KR20060050931A - Enhanced process and profile simulator algorithms - Google Patents

Abstract

Process and profile simulator algorithm enhancement methods predict the surface profile on which a given plasma treatment is performed. Active particles are first tracked. The ion flux produced by the active particles is then recorded. Local etch rates and local deposition rates are calculated from neutral flux, chemical coverage of the surface, and surface material forms, which are calculated simultaneously.

Simulator, Plasma Treatment, Neutral Flux

Description

Enhanced Process and Profile Simulator Algorithm {ENHANCED PROCESS AND PROFILE SIMULATOR ALGORITHMS}

The accompanying drawings, which are incorporated in and constitute a part of this application, illustrate one or more embodiments of the invention, and together with the description serve to explain the principles and practice of the invention.

1 is a block diagram showing a conventional plasma etching system.

2 is a block diagram illustrating a plasma etching system in accordance with an embodiment of the present invention.

3 is a system block diagram illustrating a computer according to an embodiment of the present invention.

4 is a flowchart illustrating a method of calculating material morphology, chemical coverage, neutral and ion flux for surface segments in a feature profile to calculate etch rates along a feature profile in accordance with one embodiment of the present invention.

FIG. 5 is a flowchart illustrating a method of simultaneously calculating the amount of bonding comprising chemical coverage of the surface, incident neutral flux, and surface material form in each segment along a feature profile in accordance with an embodiment of the present invention. FIG.

6 is a flowchart illustrating a method of calculating neutral flux for a surface segment in accordance with one embodiment of the present invention.

Fig. 7 is a cross sectional diagram showing surface kinematics upon plasma treatment according to the Langmuir model.

<Explanation of symbols for the main parts of the drawings>

206: TCP controller

216 TCP coil

218: RF transparent window

300: system bus

304: main system memory

The present invention relates to plasma processing of semiconductor devices. In particular, the present invention provides a method for improving the process and a profile simulator algorithm for predicting the surface profile that a given plasma process will produce.

Various forms of treatment using ionizing gases, such as plasma etching and reactive ion etching, are becoming particularly important in the field of semiconductor device manufacturing. In particular, the apparatus used for the etching process is particularly important. 1 illustrates a conventional inductively coupled plasma etch system 100 that may be used for the processing and fabrication of semiconductor devices. Inductively coupled plasma processing system 100 includes a plasma reactor 102 having a plasma chamber 104 therein. Transformer coupled power (TCP) controller 106 and bias power controller 108 control TCP power supply 110 and bias power supply 112 that affect the plasma generated inside plasma chamber 104, respectively.

The TCP power controller 106 is configured to supply a radio frequency (RF) signal adjusted by the TCP match network 114 to the TCP coil 116 located near the plasma chamber 104. An RF transmission window 118 is typically provided to allow energy to flow from the TCP coil 116 into the plasma chamber 104 while at the same time separating the TCP coil 116 from the plasma chamber 104.

The bias power controller 108 provides an RF signal adjusted by the bias match network 120 to an electrode 122 provided inside the plasma reactor 104 to accommodate a substrate 124 such as a semiconductor wafer. And generate a direct current (DC) bias above.

A gas supply mechanism 126, such as a pendulum control valve, typically provides the interior of the plasma reactor 104 with the appropriate chemistry required for the manufacturing process. The gas exhaust mechanism 128 removes particles from the interior of the plasma chamber 104 and maintains the interior of the plasma chamber 104 at a certain pressure. The pressure controller 130 controls both the gas supply mechanism 126 and the gas discharge mechanism 128.

The temperature controller 134 uses a heater 136, such as a heating cartridge, around the plasma chamber 104 to control the temperature of the plasma chamber 104 to a selected temperature set point.

In the plasma chamber 104, substrate etching is accomplished by exposing the substrate 104 to an ionizing gas compound (plasma) under vacuum. The etching process is initiated when gas is transported into the plasma chamber 104. RF power delivered by the TCP coil 116 and regulated by the TCP match network 110 ionizes the gases. The RF power delivered by the electrode 122 and adjusted by the bias match network 120 induces a DC bias on the substrate 124 to control the direction and energy of the ion bombardment of the substrate 124. During the etching process, the plasma reacts chemically with the surface of the substrate 124 to remove material not covered by the photoresist mask.

Input parameters, such as plasma reactor setup, are of fundamental importance for plasma processing. The actual amount of TCP power, bias power, gas pressure, gas temperature, and gas flow in the plasma chamber 104 greatly affect the process conditions. Severe fluctuations in the actual power delivered to the plasma chamber 104 may unexpectedly change the expected values of other process variable parameters such as neutral and ionized particle density, temperature, and etch rate.

Traditionally, the values of these input variables suitable for generating certain settings for device characteristics have been determined by trial and error methods. The development of a single process by this empirical approach is expensive and time consuming and requires continuous study of the final profile by processing several pattern wafers and electron microscopy. Small changes in one input variable may affect the profile due to an unexpected manner, such that any design from one application to another, for example, a change in device dimensions, pattern density on the wafer, or the total open area Has often required the redevelopment of processes involving the consumption of additional resources.

Recent advances in device manufacturing technology have made this approach even more cumbersome. Dimensional reduction of features requires tighter tolerances in feature dimensions and shapes, increasing the number of experiments required to optimize certain processes. Full redesign of the process, accompanied by significant changes in wafer diameter growth and diameter, has increased the number of times this empirical process has to be repeated. Increasing the use of devices manufactured exclusively for a particular application also increases the amount of development and optimization work required.

An alternative computerized approach atomically determines the final etch or deposition rate along the wafer surface from the macro-flow rate and the plasma model to account for the concentration and energy distribution of the various species in the plasma, the coupling between the macro-input variables and the macro-flow rate. And derive an input variable from a complete physical description of the plasma process including a profile simulator to calculate profile progress therefrom. Ideally, this physical description of the plasma etch and deposition process allows for the theoretical selection of the macroscopic input parameters appropriate to produce the desired profile on the substrate, eliminating the need for expensive time consuming experimental iterations.

Research in these fields contributed to the laws of scale that could explain the mechanisms in the work of plasma processes and form the physical explanations. However, despite the usefulness of calculation means that are sufficiently effective to carry out the necessary calculations based on known law of scale, the execution of this theoretical approach has been limited by the lack of data. For example, it is not yet known how some coefficient values in these laws depend on a particular given process. In some investigations, the determination of these scale coefficient values that make up a plasma process defined by a predetermined set of input variables may be achieved by applying a simulated profile that includes one or more of these coefficients as adjustable variables to the final profile produced by applying such a process. By comparison. This later evaluation may help to understand the role of certain coefficients in scale law, but for any process defined by input variable settings that differ from the settings used in the experimental process used to derive these coefficient values. It did not provide the ability to anticipate profile progress against the profile.

Thus, what is needed is a method of improving process algorithms and profile simulators to accurately predict the surface profile that a given plasma process will produce.

The present invention is a process for predicting the surface profile under which a given plasma treatment is performed and a method for improving the profile simulator algorithm. The method first tracks active particles and records the ion flux produced by the active particles. Local etch rates and local deposition rates are calculated from neutral flux, chemical coverage of the surface, and surface material forms, which are calculated simultaneously.

Neutral flux, chemical coverage of the surface, and surface material form are calculated simultaneously by first computing the neutral flux. The position balance equation is first calculated using neutral flux without self-contradiction. The etch rate of the deposited film is then calculated using the solution of the position balance equation. Based on the etch rate of the deposited film, the surface material form is adjusted to the deposited film or underlining material form. The method repeats these steps until neutral flux, chemical coverage of the surface, and surface material forms do not contradict each other themselves.

Neutral flux can be estimated analytically with greater accuracy and improved computational efficiency by calculating direct flux distribution, re-release flux, and permeation matrix for each neutral species. The previous calculation is repeated for all neutral species.

Simulation of the plasma etch or deposition process is enhanced by extrapolating the depiction of the feature profile using Cartesian coordinates having three to three degrees of freedom and three degrees of freedom for particle velocity and cylindrical coordinates symmetrically azimuthally. May be

Embodiments of the invention are described herein in connection with a profile simulator algorithm. Those skilled in the art will appreciate that the following detailed description of the invention is illustrative only and is not intended to be limiting in any way. Other embodiments of the present invention may enable those skilled in the art to readily appreciate the advantages of the present disclosure. As shown in the accompanying drawings, reference numerals are given in detail to practical examples of the present invention. The same reference numerals have been used in the drawings and the description of what are considered identical or similar parts.

For clarity, not all typical features of the actual examples described herein are shown and described. Of course, in the development of such practical embodiments, it can be seen that a number of implementation-specific decisions must be made in order to achieve the developer's specific objectives, such as compliance with application and business-related limitations, and such specific objectives may not You can see that you can change to another implementation and change the developer to another developer. In addition, while such development efforts are complex and time consuming, it will nevertheless be appreciated that they may be routine engineering tasks for those skilled in the art to understand the advantages of the present application.

According to one embodiment of the invention, the components, processing steps and / or data structures may comprise various operating systems (OSs), computing flatforms, firmware, computer programs, computer languages, and / or utility machines. Can be performed using. The method can be driven like a programmed process running on a processing circuit. The processing circuit can have a number of combined forms of processor and operating system, or a single device. The process may be performed by instructions performed by such hardware, hardware alone, or any combination thereof. The software may be stored in a program storage device readable by the machine.

In addition, those skilled in the art will appreciate that hardwired devices, field programmable gate arrays (FPGA) and complex programmable logic devices may be made without departing from the spirit and scope of the inventive concepts detailed herein. It will be appreciated that devices that do not have versatile characteristics may be used, such as field programmable logic devices (FPLD), including application devices (CPLD), application specific integrated circuits (ASIC), and the like. Can be.

According to one embodiment of the present invention, the method is available from Sun Corporation, Sun Microsystems, Inc., Palo Alto, Calif., Microsoft Corporation, Microsoft Corporation, Redmond, Wash. A personal computer running an operating system such as Windows® XP and Windows® 2000, or several versions of a Unix-powered system, such as Linux, available from a number of vendors, It may be performed on a data processing computer such as a workstation computer, mainframe computer, or a high performance server. The method may be performed in a multi-processor system or computing environment including various peripherals such as input devices, output devices, displays, printing devices, memory, storage devices, media interfaces for inputting and outputting data to and from the processor, and the like. ) May be performed. In addition, such a computer system or computing environment may be networked locally or networked via the Internet.

The semi-empirical approach of the profile or process simulator is not limited to any particular macroscopic physical description of the surface profile or the underlying plasma model. Mathematical models can incorporate scaling laws, numerical simulations, or experiments derived from basic plasma physics and chemistry. In general, the plasma model is configured to regress the flux, energy, and angular distributions that reach the substrate surface for the relevant species based on the macro variables related to the function of the reactor. The plasma model may comprise a complete plasma simulation tool with all known cross sections for gas phase chemistry and plasma surface interaction, or simply look-up table of empirically measured fluxes and observed trends. ) May be included.

The profile simulator models the evolution of the feature to be processed by preferably calculating the motion of each segment along the discretized profile surface for accuracy and flexibility. Appropriate profile simulators include local transport models that calculate local fluxes reaching each point along the substrate surface, position balances that calculate the chemical coverage of the surface (enrichment of reactive, inhibitory and / or deposited species at each point on the substrate surface). Models, rate models for calculating the resulting local etch and deposition rates, and surface evolution algorithms that convert these mechanisms into pure surface motion. In a preferred embodiment, the profile simulator derives local fluxes from the macroscopic variables provided by the plasma model using Monte Carlo methods and / or analytical methods. The ratio model and the position balance model are preferably based on the Langmuir-type model, which is the kinetics of particle-surface interactions, but exhibits the interaction between the particles and the surface mixed layer (which can be multiple single layer thicknesses). Can be generalized. The model is characterized by a species that removes material from the substrate like an etchant, depositing the material onto the substrate surface like an evaporator, and suppressing the surface reaction like an inhibitor.

Surface evolution algorithms include a shock-front-tracking approach. In general, each of these four segments contributes to the coefficient of unknown value in the profile simulator characteristic link between the processing medium and the behavior of the substrate surface, which collectively describes substrate evolution.

According to one embodiment, the plasma model characterizes the relevant species only when belonging to a general class of reactors, for example charged particles (and thus attracted to the substrate by an applied bias), or neutral species such as molecules and excited radicals. do. Physical models based on functional dependence in terms governing the empirical input variables of plasma description, such as particle flux, energy and angular distribution, such as the Maxwell and Boltzmann equations, are well known in the art (eg For example, see Liverman and Liechtenberg, Plasma Emission and Material Processing, John Willy, 1994). Based on this basic physical model with respect to empirical data, it is well known in the art how this description is scaled with the input parameters. However, the absolute value of this flux or distribution is not yet known in a given process.

Accordingly, the present invention provides an algorithm that accurately predicts the kinetics of a plasma etching system as shown in FIG. 2, particularly the interaction of energy particles impacting a semiconductor substrate such as a wafer. The method uses a model (local transmission model, position balance model, ratio model, surface evolution algorithm), which combines this model with several computing processes, where solutions are solved simultaneously and consistently.

2 shows an inductively coupled plasma etching system 200 according to one embodiment of the invention. Inductively coupled plasma processing system 200 includes a plasma reactor 202 having a plasma chamber 204. A transformer coupled power (TCP) controller 206 and a bias power controller 208 are TCP power supplies 210 and bias power supplies 212 that affect the plasma generated within the plasma chamber 204. Control each.

The TCP power controller 206 is a set point for the TCP power supply 210 configured to supply a high frequency (RF) signal tuned by the TCP matching network 214 to a TCP coil 216 located near the plasma chamber 204. Set. An RF transmissive window 218 is provided to separate the TCP coil 216 from the plasma chamber 104 while allowing energy to pass from the TCP coil 216 into the plasma chamber 204.

The bias power controller 208 sets a set point for the bias power supply 212 configured to supply an RF signal tuned by the bias matching network 220 to an electrode 222 located within the plasma reactor 204, The plasma reactor generates a direct current (DC) bias over the electrode 222, on which the substrate 224, such as a semiconductor wafer, to be processed is placed.

Gas supply devices 226, such as pendulum control valves, generally supply plasma reactor 204 with the appropriate chemicals needed for the manufacturing process. The gas exhaust mechanism 228 removes particles from within the plasma chamber 204 to maintain a certain pressure within the plasma chamber 204. The pressure controller 230 controls both the gas splice mechanism 226 and the gas discharge mechanism 228.

The temperature controller 234 uses a heater 236, such as a heating cartridge around the plasma chamber 204, to control the temperature of the plasma chamber 204 to a selected temperature set point.

In the plasma chamber 204, substrate etching is accomplished by exposing the substrate 204 to an ionized gas compound (plasma) under vacuum. The etching process begins when gas is delivered into the plasma chamber 204. The high frequency power delivered by the TCP coil 216 and tuned by the TCP matching network 210 ionizes the gas. RF power delivered by electrode 222 and tuned by bias matching network 220 includes a DC bias on substrate 224 to control the direction of ion bombardment and ion bombardment energy to substrate 224. During the etching process, the plasma chemically reacts with the surface of the substrate 224 to remove material that is not covered by the photosensitive mask.

Computer system 238 is coupled to TCP power controller 206, bias power controller 208, pressure controller 230, and temperature controller 234. Hardware components of computer system 238 are shown in greater detail in FIG. The profile simulator algorithm present in computer system 238 is shown in more detail in FIG. Computer system 238 is configured to control substrate 222 to be processed inside chamber 202 based on settings by TCP power controller 206, bias power controller 208, pressure controller 230, and temperature controller 234. Calculate the profile predictively. Thus, the desired profile of the wafer 222 predicted by the computer system 238 is based on the algorithm's predicted output based on the TCP power controller 206, the bias power controller 208, the pressure controller 230, the temperature controller 234. Is obtained by controlling

Returning to FIG. 3, FIG. 3 shows in block diagram form a hardware system incorporating one embodiment of the present invention. As pointed out herein, the system is not only the main system memory 304 but also the system bus 300 (all system elements communicate over this system bus), and mass storage devices (such as hard disks or optical storage units). 302).

The operation of the illustrated system is controlled by a central processing unit ("CPU"). The user interacts with the system using a keyboard 308 and position detection device (eg, mouse) 310. The output of another device may be used to present information or to select a particular area of the screen display 312 to control the functions to be performed by the system.

Main memory 304 stores a group of modules that control the operation of CPU 306 and interaction with the remaining hardware elements. The operating system 314 regulates the performance of low-level basic system functions such as memory allocation, file file management, and operation of the mass storage device 302. At a higher level, the interpretation module 316, which takes effect as a series of store instructions, controls the execution of the first function performed by the present invention as described below. Instructions defining the user interface 318 allow for direct interaction on the screen display 312. The user interface 318 presents a word or geometric image on the display 312 to facilitate action by the user and receive user commands via the keyboard 308 and / or the position detection device 310. In addition, main memory 304 may store one or more test or processing values of input parameters, including input variables, desired profiles, test surface profiles, and approximate primary test values in plasma models and profile simulators. Has a database 320.

Although the modules of main memory 304 are described separately, this is for clarity only and it should be understood that it is not important how they are distributed within the system and programming structure as long as the system performs all the necessary functions. .

Test surface profiles are empirically produced by testing one or more test substrates in a plasma reactor as well known and measuring the resulting surface profile using, for example, scanning electron microscopy. The desired test surface profile may be provided to the hardware system in electronic form or as a graphical hard copy, in which case the image is processed by the digitizer 322 before numerical comparison with the approximate prediction. The digitized profile is transmitted as a bit stream on the bus 300 to the database 320 of the main memory 304. Test surface profiles may be stored in the mass storage device 302 as well as in the data base 320.

As mentioned above, the execution of the critical tasks associated with the present invention is controlled by the analysis module 316, which governs the operation of the CPU 306 and includes the optimal test values corrected for the initial surface profile model. By controlling the interaction with the main memory 304 in executing the blocks necessary to provide the final mathematical surface profile model, further processing based on the final surface profile and the desired surface profile, thereby providing the desired surface on the process substrate. By determining a processing value of one or more input variables or by inserting the processing values of the input variables into a final mathematical model, by determining a plasma processing sequence suitable for generating a profile of The treatment surface profile to be produced on the treatment substrate can be calculated in a predictive manner.

According to another embodiment, the hardware system shown in FIG. 3 can be used to perform the following calibration process. An input variable test value, a test value of a fixed input parameter, and a test surface profile and, if desired, a desired surface profile and / or associated process values may be provided to the database 320 and provided to the analysis module 416. . Alternatively, module 316 may retrieve test values, approximate primary values and test surface profiles from mass storage 302 or user interface 318 in response to user commands. Alternatively, the approximate primary value may be determined by module 316 based on the input variable test value according to a predetermined algorithm.

By performing plasma modulation and profile simulation respectively, module 316 achieves an initial mathematical surface model that predicts the profile generated by the test process. Module 316 approaches the test surface profile and evaluates the error according to some predetermined criteria compared to the initial mathematical surface profile module. If the error is not small enough, the analysis module 316 uses the comparison results to adjust the test values of the plasma module and the profile simulator. New test values are retained in the database 320 for iteration of the modeling / simulation and comparison blocks. If the test surface profile and the approximate prediction are sufficiently similar, the test values used for the final iteration are stored in the database 320 as optimal values.

The analysis module uses these optimum values of the input variables for the calculation of the processing values, which can be loaded into the plasma reactor for the manufacture of a device comprising the desired profile, or for profile prediction as described above.

The speed and efficiency of the regression analysis to determine the free parameters that lead to the best fit between the test profile and the simulated profile can be improved by distributing calculations over a network of parallel computers or workstations.

Thus, it can be seen that the former is a very scalable and preferred approach for plasma processing of semiconductor devices. Relationships and expressions used herein are used as the relationship of the description and are not limiting, and there is no other intention in the use of such subsections and expressions to exclude equivalents to the features shown and described or as part thereof, It will be appreciated that modifications are possible within the scope of the invention as claimed. For example, the various modules of the present invention may be implemented as a multipurpose computer using a suitable software instruction, or a network of computers, or as a multiprocessor computer or hardware circuit, or in a mixed hardware-software combination (eg, plasma modeling). And profile simulation is performed by dedicated hardware elements).

4 shows a process and method for calculating material morphology, surface chemical coverage, neutral and ion flux of a surface segment within a feature profile to calculate an etch or deposition rate along a feature profile of the substrate. Illustrated. The method starts at 400. At 402, the trajectory of active particles, such as ion trajectories, is tracked until this trajectory leaves the simulation domain. Ions in the plasma reactor move towards the substrate located on the electrode. When ions encounter the surface of the substrate, they are reflected or absorbed. Tracking of trajectories is accomplished by integrating the equations of motion of active particles and using empirical or published theoretical models for ion-wall interactions. At 404, the deposited energy term in each surface segment for each material type, such as silicon, polymer, or deposited film, is recorded for use in the position balance and ratio model. In general, the ion flux is determined by the plasma model. The calculation at 404 is described in more detail in FIG.

At 406, the amount of bonding, such as neutral flux, chemical coverage of the surface, and surface material form in each segment along the feature surface is simultaneously calculated without contradiction. Since each quantity is interdependent with each other, the calculation includes simultaneous calculations and multiple proofs of preservation and coherence of the linking quantity calculated with each equation. Simultaneous calculations are shown in FIGS. 5 and 6.

At 408, local etch and deposition rates and new segment shapes are calculated using standard surface evolution algorithms. In addition, it is desirable that excellent features can be analyzed more accurately by using an analysis method for surface vibration. One such method known in the art and capable of modeling good feature appearance such as sharp edges is a unique method known as a shock-front-tracking algorithm. (See, eg, S. Hamaguchi, "Modeling Film Growth for Electronic Applications," San Diego Academy Press, 1996. S. Losnagel, Thin Films, vol. 22, p. 81). . (See, for example, J. A. Setian, Level Setting Methods: Development Interfaces in Geometry, Hydrodynamics, Computer Vision, and Materials Science, Cambridge University Press, 1996). The shock-front-tracking approach is distinguished Model the surface as a collection of continuous line segments (ie, the interface layer between vacuum and solid), and the rate of movement is calculated for each of them. The possibility of advancing or retracting each segment along the usual, regardless of the movement of the remaining segments, allows for multiple emotional solutions to the resulting surface. To avoid multiple solutions, this method of analysis models the points between line segments (i.e., discontinuities in the slope) as shocks and tracks the motion of the shocks as appropriate. At 410, the process is repeated until the process terminates at 112.

Figure 5 shows the calculation of the bond amount, including neutral flux, surface chemical coverage, and surface material form. The incident neutral flux is first calculated at 500. 6 shows the details of the incident neutral flux calculation.

In block 502, the chemical coverage of the surface on each surface segment is solved using a position balance equation that has previously calculated fluxes and the surface material form is given by the material of the deposited film under this surface segment. At 504, the computer checks whether the chemical coverage of the surface on the surface segment has converged. The calculation of the incident neutral flux for the surface segment estimates the chemical coverage of a particular surface, which also depends on the surface material morphology. Computing the position balance equation yields the chemical coverage of the newly calculated surface and then the chemical coverage of the surface is used to recalculate the incident neutral flux. The new incident neutral flux is then used to recalculate the position balance equation. The repetition of the calculation continues until both the neutral flux and the chemical coverage of the surface are constant with each other through the position balance model. Thus, the neutral flux and position balance equations are calculated and solved until the chemical coverage of the surface on each surface segment converges.

The etch rate of the deposited film on each surface segment forming the feature profile is then calculated at 506. A representative etch / deposition rate equation can be represented by Equation 1 below.

[Equation 1]

Etch Rate = R ^chem θ _Corrosive + Y _Ion θ _Corrosive / ρ _{Deposition Film} -S _Precipitator Γ _Precipitator

Here, the R ^chem's response rate, θ _etchant is a chemical coverage of the neutral species, Y _ions and ion output, wherein the density of the deposited film ρ _{deposited film,} S _settler is an adhesion probability of deposition neutral species, Γ _{needle Electricity} is the incident neutral flux of the deposited neutral species.

At 508, the computer determines whether the calculated etch rate is negative, i.e., deposited. Once the process determines that deposition has taken place, the material shape of the surface segment is set to the deposited film at 512. If the process determines that the etch rate is positive, ie etch, then the material type of the surface element is set to an undering material type at 510.

At 514, it is checked if the material shape of each surface segment converges along the feature profile. Since the calculation of the etch rate of the deposited film at 506 depends on the assumption that the surface material shape is the shape given by the surface segment and the material below the surface segment of the deposited film, the chemical coverage of the surface also depends on the surface material shape and the neutral flux. . By setting the material shape to the underlining material shape at 510 or to the deposited film at 512, the material shape can be changed and the process needs to be maintained in a constant solution with previously determined neutral flux and chemical coverage of the surface and they are also surface material It depends on the form.

The purpose of decision block 514 is to maintain this consistency by checking whether the material shape has converged. Once the amount of bond has converged, the calculation at 508 is likely to compute local etch and deposition rates.

Figure 6 shows the analytical calculation of the neutral flux. This calculation involves solving an equation. Where ΣΓ _i is the total flux entering the segment i, Γ _i ^direct is the flux to segment i arriving directly from the plasma chamber, and Γ _i ^re-emission is from segment i due to liberation or generation of neutral species Is the flux defined by M, where M is the matrix defined as the small amount of neutral emitted from segment j that collides with segment i, and T is the transmissive matrix defined as the small amount of neutral that is reflected on segment i and impinges on segment j. to be. The matrix Γ ^direct , M and T respectively depend on the viewable angles from discrete surface elements to the plasma chamber and from other discrete surface elements.

At 600, neutral species are selected. At 602, for each neutral species a direct flux contribution is calculated. Direct flux is defined as the flux incident on the surface that arrives directly from the plasma chamber without being reflected elsewhere along the feature surface.

At 604, the re-release flux reaching segment i is calculated by defining the flux re-emitted in segment j.

Γ _j = Γ _j ^{re-emission material released from the original material} + θ _settler Γ _j ^{re-emission from a film material} _on ^the _{original one}

Here, _{the precipitator on the original one} is the part of the underlining or original surface covered by the precipitator. The re-released flux from the surface is a function that may depend on surface coverage, surface material morphology and incident neutral and ion flux distributions. The parameters governing the re-emission can be known or determined during calibration of the profile simulator. Film material is defined as the material formed by the accumulation of precipitator species on the surface.

The transmission matrix T is calculated at 606. Reflection probability for particles = 1-Absorption probability on original material-Precipitator coverage on original material × absorption probability on film. The probability of absorption for particles incident on the surface is a function that can depend on surface coverage, surface material type, and the incident neutral and ion flux distribution. The parameter governing the absorption probability may be known or determined during calibration of the profile simulator. Film material is defined as the material formed by the accumulation of precipitator species on the surface.

At 608, the incident neutral flux is solved through a feature using the above equation. If not all of the neutral flux is calculated at 610, another neutral paper 600 may be selected and this process restarts at 602. Otherwise, the process continues to solve the position balance equation at 502.

FIG. 7 shows the flux interaction with the feature surface resulting in a solution to the chemical coverage and / or concentration of related neutral species at the surface and a calculated etch and deposition rate. Referring to FIG. 7, Γ _i ^incidence is the total incident flux of the i th species on the surface segment, Γ _i ^reflection is the incident flux portion of the i th species reflected from the surface, and Γ _i ^re-emission is a chemical reaction and / or ion Is the flux of the i th species that is re-emitted through energy collisions by the accelerated species such as In general, Γ _i ^reflection and Γ _i ^re-emission may depend on chemical coverage or concentration along the surface, incident flux distribution of neutral and ions, surface material morphology, surface temperature, and reaction rate constants. The overall occupancy of the underlining substrate surface position in response to reaction with the N reactive neutral species is represented by the portions θ ₀ , θ ₁ , θ ₂ ... θ _N. The overall occupancy of the surface position on the part of the surface covered by the deposited species, responsible for the macroscopic growth of the film material along the portion of the surface and by reaction with the N reactive neutral species, is the part η ₀ , η ₁ , η ₂ ... is expressed as η _N. The equation used to calculate the value of this surface coverage may depend on the coverage itself, the incident flux distribution of neutral and energetic ion species, the temperature of the wafer and the reaction rate.

The equations for the etch rate of the exposed material and the etch or deposition rate of the film material are shown in FIG. 7 with terms of chemical coverage on the appropriate material, the incident flux distribution of the neutral and energy ionic species, the temperature of the wafer, and the reaction rate constants. It can be expressed as. This ratio can be used to repeatedly feature the feature surface to predict the profile development of the feature.

Coverage for each neutral form is calculated twice in each segment. The first set of coverage θ is calculated based on the assumption that the material type of the segment is original / underlining material. The second set η estimates the shape of the underlining segment of the deposited film. These coverages can be used to calculate etch / deposition rates, release rates, and absorption probabilities for underlining and film material types alone. In the calculation of neutral flux permeation, the segments are treated as linear compounds in the form of two materials and the surface portion assumed in the form of a film is given by the coverage of the precipitator on the underlining / original material.

Simulation of the plasma reactor process may be enhanced by extrapolating the depiction of the feature profile using Cartesian coordinates having three to three degrees of freedom and three degrees of freedom for particle velocity and cylindrical coordinates symmetrically azimuthally. have.

While embodiments and applications of the present invention have been shown and described, those skilled in the art having the benefit of the present disclosure can appreciate that many modifications are possible without departing from the spirit of the invention. Accordingly, the invention is not to be restricted except in the spirit of the appended claims.

According to the present invention, there is provided a process algorithm and method for improving the profile simulator that accurately predicts the surface profile that a given plasma process will produce.

Claims (10)

Plasma chamber, Chamber, A first power supply, A second power supply, Gas supplies and outlets, An inductor positioned above the chamber and connected to the first power supply; An electrode located inside the chamber and connected to the second power supply; A first power controller coupled to the first power supply; A second power controller coupled to the second power supply; A pressure controller connected to said gas supply and outlet, A computer coupled to the first pressure controller, the second power controller, and the pressure controller, The computer predictively calculates a surface profile to be generated on the substrate by the plasma treatment sequence, The computer tracks active particles, The computer records a plurality of ion fluxes from the active particles, The computer simultaneously calculates a number of neutral fluxes, chemical coverage of the surface, and surface material morphology, The computer calculates a local etch rate and a local deposition rate from a plurality of ion fluxes, the plurality of neutral fluxes, chemical coverage of the surface, and the surface material type. The method of claim 1, wherein simultaneously calculating the plurality of neutral fluxes, the chemical coverage of the surface, and the surface material morphology, Computing a plurality of neutral fluxes, Solving a number of position balance equations without self-contradiction, Calculating an etch rate of the deposited film; Adjusting the surface material shape relative to the deposition film or underlining material shape; And repeating the computing, solving, calculating, and adjusting steps until the plurality of neutral fluxes, chemical coverage of the surface, and the surface material form are not incompatible with each other. The method of claim 2, wherein computing the plurality of neutral fluxes comprises: Selecting a neutral species, Calculating flux distribution directly for the neutral species, Estimating the re-release flux for the neutral species, Computing a permeation matrix for the neutral species; Solving the neutral flux for the neutral species, Repeating the selection step, calculating the direct flux distribution, calculating the re-emission matrix, computing, and solving other neutral species until all neutral species are selected. . The plasma chamber of claim 1, further comprising a network of parallel computers or workstations coupled to the computer for performing the calculation of the regression analysis. The plasma chamber of claim 1, wherein the computer tracks the active particles using three-dimensional coordinates. Method of operation of the plasma chamber, Calculating a process surface profile generated on the substrate by the plasma process sequence; Processing the substrate according to the plasma process sequence, The calculating step, Tracking active particles, Recording a plurality of ion fluxes from the active particles, Simultaneously dissolving a plurality of neutral fluxes, chemical coverage of the surface, surface material type, Computing a local etch rate and local deposition rate from the plurality of ion fluxes, the plurality of neutral fluxes, chemical coverage of the surface, the surface material form. The method of claim 6, wherein simultaneously solving the plurality of neutral fluxes, chemical coverage of the surface, and the surface material form, Computing a plurality of neutral fluxes, Solving a number of position balance equations without self-contradiction, Calculating an etch rate of the deposited film; Adjusting the surface material shape relative to the deposition film or underlining material shape; And repeating the computing, solving, calculating, and adjusting steps until the plurality of neutral fluxes, chemical coverage of the surface, and the surface material form are inconsistent with each other. Way. 8. The method of claim 7, wherein computing the plurality of neutral fluxes comprises: Selecting a neutral species, Calculating flux distribution directly for the neutral species, Estimating the re-release flux for the neutral species, Computing a permeation matrix for the neutral species; Solving the neutral flux for the neutral species, Repeating the selection step, calculating the direct flux distribution, calculating the re-emission matrix, computing, and solving other neutral species until all neutral species are selected. How does it work? 7. The method of claim 6, further comprising distributing the solving and computing by a network of parallel computers or workstations. 7. The method of claim 6, further comprising tracking the active particles using three-dimensional coordinates.

Images