KR20130064010A - Methods of characterizing semiconductor light-emitting devices based on product wafer characteristics - Google Patents

Abstract

A method of characterizing a semiconductor light emitting device (LED) based on product wafer characteristics is disclosed. The method includes measuring at least one product wafer characteristic such as curvature or device layer stress. The method also includes establishing a relationship between the at least one characteristic and the emission wavelengths of the LED dies formed from the product wafer. The relationship makes it possible to predict the emission wavelength of the LED structures formed in the device layer of similarly formed product wafers. This in turn can be used to characterize product wafers and especially LED structures formed thereon, and to perform process control in high volume LED manufacturing.

Description

METHODS OF CHARACTERIZING SEMICONDUCTOR LIGHT-EMITTING DEVICES BASED ON PRODUCT WAFER CHARACTERISTICS}

FIELD OF THE INVENTION The present invention generally relates to the manufacture of semiconductor light-emitting devices (LEDs), such as light-emitting diodes and laser diodes, and more particularly to characterizing semiconductor LEDs based on product wafer characteristics. It is about how to analyze.

Semiconductor LEDs, such as light emitting diodes and laser diodes, are rapidly replacing conventional light sources in almost a full range of light and lighting applications. As a result, they are increasingly being manufactured for a very wide range of emission wavelengths.

Examples of semiconductor LEDs for general illumination are light emitting diodes (also called "LEDs") and diode lasers. Phosphor coatings can be used to produce a white light spectrum. The phosphor reacts with a specific emission wavelength (eg, blue wavelength) of the device and Stokes shifts some of the emitted light from short to long wavelengths to provide output light in the white spectrum. The white spectrum is characterized by equal color temperatures associated with the emission light spectrum of the corresponding blackbody radiation. A "warm" white light spectrum has a color temperature of about 2800 ° K, while a "cold" white light spectrum has a color temperature of about 5000 ° K. In many applications, the warm white light spectrum is preferred.

In order to obtain a suitable color temperature, the emission wavelength λ _E of the semiconductor light emitting device needs to match the absorption and emission spectrum Δλ of the phosphor. In general, the actual emission wavelength [lambda] _E needs to be in the range of +/- 2 nm of the desired (selected) emission wavelength [lambda] _{ED in} order to properly match the absorption and emission characteristics of the phosphor. When properly matched, the LED lighting device provides "white light" with a color temperature of about 2800 K. Devices that fall outside certain wavelength specifications have significantly lower value because they produce light of poor color (“off-color”) and are therefore less preferred to consumers. LED manufacturers often sell these "off-color" LEDs for less color-sensitive applications such as flashlights or exterior parking lot facilities. However, the value of these LEDs is much lower than those sold in the general home lighting market where color temperature is sensitive. For this reason, LED manufacturers try to manufacture more LEDs per wafer that fall within the more valuable spectral range.

For optimal products and thereby optimal value and benefit, it is necessary to manufacture semiconductor light emitting devices to have the correct emission wavelength within a certain tolerance. In addition, it is necessary to know in advance the emission wavelength of the light emitting chip (die) to be the final semiconductor light emitting device.

A method of characterizing semiconductor light-emitting devices is disclosed based on product wafer characteristics. The method includes measuring one or more product wafer characteristics such as curvature or device layer stress. The method also includes establishing a relationship between the emission wavelength of the LED dies formed from the product wafer and the one or more characteristics. This relationship makes it possible to predict the emission wavelength of the LED structures formed in the device layer of the similarly formed product wafer. This in turn makes it possible to characterize product wafers and to perform process control in high volume LED manufacturing.

The method according to one aspect of the present invention measures the device-layer stress (S (x, y)) of a prototype wafer, dicing the prototype wafer to form a die, and their device-layer stress and corresponding By measuring the emission wavelength of the dies as a function of the device structure positions (x, y), the device-layer stress (S (x, y)) and emission wavelength λ _E (x, y)). The method further includes measuring a curvature change ΔC (x, y) of the prototype wafer based on the curvature measurement of the substrate performed before and after substrate processing to form the device structures. The method also includes calculating the device-layer stress (S (x, y)) of the prototype wafer based on the measured change in curvature ΔC. And the method also provides device-layer stress (S (x, y) and (x, y) locations on the prototype wafer to establish the expected emission wavelength for the device structures corresponding to the emission wavelength λ _E. Associating the device structures with the actual emission wavelength λ _E based on the relationship between)).

The method preferably further comprises the step of calculating the stress S (x, y) of the prototype wafer based on the measured change in curvature ΔC by the following equation.

S (x, y) = {M _S h _S ² / 6h _F } △ C (x, y)

Where M _S is the biaxial modulus of the substrate, h _S is the height of the substrate, and h _F is the thickness of the device structure.

In the above method, the device structures may comprise either a light emitting diode (LED) or a laser diode configuration.

The method preferably comprises establishing a desired emission wavelength; Dicing the product wafer to form a die from the device structure; And selecting the die based on the predicted emission wavelength.

The method preferably comprises defining a tolerance in the amount of change in emission wavelength; And comparing the predicted emission wavelength with a tolerance and determining from the comparison a workflow comprising any one of discarding the product wafer, reworking the product wafer, or using a selected portion of the product wafer. .

In the method, the substrate treatment may include performing an organometallic chemical vapor deposition (MOCVD) process.

Another aspect of the invention is a method of characterizing the emission wavelength of a semiconductor light emitting device (LED) structure of a product wafer having a substrate and a device layer. The method includes measuring one or more characteristics of a prototype wafer formed in substantially the same manner as the product wafer, wherein the one or more characteristics are selected from a group of characteristics including device layer stress and product wafer curvature. The method also includes dicing the one or more prototype wafers to form LED dies associated with locations on the prototype wafer. The method also includes measuring the LED emission wavelength for each of the LED dies to establish a set of LED emission wavelengths that vary with location on the product wafer. The method further includes determining a relationship between the one or more prototype wafer characteristics, the LED emission wavelength set, and the location of the LED dies; And using the relationship determined in the " determining relationship " step to predict the LED emission wavelength of the LED structures formed on the product wafer.

The method preferably further comprises binning the LED structures of the product wafer based on the relationship determined in the " determining relationship " step.

In the method, the product wafer comprises a substrate. And the method preferably further comprises measuring the curvature of the substrate prior to forming the product wafer.

The method preferably further comprises measuring at least one of the substrate curvature and the product wafer curvature using coherent gradient sensing (CGS).

In the method, the device structure has a dimension. And the method includes determining the one or more prototype wafer characteristics to a dimension substantially equal to or less than the device structure dimension.

Another aspect of the invention is a method of forming a semiconductor light emitting device (LED). The method includes forming a product wafer comprising LED structures formed on a substrate by a process having one or more processing variables, the substrate having a known initial curvature C ₀ (x) prior to forming the LED structure on the substrate. , y) The method also measures the curvature C (x, y) of the product wafer after forming the light emitting device structures on the substrate, and the curvature change ΔC (x, y) = C (x, y) -C Determining ₀ (x, y). The method also includes calculating the stress S (x, y) of the product wafer based on the measured curvature change ΔC (x, y). The method also determines the emission wavelength λ _E of the LED structures based on the relationship between the calculated stress (S (x, y)) and the (x, y) positions of the light emitting device structures on the product wafer. Associating the calculated stress S (x, y). And the method also includes comparing the emission wavelength λ _E with an emission wavelength change tolerance and storing the LED structures in one or more bins based on the tolerance.

The method preferably further comprises using only LED structures in one or more selected reservoirs to form an LED device.

The method preferably further comprises adjusting at least one of the one or more processing variables to reduce the emission wavelength variation of the LED structure as a function of the (x, y) position of the LED structure on the product wafer. To include.

In the method, the processing variables are selected from a set of processing variables including time, temperature, temperature uniformity, gas partial pressure, gas partial pressure uniformity, gas flow rate and gas flow uniformity.

The method preferably comprises performing curvature measurements of one or more prototype wafers formed in the same manner as the product wafer, dicing the one or more prototype wafers to form an LED die, the measured for emission wavelength Measuring the emission wavelength of the LED dies as a function of their (x, y) position on the one or more prototype wafers with respect to prototype wafer curvature. will be.

The method preferably determines the LED emission wavelengths of the LED structures on the new product wafer based on forming a new product wafer, measuring the curvature of the new product wafer, and based on the measured curvature of the new product wafer. Further comprising the step of determining.

Another aspect of the invention is a method of forming a semiconductor light emitting device (LED). The method includes forming a product wafer comprising LED structures formed on a semiconductor substrate by a process having one or more processing variables, measuring the curvature uniformity of the product wafer after forming the LED structure on the substrate, and curvature. Adjusting at least one of the one or more processing parameters to meet at least one of a uniformity requirement and a stress uniformity requirement.

Another aspect of the invention is a method of forming a semiconductor light emitting device (LED). The method includes forming a product wafer comprising LED structures formed on a substrate by a process having one or more processing variables, the substrate having a known initial curvature C ₀ (x) prior to forming the LED structure on the substrate. , y) The method also includes measuring the curvature C (x, y) of the product wafer after forming the LED structure on the substrate, and based on C (x, y) and C ₀ (x, y). Determining curvature uniformity. The method also includes determining a number of first dies that fall within a range of first curvature uniformity. The method also includes determining a number of first dies that fall within a range of second curvature uniformity. And the method also includes assigning a quality value to the product wafer based on the first and second numbers.

The method preferably further comprises disposing the product wafer based on the quality value imparted.

The method preferably further comprises dicing the product wafer to form an LED die associated with the first and second numbers. The method preferably further comprises using the LED die associated with the first number for a first use and using the LED die associated with the second number for a second use.

Another aspect of the invention is a method of forming a semiconductor light emitting device (LED). The method includes forming a product wafer comprising LED structures formed on a substrate by a process having one or more processing variables, the substrate having a known initial curvature C ₀ (x) prior to forming the LED structure on the substrate. , y) The method also measures the curvature C (x, y) of the product wafer after forming the LED structure on the substrate, and the curvature change ΔC (x, y) = C (x, y) − Determining C ₀ (x, y). The method also calculates the curvature C (x, y) and the LED based on the relationship between the (x, y) position of the LED structure on the product wafer and the calculated curvature C (x, y). Associating the emission wavelength λ _E of the structure. And the method also includes comparing the emission wavelength λ _E with the emission wavelength change tolerance to determine which LED structure can be used to form the LED.

Additional features and advantages of the invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention as set forth herein, including the following detailed description, As will be appreciated by those skilled in the art.

The foregoing general description and the following detailed description are intended to provide an overview or framework for presenting embodiments of the invention and for understanding the nature and properties of the invention as claimed. The accompanying drawings are part of this specification as an aid to understanding the present invention. The drawings illustrate several embodiments of the invention and together with the description serve to explain the operation and principles of the invention.

 According to the present invention, it is possible to manufacture more LEDs per wafer with an accurate emission wavelength within a certain tolerance and within a more valuable emission wavelength spectral range for optimum product and thus optimal value and benefit.

1 is a plan view of an example product wafer used in the formation of a semiconductor light emitting device,
2 is a cross-sectional view of an example product wafer used in the formation of a semiconductor light emitting device,
3 is a perspective view of an LED device showing how a die from a product wafer is included in a light emitting device structure;
4 is a collection of sapphire substrates supported in a susceptor,
5 is an enlarged cross sectional view of a portion of the susceptor showing how the substrate is supported inside the susceptor, FIG.
6 illustrates a MOCVD reactor system used to form a device layer over a substrate to form the product wafer of FIG. 2, and FIG.
7A-7D are plan views of product wafers showing contours showing changes in product wafer temperature (T), device dimensions (D), device layer stress (S), and emission wavelength (λ _E ), respectively. ,
8A-8C are plan views of product wafers showing equivalent lines showing changes in product wafer curvature C, device layer stress S, and emission wavelength λ _E , respectively;
FIG. 8D is a plan view of a product wafer similar to FIG. 8C, showing an example where a predicted emission wavelength is shown, with equivalent lines at 457 nm and 455 nm showing boundaries with device structures having a desired wavelength of 456 nm +/− 1 nm. It's a thick line, so
9 shows an example of a coherent gradient sensing (CGS) system used in examples of measuring substrate and product wafer curvature,
10 is a flowchart illustrating an example of a method of predicting the emission wavelength of dies from a product wafer based on the curvature measurements of the product wafer.
In some drawings a rectangular coordinate system is shown to provide a reference but is not intended to be limiting on orientation or configuration.

Reference will now be made in detail to preferred embodiments of the invention, which are illustrated in the accompanying drawings. Wherever possible, the same reference numbers and symbols are used for similar or identical parts throughout the figures. The claims are included in and constitute a part of the following detailed description.

In the following description, device-layer stress is the stress in the device layer 30 that defines the devices 50 formed therein, and S is the shorthand notation of S (x, y; σ _{i, j} ). Expressed as (x, y) or just S, where σ _{i, j} is a stress tensor. The stress tensor σ _{i, j} defines stress at any point in the device layer of the substrate. Coordinates (x, y) represent the wafer rectangular coordinates of the point in device layer 30 where stress is defined. The stress tensor (σ _{i, j} ) is composed of three orthogonal normal stress components (σ ₁₁ , σ ₂₂ , σ ₃₃ ) and six orthogonal shear stresses (σ ₁₂ , σ ₁₃ , σ _21, σ ₂₃ , σ ₃₁ , σ ₃₂ ). When the coordinate system used to measure stress is directed in the basic direction, the stress tensor has only three orthogonal vertical stress components. In one embodiment, S (x, y) has a single value representing stress, which single value is for example the maximum stress component, the sum of the stress components, the mean of the stress components, or any combination of stress components. Can be. In another example where the rate of change of stress at a given (x, y) position is employed, the notation S '(x, y) is used.

Also in the discussion below, the curvature of the surface C (x, y) is a short axis of C (x, y; n), where n is a vertical vector in a given direction in the xy plane and intersects the surface at right angles to the xy plane Define the plane, where C (x, y) is the curvature of the curve created when the surface intersects the plane. The curvature is generally defined as 1 / R, where R is the local radius of curvature of the surface at the point (x, y) in the plane defined by the vertical vector n. In one embodiment, n is taken along a single direction, in which case C (x, y) is a scalar function. In general, C (x, y) is a tensor because it can have multiple curvatures associated with it (ie, one in each direction relative to the unit vector n). In one embodiment the curvature C (x, y) can be determined from a function H (x, y) that defines the height of each (x, y) surface point measured for a certain reference plane (eg, a fully planar wafer). . Specifically, C (x, y) fits a circular curve to the points surrounding a given plane ‚‚ ´ point (x, y) to obtain the optimum radius R, and then determines the curvature as 1 / R, where H ( x, y) can be obtained. C (x, y) can also be obtained from the second derivative of H (x, y) using standard mathematical techniques. The notation C ₀ is used to indicate the curvature of the substrate used to form the product wafer.

In the present specification, the abbreviation “LED” is generally understood to mean “light emitting device”, but may also mean a light emitting diode, and those skilled in the art will understand the difference based on the context in which the abbreviation is used.

1 is a plan view of an embodiment product wafer used in the formation of semiconductor light emitting devices such as light emitting diodes and laser diodes, and FIG. 2 is a cross sectional view of the product wafer used in the formation of semiconductor light emitting devices. Here, the term "product wafer" generally refers to a wafer or substrate on which device structures are formed, and then the device structures, in one embodiment, emit light that can be used to form an LED product or device. Used to make devices.

The product wafer 10 includes a semiconductor substrate 20 having an edge 21, an upper surface 22, and a lower surface 24, on which a device layer 30 is formed. The semiconductor substrate 20 is made of sapphire or silicon, for example. Device layer 30 includes an array 32 of semiconductor light emitting device (LED) structures (“device structures”) 40. In one embodiment, product wafer 10 includes thousands of device structures 40, which may have a size of about 1 mm. The product wafer 10 has a diameter of 2 to 6 inches for the sapphire substrate 20 and a diameter of 6 to 12 inches for the silicon substrate. The device structure 40 has an actual emission wavelength λ _E associated therewith and an output spectrum Δλ associated with the aforementioned color temperature. The desired or selected emission wavelength for device structure 40 is denoted as λ _ED .

Once the product wafer 10 is fully processed so that the device structure 40 has a function, the product wafer 10 is cut so that the individual device structures in the array 32 are separated to form individual dies 42 (dicing). (Diced). Dies 42 are then included in the semiconductor light emitting device structure to form a light emitting device 50, such as shown in FIG. 3 as a perspective view of the LED device. The LED device 50 of FIG. 3 includes an anode 52A and a cathode 52C extending into the interior 54 of the epoxy lens sheath 56. Cathode 52C includes reflective cavity 58 where die 42 is located. Wire bond 60 electrically connects anode 52A and cathode 52C to the LED. A power source (not shown) is connected to anode 52A and cathode 52C to supply the electricity needed to operate the LED so that the LED emits light 62 at the actual wavelength λ _E. The LED device 50 of FIG. 3 is a prior art device when the die 42 is a prior art die, but is not a prior art device when the die 42 is formed using the methods disclosed herein.

An example of the device structure 40 is in the form of an LED made by growing GaN on a sapphire substrate 20. GaN is grown using metalorganic chemical vapor deposition (MOCVD). The MOCVD process is performed in a MOCVD reactor and in such a way that a multi-quantum-well (MQW) structure (not shown) is formed.

4 shows a collection of sapphire substrates 20 supported in a susceptor 70 having a substrate holder or top surface 72. 5 is an enlarged cross sectional view of the susceptor 70 showing how the substrate is supported therein. The susceptor 70 includes an opening 76 formed in the upper surface 72, the opening lip supporting the substrate 20 at the lower surface 21 near the edge 21 ( 78).

6 illustrates a MOCVD reactor system 90 having a reactor chamber 100 operatively connected to the MOCVD subsystem 96. The MOCVD subsystem 96 includes various MOCVD system components (not shown), such as a vacuum pump, gas source, exhaust system, baratron, and the like. The MOCVD chamber 100 has an interior 104 where the susceptor 70 and the substrate 20 are located when the MOCVD process is performed. After the MOCVD process is performed, the substrate 20 becomes a product wafer 10. The MOCVD reactor system 90 includes a controller 110 operatively connected to the MOCVD subsystem 96. Controller 110 is configured to control the MOCVD process (such as indicated by arrow 105) that takes place inside reactor 104.

Specifically, the controller 110 is used to carefully control the temperature T of the product wafer 10 because the actual emission wavelength (λ _E ) changes significantly as a function of the MOCVD growth conditions. 7d is the contours of product wafer temperature T (x, y), device dimension D (x, y), device layer stress S (x, y) and (actual) emission wavelength λ _E (x, y), respectively. Is a top view of a product wafer embodiment showing an embodiment of where temperature T refers to the substrate temperature or product wafer temperature, because substrate 20 is subjected to processing to form product wafer 10. The following description is referred to as "product wafer temperature" for convenience of description.

As the product wafer temperature T (x, y) changes (FIG. 7A), the device dimension D (x, y), for example, the thickness of the multi-quantum well structure (not shown) formed during the MOCVD process is corresponding. In a way (Figure 7b). This is also interpreted as a corresponding change in device layer stress S (x, y) for the product wafer 10 (FIG. 7C), which in turn results in a corresponding change in the actual emission wavelength λ _E (x, y) ( 7d).

In some cases, a change of 1 ° K in the product wafer temperature T may cause a transition (δλ _E ) of about 1 nm at the emission wavelength (λ _E ), therefore, appropriate for the product wafer temperature T To ensure control, it is desirable to first detect and finally control temperature non-uniformity and temperature repeatability in the MOCVD reactor system 90. Temperature non-uniformity on the product wafer can lead to local changes in growth conditions, thereby This can cause a change in the LED emission wavelength.

In the manufacture of today's MOCVD-based device structures 40, the actual emission wavelength λ _E and the corresponding emission uniformity are such that the device structure 40 is diced from the product wafer 10 to form the die 42. Then, the die is not known until it is included in a light emitting device, such as the LED device 50 of FIG. 3 or an equivalent test structure.

Currently, in order to estimate or predict the LED emission wavelength (λ _E ), the substrates are examined by photoluminescence technique, in which a short wavelength source (typically 248 nm) is incident on the quantum well region to emit the emission. Stimulate However, a serious limitation of this technique is that it is a point-by-point technique. In order to accurately map the entire product wafer at high spatial resolution (e.g., less than the die size), it takes between 30 and 240 minutes, depending on the substrate size used to form the product wafer.

 An additional limitation to this technique is that the emission wavelength of the photoluminescence is generally not the same emission wavelength from the LED during the electrical simulation. This difference is believed to result from the additional manufacturing steps that the LED experiences between photoluminescence inspection and the final product. Typically, there is an offset between the electrically simulated LED emission wavelength and the photoluminescence emission wavelength that are measured in advance in production. This is then used as a process monitor.

The amount of device layer stress S (x, y) of the device structures 40 formed on the product wafer 10 is directly related to the history of the product wafer temperature T (x, y). That is, T (x, y, t), where t is time. In essence, the difference in product wafer temperature (ie, non-uniform heating) during the deposition process results in a change in thickness in the deposited GaN layers and in particular the quantum well layers. After deposition and when the substrate returns to room temperature, the difference in the coefficient of thermal expansion of the component materials that make up the device structures 40 results in relative expansion or contraction of the device structures 40 relative to the substrate 20 during the thermal cycle. . Non-uniform heating leads to non-uniform device layer stress and non-uniform emission wavelength λ _E.

This relative expansion or contraction causes stress in the device structure 40 formed on the product wafer 10 such that thermal energy delivered to the product wafer is partially stored as mechanical energy in the device layer 30. Non-uniformity (ie, change) of the product wafer temperature T (x, y) during deposition results in non-uniformity (variation) in layer thickness and ultimately in device-layer stress S (x, y). The product wafer 10 with uniform but different temperatures T (x, y) will have a uniform but different device-layer stress S (x, y).

The device-layer stress S (x, y) can be monitored through the change ΔC (x, y) of the substrate curvature. The most common stress metrology system relies on the measurement of changes in product wafer geometry to calculate device-layer stress. The device-layer stress S (x, y) in the device structure 40 supported by the product wafer 10 is calculated from the curvature change ΔC (x, y) using the Stony formula:

S (x, y) = { M S h S 2 / 6h F} △ C (x, y) ( Equation 1)

Where M is biaxial modulus, h is thickness, subscript "F" is deposited film (here GaN) and subscript "s" means substrate (here sapphire or silicon). Changes in C (x, y) for the wafer indicate non-uniform device-layer stress (S).

Knowing that device-layer stress S may occur due to a mismatch of temperature and coefficient of thermal expansion α between the film (ie device layer 30) and the underlying substrate 20, device-layer stress S ) And the curvature (C) it is possible to derive the following equations:

(식 2)

(Equation 2)

(식 3)

(Equation 3)

Where ε _m is the mismatch or misfit strain between the film and the substrate, and ΔT is the temperature difference from the stress-free temperature T ₀ (ie, the temperature at which the film and substrate are matched).

The above equations show how the curvature change ΔC (x, y) is related to the change in deposition temperature. Therefore, from these equations, the product wafer curvature can be related to temperature, which is directly related to the emission wavelength λ _E of the LED.

The equations provided above are based on certain assumptions about stress and strain states, the geometry of the device structure and the thermal-mechanical properties of the component materials. More complex forms or versions of these equations can be developed for film / substrate systems with arbitrary properties. However, for any particular device / substrate system, an appropriate relationship can be developed according to the lines presented herein to relate the substrate curvature change with temperature change to the emission wavelength change.

One aspect of the invention relates to at least one product wafer characteristic, such as product wafer stress (ie, one or more components of the stress tensor), and at least one LED performance characteristic such as emission power, efficiency, wavelength, spectral bandwidth, and the like. Correlated. In one embodiment, this is accomplished by empirically establishing a relationship between the at least one measured performance characteristic of the LED and the at least one product wafer characteristic. This can be achieved by evaluating at least one performance characteristic of the LED obtained from the product wafer for the individual processing steps used to form the product wafer in one embodiment. This is also achievable by reducing at least one LED performance characteristic for the multiple processing steps used to form the product wafer.

Accordingly, one aspect of the invention is a method of employing at least one of product wafer stress and product wafer curvature as a process monitor to perform quality control to store or bin the performance of individual dies of a product wafer. It includes. The method may further comprise monitoring deposition characteristics during product wafer formation and performing process optimization. The method further includes matching the performance of a set of two or more of the same type of processing tools such that product wafers and their LED dies are also matched when formed in different processing tools. That is, product wafers formed by the different sets of matched tools are more similar than when processing tools were not matched.

8A-8C are plan views of example product wafers 10 showing examples of equivalent lines of curvature C (x, y), device layer stress S (x, y), and the actual emission wavelength λ _E (x, y), respectively. . One aspect of the invention is to determine the device layer stress S (x, y) by measuring the product wafer curvature C (x, y), and then using the determined device layer stress S (x, y), Predicting the actual emission wavelength λ _{E of the} formed (separated) dies 42 as a function of the position (x, y) of the corresponding device structures 40 of the wafer 10.

In FIG. 8C, one specific device structure 40 is shown having a location x _i , y _j , which is measured relative to a reference location, eg, the center of the product wafer. Enlarged view of the array 32 of the device structure 40 is a more detailed (i.e., closely-spaced) of the actual emission wavelength (λ _E) shows a 1 nm increments the equivalent line (λ _E). In one embodiment, the change (transition) of the emission wavelength of δλ _E of 1 nm can be predicted.

FIG. 8D is a top view of an assumed product wafer 10, showing an example of a 2 nm equivalent line of the predicted emission wavelength λ _EP , where the product wafer 10 has a concave or bowl-shaped curvature C (x, y) Assuming the desired emission wavelength of λ _ED = 456 nm with emission wavelength change tolerance (δλ) of +/− 1 nm, the predicted-wavelength equivalent of FIG. 8D dicing the product wafer 10, and the solid line Only dies 42 from device structure 40 having (x, y) positions within the predicted-wavelength equivalent line of 455 nm to 457 nm (ie, within emission wavelength change tolerance (δλ)) It is possible to use in the light emitting device 50. Thus, in one embodiment, the first plurality of LED structures 40 fall within the 455 nm to 457 nm equivalent line and the second plurality of LED structures 40 fall outside these equivalent lines. Selection of this type of LED structure 40 based on an optional specification, here the emission wavelength, may be used to deposit the formed LED die 42 for the first and second applications or for the first and second type of LED devices. Can be.

At this point, the number of dies 42 in the allowable wavelength change (i.e., δλ = +/- 1 nm) can be predicted. In addition, the number of dies 42 slightly out of this window but still useful in “off-color” applications is predictable. It is also possible to predict the number of dies 42 outside of any useful range and so not inherent in value. Knowing the value of the die 42 for each wavelength variation range and from the number of dies 42 within each wavelength variation range, it is possible to predict the inherent value of the product wafer 10. This makes it possible for the LED manufacturer to determine if there is sufficient reason to continue processing the particular product wafer 10. For example, if the die array falls outside of any valued wavelength range, there is little reason to continue processing wafer 10 and incur additional costs.

In summary, changes in deposition temperature during LED formation result in wavelength transitions of the LED's emission wavelength λ _E. These same temperature changes result in a change in stress in the deposited film, which results in a change in wafer curvature. Therefore, changes in wafer curvature C can be related to wavelength transition via the above equations.

In practice, it may be difficult to know the exact values for the material constants (such as the coefficient of thermal expansion α and the biaxial coefficient M) for the various materials at elevated temperatures used during the MOCVD process. For example, typical temperatures in the MOCVD process are often about 1000 ° C. Analytically correlating the above equations to the actual LED wavelength transition requires accurate knowledge of these material constants and the temperature history of the substrate 20. This information may not be available.

Therefore, empirically obtaining the necessary data and performing a precise measurement of the curvature (C) for all points on one or more prototype wafers 10 and the LED emission wavelengths λ _E of the LED dies 42 obtained from the product wafers 10. It may be more convenient to develop a lookup-table or correlation curve for the actual wavelength transition by measuring. This approach makes it possible to correlate the measured substrate curvature C to the actual device data and also to use the substrate curvature as a process control and inspection monitor for subsequent high volume LED fabrication.

The main advantage of using wafer curvature (C) to monitor emission wavelength (λ _E ) and emission uniformity over current photoluminescence techniques is that wafer curvature can be measured at very high spatial frequencies in a very short time. It is. The coherent-gradient sensing system (described below) allows for easy sampling of substrates with spatial sampling of hundreds of microns, lower than a typical die size of 1mm, and inspection (measurement) of 200-mm wafers in about one minute. It is possible to In comparison, it would take several hours to obtain similar spatial information for a 200-mm wafer using a photoluminescence system.

9 shows an example of a coherent gradient sensing (CGS) system 200 that can be used to measure the product wafer 10 curvature C (x, y). Details of how CGS sensing works are described in US Pat. No. 6,031,611 (hereinafter referred to as the '611 patent), the contents of which are incorporated herein by reference. 9 is based on FIG. 1 of the '611 patent.

The CGS system 200 includes a digital camera 210 along the A1 axis, a filtering lens 224 (eg, a filter combined with the lens described in the '611 patent and shown in FIG. 1 thereof), first and second axial spacing. Gratings G1 and G2, beam splitter 230, and wafer stage 240. The CGS system 200 also includes a laser 250 arranged along the optical axis A2 that intersects the A1 axis in the beam splitter 230. The CGS system 200 also includes a controller or signal processor 260 operatively connected to the digital camera 210 and the laser 250. Examples of signal processor 260 include a computer or computer having a processor 262 and a computer readable medium (memory) 264, which are product wafer curvature C (according to the methods described in the '611 patent). x, y) is set to control the operation of the CGS system 200 by the instructions recorded on the medium.

In operation, the laser 250 generates a collimated laser beam 252 which is steered to the product wafer 10 by the beam splitter 230. The collimated laser beam reflects from the product wafer 10 (particularly the device layer 30) as the reflected light 252R traveling upward through the beam splitter 230 and the gratings G1 and G2. The two gratings G1 and G2 are configured to shear apart and otherwise collimated laser beams. Light passing through gratings G1 and G2 is then focused onto digital camera 210 using filtering lens 224.

The shear and filtering process produces fringes in the digital camera 210 that are equivalent lines of constant surface tilt on the product wafer 10. As a result, two sets of orthogonal numbers to characterize surface gradients that can be numerically integrated to reconstruct substrate topography or that are numerically differential to obtain product wafer curvature C (x, y) Interferograms are used. Two camera configurations for the CGS system, such as shown in Figure 6 of the '611 patent, may also be employed to facilitate capturing two orthogonal sets of interferograms.

The curvature map, ie, product wafer curvature C (x, y), is then generated from the interference fringe pattern, using the method described in the '611 patent, such as performed in controller 260, for example. In one embodiment, the measured curvature C (x, y) of the product wafer 10 has a spatial resolution of 100 to 300 microns in size. Thus, for a device structure 40 having a dimension d of size 1 mm (see FIG. 3), there may be at least one curvature data point per device, and in some cases multiple data points per device (die) There may be. In one embodiment, the methods disclosed herein include determining a spatial change in stress S (x, y) in the product wafer to a dimension that is approximately equal to or less than the dimension d of the device structure 40.

In one embodiment, the curvature C ₀ (x, y) of the substrate 20 is measured prior to processing the substrate to form the product wafer 10. This enables the change in curvature ΔC (x, y) = C (x, y)-C ₀ (x, y) to be calculated.

Once the substrate and product wafer curvatures C ₀ (x, y) and C (x, y) are measured and the change in curvature ΔC (x, y) is calculated, the device-layer stress S (x, y) can be determined. have. In one embodiment, this is accomplished using the Stony equation (Equation 1), which relates the device-layer stress S (x, y) to the change in curvature ΔC.

For most product wafers 10, the change in curvature ΔC (x, y) changes with respect to the wafer and also changes with orientation, suggesting the presence of non-uniform device-layer stress S (x, y). do. Alternatively, if the material properties at each temperature are unknown, it is possible to measure the wavelength shift δλ relative to the curvature C measured on the sample substrate (wafer) and for future measurements on the product wafer 10. It is possible to relate these measurements together.

Other methods can be used for the measurements that enable the determination of C (x, y). Such methods include, for example, profilometry, interferometry, capacitance gauges, and laser beam deflection. In addition, the device-layer stress S (x, y) can be determined by measuring the distortion of the crystal lattice, for example by using X-ray diffraction and Raman spectroscopy.

Accordingly, one aspect of the method of the present invention includes the following.

Before processing, for example, before performing MOCVD, the shape (curvature) C ₀ (x, y) of the substrate 20 is measured.

The shape C (x, y) of the product wafer 10 is measured after the treatment.

DELTA C (x, y) = C (x, y)-C ₀ (to determine the change in curvature C associated with the product wafer 10 caused by processing the substrate 20 to produce a product wafer. x, y)

Using a stony formula or related formula (e.g., Blake's Equation) to calculate the process-induced device-layer stress S (x, y) on the product wafer based on ΔC (x, y) The device-layer stress S (x, y) of the product wafer is calculated.

Associate the device structure 40 located at position (x _i , y _j ) with the device-layer stress S (x _i , y _j ).

Associates the device-layer stress S (x _i , y _j ) with the emission wavelength λ _E (x _i , y _j ), which then of the LED structure 40 forms the LED die when the product wafer is diced. Enables measurement of product wafer curvature to predict emission wavelengths.

Step 6 above establishes a relationship between the actual emission wavelength (λ _E ), the curvature change (ΔC) and the device-layer stress S (x, y) for a given type of treatment and the device structure 40 produced. Needs one. This can be done if all material properties are well known and the relationship between temperature and deposition rate is well known. As mentioned above, these quantities may not be well known and it is necessary to empirically generate data to correlate the measured change in curvature to the emission wavelength change.

In one embodiment, this empirical "fingerprinting" process may be performed as follows:

Complete steps 1 through 4 for one or more prototype wafers 10 for a particular process to form device structure 40.

The actual emission wavelength λ _E of the LED dies obtained from the prototype wafer 10 is measured as a function of the device structure position (x, y) to establish λ _E (x, y).

Determine the relationship between the device-layer stress S (x, y) and the actual emission wavelength λ _E (x, y).

The methods of the present invention also provide process monitoring and process control of the product wafer 10 to determine (predict) the actual emission wavelength λ _E that the LED die 42 produce when the product wafer 10 is diced. Includes performing. This can be done for a given product wafer and between different product wafers for the same process (ie, the same device structures). Selected LED dies 42 characterized using the methods disclosed herein can then be used in selected types of LED devices 50 as shown in FIG. 3.

The methods of the present invention also include performing process optimization to form the device structure 40 by adjusting at least one of the process variables used to form the product wafer 10. This includes, for example:

Establish a relationship between process variables (eg temperature, temperature uniformity, gas partial pressure, gas partial pressure uniformity, flow rate, flow uniformity, time) and treatment-induced stress characteristics (eg average stress and stress uniformity) Making; And

Modifying at least one of the processing parameters to optimize the stress characteristics of the process providing the desired device characteristics, such that the emission wavelength is as close as possible to the desired emission wavelength.

The methods of the present invention directed to process optimization also include the following steps:

Establishing stress characteristics associated with the different processing tools used to carry out the same processing, such as the imagined MOCVD reactor system.

Identifying a particular processing tool that provides a minimum desirable stress characteristic, that is, a stress characteristic that causes a maximum amount of device performance change (eg, a maximum change in the actual emission wavelength).

Identifying processing tool parameters such as hardware or control settings, adjustments, etc. that affect stress characteristics (ie wafer temperature, wafer temperature uniformity, gas partial pressure, gas partial pressure uniformity, gas flow rate, gas flow uniformity, etc.).

Adjusting the processing tool parameters to obtain a reduced change in the emission wavelength λ _E , in particular in the specific example the emission wavelength λ _E of the LED dies minimizing the change.

10 illustrates an example of a method of predicting the actual emission wavelength λ _E for LED dies 42 formed from the product wafer 10 based on the curvature measurements of the product wafer and the substrate used to form the product wafer. A flowchart 300 is presented.

Flowchart 300 shows the relationship between device-layer stress S (x, y) and the actual emission wavelength λ _E (x, y) for dies 42 formed from device structure 40 on product wafer 10. The first step 302 of establishing a.

Flowchart 300 includes a second step 304 of measuring the curvature change ΔC (x, y) of the product wafer 10 based on the substrate curvature measurement C ₀ (x, y) as described above.

The flow chart 300 is based on the third step of calculating the device-layer stress S (x, y) of the product wafer 10 based on the measured curvature change ΔC (x, y), for example using the Stony equation. 306).

Flowchart 300 includes a fourth step 308 of associating the device structure 40 on the product wafer 10 on which dies 42 are formed with the calculated device-layer stress S (x, y). .

Flowchart 300 includes a fifth step 310 of associating the actual emission wavelength λ _E with the various dies 42 based on the device-layer stress S (x, y) established in step 302.

Flowchart 300 includes a sixth step 312 of adjusting at least one processing variable to reduce the change in emission wavelength λ _E of dies 42.

In another example of the methods disclosed herein, product wafer curvature measurements can be made at different stages of product wafer processing completion. This can provide insight into how product wafer curvature changes with each processing step.

Further, as described above, the flowchart 300 measures the substrate curvature C measured after precisely measuring the curvature C for a sample (set) of wafers and measuring their LED emission wavelength λ _E. As part of step 302, a sub-step may be empirically obtained from the prototype wafer 10 to correlate the actual device data with a correlation curve for actual wavelength transition or to generate a reference-table. This enables the use of product wafer curvature measurements as process control and inspection monitors for subsequent high volume LED device fabrication.

It will be apparent to those skilled in the art that various changes and modifications can be made herein without departing from the scope and spirit of the invention. Therefore, it is intended that the present invention cover the modifications and variations of this invention provided they come within the scope of the appended claims and their equivalents.

10 products wafers 20 substrates
21 Edge 22 Top
24 Bottom View 30 Device Layers
32 Device Structure Array 40 Device Structure
42 die 50 light emitting device
52A anode 52C cathode
54 inner 56 outer shell
58 Reflective Cavity 60 Wire Bond
62 Optical 70 Susceptor
72 Top face 78 Lip
90 MOCVD Reactor System 96 MOCVD Subsystem
100 reactor chamber 104 inside
110 Controller 210 Camera
230 beam splitter 224 filtering lens
240 wafer stage 250 laser
252 collimated laser beam 252R reflected light
260 signal processor 262 processor
264 memory A1, A2 optical axis
G1, G2 Grid

Claims (22)

A method of characterizing an emission wavelength of a semiconductor light emitting device structure of a product wafer having a substrate, the method comprising:
Measure the device-layer stress (S (x, y)) of the prototype wafer, dice the prototype wafer to form dies, and their device-layer stress and corresponding device structure location (x, y) By measuring the emission wavelength of the dies as a function of the relationship between the device-layer stress (S (x, y)) and the emission wavelength λ _E (x, y) for device structures formed on the product wafer Establishing;
Measuring a change in curvature (ΔC (x, y)) of the prototype wafer based on measurement of curvature of the substrate performed before and after substrate processing to form the device structures;
Calculating a device-layer stress (S (x, y)) of the prototype wafer based on the measured change in curvature ΔC; And
Relationship between (x, y) locations on the prototype wafer and device-layer stress (S (x, y)) to establish a predicted emission wavelength for device structures corresponding to emission wavelength λ _E Associating the device structures with the actual emission wavelength [lambda] _E based on the method. The method of claim 1,
And calculating the stress (S (x, y)) of the prototype wafer based on the measured change in curvature (ΔC) by the following equation.
S (x, y) = {M _S h _S ² / 6h _F } △ C (x, y)
Where M _S is the biaxial modulus of the substrate, h _S is the height of the substrate, and h _F is the thickness of the device structure. The method of claim 1,
Wherein said device structures have a light emitting diode (LED) or laser diode structure. The method of claim 1,
Establishing a desired emission wavelength;
Dicing a product wafer to form a die from the device structure; And
And selecting the die based on the predicted emission wavelength. The method of claim 1,
Defining a tolerance in the amount of change in emission wavelength; And
Comparing the predicted emission wavelength with the tolerance and determining from the comparison a workflow comprising any one of discarding the product wafer, reworking the product wafer, or using a selected portion of the product wafer. Further comprising emission wavelength characterization method. The method of claim 1,
Wherein said substrate treatment comprises performing an organometallic chemical vapor deposition (MOCVD) process. A method of characterizing an emission wavelength of a semiconductor light emitting device structure of a product wafer having a substrate and a device layer, the method comprising:
a) measuring one or more properties of a prototype wafer formed in substantially the same manner as the product wafer;
b) dicing the one or more prototype wafers to form LED dies associated with locations on the prototype wafer;
c) measuring the LED emission wavelength for each of the LED dies to establish a set of LED emission wavelengths that vary with location on the product wafer;
d) determining a relationship between the one or more prototype wafer characteristics, the LED emission wavelength set, and the location of the LED dies; And
e) using the relationship determined in step d) to predict the LED emission wavelength of the LED structures formed on the product wafer,
Wherein said at least one characteristic is selected from a group of characteristics comprising device layer stress and product wafer curvature. The method of claim 7, wherein
And emitting the LED structures of the product wafer based on the relationship determined in step d). The method of claim 7, wherein
The product wafer comprises a substrate,
The method comprises:
And measuring curvature of the substrate prior to forming the product wafer. The method of claim 9,
And measuring at least one of the substrate curvature and the product wafer curvature using coherent gradient sensing (CGS). The method of claim 9,
The device structure has a dimension,
The method comprises:
And determining the one or more prototype wafer characteristics up to a dimension substantially equal to or less than the device structure dimension. In the method of forming a semiconductor light emitting device (LED),
Forming a product wafer comprising LED structures formed over the substrate by a process having one or more processing variables;
After forming the light emitting device structures on the substrate, the curvature C (x, y) of the product wafer was measured, and the curvature change ΔC (x, y) = C (x, y)-C ₀ (x, determining y);
Calculating the stress S (x, y) of the product wafer based on the measured curvature change ΔC (x, y);
The calculated emission wavelength λ _E and the calculated stress based on the relationship between the calculated stress (S (x, y)) and the (x, y) positions of light emitting device structures on the product wafer Associating S (x, y); And
Comparing the emission wavelength λ _E with an emission wavelength change tolerance and storing the LED structures in one or more bins based on the tolerance,
In the forming a product wafer, the substrate has a known initial curvature C ₀ (x, y) prior to forming the LED structure on the substrate. 13. The method of claim 12,
And using only LED structures in one or more selected reservoirs to form LED devices. 13. The method of claim 12,
Further comprising adjusting at least one of the one or more processing variables to reduce the amount of emission wavelength variation of the LED structure as a function of the (x, y) position of the LED structure on the product wafer. Forming method. 15. The method of claim 14,
Wherein the processing variables are selected from a set of processing variables including time, temperature, temperature uniformity, gas partial pressure, gas partial pressure uniformity, gas flow rate and gas flow uniformity. 15. The method of claim 14,
Performing curvature measurements of one or more prototype wafers formed in the same manner as the product wafers;
Dicing the one or more prototype wafers to form LED dies; And
And measuring the emission wavelength of the LED dies as a function of their (x, y) location on the one or more prototype wafers relative to the measured prototype wafer curvature relative to the emission wavelength. 17. The method of claim 16,
Forming a new product wafer;
Measuring the curvature of the new product wafer; And
Determining the LED emission wavelength of the LED structures on the new product wafer based on the measured curvature of the new product wafer. In the method of forming a semiconductor light emitting device (LED),
Forming a product wafer comprising LED structures formed over the semiconductor substrate by a process having one or more processing parameters;
Measuring the curvature uniformity of the product wafer after forming an LED structure on the substrate; And
Adjusting at least one of the one or more processing parameters to meet at least one of a curvature uniformity requirement and a stress uniformity requirement. In the method of forming a semiconductor light emitting device (LED),
Forming a product wafer comprising LED structures formed over the substrate by a process having one or more processing parameters;
Measuring the curvature C (x, y) of the product wafer after forming the LED structure on the substrate, and determining the curvature uniformity based on C (x, y) and C ₀ (x, y) ;
Determining a number of first dies that fall within a range of first curvature uniformity;
Determining a number of first dies that fall within a range of second curvature uniformities; And
Assigning a quality value to the product wafer based on the first and second numbers,
In the forming a product wafer, the substrate has a known initial curvature C ₀ (x, y) prior to forming the LED structure on the substrate. The method of claim 19,
And disposing the product wafer based on the imparted quality value. The method of claim 19,
Dicing the product wafer to form an LED die associated with the first and second numbers; And
Using the LED die associated with the first number for a first use and using the LED die associated with the second number for a second use. In the method of forming a semiconductor light emitting device (LED),
Forming a product wafer comprising LED structures formed over the substrate by a process having one or more processing parameters;
After the LED structure is formed on the substrate, the curvature C (x, y) of the product wafer is measured, and the curvature change ΔC (x, y) = C (x, y)-C ₀ (x, y Determining);
The curvature (C (x, y)) and the emission wavelength of the LED structure calculated based on the relationship between the (x, y) position of the LED structure on the product wafer and the calculated curvature C (x, y) associating (λ _E ); And
Comparing the emission wavelength λ _E with an emission wavelength change tolerance to determine which LED structure can be used to form an LED,
In the forming a product wafer, the substrate has a known initial curvature C ₀ (x, y) prior to forming the LED structure on the substrate.

Images