PHOTOLUMINESCENCE IMAGING WITH PREFERENTIAL DETECTION OF

PHOTOLUMINESCENCE SIGNALS EMITTED FROM A SPECIFIED

MATERIAL LAYER OF A WAFER OR OTHER WORKPIECE

PRIORITY CLAIM

[0001] This application claims the benefit of U.S. Provisional Application

No. 60/696,853, filed July 6, 2005, which is incorporated herein by reference.

BACKGROUND

[0002] Photoluminescence imaging or photoluminescence spectroscopy is a contactless, nondestructive method of probing the electronic structure of materials, such as silicon semiconductor wafers, as well as other workpieces and materials. In a typical photoluminescence process, light is directed onto a wafer or other workpiece or sample (hereinafter collectively referred to as a "wafer"), where at least some of the light is absorbed. The absorbed light imparts excess energy into the material via a process of "photo-excitation." This excess energy is dissipated by the wafer through a series of pathways; one such pathway is the emission of light, or photoluminescence. The intensity and spectral content of this photoluminescence is directly related to various material properties of the wafer.

[0003] Photoluminescence imaging processes may be used to identify and quantify defects and contaminants present in the wafer based on spatial variations in the photoluminescence images produced. One photoluminescence imaging process, as described in International Application Number

PCT/GB97/02388 (publication number WO 98/11425), which is incorporated herein by reference, involves probing the surface and/or the sub-surface bulk region of the wafer with one or more lasers of varying excitation wavelengths. A laser of a given wavelength is directed into the wafer and penetrates the wafer to a given depth. Return light emitted from excited regions of the wafer is detected and quantified by a detection system. Images of the measured return light, including spatial images of defects and contaminants in the wafer, may then be produced by the detection system or by an associated image-producing system. [0004] While these photoluminescent images may effectively identify defects and contaminants in the wafer as a whole, it is sometimes difficult to readily identify the location and concentration of defects in a specific material layer of the wafer. For example, if two or more material layers of the wafer are penetrated by a laser, images of defects in the second material layer may obscure images of defects in the first material layer. This can be problematic if detailed defect data about only the first material layer is desired. Thus, there is a need to be able to obtain more accurate measurements of the location and concentration of defects in a specific material layer or layers of a wafer.

SUMMARY

[0005] A method and apparatus uses photoluminescence to identify defects in one or more specified material layers of a sample. One or more filtering elements are used to filter out predetermined wavelengths of return light emitted from a sample. The predetermined wavelengths are selected such that only return light emitted from one or more specified material layers of the sample

is detected. Additionally or alternatively, the wavelength of incident light directed into the sample may be selected to penetrate the sample to a given depth, or to excite only one or more selected material layers in the sample. Accordingly, defect data characteristic of primarily only the one or more specified material layers is generated.

[0006] Other features and advantages of the invention will appear hereinafter. The features of the invention described above can be used separately or together, or in various combinations of one or more of them. The invention resides as well in sub-combinations of the features described. Furthermore, many of the method steps described herein may be performed in a different order than that which is explicitly described.

BRIEF DESCRIPTION OF THE DRAWINGS

[0007] In the drawings, wherein the same reference number indicates the same element in each of the figures:

[0008] Fig. 1 is a schematic diagram illustrating the basic components of a photoluminescence imaging system, according to a first embodiment. [0009] Fig. 2 is a graphical comparison of a high-pass filter, a low-pass filter, and a band-pass filter. [0010] Fig. 3 is a side-view of a multi-layer wafer.

DETAILED DESCRIPTION OF THE DRAWINGS

[0011] Various embodiments of the invention will now be described. The following description provides specific details for a thorough understanding of

these embodiments. One skilled in the art will understand, however, that the invention may be practiced without many of these details. Additionally, some well-known structures or functions may not be shown or described in detail so as to avoid unnecessarily obscuring the relevant description of the various embodiments.

[0012] The terminology used in the description presented below is intended to be interpreted in its broadest reasonable manner, even though it is being used in conjunction with a detailed description of certain specific embodiments of the invention. Any terminology intended to be interpreted in any restricted manner will be overtly and specifically defined as such in this detailed description section.

[0013] Fig. 1 schematically illustrates the basic components of a photoluminescence spectroscopy or imaging system 10, according to one embodiment. The photoluminescence imaging system 10 includes a light source 12, such as a laser diode or other suitable light-emitting diode (LED) or device. The light source 12 generates an illumination or incident laser beam 14 or other light beam (e.g., a UV light beam) toward, either directly or via reflection off of one or more optical components, a wafer 16 or other sample. An illumination optic 13, such as a focusing lens or other optical component(s), may optionally be used to focus and/or direct the incident laser beam 114. The term "wafer," as used herein, includes any sample upon which photoluminescence imaging may be performed. The term "laser beam," as used herein, includes any light beam suitable for exciting one or more material layers in the wafer.

[0014] The wafer 16 may be a silicon semiconductor wafer or other microelectronic, micro electro-mechanical, or optoelectronic substrate or other sample. The wafer 16 may include one or more epitaxial layers or other surface layers 15 grown on, applied on, or otherwise located on a substrate region or other bulk region 17 of the wafer 16. For example, the wafer 16 may include a bulk region 17 of silicon (Si) and a surface layer 15 of silicon germanium (SiGe). The wafer 16 may include any number of layers including any number of materials and/or compositions. The various wafer layers having different materials and/or compositions will be referred to herein as "material layers." For example, in the wafer just described, the SiGe region is a first material layer, and the Si region is a second material layer.

[0015] The incident laser beam 14 penetrates the wafer 16 to a given depth, which is substantially a function of the wavelength of the incident laser beam 14. At least a portion of the incident laser beam 14 is absorbed by the wafer 16 and is emitted from the wafer 16 as return light 18. The emitted return light 18 includes data identifying defects located in the wafer 16 down to the penetration depth (and slightly beyond, due to diffusion) of the laser beam 14. The term "defects," as used herein, includes any contaminants, flaws, discontinuities, impurities, and/or other imperfections present in a wafer 16. [0016] The varying material layers of the wafer 16 each emit return light at a specific wavelength corresponding to the composition of materials in the layer. For example, a typical silicon layer in a wafer 16, at room temperature, may emit return light at a wavelength of approximately 1.07 microns. A typical layer of SiGe, having a 25% Ge concentration, for example, at room temperature, may

emit return light at a wavelength of approximately 1.17 microns. The wavelengths at which specific materials and compositions emit return light are well known, and/or can be readily calculated.

[0017] The return light 18 emitted from the wafer 16 is collected by a collector 19, such as an objective lens or other suitable optical component(s). Different wavelengths of collected return light correspond to the varying material layers in the wafer 16. The collected return light enters a filter 20 connected to and/or optically aligned with the collector 19. Depending on which wavelength(s) of return light are intended to be detected and measured in a given application, the filter 20 may be a low-pass, high-pass, band-pass, or other suitable filter. [0018] Using the values from the above example, if only wavelengths of return light emitted by the layer of SiGe are to be measured, then a high-pass filter that allows only wavelengths of light above a specified threshold to pass through the filter may be used. For example, a high-pass filter that allows passage of only wavelengths of light above 1.12 microns could be used to allow return light from the SiGe layer (approximately 1.17 microns) to pass, and to prevent return light from the silicon bulk region (approximately 1.07 microns) from passing.

[0019] The filter 20 may optionally be an adjustable or variable filter having multiple filtering elements for allowing a variety of wavelengths of light to pass through the filter 20. For example, the filter 20 may be a rotatable wheel or slidable bar element having multiple filtering elements therein, which allows a user to select which wavelength(s) of light will pass through the filter 20. The

multiple filtering elements may include one or more low-pass, high-pass, and/or band-pass filters, or may include any other suitable filtering elements. [0020] The return light that passes through the filter 20, referred to herein as filtered return light 22, is detected by a detector 24, such as a CCD array or other suitable detection system connected to and/or optically aligned with the detector 24. A microprocessor, computer, or other processing system is preferably included in, or is connected to or otherwise associated with, the detector 24 for processing data from the filtered return light 22 detected by the detector 24. Thus, only defect data or intensity values characteristic of the desired material layer (the SiGe layer in the above example) is detected and processed. Accordingly, the defect data obtained via this method of selective detection is not obscured or distorted by defect data from other material layer(s) (the Si layer in the above example) in the wafer 16.

[0021] The generated defect data may optionally be converted into a visual image indicating the locations and quantity of defects in the specified material layer of the wafer 16. The visual image may be displayed on a monitor or other suitable display screen and/or may be printed out in hard copy form via a printer or similar device. The selective detection process may optionally be performed on one or more additional material layers in the wafer 16 by adjusting the variable filter, or inserting a new filter, to allow passage of return light emitted from the selected material layer(s) in the wafer 16.

[0022] In another embodiment, if defect data corresponding to only the bulk region 17, or to another material layer of the wafer 16, is desired, the selective detection method may be used in combination with a bulk probe and a

mathematical subtraction process to generate the desired defect data. Using the sample values from above, for example, after the selective detection method is performed to generate defect data corresponding to the SiGe layer of the wafer, a second laser probe penetrating through all or most of the bulk region 17 may be performed.

[0023] The bulk probe may be performed using no filter, or using a filter that allows wavelengths of light emitted from both the surface layer 15 (SiGe) and the bulk region 17 (Si) to pass. A high-pass filter set to allow all wavelengths of light greater than 1.00 micron to pass, for example, may be used to allow passage of return light emitted from the SiGe and Si layers, while preventing passage of light having a wavelength of less than one micron, such as light from the incident laser beam, for example.

[0024] The defect data obtained from the selective detection probe (i.e., the surface layer defect data) is then mathematically subtracted from the defect data obtained from the bulk probe (which includes surface layer and bulk region defect data) to yield defect data characteristic of defects located primarily in the bulk region 17 of the wafer 16. Put another way, intensity values or defect data points common to both the selective detection probe and the bulk probe (i.e., intensity values or defect data points corresponding to defects primarily in the surface layer 15 of the wafer 16) are mathematically removed from the bulk probe data to produce new defect data characteristic of defects located primarily in the bulk region 17 of the wafer 16.

[0025] This subtraction or removal process may be performed via a software program, or any other suitable program or processor, that compares the

data points from the surface probe and the bulk probe and subtracts off or otherwise removes the data points common to both probes. The term "primarily," as used herein, means predominantly or entirely, since the generated defect data may not always correspond exactly to defects in the specified material layer(s), due to calibration issues, measurement errors, layer non-uniformity, and so forth. [0026] By removing the defect data corresponding to the surface layer 15, which is often uneven or otherwise flawed, the new defect data, which corresponds primarily to defects in the bulk region 17 of the wafer 16, is no longer distorted by the previously overlying surface layer defect data. Accordingly, a substantially clearer image of the location and concentration of defects present in the bulk region 17 of the wafer 16 may be produced using the new defect data. A visual image of these bulk region defects may be generated using an image-producing device associated with, or located in, the detector 20 and/or with other system elements. The visual image may be displayed on a monitor or other suitable display screen and/or may be printed out in hard copy form via a printer or similar device.

[0027] In another embodiment, the surface layer defect data (obtained using the selective detection process) and the bulk probe defect data are normalized relative to each other, before the subtraction process is performed, to enhance the defect images and to make them more easily mathematically processable. The normalization process may include adjusting the contrast and/or the brightness of one or both of the images produced by the selective detection and bulk probes so that they have the same, or substantially the same, relative intensity. Such a normalization process may be performed by a software

program or other suitable program or processor, which may be the same as or different than the program or processor that performs the subtraction process. [0028] The normalization process may alternatively or additionally include plotting data points from the selective detection probe versus data points from the bulk probe and removing any mathematically outlying data points, thus yielding images that are more readily mathematically processable. For example, a software program or other suitable program or processor, which may be the same as or different than the program or processor that performs the subtraction process, may curve-fit lines to their respective data points and then remove any outlying data points located beyond a predetermined tolerance limit or distance from the lines.

[0029] By normalizing the surface layer defect data relative to the bulk defect data before performing the subtraction process, a clearer, more precise new data set, corresponding to defect data located primarily in the bulk region 17 of the wafer 16, may be obtained. Any other suitable method(s) for normalizing the bulk probe and surface probe data and/or corresponding images relative to each other may additionally or alternatively be used.

[0030] A similar selective detection and mathematical subtraction process may be performed to generate optimized defect data characteristic of any material layer in the wafer 16. For example, a low-pass filter may be used to generate first defect data corresponding primarily to the bulk region 17 of the wafer, after which a bulk probe may be performed to generate second defect data corresponding to both the surface layer 15 and the bulk region 17 of the wafer 16. These two sets of data may optionally be normalized relative to each

other, and the first defect data may be subtracted from the second defect data to yield new defect data characteristic of defects located primarily in the surface layer 15 of the wafer 16. A similar selective detection and subtracting process may be performed on any material layer of a multi-layer wafer 16, as well. [0031] In another embodiment, the photoluminescence imaging system 10 may be used to perform a selective excitation process, instead of or in conjunction with performing a selective detection process. In this embodiment, the wavelength of the incident laser beam 14 is selected such that it penetrates the wafer 16 approximately to a desired depth and/or to excite one or more selected material layers in the wafer 16. For example, if only defect data from a surface layer 15 having a thickness of 50 nm is desired, then the wavelength of the incident laser beam 14 may be selected so that it penetrates the wafer to a depth of approximately 50 nm. A variable wavelength laser source 12, or a device including multiple laser sources 12, may be used to produce the desired wavelength(s).

[0032] Return light 18 emitted from the wafer 16 will therefore include defect data corresponding only to the region of the wafer 16 having a depth approximately equal to the penetration depth (actually slightly beyond, due to diffusion) of the incident laser beam 14. Thus, in the 50 nm surface layer example, the wavelength of the incident laser beam 14 is selected to penetrate the wafer to approximately 50 nm, such that the emitted return light 18 includes defect data primarily characteristic of the surface layer 15 of the wafer 16. The emitted return light 18 is collected by the collector 19, after which it may optionally be passed through a filter 20 to further limit which wavelength(s) of

light are allowed to enter the detector 24, as described above. The microprocessor, computer, or other processing system may then process data in the detected return light.

[0033] In another selective excitation embodiment, the wavelength of the incident laser beam 14 may be selected such that it is shorter, or has higher energy, than the energy band gap of the material(s) intended to be excited, but is longer, or has lower energy, than the energy band gap of the material(s) not intended to be excited. The shorter the wavelength of the excitation laser, the higher the energy of the laser. The laser's energy must be high enough to excite at least one of the material layers. Using the sample values above, the wavelength of an incident laser beam may be selected such that it has a high enough energy to excite the SiGe layer (which emits return light at a wavelength of 1.17 microns), but does not have a high enough energy to excite the Si layer (which emits return light at a wavelength of 1.07 microns). Thus, return light emitted from the wafer 16 in this example would include defect data characteristic primarily of only the SiGe layer.

[0034] Fig. 2 is a graphical comparison of the operation of a high-pass, a low-pass, and a band-pass filter. A high-pass filter allows wavelengths of light

greater than λi to pass through the filter, while preventing wavelengths of light

less than λi from passing through the filter. A low-pass filter allows wavelengths

of light less than λ ₂ to pass through the filter, while preventing wavelengths of

light greater than λ ₂ from passing through the filter. A band-pass filter allows

wavelengths of light greater than λ ₃ and less than λ ₄ to pass through the filter,

while preventing wavelengths of light less than % _$ , or greater than λ ₄ , from

passing through the filter. Alternatively, a band-pass filter may prevent

wavelengths of light greater than Xz and less than X4 from passing through the

filter, while allowing wavelengths of light less than Xz, or greater than λ ₄ , to pass

through the filter.

[0035] A band-pass filter may be particularly useful when performing photoluminescence imaging on a wafer having more than two layers, such as the three-layer wafer 16' illustrated in Fig. 3. If defect data is desired from only material layer B, for example, and if material layer B emits light at a wavelength

λ _B that falls between the wavelengths emitted by material layers A (λ _A ) and C

(λc), then a band-pass filter that allows the passage of light emitted at

wavelengths in a first range encompassing λ _B , while excluding light emitted at

wavelengths above and below the first range, including light emitted at XA and λ _B ,

would yield the desired results.

[0036] The photoluminescence imaging systems described herein may include any number of additional or alternative components for performing photoluminescent imaging or spectroscopy on a wafer 16. For example, a photoluminescence imaging system may include a beam splitter for reflecting the incident laser beam 14 toward the wafer 16, and for transmitting the return light 18 toward the collector 19, the filter 20, and/or the detector 20. One or more additional beam splitters or other optical elements or mirrors may be included in the photoluminescence imaging system to achieve a desired size or configuration of the system, and/or to select, focus, collect, and/or direct the incident laser beam 14 and/or the emitted return light 18 throughout the system.

[0037] The photoluminescence imaging system may further include an imaging microscope having multiple laser sources, a display screen for visually displaying photoluminescent image data, a printer for printing our hard copies of photoluminescent image data, a wafer holder for supporting the wafer 16, additional filters for filtering incident or return light, and/or any other suitable components. Indeed, a first filter may be used to filter an incident laser beam generated by the light source, while a second filter may be used to filter return light emitted from the wafer 16 (i.e., a combination of selective excitation and selective detection may be used). A detailed example of a photoluminescent imaging system is described in International Application Number PCT/GB97/02388.

[0038] The photoluminescence imaging processes described above may be performed at approximately room temperature, also as described in International Application Number PCT/GB97/02388. Alternatively, any of the photoluminescence imaging processes may be performed at higher or lower temperatures, including significantly higher or lower temperatures, depending upon the requirements of a given application.

[0039] Any suitable devices and components may be used to perform the photoluminescence processes described herein. For example, various means may be used to produce lasers of varying excitation wavelengths, such as one or more laser diodes or LEDs or other suitable light-producing devices. Any suitable means for collecting emitted return light, such as an objective lens or other optical collector, may be used in the photoluminescence imaging system. Any suitable filtering means, such as a high-pass, low-pass, or band-pass filter

may be used to selectively filter emitted return light. Any suitable means for detecting return light may also be used, such as a CCD array, a spectrometer, or any other suitable detecting system or device.

[0040] Means for generating defect data based on the detected return light may include a software program or other program or processor, which may be included in or associated with the detector and/or with other elements in the photoluminescence imaging system. Such a program or processor may also be a means for performing the optional subtracting and/or normalizing processes described above. Alternatively, one or more separate programs or processors may be used to perform one or more of these processes.

[0041] Thus, while several embodiments have been shown and described, various changes and substitutions may of course be made, without departing from the spirit and scope of the invention. Many of the method steps described herein, for example, may be performed in a different order than that which is specifically described. The invention, therefore, should not be limited, except by the following claims and their equivalents.