WO1992015924A1

WO1992015924A1 - Methods for photolithography and development analysis

Info

Publication number: WO1992015924A1
Application number: PCT/US1992/001837
Authority: WO
Inventors: David Carl Debruin; Samuel Patrick Valeriano
Original assignee: Site Services, Inc.
Priority date: 1991-03-04
Filing date: 1992-03-04
Publication date: 1992-09-17
Also published as: AU1557392A

Abstract

An in-line process for determining development rates and other characteristic parameters -- such as layer thickness, contrast, and relative solubility -- during development of PAC-containing photoresist films. The process comprises the steps of: 1) during development, detecting (13) light reflected from films of the photoresist material; 2) generating signals (31) that represent the intensity of the reflected light; 3) fitting sinusoidal functions to the generated signals (33); and 4) using the fitted signals for analyzing (15) characteristic parameters of the photoresist material.

Description

METHODS FOR PHOTOLITHOGRAPHY AND DEVELOPMENT ANALYSIS

BACKGROUND OF THE INVENTION

Field of the Invention:

The present invention relates to lithographic methods for fabricating semiconductor integrated circuits and, more particularly, to methods for monitoring photoresist development rates and related variables.

State of the Art:

In lithographic processes for fabricating semiconductor integrated circuits, thin films of photoresist are formed onto substrates such as silicon wafers. The photoresist materials normally have three components: a resin component, a sensitizer component, and a solvent component. The resin component is the "resistant" portion of the photoresist material. The sensitizer is a photo-active component (PAC) that renders the photoresist material sensitive to light. The solvent component provides viscosity and, normally, is evaporated from the photoresist material after it is dispensed onto a substrate.

PAC materials are either of the positive or negative type. Positive PAC materials, upon exposure to ultraviolet or certain other wavelengths of light, undergo photo-induced chemical reactions called "bleaching." As a result of bleaching, the PAC materials become soluble in polar "developer" solutions.

After bleaching, PAC-containing photoresist materials can be "developed" by a developer solution. The developer solution removes the reaction products of the PAC-containing photoresist materials from the substrate. As a result, the developed areas of the photoresist film define patterns for semiconductor integrated circuits. The line widths in the circuit patterns can be, for instance, as narrow as one micron or less.

Negative PAC materials behave in a manner opposite that of positive PAC materials. Thus, during exposure to certain wavelengths of light, exposed areas of negative PAC materials form polymers. To develop negative PAC materials, developer solutions are selected that dissolve the non-exposed areas.

Batch processes are often used for developing several semiconductor wafers concurrently. Alternatively, in-line or "track" systems can be used for developing individual semiconductor wafers in series. The series development systems are often selected for fabricating high-density integrated circuits with critical dimensions less than »ne micron. In both batch and series systems, development of the exposed PAC materials creates sidewalls of photoresist. Ideally, the sidewalls are perpendicular to the substrate.

In processes for developing PAC-σβntaining photoresist materials, the term "development rate" refers to the rate at which the thickness of a photoresist layer diminishes during developnent. (The thickness of a photoresist layer normally is measured by its height above a substrate surface.) For a given film of PAC-containing photoresist material, the development rate depends upon factors including the material's intrinsic chemical properties and the extent of conversion of the PAC material during exposure.

The time at which a developer solution first renders a given area on a substrate free of photoresist film is called the "clear time" or development "end point." In practice, "over-develop" periods are used for allowing cleared areas to expand laterally after reaching an end point. Thus, the total development time of a PAC-containing photoresist film is the sum of the clear time plus the over-develop time.

Photoresist materials are often characterized by three parameters — Rl, R2 and R3 — or their equivalents. Also, photoresist materials can be characterized by a parameter that is referred to as "contrast" or "gamma." This parameter measures the sensitivity of the PAC-containing photoresist to exposure differences. (Mathematically, gamma values represent the rate of change of a normalized thickness with exposure energy.) Lengthy experiments, sometimes requiring many days to perform, are needed to determine such parameters.

SUMMARY OF THE INVENTION

The present invention provides in-line processes for accurately determining or "extracting" development rates and other characteristic parameters — such as layer thickness, contrast, and relative solubility — during development of PAC-containing photoresist films. These processes are especially adapted for use with "in-line" systems, i.e., systems that develop individual semiconductor wafers in series. In one embodiment, the process according to the present invention comprises the steps of: during development, detecting light reflected from films of the photoresist material; generating signals that represent the intensity of the reflected light; fitting sinusoidal functions to the generated signals; and using the fitted signals for analyzing characteristic parameters of the photoresist material.

BRIEF DESCRIPTION OF THE DRAWINGS

The present invention can be further understood with reference to the following description and the appended drawings, wherein like elements are provided with the same reference numerals. In the drawings:

Figure 1A is a functional block diagram of a system that uses interferometric methods for detecting end points during development of PAC-containing photoresist materials;

Figure IB is an illustration showing interference effects in photoresist films;

Figure 2 is a graph showing the theoretical shape of a signal detected by the system of Figure 1;

Figure 3 is a graph showing an example of the shape of an actual signal detected by the system of Figure 1; Figures 4 is a graph showing curves fitted to the actual signal of Figure 3;

Figure 5 is a graph showing an example of photoresist film thickness plotted as a function of time for a particular sample;

Figure 6 is a graph showing plots of photoresist film thickness versus time for a number of samples;

Figure 7 is a graph showing clear time plotted as a logarithmic function of exposure intensity;

Figure 8 is a graph showing normalized thickness plotted as a logarithmic function of exposure intensity when development periods are of constant duration;

Figure 9 is a graph showing average development rate as a logarithmic function of exposure dosage;

Figure 10 is a plan view of a circular substrate having a plurality of exposure locations;

Figure 11 is a graph showing a waveform detected from the sample areas on the substrate of Figure 10;

Figure 12 is a graph of a set of three variable-frequency sinusoidal curves.

Figure 13 is a graph of thickness plots that, are calculated from the sets of curves in Figure 12; Figure 14A is a graph of the intensity-time relationship for light reflected during development of an unexposed photoresist film;

Figure 14B is a graph of the intensity-time relationship for light reflected from an exposed photoresist film undergoing development under the same conditions as the film in Figure 14A; and

Figure 1 C is a graph resulting from subtraction of the intensity plot of Figure 13A from the intensity plot of Figure 13B.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

In Figure 1A, an end point detection system includes four components: an optical processing head 11, a signal processing unit 13, a central processing unit 15, and an interface system 17. The optical processing head is connected, as by fiber optic strands 18, to send the collected light to the signal processing unit. The interface system 17 connects the signal processing unit to a conventional in-line track system 19 that develops individual semiconductor wafers in series. The components of this end point detection systems are available from SITE Services, Inc. of Santa Clara, California.

Generally speaking, the system in Figure 1 detects the development end points of PAC-containing photoresist films during the fabrication of semiconductor integrated circuits. As so used, the end point detection system receives a sequential series of semiconductor wafers. Each of the wafers has a photoresist film coating for photolithographic processing. As shown in Figure 1A, the optical processing head 11 is mounted for directing collimated light (indicated by the dotted line) onto a photoresist film 21 that coats the surface of a substrate 23. In practice, the substrate 23 can be, for example, a semiconductor wafer. The optical processing head 11 also includes means for collecting light reflected from the photoresist film and substrate.

As further shown in Figure 1A, the signal processing unit 13 includes at least one photodetector array 27 for receiving collected light from optical processing head 11. The output signals from the photodetectors in array 27 indicate the intensity of the collected light. The output signals from the photodetectors in the array are provided to a multiplexer 29 for multiplexing.

As still further shown in Figure 1A, the signal processing unit 13 includes an analog-to-digital (A/D) converter 31 and a microprocessor system 33 that receives digital signals from the A/D convertor. The A/D converter can be of conventional design for converting the multiplexed analog signals to digital signals. In practice, the microprocessor system processes several thousand digital signals per second.

The purpose of processing the digital signals is to build data files. The data files, as will be further explained below, are used for analyzing the characteristic parameters of PAC-containing photoresist films during development. In practice, the parameters can be detected and analyzed during development of a series of individual semiconductor wafers. The operation of the optical processing head 11 can be further understood by referring to Figure IB. As shown, the surface of the photoresist film reflects a portion of the incident light from the optical processing head. Another portion of the incident light passes through the film and is reflected from the underlying substrate. The optical processing head collects both the surface-reflected light and the substrate-reflected light.

Figure 2 shows an example of a signal that, in theory, represents the intensity of reflected light as a function of the development time of a photoresist film on a substrate. The signal has a sinusoidal shape. The locations of the peaks and valleys in the signal vary as the thickness of the photoresist layer diminishes during the development cycle. The maxima are due to constructive interference, while the minima result from destructive interference.

The signal in Figure 2 can be further understood by reference to Figure IB. That drawing illustrates two examples of interference between incident and reflected light relative to development of a photoresist film. In particular, the drawing shows film thicknesses corresponding to one quarter and three quarter wavelengths of the illuminating radiation.

Further in Figure IB, solid dots represent peaks in the incoming and outgoing optical wave amplitudes, while hollow dots represent valleys. For the quarter wavelength thickness, outgoing waves reflected from the top and bottom of the resist film have opposite amplitudes that sum to zero. Since the observed intensity is proportional to the square of the summed amplitudes, the observed reflected intensity will be zero. This cancellation of wave amplitudes repeats at odd multiples of the quarter wave thickness, as shown for the three quarter wavelength film. Successive minima in the observed reflected intensity occur for each half wavelength change in film thickness, permitting indirect measurement of the rate of change of film thickness by counting intensity minima observed during development.

The waveform in Figure 3 is an example of a signal generated by the end point detection system during development of a photoresist film. Normally, a low-pass filter is used to remove noise from the signal. The peaks and valleys in the signal can be explained by phase shifts in the reflected light with diminishing thickness of the photoresist layer. In practice, there may be more than one hundred sample points between each peak and valley of the detected signal.

In Figure 4, a curve has been fitted to data sampled by the end point detection system. In this example, the fitted curve is a variable-frequency sinusoid. As such, the angular frequency of the sine curve varies depending upon the set of data points used for fitting. The variable-frequency sine function can be fitted, for example, by least-squares methods. In practice, every two data points is used for matching.

One reason for fitting variable-frequency sine functions to the data is to exclude noise from the detected signals. Another reason for fitting variable- frequency sine functions to the data is that such functions closely approximate the waveform shapes that are theoretically expected when development of a photoresist layer is monitored. In Figure 5, a thickness plot has been calculated from the information in Figure 4. The plot shows development rates, or development speeds, of the photoresist material. From this curve, it can be seen that the photoresist layer becomes thinner as the development period increases. In this particular example, the film thickness diminishes to zero after approximately seventy-seven seconds — which is the clear time.

Curves such as the one in Figure 5 can be generated for different exposure energies applied to the same photoresist material, with other conditions remaining constant. Data from a family of such curves can be collected to create a data file for characterizing a particular photoresist material.

In Figure 6, normalized thickness plots are shown for a number of samples developed at different exposure dosages. (The normalization reduces the effects of initial differences in thickness of the photoresist films.) The uppermost waveform is approximately horizontal and shows the development of an unexposed area of a photoresist film. The set of curves show that the photoresist layers develop more rapidly when the films have received higher exposure dosages.

In Figure 7, clear time is shown as a function of exposure intensity (dosage) for the data presented in Figure 6. The clear time is expressed in seconds, and exposure intensity is expressed in mJ/cm² on a logarithmic scale. Each box along the curve represents the speed with which a sample actually clears or is projected to clear. The information in Figure 7 indicates that the development rate is slow for exposure intensities ranging up to about 100 mJ/cm². For example, exposure energy of about 50 mJ/cm² results requires a clear time of about 1000 seconds. However, for exposure intensities ranging from about 150 mJ/cm² to about 350 mJ/cm², the clear time changes rapidly. For instance, an exposure energy of 184 mJ/cm² results in a clear time of 118.8 seconds. For exposure dosages exceeding about 500 mJ/cm², the clear time does not diminish significantly — indicating that the PAC material is bleached completely.

The information in Figure 7 can be put to several uses. For example, clear times can be compared for different substrates coated with the same photoresist. Also, clear times can be compared for various substrates that have been exposed at different exposure intensities. Also, the minimum energy exposure for development of photoresist can be determined.

In Figure 8, normalized thickness is shown as a logarithmic function of exposure intensity for a constant development period. Again, exposure intensity is expressed in mJ/cm². For a given development time, the curve shows photoresist film thickness as a function of exposure energy. From this curve, gamma values can be calculated, since those values represent the rate of change of the normalized thickness as a function of exposure energy.

In Figure 9, the average development rate is shown as a logarithmic function of exposure dosage. The development rate is calculated from the data in Figure 6 and is measured in A/second; the exposure dosage is measured in mJ/cm². For this example, the development rate is slow for exposure dosages below about 100 mJ/cm². The term "bulk development rate" is often used to refer to the midpoint of a development rate curve such as shown in Figure 9. For that curve, the bulk development rate is about 414 /second.

Based upon the information in Figure 9, parameters can be calculated that characterize photoresist materials. For example, the information can be used for calculating the minimum exposure energy that is required to initiate development (i.e., the Rl parameter), the fifty-percent exposure point (i.e., the R2 parameter), and the maximum development rate (i.e., the R3 parameter) . Also, the "optimal exposure" or "ETH" values can be calculated. The ETH value usually is defined as the center point of the bulk develop rate.

Also, based upon information in Figure 9, the effect of different development times can be calculated for different samples at constant exposure dosages. This allows the contrast parameter to be determined for different exposures and development processes. From this information, the process conditions can be selected for providing optimal contrast.

With reference to Figures 10 and 11, a process will be described for exposing several different photoresist locations on a single substrate at different exposure intensities (or for different exposure periods) and then monitoring development of the photoresist film at each location. This process, like the one described above, provides information on the photoresist film characteristics. In Figure 10, a disk-like substrate 40 has twelve exposure locations — designated by the letters A through L — arranged in a circular pattern on its surface. According to the preferred embodiment of the process for determining the photoresist film characteristics, the disk-like substrate is rotated under a light source such that the exposure locations are subjected to progressively increasing exposure times (or doses) . For instance, the exposure time at location D exceeds the exposure time at location J.

The development rate of each of the exposure locations on the disk-like substrate in Figure 10 can be monitored by a system including an optical processing head as described above. Then, output signals from the optical processing head can be correlated to tachometer signals for sorting the output information by exposure location. Using such a technique, information that characterizes the properties of the photoresist film under each of the exposure conditions can be calculated for each exposure location.

Figure 11 is a graph showing, for example, collected intensity values at each exposure location for a single revolution of the disk-like substrate of Figure 10. As is expected, the trend of the peak points of the curves indicates that the photoresist thickness decreases with increasing exposure periods. (As is also expected, the peaks reach generally constant height as the clear time is approached for the exposure locations.)

As another example of use of the disk-like substrate of Figure 10, intensity values can be recorded for any selected exposure location on the substrate during multiple rotation of the substrate. The method results in a series of collected values that can provide information equivalent to the information shown in Figures 7-9.

Figure 12 is a graph showing a set of three variable-frequency sinusoidal curves. Each curve has been fitted to data provided by the above-described endpoint detection system. The number of curves in the set is a matter of choice; typically, the set includes at least three curves.

The three curves in Figure 12 differ from one another in that each represents, for a different wavelength of light, the intensity of light reflected from a photoresist film. In this particular example, the upper curve represents light reflected at 960 nm, the middle curve represents light reflected at 830 nm, and the lower curve represents light reflected at 700 nm. As is expected from theory, the upper curve has fewer peaks and valleys than do the lower curves that were detected at shorter wavelengths. Thus, this set of cuirves provides a ready comparison of the reflective response of a photoresist film at different wavelengths.

As shown in Figure 13, thickness plots can be calculated from sets of fitted curves such as shown in Figure 12. More particularly, the drawing shows a thickness plot comprising an overlaid set of three thickness plots based upon a single sample. This method of calculating thickness plots has several advantages. For example, because a user can readily select the best signal (or groups of signals) for use in making a thickness calculation, accuracy can be increased. Also, phase relationships can be easily detected in a set of multiple wavelength curves such as shown in Figure 12. The detected phase relationships can be used for calculating the thickness of a photoresist film that remains on a substrate after removal of only a small amount of the film. In such use, wavelengths are chosen that are normally out of phase when developing thin photoresist films. Then, by comparing phases as the development approaches endpoint, it is possible to accurately predict endpoint as the data is being collected.

In Figure 14A, a "reference" signal represents the intensity-time relationship for light reflected during development of an unexposed photoresist film. In this particular case, the reflected light is monitored from a spinning substrate. Early in the monitoring period, the reflected light intensity decreases rapidly because the photoresist wets the substrate surface. Later, as the substrate stops rotating, the reflected light intensity increases toward a constant value.

In Figure 14B, the intensity-time relationship is shown for light reflected from an exposed photoresist film undergoing development. The photoresist material is the same as that in Figure 1 A, and the development conditions are the same. Thus, this signal contains the same artifacts and noise as the signal in Figure 14A. The peaks and valleys in this signal, as in the case of the signal in Figure 3, represent phase shifts in the reflected light with diminishing thickness of the photoresist layer.

In Figure 14C, the reference signal of Figure 13A has been subtracted from the intensity plot of Figure 13B. The resulting graph represents — even more accurately than Figure 14B ~ the intensity of light reflected from the exposed photoresist film during development. The increased accuracy is due to the effect of the reference signal in canceling artifacts and noise from the signal in Figure 14B.

The data on functional characteristics of photoresist films has numerous uses. For instance, characteristic data for different photoresist batches can be used off-line for quality control and quality assurance purposes. Also, the detected functional characteristics can be used for real-time control of in-line development systems. Such control systems could assure uniform development of photoresist films from different photoresist batches.

The foregoing has described the principles, preferred embodiments and modes of operation of the present invention. However, the invention should not be construed as limited to the particular embodiments discussed. For example, although the foregoing described methods for monitoring characteristics of positive PAC-containing materials, the methods can be adapted for use with negative PAC-containing materials.

Thus, the above-described embodiments should be regarded as illustrative rather than restrictive. Variations may be made in those embodiments without departing from the scope of present invention as defined by the following claims.

Claims

WHAT IS CLAIMED IS:

1. A process for determining development rates and other characteristic parameters of PAC- containing photoresist materials during development, comprising: detecting the intensity of light reflected from films of PAC-containing photoresist material; generating signals that represent the intensity of the reflected light; fitting sinusoidal smoothing functions to the generated signals, thereby providing smoothed signals; and using values of the smoothed signals for analyzing characteristic parameters of the photoresist material.

2. A method according to claim 1 wherein the characteristic parameters include layer thickness and contrast.

3. A method according to claim 1, wherein the step of using values of the smoothed signals for analyzing characteristic parameters of the photoresist material includes the steps of: detecting maxima and minima values of the smoothed signals; and from the detected maxima and minima values, calculating normalized thickness values of the photoresist films as a function of development time.

4. A method according to claim 3 -including the step of graphically displaying normalized thickness values versus development time.

5. A method according to claim 1 wherein the step of using values of the smoothed signals for analyzing characteristic parameters of the photoresist material includes the steps of: determining the clear time for samples of the photoresist film; and determining clear time as a function of the exposure energy that was used to develop the samples of the photoresist film.

6. A method according to claim 5 further including graphically displaying the clear time versus exposure energy for each of the samples.

7. A method according to claim 5 including the step of determining the bulk development rate for the photoresist material.

8. A method according to claim 1 wherein the smoothed signals are generated by exposing samples of the photoresist film at various exposure energies.

9. A method according to claim 8 wherein the step of using values of the smoothed signals for analyzing characteristic parameters of the photoresist material includes the steps of: converting maxima and minima values of the smoothed signals to normalized thickness values of the photoresist film; and determining normalized thickness as a function of exposure energy.

10. A method according to claim 8, further including determining the contrast characteristic of the photoresist film.

11. A method according to claim 1, further including: exposing samples of the photoresist film to different exposure energies; and determining the development rate of the photoresist material as a function of the exposure energies, thereby characterizing the photoresist material.

12. A method according to claim 1 wherein the step of using values of the smoothed signals for analyzing characteristic parameters of the photoresist material includes the steps of: determining a minimum exposure energy required to initiate development of the photoresist, a maximum exposure energy, and an exposure energy mid¬ point between the minimum and maximum exposure energies.

13. A method according to claim 1, wherein a number of samples of the photoresist film are analyzed on a single substrate.