JP2010219198A - Plasma processing apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To solve such a problem that the uniformity of etching shape among wafers aggravates by temperature fluctuation during continuous processing and during stoppage because the etching performance of wafers varies by temperature. <P>SOLUTION: The inner wall of an reactor is heated up to optimal temperature by plasma discharging prior to etching processing, and while taking a period of etching operation stoppage (idling) into account, the generation of plasma 402 is intermittently repeated during idling, so as to keep the surface temperature of inner wall of a vacuum processing chamber. The emission intensity of a processing gas correlated with the surface temperature of inner wall of the vacuum processing chamber is measured to control the intermittent generation of plasma 402. <P>COPYRIGHT: (C)2010,JPO&amp;INPIT

Description

  The present invention relates to an apparatus for performing semiconductor surface treatment by physicochemical treatment of activated particles by converting a source gas into plasma, and more particularly to a plasma processing apparatus for performing plasma etching.

  Conventionally, a magnetic field plasma generator has been used in a plasma processing step used in manufacturing a semiconductor device. This magnetic field plasma generator introduces a processing gas into a processing chamber which is a vacuum vessel, generates plasma by microwaves supplied through a waveguide, and etches a silicon wafer placed on a lower electrode. The lower electrode operates as a lower electrode, and an etching rate and an etching shape can be controlled by applying a bias potential.

  In Patent Document 1 and Patent Document 2, the processing chamber wall is heated under discharge conditions set in advance corresponding to the processing chamber wall temperature obtained by numerical calculation or the like prior to wafer processing, and uniform etching performance is obtained. The technique for obtaining is described. However, it is not always possible to estimate the temperature of the inner wall of the processing chamber, and there is a need for a heating and discharging means that does not require resetting the discharge conditions again even when the etching processing conditions are changed. Met.

JP 2006-210948 A JP 2006-244065 A

  The present invention relates to an improvement in yield by stabilizing short-term fluctuations in processing performance. Along with the recent miniaturization of semiconductors, the etching shape and uniformity of the etching shape for each sample have increased the impact on product yield. One example of this has emerged as a new problem. Is the reactor temperature.

  The wafer to be processed was loaded on a temperature-controlled sample stage, and the wafer temperature was optimized. However, as the processing was repeated, the temperature of the reactor inner wall heated by the heat input from the plasma was also affected. . One of the causes is a change in the reaction rate and reaction product composition on the wall surface as the temperature of the reactor inner wall changes. As a result, molecules etc. re-emitted from the wall may affect the etching performance.

  There is also heating by radiant heat from the reactor inner wall. The wafer temperature is controlled by the sample stage to greatly affect the etching performance, but the surface may be heated by heat input from plasma or the like. Although plasma is dominant as a heat source, the wafer can be heated by radiation from the inner wall of the reactor, except during plasma processing.

  Furthermore, fluctuations in the introduced gas temperature can be considered. While part of the inner wall material that constitutes the reactor also serves as a gas introduction path and is affected by temperature, the pressure gauge is arranged so that the influence of temperature is negligible, so the pressure at the measurement unit is constant. Even if controlled, the gas density in the reactor may fluctuate and affect the etching performance.

  When continuous processing is performed, due to the above-described heating by plasma, the temperature of the portion in the reactor where temperature control is not performed gradually rises, and saturates when sufficient time has elapsed. When a plurality of wafers are processed intermittently, the temperature drops during the standby time. For the reasons described above, the etching performance varies depending on the temperature. Therefore, the uniformity of the etching shape between the wafers deteriorates due to the temperature variation during continuous processing and during stoppage.

  For the above reasons, temperature control in the reactor is important, but it has been very difficult to control by directly heating or cooling the inner wall surface temperature of the reactor. Therefore, it is heated to a range that does not affect the etching performance based on the saturation temperature that is reached by continuous processing, and it is reheated immediately before the processing is resumed or maintained while the etching processing is interrupted. It is necessary to. At the same time, some conventional etching apparatuses do not have means for monitoring the inner wall temperature, and even in such a case, means for determining the end point of the heat treatment is necessary.

  According to the present invention, the reactor inner wall temperature is heated to an optimum temperature by heating by plasma discharge prior to the etching process. Alternatively, during the etching process operation stop (hereinafter referred to as “idling”), discharge for maintaining the temperature is performed.

  The idling in the present invention indicates a standby time between processing of a plurality of continuous wafers and another continuous processing of a plurality of wafers, and a standby time after the apparatus maintenance until the start of wafer processing. The plasma processing apparatus according to the present invention has a computer that determines the presence or absence of a wafer to be processed next after processing a certain wafer, a timer that measures an idling time, an emission spectrometer, and a portion that correlates with etching performance. A temperature monitor for measuring the temperature as necessary, and having a discharge condition prepared for effectively raising the temperature in advance, and an optimum time for heating to a desired temperature The optimum reactor temperature is always obtained during wafer processing.

  By the above means, the state of the vacuum processing chamber in which plasma processing is performed is always kept constant, and as a result, a plasma processing apparatus having a stable etching performance over a long period of time is provided.

  The plasma processing apparatus of the present invention includes a vacuum processing container that forms a vacuum processing chamber, a processing gas supply device that supplies a processing gas into the vacuum processing container, and a processing gas supplied into the vacuum processing container. A plasma generating means for generating plasma in the room, judging that the apparatus is in a dormant state, and repeating the generation of plasma as a heat source for maintaining the inner wall surface temperature of the vacuum processing chamber even during the pause. And

  Further, the plasma processing apparatus of the present invention further generates and discharges plasma for heating the inner wall surface of the vacuum processing chamber under a predetermined processing condition and processing time prior to wafer processing, The inner wall surface temperature of the vacuum processing chamber is controlled.

  In addition, the plasma processing apparatus of the present invention further has a specific wavelength of plasma having a correlation with the inner wall surface temperature of the vacuum processing chamber when performing plasma discharge for controlling the inner wall surface temperature of the vacuum processing chamber. By measuring the emission, the temperature change during heating of the inner wall of the vacuum processing chamber is measured, and the discharge time of the plasma discharge is controlled using a preset threshold value of the emission intensity, thereby controlling the inner wall of the vacuum processing chamber It is characterized by controlling the surface temperature.

  The plasma processing apparatus of the present invention further performs plasma discharge for controlling the inner wall surface temperature of the vacuum processing chamber when the inner wall surface temperature of the vacuum processing chamber is saturated. The threshold value for determining the end of discharge in the discharge time is automatically set again.

  By providing the above-described configuration, the present invention can provide an etching apparatus and an etching method that suppress variation in etching shape between wafers and realize a high yield.

FIG. 1 shows a plasma processing apparatus in Embodiment 1 of the present invention. FIG. 2 is a diagram illustrating a typical temperature change in a continuous process when no heat discharge is performed. FIG. 3 is a diagram illustrating a temperature change when there is a processing stop time between two continuous processes. FIG. 4 is a diagram showing maintenance of the apparatus temperature by performing intermittent discharge. FIG. 5 is a diagram showing the correlation between temperature and emission intensity during heating discharge when the gas containing F (fluorine) of Example 2 of the present invention is added. FIG. 6 is a diagram for explaining a means for correcting the threshold value for determining the end point. FIG. 7 is a graph showing the correlation between the temperature during heating and discharging and the SiF emission intensity when a small amount of gas containing F (fluorine) is added before and after idling and at the time of starting the apparatus.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

  FIG. 1 is a longitudinal sectional view showing an outline of a configuration around a vacuum processing chamber according to a plasma processing apparatus of the present invention. Microwaves oscillated from the microwave source 101 are transmitted using the rectangular waveguide 103 and connected to the circular waveguide 105 by the rectangular circular waveguide converter 104. The automatic matching machine 102 can adjust the load impedance to automatically suppress the reflected wave.

  As the microwave source 101, a magnetron having an oscillation frequency of 2.45 GHz is used. The circular waveguide 105 is connected to the cavity resonance unit 106. The cavity resonance unit 106 has a function of adjusting the microwave electromagnetic field distribution to a distribution suitable for plasma processing. A plasma processing chamber 119 is provided below the cavity resonance unit 106 through a microwave introduction window 107 and a shower plate 108.

  The plasma processing apparatus of the present invention generates a reactive gas plasma in the plasma processing chamber 119 by the microwave introduced into the plasma processing chamber 119 and the ECR (Electron Cyclotron Resonance) resonance of the magnetic field formed by the solenoid coil 120. It is a plasma etching apparatus that forms and performs an etching process.

  ECR resonance is a method in which electrons move selectively along the magnetic field lines of the magnetic field generated by the solenoid coil 120, and electrons are selectively incident on the plasma by injecting microwaves having a frequency corresponding to the rotation period. It is an effective method of heating plasma. Another advantage of using a static magnetic field is that the position where ECR resonance occurs can be controlled by changing the distribution of the static magnetic field, and the plasma generation region can be controlled.

  Furthermore, it is known that diffusion of plasma is suppressed in a direction perpendicular to the lines of magnetic force, and it is possible to control plasma diffusion and reduce plasma loss. Due to these effects, it is possible to control the distribution of the plasma and thus improve the uniformity of the plasma treatment.

  A three-stage electromagnet was used as the solenoid coil 120 which is a static magnetic field generator. The distribution of the static magnetic field can be controlled by changing the current supplied to each electromagnet.

  The reactive gas is supplied from a gas source 109 from a shower plate 108 installed on a surface facing the lower electrode 114 serving as a sample stage, and its flow rate is controlled by a mass flow controller (hereinafter referred to as “MFC”) 111. . The gas valve 110 is provided to start or end the gas supply. The gas in the plasma processing chamber 119 is exhausted from a turbo molecular pump (hereinafter referred to as “TMP”) 118, and its exhaust speed is controlled by a movable valve 117 provided upstream of the TMP 118. To control the pressure.

  A silicon wafer as a sample (hereinafter simply referred to as “wafer”) can be attracted and held on the lower electrode 114 by electrostatic attraction. Furthermore, by applying an RF wave having a frequency of 400 kHz from the RF power source 116 to the lower electrode 114 via the RF matching unit 115, the processing performance can be controlled and the processing speed can be improved. The plasma processing chamber 119, the lower electrode 114, and the TMP 118 each have a substantially cylindrical shape, and the axes of the cylinders are the same. The lower electrode 114 is supported by a beam in the plasma processing chamber 119.

  In addition, all the above configurations are connected to the control computer 113, and their timing and operation amount are controlled so that they operate in an appropriate sequence. The operation sequence is called a recipe, and an operation based on a preset recipe is performed. In order to determine the end point of etching, the plasma processing apparatus is provided with a spectroscope 112 attached via an optical fiber 121 and a control computer 113 connected to the spectroscope 112, and emits light from the plasma for wavelength / time. Decompose and measure and record. In addition, the control computer 113 determines whether or not a wafer is inserted, and if it is determined that idling is in progress, the control computer 113 operates a timer to measure the idling time. Further, when particularly required, a temperature sensor 122 is provided at a position where the temperature of the inner wall of the reactor can be measured.

  FIG. 2 shows a time change (201) in a conventional reactor without temperature control. This example is a part where a correlation is found between the process performance and the temperature, particularly, until the temperature of the measurement part is almost saturated when a plurality of wafers are continuously processed. . The temperature rises rapidly in the initial stage, and the temperature changes gradually from a certain point of time, but continues to rise gradually and eventually saturates.

  FIG. 3 shows the temperature change during the idling 302 described above. When the plasma treatment is completed (303) and the heating by the plasma is stopped (302), the temperature drops rapidly. When the plasma treatment is resumed, the temperature rapidly increases (304), and reaches a certain steady temperature when a sufficient time has elapsed.

  As a means to obtain an appropriate reactor temperature from the start of plasma processing start, it is possible to maintain temperature by inserting intermittent discharge when it is judged that idling is in progress, and to obtain stable etching performance Is possible.

  FIG. 4 is a specific example of this, and is a temperature change (402) in the case of repeating discharge for a fixed time which is a certain period during idling. Although the temperature change rate is fast and the fluctuation range is large, the high temperature is maintained on a time average as compared with the temperature change (401) in the case where the intermittent discharge is not performed, and the plasma processing is resumed. It can be seen that the temperature quickly rises to the temperature before idling even at the start of construction.

  The period of intermittent discharge and the ratio of discharge time and standby time are determined as follows. If a part having a high correlation between etching performance and temperature is known, the thermal resistance of the part can be easily estimated, and the part in contact with the atmosphere of the reactor is at room temperature. Since the amount of heat input from plasma under a certain temperature rise condition can be estimated from the result shown in FIG. 2 and the thermal resistance, it is given to the reactor on a time average by changing the ratio of the discharge time and the standby time. The amount of heat can be controlled, so that the reactor temperature during idling can be controlled.

  Here, the conditions of the heat discharge are optimized for the heating described later. If the standby time in intermittent discharge is too long, the timing at which the etching process can be performed is limited, and the production capacity is reduced. Since the efficiency is improved when switching is performed in a region where the temperature rises and falls quickly, the actual discharge / standby time is preferably about 10 minutes or less.

  The idling here is not only the stop time between continuous processing, but also applies during the start-up time after maintenance work involving the release of the atmosphere to the reactor, and plasma discharge is possible By performing intermittent discharge from the determined time point, the etching performance at the start of processing can be stabilized.

  In addition, when the etching process is started, it is possible to stabilize the etching performance by performing a process using a recipe optimized for heating the reactor immediately before the etching process.

  When the reactor is heated by the heating discharge, a mechanism for determining that the reactor temperature has reached the required temperature range, that is, determination of the end point is required. Here, when the etching process is started again during idling, the change curves when the temperature rises are substantially equal regardless of the temperature before the start of heating. Therefore, if the temperature at the start of heating is known, the optimum temperature can be obtained by controlling the heating and discharging time.

  As described above, when there is no means for measuring the temperature inside the reactor, there is a method for predetermining the heating time required in each case by dividing it into several appropriate conditions as the means for determining the heating and discharging time. is there.

  Since the temperature change curve can be modeled, it is sufficient to prepare a database having a sufficient dimension corresponding to the model. The specific procedure for building a database from experimental data is as follows.

  First, in general, a temperature change curve can be expressed by a linear combination of exponential functions. In the example of the present invention, it can be expressed by a combination of three exponential functions having different coefficients and a constant term that is a temperature at the start of idling. That is, the reactor temperature is a function having the temperature at the start of idling and the idling time as variables.

  Regarding the idling time, since the temperature change curve has three time constants, there are three elements that are dominant in the temperature change. One of them showed a temperature drop for a few seconds immediately after the start of idling, so it was ignored and the remaining two time constants were determined to be dominant and stable at a constant temperature. There are three cases, including cases after a sufficient amount of time has passed.

  The temperature at the start of idling can be estimated as follows. As shown in FIG. 3, the temperature rise curve (301) and the temperature rise curve (304) at the time of reprocessing after idling (302) are similar. Thus, it has been found that the temperature at which continuous processing ends and idling starts is a function of the number of wafers processed so far. In the case of heating discharge, it was sufficient to divide the case into three according to the number of wafers processed. Therefore, in this example, the number of processed sheets immediately before idling and the elapsed time of idling are three ways, for a total of nine ways, and the heating and discharging time can be controlled.

  The necessary amount of data can be determined by the above-mentioned procedure. For example, a database is created using data obtained by confirming the heating time required in typical cases in advance by experiments, and heating is performed before the start of plasma processing. By performing the discharge, it becomes possible to obtain stable etching performance.

  The heating discharge is a recipe optimized for reactor heating. It is the microwave power that affects the heating performance. By applying more electric power, the electron temperature of plasma rises and heating is promoted. The composition of the plasma is not particularly limited, but by using a rare gas, heating without destroying the atmosphere in the reactor is possible.

  On the other hand, by using plasma of a gas containing halogen such as F (fluorine), the reactor can be cleaned together with heating. Those that can stabilize the performance in relation to the process of processing the wafer are selected.

  Further, in the ECR plasma, a high density plasma is generated in a portion along the ECR plane, so that the portion that intersects the ECR plane is most effectively heated. By utilizing this, it is possible to continuously change the magnetic field conditions during overheating discharge to increase the heating efficiency, and at the same time to control the temperature distribution.

  Moreover, the combination with the intermittent discharge mentioned above is also possible. If the standby time during intermittent discharge is long, the processing capacity will decrease, but by just counting the elapsed time in the standby state immediately before the etching process, the appropriate heat discharge time can be selected to obtain the optimum reactor temperature. Is possible.

  The end point of the heating discharge can be determined using a specific wavelength of plasma emission. When a rare gas plasma is used as the heating discharge, a small amount of a gas containing F (fluorine) such as SF6 or CF4 is added. F (fluorine) radicals react slightly with the quartz parts inside the reactor to produce SiF. Since the reaction rate of this reaction is proportional to temperature, the SiF density in the plasma is proportional to the reactor temperature.

  Plasma emission is a function of plasma density and electron temperature, but since the electron temperature may be considered constant, the emission intensity of SiF is proportional to the reactor temperature, and the correlation between the emission intensity of SiF and the reactor temperature is measured in advance. The temperature measurement using a photometer becomes possible by having an approximate expression with an appropriate function obtained in the apparatus.

  FIG. 5 shows an example in which a small amount of the above-mentioned gas containing F (fluorine) is added and overheating discharge is performed. Reference numeral 501 shows the emission intensity (vertical axis) of SiF during heating discharge and the reactor temperature (horizontal axis) at that time. It can be seen that the emission intensity increases as the temperature rises, and that the temperature and the emission intensity can be approximated by a straight line except in the low temperature region at the beginning of discharge.

  To determine the end point of heating using this, the SiF emission intensity corresponding to the desired temperature is set as a threshold value, and during the heating discharge, the SiF emission intensity is measured, and when the set threshold value is reached It suffices to end the discharge.

  Further, there is a possibility that the correlation between the SiF emission intensity and the temperature may be shifted due to a change in the reactor atmosphere, and there is a possibility that an erroneous determination is made in the end point determination using a fixed threshold value. As a means for solving such a problem, there is a method in which processing equivalent to heating discharge is performed before the start of idling (603), the emission intensity at that time is recorded, and the value is used as a new threshold value. Details are shown in FIG. In the heating discharge (605) after idling (604), the discharge is terminated when the emission intensity measured by the threshold measurement discharge (603) before idling is reached, and the etching process (606) may be performed.

  FIG. 7 is a graph in which the horizontal axis represents the temperature at which the discharge for heating was performed, and the vertical axis represents the SiF emission intensity when idling for 30 minutes after idling for 30 minutes when overheating from the steady temperature. Show. Although there is a deviation from the proportional relationship between the temperature and the light emission intensity at the beginning of discharge, it is sufficient for any of the cases before idling (702), after idling (703), and before starting construction (701) immediately after starting the apparatus. It can be seen that if the discharge is performed for a long time, the relationship between the reactor temperature and the SiF emission intensity is aligned in a straight line, and it is possible to reset the end point of heating and determine the end point by measuring the emission intensity.

DESCRIPTION OF SYMBOLS 101 Microwave source 102 Automatic matching machine 103 Rectangular waveguide 104 Rectangular circular waveguide converter 105 Circular waveguide 106 Cavity resonance part 107 Microwave introduction window 108 Shower plate 109 Gas source 110 Gas valve 111 Mass flow controller (MFC)
112 Spectrometer 113 Control computer 114 Lower electrode 115 RF matching device 116 RF power source 117 Movable valve 118 Turbo molecular pump (TMP)
119 Plasma processing chamber 120 Solenoid coil 121 Optical fiber 122 Temperature sensor 201 Temperature change of the correlation part with etching performance in continuous processing without heating discharge 301 Same as 201 302 Temperature change during idling 303 End of plasma processing 304 After idling Temperature change during reprocessing 401 Temperature change during idling when intermittent discharge is not performed 402 Temperature change during idling when intermittent discharge is performed 501 Temperature when gas containing F (fluorine) is added and heating discharge is performed 601 Typical temperature change inside the reactor 602 Etching process 603 Heating discharge before idling 604 Idling 605 Heating discharge after idling 606 Etching process 701 Temperature and emission during heating discharge immediately after starting up the apparatus Correlation between light intensity 702 Correlation between temperature and light emission intensity during heating discharge before idling 703 Correlation between temperature and light emission intensity during heating discharge after idling

Claims (4)

  A vacuum processing chamber forming a vacuum processing chamber, a processing gas supply device for supplying a processing gas into the vacuum processing chamber, and a plasma generating means for converting the processing gas supplied into the vacuum processing chamber into plasma in the vacuum processing chamber The plasma processing apparatus is characterized in that it is determined that the apparatus is in a dormant state, and plasma generation as a heat source for maintaining the surface temperature of the inner wall of the vacuum processing chamber is repeated even during the pause.   2. The plasma processing apparatus according to claim 1, wherein plasma for heating the inner wall surface of the vacuum processing chamber is generated and discharged under preset processing conditions and processing time prior to wafer processing, and the vacuum A plasma processing apparatus for controlling an inner wall surface temperature of a processing chamber.   The plasma processing apparatus according to claim 2, wherein when performing plasma discharge for controlling the inner wall surface temperature of the vacuum processing chamber, the plasma processing apparatus has a specific wavelength of plasma correlated with the inner wall surface temperature of the vacuum processing chamber. By measuring the emission, the temperature change during heating of the inner wall of the vacuum processing chamber is measured, and the discharge time of the plasma discharge is controlled using a preset threshold value of the emission intensity, thereby controlling the inner wall of the vacuum processing chamber A plasma processing apparatus for controlling a surface temperature.   The plasma processing apparatus according to claim 3, wherein when the inner wall surface temperature of the vacuum processing chamber is saturated, plasma discharge for controlling the inner wall surface temperature of the vacuum processing chamber is performed, whereby the plasma discharge is performed. A plasma processing apparatus characterized by automatically resetting a threshold value for determining a discharge end of a discharge time.

Images