Scanning electron microscope system and method of manufacturing an integrated circuit

[0002] The present invention relates generally to scanning electron microscopes and, more particularly, to a scanning electronic microscope processor for analyzing semiconductor devices and a method of manufacturing an integrated circuit.

[0003] A large amount of activity in the microelectronics industry is directed toward developing methodologies for testing wafers during the manufacturing process. Typically, scanning electron microscopes (SEM) have been used in semiconductor manufacturing processes. Although the SEM is useful for providing some information regarding a semiconductor device that is scanned, it can be difficult for an operator to detect errors by simply viewing the waveform from the SEM. Usually, an operator operates a SEM in an automatic mode to measure critical line width. The measurement is derived from the intensity waveform produced by the SEM. The SEM is not utilized to extract processing variables from the topography of the wafer. The shape of the measured object is typically not considered.

[0004] In addition, cross-sections of wafers are performed to assess the quality of a wafer. Although this process may be useful, the process is destructive and time consuming. As a result, a small number of wafers are selected for testing. An alternative approach is to perform electrical tests at the end of wafer processing to determine if deficiencies exist and assess the quality of the semiconductor devices formed on the wafer. Although deficiencies may be found, the defects are not detected until the end of wafer processing. As a result, problems in the manufacturing process or with the equipment used during manufacturing may not be detected. Numerous defective wafers may be produced before problems with the manufacturing or measurement process can be corrected.

[0005] The present invention provides a method and system for analyzing the intensity profiles of a wafer including the step of scanning the wafer to produce a scanned signal. The scanned signal is preprocessed and compared to a reference signal to assess the wafer. The present invention is also directed to a method of manufacturing a wafer using the above method and system.

[0006] It is to be understood that both the foregoing general description and the following detailed description are exemplary, but are not restrictive, of the invention.

[0007] The invention is best understood from the following detailed description when read in connection with the accompanying drawing. It is emphasized that, according to common practice in the semiconductor industry, the various features of the drawing are not to scale. On the contrary, the dimensions of the various features are arbitrarily expanded or reduced for clarity. Included in the drawing are the following figures:

[0008] FIG. 1 is a block diagram of scanning electron microscope system 100 according to the present invention;

[0009] FIGS. 2-5 are flow chart diagrams for illustrating the operation of the scanning electron microscope system 100 shown in FIG. 1;

[0010] FIGS. 6-22 are graphs illustrating the operation of the scanning electron microscope system 100; and

[0011] FIG. 23 is a SEM picture of a metal line on a wafer illustrating an alternative embodiment.

[0012] The present invention is directed to a scanning electron microscope system that measures and analyzes the surface topology of a wafer during the manufacturing process. In other words, the present invention analyzes the shape of the scanned features. The surface topology is measured using a scanning electron microscope which produces intensity values in an array, an intensity profile. The intensity profile may then be processed using signal processing techniques and compared to standard waveforms from, for example, a standardized wafer which has been processed in the same manner. The standard waveforms are waveforms representing known shapes of scanned features. Differences and errors in the wafer are manifested in the measured topology and may be detected by the comparison. Errors in the measurement, i.e. focus and sharpness, by the SEM may also be determined. If the measured intensity topology can not be matched to the standard waveforms, it may be stored for further analysis. As a result, process monitoring may be improved and fatal errors in the wafers may be detected before further processing. Further, the above process may be implemented without adding additional steps to the manufacturing process because scanning electron microscopes are already used during processing.

[0013] Referring now to the drawing, wherein like reference numerals refer to like elements throughout, FIG. 1 is a scanning electron microscope (SEM) system 100. The SEM system 100 includes a scanning electron microscope (SEM) 150 for scanning a wafer 120 or a substrate having a surface feature and producing a wafer waveform signal y_i(t). An exemplary wafer waveform signal y_i(t) for a resist line is shown in FIG. 6. The SEM 150 is, for example, Model 8820 available from Hitachi, 3100 North 1^stStreet, San Jose, Calif. 95134 USA. The SEM 150 is coupled to a processor 110.

[0014] The processor 110 receives the wafer waveform signal y_i(t) and detects errors and deficiencies in the wafer 120 by analyzing the wafer waveform signal y_i(t). In addition, the processor 110 may detect deviations in the manufacturing process such as variations between tools. This process may be performed in-line during the manufacturing process. “In-line during the manufacturing process” means during the process of forming circuitry on the wafer 120. Consequently, process errors and degraded quality in, for example, the lithography and etching processes may be detected before manufacture of the devices is completed. Measurement errors, such as charging, may also be detected. Measurement errors can cause unnecessary reworks or scrap.

[0015] In this way, adjustments may be made in the manufacturing line to correct, for example, tool drift and tool-to-tool matching for SEMs, steppers, and etchers. This allows problems such as SEM charging, stepper out of focus, and over etch errors to be detected and corrected. Further, defective wafers may be detected and removed prior to further processing. In addition, wafer characterization may be performed to determine profile degradation across a wafer. As a result, the cost of the manufacturing process may be decreased while increasing the quality of the wafers 120 produced.

[0016] The SEM system 100 also includes a database 125 for storing reference data. The SEM system 100 may also include a tool or tools 130 that may be automatically or manually adjusted in response to the analysis performed by the SEM system 100. The components shown in FIG. 1 may be combined into one or more components and may be implemented in hardware or software. The operation of SEM system 100 is described below with reference to FIGS. 2-5. In addition, the process shown in FIGS. 2-5 is illustrated using the example shown in FIGS. 6-22.

[0017] At step 200, shown in FIG. 2, the SEM 150 acquires the wafer waveform signal y_i(t). An exemplary wafer waveform signal y_i(t) for a charged resist line is shown in FIG. 7. FIG. 8 is a diagram corresponding to a portion of the waveform shown in FIG. 7. At step 205, the processor 110 generates a processed waveform signal p₂₂(t) from the wafer waveform signal y_i(t) by implementing an auto-correlation operation.

[0018] Step 205 is described in greater detail below with reference to FIG. 3. At step 300, the processor 110 processes the wafer waveform signal y_i(t) using, for example, equation (1) below to produce a converted waveform signal Y_i(jw).

[0019] F[y_i(t)]=Y_i(jw)  (1)

[0020] F[ ] denotes any Fourier Transform. Equation (1) implements, for example, a Fast Fourier Transform (FFT). FIG. 9 is an exemplary converted waveform Y_i(jw) for the wafer waveform signal y_i(t) shown in FIG. 8. At step 305 the converted waveform signal Y_i(jw) is filtered using a low pass filter as is shown in equation (2) below to produce a filtered waveform signal Y_f(jw).

[0021] Φ[Y_i(jw)]=Y_f(jw)  (2)

[0022] For example, the filter implemented by equation (2) may pass a quarter (¼) or less of the components of the converted waveform signal Y_i(jw). The high frequency components are removed to reduce the systematic noise in the wafer waveform signal y_i(t). Other filters may be used. FIG. 10 is the filtered waveform signal Y_f(jw) corresponding to the converted waveform signal Y_i(jw) shown in FIG. 9.

[0023] At step 310 an auto-correlation operation is performed to produce a wafer auto-correlation signal R₂₂(jw) from the filtered waveform signal Y_f(jw) using equation (3) below.

[0024] R₂₂(jw)=Y_f(jw)Y_f*(jw)  (3)

[0025] In equation (3), the “*” indicates a complex conjugate. An exemplary wafer auto-correlation signal R₂₂(jw) is shown in FIG. 11. At step 315, an inverse transform is performed to produce a transformed signal r₂₂(t) using equation (4) below. $\begin{array}{cc}{r}_{22}\left(t\right)=\frac{F^{-1}\left[R_{22}\left(jw\right)\right]}{N}& \left(4\right)\end{array}$

[0026] In equation (4), N is the total number of pixels that would be used for display_ing the waveform or the total number of input quantities (samples). For example, an exemplary transformed signal r₂₂(t) is shown in FIG. 12.

[0027] At step 320 the maximum value MAX₂₂of the transformed signal r₂₂(t) is determined at a phase or lag equal to zero (0) as is shown in equation (5) below.

[0028] MAX₂₂=r₂₂(t)(t=0)  (5)

[0029] At step 325, the transformed signal r₂₂(t) is normalized according to equation (6) below to produce the auto-correlation waveform signal p₂₂(t). $\begin{array}{cc}{P}_{22}\left(t\right)=\frac{r_{22}\left(t\right)}{{\mathrm{Max}}_{22}}& \left(6\right)\end{array}$

[0030] An exemplary auto-correlation signal p₂₂(t) is shown in FIG. 13 where the transformed signal r₂₂(t) is normalized. The result is a function of probability densities for a lag of t. This is shown for completeness, but is not necessary for further processing. The maximum value MAX₂₂is, for example, 332.164.

[0031] Returning to FIG. 2, at step 210 the wafer waveform signal y_i(t) is compared to a standard waveform signal x_i(t) using the auto-correlation signal p₂₂(t). The standard waveform signal x_i(t) is used as a bench mark to determine whether other wafers 120 have deficiencies and satisfy quality standards and to detect variations in the manufacturing process. The standard waveform signal x_i(t) is derived from a scan of a standard wafer (not shown). The standard wafer is a wafer that satisfies the desired manufacturing criteria for producing the wafer 120. In other words, the wafer 120 is acceptable if the wafer 120 is within a specified range of the standard wafer. The process for deriving the standard auto-correlation signal p₁₁is the same as the process for producing the wafer auto-correlation signal p₂₂except the process is performed on a scanned signal rather than standard wafer. The process for deriving the standard auto-correlation signal p₁₁is shown in FIG. 4. A description of FIG. 4 is omitted for the sake of brevity.

[0032] FIGS. 14-19 are exemplary waveforms corresponding to the process steps shown in FIG. 4. FIG. 14 is an exemplary standard waveform signal x_i(t) for a resist line on the standard wafer. The resist line is not charged. FIG. 15 is an exemplary converted waveform X_i(jw) for the wafer waveform signal x_i(t) shown in FIG. 14. FIG. 16 is the filtered waveform signal X_f(jw) corresponding to the converted waveform signal X_i(jw) shown in FIG. 15. FIG. 17 is the wafer auto-correlation signal R₁₁(jw) for the filtered waveform signal X_f(jw) shown in FIG. 16. The transformed signal r₁₁(t) of the auto-correlation signal R₁₁(jw) is shown in FIG. 18. Finally, FIG. 19 shows the standard auto-correlation waveform signal p₁₁(t) corresponding to the transformed signal r₁₁(t) shown in FIG. 18. The maximum value MAX₁₁used to produce the standard waveform signal p₁₁(t) shown in FIG. 12 is 551.405.

[0033] Returning to FIG. 2, at step 215, the processor 110 determines whether the comparison waveform signal p₁₂is within a predetermined range. Step 215 is described in greater detail in FIG. 5. At step 500, a cross-correlation signal R₁₂(jw) is generated from the wafer converted waveform signal y_f(jw) and the standard converted waveform signal x_f(jw), calculated in earlier steps, using equation (6) below.

[0034] R₁₂=X_f(jw)Y_f*(jw)  (6)

[0035] “*” indicates the complex conjugate. FIG. 20 is an exemplary cross-correlation signal R₁₂(jw) of the processed waveform signal p₂₂and the standard waveform signal p₁₁.

[0036] At step 505, the cross-correlation signal R₁₂(jw) is converted to the time domain using equation (7) below to produce the unnormalized cross-correlation signal r₁₂(t). $\begin{array}{cc}{r}_{12}\left(t\right)=\frac{F^{-1}\left[R_{12}\left(jw\right)\right]}{N}& \left(7\right)\end{array}$

[0037] N is the same value for both signals. FIG. 21 is the unnormalized cross-correlation signal r₁₂(t) corresponding to the cross-correlation signal R₁₂(jw) shown in FIG. 20.

[0038] At step 510, the unnormalized cross-correlation signal r₁₂(t) is normalized according to equation (8) below. $\begin{array}{cc}{p}_{12}\left(t\right)=\frac{r_{12}\left(t\right)}{{\mathrm{Max}}_{\mathrm{func}}}& \left(8\right)\end{array}$

[0039] The value Max_funcis defined in equation (9) below. $\begin{array}{cc}{\mathrm{Max}}_{\mathrm{func}}=\{\begin{array}{cc}{\mathrm{Max}}_{11}& \mathrm{if}{\mathrm{Max}}_{22}<{\mathrm{Max}}_{11}\mathrm{and}\mathrm{to}\\ & \mathrm{compare}\mathrm{shape}\mathrm{and}\mathrm{amplitude}\\ {\mathrm{Max}}_{22}& \mathrm{if}{\mathrm{Max}}_{11}<{\mathrm{Max}}_{22}\mathrm{and}\mathrm{to}\\ & \mathrm{compare}\mathrm{shape}\mathrm{and}\mathrm{amplitude}\\ \mathrm{or}& \\ \sqrt{{\mathrm{Max}}_{22}·{\mathrm{Max}}_{11}}& \mathrm{if}\mathrm{only}\mathrm{to}\mathrm{compare}\mathrm{shape}\end{array}& \left(9\right)\end{array}$

[0040] FIG. 22 shows the normalized signal p₁₂(t) normalized using a Max_funcof 551.405 (Max₁₁). Amplitude is not considered in this case. The shape of the waveforms are only compared if there are scaling errors present, such as those caused by a degradation in the SEM 150 or when the SEM 150 is not matched to the SEM that measured the standardized wafer. At step 515, the maximum value p_maxof the comparison waveform signal p₁₂(t) is determined. The maximum value p_maxfor the normalized signal p₁₂(t) shown in FIG. 22 is 0.767. Phase errors (picture offsets) which may occur in the SEM 150 are eliminated. As a result, the wafer waveform signal y_i(t) may later be compared to a standard waveform signal x_i(t) without this possible source of error.

[0041] At step 520, the processor 110 determines whether the comparison waveform signal p₁₂(t) is within specification. For example, if the absolute value of the maximum value p_maxis greater than 0.9 and less than 1 (0.90<) |p_max|<1), the wafer 120 is considered acceptable. In a production line arrangement, one or more wafers 120 may be tested to determine if an entire lot is acceptable. Otherwise, the lot may be rejected. Typically, acceptable wafers 120 have been found to have a maximum value p_maxbetween 0.95 and 1 (0.95<|p_max|<1).

[0042] Returning to FIG. 2, at step 220, the lot or tools 130 are indicated as passing if the maximum value p_maxis within the specified range. Otherwise, at step 230, the wafer waveform signal y_i(t) is compared to data stored in database 125 using the same methods described earlier to determine what error has occurred and how the error may be corrected. The database 125 includes data and/or instructions for modifying the production process to eliminate the errors. At step 235, the processor 110 provides instructions to the tools 130, equipment, etc. using the data from the database 125 to correct the errors. In addition, the lot of wafers is disposed of is appropriate. Alternatively, the information may be provided to an operator via a user interface (not shown). The operator makes adjustments to the manufacturing process or measurement in response to the information.

[0043] At step 240, if there is no corresponding instructions for correcting the error, the error and the associated data are stored in the database 125 for future analysis and comparison. For example, the wafer waveform signal y_i(t), the comparison waveform signal p₁₂, and/or any of the other signals produced or used during the analysis of the wafer 120 may be stored.

[0044] Although the above exemplary embodiment utilized a one dimensional waveform, a multi-dimensional process may be used or multiple waveforms may be compared. In addition, regions of the intensity topology may be averaged, summed, or in general signal processed, and compared. Further, the orientation of a feature may be analyzed. In this case, the wafer may be rejected if the feature does not have the proper orientation. The particular comparison that is chosen is dependent upon the particular feature to be analyzed.

[0045] For example, consider FIG. 23 which is a picture of a metal line 710 formed on a substrate. One process for determining the roughness of the edge 700 of the line 710 is to compare multiple average intensities in the y-direction. An average intensity is derived by averaging intensity values of a segment 720 extending in the x-direction. If the edge 700 varies, i.e. has bulges or inflections, the average intensity varies. Variations in the edge 700 may be detected and compared to an expected or standard edge profile. The average intensities may be considered to form a waveform signal y_i(t) extending in the y-direction. This waveform signal y_i(t) may be processed using the process shown in FIG. 2.

[0046] The above invention is not limited to comparing the intensity profile in the x or y-directions or in straight lines. The selected comparison may include comparisons which traverse the wafer in a manner suitable to assess the particular feature. For example, the pattern may be selected to correspond to the edge of a feature which is not solely oriented in the x and y-directions.

[0047] Although the invention has been described with reference to exemplary embodiments, it is not limited to those embodiments. Rather, the appended claims should be construed to include other variants and embodiments of the invention which may be made by those skilled in the art without departing from the true spirit and scope of the present invention.