TWI592645B - Non-contacting system and method for measuring the dopant content of semiconductor material and method of determining the impact of a semiconductor material fabrication line upon semiconductor wafer - Google Patents

Description

Non-contact systems and methods for measuring dopant content of semiconductor materials and methods for determining the effect of semiconductor material lines on semiconductor wafers

The present invention relates generally to the measurement of dopant content in one or more layers of a semiconductor component, and more particularly to the non-dopant content of such components in a related industrial line. Systems and methods of contact measurement, such as for photovoltaic (PV) solar cells, LEDs, and any other semiconductor component that employs a doped layer of diffusion, implantation, or epitaxial deposition.

We will describe the crystalline germanium (c-Si) PV cell process and the semiconductor LED process as a background.

In order to manufacture c-Si PV cells, a wafer is subjected to a series of process steps in a battery production line. Each incoming wafer is slightly doped with atoms that will produce a "free carrier" (semiconductor slang) with a positive potential ( n -type wafer) or a negative potential ( p -type wafer) (ie, Overall diffusion). The first step (after the incoming inspection to discard the defective wafer or sort the wafer into batches) is to pass the wafer through a wet chemical etching process to remove saw marks and other surface defects and contamination. . Each wafer is then isotropically textured (another wet process) to microscopically roughen its surface, enhancing its ability to capture incident photons. After forming the texture, the wafer is then doped with a chemical that produces a "free carrier" (semiconductor lingo) having a potential opposite to the body doping in a layer on the surface of the wafer. In today's practice, such doping can be produced in one of the following ways: an "in-line" method or a "batch" method. The in-line method deposits dopant chemicals, typically carried in liquid form, on the top surface of the wafer. (In the case of a phosphorus dopant, this carrier is most often phosphoric acid). The deposited dopant carrier is then dried and the resulting product is then diffused (using a high temperature furnace) into each wafer to form a semiconductor junction that is exposed when exposed to sunlight. The surface will allow the wafer to generate electricity. In this in-line method, the wafer is continuously conveyed through the apparatus performing the steps, which is typically performed by a machine that is first a "doping machine" to apply the liquid carrier, followed by a "dryer" machine. The carrier is formed by leaving the dopant chemical on the surface and a third machine, an in-line diffusion furnace that diffuses the dopant into the wafer. In the batch method, the wafer is loaded into a cassette (most commonly made of quartz and the semiconductor colloquial is called a "boat"), which is inserted into a "tubular" diffusion furnace. The wafer is then sealed and the wafers are simultaneously exposed to a gas type dopant carrier (most commonly phosphonium chloride) and heated to diffuse the dopant into the wafer. The wafers are then removed from the furnace, removed from the boat and moved to the next portion of the line. In both methods, the amount of dopant introduced, the time spent in the diffusion process, and the temperature of the diffusion process determine the depth of penetration of the second dopant and the concentration depending on the depth. Furthermore, the second dopant is introduced and diffused into all surfaces of the wafer depending on the nature of the diffusion process. It is noted that, unless otherwise specifically mentioned, "dopant" as used herein refers to such a second dopant that is introduced onto the surface of a body doped wafer. Each wafer is then wet etched again to remove the phosphorous glass (also known as PSG, a by-product of the dopant diffusion step) and can be etched to be patterned or removed on the "back" side All or part of the dopant to avoid shunting. After this step, a coating (most typically tantalum nitride) is applied to the top surface of the wafer to reduce reflection and passivate the surface. This coating is typically applied using a plasma enhanced chemical vapor deposition apparatus. Thereafter, the wafer has metal contacts printed on its top and bottom surfaces, wherein the top contact pattern is designed to minimally interfere with light exposed to the Si material while providing a minimum resistance The path gives the current flowing out of the wafer. These metal contacts, which are printed in the form of a metal paste, are dried using a furnace and then diffused into the wafer. Thereafter, if a portion of the dopant on the back side of the wafer has not been completely or partially removed, a laser or mechanical device is used to cut a trench around the periphery of the wafer to Avoid diversion. Finally, the wafer (which is now a completed PV cell) is tested and graded.

The dopant concentration itself as a function of its distribution within the volume of the wafer plays a central role in determining the quantum efficiency and other electrical characteristics of the resulting PV cell that ultimately produces its power output capacity and market value. character of. Therefore, the steps associated with the amount and distribution of dopants that are diffused into the wafer during the PV cell process are of the highest importance. Specifically, these steps are: (a) the initial "base" doping of a blank wafer as provided by the wafer manufacturer (in most cases, the blank wafer is boron). And (b) subsequent doping in the outer region of the wafer (in most cases, this utilizes the negative doping of phosphorus). This second doping step forms what is known as an "emitter". The term "base" will be used to refer to the doping of a blank wafer, and the term "emitter" refers to the semiconductor produced by the second doping step.

To ensure that the process of emitter formation is within the required specifications, some measurements are made to provide a representation of the base dopant concentration and emitter dopant concentration of the blank wafer. In current practice, photovoltaic (PV) wafers are often inspected in a PV cell process at varying intervals, either manually or by a single point vision measuring device using an industrial camera of visible spectrum. In addition to the raw material acceptance phase (at the beginning of the production line) and the final inspection and grading (at the end of the production line), continuous on-line measurements of the wafer are often limited by the range and coverage area, thus using offline Non-continuous sampling, especially for industrial cameras using visible spectrum, cannot be checked for the nature of interrogation. When off-line sampling is used, there may be hundreds of wafers passing through one or more steps of interest in the process during the time interval between samples. This situation is a common process step in determining the application, concentration, and distribution of dopants in a PV wafer, and these steps are therefore not currently well controlled, limiting the acceptable yield of the PV cell manufacturing facility. . In order to increase yield, the industry is now looking for continuous on-line measurements, ideally on 100% wafers for better control of the concentration and distribution of dopants in PV wafers.

In addition to the established commercial PV cell structures and processes described above, certain novel PV cell structures and associated processes are now being introduced into commercial production. These include selective emitter cells, emitter-wrap-through cells, and interdigitated back contact solar cells (IBC cells). The selective emitter cell system changes the emitter dopant concentration in the immediate vicinity of the front metal contact to achieve optimum conduction efficiency (meaning heavier doping in these regions) while limiting to the contacts The undesired carriers are recombined (meaning lighter doping in these regions). The emitter bypass and the IBC battery eliminate the shadow loss by placing the emitter and base contacts on the back of the cell. The invention described herein can be utilized to measure the PV cell geometry as well as the dopant levels of the more common front and back contact geometries described above.

A semiconductor light emitting diode (hereinafter abbreviated as "LED") performs a function opposite to that of a PV cell. Instead of absorbing photons to generate electricity, LEDs use electricity to emit photons (a phenomenon called electroluminescence). In LED manufacturing, the wafer is composed of a neutral substrate such as sapphire. Compared to PV cell fabrication, the wafers are polished rather than textured, each wafer contains a plurality of LEDs, and the dopants used to create the semiconductor are deposited on the surface of the wafer. The layer is not diffused by diffusion processes used in PV cell fabrication. Despite these differences in structure and fabrication, these dopant layers can be examined by the same methods disclosed in the present invention. From hereafter, for the sake of simplicity and clarity, the PV cell structure will be described without limiting the application of the invention to other doped semiconductor structures.

In the manufacture of PV cells, some existing and novel techniques have been proposed for in-line measurements of emitter doping, but all have serious limitations. Diffusions for measuring diffusion, which are surface length photovoltage (SPV) measurements of diffusion length, eddy current measurements of surface resistance, and an infrared method developed for surface resistance measurement developed by the Fraunhofer Solar Research Institute in Germany. (J. Isenberg, D. Biro, and W. Warta, "Fast, Contactless, and Spatially Analytical Measurements of Surface Resistance by an Infrared Method," "Photovoltaic Progress: Research and Applications", 2004; 12:539 -552 pages). To the best of our knowledge, there is no method for measuring a wet dopant carrier film.

SPV measurements have been used in the laboratory to measure the diffusion length (how long the average travel distance is before an excess carrier is recombined in a bulk semiconductor to achieve a balanced carrier concentration). See, for example, D. K. Schroder, "Surface Voltage and Surface Photovoltage: History, Theory and Applications", "Measuring Science and Technology", 2001 12 R16-R31. SPV measurements are typically made by placing a wafer on a ground electrode (although a contactless method without a back sensor plate is also possible) and positioning a capacitive probe at a small distance above the sample. Come and do it. Because the measurement is capacitive, the measurement area is very limited, the maximum working distance is extremely small, and there is little tolerance for wafer bending or vertical movement. Moreover, in the manufacturing operation of the conveyor belt feeding, due to the limited working distance, there is a relatively high chance of "collision" if any wafers are stuck together (a non-rare situation), if the wafer breaks And the fragments on the conveyor belt are not flat (also non-rare), or if any foreign objects are accidentally brought to the conveyor belt, or if the conveyor belt itself encounters a small vertical oscillation that exceeds the working distance of the sensor This will cause blockages on the conveyor belt. Finally, due to the need for dedicated wafer transport for SPV measurements and the need for very close working distances, the introduction of such technology into existing production lines may require significant line modifications that may make their use expensive. And it is not practical.

Eddy current measurements have many of the same limitations as SPVs and have previously been shown to be unsuitable for in-line measurements of emitter doping (using surface resistance measurements as a measure). (Rueland, E.; Fath, P.; Pavelka, T. Pap, A.; Peter, K.; Mizsei, J. "Comparative study of emitter surface resistivity measurements for on-line quality control", “Photovoltaic Energy Conversion”, Third World Conference Record 2003, Issue 2, May 12-16, 2003, pages 1085-1087: Volume 2)

Although the Fraunhofer method is suitable for use in laboratories, there are many requirements that make it unsuitable for practical online use, most notably the strict requirement for spurious heat or light, so it is provided in an online manufacturing environment. This condition is extremely difficult and expensive.

In summary, while it is important to develop a commercially viable technique to allow for on-line measurement of electrical properties in PV wafers determined by dopant content, there is currently no known technique that is fully configurable for use in production lines. It is robust enough to operate reliably and cost-effectively at all points in the industry.

There is therefore a need for a method and apparatus that is useful for blank wafer dopant concentrations, the amount and distribution of wet dopant films produced by in-line doping devices, and any steps after diffusion in a production line. On-line measurements of dopant concentration in the emitter are resilient, configurable, robust, and cost effective.

There is a further need to define a particular repeatable sampling position for each wafer to enable enantioselective emitters, pass-through contacts, and IBC cell doped structures as well as conventional uniform doping. Therefore, there is also a need for an apparatus and method having the ability to vary the scanning "strength" (the number of samples taken per unit length in a direction transverse to the machine during a certain period of time) to allow the operator to execute the cycle. Sexual measurements, or if necessary, perform in-depth measurements that are not scheduled in advance.

One aspect of the invention is a non-contact system for measuring the dopant content of a semiconductor material, comprising:

a source of infrared radiation configured to focus infrared radiation on a point on the material;

b. a modulator for modulating radiation from the source prior to the radiation striking the material;

c. a first lens arranged to collect radiation reflected from the material and to focus the radiation;

d. a first band pass filter arranged to receive the radiation from the first lens, the first filter being configured such that a narrow wavelength band of the radiation passes through the filter and reflects the radiation The remaining part;

e. a second band pass filter configured to receive radiation reflected from the first filter, the second filter being configured to pass a narrow wavelength band of the radiation through the second filter And configured such that the pass wavelength band is different from the narrow wavelength band pass through which the first filter passes;

f. a first radiation detector configured to receive radiation passing through the first band pass filter and configured to determine a first level of energy;

g. a second radiation detector configured to receive radiation passing through the second filter and configured to determine a second level of energy;

h. A calculator configured to compare the first and second levels and utilize a correlation curve to return a dopant content value based on a series of known different levels for a series of A comparison of the energy levels of the semiconductor materials of the same dopant content to the two sensors correlates to the dopant content of the semiconductor material.

Another aspect of the invention is a non-contact system for measuring a dopant content of a semiconductor material, the method comprising:

a source of broadband infrared radiation that is configured to focus infrared radiation at a point on the crucible;

b. a modulator for modulating the infrared radiation from the source before the radiation strikes the crucible;

c. a first lens arranged to collect infrared radiation reflected from the crucible and to focus the radiation;

d. a beam splitter disposed adjacent to a focus of the first lens, the beam splitter configured to separate the radiation into two radiation streams;

e. a first narrow band pass filter configured to receive the first radiation stream of the two streams at a first predetermined wavelength of radiation;

f. a second narrow band pass filter configured to receive the second radiation stream of the two streams at a second predetermined wavelength, the second predetermined wavelength being different from the First preset wavelength;

g. a first infrared detector configured to receive and configured to determine an energy level of radiation passing through the first band pass filter;

h. a second infrared detector configured to receive and be configured to determine an energy level of radiation passing through the second band pass filter;

i. A calculator configured to compare the first and second levels and utilize a correlation curve to return a dopant content value based on a series of known different levels A comparison of the energy levels of the semiconductor materials of the same dopant content to the two sensors correlates to the dopant content of the semiconductor material.

Another aspect of the invention is a method for non-contact measurement of dopant content of a semiconductor material, comprising the steps of:

a. directing a modulated infrared radiation source to a measurement point on the material;

b. directing radiation reflected from the material onto a first band pass filter, the first band pass filter being configured such that a range of wavelengths of the radiation passes through the filter and reflects the remainder of the radiation;

c. directing radiation reflected from the first filter onto a second band pass filter, the second band pass filter being configured such that a range of wavelengths of the radiation passes through the second band pass filter And configured such that the wavelength range is different from the wavelength range through which the first band pass filter passes;

d. determining the energy level of the radiation passing through the first band pass filter;

e. determining the energy level of the radiation passing through the second band pass filter;

f. comparing the energy levels determined in steps d and e;

g. According to the comparison, a correlation curve is used to calculate the dopant content of the material, the correlation curve being based on the arrival of the same semiconductor material for a series of dopant levels having known different levels. A comparison of the energy levels of the detector correlates to the dopant content of the semiconductor material.

Another aspect of the invention is a method for determining the effect of one or more process steps on a semiconductor wafer in a semiconductor material line, comprising the steps of:

a. setting the first system in the first aspect of the line at one of the selected ones of the lines to select the upstream point of the wafer to determine the dopant content of the wafer at the upstream point in the line Level

b. setting a second system of the above-described aspect at one or more processing steps of the line at a selected downstream point of the location to determine a dopant content of the wafer at a downstream point in the line Level

c. operating the production line to move a series of wafers from the upstream point to the downstream point through the one or more process steps;

d. utilizing the first system to determine a dopant level level of the wafer at a point upstream of the line;

e. utilizing the second system to determine a dopant level level of the wafer at a downstream point of the line;

f. comparing the dopant content level of the wafer at the downstream point with the dopant content level at the upstream point to obtain a wafer at the upstream point and between the wafers at the downstream point A difference in dopant level.

A wet film system on a wafer (or any substrate) has electromagnetic absorption and reflection characteristics. In particular, water molecules in aqueous membranes have unique absorption peaks at the infrared ("IR") wavelength. This is depicted in the water absorption spectrum of Figure 1. The doped germanium (or even any semiconductor) wafer also has absorption, reflection amplitude, and reflected phase/polarization that are unique to one of the spatial concentrations of free charge carriers due to doping. In particular, as shown in Figures 2 and 3, the n-doped germanium exhibits significantly different absorption (or hence reflection coefficient) of the infrared spectrum of the free carrier at different doping levels.

2 is a graph of absorption versus wavelength for free carriers that diffuse different negative dopants of a negatively doped germanium substrate ( n- Si) at 300[deg.] K at different concentrations. Referring to the numbering on the graph of Figure 2, the dopant concentration (in cubics per cubic centimeter) is: 1) 1.4x10 ¹⁶ cm ^-3 (arsenic dopant); 2) 8x10 ¹⁶ cm ^-3 (锑); 3) 1.7x10 ¹⁷ cm ^-3 (锑); 4) 3.2x10 ¹⁷ cm ^-3 (phosphorus); 5) 6.1x10 ¹⁸ cm ^-3 (arsenic-tin alloy); and 6) 1x10 ¹⁹ cm ^-3 (arsenic) ).

Figure 3 is a graph of the reflectance of two polycrystalline (poly c-Si) wafers based on the difference in infrared wavelength, where one wafer (W1) is only doped with boron and the other wafer (W16) A layer of phosphorus is also diffused into its top surface. Numbers 1-5 after W1 and W16 identify the segments that are inspected on each wafer. The measurements on this graph are normalized relative to a pure crystalline 矽 reference sample. The figure shows that as the incident infrared wavelength is lengthened, the corresponding reflection coefficient of the wafer with the phosphorus doped layer is significantly stronger than the bulk doped wafer compared to the reference sample, thus indicating the increase The dopant layer affects the reflection coefficient as a function of the infrared wavelength, and thus the slope of the normalization of the infrared reflection coefficient relative to the wavelength can be utilized to determine the degree of doping of this layer.

In addition, any chemical layer or film (not just phosphorus, whether diffused or not) present on a different substrate will cause refraction, reflection, wavelength shift and phase change, which can be utilized to determine the layer/film Thickness and/or condition at the boundary. The magnitude, phase, polarization and wavelength of such absorption and reflection are dependent upon the particular film or dopant used, the density and thickness of the film or doping, and the nature of any underlying substrate.

By transmitting infrared radiation at a known wavelength and intensity level onto a wafer or substrate, the absorption of the characteristic wavelength can be measured as a function of the reflected value observed at the receiver. Phase shifts, wavelength changes, and polarization changes can also be measured. Since the amount of absorbed energy varies with the amount and composition of the wet film or the concentration of the emitter doping (depending on the specific conditions), the wet film concentration, depth and distribution, or emitter density can be It is measured by measuring the difference between the energy of the illumination and the energy of the reflection.

It is desirable to take measurements or samples from a plurality of specific locations on a wafer or substrate. This is because a single sample may exhibit broad variations and may have to smooth out these variations, and also because the wafer or substrate may have a deliberately differentiated deposition of wet film or diffusion of dopants.

Furthermore, for each sample, the apparatus and method described herein can tolerate light, heat and vibration from the factory environment and compensate for temperature, varying working distances by utilizing simultaneous differential queries at the sampling location. And varying angles of incidence.

The apparatus 10 described in Figure 4 is for a multi-belt conveyor-fed photovoltaic ("PV") battery manufacturing facility. Although it should be appreciated that the configuration of single and/or non-conveyor belts is also feasible for LEDs and other semiconductor manufacturing facilities.

One or more transmitters and receivers are mounted above the area where the PV wafer 12 will be measured. Each receiver is composed of two or more sensors whose purpose is to capture the differential signal data as explained above. For simplicity and clarity, a non-contact system for measuring the dopant content of doped germanium will be described with respect to a device 14 consisting of a single transmitter 16 and a single receiver 18, the receiving The device 18 is composed of two sensors 20, 22. This is depicted with reference to FIG. 5, which is a block diagram that schematically depicts an alternate embodiment of the present disclosure.

A sensor housing containing the device in the block diagram is located approximately 50-150 mm above the surface of the wafer 12.

The transmitter has at least three possible embodiments, each of which contains a different source of infrared radiation. In a first embodiment, the source is comprised of one or more continuous broadband infrared sources mounted in an elliptical reflector. In the second embodiment, the source is composed of a multi-wavelength infrared laser. In a third embodiment, the source is comprised of two single wavelength infrared lasers.

Referring to the first embodiment of the transmitter, the elliptical reflector 24 of the infrared source 16 focuses the infrared radiation from the broad spectrum of the infrared source to a single point in space. A chopper wheel 26 is positioned at the focus of the ellipse, which modulates infrared radiation at approximately 1 kHz, although the radiation may be by any suitable method including amplitude, frequency, pulse, or phase shift modulation or A combination of methods to modulate. The use of modulation is necessary when the detector is responsive to changes in the detected signal, and because the modulation distinguishes the transmitted infrared signal from the background infrared radiation and enhances the signal to noise ratio. The modulation can also be utilized to effect an effect on the modulation by measuring the change in dopant content through the reflected signal to produce information about the dopant content.

An off-axis elliptical reflector 28 is shown facing the infrared source 16 to receive the modulated radiation. The elliptical reflector 28 is incident at about 45 degrees to the wafer 12 and aligns the peak of the radiation at the center of the first lens 32 of the receiver 18 (discussed below), which will be tuned from the calender wheel 26. The altered radiation is focused onto a measurement point 30 of the wafer 12. Although it will be appreciated that the reflector 28 is not necessary in the second and third embodiments of the transmitter 16, since the lasers are already in a collinear form.

There are at least two possible embodiments for each receiver 18. In a first embodiment of the receiver 18, the receiver 18 is mounted above the measurement point 30 where the infrared radiation strikes the wafer 12. The reflected infrared radiation is diffused and collected by the first lens 32 and directed to the first narrow band pass filter 34. The first filter is radiated by a narrow band of infrared rays centered at one of the wavelengths selected in the infrared spectrum. This wavelength is chosen such that the effect of the isotropic texture on the nature of the received signal of interest is not important. The other portion of the received radiation is reflected by the first filter 34.

The reflected radiation is directed to a second narrow bandpass filter 36 whose center is set at a different selected wavelength such that the wavelengths of the two bands do not overlap. Similarly, this second wavelength is chosen such that any effect of the isotropic texture on the nature of the received signal of interest is not important. In a preferred embodiment, a bandpass filter 34 or 36 has a central passband having a passband of +/- 125 nm at approximately 8 microns and another filter 34 or 36 having approximately A central passband with a passband of +/- 175 nm at 10.5 microns.

In another preferred embodiment, the pass band of each band pass filter is between 50 nanometers and 500 nanometers. And in another preferred embodiment, the center of the pass band of one of the filters is between 1 and 20 microns. In another preferred embodiment, the passband center wavelength of the second filter is between 1 and 20 microns and is different from the passband center wavelength of the first filter.

In another preferred embodiment, the difference between the center wavelengths of the first and second filters is between 1 and 10 microns. And in another preferred embodiment, the difference between the center wavelengths of the first and second filters is 2 microns.

In another preferred embodiment, the center wavelength of the first filter is set at 8.06 microns, and the center wavelength of the second filter is set to 10.5 microns, wherein each filter has a range of 200 to 400. The width of the pass between the nanometers.

The radiation passing through the first filter 34 is focused by a second lens 38 onto a first infrared detector or sensor 20, the first detector 20 generating a proportional to the arrival of the infrared radiation. A low voltage signal of the intensity of the first detector 20. The radiation passing through the second filter 36 is focused by a third lens 40 onto a second infrared detector or sensor 22, which generates a proportional to the arrival of the infrared radiation. A low voltage signal of the intensity of the second detector 22.

The low voltage signals of each of the detectors 20, 22 are amplified by individual amplifiers 42, 44 and then acquired by an analog to digital data acquisition board 46, which is synchronized to the transmission. The chopper frequency in the device 16 is controlled by a computer 48. Thus, the sensors 20, 22 produce two voltage values that are proportional to the infrared energy in the two narrow frequency bands that pass through the first and second filters 34, 36, respectively.

The computer 48 uses the voltage from each of the detectors 20, 22 to calculate the slope and/or ratio between the amounts of energy received in each frequency band, the slope and/or ratio being proportional to the above disclosed The energy absorbed by the dopants in the top layer of wafer 12. The dopant content is calculated by passing the slopes according to a model of infrared reflection of the wafer material at different dopant levels, particularly but not limited to, through a correlation curve as illustrated in FIG. Or look up the table to decide.

In a second embodiment of the receiver, a beam splitter is used to separate the reflected IR energy at the focus of the first lens into equal portions and to direct the generated equal portions to a detector On the array, each detector array has a different bandpass filter in front of the detector. Each detector delivers a voltage proportional to the intensity of the infrared radiation reaching each detector. Therefore, multiple points on the correlation curve of doping with respect to wavelength are measured, which improves the correctness of the slope measurement (because the slope may vary with wavelength) and thus improves the semiconductor material or semiconductor Measurement of the dopant content on the material.

In another embodiment of the device 14, it is not that the bandpass filter 34 is disposed behind the lens 32, but a beam splitter is disposed behind the lens 32. This splits the beam from lens 32 into two beams that are directed to individual bandpass filters 34, 36, individual lenses 38, 40, and then individual sensors 20, 22.

The surface of the object shown as a single semiconductor wafer 12 in FIG. 5 can also be a plurality of semiconductor wafers on a conveyor belt, a stationary wafer, or a monolithic surface such as a film on a substrate. The surface can have any size.

A preferred embodiment of a non-contact system for measuring the dopant content of semiconductor material 10 is shown in Figure 4 in the form of a schematic block diagram. A plurality of sensing heads 50 are mounted between 5 mm and 250 mm above a wafer transfer belt (not shown) aligned in a direction perpendicular to the direction of travel of the conveyor belt. Each of the sensing heads 50 includes a housing in which the components of the apparatus of Figure 5 including a single transmitter 16 and a single receiver 18 (from Figure 5) are contained. The receiver 18 includes the two sensors 20, 22 (Fig. 5). Again, the components are configured to operate in the manner previously described with reference to FIG. In particular, within each of the sensing heads are an infrared source 16, a neon wheel 26, a focusing reflector 28, a lens for collecting reflected infrared radiation and directing the infrared radiation onto a bandpass filter 34 or beam splitter. 32. Two detectors 20, 22 that produce a voltage proportional to the magnitude of the infrared radiation in a given frequency range, and a frequency for amplifying with a frequency synchronized with the gap in the stop wheel 26 42 , 44 and conversion 46. This voltage becomes a device for a digital signal.

Each of the sensing heads 50 is erected over a wheel 52 in a precise track 54 that is perpendicular to the direction of travel of the conveyor belt. The track 54 is supported by a support beam 56 that is secured to the equipment frame 58 or supported from the floor. Power to each of the sensing heads is communicated by a corresponding power line 60 from the power supply and terminal cabinet 62. The power cords 60 are configured such that the sensing head 50 is free to move along the track 54 over a defined measurement range. The array of sensing heads 50 are moved together along the track 54 in the direction of arrow 66 by a linear actuator 64 that positions each of the sensing heads 50 underneath and Placed on top of the corresponding wafer 12 on the conveyor belt. The combination of movement of the conveyor belt and linear actuator 64 allows a pattern to be measured across the wafer.

When in action, the linear actuator 64 and the conveyor belt are moved in a direction at right angles to each other. This allows the pattern of measurement points 30 to be diagonal in nature, i.e. as depicted in FIG. The conveyor belt moves in the direction of arrow 68. However, if the actuator 64 is moved much faster than the conveyor belt, it is possible to measure each wafer 12 at several points across the wafer 12. This is exemplified by the pattern of measurement points 30 shown by the points in FIG. 6, some of which are labeled component symbols 30. It can be seen from Figure 6 that when the linear actuator 64 is moved in the opposite direction, a diagonal pattern of another measurement point 30 can be made. This can be repeated multiple times as the wafer 12 is moved in the direction of arrow 68 by the conveyor belt. For a fixed conveyor speed, the array of measurement points 30 and its location across the wafer 12 is a function of the sampling rate and the speed of the linear actuator 64.

At each measurement point 30, the amplified voltage from the two detectors 20, 22 of the receiver 18 of each of the sensing heads 50 utilizes multiplexed in the sensing head 50 (Fig. 5). Analog to digital converter board 48 and embedded computer 48 are converted to a digital signal. The resulting values are transmitted to a power and terminal cabinet 62 on a fieldbus or LAN cable that can be combined with power line 60. The two measurements produced at each measurement point 30 and the location of the measurement point corresponding to the position of the linear actuator 64 are transmitted to the computer 72 and stored for each measurement point 30. The presence of wafer 12 on the conveyor belt is known in terms of an increase in steps over the overall signal level of sensors 20, 22.

The sampling position and/or sampling rate on a particular wafer 12 or other substrate can be defined to follow a particular pattern. Furthermore, the mode can be predefined and more than one mode can be predefined. On a series of samples, one or more modes can be utilized, or the sampling position (measuring point 30) and the sampling rate can be arbitrarily changed. Such a variable sampling technique is depicted in Figure 6. In addition, the sampling position can be changed in the "direction of travel" by utilizing the movement of the wafer 12 or other substrate on the conveyor belt.

In order for the sampling position to be reproducible between wafer 12 and wafer 12, the locations must be offset from a particular two-dimensional location defined on the surface of the object. In the case where the surface of the object is composed of a plurality of wafers 12, the two edges of each wafer 12 are used as a reference for all sampling locations on the wafer 12. These edges are found by detecting changes in the level of radiation received by the wafer 12 relative to the signal at which only the conveyor belt is present.

The ratio or difference in voltage from each of the sensors 20, 22 of the receiver 18 in a sensing head 50 is used as a dependent variable in an associated curve, which will This ratio/difference is related to the independent variable of the dopant content of the wafer. The correlation curve is determined by having a known dopant content (measured using a laboratory-based contact four-point probe or other off-line measurement techniques such as electrochemical capacitive voltage analysis) The wafer passes under the sensing head 50 and measures the signals generated at the two sensors 20, 22 and performs a least squares regression method, which is the ratio/difference in the observed voltage. The value is related to the known dopant content from laboratory measurements. The correlation curve of the type shown in Figure 10 is thereby generated and stored in the memory of the computer for reference.

If the wafers 12 are staggered, or where different modes are desired to be measured on each wafer 12, an alternate embodiment includes a linear actuator for each of the sensing heads 50, and each sensing The head 50 is on a separate track. However, in this embodiment, the overall measurement system size may increase in the direction of travel of the wafer 12 on the conveyor belt.

An alternate embodiment of a non-contact system for measuring the dopant content of a semiconductor material is schematically illustrated in FIG. A single transmitter 74 (e.g., transmitter or source 16 of Figure 5) housing a single source of infrared radiation is positioned on one side of a conveyor belt (not shown) that supports and transports wafer 12 as One is for example part of a production line for PV cell production lines. The source can be a broadband source with a focusing lens or a laser with an optional wavelength. The source can be a continuous source of broadband infrared. The focused beam is modulated by a light wheel or by electronically modulated laser light onto a steering reflector that directs and focuses the radiation beam onto the wafer Above one of the selected points. It is all as previously discussed with reference to FIG.

In this embodiment, a steered reflector is rotatable about an axis to change the transmitted signal 76 to a selected location on the surface 78 of the wafer 12 at selected intervals for continuous guidance and Focus the beam onto a group of wafers in a row. Although FIG. 7 depicts a plurality of transmitted signals 76 and a corresponding plurality of received signals 82, it should be understood that the system operates continuously, and thus the signals are not simultaneously generated and they are not simultaneously received. Similarly, if a laser is used as the source, the redirecting reflector rotates about an axis to move the beam to contact selected points on the wafer 12 of the group moving on the conveyor belt.

A receiver 80 is disposed on the other side of the conveyor belt, with a focusing element and a reflector being adjusted to see the same point on the wafer illuminated by the source beam 76. The beam of light produced by the radiation 82 is directed by a focusing element to a detector.

When transmitter 74 and receiver 80 are directed to a particular sampling location, the transmitter transmits a radiation beam 76 and the receiver receives such signal 82 reflected from wafer surface 78. Such transmission and reception occurs during a specific period of time, i.e., "sampling period." (The number of samples taken during a defined period of time is known as the "sampling rate"). The shape and size of the portion of the surface 78 of the wafer 12 at the sampling position that is observed is the "sampling area." Within a sampling area, there may be a sub-area defined by the shape and size of a particular area that the receiver can see at any time. This is called a sampling "point".

If the source is a broadband source containing a broad spectrum of infrared energy (e.g., a broadband infrared source), it is necessary to separate the received signal into two equal portions using a splitter that is part of the receiver 80. Then, each half of the signal is focused onto two narrow bandpass filters in receiver 80, each filter having a different center wavelength. The energy of each narrow bandpass filter is focused onto a corresponding detector of the two detectors, converted to a voltage, amplified and converted into a digital signal corresponding to the energy in each frequency band. The slope or ratio between the two measurements is calculated and stored for a given signal position defined by the position of the steering reflector. This is done in the same manner as previously described with reference to FIG. The source beam then moves to a new point on the wafer and the receiver is set to see the same point and the process is repeated for the next location.

If the source is a laser with an optional wavelength, the laser system alternates between two or more wavelengths and is focused at a point with the steerable reflector. The receiver is comprised of a focusing element, and the reflector focuses the received energy on a single detector whose voltage is amplified and at a frequency corresponding to the laser modulation frequency. Was sampled.

An alternate embodiment of a non-contact measurement system for measuring dopant content is schematically illustrated in Figure 8 for measurement at the beginning of a process step or during a continuous series of process steps One or more wafers are then measured at the end of the process step and a change in the infrared reflectance of the wafer is calculated. This change is used to determine the exact impact of the process on each wafer.

This embodiment can be utilized as long as a dopant or a dopant carrier (eg, phosphoric acid) is applied to a wafer surface, dried or diffused into a wafer, implanted in a wafer, deposited into one or Multiple epitaxial layers, or semiconductor processes etched from the surface of a wafer. It can also be utilized in situations where only one wafer is processed to create a surface texture.

In this configuration, the wafer 12 is carried on the conveyor belt 86 in the direction of arrow 88. The wafer is measured by a system as described with reference to Figure 7 (having the same component symbols as in Figure 8) before and after the process or process series, by means of the machine or continuous The machine group (shown as a single entity) 84 is implemented. This configuration measures the reflection coefficient of the base wafer 90 prior to the process and then measures the reflection coefficient of the wafer 92 after the process. Computer 94 controls the process of measurement and comparison. The system described with reference to Figure 4 can be utilized in this system instead of the system described with reference to Figure 7.

Without limiting the generality of the foregoing, examples of the use of this embodiment in certain PV cell fabrication steps will now be described. In the first example, the machine (84) is just a machine for doping the device, and this embodiment is used to measure the wet dopant carrier deposited on the wafer. In the second example, the machine (84) is only an in-line diffusion furnace, and this embodiment is used to measure the furnace to diffuse the dried dopants on the surface of the wafer into the wafer. effect. In a third example, the machine (84) is a diffusion furnace followed by a PSG etching machine, and this embodiment is used to measure the combination of dopant diffusion and etching processes.

An alternate embodiment of a non-contact system for measuring the dopant content of a semiconductor material is schematically illustrated in FIG. Instead of the embodiment described with reference to Figure 4, all of the transmitter 74 and receiver 80 (as in Figure 7) are located in a single support structure 96, and the support structure 96 is moved back and forth together. In the direction of arrow 98, wafer 12 is inspected in a pattern as exemplified in FIG.

One example of the Applicant's method for comparing samples at sensors 20 and 22 is to calculate the difference in signal amplitudes received at sensors 20 and 22, divided by and corresponding sensors 20 and 22 The difference between the centers of each of the passbands of the associated bandpass filters 34 and 36. The slope of the line crossing the center of the passband of the bandpass filters 34 and 36 associated with the corresponding sensors 20 and 22 is plotted on a plot of the reflection coefficient as a function of wavelength. For further clarification, for example, the passband center of the bandpass filter 34 associated with the sensor 20 may be at 8 microns, while the passband center of the bandpass filter 36 associated with the sensor 22 may be at 10 Micron. If the signal amplitude received at the sensor 20 is " x " and the signal value received at the sensor 22 is " y ", the slope is ( y - x )/2. The different slopes represent the amount of different detected dopants, and by utilizing this slope, the effects of amplitude variations caused by the factors described herein are mitigated.

A similar mitigation can also be achieved by utilizing the ratio of signal amplitudes measured at sensors 20 and 22. In this case, the ratio is defined as y/x. Similarly, the difference or ratio between the received signal phase of the sensors 20 and 22 or the received signal polarization can also be utilized.

10 is a diagram of an example correlation curve for a non-contact system for measuring dopant content of a semiconductor material in accordance with an embodiment of the present disclosure. In this example, the dopant content is expressed as surface resistance. The plot of the plot (in this case, a line) is the offline four-point probe measurement of the surface resistance (y-axis) and the measurement of the line slope between the two voltage readings from the two detectors (x-axis). The relationship between. It is produced by placing a series of known and incrementally doped wafers on the conveyor belt and measuring the voltage generated from each of the sensors 20, 22, calculating the line between the two points. The slope (or the ratio of the two voltages) is generated by fitting to a linear model using least squares regression. The observed data points are displayed using diamond-shaped marks, and the best fit is displayed using the line. The R ² value represents the degree to which the calculated line meets the observed measurements, and the closer the value is to 1.0, the better the fit of the line to the observed data. In the example of Figure 10, the R ² value is 0.9486. This line is used to calculate the surface resistance y corresponding to the observed slope X. For example, refer to Figure 10: y = -575.65x + 17.391.

If the slope is -0.1, the surface resistance is: y = -575.65 (0.1) + 17.391 = 74.9 ohms per square area.

Some samples were taken on a sample area on a wafer or substrate. The values of these samples are processed in their entirety (for example, the average is calculated, but this is not the only way) to provide a meaningful measurement. Each sampling area can be clearly defined, and individual samples do not need to be repetitively in the same position between wafer-to-wafer or substrate-to-substrate to obtain wafer-to-wafer or substrate-to-substrate Statistically valid and comparable measurements between the substrates.

The passbands of filters 34 and 36 are selected to be sensitive to the amplitude of the reflected signal. By utilizing a comparison between two different values rather than a single absolute measurement, the measurements are normalized to eliminate variations due to any one or more of the following items: due to spanning multiple samples The positional scan causes changes in the length of the incident and reflected path and the sample to the sample, due to vibration or changes in the three-dimensional position on the surface of the object (for example, due to the irregularity of the belt of the conveyor belt) ") causes sample-to-sample variations in path length, attenuation, or sampling area due to surface texture, crystal boundaries, or other surface artifacts, such as oxides, phosphors, anti-reflective coatings, Changes in pollutants caused by changes in the nature of the signal, from sample to sample, due to changes in temperature caused by changes in the surface temperature of the object, signal attenuation due to changes in atmospheric humidity and/or dust, Phase or polarization, varying ambient light and heat, generating electrical noise in the sensor, drifting in wavelength and/or amplitude on the transmitted signal, And the reference wavelength drift in the receiver, any other source of signal impairment in the measurement environment.

Applicant's invention does not use an analog system. The present invention utilizes a bandpass optical filter to look only at a particular narrow segment of the reflected spectrum and use the difference or ratio between the energies in these two narrow bands to directly infer the free carrier of the semiconductor material concentration. Instead of using a reference spectrum, the present invention isolates energy in two discrete narrow bands. The present invention uses a particular region of the optical frequency that is found to be sensitive to interaction with the doped layer.

In general, the present invention operates by isolating a narrow band of infrared radiation reflected from a semiconductor material and measuring the energy in this band compared to another different narrow band. The applicant is hereby able to directly deduce the amount of dopant in the semiconductor material. Applicants do this by measuring the energy in each frequency band on a set of samples with known doping levels, rather than by comparing the energy in the entire spectrum with a reference spectrum. For a given dopant level from one of each sample as an independent variable, the applicant takes the difference or ratio of the energy levels as a dependent variable, performs a linear regression method and can then be based on each narrow frequency band. The measured energy level directly infers the doping level of a new item.

It will be apparent that the particular embodiments of the invention have been described herein for the purposes of illustration and description Furthermore, although various advantages associated with certain embodiments of the present invention have been described in the context of these embodiments, other embodiments may exhibit such advantages, and not all embodiments necessarily require such Advantages fall within the scope of the present invention. Accordingly, the invention is not limited by the scope of the appended claims.

10. . . Device

12. . . Wafer

14. . . Device

16. . . Transmitter

18. . . receiver

20, 22. . . Sensor

twenty four. . . reflector

26. . . Twilight wheel

28. . . reflector

30. . . Measuring point

32. . . First lens

34. . . First filter

36. . . Second filter

38. . . Second lens

40. . . Third lens

42, 44. . . Amplifier

46. . . Data acquisition board

48. . . computer

50. . . Sensor head

52. . . wheel

54. . . track

56. . . Support beam

58. . . Equipment framework

60. . . power cable

62. . . Power supply and terminal cabinet

64. . . Linear actuator

66. . . arrow

68. . . arrow

72. . . computer

74. . . Transmitter

76. . . Signal sent

78. . . surface

80. . . receiver

82. . . signal

84. . . machine

86. . . conveyor

88. . . arrow

90. . . Basic wafer

92. . . Wafer

94. . . computer

96. . . supporting structure

98. . . arrow

Figure 1 is an absorption spectrum of water;

Figure 2 is a graph of free carrier absorption versus wavelength at different doping levels ( n- Si);

3 is a graph of differential reflectance coefficients of an undoped (W1) c-Si wafer and a doped (W16) c-Si wafer;

4 is a schematic block diagram of a non-contact system for measuring a dopant content of a semiconductor material in accordance with an embodiment of the present disclosure;

5 is a schematic block diagram of a contactless system for measuring a dopant content of a semiconductor material, the system being composed of a single transmitter and a single receiver, in accordance with an alternative embodiment of the present disclosure. Is composed of two sensors;

6 is a schematic top plan view of a method of sampling at various test locations and sampling modes of a wafer in accordance with an embodiment of the present disclosure;

7 is a block diagram of a non-contact system for measuring a dopant content of a semiconductor material in accordance with an alternate embodiment of the present disclosure;

8 is a block diagram of a non-contact system for measuring dopant content of a semiconductor material in accordance with an alternate embodiment of the present disclosure, wherein a pair of systems as depicted in FIG. 7 are used in a blend. Both sides of the compartment;

9 is a block diagram of a non-contact system for measuring a dopant content of a semiconductor material in accordance with an alternate embodiment of the present disclosure, wherein a plurality of transmitters and receivers are disposed in a series of semiconductor material crystals Above the circle;

10 is an example of a non-contact system having a dopant content for measuring a layer of semiconductor material and an area resistance of a layer of semiconductor material measured by a four-point probe, in accordance with an embodiment of the present disclosure. Graphical diagram.

12. . . Wafer

14. . . Device

16. . . Transmitter

18. . . receiver

20, 22. . . Sensor

twenty four. . . reflector

26. . . Twilight wheel

28. . . reflector

30. . . Measuring point

32. . . First lens

34. . . First filter

36. . . Second filter

38. . . Second lens

40. . . Third lens

42, 44. . . Amplifier

46. . . Data acquisition board

48. . . computer

Claims (23)

A non-contact system for measuring a dopant content of a semiconductor material, comprising: a. an infrared radiation source configured to focus infrared radiation on a point on the material; b. And modulating radiation from the source before the radiation strikes the material; c. a first lens arranged to collect radiation reflected from the material and to focus the radiation; d. a bandpass filter configured to receive the radiation from the first lens, the first filter being configured such that a narrow wavelength band of the radiation passes through the filter and reflects a remaining portion of the radiation; e. a second band pass filter configured to receive radiation reflected from the first filter, the second filter being configured to cause a narrow wavelength band of the radiation to pass through the second filter and configured Having the passed wavelength band different from the narrow wavelength band through which the first filter passes; f. a first radiation detector configured to receive the first band pass filter Radiation and is matched Determining a first level of energy; a second radiation detector configured to receive radiation passing through the second filter and configured to determine a second level of energy; and h. a calculator configured to compare the first and second levels and utilize a correlation curve to return a dopant content value based on a series of dopants having known different levels The same half of the content A comparison of the energy levels of the conductor material to the two sensors correlates to the dopant content of the semiconductor material. The system of claim 1, further comprising a focusing device between the modulator and the wafer to focus the radiation at the point of the material, the focusing device being selected from the group consisting of Group: a. a parabolic reflector; b. an adjustable reflector; c. an elliptical reflector; d. a parabolic lens; and e. an optical lens. The system of claim 1, wherein the modulator is selected from the group consisting of: a. a modulator utilizing a high speed crucible wheel; b. a modulator utilizing the pulse modulation of the source And c. a modulator that utilizes the frequency modulation of the source. The system of claim 1, wherein the semiconductor material is selected from the group consisting of: a. a doped germanium material; b. an undoped germanium material; c. a doped germanium material; d. an undoped germanium material e. a doped indium material; f. an undoped indium material; g. a combination of aluminum, boron, gallium, indium, phosphorus, arsenic and antimony doped or Yes An undoped tantalum or niobium material; and h. a film having any of the above materials on a substrate, which may be a semiconductor or may be a non-conductive material. The system of claim 1, further comprising a second lens configured to receive radiation through the second band pass filter and configured to focus the radiation on the first detector . The system of claim 1, further comprising a third lens configured to receive radiation reflected from the first band pass filter and configured to focus the radiation on the second detector on. A system of claim 5, further comprising a third lens configured to receive radiation reflected from the first band pass filter and configured to focus the radiation on the second detector on. The system of claim 1, wherein the source of radiation is a laser. The system of claim 1, wherein the source of radiation is a source of broadband infrared radiation. A system as claimed in claim 1, wherein the pass band of each band pass filter is between 50 nm and 500 nm in width. As in the system of claim 1, the center of the pass band of a filter is between 1 and 20 microns. The system of claim 1, wherein the center wavelength of the pass band of the second filter is between 1 and 20 microns and is different from the center wavelength of the pass band of the first filter. Such as the system of claim 1 of the patent scope, wherein the first and the The difference between the center wavelengths of the two filters is between 1 and 10 microns. The system of claim 1, wherein the difference between the center wavelengths of the first and second filters is 2 microns. The system of claim 1, wherein a center wavelength of the first filter is set at 8.06 microns, and a center wavelength of the second filter is set to 10.5 microns, wherein each filter system has a medium The width of the passband between 200 and 400 nm. A system as claimed in claim 1, wherein a band pass filter has a central pass band having a pass band of +/- 125 nm at about 8 microns and another filter having a pass band of about 10.5 microns A central passband with a +/-175 nm passband. A system as claimed in clause 1, wherein the detector after each filter is a mass spectrometer detector capable of measuring power in a narrow wavelength band selected by a user. The system of claim 2, wherein the focusing device is configured to direct the radiation at the point at an angle of 45 degrees relative to the surface of the material. The system of claim 1, wherein the source focuses the radiation on a selected focus area, and wherein the modulator is a light wheel, the light wheel train configured to modulate the focus area Radiation. A non-contact system for measuring a dopant content of a semiconductor material, comprising: a. a broadband infrared radiation source configured to focus infrared radiation on a point on the crucible; b. a modulator for modulating the infrared radiation from the source before the radiation strikes the crucible; c. a first lens configured to collect infrared radiation reflected from the crucible and focus The radiation; d. a beam splitter disposed adjacent to a focus of the first lens, the beam splitter configured to separate the radiation into two radiation streams; e. a first narrow band pass filter, Is configured to receive a first radiation stream of the two radiation streams at a first predetermined wavelength; f. a second narrow band pass filter configured to receive the center in one a second radiation stream of the two radiation streams of the second predetermined wavelength, the second predetermined wavelength is different from the first predetermined wavelength; g. a first infrared detector, which is set Receiving and configured to determine an energy level of radiation passing through the first band pass filter; h. a second infrared detector configured to receive and configured to determine by the second band pass filter The energy level of the radiation of the device; and i. a calculator that is configured to compare The first and second levels and using a correlation curve to return a dopant content value based on the arrival of the same semiconductor material for a series of dopant levels having known different levels A comparison of the energy levels of the two sensors correlates the dopant content of the semiconductor material. A method for non-contact measurement of dopant content of a semiconductor material, comprising the steps of: a. directing a source of modulated infrared radiation to a measurement point on the material; b. directing radiation reflected from the material onto a first band pass filter, the first band pass filter being configured such that a range of wavelengths of the radiation passes through the filter and reflects the remainder of the radiation; c. directing radiation reflected from the first filter onto a second band pass filter, the second band pass filter being configured such that a range of wavelengths of the radiation passes through the second band pass filter And configured such that the wavelength range is different from the wavelength range through which the first band pass filter passes; d. determining an energy level of radiation passing through the first band pass filter; e. The energy level of the radiation of the second band pass filter; f. comparing the energy levels determined in steps d and e; and g. according to the comparison, using a correlation curve to calculate the dopant content of the material, The correlation curve correlates the dopant content of the semiconductor material based on a comparison of the energy levels of the same semiconductor material for a series of dopant levels having known different levels to the two sensors. A method for determining the effect of one or more process steps on a semiconductor wafer in a semiconductor material line, comprising the steps of: a. setting a first system as in claim 1 One or more processing steps of the line are at a selected upstream point of the location to determine a dopant level level of the wafer at an upstream point in the line; b. setting as in claim 1 The second system has one or more processing steps of the line at a selected downstream point of the location to determine a dopant level of the wafer at a downstream point in the line; c. operating the production line to move a series of wafers from the upstream point to the downstream point through the one or more process steps; d. utilizing the first system to determine wafer doping at a point upstream of the line a level of impurity content; e. utilizing the second system to determine a dopant level of the wafer at a downstream point of the line; and f. comparing a dopant level of the wafer at the downstream point A level of dopant content at the upstream point is obtained to obtain a difference in dopant level between the wafer at the upstream point and the wafer at the downstream point. The method of claim 22, wherein the one or more process steps are performed by one or more or a combination of: a wet dopant chemical application machine, a wet type Doping carrier drying machine, one-line diffusion furnace, one batch diffusion furnace, one laser annealing machine, one ion implantation machine, one epitaxial layer deposition machine, one PSG etching machine, one wafer etching machine And a wafer-forming machine.