KR102712312B1 - Selective collection system and method for light-emitting elements - Google Patents

Abstract

A method for selectively collecting microLED elements from a carrier substrate is provided. The carrier substrate includes a predetermined defect region of a plurality of adjacent defective microLED elements. A content-soluble colloidal material covers the predetermined defect region, and the exposed adhesive is dissolved by an adhesive dissolving solvent. Defect-free microLED elements located outside the predetermined defect region are separated from the carrier substrate, while at the same time maintaining adhesion between the microLED elements and the carrier substrate within the predetermined defect region. In addition, a method for distributing microLED elements in a light-emitting display panel is provided by optically measuring a microLED suspension to determine the suspension uniformity and calculate the number of microLEDs per unit volume. Knowing the number of microLED elements collected in the suspension, the number of microLED elements per unit volume of the suspension can be calculated.

Description

Selective collection system and method for light-emitting elements

The present invention relates generally to a light emitting diode (LED) having a size of less than 100 μm, and more specifically, to a manufacturing process and system for a micro-light emitting diode (microLED).

Red/Green/Blue (RGB) displays consist of pixels that emit light at three wavelengths corresponding to the visible colors red, green, and blue. The RGB components of the pixel (each called a sub-pixel) are excited in a systematic manner to produce colors in the visible spectrum. There are several types of displays that produce RGB images in different ways. The most widely used technology, liquid crystal displays (LCDs), produce RGB images by irradiating a white light source (usually white LEDs generated by phosphor powder) and passing the RGB images through color filters in the sub-pixels. Some of the white light wavelengths are absorbed and some are transmitted through the color filters. Therefore, the efficiency of LCD displays can be less than 4%, and the contrast ratio is limited by light leakage through the liquid crystals. Organic light-emitting diode (OLED) displays produce RGB light by directly emitting light at each of these wavelengths at the pixel level within the organic light-emitting material. Although OLED materials emit light directly, they display high contrast ratios, and thus the organic materials can deteriorate over time, resulting in image burn-in.

A third display technology and one described herein is a microLED display, which uses micro-sized (5 to 100 microns (μm) in diameter) inorganic LEDs to directly emit light at the sub-pixel level. Inorganic microLED displays have several advantages over competing displays. MicroLED displays have very high contrast ratios, exceeding 50,000:1, and higher efficiency compared to LCD displays. Inorganic LEDs can achieve much higher brightness without burn-in, unlike OLED displays.

MicroLEDs are fabricated from metalorganic chemical vapor deposition (MOCVD) wafers similar to the wafers used to fabricate LEDs for general lighting, resulting in a very low cost per device, but some inherent challenges to microLED technology arise. The structure for fluidic assembly of microLEDs is described in detail in parent patent US10,643,981, which is incorporated herein by reference. For general lighting applications, the most important characteristic of the device is that each photon be generated at a low cost, so as to minimize the cost per bulb.

Due to these limitations, LED manufacturing has dealt with process variability and defects using a process called binning. Simply put, the binning process tests all LEDs after packaging and groups LEDs with similar efficiency and emission wavelength characteristics into the same group, while discarding defective devices. The binning process reduces defects and minimizes process control costs, making MOCVD manufacturing more cost-effective.

Recent characterization of 40 μm microLEDs fabricated with conventional gallium nitride (GaN) on MOCVD wafers showed that 0.25% of the devices were shorted and 0.75% were opened. These defects result in non-luminous pixels, making them unacceptable for display applications. The microLEDs are unpackaged and their device size, especially the electrodes, is very small, making device processing and functional testing difficult. Since an ultrahigh definition (UHD) display requires at least 24.8 million (3 × 3840 × 2160) microLEDs, testing times are astronomical. Therefore, conventional binning techniques are impractical for identifying and discarding defective microLEDs. Novel structures and methods are needed to prevent defective microLEDs from generating defective sub-pixels. As described in parent applications Ser. Nos. 16/125,671, 16/595,623 and 16/693,674, which are incorporated herein by reference, it is possible to remove and replace defective microLEDs, but purchasing and operating mechanical pick and place tools is costly. It would be more desirable to identify defective microLEDs and prevent them from entering the suspension for fluidic assembly.

Because LEDs used in general lighting are much larger than those used in microLED displays (each side is up to 3–4 mm and the diameter is 5–100 μm), their demands on patterns and electrodes are significantly different. MicroLEDs are bonded to substrate electrodes using solder or anisotropic conductive films, whereas larger, general-purpose light-emitting LEDs are typically bonded to lead frames by wire bonding or solder paste. Because microLEDs are so small, the technologies used to process the devices, especially those used to assemble microLED displays, are very different from those developed for ultra-large LEDs used in general lighting.

To fabricate a microLED display, green and blue GaN microLEDs are fabricated on a sapphire substrate, and red aluminum gallium arsenide phosphide (AlGaAsP) microLEDs are fabricated on a GaAs substrate. After fabrication and segmentation, the microLEDs must be transferred to a second substrate, which becomes the light-emitting display. The second substrate can be a silicon (Si) wafer (or chip) with built-in control circuitry, or a glass or flexible plastic substrate with thin film transistors. A conventional transfer method is a mechanical pick-and-place system that uses a pick head to capture and position the devices on the display substrate. Another mechanical transfer method that simultaneously transfers blocks of microLEDs using a stamp or the like is called mass transfer. An alternative technique, as described herein, is to position the microLEDs using fluidic assembly.

In simple terms, the fluidic assembly process involves applying a microLED suspension to a substrate having an array of capture trap sites (wells), and moving the suspension so that microLEDs assemble at the trap sites. For successful fluidic assembly, defect-free microLEDs must be collected from the growth substrate, a microLED suspension with a known concentration must be prepared, and the suspension must be uniformly distributed onto the display substrate.

The treatment of micron-sized particle suspensions has gained wide acceptance in systems such as cell culture in life sciences or slurry grinding in industrial applications. In all cases, the goal of a suspension treatment system is to achieve a highly uniform suspension and to transfer the suspension to the target process with a high degree of volume and concentration control. Suspension uniformity is usually achieved by direct mechanical mixing using submerged impellers or by active circulation by pumps. Transfer of the well-mixed suspension to the target process is usually accomplished by means of a pipe pumped downstream of the feed tank or by pressurizing a sealed tank. Volume control in a good suspension system is achieved by controlling the flow rate by differential pressure and by metering the net flow rate with valves. Since it may be desirable to control the suspension concentration, especially when reusing the suspension, the transfer pipe usually includes a component with multiple inlets to allow a pure carrier liquid to equilibrate with the suspension.

Unfortunately, existing suspension processing technologies are not compatible with the characteristics of microLEDs or the requirements of fluidic assembly technology. Specifically, microLED suspensions have the salient features described below, and alternative methods must be developed.

Figure 1 shows a microLED suspension with a uniform distribution after stirring, which settled to about half the height of the liquid column after time t1 and settled completely after time t2. Unlike abrasive slurries, which are formulated to have prolonged uniform mixing and have settling times measured in months, the microLED suspensions are formulated to have relatively short settling times. Since microLEDs can only be assembled by fixing them to the target substrate surface, microLED suspensions for fluidic assembly typically set completely within minutes and lose their uniformity shortly after mixing is stopped. As an example, a 42 μm diameter, 5 μm thick disk-shaped microLED fabricated according to the center table design described in parent application Ser. No. 16/406,080 (incorporated herein by reference) has a hydrodynamic diameter of 18.9 μm. When gravity is in equilibrium with the viscosity of the liquid, an object of this size has a terminal velocity in the liquid, as given by:

Here, D _LED is the hydrodynamic diameter, ρ is the liquid density, ρ _LED is the density of the microLED, and μ is the liquid viscosity. For water, the terminal velocity is 1.1 mm/sec, so in a typical vessel such as a 50 milliliter (ml) Falcon tube, the microLED completely sinks in about 1 minute.

MicroLEDs generally have surfaces including metals, inorganic materials and organic materials. Therefore, it is almost impossible to prevent temporary sticking of the solid surface in contact with the microLED suspension. Therefore, the container containing the microLED suspension is generally made of hydrophobic materials such as acetal homopolymer, polytetrafluoroethylene (PTFE), polypropylene, etc., to minimize the stickiness. The final state in Fig. 1 shows the interaction effect between the microLED and the container wall, where the bottom cone is smaller than the angle of repose, so that some microLEDs stick to the container wall and do not settle in a uniform layer on the bottom.

Microorganisms in biological applications are generally robust enough to withstand internal mixing (e.g., using a stir bar), whereas industrial abrasive suspensions (e.g., chemical mechanical polish (CMP) slurries) can be suspended by drum circulation or impeller mixing without damage. In contrast, microLEDs are fragile and can be ruptured by direct mechanical mixing or pumping. Damaged microLEDs cannot be removed from suspension by filtration because their primary dimensions are similar to those of intact microLEDs, and they can partially block capture trap sites, interfering with fluid assembly.

MicroLEDs account for a significant portion of display manufacturing costs, and the low utilization and recovery efficiency of μLEDs have a significant impact on cost. The components in suspension are more valuable than in biological and industrial abrasive applications.

Unlike conventional suspensions, since each element constitutes a sub-pixel, the performance characteristics of individual microLEDs are very important. The number of microLEDs that can be used for assembly must be strictly controlled to control the emission distribution of the entire display. Therefore, the suspension processing must be designed to prevent cross-contamination.

The characteristics of microLEDs and the stringent demands on display fabrication preclude conventional industrial systems and suspension handling methods. Controlled and efficient dispensing of clean, high-quality assemblies is critical for fluidic assembly, as the forces involved in fluidic assembly are limited by the threshold for separation of the assembled assemblies. Furthermore, rapid fluidic assembly relies on a short transport path on the substrate between the microLEDs and the final assembly (capture trap) location. Therefore, optimal dispensing of the microLED suspension on the display substrate must be fast and highly uniform, as well as loss-free and damage-free.

It would be advantageous if a method existed for collecting and dispensing inorganic microLEDs dedicated to fluid assembly processing.

This specification describes systems and methods for preparing and manipulating suspensions of micro light emitting diodes (microLEDs) applicable to the fluidic assembly of microLED displays. An optional collection method produces a microLED suspension comprising known good LEDs at a determined concentration in a suitable liquid. A dispensing system is used to deliver the microLED suspension, which minimizes damage and loss of the microLEDs while uniformly distributing the devices at a controlled density on a display substrate. These optimal initial conditions are critical to the successful fluidic assembly of microLED displays.

Accordingly, a method for selectively collecting microLED devices from a carrier substrate is provided. The method provides an inorganic microLED device adhesively attached to the carrier substrate. The defect region is a predetermined region (e.g., a wafer edge) comprising a plurality of adjacent defective microLED devices. A content-soluble colloidal material covers the predetermined defect region, and the exposed adhesive is dissolved by an adhesive dissolution solvent. Some embodiments of the adhesive dissolution solvent include acetone, toluene, trichloroethane, N-methylpyrrolidone (NMP), xylene, cyclohexanone, butylacetate, or combinations thereof.

A defect-free microLED element located outside a predetermined defect region is separated from a carrier substrate, while at the same time maintaining adhesion between the microLED element and the carrier substrate within the predetermined defect region. In response to the separation of the microLED element from the carrier substrate, the valid microLED element is collected in a collection vessel. In one variation, only a portion of the carrier substrate is exposed to the adhesive dissolving solvent, such that the microLED is separated only from the selectively exposed portion of the carrier substrate.

Additionally, the carrier substrate can be inspected to locate defective microLED devices in undetermined defect regions, and furthermore, a content-soluble colloidal material can be formed in these undetermined defect regions. In one embodiment, the inspection can locate individual defective microLED devices in undetermined defect regions, and a laser trimming process can be used to release the individual defective microLED devices. The inspection can be performed using optical comparison, electroluminescence, photoluminescence, or cathodoluminescence testing.

The microLED devices collected in the collection vessel are typically a suspension of valid microLED devices having an average cross-sectional size s. However, impurities also exist in the suspension. In one embodiment, a filtration step is performed using mechanical screening, elution, classification, or a combination thereof to remove impurities having a maximum cross-sectional size greater than t, where t>s. Similarly, a separate filtration step is performed to remove impurities having a maximum cross-sectional size less than p, where p<s. In one embodiment, prior to filtration, the adhesive dissolution solvent in the collected microLED suspension is replaced with a filtration solution having a lower viscosity than the adhesive dissolution solvent. Optionally, or after filtration, the fluid in the collected microLED suspension can be replaced with an assembly solution having a lower polarity or a faster evaporation rate. In one embodiment, a surfactant, such as an anionic, a cationic, a nonionic surfactant, or a combination thereof, can be added.

Also provided is a method of dispensing microLED elements on a light-emitting display panel. A suspension of the collected microLED elements is transferred to a transparent first vessel and stirred. Some examples of the stirring process include external vibration of the first vessel, generation of fluid flow in the suspension, and flow of gas in the first vessel. The opacity of the suspension is optically measured at a plurality of first vessel heights to determine the uniformity of the suspension. The determined uniformity is greater than a uniformity minimum threshold, and the suspension can be dispensed on the uppermost surface of the light-emitting display panel. Some examples of the dispensing process include single-step bulk decantation, multi-step pipette translation, nozzle-limited vessel translation, and pipetting.

Knowing the number of microLED elements collected in the suspension, the number of microLED elements per unit volume (volume) of the suspension can be calculated. As a result, in response to dispersion of the first volume of the suspension, a known first number of microLED elements can be deposited on the light-emitting display panel. Advantageously, after determining the number of assembly locations within the first region of the uppermost surface of the light-emitting display panel, the number of the first known number of microLED elements deposited is at least equal to the number of assembly locations within the first region.

An optical measurement of the opacity of the suspension is performed by arranging a plurality of light-emitting elements having a predetermined output light intensity, which are spaced apart from each other by a first predetermined distance along a first vertical axis and directed toward the central axis of the first vessel. A plurality of photodetectors are spaced apart from each other by a first predetermined distance along a second vertical axis, each photodetector having an input terminal directed toward a corresponding light-emitting element output terminal. The light intensities received by the photodetectors are then compared.

In one embodiment, in response to determining the opacity of the suspension at a plurality of first container heights, a first number of microLED elements per unit volume of the suspension can be calculated. An equal volume of the suspension can be dispersed on a top surface of the light-emitting display panel, and then the optical measurement can be repeated to calculate a second number of microLED elements per unit volume of the suspension. A known equal volume of the suspension can be transferred to a second container, and a predetermined amount of fluid can be added to (or removed from) the second container, and then a third number of microLED elements per unit volume of the suspension in the second container can be calculated. A fluid can be added to (or removed from) the suspension in the first container, and then the suspension can be exchanged, and then the density of the suspension can be measured again optically to calculate a fourth number of microLED elements per unit volume of the suspension.

Other details of the above method, as well as the system for isolating regions of the microLED carrier substrate, the system for collecting the microLEDs, and the system for characterizing the microLED suspension will be described later.

Figure 1 shows a microLED suspension with a uniform distribution after stirring, which settled to about half of the liquid column height after time t1 and completely settled after time t2.
Figure 2 is a partial cross-sectional view of a system for characterizing a microLED suspension.
FIGS. 3A and 3B are partial cross-sectional schematic diagrams of an assembly of a system for selectively collecting microLED elements from a carrier substrate.
FIG. 4 is a partial cross-sectional view of a system for selectively separating an area of a carrier substrate of a microLED.
Figure 5 is a schematic diagram showing the fabrication of a group of microLEDs for fluidic assembly consisting of three sequential steps.
Figure 6 is a partial cross-sectional view of a typical microLED on a carrier wafer after device processing is completed.
Figures 7a and 7b show a wafer map of known defect locations from microLED fabrication (Figure 7a) and a wafer map of known defect locations for a typical aligned structure (Figure 7b).
Figure 8 is a graph showing the cathode ray emission spectrum of GaN microLED.
Figure 9 is a composite fault map that can be used to guide the fault control process.
Figure 10 is a plan view of the wafer with the emission wavelengths depicted as contours of equal width.
Figures 11a and 11b are partial cross-sectional views showing some representative microLED defects and corrective actions, respectively.
Figures 12a, 12b, and 12c are partial cross-sectional views of microLEDs collected in a solvent.
Figure 13 is a partial cross-sectional view showing a suitable storage container.
Figure 14 is a graph showing the main diameter of an exemplary microLED and a contaminant.
Figure 15 is a schematic diagram of a dissolution splitting tank.
Figure 16 is a schematic diagram of a continuous flow classification filtration method.
FIGS. 17A through 17C illustrate exemplary measurement graphs of optical transmittance over time for a microLED suspension having a diameter of 42 μm in 20 mL of isopropanol (IPA), exemplary measurement graphs of optical transmittance over time for 1.3 million microLEDs in different volumes of IPA, and a calibration curve of optical transmittance versus concentration of the microLED suspension of the system, respectively.
Figures 18a and 18b show the distribution density gradients for unidirectional and bidirectional/double-speed distribution paths, respectively.
Figures 19a to 19c are schematic diagrams showing that a well-mixed suspension from a container, a nozzle, and a tube is directly transferred from an initial source container to an assembly substrate by gradient separation, respectively.
Figure 20 is a schematic diagram showing controlled volume pipetting from a well-mixed suspension onto an assembly substrate.
Figure 21 is a schematic diagram showing the subsequent suction and distribution through the compensated suspension and bubble mixing and distribution head in the intermediate port (tundish).
Figure 22 is a schematic diagram of a parallel distribution method using a distribution head array.
FIG. 23 is a flowchart illustrating a method for selectively collecting microLED elements from a carrier substrate.
FIGS. 24A through 24C are flowcharts illustrating a first method of distributing microLED elements in a light-emitting display panel.
FIG. 25 is a flowchart illustrating a second method for distributing microLED elements in a light-emitting display panel.

FIG. 2 is a partial cross-sectional view of a system for characterizing a microLED suspension. The system (200) includes a transparent vessel (201) having a vertical central axis (202). A plurality of light-emitting elements (204a-204n) (an LED array) are shown, each of the light-emitting elements having a predetermined output light intensity and spaced apart a first predetermined distance (206) along a vertical first axis (208) that is oriented toward and parallel to the central axis (202) of the vessel (200), where n is an integer greater than 1. A plurality of photodetectors (210a-210n) (an array of photodetectors) are spaced apart a first predetermined distance (206) along a vertical second axis (212) that is parallel to the central axis (202). Each of the photodetectors (210a-210n) has an optical input terminal directed to a corresponding light emitting element output terminal, and an output terminal respectively in a circuit (214a-214n) and provides an electro-optical density signal responsive to the measured light intensity. A monitoring device (216) has an input terminal for receiving the optical density signals (214a-214n). The monitoring device (216) compares the light intensity associated with the optical density signal and provides an output in the circuit (218) responsive to the comparison. For simplicity, the drawing shows that the number of light emitting elements is equal to the number of photodetectors and is spaced equally apart. Additionally, the light emitting elements output light of the same intensity. However, it should be understood that, once calibrated, similar systems may be operated without these explicit limitations.

The vessel (201) comprises a microLED suspension (228). The monitoring device (216) can provide a microLED uniformity measurement, or can provide a calculation of the microLED counts per unit volume of the suspension in the circuit (218) determined based on the uniformity (density) measurement. In one embodiment, the monitoring device (216) comprises a non-transitory memory (220) having a stored calibration curve (222). In this case, the monitoring device (216) can provide the microLED counts per unit volume of the suspension in the circuit (218) in response to a comparison of the optical density signals from the circuit (214a-214n) with the calibration curve (222). As part of the calibration curve, in one embodiment, the monitoring device (216) can receive and store data regarding the volume of the vessel (201).

In another embodiment, the monitoring device (216) has an input to the circuit (224) for receiving a calibration input signal representing the total number of microLEDs in the suspension. In this case, the monitoring device can provide a microLED count per unit volume of the suspension in response to comparing the optical density signals (214a-214n) to the total number of microLEDs. For example, the total number of microLEDs can be determined by counting the number of valid microLEDs collected from the carrier substrate. The input to the circuit (224) can alternatively or additionally receive a running measurement of the suspension volume.

In one variation, the monitoring device (216) receives a set of optical density signals (214a-214n) collected over a period of time and provides an output of the microLED settling time or microLED size to the circuit (218). Additionally, the vessel (200) may be divided into a plurality of notches (226a-226n) and may contain a uniform suspension of microLEDs (the suspension (228) shown is not uniform). In this case, the photodetectors (210a-210n) detect changes in the suspension level as measured relative to the vessel notches (226a-226n). The monitoring device provides an output of the number of microLED elements dispersed from the vessel (200) or the volume of suspension dispersed from the vessel to the circuit (218). Advantageously, the output may be provided in real time.

To aid in the above measurement, a stirring device can be used to homogenize the suspension. Various homogenization mechanisms are described in more detail below. In one embodiment shown, a solution (solvent) or a gas can be used to mix the suspension. Once the suspension volume is known, the stirring mechanism can be adjusted to optimize mixing.

FIGS. 3A and 3B are partial cross-sectional schematic diagrams of an assembly of a system for selectively collecting microLED devices from a carrier substrate. As shown in FIG. 3A, the system (300) includes a turntable or vacuum chuck (302) having a rotating (spindle) interface for mounting a carrier substrate (304) comprising inorganic microLED devices (306) attached to the carrier substrate by an adhesive (308). An elbow (310) is connected to the rotating vacuum chuck (302) having a plurality of optional settings for determining a rotational angle (tilt) (312) of the turntable in a plane in which the x- and z-axes are located. A gantry (314) is connected to the elbow (310) having a plurality of optional settings for determining a height of the turntable along the z-axis. A tray (316) contains an adhesive dissolving solvent (318) and has a top opening for receiving the carrier substrate (304). The controller (320) has outputs connected to the gantry (314) and the elbow (310) respectively in the circuits (322 and 324) and provides height and angle settings. As illustrated in the drawing, the system (300) allows selected radial portions of the carrier substrate (304) to be exposed to the adhesive dissolving solvent (318) in response to the settings of the gantry (314) and the elbow (310). MicroLED elements (306) separated from the selectively exposed portions of the carrier substrate (304) are collected on the tray (316). In this example, a λ ₁ region of the carrier substrate is being collected (see FIG. 10 ).

In one embodiment, the controller (320) has an input terminal in the circuit (326) for receiving a first map of the microLED performance region, and provides setting of the gantry (314) and setting of the elbow (310) for selecting a radial region exposed to the adhesive dissolving solvent on the carrier substrate (304) in response to the first map.

As illustrated in FIG. 3B, the system (300) may include an inspection subsystem (328) having an optical input (330) and an output connected to the circuitry (332) of the controller (320) for identifying individual defective microLED elements (306) in the carrier substrate (304). A trimming laser (334) has an input connected to the circuitry (336) of the controller (320) for receiving a second map of the defective microLED elements, and an output (338) for expelling the defective microLED elements from the carrier substrate (304) via laser radiation in response to the second map.

FIG. 4 is a partial cross-sectional view of a system for selectively isolating a region of a carrier substrate of microLEDs. The system (400) includes a controller (402) having an output terminal from a circuit (404) for providing a first map of a predetermined defect region (406) on a carrier substrate (408). A printer (410) has an input terminal for receiving the first map from the circuit (404) and a nozzle (412) for applying a content-soluble colloidal material (414) to a selected region of the carrier substrate (408) in response to the first map. The microLEDs (415) in the selected region (406) remain attached to the carrier substrate (408) despite being exposed to an adhesive dissolving solvent (not shown). Some examples of content-sensitive colloidal materials (414) include SU-8, epoxy resin, polyethylene terephthalate (PET), acrylonitrile butadiene styrene (ABS), and polyimide.

Optionally, the system (400) can further include an inspection subsystem (416) having an optical input (418) and an output connected to a controller of the circuit (420) for identifying an area (422) of a microLED device having a non-determined defect in the carrier substrate (408). The input of the circuit (404) of the printer can receive a second map of the area (422) of the non-determined defect in the microLED device from the controller (402), and can apply a solvent-soluble colloidal material to the area (422) of the detected defect in the microLED device in response to the second map (the solvent-soluble colloidal material is not applied in FIG. 4 ).

Figure 5 is a schematic diagram showing the fabrication of a group of microLEDs for fluidic assembly, which consists of three sequential steps. First, the microLED wafer is inspected to determine the location of defective microLEDs. According to the defect map, defective microLEDs and other fragment particles (debris) are removed from the carrier substrate or packaged in the carrier substrate to prevent them from being collected in the suspension. Next, the carrier substrate is immersed in a solvent to dissolve the adhesive on the substrate known to be well fixed to the microLEDs, and the microLEDs are rinsed into a holding vessel. After the microLEDs are settled, the solvent containing the dissolved adhesive is carefully poured out, and the solvent is exchanged several times to remove any remaining adhesive residues. The obtained suspension is filtered to remove particles whose size is significantly different from the microLEDs, and the amount of solvent is adjusted so that the density of the microLEDs in the suspension is suitable for subsequent mixing and dispensing operations.

Using microLED suspensions, three alternative dispensing systems can be applied to the display substrate, each taking a different approach with varying trade-offs in terms of dispensing speed, volume control, and complexity. In each case, multiple individual isotropic transfers are performed to cover the substrate. The isotropic transfers can be performed directly onto the substrate, or via a controlled-volume intermediate in the form of a “suspension tank,” which not only dilutes the suspension, but also actively mixes the microLED suspension using a dispensing head prior to dispensing onto the substrate. For suspensions with sufficiently long settling times, direct transfer is preferred. Finally, the uniformity of the substrate can be checked, and additional small-area dispensers can be used to fill low-density regions.

FIG. 6 is a partial cross-sectional view of a typical microLED on a carrier wafer after device fabrication is complete. As described in parent patent US 10,643,981, which is incorporated herein by reference, the microLEDs are typically fabricated from a sapphire substrate and transferred to the carrier wafer by laser lift off (LLO). As illustrated in the drawing, the resulting wafer has millions of microLEDs embedded in an adhesive layer of the carrier wafer. Unfortunately, the wafer may have several different defect types (including process control structures) that can negatively impact subsequent processing, so selective collection is used to ensure that only known good microLEDs are utilized by the fluidic assembly process.

Figures 7a and 7b show a wafer map of known defect locations from microLED fabrication (Figure 7a) and a wafer map of known defect locations for a typical alignment structure (Figure 7b). There are several known defects or process control structures in the device fabrication for microLED fabrication, which systematically appear at the same locations due to the nature of the process depicted in the figure. Each wafer has an identification mark near the wafer edge, which creates a concave hole in the substrate surface, damaging the microLED pattern. In the processing steps to form the microLED structure, all photoresist coatings are used to transfer the pattern to the wafer. The photoresist coating is not perfect, and therefore produces an annular pattern defect known as an edge bead, especially at the edge of the wafer. The photolithography process uses a series of marks to determine the location of each layer relative to the previous layer. The alignment marks and metrology structures (Figure 7b) used are necessary for microLED fabrication, but since these structures are incorporated into the solution collected in the microLED, they are defects. These defects are much larger than typical microLED sizes, with wafer edge annular pattern defects being as wide as 2–3 mm. The first configuration of the wafer defect map is the location of these large system layout structures.

The second type of defect is a random processing defect, such as a chemical mechanical polishing (CMP) scratch on the substrate, a large residual gallium nitride (GaN) chunk caused by particles falling from the isolation etch, or a missing metal electrode. These larger defects can be identified by optical scanning, which compares the differences in adjacent images that do not match the expected microLED pattern. This part of the defect map consists of a set of coordinates describing the area and location of each defect.

The most important defect types are functional defects that affect the electrical performance and optical emission of microLEDs. Mapping of these defects can be achieved through four different complementary techniques.

l) Perhaps the most ideal technique would be an electroluminescence (EL) test, which detects each microLED and measures the light emitted. This test directly identifies weak devices with little light emission, as well as devices that are shorted or open. The downside is that this technique is slow and particularly difficult to detect for small electrodes without damaging the electrode surface. This technique can be used to measure some representative devices.

2) Photoluminescence (PL) uses a light source (typically a laser) whose wavelength excites a transition in the LED structure and the resulting wavelength and intensity can be measured. The technique can identify cracked or shorted microLED devices as well as metal-organic chemical vapor deposition (MOCVD) defects, but cannot identify missing or open contacts in the metal.

3) For optical comparison, the common method is to compare two images and find differences between them, the differences being defects. Optical images can also be compared to patterns (die-to-database).

4) Cathodoluminescence (CL) is explained below.

Figure 8 is a graph showing the GaN microLED cathodoluminescence spectrum. Micro-cathodoluminescence uses an electron beam from a scanning electron microscope (SEM) to excite transitions in the LED structure and measure the wavelength and intensity of the light produced. As shown in the figure, GaN microLEDs have a number of characteristic emission lines, so differences in the spectra can identify different defect mechanisms. A weak or absent emission peak at 455 nm indicates a shorted LED or multiple quantum well (MQW) problem. High emission at a broad peak near 570 nm can indicate etch damage or poor quality of the n-GaN. A low-intensity exciton peak can also indicate poor initial growth quality or dopant deficiency. Any significant deviation from the typical microLED spectrum indicates that the device may not function properly in a display. Figure 9 is a composite defect map that can be used to guide the defect control process. Since functional testing can identify individual defective microLEDs, this portion of the defect map consists of the X-Y coordinates of each defective LED. Electroluminescence and photoluminescence can also produce maps of LED performance characteristics (including emission wavelength, efficiency, and threshold voltage). Since these techniques can identify areas of different performance, selective collection techniques can be applied to collect small areas of the wafer to produce microLED suspensions with more closely distributed microLED device performance uniformity.

Figure 10 is a plan view of the wafer with the emission wavelengths outlined in equal width. In some cases, each radial band of the carrier substrate must be collected separately so that each individual suspension (e.g., the four bands shown in the figure) has a narrow wavelength or efficiency distribution. This simple binning technique can be used to collect suspensions from different wafers with the same emission wavelength to produce a display without color spots.

Figures 11a and 11b are partial cross-sectional views showing some representative microLED defects and corrective actions, respectively. Using the defect size information in the composite defect map, the wafer is processed in two ways as shown in Figure 11b to remove the defects. For large defect areas such as wafer edge and pattern of area defects, a material resistant to the solvent used in the collection process is coated so that the defective areas are still captured on the carrier wafer after collection. Suitable coating materials include SU8, epoxy resin, polyethylene terephthalate (PET), acrylonitrile butadiene styrene (ABS), or polyimide, which can be patterned by spray, inkjet, or coated with a brush or pen. The wafer can be baked after processing to harden the retaining material. For a single element identified as a small defect, the defective MicroLED can be removed by bombarding the single element with the small defect using a pulsed laser beam whose size is smaller than the microLED diameter. By selecting an appropriate laser wavelength and energy, the target element is ejected from the adhesive layer by accelerating heating by absorption of the defective microLED. An auxiliary vacuum nozzle close to the laser target area is used to capture and process the ejected element, thereby preventing the defective microLED from being re-deposited on the carrier wafer.

Figures 12a, 12b and 12c are partial cross-sectional views of microLEDs collected in a solvent. In Figure 12a, a carrier wafer with known good microLEDs (a carrier wafer that has undergone calibration) is placed in a solvent-resistant container, and immersed in a solvent or a solvent mixture that can dissolve the adhesive that holds the microLEDs. The solvent may be acetone, toluene, trichloroethane, N-methylpyrrolidone (NMP), xylene, cyclohexanone, butyl acetate, etc., and the solvent may be heated and gently stirred to accelerate the dissolution of the adhesive and the collection of the microLEDs. It should be noted that mechanical contact between the container and the microLEDs in the carrier is prevented during this process, as this may damage the microLED devices. By maintaining the vertical position of the carrier wafer in the container, the solvent can circulate well and the microLEDs can settle to the bottom of the container when the adhesive is dissolved.

Another way is, as shown in Fig. 12b, when the carrier wafer is placed horizontally in a shallow container, the edges of the carrier wafer are supported by the narrow shelves of the container. From this, it can be seen that the printed or coated content-sensitive colloidal material can protect the adhesive from being affected by the solvent, so that the MicroLEDs within the defective capture region exceeding a certain size will remain on the carrier wafer. When all the microLEDs have been released and settled to the bottom of the container, the carrier wafer can be removed and examined to see if there are any good microLEDs remaining. In some cases, if the carrier substrate is not immediately removed from the solvent, the solvent may penetrate from underneath the defective capture region coated with the capture material, dissolve the adhesive, and release large clumps of capture medium into the container.

The above "complete" wafer collection method is quick and simple, but in some cases, selective collection techniques may be used. As shown in FIG. 12c, since the microLEDs are collected only from the area in contact with the adhesive dissolving solvent, a simple selective collection can be performed by dropping a small amount of solvent onto a horizontal carrier wafer. After the adhesive is dissolved, the collection area is rinsed to remove and collect the released microLEDs from the carrier wafer.

Another system for selectively collecting a radial region of a wafer can use the same principle of exposing a controlled region of a carrier wafer to an adhesive dissolution solvent, which is illustrated in the system of FIG. 3a . For example, based on the emission wavelength contour map of FIG. 10 , the tilt of the carrier substrate and the height along the z-axis are adjusted so that only region λ ₁ (see FIG. 10 ) is immersed in the solvent. The carrier substrate is slowly rotated to expose a radial band to the solvent, and the emitted microLEDs are allowed to settle to the bottom of the solvent tank. Once the first region is collected, the carrier is taken out of the solvent, and the microLEDs with the λ ₁ wavelength distribution are removed from the solvent. The carrier can then be repositioned to collect the λ ₂ region, and subsequent microLED collections are performed in the same manner.

Typically, the collection solvent is heavily contaminated by residual and adhesive-covered microLEDs, so the microLED suspension must undergo a series of solvent exchanges to remove the contaminants. One solvent exchange cycle is as follows.

1) Indirectly stir the solvent using a vortex mixer or ultrasonic transducer to completely homogenize the suspension and break up any remaining adhesive lumps.

2) You need to wait several times for the settling time so that all the microLEDs are collected at the bottom of the container.

3) Carefully pour 80% to 90% of the solvent without affecting the sunken microLED.

4) Add new solvent to the container.

5) Repeat steps 1) to 4).

Typically, three or more solvent circulation exchanges are performed to ensure removal of the adhesive component. The solvent used for adhesive removal is selected solely based on its ability to dissolve the adhesive without damaging the electrodes of the microLED, and therefore is not the most suitable choice for subsequent filtration and fluid assembly operations. To replace with a new solvent selected for washing or fluid assembly, the solvent exchange step can be performed in step 4) above. The solvent circulation exchange can be performed at least three times to ensure that the solvent rinses the microLEDs out of the vessel walls. After the solvent exchange, the microLED solution is transferred from the collection vessel to a clean vessel capable of storing and transporting the suspension. The vessel should be chemically stable and hydrophobic to the suspension to minimize adhesion of the microLEDs to the surface. Some suitable materials include acetal homopolymer, polytetrafluoroethylene (PTFE), polypropylene, polystyrene, etc.

Figure 13 is a partial cross-sectional view showing a suitable storage vessel. Typically, the vessel is tall and thin with vertical walls and a large angled surface to minimize accumulation and to deposit the microLEDs at a single intended location. This control over the accumulation point allows for efficient transfer of the microLEDs from the vessel. If the vessel is transparent, optical density measurements can be conveniently used to monitor the uniformity of stirring the suspension and to adjust the microLED concentration by adding or removing solvent.

FIG. 14 is a graph showing the main diameters of exemplary microLEDs and contaminants. By filtering to remove particles and fragments, the quality of the microLED suspension can be further improved. The exemplary microLED has a diameter of 42 microns (μm). Small particles of about 10 μm or less may be electrode metal pieces that fall off during the electrode exfoliation process, pieces of interlevel dielectric (ILD), or dust in the air. These particles may be trapped in the well structure during fluid assembly, which may interfere with the assembly of the microLED or affect the bonding between the electrodes of the microLED and the substrate electrode, resulting in yield loss. Larger fragments may be GaN fragments or small defects wrapped in the capture medium released during collection. During the filtration process, all objects except a band centered around the diameter of the microLED are removed.

It is clear from the drawings that damaged microLEDs with similar size to good microLEDs cannot be removed by filtration, so selective collection is very important for capturing or removing damaged microLEDs prior to collection. It is also important that the microLED suspension does not generate new damaged microLEDs due to excessive mechanical interactions between microLEDs or between microLEDs and the container and fixture.

A simple filtration method uses a mesh filter developed for cell collection to create a desired bandpass around the size of the microLEDs. First, the suspension is filtered through a 40 μm mesh to remove any larger particles that escape from the carrier wafer during collection, including any capture media. Next, the suspension is filtered through a 20 μm mesh to capture the microLEDs and allow the smaller particles to pass through into a waste container. The microLEDs are backwashed into a clean container with the wash solvent from the filter.

Figure 15 is a schematic diagram of the dissolution splitter tank. Although the mesh filtration method is inexpensive and effective, there is a significant mechanical interaction between the microLEDs and the filter, and the elements are affected by shear forces from accumulation on the filter membrane during the microfiltration process. The fluid flow-based filtration method helps to prevent potential mechanical damage. Since the hydrodynamic diameter of the microLEDs and the viscosity of the suspension solvent are known, the dissolution splitter tank used as a filter can be fabricated as shown in the figure. The microLED suspension is introduced into the top of the first filtration column (μLED feed), and the solvent is introduced from the bottom of the column and is introduced at a rate that pushes all particles smaller than a critical size of about 50 μm upward from the column (stream 1), while the larger particles settle to the bottom of the column where they can be collected and discarded. The small-sized fraction flows through the transfer channel to the top of the second dissolution column. In the second column, flow 2 is adjusted so that particles smaller than about 30 μm are forced upward into the waste channel, while the microLEDs settle to the bottom of the column and can be collected for assembling the microLED display screen.

Figure 16 is a schematic diagram of a continuous flow classification filtration method. The continuous flow classification separation method is similar to the dissolution division tank, and separates good microLEDs from small particles by utilizing the difference in sedimentation speed of different particles. As shown in the figure, the current carrying is adjusted so that particles smaller than about 30 μm flow out from the upper waste gate, and microLEDs with a faster falling speed flow out from the lower port.

For highly efficient fluidic assembly, the microLEDs should be distributed uniformly on the display substrate and the number of microLEDs should be sufficient to fill all available assembly sites (also called trap sites or wells). In practice, the optimal number of microLEDs is greater than the number of assembly sites. If the number of microLEDs is less than the optimal number, the assembly time increases because the microLEDs have to travel further to reach an empty well for assembly. However, if the number of microLEDs is more than the optimal number, the devices often tend to clump together, which hinders the assembly process. In addition, since all remaining microLEDs need to be removed after assembly, too many distributed microLEDs increase the cleanup time and involve more microLEDs in the recovery process. Therefore, it is very important that the suspension distribution process is based on a microLED suspension with a well-controlled number of microLEDs per unit volume.

Since the variation of the number of microLEDs in an aliquot increases with the suspension concentration and non-uniformity, the concentration of microLEDs in the suspension should be adjusted to ensure that an accurate number of microLEDs are transferred to the display substrate. The number of microLEDs collected from the carrier wafer can be well determined by calculating the collection area after the defective region is removed. Then, the concentration of the suspension can be simply set by adding an appropriate volume of solvent during the final exchange process. However, the variation of the concentration occurs due to solvent evaporation, removal of the aliquot for distribution, and returning the recovered microLEDs to the suspension. In order to control the concentration, a system to accurately determine the concentration of the suspension is required.

Referring back to Figure 2, for a radially symmetric transparent vessel, the concentration only varies in the z-axis, so to quantify the microLED concentration at different heights in the suspension vessel, a pair of collimating LEDs (or laser diodes) and a photodetector are used to emit light through the suspension and measure the density (optical opacity), which is expressed as log(Iin/Iout). Since the light attenuation at different heights in the suspension is directly proportional to the microLED concentration at that height, a calibration curve based on the microLED size is used to convert the optical density measurements into a calculated concentration. To determine the concentration, the optical density measurements are started immediately after stirring to create a homogeneous suspension. The relationship between time and height of the measured vessel directly provides the sedimentation rate.

After about half of the settling time, detectors (210a and 210b) receive full intensity, detector (210c) returns to an intensity representing 50% to 60% of the uniform microLED density, and detector (210n) sees a nearly uniform microLED density. If the suspension is left undisturbed for a long enough period of time, all of the microLEDs will be collected at the bottom of the vessel relative to the settling time, minimizing light scattering and all intensity measurements will be at a maximum. If the suspension is sufficiently stirred, the microLEDs will be uniformly distributed throughout the liquid column, minimizing light scattering at each height. After stirring is stopped, the microLEDs will begin to fall under the influence of gravity until they reach terminal velocity. Over time, the microLED concentration at the top of the liquid column will decrease and the detector intensity will increase.

FIGS. 17A through 17C illustrate exemplary measurement plots of optical transmittance versus time of a suspension of microLEDs having a diameter of 42 μm in 20 mL of isopropanol (IPA), exemplary measurement plots of optical transmittance versus time of 1.3 million microLEDs in different volumes of IPA, and a calibration curve of optical transmittance versus concentration of the microLED suspension of the system, respectively. Optical density is defined herein as the reciprocal of transmittance, and the data are divided by the final intensity for normalization purposes. In FIG. 17A , about 1.2 million microLEDs are suspended in a cylindrical translucent tube with a diameter of 27.5 millimeters (mm), and the optical density is measured at five different vertical positions. At time 0, the tube is mechanically vibrated to agitate the suspension and ensure that the microLEDs are evenly distributed in the fluid column. Noticeable noise occurs during agitation and during the first few seconds after agitation due to bubbles in the liquid. The measured light intensity after stirring is about 60% lower than that of the settled suspension. The recovery to full intensity when all the microLEDs have settled from the sensing holes is a function of the distance from the sensor to the top of the fluid column. Thus, for the upper sensor located 12 mm from the top of the liquid column, t1 is 23 s, whereas for the lower sensor located 33 mm from the top of the liquid column, it recovers in 64.5 s (t5). Therefore, since the terminal velocity of these microLEDs in IPA is 0.51 mm/s to 0.56 mm/s, the settling time for the 45 mm liquid column is about 85 s, taking several seconds to settle after being distributed in a thin liquid layer.

The transmittance measurement system can also be used to determine the number density of microLEDs in the suspension, which is key information for accurate handling of the suspension. In Fig. 17b, the transmittance is measured as a function of time after stirring for five different dilutions of the suspension. A microLED with a diameter of 42 μm is measured at 13 mm in 20 mL of IPA, and then a measured amount of liquid is added to prepare a suspension with lower density. In Fig. 17b, the light intensity after stirring is lowest at the highest concentration (C1) and increases as the density decreases continuously. As the liquid column height increases, the settling time at the fixed location in the vessel also increases.

Figure 17c is a graph of multiple tests showing a calibration curve for determining the number of microLEDs per milliliter of liquid based on the optical transmittance of the uniform suspension. It can also be seen that the uniformity of the suspension is very high within a few seconds after stirring, and the duration of this state decreases as the distance to the top of the fluid column decreases. It is ideal to extract aliquots from the system at least 20 mm below the top surface after a few seconds of stirring. This characterization system can be used to select optimal process parameters for different microLED sizes, vessel sizes and vessel shapes, and suspensions.

Mixing of microLED suspensions is a balance between the force required to achieve high suspension uniformity and the shear force to prevent damage. Mixing can be accomplished by external agitation, where the smooth vessel walls impact the suspension, creating fluid flow that agitates the microLEDs within the suspension vessel. Internal agitation can be accomplished by introducing a solvent or gas stream into the suspension vessel to induce turbulent fluid flow. Mixing can also be accomplished by rapid pipetting, for example, to remove and inject liquid. The goal, of course, is to create microLEDs that are uniformly distributed in vertical columns in the vessel without damaging the microLED devices.

A controlled volume containing a number of LEDs for a unidirectional distribution path can be drawn from a well-mixed suspension. The distribution path can be a single point, a single line segment, a tortuous path, or some combination of paths. Multiple distribution paths are used to ensure complete and uniform distribution in the display assembly area. Due to the very short settling time of the LEDs, especially in the thin fluids used for assembly, the lateral diffusion of the distribution paths is limited to the millimeter level. Therefore, multiple distribution paths that are relatively close together are required for uniform distribution of the microLEDs.

Figures 18a and 18b show distribution density gradients for unidirectional and bidirectional/double-speed distribution paths, respectively. The exact method by which the amount of dispensed material is transferred from a well-mixed suspension is determined by the characteristics of the system and the product requirements. In particular, the size of the assembly substrate, the moving speed of the distribution head, the thickness of the assembly fluid on the substrate, the uniformity of the mixing, and the control over the transfer volume are all important. These considerations are generally a trade-off between uniformity and cost (including system cost, processing time, and product yield). Defects in distribution uniformity can also be compensated for during the assembly process, and therefore a series of methods have been proposed. For example, a given path can be rapidly traced multiple times along a single line segment to compensate for uneven distribution speeds (which may occur, for example, due to sedimentation within the distribution head), and achieve a uniform density along the path illustrated in Figure 18b.

Figures 19a-19c are schematic diagrams showing a well-mixed suspension from a vessel, nozzle, and tube being transferred directly from an initial source vessel to an assembly substrate by decantation, respectively. To limit or completely avoid the use of tubes and components for transferring the well-mixed suspension, either direct transfer from the source tank or discrete aliquots of the source suspension can be removed using a pipette tip. This approach prevents dead space accumulation of microLEDs in the component and limits the interaction of the microLEDs with the suspension to a surface that is a) small, b) cleanable, and c) disposable. The pipette tip can draw the suspension from the tank by volume displacement (e.g., a plunger) or by actively applied vacuum. The advantage of applying continuous vacuum is that once the aliquoted suspension is no longer in contact with the source suspension, the surrounding atmosphere is drawn into the aliquot sample as bubbles, actively mixing the suspension during the transfer process.

To limit yield loss, a high concentration suspension is preferred prior to introduction to the substrate. In this regard, the preparative method can be combined with an intermediate small volume vessel, in which additional liquid is replenished to the dispensed suspension. This can be accomplished by filling the suspension tank with a known volume of pure liquid using a standard system (including pipes, valves and fittings), since the pure liquid is not limited by the microLED suspension. The suspension preparative sample can be calibrated by aspirating the liquid into the dispense head, or by allowing the suspension to settle into the suspension tank, and then drawing the mixture back into the dispense head. The diluted suspension can then be transferred to the substrate.

The above mixing, transfer and dilution processes are general and variations of them can be selected to configure an integrated system and application method optimized for the type of microLED display being produced. Some complete dispensing processes are detailed below as examples.

In Fig. 19a, a desired number of LEDs can be dispensed directly from a well-mixed vessel onto an assembly substrate. This method provides the fewest transfer steps and the least surface exposure to unwanted microLED static friction. Since the fully settled state of the suspension is not uniform, the system mixes the suspension to a uniform suspension density, so that the transfer of a volume of suspension corresponds to the transfer of a number of microLEDs. The well-mixed suspension can then be transferred directly onto the substrate. Decanting the microLEDs onto the substrate creates significant flow throughout the fluid on the substrate, which can rapidly transfer the LEDs across the entire surface area. However, the uniformity is poor, which may require additional assembly time to compensate.

In Figure 19b, a small area nozzle paired with appropriate pressure control in the top space of the vessel can transfer the suspension in a more controlled manner, but this method is much slower than the decanting method. This method also requires more precise translation of the microLED vessel on the assembly substrate than the decanting method.

In Fig. 19c, instead of the container translation, a tube can be used in which one end is immersed in the suspension and the output end is translated to the assembly substrate. The advantage of this method is that the suspension container can be actively mixed during the dispensing process, and the translation of the tube end is much easier than the container translation. Since multiple tubes can be used in the same microLED suspension container, the area covered by the unidirectional distribution is significantly increased. The disadvantage of this method is that the surface area interacting with the suspension is very large, resulting in the trapping of the microLEDs in the pipe. For applications where cross-contamination is not a concern, this method using a limited pipe is the best option.

Figure 20 is a schematic diagram showing controlled volume pipetting from a well-mixed suspension to an assembly substrate. As shown in the schematic diagram, the pipette can draw an aliquot sample from a well-mixed suspension and dispense it directly onto the substrate, rather than dispensing directly from the vessel. This method uses a mature technology in biology to stably transfer an aliquot sample with high precision. In addition, this method does not require additional configuration of the suspension vessel, such as an inlet, nozzle, etc.

When accurate volumes and avoidance of cross-contamination are top priorities, such as when minimizing excess microLEDs used in assembly or when sequentially assembling microLEDs of different sizes, the pipette transfer method is preferred. The trade-off is that pipette dispensing is slower than other methods, as the pipette must return to the suspension vessel after each dispensing pass. Multi-pipettes exist, but given the concentration of the suspension vessel and the mixing limitations of the suspension vessel, they are not suitable for pipetting from a single source.

Figure 21 is a schematic diagram showing the subsequent aspiration and distribution through the dispense head and mixing and dispensing of the suspension and bubbles corrected at the intermediate port. Another limitation of the direct transfer method and the pipette transfer method is that the volume dispensed is equal to the volume withdrawn from the source suspension container. In order to limit the size of the suspension container for a large dispensing area, it may be necessary to reduce the concentration of the suspension dispensed onto the assembly substrate. As shown in the figure, one way to implement this dilution is to draw a concentrated aliquot sample from the suspension supply container and compensate with pure liquid at the dispense head or intermediate port.

Figure 22 is a schematic diagram of a parallel dispensing method using an array of dispense heads. The use of an intermediate port has the potential advantage of separating the dispense head pickup from the suspension container pickup. Thus, the system depicted in the figure is scalable proportionally to an array of dispense heads drawn from an array of suspension tanks. The generation of racks of suspension tanks with precisely controlled microLED concentrations can be done in a separate step, and the racks are then loaded into an assembly tool together with the assembly substrate.

All three methods described are variations on the core concept of efficient transfer of microLEDs. Fluidic assembly can be used for a variety of microLED sizes, assembly areas, and pixel spacings. Here are some examples of how changes in assembly requirements can affect method selection.

For single-color assembly on small area substrates, direct dispensing may be required.

The batch assembly of multiple substrates in parallel from a single suspension vessel suggests the use of a decanting method.

Continuous repetitive assembly of substrates may be best suited to the nozzle method.

Large-area assembly containing large volumes of well-suspended microLEDs has low wavelength variation between batches. Using tube-mediated transfer from a source vessel of the suspension may be most economical.

For medium-sized substrates (a few centimeters on a side) with valuable microLEDs and low cross-contamination tolerance, such as sequentially assembled RGB triple-emitter color displays, pipette dispensing may be the preferred choice, especially when pipette dispensing time is not a critical part of the production time.

For very large substrates exceeding the 2nd generation size (360×465mm), the single head distribution becomes abnormally slow, and the production demand dictates the parallel distribution of multiple head arrays in the distribution system. For fast settling suspensions, the ability to vacuum mix the suspension is very important to improve distribution uniformity. In addition, for large substrates, the overall distribution requirement becomes very high, and correcting the concentrated source suspension can improve processing and mixing uniformity.

Some of the key requirements for dispensing microLEDs from a suspension onto a substrate are to limit microLED waste due to damage, surface adhesion, and non-uniform application. Therefore, it is desirable that the suspension does not encounter valves, pumps, or components during the collection, filtration, overall mixing, dispensing, and recovery operations. Although some loss is inevitable in the suspension vessel itself, thorough rinsing combined with the capture and recovery process greatly mitigates this loss. In the case of the pipette transfer method, only the pipette tip comes into contact with the suspension. The pipette can be rinsed inside and out to recover the microLEDs or discarded to prevent cross-contamination of the microLEDs.

The vacuum mixing dispense head using disposable tips and the intermediate suspension tank can be rinsed for recovery, reused if cross-contamination is not a concern, or replaced to prevent cross-contamination. The embodiment illustrated in Fig. 19c uses a tube that cannot be washed, and this method is only optional if the microLEDs are in good suspension without considering cross-contamination, thereby minimizing contact with the tube sidewalls as much as possible.

FIG. 23 is a flow chart illustrating a method for selectively collecting microLED elements from a carrier substrate. For clarity, the method is described in numbered steps, but the numbering does not necessarily indicate the order of the steps. It should be understood that these steps may be skipped, performed in parallel, or performed without maintaining a strict order. However, the method generally follows the numerical order of the steps presented. The method begins at step 2300.

In step 2302, an inorganic microLED element is provided that is adhesively attached to a carrier substrate. The defect regions of step 2304 are predetermined, wherein each defect region includes a plurality of adjacent defective microLED elements or process control structures (e.g., CMP scratches). In step 2306, a content-soluble colloidal material is formed that covers the predetermined defect regions. In step 2308, the exposed adhesive is dissolved with an adhesive dissolving solvent. The adhesive dissolving solvent includes acetone, toluene, trichloroethane, N-methylpyrrolidone (NMP), xylene, cyclohexanone, butylacetate, or a combination thereof. In step 2310, the microLED element located outside the predetermined defect region is separated from the carrier substrate. In step 2312, the attachment of the microLED element and the carrier substrate within the predetermined defect region is maintained by adhesion. In response to the separation of the microLED device from the carrier substrate, the valid microLED device is collected into a collection vessel at step 2314.

In one embodiment, a carrier substrate is inspected in step 2305a to locate a defective microLED device, and an undetermined defective region including a plurality of adjacent defective microLED devices is located in step 2305b. The inspection process can be performed via optical comparison, electroluminescence, photoluminescence or cathodoluminescence testing. Then, in step 2306, a content-soluble colloidal material is formed to cover the undetermined defective region. In another embodiment, in response to the inspection in step 2305a, an undetermined individual defective microLED device is located in step 2305c. Then, a laser trimming process is used in step 2307 to release the individual defective microLED device.

In one embodiment, at step 2309, additional power, such as fluid circulation, thermal energy, gravity, vibration, or a combination thereof, is applied, and in response to the additional power, at step 2310, the microLED element is at least partially separated from the carrier substrate.

In one embodiment, the step of dissolving the exposed adhesive in step 2308 comprises selectively exposing a portion of the carrier substrate to an adhesive dissolving solvent. Subsequently, the step of separating the microLED device from the carrier substrate in step 2310 comprises separating the microLED device from the selectively exposed portion of the carrier substrate. More specifically, the step of selectively exposing the portion of the carrier substrate to the solvent may comprise the following substeps: In step 2308a, the carrier substrate is rotated in the solvent. In step 2308b, a radial portion of the carrier substrate having a radius greater than d is exposed to the solvent. Subsequently, the step of separating the microLED device from the exposed portion of the carrier substrate in step 2310 comprises separating the microLED device from the radial portion of the carrier substrate.

In another embodiment, the step of collecting valid microLED devices in the collection vessel in step 2314 comprises the step of replacing the adhesive dissolving solvent with another liquid. If the valid microLED devices collected in the collection vessel in step 2314 have an average cross-sectional physical size s and the fluid contains impurities, then in step 2315a, the impurities having a maximum cross-sectional physical size greater than t are filtered to remove them, where t>s. Optionally, or in step 2315b, the impurities having a maximum cross-sectional physical size less than p are filtered to remove them, where p<s. The filtration methods in steps 2315a and 2315b can use mechanical screening, elution, classification, or a combination thereof. For example, in order to perform high-pass filtration and low-pass filtration simultaneously, the mechanical filtration can use two different mesh sizes. In the case of elution and classification, it should be changed according to the flow rate, and the output port can be either product or waste. Besides, there is no reason to use the same filtration method for the two types of filtration. For example, a mesh filter can be used to remove large contaminants, and then a classification tank can be used to remove small particle contaminants.

In one embodiment, the step of replacing the adhesive dissolving solvent with another liquid in step 2314 comprises the step of replacing the adhesive dissolving solvent with a filter solution having a lower viscosity than the adhesive dissolving solvent, and using the filter solution to filter in step 2315 to remove impurities from the filter solution.

In another embodiment, the method comprises replacing the adhesive dissolving solvent in step 2314 with an assembly solution having a lower polarity or a faster evaporation rate than the adhesive dissolving solvent. A surfactant, such as an anionic, cationic, nonionic surfactant, or a combination thereof, may be added to the assembly solution.

FIGS. 24A-24C are flow charts illustrating a first method for dispensing microLED elements in a light-emitting display panel. The method begins at step 2400. At step 2402, a suspension of collected microLED elements is added to a transparent first vessel. At step 2404, the suspension is stirred. Some examples of the stirring process include external vibration of the first vessel, generation of fluid flow in the suspension, and flow of gas in the first vessel. At step 2406, the opacity of the suspension is optically measured at various heights in the first vessel. In response to the optical measurement, the uniformity of the suspension is determined at step 2408. In response to the determined uniformity being greater than a uniformity minimum threshold, the suspension is dispersed on a top surface of the light-emitting display panel at step 2410.

In one embodiment, the number of microLED elements collected in step 2401a is determined. For example, the number of microLED elements collected on the carrier substrate may be known. In step 2409a, calculating the number of microLED elements per unit volume of the suspension, and in step 2410, dispersing the suspension onto the top surface of the light-emitting display panel comprises depositing a known first number of microLED elements in response to dispersing the first volume of the suspension.

In one embodiment, the step of optically measuring the opacity of the suspension at various heights in the first vessel in step 2406 comprises sub-steps. In step 2406a, a plurality of light-emitting elements having a predetermined output light intensity are arranged facing the central axis of the first vessel and spaced apart from each other by a first predetermined distance along a first vertical axis. In step 2406b, a plurality of photodetectors are arranged spaced apart from each other by a first predetermined distance along a second vertical axis, each photodetector having an input terminal directed toward a corresponding light-emitting element output terminal. In step 2406c, the light intensities received by the photodetectors are compared.

Another alternative approach starts with a known number of microLEDs (step 2401a), calculates a first number of microLED elements per unit volume of suspension in step 2409a, and determines the uniformity of the suspension in step 2408. In step 2412, a predetermined amount of fluid is changed to vary the ratio of fluid to LED elements in the suspension, and in step 2414, the opacity of the suspension is optically measured to calculate a second number of microLED elements per unit volume of suspension.

In step 2410, the step of dispersing the suspension onto the luminescent display panel comprises a step of using one of a single-step bulk decanting, a multi-step pipette translation, a nozzle-limiting vessel translation, and a pipetting dispersion process. The multi-step pipette lateral dispersion process comprises the following substeps: In step 2410a, the uniformity of the suspension is maintained such that it is greater than a uniformity minimum threshold value in the first vessel. In step 2410b, a predetermined aliquot volume is repeatedly withdrawn from the first vessel using a pipette. After each aliquot sample is withdrawn, in step 2410c, the pipette is translated a predetermined distance relative to the top surface of the luminescent panel. In step 2410d, a predetermined amount of aliquot sample is ejected per second during transcription.

In another aspect, the first vessel is pressure controlled and includes a nozzle, and the nozzle-limiting vessel translational dispersion process includes the following substeps: In step 2410e, the uniformity of the suspension is maintained to be greater than a uniformity minimum threshold value in the first vessel; In step 2410f, the first vessel is translated a predetermined distance relative to the top surface of the light-emitting panel, and in step 2410g, a predetermined amount of suspension is emitted from the nozzle per second during the translation.

In one embodiment, the first vessel comprises an output port that is pressure controlled and connected to one or more transport tubes, and the pipetting dispersion process comprises the following substeps: In step 2410h, the uniformity of the suspension is maintained to be greater than a uniformity minimum threshold value in the first vessel; In step 2410i, the one or more transport tubes are translated a predetermined distance relative to the top surface of the light-emitting panel; and in step 2410j, a predetermined amount of suspension is discharged from the one or more transport tubes per second during the translation.

The single-step mass decantation dispersal process includes the following sub-steps: in step 2410k, the uniformity of the suspension is maintained to be greater than a uniformity minimum threshold value in the first vessel, and in step 2410m, the suspension is discharged from the first vessel to the top area of the light-emitting panel through a fixed location center area discharge or area translational discharge.

In one embodiment, the number of assembly locations within a first region on the uppermost surface of the light-emitting display panel is determined in step 2401b. Then, in step 2410, the step of dispersing the suspension onto the uppermost surface of the light-emitting display panel includes the step of depositing the suspension onto a first number of microLED elements that is at least equal to the number of assembly locations within the first region.

In another embodiment, the number of repetitions of the translational path of the first region of the light-emitting display panel is determined in step 2409b, and the translational speed is determined in step 2409c. Then, the step of dispersing the suspension in step 2410 includes the step of calculating a dispersion speed of the suspension of the first volume in response to the number of repetitions of the path and the translational speed, thereby generating a uniform suspension density in the first region of the light-emitting display panel.

FIG. 25 is a flow chart illustrating a second method for distributing microLED elements in a light-emitting display panel. The method begins at step 2500. In step 2502, a suspension of collected microLED elements is added to a transparent first vessel. In step 2504, the suspension is stirred. In step 2506, a density of the suspension is optically measured at a plurality of first vessel heights. In response to the optical measurement, a first number of microLED elements per unit volume of the suspension is calculated in step 2508. In step 2510, an aliquot of the suspension is dispensed onto a top surface of the light-emitting display panel. In step 2512, the optical measurement is repeated, and in step 2514, a second number of microLED elements per unit volume of the suspension is calculated.

In step 2509a, a known aliquot volume of the suspension is transferred to a second container. In step 2509b, the amount of liquid in the second container is adjusted to a predetermined amount, and in step 2509c, a third number of microLED elements per unit volume of the suspension in the second container is calculated. In step 2509d, the suspension in the second container is dispersed on the top surface of the light-emitting display panel.

In another embodiment, the amount of the suspension in the first container is modified in step 2516, the suspension is stirred, and then the density of the suspension is optically measured in step 2518 to calculate the fourth number of microLED elements per unit volume of the suspension.

Systems and methods for collecting and dispersing microLEDs have already been provided. Examples of specific process steps and hardware units of the present invention have already been provided. However, the present invention is not limited to these examples. Those skilled in the art will devise other variations and embodiments of the present invention.

Claims (22)

A method for selectively collecting microLED elements from a carrier substrate,
The above method,
A step of providing an inorganic microLED element adhesively attached to a carrier substrate;
A step of predetermining a defective region comprising a region composed of a plurality of adjacent defective microLED elements or a region composed of a process control structure;
A step of forming a content-sensitive colloidal material covering a predetermined defective area;
A step of dissolving the exposed adhesive with an adhesive dissolving solvent;
A step of separating a microLED element located outside a predetermined defect area from a carrier substrate; and
A method for selectively collecting microLED elements from a carrier substrate, characterized by comprising a step of maintaining attachment by adhesion of the microLED elements and the carrier substrate within a predetermined defect area.
In the first paragraph,
a step of inspecting the carrier substrate to determine the location of a defective microLED element; and
Further comprising the step of locating a non-predetermined defect region comprising a plurality of adjacent defective microLED elements;
A method for selectively collecting microLED elements from a carrier substrate, characterized in that the step of forming the content-sensitive colloidal material includes the step of covering a non-determined defect area with the content-sensitive colloidal material.
In the second paragraph,
In response to the above inspection, a step of positioning a microLED element having an individual defect that is not predetermined; and
A method for selectively collecting microLED devices from a carrier substrate, further comprising the step of releasing said individual defective microLED devices using a laser trimming process.
In the second paragraph,
A method for locating a defective microLED element in response to an inspection,
A method for selectively collecting microLED devices from a carrier substrate, characterized by comprising a step of using an inspection process selected from optical comparison, electroluminescence, photoluminescence or cathodoluminescence testing.
In the second paragraph,
A method for selectively collecting microLED devices from a carrier substrate, characterized in that the method further comprises the step of collecting valid microLED devices into a collection vessel in response to separation of the microLED devices from the carrier substrate.
In paragraph 5,
Further comprising the step of applying additional power selected from the group consisting of fluid circulation, thermal energy, gravity, vibration and combinations thereof;
A method for selectively collecting microLED devices from a carrier substrate, wherein the step of separating the microLED devices from the carrier substrate comprises the step of at least partially separating the microLED devices in response to the additional power.
In paragraph 5,
A method for selectively collecting microLED devices from a carrier substrate, wherein the step of collecting valid microLED devices into the collection vessel comprises the step of replacing the adhesive dissolving solvent with another liquid.
In Article 7,
The step of collecting valid microLED devices into the collection vessel comprises the step of generating a suspension of valid microLED devices having an average cross-sectional physical size s, wherein the suspension includes impurities;
The above method,
A method for selectively collecting microLED elements from a carrier substrate, characterized in that after collecting the valid microLED elements, the method further comprises a step of filtering to remove impurities having a maximum cross-sectional physical size greater than t, wherein t>s.
In Article 7,
The step of collecting the valid microLED devices into the collection vessel comprises the step of generating a suspension of the valid microLED devices having an average cross-sectional physical size s, wherein the suspension includes impurities;
The above method,
A method for selectively collecting microLED elements from a carrier substrate, further comprising a step of filtering to remove impurities having a maximum cross-sectional physical size smaller than p, wherein p<s.
In Article 8,
A method for selectively collecting microLED elements from a carrier substrate, wherein the filtration process is selected from the group consisting of mechanical screening, elution, classification, or a combination thereof.
In Article 7,
The step of replacing said adhesive dissolving solvent with another liquid comprises the step of replacing said adhesive dissolving solvent with a filtered solution having a lower viscosity than said adhesive dissolving solvent;
The above method,
A method for selectively collecting microLED elements from a carrier substrate, characterized in that it further comprises a step of filtering to remove impurities from the filter solution.
In Article 7,
A method for selectively collecting microLED elements from a carrier substrate, characterized in that the step of replacing the adhesive dissolution solvent with another liquid comprises the step of replacing the adhesive dissolution solvent with an assembly solution selected from the group consisting of an assembly solution having a lower polarity than the adhesive dissolution solvent or an assembly solution having a faster evaporation rate than the adhesive dissolution solvent.
In Article 12,
A method for selectively collecting microLED elements from a carrier substrate, wherein the step of replacing the adhesive dissolving solvent with the assembly solution comprises the step of adding a surfactant to the assembly solution, wherein the surfactant is selected from the group consisting of anionic, cationic, nonionic surfactants, or combinations thereof.
In the first paragraph,
The step of dissolving the exposed adhesive comprises the step of selectively exposing a portion of the carrier substrate to the adhesive dissolving solvent;
A method for selectively collecting microLED devices from a carrier substrate, wherein the step of separating the microLED devices from the carrier substrate comprises the step of separating the microLED devices from selectively exposed portions of the carrier substrate.
In Article 14,
The step of selectively exposing a portion of the carrier substrate to the solvent comprises:
A step of rotating the carrier substrate in a solvent bath; and
A step of exposing a radial portion having a radius greater than d on the carrier substrate to the solvent bath;
A method for selectively collecting microLED devices from a carrier substrate, wherein the step of separating the microLED device from the exposed portion of the carrier substrate comprises the step of separating the microLED device from the radial portion of the carrier substrate.
In the first paragraph,
A method for selectively collecting microLED devices from a carrier substrate, wherein the step of exposing the microLED devices to the adhesive dissolving solvent comprises a step of selecting the adhesive dissolving solvent from the group consisting of acetone, toluene, trichloroethane, N-methylpyrrolidone, xylene, cyclohexanone, butyl acetate, or combinations thereof.
A system for selectively collecting microLED elements from a carrier substrate,
The above system,
A vacuum chuck having a rotational interface for mounting a carrier substrate comprising inorganic microLED elements adhesively attached to the carrier substrate;
An elbow connected to said vacuum chuck, said elbow having a plurality of optional settings for determining a rotation angle of said vacuum chuck in a plane where the x-axis and the z-axis are located;
A gantry connected to said elbow having a plurality of optional settings for determining the height of said vacuum chuck along the z-axis;
A tray of adhesive dissolving solvent having a top opening for receiving the carrier substrate; and
A controller is provided having output terminals connected to the gantry and the elbow to respectively provide setting of the height and setting of the angle;
In response to the setting of the gantry and the elbow, a selected radial portion of the carrier substrate is exposed to the adhesive dissolving solvent;
A system for selectively collecting microLED elements from a carrier substrate, characterized in that microLED elements separated from selectively exposed portions of the carrier substrate are collected on the tray.
In Article 17,
A system for selectively collecting microLED devices from a carrier substrate, wherein the controller has an input for receiving a first map of a microLED performance region, and in response to the first map, the controller provides settings of the gantry and settings of the elbow for selecting a radial region exposed to an adhesive dissolution solvent on the carrier substrate.
In Article 17,
Further comprising an inspection subsystem having an optical input terminal and an output terminal connected to the controller for identifying individual defective microLED elements in the carrier substrate;
The above system,
A system for selectively collecting microLED elements from a carrier substrate, further comprising a trimming laser having an input terminal connected to said controller to receive a second map of defective microLED elements, and an output terminal responsive to said second map to emit the defective microLED elements from the carrier substrate via laser radiation.
A system for selectively separating an area of a carrier substrate of a microLED,
The above system,
A controller having an output terminal for providing a first map of a predetermined defect area on a carrier substrate;
A printer comprising an input terminal for receiving the first map, and a nozzle for printing a content-sensitive colloidal material on a selected area of the carrier substrate in response to the first map;
A system for selectively separating a region of a carrier substrate from microLEDs, wherein the microLEDs in the selected region are still attached to the carrier substrate despite being exposed to an adhesive dissolving solvent.
In Article 20,
A system for selectively separating an area of a carrier substrate of a microLED, wherein the content-sensitive colloidal material is selected from the group consisting of SU-8, epoxy resin, polyethylene terephthalate, acrylonitrile butadiene styrene or polyimide.
In Article 20,
Further comprising an inspection subsystem having an optical input terminal and an output terminal connected to the controller for identifying an area of the microLED element having a non-determined defect on the carrier substrate;
A system for selectively separating an area of a carrier substrate of microLEDs, characterized in that the printer receives a second map of an area of the microLED element having the non-determined defect from the controller, and prints the content-sensitive colloidal material in response to the second map on an area of the microLED element having the detected defect.