Vacuum processing device and vacuum processing method

Abstract

A vacuum processing apparatus includes: a processing unit including: a processing chamber disposed in a vacuum container; and a detector for detecting a thickness or an end point of a target film on a wafer during processing of the wafer using light from the wafer, the detector detecting the thickness or the end point by comparing a pre-acquired data pattern indicating light intensities of a plurality of wavelengths related to film thickness and using wavelength as a parameter with a true data pattern indicating light intensities of the plurality of wavelengths obtained at a specific time during processing, the data pattern being obtained by dividing differential coefficient values ​​of time series data of light intensities of the plurality of wavelengths by time series data indicating light intensity values ​​of the plurality of wavelengths.

Description

Technical Field

[0001] The present invention relates to a vacuum processing apparatus or a vacuum processing method in which, in a manufacturing process of a semiconductor device, a film layer to be processed on the surface of a sample in the shape of a substrate such as a semiconductor wafer disposed in a processing chamber is processed using plasma formed in a processing chamber, and the present invention relates to a vacuum processing apparatus or a vacuum processing method in which the processing is performed while detecting the remaining film thickness and processing depth of the film layer on the sample surface. Background Art

[0002] In semiconductor device manufacturing, dry etching is widely used to remove pre-formed layers of various materials and dielectric materials on the surface of semiconductor wafers prior to the process of forming circuit patterns. To improve the yield rate of semiconductor device manufacturing by using this dry etching method to form patterns with high precision, it is necessary to stop the etching process at a desired remaining film thickness or desired etching depth during the etching process of the target film layer. In other words, to improve the processing accuracy of semiconductor device circuits, it is necessary to more accurately detect the etching endpoint during the dry etching process.

[0003] In view of the above requirements, as the target film layer is etched and the remaining film thickness changes (decreases during etching), a technique for determining the etching endpoint based on the intensity of light of a specific wavelength emitted from the film, detected during the etching process, or changes therein, is known in the related art. For example, the technique disclosed in JP-A-2007-234666 (Patent Document 1) is known in the prior art. This prior art describes a method for determining the endpoint based on temporal changes in the amount of reflected light from a wafer during etching.

[0004] Specifically, Patent Document 1 discloses a technique in which, at each sampling time during wafer processing, the intensity of interference light of multiple wavelengths, formed by light reflected in a processing chamber from the upper and bottom surfaces of a film to be processed, or the rate of change thereof over time, is detected. The remaining film thickness or etching depth of the film during processing is detected by comparing the pattern of the intensity or rate of change, with the wavelength detected at each time as a parameter, with reference data to be used as a comparison, thereby determining whether the etching process has reached an end point. The reference data to be used as a comparison indicates the correlation between the remaining film thickness of the film to be processed and the pattern of the value of the intensity of the interference light, with the wavelength as a parameter, or the rate of change thereof, obtained before processing the wafer. A sample wafer for obtaining the data is pre-etched under the same conditions as those used for processing a product wafer, wherein the semiconductor wafer has a structure substantially the same as or similar to that of a wafer used to manufacture a semiconductor device as a product, and the reference data is obtained during etching of the sample and stored as a database in a storage device configured to communicate with an apparatus for performing the etching process.

[0005] However, the above-mentioned technology in the related art has problems because the following points are not fully considered.

[0006] That is, (A) when detecting the endpoint of a process performed in a process chamber different from the process chamber in which the reference data for determining the remaining film thickness stored in the database was obtained, since the balance of the spectrum of the interference light generated in each process chamber varies from process chamber to process chamber, even if the endpoint of the process is determined in each process chamber using the same database data, the data related to the intensity of the interference light obtained from the process chamber when the target remaining film thickness or the process endpoint is actually reached will differ from the value of the reference data in the database. Therefore, even if the determination accuracy is within the allowable range for determining the process endpoint in one process chamber and the determination accuracy is outside the allowable range for determining the process endpoint in another process chamber, a phenomenon of determining that the correct endpoint has been reached may occur, and the accuracy of determining that the process endpoint or the target remaining film thickness has been reached may differ in each of the multiple process chambers.

[0007] Furthermore, (B) even when the atmosphere in the processing chamber differs from the atmosphere at the time of data acquisition, differences in the spectral balance of interference light may occur when processing multiple wafers in the same processing chamber. Therefore, when conditions such as the surface or gas in the processing chamber differ from those when reference film thickness data was acquired by etching multiple batches in the processing chamber and attaching deposits to the chamber, the accuracy of determining whether the target film thickness or endpoint has been reached may vary.

[0008] Therefore, in the above-mentioned prior art, the problem that variations in processed shapes increase due to wafer processing and the yield of the process of manufacturing semiconductor devices decreases is not considered. Summary of the Invention

[0009] An object of the present invention is to provide a vacuum processing apparatus or a vacuum processing method for improving the processing yield.

[0010] The above object is achieved by a vacuum processing apparatus comprising: a processing unit including: a processing chamber disposed in a vacuum container and in which plasma is formed; a sample stage disposed below a space in the processing chamber in which the plasma is formed, with a wafer having a film to be processed mounted on its upper surface; a light receiving unit disposed above the processing chamber and configured to receive light from the wafer during processing of the wafer using the plasma; and a detector connected to the light receiving unit and configured to detect the thickness of the film to be processed during the processing or the arrival of an end point of the processing. The detector is configured to detect the thickness of the film to be processed by comparing a pattern for detection with a true pattern, wherein wavelengths of predetermined values ​​correlated with the intensities of light of a plurality of wavelengths, obtained in advance and varying with respect to the thickness of the film to be processed, are used as parameters, the true pattern being a true pattern of predetermined values ​​correlated with the intensities of the light of the plurality of wavelengths obtained at specific times during the processing, and the predetermined value correlated with the intensity of the light is obtained by dividing a differential coefficient value of time series data of the intensities of the light of the plurality of wavelengths by time series data indicating the intensity values ​​of the light of the plurality of wavelengths.

[0011] According to the present invention, a vacuum processing apparatus or a vacuum processing method that improves the processing yield can be provided. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1A and Figure 1B 1 is a diagram showing an outline of the overall structure of a vacuum processing apparatus according to an embodiment of the present invention.

[0013] Figure 2 is a schematic diagram showing the Figure 1A and Figure 1B 1 is a diagram schematically illustrating the configuration of a vacuum processing unit of a vacuum processing apparatus according to an embodiment shown in FIG.

[0014] Figure 3 It shows that according to Figure 1A and Figure 1B Flowchart of the flow of the operation of detecting the etching amount in the plasma processing apparatus of the embodiment shown in FIG.

[0015] Figure 4 1 is a flowchart showing the flow of operation of a conventional vacuum processing apparatus.

[0016] Figure 5A and Figure 5B It is shown that according to Figure 1A and Figure 1B Graphs showing wavelength distributions of luminous intensity obtained during wafer etching processes performed by the vacuum processing apparatus of the embodiment shown in FIG. 1 and the prior art. DETAILED DESCRIPTION

[0017] Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings.

[0018] In this embodiment, as at least one process for manufacturing a semiconductor device, a vacuum processing apparatus includes multiple vacuum processing units. Each of the vacuum processing units includes a processing chamber within a vacuum container within any of the processing units. A sample in the form of a substrate to be processed using plasma (such as a semiconductor wafer) is placed within the processing chamber. The target film layer of the film structure includes a mask layer pre-formed and disposed on the surface of the sample and a target film layer located beneath the mask layer. The target film layer of the film structure is etched using plasma generated within the processing chamber. In this case, during processing of a sample (real wafer) used for device manufacturing, the intensity of interference light of multiple wavelengths originating from the film structure on the wafer surface is detected, and a pattern (hereinafter referred to as a real pattern) is calculated, indicating the dependence of the differential value of the intensity of the interference light on wavelength (with the wavelength set as a parameter) over time. Furthermore, before processing the real wafer, the etching amount of the real wafer is calculated by comparing the real pattern with a pattern (hereinafter referred to as a standard pattern) indicating the wavelength dependence (wavelength being set as a parameter) of the differential value of interference light with respect to the remaining film thickness and etching depth (etching amount) of the film layer to be processed, which is obtained by etching a sample (sample wafer) having a structure substantially equivalent to that of the real wafer in the processing chamber of the same vacuum processing unit, and determining whether or not the etching end point has been reached.

[0019] [Example 1]

[0020] In the following, reference will be made to Figures 1A to 4 to describe embodiments of the present invention.

[0021] First, refer to Figure 1A and Figure 1B The overall structure of a vacuum processing apparatus including a plurality of vacuum processing units including a structure for detecting an etching amount (here, an actual wafer etching depth or a remaining film thickness) according to the present embodiment will be described. Figure 1A and Figure 1B : is a diagram showing an outline of the overall structure of a vacuum processing apparatus according to an embodiment of the present invention. Figure 1A is a perspective view of the vacuum processing apparatus 100, and Figure 1B1 is a cross-sectional view of the vacuum processing apparatus 100 as viewed from above.

[0022] The vacuum processing apparatus 100 according to an embodiment of the present invention shown in the figure is roughly divided into an atmospheric-side block 101 and a vacuum-side block 102. The atmospheric-side block 101 is a portion for transporting substrate-shaped samples, such as semiconductor wafers, as processing materials, at atmospheric pressure and for performing positioning and other operations. The vacuum-side block 102 is a block for transporting substrate-shaped samples, such as wafers, at a pressure reduced from atmospheric pressure and for performing processing in a predetermined vacuum processing chamber. Furthermore, between the position of the vacuum-side block 102 where the transport and processing are performed and the atmospheric-side block 101, a portion is provided that connects the vacuum-side block 102 to the atmospheric-side block 101 at this position, thereby increasing and decreasing the pressure between atmospheric pressure and vacuum pressure while the sample is inside.

[0023] The atmospheric side block 101 includes: a roughly rectangular shell 106, which includes an atmospheric side transfer robot 109 inside; and a plurality of box stages 107, which are attached to the front surface side of the shell 106 and on which boxes are placed, in which samples in the shape of substrates to be processed such as semiconductor wafers for processing or cleaning (hereinafter, wafers 126) are contained.

[0024] The vacuum-side block 102 is disposed between the first and second vacuum transfer chambers 104-1, 104-2, and the atmospheric-side block 101, and includes one or more lock chambers 105 for switching pressure between atmospheric pressure and vacuum pressure while wafers, the materials to be processed and exchanged between the atmospheric and vacuum sides, are positioned within the lock chambers. The lock chambers are vacuum containers whose internal space can be adjusted to the aforementioned pressures and are equipped with a passage for transferring wafers within the chambers, as well as a valve that can open and close the passage and hermetically seal the passage at a connection point (not shown), thereby airtightly separating the atmospheric and vacuum sides. Furthermore, the internal space is provided with a storage unit capable of accommodating and holding multiple wafers with vertical gaps therebetween. When the wafers are accommodated, the lock chambers are closed by valves and hermetically separated.

[0025] The first vacuum transfer chamber 104-1 and the second vacuum transfer chamber 104-2 are units each including a vacuum container having a substantially rectangular planar shape, and are two units having structural differences that can be considered to be substantially the same. A purge gas, such as an inert gas, whose flow rate is adjusted based on a command signal from a control unit (not shown) of the vacuum processing apparatus 100, is introduced into the vacuum transfer chambers to evacuate the vacuum transfer chambers. As a result, the first vacuum transfer chamber 104-1 and the second vacuum transfer chamber 104-2 are configured to be adjustable to a pressure value relatively lower than that of the other vacuum processing chamber 103 and the intermediate vacuum transfer chamber 111, to the same pressure value or a pressure value that can be considered to be the same pressure value, or to a pressure value relatively higher than that of the vacuum processing chamber 103 and the intermediate vacuum transfer chamber 111.

[0026] The vacuum transfer intermediate chamber 111 is a vacuum container whose interior can be decompressed to a relatively higher pressure than that of the other vacuum processing chambers. The vacuum transfer chambers are connected to one another, and their internal chambers communicate with one another. A valve (not shown) is provided between the vacuum transfer intermediate chamber 111 and a vacuum transfer chamber connected and located at an end in the front-to-back direction (left-to-right in the figure). Closing the valve seals the vacuum transfer intermediate chamber and the vacuum transfer chambers. In this embodiment, the vacuum transfer intermediate chamber 111 is not equipped with a mechanism for purging or evacuating the space containing the wafers. Instead, the transfer intermediate chamber 111 serves as a path for transferring wafers between one vacuum transfer chamber and another vacuum transfer chamber connected to each other, sandwiching the transport intermediate chamber 111.

[0027] Furthermore, within the interior of the vacuum transfer intermediate chamber 111, a storage unit is provided. Multiple wafers are placed at both ends of each wafer, with the gap between the upper and lower surfaces opened and held horizontally. The interior chamber also functions as a relay chamber, temporarily accommodating the wafers while they are being transferred between the first vacuum transfer chamber 104-1 and the second vacuum transfer chamber 104-2. Specifically, wafers placed in the storage unit are loaded by one of the vacuum transfer robots 108 in one vacuum transfer chamber and unloaded by one of the vacuum transfer robots 108 in the other vacuum transfer chamber, where they are transferred to the vacuum processing chamber 103 or the lock chamber 105 connected to that vacuum transfer chamber.

[0028] A vacuum transfer intermediate chamber 111 is disposed between the sidewalls of the first vacuum transfer chamber 104-1 and the second vacuum transfer chamber 104-2. Each of the sidewalls corresponds to a surface located opposite each other, and the first vacuum transfer chamber 104-1 and the second vacuum transfer chamber 104-2 are connected to each other. Furthermore, a vacuum processing chamber 103, which depressurizes the interior, transfers wafers therein, and processes the wafers, is connected to the other surfaces. In this embodiment, the vacuum processing chamber 103 represents an entire unit including an electric and magnetic field generating unit, which includes a vacuum container and an exhaust unit including a vacuum pump that exhausts the depressurized space within the container. The semiconductor wafers 126 are then subjected to etching, ashing, or other processing within the internal processing chamber.

[0029] Furthermore, each vacuum processing chamber 103 is connected to a pipeline through which process gases, supplied according to the process to be performed, flow. Each vacuum processing chamber 103 can be depressurized to a relatively lower pressure than that of the other vacuum transfer chambers and the vacuum transfer intermediate chamber. Furthermore, the internal pressure of each vacuum processing chamber can be adjusted by receiving commands from a command device (not shown) depending on the process performed within each vacuum processing chamber. In other words, if the types of processes to be applied to wafers within the multiple vacuum processing chambers differ, the pressure within each vacuum processing chamber may not be exactly the same. However, the pressure within each vacuum processing chamber can be varied and adjusted to an optimal pressure for each process.

[0030] The first vacuum transfer chamber 104 - 1 and the second vacuum transfer chamber 104 - 2 are respectively configured to be connectable to two vacuum processing chambers 103 , and in this embodiment, two vacuum processing chambers 103 are connected to the first vacuum transfer chamber 104 - 1 and the second vacuum transfer chamber 104 - 2 .

[0031] The interior of each of the first vacuum transfer chamber 104-1 and the second vacuum transfer chamber 104-2 is a transfer chamber, and in the first vacuum transfer chamber 104-1, a vacuum transfer robot 108 is provided in the central portion of its internal space for transferring wafers 126 between the lock chamber 105 and one of the vacuum processing chamber 103 and the vacuum transfer intermediate chamber 111 under a vacuum state. In the second vacuum transfer chamber 104-2, the vacuum transfer robot 108 is provided in the central portion of the interior of the second vacuum transfer chamber 104-2 in a manner similar to that described above, and transfers the wafers 126 to one of the vacuum processing chamber 103 and the vacuum transfer intermediate chamber 111. The vacuum transfer robots 108 have the same configuration.

[0032] In the vacuum transfer robot 108, wafers are placed on its arm, and in the first vacuum transfer chamber 104-1, the vacuum transfer robot 108 carries wafers in and out between the wafer stage provided in the vacuum processing chamber 103 and one of the lock chamber 105 and the first vacuum transfer intermediate chamber 111. As described above, although not shown, a passage is provided that communicates with each of the vacuum processing chamber 103, the lock chamber 105, the vacuum transfer intermediate chamber 111, the first vacuum transfer chamber 104-1, and the second vacuum transfer chamber 104-2 through a valve that can be airtightly closed and opened, and the passage is opened and closed by the valve.

[0033] Next, an operation of performing processing on a wafer in such a vacuum processing apparatus 100 will be described below.

[0034] A plurality of wafers housed in a cassette placed on one of the cassette stages 107 receive commands from a command device (not shown) connected to the vacuum processing apparatus 100 via some communication unit to adjust the operation of the vacuum processing apparatus 100, or receive commands from a command device or the like of a production line in which the vacuum processing apparatus 100 is installed, and start their processing. The atmosphere-side transfer robot 109, having received the command from the command device, removes a specific wafer from the cassette and transfers the specific wafer to the lock chamber 105.

[0035] For example, in the lock chamber 105 that transfers and stores wafers, a valve (not shown) on the atmospheric side is closed and sealed while the wafers being transferred are stored, and the pressure is reduced to a predetermined level. Next, the valve on the side of the lock chamber 105 facing the first vacuum transfer chamber 104-1 is opened, thereby connecting the first lock chamber 105 and the transfer chamber of the vacuum transfer chamber 104-1 to each other.

[0036] The vacuum transfer robot 108 extends its arm into the lock chamber 105, receives the wafer in the lock chamber 105 from the wafer 126 support portion at the distal end of its arm, and carries the wafer out to the first vacuum transfer chamber 104-1. Furthermore, when a wafer is removed from a cassette, the first vacuum transfer robot 108 transfers the wafer placed on its arm to a vacuum processing chamber 103 or a vacuum transfer intermediate chamber 111 connected to the first vacuum transfer chamber 104-1 along a transfer path pre-specified by a command device. For example, the wafer transferred to the vacuum transfer intermediate chamber 111 is then carried out of the vacuum transfer intermediate chamber 111 to the second vacuum transfer chamber 104-2 by the vacuum transfer robot 108 disposed in the second vacuum transfer chamber 104-2, and then carried into one of the vacuum processing chambers 103 that is the destination of the predetermined transfer path.

[0037] After the wafer is transferred to one of the vacuum processing chambers 103, the valve that opens and closes between the vacuum processing chamber 103 and the first vacuum transfer chamber 104-1 connected to the vacuum processing chamber 103 is closed to seal the vacuum processing chamber. Subsequently, a process gas is introduced into the processing chamber, and the pressure in the vacuum processing chamber is adjusted to a pressure suitable for processing. An electric field or a magnetic field is applied to the vacuum processing chamber to excite the process gas, forming a plasma in the processing chamber, and processing the wafer.

[0038] The valves that open and close between one of the vacuum processing chambers 103 for processing wafers and the vacuum transfer chamber connected to the vacuum processing chamber 103 are opened in the following state: the other valves that can open and close the space including the vacuum transfer chamber and the space connected to the vacuum transfer chamber and communicating with each other are closed by receiving commands from the command device. For example, before the valve defined between one of the vacuum processing chambers 103 and the vacuum transfer chamber to which the vacuum processing chamber 103 is connected is opened, the command device commands or confirms the closing of any valve that opens and closes a gate provided on a passage for transferring wafers 126 to each vacuum processing chamber. The closure of the valves prevents the vacuum processing chamber and the other vacuum processing chamber from communicating with each other, and after this confirmation, the valve that seals the one of the vacuum processing chambers 103 is opened.

[0039] When completion of the etching process of the wafer is detected, after confirming that the valve between the other vacuum processing chamber 103 and the second vacuum transfer chamber 104-2 is closed to hermetically seal the two, the valve opened and closed between one of the vacuum processing chambers 103 and the second vacuum transfer chamber 104-2 connected to the vacuum processing chamber 103 is opened, and the vacuum transfer robot 108 moves the processed wafer out to its interior, and moves the wafer into the lock chamber 105 through a transfer path opposite to the transfer path when the wafer was moved into the processing chamber.

[0040] When the wafer is transferred to the lock chamber 105, the valve that opens and closes the passage connecting the lock chamber 105 with the transfer chamber of the first vacuum transfer chamber 104-1 is closed, and the pressure in the lock chamber 105 is increased to atmospheric pressure. Thereafter, the valve defining the inside of the housing 106 is opened, the interior of the lock chamber 105 and the interior of the housing 106 are connected to each other, and the atmosphere-side transfer robot 109 transfers the wafer from the lock chamber 105 to the original cassette and returns the wafer to its original position in the cassette.

[0041] Next, we will refer to Figure 2 Describe in more detail Figure 1A and Figure 1B 1 . The structure of any one of the four vacuum processing units 103-1, 103-2, 103-3, and 103-4 of the vacuum processing apparatus 100 shown in FIG. In the drawing, any one of the vacuum processing units will be described with reference numeral 103. Figure 2 It is schematically shown Figure 1A and Figure 1B 1 is a diagram schematically illustrating the configuration of a vacuum processing unit of a vacuum processing apparatus according to an embodiment shown in FIG.

[0042] The plasma processing unit 103 shown in the figure includes a processing chamber 2 within the reduced pressure of any of the vacuum transfer chambers connected to the plasma processing unit 103. The processing chamber 2 is held at the distal end of the arm of a vacuum transfer robot 108 positioned within the vacuum transfer chamber and is carried in through a door with an open valve. The vacuum processing unit 103 of this embodiment generally comprises a vacuum container housing the processing chamber 2; a plasma formation unit (not shown) disposed outside the upper portion of the vacuum container and providing an electric or magnetic field for forming plasma in the processing chamber 2; and an exhaust unit (not shown) disposed below the vacuum container and including a vacuum pump (not shown) communicating with the processing chamber 2 and connected to an exhaust port at the bottom of the processing chamber 2. The vacuum container houses the processing chamber 2, which serves as a space for etching wafers 4, such as semiconductor wafers, that are transported and positioned therein. A sample stage 5 is disposed below the interior of the vacuum container, on which the wafers 4 transferred onto its upper surface are placed. In addition, the plasma processing device 1 is provided with an etching amount detection device 8 connected to the vacuum container, and the etching amount detection device 8 detects the etching amount of the etching object film using light received from the inside of the processing chamber 2 during the etching process of the etching object film pre-configured on the upper surface of the wafer 4 using the plasma 3 formed in the processing chamber 2.

[0043] In the processing chamber 2, an etching processing gas is introduced from a gas introduction unit (not shown), atoms or molecules of the etching processing gas are excited by an electric field or a magnetic field generated by a plasma forming unit that receives power from a high-frequency power supply (not shown), and are ionized or dissociated to form a plasma 3, and a wafer 4 such as a semiconductor wafer set on a sample stage 5 below the processing chamber 2 is etched by the plasma 3.

[0044] The film structure to be etched, pre-applied on the surface of wafer 4 in this embodiment, comprises multiple layers of films arranged in a vertical direction, including a resist (comprising an organic material) layer as an upper mask layer and a film layer to be etched below the resist layer. The process for processing wafer 4 illustrated in this embodiment constitutes a semiconductor device manufacturing process in which a process gas containing carbon (C) is supplied to processing chamber 2 to form a plasma. When the film to be etched is viewed from above, grooves or holes are formed in the shape of the mask pattern to form a circuit pattern. It is known that in such a structure, the attenuation of light having a wavelength in the ultraviolet region of the interference light emitted by the film structure is relatively large.

[0045] Light emitted from plasma 3 during the etching process on wafer 4 is reflected at the interface between the film structure of the etching target film (processing target film) layer on the upper surface of wafer 4 and the film structure existing below it, forming interference light 6. Specifically, interference light 6 includes a component caused by interference due to the remaining film of the etching target film, and changes in its intensity can be detected to determine the film thickness. Interference light 6 and non-interference light emissions are received by a light receiver positioned above processing chamber 2 and transmitted via optical fiber 7 to etching amount detection device 8.

[0046] For each predetermined time (sampling interval) during the processing of the wafer 4, the etching amount detection device 8 calculates the light intensity pattern of the interference light using the wavelength of the interference light as a parameter, using a signal indicating the light intensity of multiple wavelengths obtained from the light emission output from the spectroscope 9. The pattern is then compared with data pre-stored in a differential waveform pattern database 15 showing the correlation between film thickness and the intensity pattern of the interference light, thereby calculating the remaining film thickness and etching amount of the etching target film. Specifically, the etching amount detection device 8 includes the spectroscope 9, a first digital filter 10, a differentiator 11, a second digital filter 12, a differential waveform divider 13, a differential waveform comparator 14, a differential waveform pattern database 15, a regression analyzer 16, an endpoint determination device 17 that determines the etching endpoint based on the comparator's results, and a display 18 that displays the endpoint determination device's determination results. The light emission transmitted through the optical fiber 7 and introduced into the spectroscope 9 is separated into multiple predetermined wavelengths, and the intensity of the light at each wavelength is detected and converted into a digital signal indicating the intensity.

[0047] It should be noted that Figure 1A and Figure 1B The functional structure of the etching amount detection device 8 is shown. In addition to the display 18 and the spectroscope 9, the actual structure of the etching amount detection device 8 can also be configured with a storage device, including: a CPU; a ROM for storing various data (for example, etching depth and film thickness detection processing program, differential waveform pattern database of interference light 6); a RAM for storing detection data; an external storage device; a data input and output device; and a communication control device. This also applies to Figure 2 and Figure 3 .

[0048] The interference light from the film structure on the upper surface of the wafer 4 captured by the spectroscope 9 during processing includes multiple wavelengths and is converted into a current detection signal corresponding to the intensity of each wavelength of light, and further converted into a voltage signal. The interference light of multiple wavelengths received and sent to the spectroscope 9 at an arbitrary sampling time i is output as a sampling signal indicating the intensity of each of the multiple wavelengths at time i, and is output as time series data y ijStored in a storage device (not shown) such as RAM.

[0049] Next, the time series data y from the spectroscope 9 ij The data is transmitted to the first digital filter circuit 10, and is smoothed by removing data of a predetermined frequency or higher in the data, and the data obtained by the processing is used as the smoothed time series data Y ij is stored in a storage device such as RAM. Smoothed time series data Y ij The time series data d is transmitted to the differentiator 11, which calculates the differential coefficient value (first differential value or second differential value) at the predetermined sampling time. ij And store it in a storage device such as RAM. Time series data of differential coefficient value d ij is transmitted to the second digital filter circuit 12, is smoothed again, and is used as the smoothed differential coefficient time series data D ij Stored in a storage device such as RAM.

[0050] Here, the smoothing differential coefficient time series data D will be described. ij As the first digital filter circuit 10, for example, a quadratic Butterworth type low-pass filter is used. Using the quadratic Butterworth type low-pass filter, the smoothed time series data Y is obtained by equation (1): ij .

[0051]

[0052] Here, the coefficients a and b have different values ​​depending on the time of each sampling time (sampling interval or sampling frequency) and the cutoff frequency. For example, the coefficient values ​​are a2 = -1.143, a3 = 0.4128, b1 = 0.067455, b2 = -0.013491, and b3 = 0.067455 (sampling frequency is 10 Hz and cutoff frequency is 1 Hz).

[0053] The time series data di of the quadratic differential coefficient value is obtained by using the time series data Y of 5 points. i The polynomial adaptive smoothing differentiation method is calculated by the differentiator 11 according to equation (2) as follows.

[0054] j = 2

[0055]

[0056] j = -2

[0057] Here, as the weighting factor w j Examples of values ​​for w -2 =2,w-1 =-1, w0=-2, w1=-1, w2=2.

[0058] Smoothed differential coefficient time series data D ij The second digital filter circuit 12 is formed by, for example, a quadratic Butterworth type low-pass filter, by using the time series data d of the differential coefficient value. i It is calculated as follows according to equation (3).

[0059]

[0060] Then, the smoothed differential coefficient time series data D is divided by the differential waveform divider 13. ij Divide by the smoothed time series data Y ij , the true mode Q indicating the wavelength dependence of the differential value of the interference light intensity is determined ij (The wavelength is set as a parameter).

[0061] On the other hand, the differential waveform pattern database 15 stores the interference light pattern data Ps in advance. j , interference light pattern data Ps j It is obtained when the wafer 4, which is the processing object for manufacturing semiconductor devices, and a test wafer having a surface film structure identical to that of the wafer 4 in material, shape, and structure are etched under the same conditions as the wafer 4. Interference light pattern data Ps j This includes a mode in which the wavelength of the intensity of interference light from the etching target film corresponding to different remaining film thicknesses of the etching target film or the value of data indicating different remaining film thicknesses is set as a parameter. The differential waveform pattern database 15 is stored in a storage device such as a RAM or ROM (not shown), a hard disk, or a DVD in the etching amount detection device 8.

[0062] The differential waveform comparator 14 compares the differential waveform pattern database 15 with the true pattern Q of the interference light corresponding to a predetermined sampling time. ij That is, calculate the pattern data Ps stored in the differential waveform pattern database 15 j With real mode Q ij The difference between the two, determining the data of the pattern with the minimum value of the difference, detecting the remaining film thickness corresponding to the data of the pattern as the instantaneous film thickness value Z at the sampling time i , and the instantaneous film thickness value Z i The value of is stored as time series data in the storage device in the etching amount detection device 8.

[0063] In the regression analyzer 16, the output from the differential waveform comparator 14 is received or the instantaneous film thickness Z at the sampling time i stored in the storage device is read. i The instantaneous film thickness value before time i is read from the storage device, and regression analysis using the output or data of the instantaneous film thickness Zi and the instantaneous film thickness value before time i is performed, and the film thickness value at time i is calculated based on the result of the regression line approximation. That is, the linear regression line is obtained by the regression analyzer 16. (Y: Remaining film amount, t: Etching time, X a : absolute value of Xb, is the etching rate, Xb: initial film thickness), and the film thickness Y at sampling time i is calculated based on the regression line i (Calculated film thickness) value.

[0064] Next, the calculated film thickness Y obtained will be indicated i The data of the value of is transmitted to the end point determination device 17, in which the film thickness Y is i The value of is compared with the film thickness (target film thickness) value as the target of the etching process, and the film thickness Y is determined. i When the etching amount of the target film of the wafer 4 is equal to or less than the target film thickness value, the etching amount reaches the target, and the result is displayed on the display 18. Thereafter, the generation of the electric field or magnetic field in the plasma forming unit is stopped, the plasma 3 disappears, and the etching process of the target film of the wafer 4 is completed, or the processing conditions such as the etching process gas and pressure are changed and the processing of the target film is continued.

[0065] In this embodiment, the differential waveform divider 13 divides the smoothed differential coefficient time series data D output from the second digital filter 12 into ij Divided by the smoothed time series data Y output from the first digital filter 10 ij , and calculate the true pattern Q indicating the wavelength dependence of the differential value of the intensity of the interference light ij (The wavelength is set as a parameter.) In addition, according to the actual mode Q ij The data Ps of the database of patterns with the wavelength of the interference light obtained in advance as a parameter j Therefore, even when processing is performed in a processing chamber with a different atmosphere or a different optical system, endpoint determination can be performed using the same differential waveform pattern database without reducing accuracy.

[0066] Next, refer to Figure 3 Flowchart describing the Figure 2 The etching amount detection device 8 calculates the etching amount of the processing target film when the etching process is performed. Figure 3 It shows that according to Figure 1A and Figure 1B Flowchart of the operation flow of detecting etching amount of the plasma processing apparatus of the embodiment shown in . Flowchart of the operation flow of the etching amount detecting apparatus 8 is mainly shown.

[0067] In this embodiment, before processing the wafer 4, the target remaining film thickness of the etching target film and the pattern data for detecting or determining the remaining film thickness are set in the differential waveform pattern database (step 301). In the differential waveform pattern database, the interference light pattern (standard pattern) data Ps is used. j The interference light pattern data uses wavelength as a parameter and is obtained when a wafer 4, which is a processing object for manufacturing a semiconductor device, and a test (sample) wafer having a surface film structure identical to that of wafer 4 in material, shape, and structure are etched under the same conditions as wafer 4.

[0068] Next, plasma 3 is formed in processing chamber 2 to begin etching the target film on wafer 4. Interference light emitted from the target film during the etching process is detected at predetermined sampling intervals (e.g., 0.1 to 0.5 seconds) (step 302). At this point, a sampling start command is issued with the start of the etching process. During the etching process, the intensity of the multi-wavelength interference light, which changes with the etching process, is transmitted to spectroscope 9 of etching amount detection device 8 and detected and output by the photodetector of spectroscope 9 as a photodetection signal having a voltage corresponding to the intensity of each predetermined frequency of light.

[0069] The photodetection signal of the spectroscope 9 is converted into a digital signal, and a sampling signal y is obtained. ij Next, the first-stage digital filter circuit 10 makes the multi-wavelength output signal y from the spectroscope 9 ij Smooth and calculate the time series data Y at any time ij (Step 303).

[0070] Next, the time series data Y ij The signal is transmitted to the differentiator 11 and the time series differential coefficient dij is calculated by the Savitzky-Golay method (SG method) (step 304). That is, the coefficient (primary or secondary) d of the signal waveform is detected by the differential processing (SG method). i .

[0071] The differential coefficient dij is transmitted to the second-stage digital filter circuit 12, and the smoothed differential coefficient time series data D is calculated. ij (Step 305). The obtained smoothed differential coefficient time series data Dij Transmitted to the differential waveform comparator 14.

[0072] Furthermore, by dividing the smoothed differential coefficient time series data D in the differential waveform divider 13 ij Divide by the smoothed time series data Y ij , the true mode Q indicating the wavelength dependence of the differential value of the interference light intensity is determined ij (Wavelength is set as a parameter) (Step 306). When the processing chamber 2 for obtaining pattern data of the differential value of the intensity of interference light with wavelength as a parameter using the wafer 4 as a real wafer is different from the processing chamber for obtaining standard pattern data using the test wafer, or when the conditions of the gas supplied to the processing chamber 2 are different even if the processing chamber is the same processing chamber 2, the differential waveform divider 13 is used to divide the smoothed differential coefficient time series data D ij Divide by the smoothed time series data Y ij By performing this step, it is possible to reduce the influence of the difference on the spectrum balance obtained by receiving light from the processing chamber 2 and prevent the loss of detection accuracy of the remaining film thickness and etching amount.

[0073] Here, the real mode Q ij It is achieved by smoothing the differential coefficient time series data D ij Divide by the smoothed time series data Y ij But it can also be obtained by dividing the smoothed time series data y ij In addition, the smoothed time series data y of the differential waveform pattern database can be used ij or Y ij To remove D ij , the differential waveform pattern database is obtained when the chip 4 and the test processing object material whose surface film structure is equivalent to the chip 4 in material, shape and structure are etched under the same conditions as the chip 4, and is obtained in advance.

[0074] In addition, Figure 2 In the example, by using Y ij or y ij To divide the time series data of the differential coefficient value d ij The obtained value can be used as Q ij Here, Y is the divisor. ij or y ijThe differential waveform pattern database is obtained by using a differential waveform pattern database obtained in advance when a test processing object material having a wafer 4 and a surface film structure equivalent to that of the wafer 4 in terms of material, shape, and configuration is etched under conditions equivalent to those of the wafer 4, or is obtained by using luminescence data obtained when light is irradiated into a processing chamber for performing etching processing or a processing chamber having an atmosphere equivalent to that of the processing chamber for performing etching processing. In addition, the value dij calculated using the SG method is used as the dividend here, but any value can be used as long as it is time series data reflecting the difference in the atmosphere in the processing chamber, such as Y ij Itself or against Y ij The value calculated using the least squares method. This also applies to the divider. Furthermore, pre-acquired time series data can be used in the divider.

[0075] In the differential waveform comparator 14, the calculation and calculate the value of The minimum value of (step 307). Calculate and calculate The minimum value of the remaining film thickness corresponding to the instantaneous film thickness data Z at any sampling time (current time) i i (Step 308).

[0076] In the endpoint determination device 17, the remaining film amount of the current processing target film is compared with the preset target remaining film thickness value (set in step 301) to determine whether the endpoint has been reached (step 308). If the thickness is determined to be equal to or less than the target remaining film thickness value, the target is determined to have been reached, and a signal indicating the completion of the etching process is transmitted to the plasma processing apparatus 1. If the target is determined not to have been reached, the process returns to step 303 and the subsequent processes are repeated. The etching amount, such as the etching depth, is determined, and if it is determined that the etching amount is sufficient, the sampling end setting is finally executed (step 309).

[0077] Will refer to Figure 4 An operation flow of a vacuum processing apparatus in the related art will be described as a comparative example. Figure 4 1 is a flow chart showing the operation flow of a vacuum processing device of the prior art. Figure 3 In the flowchart showing the operation flow of the vacuum processing apparatus of this embodiment, step 306 is omitted, and steps 401 to 408 are equivalent to steps 301 to 305 and steps 307 to 309.

[0078] exist Figure 4 The flowchart does not have the time series data D by smoothing the differential coefficient ij Divide by the smoothed time series data Y ijTo calculate the true mode Q indicating the wavelength dependence of the differential value of the interference light intensity ij (Wavelength is set as a parameter) step (step 306). Therefore, when wafer 4, which is an actual wafer, is processed in a process chamber 2 different from the process chamber in which the differential waveform pattern database is obtained, or when wafer 4, which is an actual wafer, is processed in the same process chamber 2 with different supplied gas conditions (such as different types and compositions), the influence of the difference in the conditions in process chamber 2 on the spectral balance cannot be reduced, the accuracy of detecting the target film thickness or the arrival of the process endpoint deteriorates, and the process yield deteriorates.

[0079] Will refer to Figure 5A and Figure 5B To describe the effects obtained by this embodiment. Figure 5A and Figure 5B In this paper, the following two are compared: one is in Figure 1A and Figure 1B The intensity distribution of light of multiple wavelengths (wavelength distribution of light intensity) obtained during the etching process of the processing target material 4 processed in the embodiment shown in , and the other is the wavelength distribution of light emission intensity obtained during the etching process using the technology in the conventional art. Figure 5A and Figure 5B It is shown that according to Figure 1A and Figure 1B Graphs showing wavelength distributions of luminous intensity obtained during wafer etching processes performed by the vacuum processing apparatus of the embodiment shown in FIG. 1 and the prior art.

[0080] Figure 5A and Figure 5B Data obtained during processing of each of two process chambers A and B, each of which has an identical structure, for real wafers 4 is shown. Among the processing conditions for the wafers 4 processed in each process chamber, the two process chambers differ in the processing conditions of process chamber 2 (including at least the composition and flow rate of the supplied gas, and the pressure within the process chamber). The portion of each etching amount detection device 8, which receives light from process chamber 2 and includes spectroscope 9 and optical fiber 7, has substantially the same structure, but has different values ​​depending on the shape, size, and material. These different values ​​include differences within tolerance. Furthermore, within the process chambers, a standard pattern obtained when processing a sample wafer 4 under conditions identical to those of the real wafer 4 in a process chamber 2 including any vacuum processing unit 103 other than vacuum processing apparatus 100, is used to detect the remaining film thickness and etching amount, as shown in step 308.

[0081] Figure 5A is shown when any wafer 4 along as a real wafer Figure 4Graph showing the differential value of the intensity of interference light of each of a plurality of wavelengths from the surface of the wafer 4 obtained from the interior of the processing chamber 2 at any time during the process when the flowchart shown in is processed. The horizontal axis represents the wavelength, and the vertical axis represents the smoothing coefficient time series data D ij On the other hand, Figure 5B is shown when another wafer 4 is used as a real wafer along Figure 3 When the flowchart shown in FIG is processed, a graph of the differential value of the intensity of interference light of each of a plurality of wavelengths from the surface of the wafer 4 obtained from the inside of the processing chamber 2 at any time during the process is obtained. The horizontal axis represents the wavelength, and the vertical axis represents the time series data Y smoothed by dividing ij Divide the time series data by the smoothing coefficient D ij The value Q obtained ij The value of .

[0082] exist Figure 5A It was found that the treatment chambers A and B were in the region of D with a wavelength of 400 nm or less. ij The difference in value is greater than that of D in the region with a wavelength of 400 nm or more. ij On the other hand, Figure 5B The Q of the process chambers A and B in the region of wavelength 400 nm or less is ij The difference in values ​​is suppressed to be smaller than Figure 5A As mentioned above, it is found that by smoothing the differential coefficient time series data Y ij Divide by the smoothed time series data D ij Or normalize it to obtain smoothed time series data Q ij In, prevented Figure 5A The effects of different states of the inner wall surfaces of the processing chambers A and B on the spectral balance are shown in Figure 5B As shown in .

[0083] Therefore, with the configuration of this embodiment, when using a pattern (standard pattern) of differential coefficients of interference light intensity having the same wavelength as a parameter in each processing chamber of different processing units to determine the endpoint of etching processing on wafer 4, which is a real wafer, or to detect the remaining film thickness of a film to be processed, even if the surface condition or processing conditions of the processing chamber in which wafer 4 is to be determined or detected differ from those in the processing chamber in which the standard pattern was obtained, degradation in detection or determination accuracy can be reduced, and processing can be performed with high accuracy. Therefore, even in a vacuum processing apparatus 100 that processes multiple wafers 4, which are real wafers, using multiple processing chambers, a high processing yield can be achieved without having to pre-acquire and use a standard pattern for determination and detection for each processing chamber. Furthermore, the number of sample wafers used to obtain the standard pattern is reduced, and the operating cost of the vacuum processing apparatus 100 can be reduced.

[0084] Furthermore, according to the present invention, since the end point can be accurately determined even if the atmosphere in the processing chamber changes, the yield of the processing target material 4 for manufacturing a semiconductor device as a product can be improved.

[0085] In the above embodiment, the description is given under the assumption that light originating from a film structure including a film to be processed on the surface of wafer 4 disposed in processing chamber 2 and transmitted to beam splitter 9 via optical fiber 7 is emitted from plasma formed in the space above wafer 4 in processing chamber 2, irradiated with and reflected from the film structure. Alternatively, the light may be light having a wavelength within a predetermined range, irradiated onto the surface of wafer 4 from above, from a light source such as a lamp or diode disposed outside the vacuum chamber above processing chamber 2. It is desirable that the wavelengths of light within the predetermined range from such a light source include a plurality of wavelengths, and it is desirable that the vacuum chamber of each vacuum processing unit of vacuum processing apparatus 100 be equipped with a light source having the same structure and specifications, and having the same intensity distribution for the same wavelength (frequency).

[0086] Furthermore, the detection of the etching amount and determination of the processing endpoint are not limited to the detection of the etching amount and determination of the processing endpoint when processing wafers 4, serving as actual wafers, is performed in the processing chamber 2 of each of the multiple vacuum processing units 103. When etching is performed individually, intermittently, and continuously on a batch of wafers 4 in any vacuum processing unit 103, conditions such as the adhesion or distribution of products, their amount, the thickness of deposits on the surfaces of components disposed in the processing chamber 2 and facing the plasma, the surface roughness, and the physical properties of the components' materials naturally vary between the initial processing and subsequent processing within the batch (of wafers 4). Consequently, these differences in surface conditions can alter the distribution or balance of the spectrum of light emitted by the plasma or from the surface of the wafers 4. Thus, even if the spectral balance of the interference light from the wafers 4 obtained during processing changes as the number of wafers 4 processed in the processing chamber 2 increases and the cumulative time during which processing is performed, the accuracy of endpoint determination and etching amount detection can be improved by applying the configuration of the above-described embodiment.

Claims

1. A vacuum processing device comprising: Processing unit, including: a processing chamber disposed in the vacuum container and in which plasma is formed; a sample stage provided below a space in the processing chamber where the plasma is formed, and having a wafer having a film to be processed mounted on an upper surface of the sample stage; a light receiving unit disposed above the process chamber and configured to receive light from the wafer during processing of the wafer using the plasma; and a detector connected to the light receiving unit and configured to detect the thickness of the process object film during the process or the arrival of the end point of the process, wherein the detector is configured to detect the thickness of the film to be processed by comparing a standard pattern in which wavelengths of predetermined values ​​correlated with intensities of light of a plurality of wavelengths relative to thickness variations of the film to be processed, obtained in advance, are used as parameters, with a real pattern being a real pattern of predetermined values ​​correlated with intensities of light of the plurality of wavelengths obtained at a specific time during the process; A predetermined value related to the intensity of light is obtained by dividing the differential coefficient value of the time series data of the intensity of the light of the plurality of wavelengths by the time series data indicating the intensity values ​​of the light of the plurality of wavelengths.

2. The vacuum processing apparatus according to claim 1, wherein A predetermined value related to the intensity of light is obtained by dividing the differential coefficient value of the time series data from which noise is removed by the time series data from which noise is removed from the time series data indicating the intensity values ​​of the light of the plurality of wavelengths.

3. The vacuum processing apparatus according to claim 1, wherein: Each detector of the plurality of processing units detects the thickness of the processing target film by using the same standard pattern.

4. A method for vacuum processing a wafer, the method comprising: mounting a wafer including a film to be processed on an upper surface of a sample stage in a processing chamber; processing the wafer using plasma formed in the processing chamber; receiving light from the film to be processed; and detecting the intensity of the light; the processing chamber being provided in a vacuum container constituting a processing unit; the method comprising: A step of detecting the thickness of the film to be processed or the arrival of the end point of the processing by comparing a real pattern obtained at any time during the processing of the wafer and having predetermined values ​​associated with the intensities of light of a plurality of wavelengths from the film to be processed with a standard pattern, wherein in the standard pattern, wavelengths of predetermined values ​​associated with the intensities of the light of a plurality of wavelengths, which vary with respect to the thickness of the film to be processed, are used as parameters, wherein: A predetermined value related to the intensity of light is obtained by dividing the differential coefficient value of the time series data of the intensity of the light of the plurality of wavelengths by the time series data indicating the intensity values ​​of the light of the plurality of wavelengths.

5. The vacuum processing method according to claim 4, wherein: A predetermined value related to the intensity of light is obtained by dividing the differential coefficient value of the time series data from which noise is removed by the time series data from which noise is removed from the time series data indicating the intensity values ​​of the light of the plurality of wavelengths.

6. The vacuum processing method according to claim 4, wherein: In each of the plurality of processing units, the thickness of the processing target film is detected using the standard pattern.

Images