JPH09205076A - Method for monitoring fabrication process of semiconductor element - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a plasma etching process control method for removing silicon dioxide from the surface of a semiconductor water at the time of fabricating a semiconductor element. SOLUTION: In this system, a semiconductor wafer covered with a patterned photoresist layer is set in a reaction chamber and fluorocarbon is introduced into the chamber. A sufficient quantity of energy is then imparted to the gas in order to generate plasma for etching silicon dioxide deposited on the surface of the wafer. Optical emission spectrum(OES) of the plasma is monitored as etching proceeds and some kinds of radiation intensity of the plasma are observed individually at specified wavelengths. Radiation intensity thus observed is compared with a calibration curve and an etching rate of photoresist is determined.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to control of plasma etching used in manufacturing semiconductor devices.

[0002]

2. Description of the Related Art In the process of manufacturing semiconductor devices, dry etching is a process often used because it is convenient for removing material from the surface of a semiconductor substrate.
Dry etching either selectively removes some of the material layers on the substrate, thereby forming a patterned layer or transferring the pattern of the top layer to the layer below it. Used for. Selective removal of material from the substrate surface is an integral part of forming integrated circuit patterns on semiconductor substrates. The plasma process is used to perform a dry etching step in forming an integrated circuit device. This plasma is generated by applying energy to the reactive gas, thereby generating high energy species (ions, neutral groups, atoms, molecules). When plasma is applied to the surface of the semiconductor substrate, the high energy species separate the material from the substrate.

In the plasma etching process, a part of the substrate surface which should not be removed by etching is covered with a protective material called a mask. This mask exposes only part of the substrate surface to be removed.
However, it is not preferred to remove the material beneath the normally removed material layer. Therefore, it is desirable to accurately detect the end point of the plasma etching process.

One way to detect the end point of the plasma etching process is to monitor the emission intensity of the plasma. The end point of this plasma etching process can be determined by observing changes in the emission intensity of plasma. In the process of plasma etching silicon dioxide using a fluorocarbon (fluorinated hydrocarbon) -based reactive gas, the emission intensity of carbon oxide (a reaction product of high-energy plasma species and silicon dioxide) is monitored. . A sharp decrease in the emission intensity of carbon oxide in the plasma indicates that the plasma is at or past the end point.

According to US Pat. No. 5,322,590, the problem with observing changes in the emission intensity of etching reaction products is that in current processes, the amount of surface etched is 10 times the surface of the wafer. % Or less, and further 1% or less. The rest of the wafer surface is covered by the mask. Therefore, the reaction products in the plasma are dominated by the reaction products from the interface between the mask and the plasma. The change in the emission intensity of the reaction product of the etching from immediately before to immediately after the etching end point cannot be easily observed because the amount of other species in the plasma is large and the emission spectrum of other species is , Because it interferes with the emission spectrum of the monitored species in the plasma.

The above-cited patent proposes observing the active etchant species, ie the groups that actually react with silicon dioxide in the plasma, and determining the end point from the change in the emission intensity of the species. Furthermore, in the above-mentioned patent, the emission intensity of the active etchant species was observed when it was reacting with silicon dioxide, compared with when the material (usually silicon) underneath was exposed after the silicon dioxide was etched. It says it is as low as possible. When the underlying silicon is exposed, the active etchant species do not react with the silicon and are no longer consumed by the reaction. In addition, in the above-mentioned patent, comparison of such improvement in observation accuracy is made based on the theory that the change in the emission intensity of the active etchant species and the change in the emission intensity of the reaction product and the reaction species have a clearer effect. It can be improved by doing.

[0007]

However, the change in emission intensity of an individual species detects the end point in processes where the surface being etched is relatively small, i.e. less than 10% of the total substrate surface. Since it is not possible to provide enough information to do so, another process is desired that controls the plasma process and determines the end point of this process. Accordingly, it is an object of the present invention to provide a method of controlling a plasma etching process for removing silicon dioxide from the surface of a semiconductor wafer during the formation of semiconductor devices.

[0008]

SUMMARY OF THE INVENTION In the process of the present invention, a semiconductor wafer covered with a patterned photoresist layer is placed in a reaction chamber in which a plasma is generated. The patterning of the photoresist layer is done throughout its thickness, some parts of the wafer being covered by the mask and other parts not being covered. Silicon dioxide is formed on at least the unmasked portion of the wafer surface. Fluorocarbon,
That is, fluorocarbons (Freon, Du Pont de Nemours
A registered trademark of Co.) is introduced into this chamber. A sufficient amount of energy is added to the gas to generate plasma. Silicon dioxide is etched by this plasma.
The optical emission spectrum of this plasma (optical emissi
on spectrum (OES) is monitored as the etching progresses and is based on observing certain radiant intensities in the plasma individually at a given wavelength, and comparing this radiant intensity with a given calibration curve. Then, the etching rate of the photoresist is determined. The etch rate of the contact holes in the silicon dioxide layer can also be determined in this way.

Using the process of the present invention, a diameter of 0.5
It is possible to determine the etching rate of the contact hole of μm or less. The photoresist etch rate and / or contact hole etch rate can be observed by determining whether the observed etch rate is within a predetermined range. If the etching rate is within the predetermined range, the etching is advanced. If it is not within the predetermined range, the etching is stopped or the process conditions are changed to bring the etch rate back within the predetermined range. Although the present invention describes a photoresist layer formed on a silicon dioxide layer as an example, the present invention describes a large photoresist layer between the surface area of the photoresist mask and the unmasked exposed underlying surface area. It is also used to monitor the plasma etching of the wafer when there is a difference. When there is a large difference, 80% or more of the wafer is masked by the photoresist layer and 20% or less of the oxide-coated wafer portion is used.

In one embodiment, the photoresist coated wafer is exposed to etching conditions using a C ₂ F ₆ plasma. C ₂ F ₆ plasma is conventionally used to etch silicon dioxide during the semiconductor device formation process. The mechanism and conditions for generating the C ₂ F ₆ plasma are conventionally known.

When a wafer having a photoresist mask formed on silicon dioxide is etched with plasma containing fluorocarbon, the optical emission spectrum (OES) has a peak of emission intensity at a certain wavelength.
This peak represents the presence and relative concentration of certain species in the plasma. This OES can be used to determine the etch rate of the photoresist or contact hole. The reason is that some emission intensity in the plasma is related to the etch rate.

The species monitored and the method of monitoring with OES is a function of the etch rate to be determined. When using OES to monitor the etch rate of a photoresist, the emission intensity of at least two species of OES must be monitored. One of the plurality of species is a by-product of the interaction of the fluorocarbon plasma with the carbon species on the walls of the photoresist and the etcher. Desirably, the species is C ₂ . Other monitored species relate to plasma intensity. For example, the OES spectrum during the fluorocarbon etch of a photoresist masked wafer has a peak intensity at the 440 nm wavelength associated with SiF species in the plasma. As the plasma intensity increases, so does the intensity of this peak. As the plasma intensity decreases, so does the intensity of this peak. Therefore, the peak of SiF species is related to the intensity of the plasma. Other species, such as Si or F, can be monitored because the peak of the OES spectrum associated with this species is related to plasma intensity. OES has several peaks associated with C ₂ in the plasma, but it is preferred to monitor the peak at 515 nm. To monitor the presence of SiF species in the plasma, it is preferable to observe the 440 nm peak in OES. When monitoring the presence of Si species in the plasma, 288 nm optical radiation is used. Optical radiation at 703 nm is used when monitoring the presence of F species in the plasma. In the following description, C ₂ optical emission (at 515 nm) and SiF optical emission (440 n) are used.
The invention is illustrated by the example in which the (in m) is observed.

The optical emission of these two species in the plasma is related to the photoresist etch rate by measuring the thickness of the photoresist layer on the wafer prior to the start of etching. The wafer is then plasma etched and the optical emission of the plasma at some relevant wavelength is observed as a function of time. After etching is complete, the thickness of the remaining photoresist layer is measured to determine the etch rate for the particular set of process conditions.

When using OES to monitor the etch rate of a contact hole, one species of optical radiation in the plasma is monitored. The species can be any species in the plasma that is associated with the contact hole etch rate. Optical emission of C ₂ or SiF species in the plasma should be monitored. The reason is the S / N of these signals
The ratio is high enough during the etching process. This optical emission is related to the etch rate of the contact holes by measuring SEM (scanning electron microscope) cross sections of the patterned wafers after a certain etching treatment time. After this certain etching treatment time, the wafer has contact holes etched therein. After a certain etching treatment time, some of the remaining photoresist mask is not removed from the wafer before SEM measurement. The depth of the contact hole can be measured from this SEM. Thereafter, this etch rate is calculated from the depth of the contact hole and the etching time. This etch rate is then determined by the OES obtained during etching.
Associated with trace.

The above process for associating the etch rate with the OES parameters is repeated a plurality of times to change the individual process parameters (eg temperature, flow rate, pressure, radio frequency bias, source power, O ₂ clean time) and photoresist. It is desirable to change the etch rate of the contact holes and / or contact holes and to determine the correlation between the optical radiation effect and the photoresist etch rate and / or the contact hole etch rate. By varying the parameters selected to change the process parameters and the amount of that parameter, the effect over the etch rate can be obtained. Therefore, only the parameters that affect the etch rate when the process parameters are changed can be selected for the purposes of the present invention. It is desirable to be able to provide an etch rate such that the parameters selected to change the process parameters change by a factor of 2, 3 or more relative to each other.

For example, the effect of changing the plasma source power on the photoresist etch rate can be observed by changing the plasma source power for each etching and otherwise etching a plurality of wafers under the same conditions. . The effect of the intensity of the relevant wavelength is obtained in OES over time and this effect is related to the etching rate of the photoresist. Alternatively, the process is repeated for other process parameters such as chamber temperature.

According to one embodiment of the invention, the chamber O
₂ Change in clean duration (etching chamber
It is cleaned with O ₂ after every etching to remove deposits on the walls of the chamber during etching)
The effect on the contact hole etching rate can be observed. There is a complex relationship between the O ₂ cleaning time and the contact hole etch rate. If the O ₂ cleaning time is insufficient to remove all deposits in the chamber, the etch rate of the contact rate holes will vary from etch to etch. Such effects are cumulative and are apparent with each etch. The relationship between O ₂ cleaning time and etch rate is measured as a function of the number of wafers etched. OES traces are obtained at a selected wavelength many times for each etch. This trace is related to the etch rate in the manner described above.

After the traces have been obtained for various etch rates as described above, the ratio of the signal from each trace is related to the particular etch rate. This can be done in various ways. In one embodiment, the ratio of the two signal intensities associated with two different species in the plasma (eg, C ₂ , SiF) (at different time points in the etching process, eg, 75 seconds after entering the etch). ) Is obtained for each trace. This ratio is then mapped as a function of photoresist etch rate. Using such information, the ratio is observed in real time during etching, and the etch rate is monitored by referring to the correlation between the ratio and the etch rate.

In another embodiment of the invention, the signal strength associated with a single species (C ₂ , SiF) in the plasma is monitored and two different time points (t ₁ , t ₂ ) during the etching process are monitored. ), Get the ratio of this intensity in. This ratio is then mapped as a function of contact hole etch rate. This etch rate is obtained by obtaining the ratio of the signal strengths at times t ₁ and t ₂ from the OES trace.
It can then be determined in real time from previously obtained calibration information and by finding the etch rate associated with these ratios.

In the third embodiment of the present invention, the integral value of the ratio of the signal intensity of the first type and the signal intensity of the second type is calculated over the entire etching time. This integrated value is related to the etching rate in the manner described above. In all the examples described above, once a ratio is associated with an etch rate, the relationship between this ratio and the etch rate can be determined by plotting this ratio as a function of etch rate. This plot is a calibration curve used to control the process of the present invention.

In one embodiment of the present invention, the signal-to-noise ratio of the C ₂ -related signal in the OES trace is determined by the photoresist or dioxide formed on the silicon wafer in the RF-generated fluorocarbon plasma. It can be reduced by etching a silicon wafer on which silicon is not formed. Observe the C ₂ signal of this OES trace. This C ₂ signal contains contributions from fluorocarbon gas and residual carbon containing species inside the chamber. Therefore, with a blank silicon wafer, we obtain the C ₂ signal in the plasma and this signal versus the C ₂ obtained when etching the wafer covered with photoresist.
By taking the ratio with the signal, the background effect of the fluorocarbon and the background effect inside the chamber in the C ₂ signal of the OES trace (obtained under the same etching conditions as those used to obtain the background signal) are obtained. Can be removed.

Once the calibration curve is obtained as described above, the OES obtained during etching can be used to perform real-time process control. For example, the ratio of the C ₂ peak at 515 nm to the SiF peak at 440 nm can be determined from OES at time t, and this ratio is used to determine the photoresist etch rate using the calibration curve obtained as described above. it can. Once the etch rate is determined, this information is used to control the process. Time t
If the photoresist etch rate at is out of specification, ie higher or lower than a given value, then this process is problematic and should be stopped before the problem is detected. Corrects the problem without having to process a few (possibly many) wafers with the wrong processing sequence. For example, when the etch rate at time t is known, the operator adjusts the etch conditions, such as pressure, flow rate, composition of reactive gas, etc., as necessary to bring the etch rate back within a predetermined range. .

In the embodiments of the present invention, the photoresist etch rate and the contact hole etch rate have been described as examples, but the actual determination of the etch rate is not important to the present invention. Rather, the etch rate is a function of the ratio of C ₂ and SiF signals in the plasma. The acceptable range of this ratio for a particular process can be determined using the calibration method described above, and the process can be controlled in real time by monitoring this ratio. This method of the present invention has particular advantages in pattern transfer processes and is particularly useful in pattern transfer processes where the surface area of the upper layer is large compared to the unmasked portion of the lower layer, and vice versa. Is. For example, the process of the present invention can be used to monitor contact holes (or biases) in the silicon dioxide layer under the mask. C
By controlling the plasma conditions to maintain the ₂ to SiF ratio within a predetermined range, over-etching or under-etching of the silicon dioxide under the mask can be avoided. By monitoring C ₂ and SiF species periodically or continuously during the etching process, the etch process can be controlled in real time.

[0024]

BEST MODE FOR CARRYING OUT THE INVENTION The present invention relates to a process of a semiconductor device in which a contact hole is etched in silicon dioxide. In this specification, a patterned photoresist mask is an oxide layer. Shall be formed on the silicon wafer formed thereon. Only 50% or less of the wafer surface is masked,
Preferably, 50% or more is patterned so that it is covered with a photoresist mask. The process of the present invention uses OES to monitor the plasma used to transfer holes in the photoresist pattern into the underlying oxide layer.

The holes are transferred to the underlying oxide layer, usually using a fluorocarbon-containing plasma. The OES of this plasma environment can be monitored to control the plasma etching process in real time. Methods and apparatus for producing a fluorocarbon-containing plasma are well known in the art and will not be described in detail here. A suitable apparatus for etching with a fluorocarbon plasma is commercially available from Applied Materials, Inc., Sunnybell, Calif.
300 HDP Etcher (also known as Omega)
It is. The device is operated at pressures below 75 mTorr and electron densities above 10 ¹¹ / cm ³ .

The OES of the plasma can be monitored using a conventional device that separates the light exiting the plasma and entering the component wavelengths and measuring the light intensity. An example of an OES system is a spectrograph (Model CP140 from Instruments SA of E
CCD array detector (model LS2000C from Alton Instruments of Gard) connected to dison, New Jersey
en Grove, Ca). About 200nm to 850n
It has a radiation detection function in the range of m, its resolution is 2 nm or less, and its scan time is 0.1 second or less.

This OES is inadvertently monitored during that time at a selected wavelength as the plasma is used to etch the wafer. To carry out the process of the present invention, the relationship between the photoresist etch rate and / or contact hole etch rate and the OES signal strength at a given wavelength must be determined. The photoresist thickness is measured before and after being etched for a period of time to determine the photoresist etch rate. The photoresist etch rate is determined by dividing the change in thickness by time. The contact hole etch rate is
It is determined by measuring the depth of the contact hole after etching for a period of time. The depth of the contact hole is measured using SEM. The contact hole etch rate is determined by dividing the depth of the contact hole by the etching time.

The relationship between etch rate and OES signal can be determined in various ways. In one embodiment suitable for monitoring the photoresist etch rate during plasma etching, the wavelengths associated with at least two species in the plasma are monitored during etching. One of the monitored wavelengths is associated with the peak intensity of the species associated with the interaction between the plasma and the photoresist, and one of the species is associated with the intensity of the plasma. Examples of suitable species pairs are C ₂ and SiF, C ₂ and Si, C ₂ and F.
Some of these species are associated with peaks at various different wavelengths within the OES spectrum of fluorocarbon plasmas. For example, C ₂ F ₆ plasma OES spectrum used for etching the masked wafer as described above, as shown in FIG. 1 has an intensity peak at a plurality of different wavelengths associated with C _2. These spectra are obtained using the device described above. To carry out the process of the present invention, only the wavelength associated with one of each of the above species of pairs needs to be monitored. In one embodiment, the 440 nm OES signal (related to SiF species) and the 515 nm OES signal (related to C ₂ species) are monitored.

C ₂ at each selected wavelength
The signal and SiF signal intensities are monitored as a function of etching time. To obtain the ratio of the integrated signals by selecting the ratio of the C ₂ signal and the SiF signal as the ratio of the _two signals at a certain point of etching, or by integrating the value of each signal over a certain time. Determined by This ratio is related to the wafer etch rate.

This process is repeated for etching processes of various etching rates. In the embodiment of the present invention in which a calibration curve for photoresist etch rate is obtained, the etch rate is varied by changing individual process parameters such as plasma source power or chamber ceiling temperature. In one embodiment, the etch rate varies from 900 A / min to 3 by varying the source power and ceiling temperature.
It varied within the range of 400 A / min (A is an abbreviation for angstrom). It has four different source powers: 2350W, 2500W, 2650W, 28
It was done by performing a series of etchings with 00W. And for each source power, three kinds of ceiling temperatures are 240 ℃, 250 ℃, 260 ℃, respectively 3
Kinds of etching were performed. Other etching parameters, such as bias and pressure, were held constant when changing the above parameters. This provided a calibration curve that allowed a specific relationship between C ₂ and SiF signal to be related to the photoresist etch rate.

In order to monitor the etch rate of the contact holes, in one embodiment of the present invention for obtaining a calibration curve, the chamber cleaning time by O ₂ is appropriately selected during the etching process to set the contact hole etch rate for each wafer. Changed to. Other etching parameters were kept constant for each wafer. 40 with O ₂
Cleaning for seconds or less did not completely clean from wafer to wafer, and the etch rate varied from wafer to wafer as a result of chamber residue. A wash time of about 40 seconds is preferred because under these conditions the contact hole etch rate will change from wafer to wafer, but the photoresist etch rate will not. The wafer loaded cassette (25) was etched between chambers with the O ₂ chamber cleaning time described above. OE for each wafer
An S trace was obtained. Species in plasma (C ₂ or S
The signal intensity associated with iF) is measured at _two time points (t ₁ and t ₂ ) during etching, and the ratio of these two intensities is associated with the etch rate as previously described.

The parameters were changed in order to obtain this calibration curve. The fluctuation amount of each parameter is mainly a design choice, but the etching rate is about ± 50% of the target etching rate. The case where it varies is preferable. C ₂
The signal to SiF signal ratio is preferably obtained from OES of blank wafers etched under the same conditions.
By dividing this "photoresist wafer ratio" by the "blank wafer ratio", a background-free ratio is obtained, and the photoresist etch rate is obtained from this ratio.

Once the calibration curve is obtained, the OES spectrum monitors the signal strength of the spectrum at a given wavelength during the etching process and calibrates the etch rate based on the ratio of the two signals to form a plasma etch process. Is obtained for the purpose of real-time process control. Being able to determine the photoresist etch rate in real time allows for better process control because the photoresist adjusts its process conditions in real time if it is observed to be out of range. Alternatively, the subsequent wafer can be treated before etching to bring the photoresist etch rate back within the desired range.

The calibration curve thus obtained is used to control a process similar to the process conditions used to obtain the calibration curve. Changes in process conditions that require new calibration curves are known to those skilled in the art. For example, process conditions are varied (eg source power and temperature are varied) over the range of conditions that are varied to obtain a calibration curve, while other process conditions are varied to create a new calibration curve. Is required.
For example, if the preheating condition of the etching apparatus is changed, a new calibration curve is needed. Further, for example, when the preheating time used for creating the calibration curve is 5 seconds, a new calibration curve is required when the preheating time is changed to 30 seconds. However, it is not necessary when the preheating time is changed to 10 seconds. If the composition of the etchant is changed from the composition used in the monitored etching process to obtain the calibration curve, then a new calibration curve is needed for the new etchant. The surface area of the wafer covered by the photoresist is ± 10 from the amount (surface area) of the wafer etched to obtain the calibration curve.
If it fluctuates by more than%, a new calibration curve is required. Similarly, when the rf power, the pressure, the etching gas flow rate, etc. are largely changed from the etching conditions used to obtain the calibration curve, a new calibration curve is required.

[0035]

【Example】

Experimental Example 1 In order to monitor the etch rate of the photoresist during plasma etching, the wafer etched to obtain the calibration curve was a silicon wafer 100% covered with a 8000 A thick I-line photoresist layer.
The Shipley 1800 is commercially available from Shipley, Inc. of Marlborough, Massachusetts. The wafer etched to obtain the calibration curve to monitor the contact hole etch rate during the plasma etching process is a silicon wafer with a 1.02 μm thick oxide layer formed thereon. The surface of these wafers was coated with more than 98% of the coated A
It has a deep UV (ultraviolet) resist layer of RCH II®. ARCH II photoresist is OC
It is manufactured by G Corp.

These wafers are manufactured by Applied Materials Mo
It was etched by a del 5300 HDP (Omega) oxide etcher. The chamber pressure is 4 mTorr and the electron density in the plasma is between 10 ¹¹ and 10 ¹² / cm ³ . The etching apparatus includes an RF source coil, an RF bias plate that generates a plasma, and controls ion implantation energy. The temperature of the quartz side walls was maintained at 220 ° C. The chamber was preheated for 5 seconds before starting the etch with plasma source power and without bias power. The flow rate of C ₂ F ₆ plasma was 25 sccm.

A calibration curve for monitoring the photoresist etch rate was generated by varying the source power from 2350W to 2800W with a bypass power increment of 600W. The etching apparatus has a heated chamber with a top plate made of silicon and side walls made of quartz. The temperature of the silicon top plate was varied from 240 ° C to 260 ° C.

The optical emission of C _{2 at} a wavelength of 515 nm is shown in FIG. 2, where the intensity of the C ₂ peak is
A function of time for three types of wafers, a wafer 10 having a surface 100% covered by photoresist, a bare (uncovered) silicon wafer 20, and a silicon wafer 30 having an oxide layer formed thereon. As shown. The data in FIG. 2 shows that the source power is 25
O of plasma generated at 00W and a ceiling temperature of 250 ° C
Obtained from ES. Under these conditions, the oxide etch rate was 7548 A / min and the photoresist etch rate was 1965 A / min. FIG. 3 was obtained using the same etching conditions to obtain the graph of FIG.
The intensity of the SiF signal at a wavelength of 440 nm for the above three types of wafers is shown.

For each set of etching conditions (each source power used at each temperature), OES at the above wavelengths
The ratio of the C ₂ signal to the SiF signal in was determined. This ratio was plotted as a function of photoresist etch rate. The photoresist etch rate was determined by measuring the thickness of the film before and after etching, and by determining its thickness variation as a function of time. The relationship between the etch rate ratio and the photoresist of C ₂ / SiF shown by white circles 110 in FIG. C ₂ / SiF ratio was determined for bare silicon wafer was exposed to the same etching conditions. These ratios are indicated by box 120 in FIG. C ₂ / SiF ratio was determined for coated wafer with an oxide layer that was exposed to the same etching conditions. These ratios are indicated by triangles 130 in FIG. White circles 110, squares 120, triangles 130 on the same vertical axis respectively represent a wafer covered with photoresist exposed to the same etching conditions, a bare silicon wafer, and a wafer covered with an oxide layer. .

A second ratio is obtained by subtracting the background noise from the C ₂ / SiF ratio. This second ratio is the C ₂ / SiF ratio obtained from the plasma OES used to etch the photoresist wafer from the plasma OES when etching a bare silicon wafer under the same etching conditions. The obtained C ₂ / S
It is obtained by multiplying by the iF ratio. This ratio _{(C 2 / SiF) photo resist} / (C 2 / SiF)
_{Bare silicon} is plotted as a function of photoresist etch rate. The calibration curve thus obtained is shown in FIG.

Once the calibration curve is obtained as described above, it is used to monitor the etch process as follows. Monitor the OES of the plasma used in the etch patterning the photoresist pattern into the underlying silicon dioxide. The intensities of C ₂ species and SiF species at an appropriate wavelength are observed, and the ratio of these two intensities is determined. This ratio is divided by the ratio of C ₂ to SiF obtained for a blank silicon wafer obtained by exposing a blank silicon wafer to the same etching conditions. The ratio thus obtained is used to calculate the photoresist etch rate in real time with reference to the calibrated relationship between this ratio and the photoresist etch rate. The etch rate of this photoresist is within a predetermined operating window (1500 A / min to 2500
If it is within (A / min), the process proceeds as it is. If the etch rate is determined to be outside the desired range above, then appropriate remedial action is taken to correct the problem. This remedial action involves either adjusting the process conditions to bring the etch rate back within the aforementioned window, or stopping the process until the problem is resolved.

A calibration curve for monitoring the etch rate of contact holes is obtained by successively etching 25 ARCHII ™ -coated wafers in a chamber. The chamber cleaning time of O ₂ between the edge treatments was about 40 seconds. With this 40 second cleaning time, the contact hole etch rate varies from wafer to wafer because it is not sufficient to remove all of the impurities in the chamber from the previous etch process. The contact hole etch rate for each wafer can be determined by timing each etching process and using SEM to measure the depth of the contact holes after etching. As shown in FIG. 6A, the depth of a contact hole varies with the diameter of the contact hole and is a function of wafer placement in a row of wafers. FIG.
A represents the depth of the contact holes of the 1st, 11th and 21st wafers in the sequence of 25 wafers. According to FIG. 6A, for contact holes with a diameter of 0.5 μm or less, the contact holes are cumulatively shallower for each wafer in the wafer sequence.

Since the etching time is constant from wafer to wafer, FIG. 6A shows that the etch rate of the contact holes is cumulatively slower for each wafer in the wafer row. OES traces are obtained during each etch. In this example, OES is 515 nm wavelength (signal related to C ₂ species) and 440 nm (signal related to SiF species).
Obtained in both. However, to make a calibration curve, 1
Only one signal is required. Time t ₁ (0 seconds) and time t ₂ (90
Signal strength in seconds) and these two signals for each trace were related to the etch rate. FIG. 6B
As shown in, the ratio varies from wafer to wafer in a 25-wafer sequence. Since the photoresist etch rate is constant from wafer to wafer, the ratio of C ₂ signals and the ratio of SiF signals also depend on the wafer sequence.

FIG. 7 is a calibration curve obtained from the information of FIG. In particular, the ratio of C ₂ at 90 seconds to C _{2 at} 0 seconds is mapped as a function of etch rate.

[0045]

The calibration curve can be used to monitor the etch rate of the contact hole by monitoring the OES C ₂ related signal during the plasma etching process. At 90 seconds, C ₂ (90 seconds) versus C ₂ (0 seconds) is obtained. The etch rate of the contact hole can be determined using the calibration information for determining the etch rate based on this ratio. If the etch rate is within the given range, no adjustments to the process are required. However, if this etch rate is outside the specified range,
The etching process must be stopped or the etching conditions must be adjusted to bring the etch rate back within specifications.

[Brief description of drawings]

FIG. 1 C ₂ F ₆ used to etch a wafer having a photoresist mask over a silicon dioxide layer.
Diagram showing the OES spectrum of plasma

FIG. 2: 515 during plasma etching of photoresist coated wafers, oxide coated silicon wafers and bare silicon wafers.
Diagram showing the intensity trace of C ₂ emission in nm

FIG. 3: 440 during plasma etching of photoresist coated wafer, oxide coated silicon wafer and bare silicon wafer.
Diagram showing the intensity trace of SiF radiation in nm

FIG. 4 is a diagram illustrating the relationship between C ₂ and SiF emission ratios and etch rates for photoresist-coated wafers, oxide-coated silicon wafers, and bare silicon wafers.

FIG. 5 is a plot of the etch rate of a photoresist coated wafer and the C ₂ and SiF emission ratio divided by that of a bare silicon wafer.

FIG. 6A shows a difference in depth of contact holes observed for each wafer when the O ₂ cleaning time for each wafer is insufficient to completely clean the chamber. FIG. Diagram showing the ratio of the intensities of signals (C ₂ and SiF) at two time points and the change of this ratio from wafer to wafer

FIG. 7 shows a calibration curve obtained from the data shown in FIG. 6 for controlling the etch rate of contact holes.

 ──────────────────────────────────────────────────続 き Continuation of the front page (71) Applicant 596077259 600 Mountain Avenue, Murray Hill, New Jersey 07974-0636 U.S.A. S. A. (72) Inventor Susan Krady McNevin United States, 07974 New Jersey, New Providence, Springfield Avenue 1584

Claims (12)

[Claims] 1. A step of: (A) disposing a substrate in a reaction chamber for plasma generation, wherein the substrate has a silicon dioxide layer formed on at least a part of its surface, and a silicon dioxide layer is formed on the silicon dioxide layer. A resist mask and at least a portion of the silicon dioxide layer is exposed; (B) introducing a fluorocarbon-containing gas into the reaction chamber; and (C) exposing the exposed silicon dioxide layer to the substrate. Generating a plasma in the reaction chamber for removal from the surface; (D) monitoring the optical emission of the plasma and measuring the brightness intensity at a wavelength associated with the species in the plasma; E) a step of calculating a value from the measured luminance intensity, and (F) a step of comparing the calculated value with a value in a predetermined range. And the predetermined value is predetermined in relation to the etch rate of one of the group of photoresist mask and silicon dioxide layer, and (G) controlling the process based on the comparison. A method for monitoring a semiconductor element manufacturing process, comprising: 2. The method according to claim 1, wherein the control step (G) is performed by adjusting a process parameter when the calculated value is not within a predetermined range. 3. A brightness intensity of a pair of species in the plasma is monitored, a first species of the pair of species is associated with photoresist-plasma interaction, and a second species is Related to the intensity of the plasma, the calculated value is the ratio of the intensity associated with the first species and the intensity associated with the second species during etching, and the predetermined value is the first 3. The method of claim 2, wherein there is a predetermined ratio of species-related intensity to the second species-related intensity, each of the predetermined ratios being associated with an etch rate. 4. The method of claim 3, wherein the first species is C ₂ and the second species is selected from the group consisting of SiF, Si, F. 5. The first species is C ₂ monitored at 515 nm wavelength and the second species is SiF monitored at 440 nm wavelength.
the method of. 6. The method of claim 3, wherein the predetermined range of ratios is determined from a calibration curve relating a ratio to a photoresist etch rate for a process condition. 7. The method of claim 6, further comprising the step of: (H) adjusting the plasma etch conditions if the ratio is not within the predetermined range of ratios. 8. The intensity intensity of the species in the plasma is monitored and the calculated value is the ratio of the intensity associated with the species at time t ₁ to the intensity associated with the species at time t ₂ . Times t ₁ and t ₂ are times during the etching duration, the predetermined value is a predetermined ratio of species-related intensities at times t ₁ and t ₂ , and the predetermined ratio is an etch rate. The method of claim 2, wherein the method is associated with. 9. The method of claim 8, wherein the controlled etch rate is a contact hole etch rate. 10. The method of claim 9, wherein the seed is selected from the group consisting of C ₂ and SiF. 11. The method of claim 10, wherein the predetermined range of ratios is determined from a calibration curve relating a ratio to a contact hole etch rate under certain process conditions. 12. The method of claim 11, further comprising the step of: (H) adjusting the plasma etch conditions if the ratio is not within the predetermined range of ratios.