore contains a porous region which remains where there was previously n-doped III-nitride material, and optionally a non-porous intermediate layer overlying the porous region.

(100) The degree of porosity of the porous region is controlled by the electrochemical etching process and may be between 1%-99% porosity, preferably between 20% to 90% porosity or between 30%-80%, though lesser or greater porosities could also be employed.

(101) The thickness of the porous region following porosification is preferably greater than 1 nm, more preferably greater than 10 nm, particularly preferably at least 40 nm or 50 nm or 100 nm. However, the thickness of material required to obtain the strain relaxation benefit provided by the porous region may vary depending on the type of III-nitride material from which the porous region is made.

(102) The porous region created by the porosification process may be a bulk layer of a III-nitride material having a uniform composition and a uniform porosity throughout the layer. Alternatively the porous region may comprise multiple layers of porous material of different compositions and/or porosities, forming a porous stack of III-nitride material. For example the porous region may be a continuous layer of porous GaN, or a continuous layer of porous InGaN, or a stack comprising one or more layers of porous GaN and/or one or more layers of porous InGaN. The inventors have found that the strain relaxation benefit of the porous region for overgrowth is obtainable across a wide range of porous regions having different thicknesses, compositions, and layered stacks.

(103) In the embodiment illustrated in the Figures, the porous region is a single porous layer.

(104) Where there is an undoped cap layer of III-nitride material over the doped region, the undoped region remains non-porous following through-surface porosification of the doped region below. The thickness D of this non-porous cap layer may preferably be at least 2 nm, or at least 5 nm, or at least 10 nm, preferably 5-3000 nm. Providing an undoped cap layer over the doped region advantageously leads to a non-porous layer of III-nitride material covering the porous region following porosification. This non-porous cap layer may advantageously allow better overgrowth of further material above the porous region.

(105) As the porosification method of PCT/GB2017/052895 (published as WO2019/063957) and PCT/GB2019/050213 (published as WO2019/145728) can be carried out on entire semiconductor wafers, no processing/patterning/treatment is needed to prepare the template for porosification.

(106) FIG. 4Connecting Layer

(107) After formation of the porous layer, a III-nitride LED epitaxy structure can be grown onto the porous template/pseudo-substrate provided by the porous layer and the non-porous cap layer.

(108) The first layer for growth of the LED structure onto the template may be termed a connecting layer 1.

(109) Although it is possible for an LED epitaxial structure to be grown directly onto the non-porous cap layer, it is preferable that a connecting layer 1 is provided over the cap layer before overgrowth of the LED structure. The inventors have found that the use of a III-nitride connecting layer 1 between the porous region and the LED epitaxy structure may advantageously ensure a good epitaxial relationship between the LED and the porous template/substrate. The growth of this layer makes sure that subsequent overgrowth on top of the connecting layer is smooth and epitaxial and suitably high quality.

(110) The connecting layer 1 is formed of III-nitride material and may contain one or a combination of these elements: Al, Ga, In (binary, ternary or quaternary layer).

(111) The connecting layer can be a doped or un-doped layer. The connecting layer can optionally be doped with suitable n-type dopant materials, e.g Si, Ge, C, O. The III-nitride layer may have a doping concentration between 110.sup.17 cm.sup.3-510.sup.20 cm.sup.3.

(112) The thickness of this connecting layer is preferably at least 100 nm, and can be for example between 100-10000 nm.

(113) FIG. 5N-Doped Region

(114) After the growth of the connecting layer, a bulk n-doped III-nitride region 2 is grown.

(115) The n-doped region 2 may comprise or consist of a III-nitride layer containing Indium, or a stack of thin III-nitride layers with or without indium, or a bulk layer or stack of III-nitride layers with a variation in atomic percentage of indium across the layer or stack is grown. For example, the n-doped region may be a layer of n-GaN, or a layer of n-InGaN, or alternatively the n-doped region may be a stack of n-GaN/n-InGaN alternating layers, or a stack of n-InGaN/n-InGaN alternating layers having different quantities of indium in alternating layers.

(116) Preferably the n-doped region 2 comprises indium, so that the crystalline lattice of the n-doped region has similar lattice parameters to the lattice of the InGaN light emitting layer in the LED. The Indium atomic percentage in the n-doped region may vary between 0.1-25% for example.

(117) In preferred embodiments, the indium content of the n-doped region is within 20 at %, or within 15 at %, or within 10 at %, or within 5 at % of the indium content of the InGaN light emitting layer. This may advantageously ensure that the lattice parameters of the n-doped region are sufficiently similar to those of the InGaN light emitting layer to avoid excessive strain between these layers.

(118) The total thickness of the n-doped region may be at least 2 nm, or at least 5 nm, or at least 10 nm, or at least 20 nm. The thickness of the n-doped region may vary between 2 nm-5000 nm, or even thicker, for example. If the n-doped region comprises a stack of layers, the thickness of each individual layer in the stack is preferably between 1-40 nm.

(119) The n-doped region preferably has an n-type doping concentration between 110.sup.17 cm.sup.3-510.sup.20 cm.sup.3, preferably between 110.sup.18 cm.sup.3-510.sup.20 cm.sup.3, particularly preferably greater than 110.sup.18 cm.sup.3.

(120) FIG. 6Light-Emitting Region

(121) After growth of the n-doped region 2, an underlay or pre-layer or pre-well (not labelled in FIG. 6) may be grown, in order to release the strain in the light emitting layer(s). The underlay can be a single layer or stack/multi-layers of GaN, InGaN, or GaN/InGaN, or InGaN/InGaN. Alternatively, the underlay may have a structure similar to InGaN QW/GaN quantum barrier but with a lower proportion of indium. For example, before depositing the light emitting layer having a relatively high proportion of indium, an underlay consisting of a layer of bulk InGaN having a lower proportion of indium than the light emitting layer may be grown. Alternatively, the underlay may take the form of an InGaN dummy QW with a lower proportion of indium than the light emitting layer, and one or more GaN quantum barriers.

(122) After growth of the n-doped region 2 and optionally the underlay, a light-emitting region 3 containing an InGaN light emitting layer is grown.

(123) The light-emitting region 3 may contain at least one InGaN light emitting layer. Each InGaN light emitting layer may be an InGaN quantum well (QW). Preferably the light-emitting region may comprise between 1-7 quantum wells. Adjacent quantum wells are separated by barrier layers of III-nitride material having a different composition to the quantum wells.

(124) The light emitting layer(s) may be referred to as quantum wells throughout the present document, but may take a variety of forms. For example, the light emitting layers may be continuous layers of InGaN, or the layers may be continuous, fragmented, broken layers, contain gaps, or nanostructured so that the quantum well effectively contains a plurality of 3D nanostructures behaving as quantum dots.

(125) The quantum wells and barriers are grown in a temperature range of 600-800 C.

(126) Each quantum wells consists of an InGaN layer with atomic indium percentage between 15-40%. Preferably the light-emitting indium gallium nitride layer(s) and/or the quantum wells have the composition In.sub.xGa.sub.1-xN, in which 0.05x0.40, preferably 0.12x0.35 or 0.22x0.30, particularly preferably 0.22x0.27.

(127) The thickness of each quantum well layer may be between 1.5-8 nm, preferably between 1.5 nm and 6 nm, or between 1.5 nm and 4 nm. The quantum wells may be capped with a thin (0.5-3 nm) III-nitride QW capping layer, which may contain one or a combination of these elements: Al, Ga, In (ternary of quaternary layer)

(128) The QW capping layer, which is the layer added immediately after QW growth, can be AlN, AlGaN of any Al % 0.01-99.9%, GaN, InGaN of any In % 0.01-30%.

(129) The III-nitride QW barriers separating the light emitting layers (quantum wells) may contain one or a combination of these elements: Al, Ga, In (ternary of quaternary layer). The QW barrier can be AlN, AlGaN of any Al % 0.01-99.9%, GaN and InGaN of any In % 0.01-15%. Preferably the QW barrier layers contain AlN and/or AlGaN.

(130) The QW capping layer(s) and QW barriers are not indicated with individual reference numerals in the Figures, as these layers form part of the light-emitting region 3.

(131) The QW capping layers may be grown after each QW but before the barrier growth. For example, if an LED contains 3 QWs then each of these QWs may be overgrown with a QW capping layer and then a QW barrier layer, so that the light-emitting region contains 3 such QW capping layers and three such QW barrier layers. 1. One can grow the cap at the same conditions as the QW. 2. One can ramp without growth to higher temperature, and grow this cap (effectively this is an annealing step) and here the ramp can be carried out in a different gas mixture. 3. One can ramp and grow during the temperature ramp.

(132) The design of the light-emitting region may be varied according to parameters that are well understood in the art and conventional in LED design. For example, depending on the target EL emission wavelength of the LED, the composition, thickness and number of light-emitting layers and barrier layers may be varied. As described earlier in the application, the indium content of InGaN light-emitting layers may be increased when longer-wavelength emission is desired.

(133) As described above, the present invention may be provided by growing a known LED structure, known to emit at a first wavelength under electrical bias, over a template containing a porous region. The strain relaxation caused by the porous region beneath the LED structure enables incorporation of more indium into the light-emitting layer(s) under the same growth conditions, so the wavelength of the resulting LED is red-shifted when compared to the same LED structure grown under the same conditions over a non-porous substrate. A greater variety of emission wavelengths may therefore be achieved using the present invention than has been possible in the prior art, and in particular, longer wavelengths can be achieved at higher InGaN growth temperatures. This leads to superior quality crystal structures in the LED, and thus higher performance LEDs.

(134) For the manufacture of longer-wavelength LEDs the large amount of Indium in the light emitting layer(s) makes the capping layer even more important, as previous attempts to manufacture longer wavelength yellow, orange or red LEDs have failed due to not enough Indium being incorporated. So capping is very important to make sure that there is sufficient Indium trapped within the light-emitting region.

(135) FIG. 7Cap Layer

(136) After growth of the light emitting layer(s) a non-doped cap layer 4 is grown. Non-doped cap layer 4 may be termed a light-emitting-region cap layer, as this layer is formed after growth of the complete light-emitting region, for example after the growth of the stack of QWs, QW capping layers and QW barrier layers.

(137) The cap layer (light-emitting-region cap layer) 4 is a standard layer which is very well known in the growth schemes for III-nitride LEDs.

(138) The thickness of the cap layer can be between 5-30 nm, preferably between 5-25 nm or 5-20 nm.

(139) The purpose of the light-emitting-region cap layer 4, is to protect the indium in the light-emitting region (QW stack) and prevent it from desorbing/evaporating during subsequent processing. Because the InGaN QW is normally grown at lower temperature, that is not favourable for GaN/AlGaN, there is typically a temperature ramp step needed before further layers can be overgrown above the light-emitting region. The cap layer is used to ensure that the InGaN light emitting layer(s) are properly capped and protected, so that there is a chance and time window to change the p-doped layer growth conditions for better material quality. The light-emitting-region cap layer 4 also ensures that no Mg dopant is entering the QW region during the growth of p-type layers.

(140) Electron Blocking Layer (EBL)

(141) After the growth of quantum wells, capping and barrier layers, an electron blocking III-nitride layer (EBL) 5 containing Aluminium is grown. The Al % can be between 5-25% for example, though higher Al content is possible.

(142) The EBL is doped with a suitable p-type doping material. The p-type doping concentration of the EBL is preferably between 510.sup.18 cm.sup.3-810.sup.20 cm.sup.3

(143) The thickness of the EBL can be between 10-50 nm, preferably 20 nm.

(144) FIG. 8P-Doped Layer

(145) A p-doped layer 6 is grown above the electron blocking layer (EBL) 5.

(146) The p-type region is preferably doped with Mg, and the p-type doping concentration of the p-type layer is preferably between 510.sup.18 cm.sup.3-810.sup.20 cm.sup.3.

(147) The p-doped III-nitride layer may contain In and Ga.

(148) The doping layer is preferably between 20-200 nm thick, particularly preferably between 50-100 nm thick. The doping concentration may vary across the p-type layer and can have a spike in doping levels in the last 10-30 nm of the layer towards the LED surface, in order to allow better p-contact.

(149) For activation of Mg acceptors in the p-doped layer, the structure may be annealed inside of MOCVD reactor or in an annealing oven. The annealing temperature may be in the range of 700-850 C. in N.sub.2 or in N.sub.2/O.sub.2 ambient.

(150) As both the EBL and the p-doped layer are p-type doped, these layers may be referred to as the p-doped region.

(151) FIG. 9Transparent Conducting Layer

(152) The stack of active semiconductor layers is covered with a transparent conducting layer 7. The transparent conducting layer can be made of Ni/Au, indium tin oxide, indium zinc oxide, graphene, Pd, Rh, silver, ZnO etc., or a combination of these materials.

(153) The thickness of the transparent conducting layer can be between 10-250 nm.

(154) Transparent conducting layers are well known in the art, and any suitable material and thickness may be used.

(155) An annealing step may be required for making the p-contact ohmic.

(156) FIG. 10

(157) Depending on the LED structure being manufactured, the semiconductor structure may be processed into LED, mini-LED or micro-LED devices.

(158) Normal LEDs are typically larger than 200 m (referring to the lateral dimensions of width and length of the LED structure. Mini-LEDs are typically 100-200 m in lateral size, while Micro-LEDs are typically less than 100 m in size.

(159) FIG. 10 onwards illustrates the semiconductor structure following etching of layers 2-7 of the semiconductor structure into multiple discrete LED stacks, or mesas, each having the same structure.

(160) The steps of LED fabrication are conventional and well-known to those skilled in the art. The order of the following fabrication steps are not specific to the present invention, and the skilled person will be appreciated that LED devices within the scope of the present invention may be prepared using alternative fabrication steps to those illustrated below. For the purposes of illustration only, however, one preferred fabrication route to prepare LEDs according to the present invention is described below.

(161) In the next step, the transparent conducting layer 7 is structured in such a way that it covers only the top surface of the active emission element. The structuring can be done using standard semiconductor processing methods that included resist coating and photolithography. The transparent conducting layer is etched by using wet chemistry or a sputter etch process using Argon. This step is followed by wet or dry etching of the III-nitride structure. An inductively couple plasma reactive ion etching, only reactive ion etching or neutral beam etching is used to create mesas in the III-nitride layer. The dry etch process may include either one or more of Cl, Ar, BCl.sub.3, SiCl.sub.4 gases.

(162) The purpose of this step is to isolate the individual emitting elements and access the buried n-doped layer of the p-n junction.

(163) After the dry etch process a wet etch process is done to remove the dry etching damage from the sidewalls of the mesa. The wet chemistry may involve KOH (1-20%), TMAH or other base chemistries.

(164) FIG. 11Passivation

(165) The next step is to deposit a passivation layer 8 or a combination of passivation layers. The starting passivation layer can be Al2O3 (10-100 nm) (deposited by atomic layer depositions) followed by sputtered or plasma enhanced chemical vapor deposited SiO2, SiN or SiON (50-300 nm).

(166) The Al2O3 can be deposited between 50-150 C.

(167) The SiO2, SiN and SiON can be deposited between 250-350 C.

(168) The sputter process can be done at room temperature.

(169) FIGS. 12-13

(170) The next step is to create openings in the oxide passivation layer 8 to expose the top of the LED structure. This can be done via wet or dry etching or a combination of both.

(171) For wet etching buffered oxide etch, diluted hydrofluoric acid phosphoric acid or a mixture of these can be used.

(172) Channels are also etched through the connecting layer 1 between the LED structures, followed by electrically isolating the LED structures from one another by depositing dielectric mask material 8 into the channels, so that the LEDs are operable independently from one another.

(173) The next step in device fabrication is to cover the transparent conducting layers 7 on the p-doped layers 6 with metal layers to act as electrical p-contacts 9. The covering can be done with a single step or multiple steps. The metals can be covering the pixels completely or partially. In this example a single step is used to simplify the details.

(174) The metal contacts 9 may contain Ti, Pt, Pd, Rh, Ni, Au. The thickness of the complete metal stack can be between 200-2000 nm.

(175) FIG. 14Exposing Connecting Layer

(176) Standard photolithography techniques can be used to create openings in the second mask layer 8 to expose a plurality of regions of the connecting layer 1. The size of the openings can vary between 200 nm-50000 nm. This distance between the openings can be between 500 nm-30000 nm. The opening are creating only in the regions of the wafer that are not occupied by LED structures.

(177) Dry etching is preferably used to etch the second mask layer 8 using fluorine based gases.

(178) FIG. 15N-Contacts

(179) The next step in device fabrication is to cover the openings in the oxide 8 with metal contacts 10 to access the connecting layer 1, which is in electrical contact with the n-doped layers of the LED structures. The covering can be done with a single step or multiple steps. The metals can be covering the pixels completely or partially. In this example a single step is used to simplify the details.

(180) The metal may contain Ti, Pt, Pd, Rh, Ni, Au. The thickness of the complete metal stack can be between 200-2000 nm.

(181) FIGS. 16-18

(182) After this processing, the substrate can be thinned, and/or the porous region can be removed so that the connecting layer 1 is exposed.

(183) Surface structuring or texturing can be done on the substrate, at the porous region, or layer 1 to enhance the light output and control the emission angle, as well as other optical engineering and design.

(184) Finally, the wafer/devices can be flipped, and bonded to another carrier substrate either can be silicon/sapphire or any type as passive devices, alternatively, the devices can be bonded to a CMOS silicon backplane for active matrix micro-LED display panel.

(185) As shown in FIG. 16, the top side of the device may be bonded to another carrier wafer/substrate/backplane 11, or to a microdriver circuit board to form an array of pixels.

(186) As shown in FIG. 17, the substrate may then be removed from the device, and the bottom-side of the device may be bonded to a cover glass or transparent material 12.

(187) As shown in FIG. 18, the substrate and the porous and non-porous region may be removed from the device. The top side of the device may be bonded to another carrier wafer/substrate/backplane 11, or to a microdriver circuit board to form an array of pixels. The bottom-side of the device may be bonded to a cover glass or transparent material 12.

(188) The skilled person will understand that the emission wavelengths of the individual LED structures may be controlled by altering the composition and layer structures of the LED structures according to known principles of LED construction. Thus a variety of variable-wavelength LED devices emitting over different emission wavelength ranges may be provided using the present invention, and colour combinations other than green to red may be provided.

(189) FIGS. 19-23

(190) FIG. 19 shows an example of an InGaN LED over a porous layer that emits at a peak wavelength of around 625 nm due to the wavelength red-shift caused by the porous region.

(191) FIGS. 20 and 21 compare the emission characteristics of an InGaN LED on a non-porous substrate (FIG. 20) and the same InGaN LED grown on a template comprising a porous layer of III-nitride material. Comparison of these two graphs demonstrates the shift towards longer emission wavelengths caused by the porous underlayer, as the emission of the LED on the porous template is consistently between 21 nm and 45 nm longer than that of the same LED on the non-porous template.

(192) FIGS. 22 and 23 compare the I-V characteristics of yellow InGaN micro-LEDs on a non-porous substrate (FIG. 22) with the same InGaN micro-LEDs grown on a template containing a porous layer. On the porous template, the InGaN micro-LEDs emit red light, as shown in the inset image.

(193) FIG. 24 is an I-V curve measured for InGaN micro-LEDs of different pixel sizes (10 m10 m; 20 m20 m; 30 m30 m; 50 m50 m) on a porous substrate. FIG. 25 shows the I-V characteristics of the same pixels, with the axes altered to focus on low currents from 110.sup.6 A to just over 100 A.

(194) FIG. 26 is a series of five EL images of the same variable-wavelength MicroLED InGaN pixel being driven at different currents in constant wave mode (CW), showing five different colours of emission. In the left-hand image, the micro-LED emission colour is seen to be red at a driving current of 50 A. In the second image from the left, the micro-LED emission colour is seen to be red-orange at a driving current of 100 A. In the third image from the left, the micro-LED emission colour is seen to be orange at a driving current of 1 mA. In the fourth image from the left, the micro-LED emission colour is seen to be yellow-green at a driving current of 10 mA. In the right-hand image, the micro-LED emission colour is seen to be green at a driving current of 20 mA.

(195) By varying the driving current between 50 A to 20 mA, the same micro-LED is therefore capable of emitting at wavelengths ranging from red to green. The spectral width of this emission wavelength range is on the order of 90 nm (from around 570 nm to around 660 nm). This is a far greater range of emission wavelengths than has ever been achievable with a single LED in the prior art.

(196) FIG. 27A is an emission wavelength vs current density plot for a 25 m25 m InGaN LED pixel array (100100 array, containing 10,000 pixels) driven in pulsed mode with a 100 s pulse at 1% duty cycle. FIG. 27B is an emission wavelength vs current density plot for a 30 m30 m InGaN LED pixel array (100100 array, containing 10,000 pixels) driven in pulsed mode with a 100 s pulse at 1% duty cycle.

(197) Both of these plots show the controllability of the peak emission wavelength with a pulse driven power supply. In particular, the wavelength is linearly dependent on the current density (plotted on a logarithmic scale). This linearity can equally be manipulated when driving with a pulsed voltage power supply. The variable emission wavelengths of the LED can therefore be controlled with either voltage or current driving schemes in either CW or pulsed mode, all of which are standard ways of display driver IC.

(198) This linear relationship between the driving current density and the resulting emission wavelength is highly advantageous for the purposes of LED display design, as it enables accurate control of the emission wavelengths by varying the current density of the power supply.

(199) FIG. 28 is a plot of intensity vs wavelength for a variable-wavelength InGaN LED driven at different DC currents. The power supply is operated in pulsed driving mode with a 100 s pulse at 1% duty cycle.

(200) FIG. 28 again reflects a gradual, continuous transition of the peak emission wavelength of the LED as the current of the power supply is varied. At a driving current of 200 mA, the peak emission wavelength is around 575 nm, with an intensity of around 10 W/nm. As the driving current is reduced, however, the peak emission wavelength moves gradually to longer wavelengths, and to lower emission intensities. When the driving current reaches 7 mA, the peak emission wavelength is approximately 675 nm, with an intensity of around 0.1 W/nm.

(201) FIG. 29A is a schematic plot of an LED structure over a porous template comprising a porous region of III-nitride material. The LED comprises a current constraining layer 100 positioned between the n-doped portion (labelled n-GaN in this example) and the light-emitting region of the LED. The light-emitting region is labelled the MQW (multiple quantum well) region of the LED. Although an SiN current constraining layer 100 is indicated in the Figures, the current constraining layer 100 (which may alternatively be termed a current constricting layer) may be formed of another dielectric material.

(202) A circular aperture 110 is provided through the centre of the dielectric current constraining layer 100. The aperture extends through the thickness of the current constraining layer, providing a conductive pathway between the n-doped region and the light-emitting region of the LED. In the illustrated embodiment, the diameter of the aperture is approximately 33% of the lateral width of the LED structure, but the width of the aperture may be varied to modify the local current density passing through the aperture.

(203) FIG. 29B is a schematic plot of an LED structure comprising a current constraining layer 100 in an alternative position, positioned between the p-doped portion (labelled p-GaN in this example) and the light-emitting region (MQW region) of the LED. An aperture 110 extends through the thickness of the current constraining layer 100, providing a conductive pathway between the p-doped region and the light-emitting region of the LED.

(204) FIGS. 30-35

(205) FIGS. 30-34 are schematic cross-sections of an LED structure formed over a porous template according to a preferred embodiment of the present invention.

(206) The LED structure comprises a substrate, which as described above may be Silicon, Sapphire, SiC, -Ga2O3, GaN. Substrate size can be as small as 11 cm-2, 50 mm, 100 mm, 150 mm, 200 mm, 300 mm or larger in diameter.

(207) A porous region is formed over the substrate, and a connecting layer (layer 1) of (Al,In) GaN is formed over the porous region. An n-type layer of n (Al,In) GaN (layer 2) is positioned over the connecting layer, and forms the n-type portion of the LED device. A pre-strain layer (layer 3) is formed over the n-type layer and an active region (layer 4) containing a multiple quantum well (MQW) is positioned over the pre-strain layer and below a p-type layer (layer 5) of p-(Al,In) GaN.

(208) This is a typical LED structure formed over a porous semiconductor template. The porous region can be a uniform porous layer or region, and can also be a partly patterned porous region.

(209) The porous region may be any porous region as described abovea variety of thickness, composition and configuration are possible within the scope of the invention.

(210) In a preferred embodiment of the invention illustrated schematically in FIG. 31, two v-shaped pits are formed in the LED structure. These pits are v-shaped in cross section, and create v-shaped voids through the upper layers of the LED structure.

(211) As illustrated, a first v-shaped pit 311 extends from the connecting layer (layer 1) to the upper surface of the LED device, which is formed by the outer surface of the p-type layer (layer 5). The narrow point of the v-shaped pit is located in the connecting layer, and the width of the pit enlarges with each layer that is epitaxially grown over the connecting layer, reaching its widest point at the surface of the p-type layer.

(212) A second v-shaped pit 312 extends from the pre-strain layer (layer 3) to the upper surface of the LED device, which is formed by the outer surface of the p-type layer (layer 5).

(213) As shown in FIG. 32, the first and second v-shaped pits 311, 312, extend through the active region MQW of the LED structure. The v-shaped pits thus create gaps or voids in the semiconductor structure. V-shaped pits can be generated either in layer 1 or layer 3, but have to go through layer 4, which is the MQW region.

(214) Active light emitting multiple quantum wells (MQWs) and quantum barriers (QBs) can be made of (Al,In) GaN and (Al,In) GaN, respectively, and effectively any combination of materials, composition, and thickness. The MQW can have any period, 1QW, 2QW, 3QW, 4QW, 5QW, all the way to 10QWs or more.

(215) In some preferred embodiments the QWs are continuous. In some preferred embodiments the QWs are fragmented.

(216) V-shaped pits can be originated or caused by the existence of threading dislocations. V-shaped pits can alternatively be formed by different growth modes in the epitaxy process, i.e. 3-dimensional growth.

(217) The growth of v-shaped pits is described, for example, in The effect of nanometre-scale V-pits on electronic and optical properties and efficiency droop of GaN-based green light-emitting Diodes; Zhou et al; Scientific Reports|(2018) 8:11053|DOI: 10.1038/s41598-018-29440-4.

(218) As shown in FIG. 33, the second v-shaped pit 312 originates from a threading dislocation 330. The threading dislocation originates at the porous region or the connecting layer (layer 1), and extends upwards through sequentially deposited layers of semiconductor material (layers 1&2). At the pre-strain layer (layer 3) the threading dislocation 330 begins to widen into a v-shaped pit 312, and as further semiconductor layers are epitaxially grown over layer 3 the v-shaped pit 312 gets wider and wider.

(219) As the v-shaped pits run through the MQW active region during epitaxial growth, there will also be MQWs being grown on the side walls 340 of the v-shaped pits. The MQWs deposited which will be of different thickness and composition comparing to the planar MQWs at which no such pits are present.

(220) As the v-shaped pits have begun to form before the active region including the MQWs is deposited, the semiconductor material that is epitaxially deposited over the layers below is deposited into the v-shaped pits. At the v-shaped pits, the layers of material forming the MQWs are therefore distorted as the layers are stretched down the sidewalls 340 of the v-shaped pits. This is shown by the transmission electron microscopy (TEM) image of FIG. 35.

(221) As shown in FIG. 35, on either side of the v-shaped pit 312, layers of semiconductor structure are grown as flat planar layers. The active MQW region is thus planar around the v-shaped pit (on either side as shown in cross-section). In the location of the v-shaped pit, however, the MQW layers are distorted and stretched downwards along the sidewalls 340 into the v-shaped pit. This stretching effect changes the thickness of the QWs on the sidewalls 340 of the pit, so that they are different in thickness compared to the planar QW layers formed over the rest of the LED structure.

(222) The inventors have found that v-shaped pits can create local strain relaxation, and MQWs deposited on the sidewall of these v-pits will have different thickness and composition compared to the rest of the MQW, hence the MQW in the region of the v-shaped pits will produce a different emission wavelength.

(223) FIGS. 36 & 37

(224) FIG. 36 is a graph of normalised intensity vs peak emission wavelength, for a variable-wavelength LED containing v-shaped pits as shown in relation to FIGS. 31-35. As the driving conditions applied to the LED are varied, the LED emits light at different peak emission wavelengths. As shown in FIG. 36, the peak emission wavelength of the LED can be varied across a continuous wavelength range from around 530 nm to around 640 nm,

(225) FIG. 37 is a graph of peak emission wavelength vs driving current, for a variable-wavelength LED according to the present invention. FIG. 37 shows that as the driving current applied to the LED is increased, the peak emission wavelength changes smoothly from around 550 nm to around 635 nm, Higher driving currents cause the LED to emit at shorter wavelengths, while lower driving currents cause the LED to emit at lower wavelengths. The variation in peak emission wavelengths is continuous and consistent as the driving current is varied, leading to simple calibration of the LED device.

(226) FIG. 38

(227) FIG. 38 is a graph of peak emission wavelength for variable-wavelength LEDs of different pixel sizes. As shown in FIG. 38, changing the size of LED pixels affects the peak emission wavelengths that the pixels emit. Under the same driving conditions, pixels of different sizes (which are otherwise identical) will emit at different peak wavelengths.

(228) FIGS. 39 & 40

(229) FIGS. 39 and 40 are a schematic side-on cross-sections of a variable-wavelength LED structure according to two alternative embodiments.

(230) In the following examples, the LEDs and LED subpixels are preferably variable-wavelength LEDs as described above.

(231) The LED structures in FIGS. 39 and 40 are shown with simplified diode structures, in which a light-emitting region containing a multiple quantum well (MQW) is positioned between 1.sup.st and 2.sup.nd semiconductor layers. The LED structure is provided on a substrate which preferably contains a porous region of semiconductor material.

(232) FIGS. 39 and 40 illustrate alternative arrangements for the electrical contacts attached to the LEDs. In both embodiments, a 1.sup.st electrical contact is shown positioned over the 1.sup.st semiconductor layer. An electrically-insulating masking layer 390 (which may be called a passivation layer) is positioned partially between the 1.sup.st semiconductor layer and the 1.sup.st contact layer, so that the 1.sup.st contact layer contacts the 1.sup.st semiconductor layer only through an aperture in the masking layer 390. The size of the aperture determines the contact area over which the 1.sup.st semiconductor layer and the 1.sup.st contact layer are in contact with one another. When a driving current is applied through the electrical contacts, this contact area restricts the area through which the driving current can pass into the LED structure. The size of the contact area and the magnitude of the driving current determines the current density experienced by the LED structure. The current density experienced by the LED structure in turn determines the peak emission wavelength of the light emitted by the MQW.

(233) The 2.sup.nd contact layers are arranged differently in FIGS. 39 and 40.

(234) In FIG. 39, the 2.sup.nd contact layer is arranged to contact the 2.sup.nd semiconductor layer through an aperture in the masking layer 390. As for the 1.sup.st contact layer, the size of the aperture determines the contact area shared by the 2.sup.nd contact layer and the 2.sup.nd semiconductor layer.

(235) In FIG. 40, the 2.sup.nd contact layer is positioned on the substrate instead of on the 1.sup.st semiconductor layer.

(236) The LED structures of FIGS. 39 and 40 may be formed as described above in relation to FIGS. 1 to 18, with the 1.sup.st and 2.sup.nd semiconductor layers acting as p-type and n-type layers so that the MQW emits at a peak wavelength when a driving current is applied between the two electrical contacts.

(237) The first semiconductor layer may include but is not limited to p-Gan and n-Gan.

(238) The first contact layer may include but is not limited to titanium, platinum, chromium, aluminium, nickel, gold and some compounds such as ITO (indium tin oxide).

(239) Because the emission wavelength will blue-shift with increasing injection current density, in order to control the current density, there will be different contact areas, as shown in FIG. 41.

(240) FIG. 41 illustrates three different contact areas between a contact layer and a semiconductor layer in an LED diode. A 1.sup.st electrical contact layer is shown positioned over a 1.sup.st semiconductor layer. An electrically-insulating masking layer 390 is positioned partially between the 1.sup.st semiconductor layer and the 1.sup.st contact layer, so that the 1.sup.st contact layer contacts the 1.sup.st semiconductor layer only through an aperture in the masking layer 390. The size of the aperture determines the contact area over which the 1.sup.st semiconductor layer and the 1.sup.st contact layer are in contact with one another. When a driving current is applied through the electrical contacts, this contact area restricts the area through which the driving current can pass into the LED structure. The size of the contact area and the magnitude of the driving current determines the current density experienced by the LED structure. The current density experienced by the LED structure in turn determines the peak emission wavelength of the light emitted by the MQW.

(241) FIG. 42 is a plan view of an electrical contact positioned over an LED subpixel mesa. The contact area A.sub.contact is necessarily smaller than the area of the subpixel mesa A.sub.mesa, as the largest possible contact would contact the LED subpixel mesa over its entire area, giving a contact ratio of 1:1.

(242) FIGS. 43A-43C illustrate three alternative ways of obtaining RGB subpixels using three subpixels having the same diode structure.

(243) In FIG. 43A, three subpixel mesas have the same subpixel mesa area, but different sizes of electrical contact pad. The different sizes of electrical contact pad means that each of the three subpixels has a different contact area, and thus each of the subpixels has a different contact ratio (the ratio of contact area:light-emitting area, which because the light-emitting region spans the entire subpixel mesa, is the same as the ratio of contact area:subpixel mesa area). The subpixel on the left has the largest contact area, and therefore the largest contact ratio. The subpixel on the right has the smallest contact area, and therefore the smallest contact ratio. And the subpixel in the centre has a contact area and contact ratio between the other two.

(244) The smaller the contact ratio, the smaller the contact area through which driving current is provided to the subpixel, and thus the higher the current density experienced by the subpixel. As shown in FIGS. 27A and 27B, higher current densities lead to shorter peak emission wavelengths, so of the three subpixels in FIG. 43A, the subpixel on the left emits at the longest peak wavelength and the subpixel on the right emits at the shortest peak wavelength.

(245) In particularly preferred embodiments, the contact ratios of the three subpixels, and the driving current, are chosen so that the three subpixels emit at red (left subpixel), green (centre subpixel) and blue (right subpixel) in response to a single driving current.

(246) FIG. 43B shows an alternative way of arriving at the same three contact ratios as FIG. 43A. In FIG. 43B, the three subpixels again have the same mesa area, and thus the same light-emitting area. In FIG. 43B, the three electrical contact pads also have the same area when viewed from above. However, masking layers 390 are positioned between the subpixels and the electrical contact layers, with the masking layers on the three subpixels containing apertures of different sizes. The contact layers can only contact the subpixel diode structure through the apertures, so the sizes of the apertures control the contact areas. The subpixel on the left has the largest aperture in the masking layer, and thus the largest contact area, which leads to the largest contact ratio. The subpixel on the right has the smallest aperture through the masking layer, and thus the smallest contact area and the smallest contact ratio. The subpixel in the centre has a masking layer aperture with a size in between the other two, which leads to a contact area and contact ratio between those of the other two subpixels.

(247) The sizes of the apertures in the masking layers are selected to create the same contact areas as the subpixels in FIG. 43A. Thus in response to the same driving current, the three subpixels in FIGS. 43A and 43B will emit at three corresponding peak emission wavelengths, preferably RGB wavelengths.

(248) FIG. 43C shows a third way of creating three subpixels having the same contact ratios as those shown in FIGS. 43A and 43B.

(249) In FIG. 43C, the three subpixel mesas each have different subpixel mesa areas, and each have different sizes of electrical contacts creating different contact areas on the three subpixels. In FIG. 43C however, the relative sizes of the subpixels and contact areas are the same as those shown in FIGS. 43A and 43B. Thus the red subpixel shown on the left in FIG. 43C has the same contact ratio as the red subpixels on the left of FIGS. 43A and 43B; the green subpixel shown in the top right in FIG. 43C has the same contact ratio as the green subpixels in the centre of FIGS. 43A and 43B; and the blue subpixel shown in the bottom right in FIG. 43C has the same contact ratio as the blue subpixels on the right of FIGS. 43A and 43B. As the peak emission wavelength at a given driving current is determined by the contact ratio (the ratio of contact area:light-emitting area, which because the light-emitting region spans the entire subpixel mesa, is the same as the ratio of contact area:subpixel mesa area), the subpixels of all of FIGS. 43A-43C will emit at the same three wavelengths in response to the same driving current.

(250) FIGS. 44 and 45 show three LED subpixels formed from a single LED structure, similar to the embodiment shown in FIG. 13. In the embodiments shown in FIGS. 44 and 45 however, instead of each LED subpixel having a uniform metal contact layer 9, the three subpixels each have differently sizes electrical contacts 9A, 9B, 9C, which create different contact areas between the contacts and the three subpixel diode structures.

(251) The embodiment of FIG. 44 corresponds to the example of FIG. 43A, in which the three subpixels all have the same mesa size, and thus the same light-emitting area, but in which the three electrical contacts are formed in differently sized apertures in an electrically-insulating masking layer 8. The subpixel on the left has the largest electrical contact 9A, and thus the largest contact area and the largest contact ratio. The subpixel on the right has the smallest electrical contact 9C, and thus the smallest contact area and the smallest contact ratio. The subpixel in the centre has an electrical contact 9B having a size between the sizes of the other contacts 9A, 9C, and thus the centre subpixel has a contact area and contact ratio between those of the other two subpixels. These three subpixels will emit at three different peak wavelengths in response to the same driving current. The subpixel on the left will emit at the longest wavelength, due to its contact ratio being the highest, and the subpixel on the right will emit at the shortest wavelength, due to its contact ratio being the lowest.

(252) FIG. 45 is an analogue of FIG. 43B, in which the electrical contacts 9A, 9B, 9C have the same size when viewed from above, but the contact areas between the electrical contacts and the three subpixels are controlled by the sizes of apertures through the masking layer 8. The contact areas between the electrical contacts 9A, 9B, 9C and the three subpixels are the same as those in FIG. 44, so the three subpixels will emit at corresponding wavelengths in response to the same driving current.

(253) FIGS. 46A-G illustrate alternative embodiments of light-emitting regions of variable-wavelength LEDs according to the present invention.

(254) Examples of MQWs:

(255) 1. Continuous MQWs 2. V-pits 3. Broken QWs, gappy QWs, fragmented QWs 4. QDs 5. Well-width fluctuation 6. Alloy composition 7. Different combinations of MQWs and underlayers

(256) These structural characteristics can be identified and examined by standard material characterisation techniques, such as cross-sectional transmission electron microscopy (TEM), X-ray diffraction (XRD), Energy Dispersive X-ray Spectroscopy (EDX or EDS), 3D atom probe (3DAP).

(257) FIG. 46A shows a continuous MQW light-emitting region of an LED, in which three identical QWs are provided between four identical quantum barriers (QBs).

(258) FIG. 46B shows the continuous MQW of FIG. 46B, with a V-shaped pit propagating through the light-emitting region. The v-shaped pit terminates in a threading dislocation, and has QWs on its semi-polar facets.

(259) FIG. 46C shows a MQW in which the QW layers comprise discontinuities or gaps in the semiconductor material.

(260) FIG. 46D shows a MQW in which quantum dots (QDs) create non-uniformities in the MQW. QDs may be provided on or in the QB or QW layers, for example in gaps in the QW structure.

(261) FIG. 46E shows a MQW with well-width fluctuation, in which the thicknesses of the QW layers are not uniform across the light-emitting region. The QWs may have different widths from each other, and also varying widths within a single QW.

(262) FIG. 46F shows a MQW with fluctuations in alloy composition in the light-emitting region. The compositions of the QBs and the QWs differ from layer to layer. In particular, the indium In % composition is varying within the same QWs, i.e. in QW2, In % is varying between 10-12% or 10-15%, or 10-25%, or 10-35%.

(263) FIG. 46G shows a MQW containing different combinations of MQWs and underlayers. In % composition is different across different QWs. For example In % in QW1 is 15%, In % in QW2 is 25%, and In % in QW3 is 30%. In embodiments of the present invention, the lower In % QW is preferably positioned at the bottom of the MQW, due to its strain and thermal effect, while the high In % QWs is preferred to be on the top. In a preferred embodiment, for example, QW1 is a blue emitting QW, QW2 is a green emitting QW, QW3 is a red emitting QW.

(264) FIG. 47A is a TEM image of a cross-section of a conventional non-variable-wavelength LED. Non-variable LEDMQWs are uniform and smooth in both upper and lower interface (5 MQWs shown here clearly)

(265) FIGS. 47B and 47C are TEM images a variable-wavelength LED comprising v-shaped pits, according to an embodiment of the present invention. Variable wavelength LEDMQWs are non-uniform, which induced by various methods, one example is v-pits and the semi-polar facets which would incorporate more indium and thinner QWs. Another example is also evident in FIG. 47B, that the MQWs are not uniform, in terms of broken QWs, discontinuous QWs, fragmented QWs, QWs with well-width or In composition fluctuation

(266) FIG. 47C shows a cross section of the variable-wavelength LED of FIG. 47B, showing a porous region and a light-emitting region comprising a plurality of v-shaped pits, according to a preferred embodiment of the present invention.

(267) In this structure, the light-emitting region contains multiple emission wavelength regions that are deliberately introduced such as multiple types of QW region with v-shaped pits extending through the light-emitting region.

(268) V-shaped pits (V-pits) are actually hexagonal pits looking from the above, v-shape is when looking at the cross-section. V-pits can be initialize at each site of dislocations under special epitaxy growth conditions during the growth of InGaN, GaN, InGaN/InGaN superlattice, or InGaN/GaN superlattice structures underlying the MQWs, such as low growth temperature (e.g. <1000 C., or <900 C., or <800 C., or <700 C.) and nitrogen ambient.

(269) FIG. 48A is a graph of peak emission wavelength vs driving current density for a conventional non-variable wavelength LED. By varying the driving current density applied to the LED, the emission wavelength can be slightly varied, across an emission wavelength range of around 15 nm.

(270) FIG. 48B is a graph of peak emission wavelength vs driving current density for a variable-wavelength LED according to an embodiment of the present invention. In the variable-wavelength LED, varying the current density of the driving power supply creates a much larger variation in the peak emission wavelengths (WLP) emitted by the LED. In this embodiment, varying the driving current density between roughly 0.1 and 100 A/cm.sup.2 varies the peak emission wavelength from around 635 nm to around 550 nman emission wavelength range of around 85 nm.

(271) FIG. 48C is a graph of peak emission wavelength vs driving current density for a variable-wavelength LED according to another embodiment of the present invention. In this embodiment, varying the driving current density varies the peak emission wavelength from around 720 nm to around 580 nman emission wavelength range of around 140 nm.

(272) FIG. 49A is a graph of peak emission wavelength vs driving current density for another variable-wavelength LED according to the present invention. In this embodiment, varying the driving current density between roughly 0.1 and 200 A/cm.sup.2 varies the peak emission wavelength from 615 nm to 508 nman emission wavelength range of around 100 nm. The data for this graph only goes to 514.5 nm due to a limitation on the testing capabilities, The current density for 508 nm is therefore estimated. However, the obtainable range of emission wavelengths can be pushed either way significantly.

(273) FIGS. 49B-D are photographs of the variable-wavelength LED of FIG. 49A, showing the same variable-wavelength LED emitting at four different wavelengths across its emission wavelength range. The inset emission spectra show the different peak emission wavelengths at different driving current densities. This shows the same variable-wavelength LED emitting at peak emission wavelengths in the orange (615 nm), yellow (556 nm), green (534 nm) and blue (508 nm) in response to different driving current densities.

(274) Display Devices

(275) Conventionally, colour display pixels are made up of multiple single-colour subpixels: a blue subpixel, a green subpixel and a red subpixel. The observed pixel chromaticity is the spatial combination of the light emitted by the three subpixels, as shown in FIG. 50B.

(276) With the variable-wavelength LED of the present invention, individual LED subpixels can display colours over a wide spectral range, for example from blue to red. This allows different colours to be achieved by using a single LED chip and different driving time frames. The observed pixel colour produced is the temporal combination of light emitted by the LED subpixel.

(277) A variety of display device pixels may be configured to incorporate one or more variable-wavelength LEDs according to the present invention. In all of the following pixel embodiments, the overall emitted colour perceived by a viewer is the spacial and temporal combination of the light emitted by the subpixels in any given device pixel.

(278) FIG. 51B illustrates a device pixel consisting of a single variable-wavelength LED. By varying the driving conditions supplied to the LED, the peak emission wavelength from this pixel can be varied across the emission wavelength range, the width and absolute wavelengths of which are determined by the driving conditions and the LED diode size and structure.

(279) FIG. 50C illustrates a device pixel comprising two subpixels, both of which are variable-wavelength LEDs according to the present invention. The two subpixels may be controlled separately by separately controlling the driving current provided to each subpixel, so that the subpixels may be controlled to emit at different wavelengths.

(280) FIG. 50D illustrates a device pixel comprising three variable-wavelength LED subpixels. By controlling the driving currents provided to the three separate subpixels, the peak emission wavelengths of each subpixel may be individually varied.

(281) FIG. 50E illustrates a device pixel comprising one variable-wavelength subpixel and one fixed-emission-wavelength subpixel.

(282) FIG. 50F illustrates a device pixel comprising one variable-wavelength subpixel and two fixed-emission-wavelength subpixels.

(283) FIG. 50G illustrates a device pixel comprising two fixed-emission-wavelength blue and red subpixels, and two fixed-emission wavelength green subpixels configured to emit at different peak wavelengths within the green range of the spectrum. The green subpixels may be variable-wavelength subpixels configured to receive two different fixed driving current densities, corresponding to different peak emission wavelengths in the green.

(284) Display Devices with Expanded Colour Gamut

(285) Conventional LED displays typically show colours by combining light from subpixels with different primary colours, as shown in FIG. 50A. Conventional pixels include red, green and blue subpixels.

(286) Combinations of light from three primary colour subpixels allow any colour within the triangle (defined by the primary colours) shown in FIG. 51A to be displayed. The triangle shown in FIG. 51A defines the colour gamut obtainable with such a display. The colour space outside of the triangle is not accessible by a conventional three subpixel pixel, and therefore the colour gamut achievable by the pixel is limited.

(287) Solutions to increase colour gamut: Including additional sub-pixels with different colours, the colour gamut then being defined by a quadrilateral as shown in FIG. 51B. This requires additional cost and complexity. Further expansion to 5 or more sub-pixels would further expand gamut but increase complexity and cost.

(288) By incorporating the variable-wavelength LED of the present invention into an LED display device, preferably there is provided a display device with controllable chromaticity comprising: an LED structure whose emission spectrum is strongly dependent on drive current density, where the shift in peak emission wavelength is greater than 20 nm/decade, where the peak emission wavelength is controllable in the range 450 nm to 630 nm or wider.

(289) The display comprises subpixels formed of LED devices with controllable chromaticity (variable emission wavelengths).

(290) Preferably subpixels have the same diode structure, and the peak emission wavelength from any given subpixel is controlled only by current density provided to that subpixel during use.

(291) Subpixels may have constant chromaticity (for example in response to a fixed driving current) or may change chromaticity (in response to a varying driving current) such that a greater colour gamut may be achieved. The achievable colour gamut is equal to or greater than that defined by sRGB primaries.

(292) In some embodiments, all subpixels may change chromaticity dynamically from frame to frame. In other embodiments, only some subpixels may change chromaticity dynamically from frame to frame.

(293) A wide colour gamut is thus achievable with a display device incorporating one or more variable-wavelength LEDs according to the present invention.

(294) FIG. 51C illustrates the expanded colour gamut achievable with a variable-wavelength LED device when peak emission wavelength is controlled from 450 nm to 620 nm

Embodiment 1 (FIGS. 52A-D)

(295) FIG. 52A illustrates a display pixel comprising three subpixels. Each subpixel may be an LED device with controllable peak wavelength of light emitted.

(296) When viewed at a distance the observed chromaticity of light emitted is a spatial and temporal combination of the sub-pixel light emission.

(297) Blue (B) and red (R) subpixel are operated with fixed emission corresponding to fixed observed chromaticity. The peak emission wavelengths of the B and R subpixels may be fixed by providing driving currents having fixed magnitudes corresponding to the current densities required for emission at the desired blue and red wavelengths.

(298) Green (G) subpixel can operate in two modes with different peak wavelengths, by providing two separate driving current modes having different magnitudes corresponding to the current densities required for emission at the two desired green wavelengths. The green subpixel can be operated in either of the green modes by switching to the desired drive current mode.

(299) In combination with the B and R subpixels, each of the two G sub-modes enables a different gamut of colours to be displayed.

(300) The effective gamut of the display is that which can be achieved with G sub-pixel in either mode

(301) By extension, when variable-wavelength LEDs are used for the R or B subpixels, those subpixels could also be switched between two or more modes. All pixels could be switched between two or more modes allowing the gamut to be set once or dynamically controlled during regular operation.

(302) In an alternative embodiment, conventional blue and red LEDs may be incorporated into the pixel to act as the blue and red subpixels, with a variable-wavelength LED forming the green subpixel.

Embodiment 2 (FIG. 53)

(303) A display formed with each display pixel comprising three subpixels. Each subpixel is preferably a variable-wavelength LED device with controllable peak wavelength of light emitted, as described above.

(304) During operation of the display device, subpixels are switched from emitting one peak wavelength to a different peak wavelength within a unit of time, as shown in FIG. 53. The peak emission wavelengths of any given subpixel is varied by varying the magnitude of the drive current provided to that subpixel. A unit of time may be a display frame or shorter such that a number of time units occur in a single display frame.

(305) Emitting peak wavelengths which require high current density can cause significant localised heating which can affect performance and reliability of devices and the whole display.

(306) The advantage of this method of controlling the subpixels is that heating of sub-pixels is shared more uniformly distributed over the display pixel and hot spots on particular sub-pixels are avoided.

(307) In the four time units of time illustrated in FIG. 53, B and R subpixels are shown as swapping, but if the G subpixel is a variable-wavelength LED then the G subpixel can also switch emission wavelengths. Any combination of switching is possible with the present invention, simply by varying the drive currents provided to the separate subpixels.

Embodiment 3 (FIGS. 54A & 54B)

(308) FIG. 54A schematically illustrates a pixel of a display device formed with pixels comprising four subpixels. Each subpixel is preferably a variable-wavelength LED device with controllable peak wavelength of light emitted, as described above, though in the illustrated embodiment the B and R subpixels could alternatively be provided by conventional blue and red LEDs.

(309) The display comprises a plurality of subpixels within the same colour (red, green or blue for example, as each of these colours is commonly accepted as extending over a range of wavelengths in the visible spectrum), with the plurality of subpixels configured to emit at different peak wavelengths within that colour.

(310) For example in preferred embodiments pixel colours may be: 400 nm-450 nm (violet); 450 nm-500 nm (blue); 500 nm-570 nm (green); 570 nm-590 nm (yellow), 590 nm-610 nm (orange), or 610-700 nm (red). Where there are a plurality of subpixels within a given colour, those subpixels may all be configured to emit at wavelengths somewhere in one of these ranges.

(311) In the preferred embodiment illustrated in FIG. 54A, the display pixel comprises two green sub-pixels with different peak wavelengths.

(312) The advantage of this pixel design is to expand the gamut of the display without the need to switch a single LED subpixel between G subpixel operating modes (as is required in FIGS. 52A-D.

(313) In the illustrated embodiment in FIGS. 54A and 54B, the peak emission wavelength of green (G) subpixel G1 is different from the peak emission wavelength of green subpixel G2.

(314) By extension there could (in addition, or alternatively) be two R or two B pixels with different peak wavelengths. For example the pixel could contain two red subpixels R1 and R2 having different peak emission wavelengths in the red. And/or the pixel could contain two blue subpixels B1 and B2 having different peak emission wavelengths in the blue.

(315) In all cases the LED devices may preferably have the same diode structure (N, active region, P) and only the drive current density is used to control the peak wavelength of emission. This may apply to all embodiments. Thus the two green subpixels G1 and G2 for example may be identical to one another in structure, but driven at different drive current densities. The difference in drive current densities leads to G1 emitting at a peak emission wavelength different from that of G2.

Embodiment 4 (FIGS. 55A-D)

(316) FIG. 55A schematically illustrates a pixel of a display device comprising two subpixels, where each subpixel is a variable-wavelength LED device with controllable peak wavelength of light emitted.

(317) When viewed at a distance the observed chromaticity of light emitted is a spatial and temporal combination of the sub-pixel light emission.

(318) Both subpixels are operated with controllable emission corresponding to a particular observed chromaticity.

(319) Both subpixels may preferably be controllable to emit at a peak emission wavelength from 450 nm to 630 nm, by varying the driving current provided to the subpixel. FIG. 55B illustrates the chromaticity achievable by the pixel when the peak emission wavelengths of the two variable-wavelength subpixels are varied from 450 nm to 620 nm.

(320) For any set of observed chromaticity points that can be displayed by the two subpixels, all chromaticities which exist on a straight line between the points can be displayed, as shown in FIG. 55C.

(321) The effective gamut of the display is shown in FIG. 55D, and is defined by the line corresponding to the chromaticity of the LED device when the peak emission wavelength is controlled from 450 nm to 630 nm, and the straight line between the highest and lowest peak emission wavelength.

Embodiment 5 (FIG. 56)

(322) FIG. 56 illustrates a pixel of a display formed with pixels comprising two subpixels. Each subpixel is preferably a variable-wavelength LED device with controllable peak emission wavelength, as described above, though in the illustrated embodiment the R subpixel could alternatively be provided by a conventional red LED.

(323) In FIG. 56, one subpixel has controllable emission between a range of wavelengths, for example from 450 nm to 530 nm (B to G). The other subpixel is operated with a fixed emission wavelength corresponding to fixed observed chromaticity, for example at 630 nm. Even if the R subpixel is a variable-wavelength LED, the peak emission wavelength of the R subpixel may be fixed by providing a driving current having a fixed magnitude corresponding to the current density required for emission at the desired red wavelength.

(324) This is particularly effective because higher drive current is needed to achieved blue emission which results in increased radiant flux, however the human eye is less sensitive to blue light and therefore luminous flux from a green-blue pixel will change less than expected as the peak wavelength shifts between 530 nm and 450 nm, enabling a more efficient display with decreased complexity to control.

Embodiment 6 (FIG. 57)

(325) FIG. 57 illustrates a pixel of a display formed with pixels comprising three subpixels, where each subpixel is a variable-wavelength LED device with controllable peak wavelength of light emitted.

(326) When viewed at a distance the observed chromaticity of light emitted is a spatial and temporal combination of the subpixel light emission.

(327) Subpixel SP1 is configured to emit light at wavelengths between 440 nm and 480 nm by varying the magnitude of the drive current supplied to SP1 during use. Subpixel SP2 is configured to emit light at wavelengths between 500 nm and 540 nm by varying the magnitude of the drive current supplied to SP2 during use. Subpixel SP3 is configured to emit light at wavelengths between 580 nm and 620 nm by varying the magnitude of the drive current supplied to SP3 during use. Thus SP1, SP2 and SP3 operate as blue (B), green (G) and red (R) subpixels respectively, and the peak emission wavelength of each subpixel can be varied across an emission wavelength range of 40 nm by varying the magnitude of the drive current to each subpixel.

(328) R, G and B can each operate in a number of modes with peak wavelength switchable by 20 nm lower and higher than a central wavelength.

(329) In combination each of the modes enables a different gamut of colours to be displayed

(330) As shown in FIG. 57, the effective gamut of the display is larger than that which can be achieved R, G and B sub-pixels having fixed emission wavelengths.

Generalisation of Embodiment 6 (FIG. 58)

(331) FIG. 58 illustrates a pixel of a display formed with pixels comprising three subpixels, where each subpixel is a variable-wavelength LED device with controllable peak wavelength of light emitted.

(332) When viewed at a distance the observed chromaticity of light emitted is a spatial and temporal combination of the sub-pixel light emission.

(333) Each variable-wavelength subpixel is controllable to emit across a wide emission wavelength range, so that each subpixel can be operated as a red, green or blue subpixel depending on the drive current provided to the subpixels. The subpixels are controllable to operate in a number of modes with different peak wavelength.

(334) In combination each of the modes enables a different gamut of colours to be displayed.

(335) As shown in FIG. 58, the effective gamut of the display is that which can be achieved with R, G and B sub-pixels in any mode.

(336) For each chromaticity within the achievable colour space there exists a continuous range of combinations of primary colours that enable this chromaticity to be displayed.

(337) When controlling the display device, the choice of which combination to choose is preferably made by calculating the efficiency of each combination and choosing the highest efficiency.

(338) The calculation of efficiency may advantageously take into account: Efficiency of light emission from each subpixel at a particular peak emission wavelength; Amount of light emission needed for each sub-pixel to achieve the chosen chromaticity and luminance; Efficiency of light extraction from the active region of the LED device to the observer; Efficiency of delivering electrical power from the display device driver to the LED device.
Temporal Colour Control

(339) In a typical display different colours cannot be shown at the same position. Therefore each light emitting area (pixel) is divided into separately addressable single-color regions (subpixels). When viewed at a distance the colour seen is a spatial combination of the sub-pixel colours. Subpixels emit a fixed chromaticity that is different to other subpixels.

(340) By adjusting the proportional amount of light emitted from each sub-pixel the observed chromaticity of a pixel is set, as illustrated in FIG. 50A.

(341) Typically three or more subpixels are required to enable the display to show a wide range of colours (large colour gamut). Reducing the number of sub-pixels is advantageous to reduce cost and complexity, however doing so will impact the achievable colour gamut.

(342) In some prior art display technologies not all coloured subpixels can be created in the same material, therefore significant cost and complexity is required to combine and arrange subpixels formed from different semiconductor materials, and to form the pixels into a display.

(343) Subpixels with differing emission properties will have different efficiencies resulting in non-uniform localised heating.

(344) Subpixels with higher efficiency will have shorter driving time (on time) than sub-pixels with lower efficiency, in order to achieve particular observed chromaticity. It is inefficient to have a system where a large quantity of sub-pixels are not emitting light for a significant proportion of the time.

(345) By incorporating the variable-wavelength LED of the present invention into a display device, there is provided a display where each pixel can emit a broad range of colours.

(346) Each pixel is comprised of a number of subpixels. In some embodiments, however, one single variable-wavelength LED may be used to form a pixel having only one subpixel as shown in FIG. 59A.

(347) The peak wavelength and therefore chromaticity of light emitted from a sub-pixel is dependent on the drive current. By choosing a drive signal where the drive current is changed during one display frame, the observed chromaticity of a pixel is determined by the temporal combination of the colours sequentially emitted by the one or more subpixels.

(348) Altering duty cycle of subpixels is used to achieve grayscale control at each chromaticity.

(349) The present invention may advantageously provide display with fewer than three subpixels that can display a wide colour gamut, which has significant benefits over prior art displays.

(350) Benefits of the invention and all advantages over existing solutions: Fewer sub pixels reduces the complexity and cost. The pixel can be emitting light at all times, increasing the system efficiency and therefore observed brightness The pixel will heat during high current operation, operating at low drive current allows the pixel to cool

Embodiment 7

(351) FIG. 59A illustrates a preferred embodiment in which the pixel of a display device consists of a single variable-wavelength LED. By varying the drive current provided to the pixel, the peak emission wavelength of the LED may be varied across a broad emission wavelength range, as discussed above, so that colours from blue to red may be emitted by the device pixel.

(352) Where there is a single pixel, duty cycle control is used to access different grayscale levels. The duration for which current pulses are applied to the pixel may be varied to control the observed pixel brightness.

(353) In the preferred embodiment illustrated in FIGS. 59B and 59C, the display device is configured to provide three discrete drive current modes to the LED pixel. The power supply can be operated to provide any one of I.sub.blue, I.sub.green or I.sub.red to the LED at any one time.

(354) The pixel colour observed by a viewer is the temporal combination of the light emitted by the pixel during a display frame. In FIG. 59B, each of the three drive modes is supplied for one third of a display frame, so the observed pixel colour will be an equal mix of blue, green and red wavelengths.

(355) FIG. 59C illustrates an alternative control mode, in which each of the three drive modes is supplied for one sixth of a display frame. As the duty cycles of each drive current are still equal to one another, the observed pixel colour will be an equal mix of blue, green and red wavelengths exactly like FIG. 59B. However, as the subpixel has a lower overall on-time in FIG. 59C, the brightness of the pixel will be lower using the control mode of FIG. 59B.

(356) FIGS. 60A-D show the three drive current modes I.sub.blue, I.sub.green or I.sub.red being provided to the LED for the same overall duty cycles, but in different orders. Where there is a single pixel, the four drive scheme variations illustrated in FIGS. 60A-60D give an equivalent result, as temporal averaging is experienced by an observer.

Embodiment 8

(357) In some preferred embodiments, a pixel may comprise a plurality of variable-wavelength subpixels which are configurable to emit at different peak emission wavelengths in response to different driving currents. For example a single subpixel may be drivable to emit at wavelengths from blue to red in response to variations in the driving current provided to that subpixel.

(358) FIG. 61A illustrates a display device pixel containing two subpixels, each of which is a variable-wavelength LED with an emission wavelength range encompassing blue to red wavelengths.

(359) Drive Scheme A in FIG. 61B illustrates a driving scheme for the upper subpixel, and a different driving scheme for the lower subpixel. Both the upper and lower subpixel are driven in the blue driving mode for one quarter of the display frame, in the green driving mode for one quarter of the display frame, and in the red driving mode for one half of the display frame. As the pixel colour observed by a viewer is the temporal combination of the light emitted by the pixel during a display frame, both the upper and lower subpixels will appear to emit the same colour of light.

(360) Drive Scheme B in FIG. 61C illustrates an alternative driving scheme which uses shorter current pulses to achieve the same result as Drive Scheme A. Despite the shorter current pulses, the upper and lower subpixels are still driven in blue, green and red modes for the same total duty cycles, so the observed colours will be the same.

(361) In all of these drive schemes, the pixel colour produced is the spatial and temporal combination of the emitted colours.

Embodiment 9

(362) In some preferred embodiments, a pixel may comprise a plurality of subpixels, at least one of which is a variable-wavelength subpixel which is configurable to emit at different peak emission wavelengths in response to different driving currents. For example a single subpixel may be drivable to emit at wavelengths from blue to red in response to variations in the driving current provided to that subpixel.

(363) In some particularly preferred embodiment, each pixel may comprise two sub-pixels in which: One sub-pixel is operated with fixed colour One sub-pixel changes colour

(364) The advantage of this arrangement is that a less efficient sub-pixel can have a longer on-time.

(365) FIG. 62 illustrates a display device pixel containing two subpixels. The upper subpixel is a variable-wavelength LED with an emission wavelength range encompassing blue to green wavelengths, while the lower subpixel is a red subpixel. The red subpixel may be a conventional red LED subpixel incorporated into the display device, but preferably the red subpixel is a variable-wavelength LED with an emission wavelength range encompassing red wavelengths.

(366) FIG. 62 shows that the upper subpixel may be driven by applying pulses of drive current in the blue mode I.sub.blue and green mode I.sub.green sequentially during the display frame, while the red lower subpixel may be driven by applying a continuous drive current in the red drive mode I.sub.red.

(367) This arrangement may advantageously provide a simplified RGB pixel.

(368) The indicated emission colours are by way of example only, as subpixels with fixed emission wavelengths in colours other than red may equally be provided, and subpixels with emission wavelengths controllable across ranges other than blue to green may be provided.

Embodiment 10

(369) In some preferred embodiments, a pixel may comprise a plurality of subpixels which are variable-wavelength subpixels which are configurable to emit at different peak emission wavelengths in response to different driving currents.

(370) In the embodiment illustrated in FIG. 63, for example, a single upper subpixel is drivable to emit at wavelengths from blue to red in response to variations in the driving current provided to that subpixel. Another lower subpixel is drivable to emit at wavelengths from green to red in response to variations in the driving current provided to that subpixel.

(371) In some particularly preferred embodiments, each pixel may comprise two sub-pixels in which a first variable-wavelength subpixel is drivable to emit at wavelengths across a first range in response to variations in the driving current provided to that subpixel (for example the driving current being controlled across a first driving current range); and a second variable-wavelength subpixel is drivable to emit at wavelengths across a second range in response to variations in the driving current provided to that subpixel (for example the driving current being controlled across a second driving current range). Preferably the first wavelength range covers a different range of wavelengths than the second wavelength range, though the first wavelength range may overlap the second wavelength range.

(372) In this embodiment both sub-pixels change colour.

(373) The advantage of this arrangement is that higher drive currents which generate significant localised heating are followed by lower drive currents which do not, allowing the sub-pixels to maintain a more stable temperature.

(374) FIG. 64A is a schematic illustration of driving current conditions for a variable-wavelength LED according to the present invention. Within a single display frame, the driving current is activated in three different non-zero modes. The duty cycle (the duration of a pulse of driving current relative to the display frame) of each mode is different, and individually controllable.

(375) The current gain of the driving current determines the wavelength produced by the variable-wavelength subpixel, while the duty cycle of the driving current determines the grayscale level produced by the subpixel.

(376) The length of a display frame may be varied to correspond to any predetermined frame rate. The length of a display frame may be controlled by controlling the LED driving conditions provided by the power supply.

(377) Where there is a single LED forming a pixel, duty cycle control is used to access different grayscale levels.

(378) The following embodiments illustrate a variety of possible driving conditions usable to control a display device comprising one or more variable-wavelength LED subpixels according to the present invention.

(379) These methods of control advantageously allow dynamic pixel tuning of the LEDs in the display device.

Embodiment 11 (FIG. 65)

(380) Single pixel can be one subpixel or more, and each subpixel can change colors by the image information (this subpixel can display colors from blue to red)

(381) Each pixel can determine its own color by modulating the signal and produce image information

(382) Pixel color and brightness are a combination of signal pulse duty cycle and amplitude

Embodiment 12 (FIG. 66)

(383) Single pixel can be three or more subpixels (can display colors from blue to red).

(384) The subpixel size will be the same.

(385) Each subpixel can determine its own color by modulating the signal, and Image information for each pixel can be determined by combining multiple sub-pixels.

(386) Pixel color and brightness are a combination of signal pulse duty cycle and amplitude.

(387) Where there is a single pixel, these drive scheme variations are equivalent as temporal averaging is experienced by an observer

Embodiment 13 (FIG. 67)

(388) Single pixel can be three or more subpixels (can display colors from blue to red).

(389) The chip size is changed to adjust the magnitude of the current and pulse width.

(390) Each subpixel can determine its own color by modulating the signal, and Image information for each pixel can be determined by combining multiple sub-pixels.

(391) Pixel color and brightness are a combination of signal pulse duty cycle and amplitude.

Embodiment 14 (FIG. 68)

(392) Single pixel can be three or more subpixels (can display colors from blue to red).

(393) Chip sizes are optimised to achieve the same driving mode for each subpixel.

(394) Each subpixel can determine its own color by modulating the signal, and Image information for each pixel can be determined by combining multiple sub-pixels.

(395) Pixel color and brightness are a combination of signal pulse duty cycle and amplitude.

Embodiment 15 (FIG. 69)

(396) Some sub-pixels can focus on producing specific colors through signal modulation, while others can change colors as needed by the image information (can display colors from blue to red).

(397) The subpixel size can be the same or different.

(398) Each subpixel can determine its own color by modulating the signal, and image information for each pixel can be determined by combining multiple sub-pixels.

(399) Pixel color and brightness are a combination of signal pulse duty cycle and amplitude.

(400) Spectral Reconstruction

(401) In many applications it is desirable to be able to reproduce a particular spectrum of light.

(402) In the prior art, this is achieved by: Modulating the intensity of one or more illumination sources with fixed broad band emission spectra. This is inherently inefficient due to the subtractive nature. Combining the emission from multiple narrow band emission sources. This provides limited tunability due to the fixed number of emission sources. Using filters to modify the light from a high power broad spectrum source. This suffers from inherent inefficiency, and limited tunability due to fixed number of tuning elements.

(403) As illustrated in FIGS. 70 to 79, an aspect of the present invention relates to utilising a variable-wavelength LED emission source in which the emission wavelength can be continuously tuned through control of the applied current.

(404) The variable-wavelength LED source may be driven sequentially with a number of different current pulses. Different current pulses may have different magnitudes, or amplitudes. The time at which a particular current pulse is applied may be different to the time at which other current pulses are applied to the variable-wavelength LED.

(405) As the emission wavelength emitted by the variable-wavelength LED correlates to the driving conditions applied to the LED at any given time, each applied current pulse with a different amplitude will produce a different peak emission wavelength.

(406) In a preferred embodiment, the total length of pulses applied at different current amplitudes is faster than the response time of the detector (50 ms for the human eye as a detector). In that case the emission spectrum perceived by the detector (preferably the human eye) will be a temporal average of the spectrum emitted by the variable-wavelength LED, that is a temporal average of the emission spectra created by each current pulse.

(407) Current control can be done either using an analogue or digital pulse form.

(408) FIGS. 70 to 72 illustrate an example of a digital pulse comprising a discrete sequence of current pulses.

(409) FIG. 70 illustrates the emission spectrum of a variable-wavelength LED according to a preferred aspect of the invention, overlaid with lines indicating five discrete driving currents I.sub.1-I.sub.5. Each of driving currents I.sub.1-I.sub.5 differs in magnitude from the other driving currents. The intersection between the overlaid lines I.sub.1-I.sub.5 and the LED's emission spectrum indicates the peak emission wavelength emitted by the variable-wavelength LED in response to each of discrete driving currents I.sub.1-I.sub.5.

(410) FIG. 71 illustrates an exemplary sequence of the current pulses being applied to the variable-wavelength LED. Each of I.sub.1-I.sub.5 has its own discrete magnitude (amplitude), and thus produces its own discrete peak emission wavelength when applied to the variable-wavelength LED. The sequence in which the pulses are applied to the LED, as well as the temporal duration of each current pulse, will therefore determine the overall emission spectra produced by the LED within the time of a given display frame. The order and duration of the current pulses may be controlled to achieve a huge variety of different perceived emission spectra.

(411) FIG. 72 illustrates the combination of five discrete emission spectra having different peak emission wavelengths, which correspond to the five spectra produced by the five current pulses of FIG. 70. The overall output spectrum which is perceived by a viewer is the combination of these five discrete spectra, as shown by the output line in FIG. 71.

Embodiment A

(412) FIGS. 73A-73D illustrate spectral reconstruction using multiple current set points (digital pulses), using the variable-wavelength LED described above.

(413) A plurality n of current set points (n=5 in the illustrated example of FIG. 73A) are selected across tuning range of the LED (the range of peak emission wavelengths over which the variable-wavelength LED can emit).

(414) A target spectrum (shown in FIG. 73B) is reconstructed as a linear combination of the Emission spectra of the LED at the chosen current set points (shown in FIG. 73C). The intensity of each constituent peak wavelength is converted into a time (pulse duration) at each of the n set currents to account for the LED emission brightness at each set current, to arrive at a pattern of digital pulses which completes within the duration of a display frame (illustrated in FIG. 73D). In use, the pulse pattern is repeated for as long a time as the target emission spectrum is to be displayed, until a change in output spectrum is desired. At that time the LED may be driven with a different pulse pattern to produce a different perceived spectrum.

Embodiment B (FIGS. 74-77)

(415) In the event where the number of current set points is large (.fwdarw.) the spectral reconstruction can be considered using an analogue current pulse.

(416) To recreate a particular desired emission spectrum a total current pulse can be calculated such that the total emitted light at each wavelength in the tuning range of the variable-wavelength LED matches the desired target emission spectrum.

(417) An exemplary target emission spectrum is shown in FIG. 74.

(418) In order to produce the target emission spectrum of FIG. 74, a variable-wavelength LED with the emission characteristics illustrated in FIG. 75 may be driven using an analogue current pulse in which the amplitude of the current pulse varies over time. FIG. 76 shows an illustrative example of an analogue pulse of driving current with an amplitude which varies during a display frame. When such an analogue pulse is used to drive the variable-wavelength LED, the LED produces different peak emission wavelengths as the amplitude of the driving current pulse varies within the display frame, resulting in an output spectrum such as that shown in FIG. 77.

(419) Similarly to the digital pulses described above in Embodiment A, the target spectrum shown in FIG. 74 is reconstructed as a temporal combination of emission spectra of the LED, with the analogue driving pulse behaving like a very high number (n.fwdarw.) of constituent driving currents.

Embodiment C (FIG. 78)

(420) One or more variable-wavelength LEDs may provide a large area spectrally tuneable illumination source for use as a hyperspectral light source or for general illumination. This may be achieved using the concepts of Embodiments A and B described above.

(421) The spectra produced can either be narrow or broad spectrum, depending on the driving conditions applied to the one or more LEDs.

(422) The illumination spectrum can either be fixed between illumination frames (by keeping the driving conditions the same, or repeating the same driving pulse during each display frame) or modified between frames (by modifying the pulse pattern/sequence or pulse shape between frames), where the illumination frame is a time shorter than the response time of the detector used.

(423) This may be particularly useful for applications requiring a particular controlled spectrum including medical imaging, phototherapy, specialty lighting, agritech.

Embodiment D (FIG. 79)

(424) In an aspect of the invention may provide one or more spectrally tuneable pixels for a spectrally correct display, where each individual pixel (or each individual subpixel) behaves as described above in Embodiment A or B.

(425) As shown in FIG. 79, a plurality of variable-wavelength LEDs (tuneable LEDs) may be provided in an array to form a display device. Each variable-wavelength LED is preferably arranged to form a pixel (or a subpixel) of the display device. The pixels of the display device may each be configured to receive their own driving current from a pulse current source which is configured to generate current pulses (either series of digital pulses as described in relation to embodiment A, or analogue pulses as described in relation to embodiment B). The driving current provided to each pixel is preferably controllable independently from that applied to the other pixels. The pulse current source may be configured to provide pulses of driving current to a multiplexer which is connected to the individual pixels in the display device.

(426) The current pulses can be designed for two main applications: To accurately represent the spectrum of the image being recreated; or To correct for non-uniformity arriving during the manufacturing process by adjusting the current pulse for each sub pixel such that they all emit identical emission wavelength and intensity across the entire display. The tuned display would then be driven with the adjusted current pulses with the emission from two or more sub pixels combining to produce the perceived colour for each pixel.
Pixel Size & Geometry

(427) The embodiments set out below illustrate aspects of the invention in terms of pixels comprising two subpixels with different areas and different emission wavelengths. The skilled person will understand, however, that pixels having other numbers of subpixels may equally be provided using the same principles.

(428) Embodiments E to H use the same variable-wavelength LED structures described throughout the present application. By controlling the light-emitting area of the LED mesas during manufacture, the current densities experienced by the mesas can be controlled. As the current density governs the emission wavelength of a given LED, as described above, by controlling the light-emitting area of the LED mesas different subpixels with identical diode structures can be made to emit at different wavelengths in response to the same drive current.

(429) FIG. 80A is a plan view of a display pixel having two variable-wavelength subpixels with the same diode structure but different sizes, which emit at different peak emission wavelengths in response to the same fixed-magnitude driving current. Separate variable-wavelength LEDs having different areas but otherwise identical diode structures (identical layered LED structures of n-type layer(s), active layers and p-type layers) will thus emit different peak emission wavelengths in response to the same absolute drive current. In the illustrated embodiment, the larger subpixel emits at a red peak emission wavelength .sub.1, while the smaller subpixel experiences a higher current density and emits at a shorter green peak emission wavelength .sub.2. Together, the red and green subpixels form a combined red-green (RG) pixel package which may be integrated onto a device back plane driver as shown in FIG. 80B.

(430) In FIG. 80B an array of combined red-green (RG) pixel packages is mounted on a back plane driver integrated circuit, with an array of blue LED subpixels. Together, each red-green (RG) pixel package combines with a blue subpixel to form a RGB pixel of a display device.

Embodiment E

(431) LED geometry design to achieve multiple wavelengths at specified current densities.

(432) For a given absolute current applied to an LED (or an LED subpixel), the current density experienced by the LED will depend on the area of the LED. The peak emission wavelength of a variable-wavelength LED can be varied by controlling the current density applied to the LED. The size, area and geometry of the LED will therefore affect the current density that results from any given absolute current, and thus the peak emission wavelength that is emitted.

(433) For a display device configured to apply a fixed drive current I, for example, providing a plurality of LEDs having light-emitting layers with different areas will result in those LEDS emitting at different peak emission wavelengths. Although the absolute drive current applied to the different LEDs may be the same, the different surface areas of the LEDs will mean that different LEDs experience different drive current densities, which will result in emission at different peak emission wavelengths.

(434) Thus even when different LED subpixels have the same diode structure (the same arrangement and composition of semiconductor layers), by varying the light emitting area of different subpixels, those subpixels may be configured to emit at different peak emission wavelengths even when all of those subpixels receive the same absolute drive current.

(435) By providing display devices configured to apply more than one fixed drive currents to the plurality of LEDs, the variety of peak emission wavelengths emittable by the LEDs may be increased. For example the drive current may be switchable between two different drive current modes, such that each LED subpixel is driveable to emit at two different peak emission wavelengths. The absolute value of the peak emission wavelengths are determined by the light-emitting area of the individual subpixels, and the magnitude of the drive currents in the drive current modes

(436) Using the drive current density vs wavelength curve shown in FIG. 81 (for an exemplary variable-wavelength LED) the required Current density for a given peak emission wavelength can be found.

(437) From this, the pixel surface area for each wavelength can be calculated using:

(438) $A_i=\frac{I_i}{{}_i}$

(439) LED mesas with the corresponding area required for a desired peak emission wavelength can then be etched into the LED wafer, leaving multiple LEDs with varying emission wavelengths connected to a common substrate.

(440) FIGS. 82A and 82B illustrate an exemplary device pixel comprising two subpixels with the same diode structure but different areas, which emit at different peak emission wavelengths in response to the same driving current.

(441) In the example illustrated in Figured 82A and 82B, a first subpixel mesa (Mesa 1) forms a first subpixel having a first area A.sub.1, and a second subpixel mesa (Mesa 2) forms a second subpixel having a second area A.sub.2. The same drive current may be applied to both subpixels, but the difference in subpixel area means that the two subpixels will experience different current densities. These different current densities will drive the two subpixels to emit light at different wavelengths, even if the first and second subpixels have identical diode structures.

(442) The pixel of FIGS. 82A and 82B is preferably manufactured by growing an LED structure on a semiconductor template comprising a porous region of III-nitride material. The LED structure may be grown as a uniform diode structure across the entire semiconductor wafer.

(443) The wafer-scale LED structure may then be processed into a multi-colour display device, or into a plurality of multi-colour LED pixels to be separated and integrated into display devices, as follows.

(444) In order to provide LED pixels each containing two LED subpixels which emit at different peak emission wavelengths from one another in response to a fixed drive current I.sub.1, the two desired peak emission wavelengths .sub.1, .sub.2 are chosen. Based on the characteristic relationship between peak emission wavelength and drive current density for the grown LED structure, the drive current densities .sub.1 and .sub.2 required for each subpixel to produce the desired peak emission wavelengths are calculated. Based on the fixed drive current I.sub.1 that will be used to drive the display, the formulae set out above are used to calculate the LED areas A.sub.1, A.sub.2 which will result in the two subpixels experiencing drive current densities .sub.1 and .sub.2.

(445) Conventional semiconductor etching techniques are used to etch the LED structure into a plurality of pixels, in which each pixel contains two mesas of the LED diode structure which form two discrete LED subpixels: a first mesa having area A.sub.1, which will experience drive current density .sub.1 when drive current I.sub.1 is applied, and will thus emit at a desired peak emission wavelength .sub.1; and a second mesa having area A.sub.2, which will experience drive current density .sub.2 when drive current I.sub.2 is applied, and will thus emit at a desired peak emission wavelength .sub.2.

(446) Drive currents I.sub.1 and I.sub.2 provided to the two subpixels may be the same drive current, or they may be different in magnitude. Drive currents I.sub.1 and I.sub.2 may be fixed-magnitude drive currents, so that the subpixels act as fixed-wavelength LEDs in use. Alternatively, Drive currents I.sub.1 and I.sub.2 may be variable-magnitude drive currents, such that the variable-wavelength subpixels can be driven to emit at a range of different wavelengths.

(447) As the two LED subpixel mesas are grown on the same substrate, both are positioned on a shared n-type substrate. N-contacts and p-contacts are applied according to conventional methods, so that the two subpixels are driveable separately.

(448) Although the description above relates to pixels comprising two subpixels, the invention may apply to pixels comprising any number of subpixels, and the same process may be used to form LED pixels comprising different numbers of subpixels.

Embodiment F

(449) LED geometry design to produce a given emission Intensity at given emission wavelengths.

(450) Using the drive current density vs wavelength curve shown in FIG. 81 (for an exemplary variable-wavelength LED) the required Drive Current Density (.sub.1, .sub.2) is selected to achieve two desired emission wavelengths (.sub.1, .sub.2).

(451) For a given LED diode structure, the relationship between emission efficiency and emission wavelength is known or can be easily characterised, as shown in FIG. 83.

(452) The Emission Efficiency at a given wavelength is given by .sub.i.

(453) In this embodiment: 1. Wavelength is selected and used to determine drive current density ; 2. Emission efficiency at (or ) is used to scale the emitting area A of the sub-pixel to achieve a desired output density (emission intensity); 3. The drive current that is required is found from Area.

(454) Thus once an emission wavelength is selected, the current density is fixed, and then the sub-pixel area is scaled according to efficiency at that wavelength, in order to achieve a required total flux (intensity) from the sub-pixel.

(455) Using this method, it can be ensured that subpixels emit not only at the intended wavelength, but also at a desired emission intensity.

(456) At the desired emission wavelengths (.sub.1, .sub.2) for a two-subpixel pixel, the drive current densities (.sub.1, .sub.2) required to emit with the desired emission wavelengths (.sub.1, .sub.2) is known. The Emission Efficiency at those wavelengths is given by .sub.i. By taking into account the emission efficiency, the light-emitting areas (A.sub.i) required for the two subpixels to emit at the desired emission intensity is calculated.

(457) From the subpixel areas (A) and the drive current densities, the total drive Current required (I.sub.i) to give the desired emission intensity is calculated.

(458) As described above in relation to Embodiment E, mesas may be etched into an LED structure to give mesas having the correct light-emitting area (A.sub.1, A.sub.2) that will emit at the desired peak emission wavelengths (.sub.1, .sub.2). As both subpixels are etched from the same LED structure, both subpixels have the same diode structure, but different areas (footprintsthe area of a subpixel is the surface area of the mesa when viewed from above).

(459) As the drive current (I.sub.i) required to give a desired emission intensity has been calculated separately for each subpixel, the subpixels will emit at the desired emission intensity whenever the drive current is applied.

(460) The drive current (I.sub.i) applied to each separate subpixel may optionally be different. This may advantageously compensate for the difference in emission efficiencies at different emission wavelengths. Thus a first subpixel (Mesa 1 in FIGS. 82A & 82B for example) may be driven by a first drive current I.sub.1 to produce the emission intensity desired from the first subpixel, while a second subpixel (Mesa 2 in FIGS. 82A & 82B for example) may be driven by a second drive current I.sub.2 to produce the emission intensity desired from the second subpixel.

Embodiment G

(461) LED geometry design to produce a given Luminosity at given emission wavelengths

(462) Using the drive current density vs wavelength curve shown in FIG. 81 (for an exemplary variable-wavelength LED) the required Drive Current Density (.sub.1, .sub.2) is selected to achieve two desired emission wavelengths (.sub.1, .sub.2) from two separate LED subpixels.

(463) For a given LED diode structure, the relationship between emission efficiency and emission wavelength is known or can be easily characterised, as shown in FIG. 83.

(464) At the design drive current densities required to emit with the desired emission wavelength, the Emission Efficiency is given by .sub.i.

(465) This can then be combined with the Photopic Luminosity function. For a given LED diode structure, the photopic luminosity function for emission at different wavelengths is known or can be easily characterised, as illustrated for an exemplary subpixel in FIG. 84.

(466) From this the total Current required (I.sub.i) for an LED subpixel to give a desired emission luminosity can be calculated.

(467) Finally from this required current (I.sub.i) and the current density, the formula above can be used to calculate a Mesa area for an LED subpixel such that the requirements are met and the subpixel will emit light at the desired peak emission wavelength and the desired luminosity when the drive current is applied.

(468) Similarly to the embodiments described above, a plurality of LED subpixels having different areas (footprints) may be formed in the same semiconductor wafer by etching the wafer into a plurality of subpixel mesas having mesa areas corresponding to the desired subpixel areas.

Embodiment H

(469) LED geometry Optimised for Concentric emission of two or more wavelengths.

(470) FIGS. 85A and 85B illustrate an exemplary embodiment of a two-subpixel device pixel in which the two subpixels are provided in the form of a ring-shaped subpixel 1 positioned concentrically around a circular subpixel 2. Both subpixels are provided on a shared conductive substrate and connected to a common n-contact, but subpixels 1 and 2 are connected to separate p-contacts through which the subpixels are bonded to a driver backplane IC.

(471) FIGS. 85A and 85B show two concentric sub pixels, however the number of emitting wavelengths can be as large as required, with each subsequent sub pixel forming a ring centered around the smallest sub pixel.

(472) The additional freedom in controlling the layout of the devices in a single layer rather than mass transfer allows for: 1. Correction of chromatic aberration of imaging optics; and 2. Generation of multi colour pixels with the same apparent centre.

(473) The principles set out above may be used to calculate the areas required for different subpixels to emit light at the desired wavelengths. Once those subpixel areas are known, the subpixels may be formed in a wide variety of shapes, with a variety of subpixel geometries. In the example illustrated above, for example, subpixels may be formed as concentric rings, one around the other. This concentric arrangement may be converted into a display device by etching concentric circles having the desired footprint area into an LED wafer comprising a shared n-type conductive layer, and then flipping the wafer before bonding p-contacts to connect each individual subpixel to a driver backplane IC. The driver backplane IC is configured to apply a drive current, or a plurality of different drive currents, to each subpixel, such that the discrete subpixels can be controlled separately from one another.