Method of anisotropic dry etching of thin film semiconductors

Abstract

A microwave ECR plasma etching method and apparatus employs a combination reactive gas medium supplied to a microwave excitation ECR plasma chamber coupled to a treatment chamber containing a Group II-VI sample to be etched. A reactive gas plasma is formed containing reactive species in the plasma chamber and the Group II-VI sample is irradiated with a beam of reactive species formed from the reactive gas plasma. The reactive gas medium is selected from the group consisting of a hydrogen halogenate, a mixture of a halogen gas and an inert gas, a mixture of a halogen gas and hydrogen gas, a mixture of a halogen gas, an inert gas and hydrogen gas, and a mixture of a halogen gas and nitrogen gas.

Description

CROSS REFERENCE TO RELATED APPLICATIONS

This application contains subject matter related to the subject matter of patent application Ser. Nos. 07/484,786 and 07/484,222 both filed Feb. 22, 1990 and assigned to the same assigned herein and incorporated herein by reference thereto.

BACKGROUND OF THE INVENTION

This invention relates generally to methods for etching treatment of semiconductor structures and more particularly to microwave plasma anisotropic dry etching of thin film semiconductor structures and, specifically, to microwave generated plasma methods of etching Group II-VI compound semiconductor thin films or substrates employing a halogen reactive gas medium to chemically etch patterned semiconductor layers or substrates of semiconductor structures without significant surface or structural damage.

The manufacture of semiconductor devices and structures usually requires the selective etching of particular patterns in specific layers thereof. Previously, most etching of this type was wet etching, i.e., employing wet chemical materials that are applied to the patterned surface. In wet etching of Group II-VI compound semiconductors, the etching solutions primarily used are a solution of sodium hydroxide; hydrochloric acid; and a compound solution of nitric acid, hydrochloric acid, and water. These etching solutions are used at appropriate temperatures and with appropriate composition ratios as known in the art to achieve desired etching rate. However, the use of wet etching is isotropic in nature so that the resultant linewidth and pattern resolution of the wet etched pattern is not the same as the originally desired resist pattern. Further wet etching requires additional treatment steps of rinse and drying. Because of these mentioned factors, semiconductor structure yields are not uniform or high.

A common problem relative to wet etching is the lack of reproducibility. A predetermined etching rate cannot be achieved unless the temperature and the composition of etching solution are closely controlled. Also, where the etching solution contains a volatile material, the composition of the solution significantly changes over time. Therefore, etching rate of the wet etchant at the time when the solution is prepared and etching rate of the wet etchant at some later time are dramatically different. Furthermore, in wet etching, the pattern to be etched in a semiconductor layer or film cannot be formed to be the same as pattern of the mask because the etching is isotropic in nature, including side etching, resulting in a larger overall etched pattern. Also, the etching of patterns of processed sections are limited, for example, formation of vertical sections or deep vertical grooves with large length/breadth ratio is difficult to obtain.

Wet etching of Group II-V group compound semiconductors present more problems than wet etching of other semiconductors, such as, Group III-V group compound semiconductors. For example, in the case wherein etching of ZnSe is performed employing a hydrochloric acid and nitric acid etching solution, the etching solution penetrates into ZnSe and it is very difficult to completely remove etchant even with long periods of rinsing. As a result, there is a substantial degradation of film characteristics. Also, in the case where etching of ZnSe or ZnS.sub.x Se.sub.1-x (0<x.ltoreq.1) is performed in the solution of NaOH, the surface morphology worsens extremely, therefore, these compounds are not particularly suitable for precision wet etching compared to Group III-V compound semiconductors. In the case where hydrochloric acid is employed as an etchant, the etching rate is very slow and, therefore, HCl is not practical for use in the fabrication and selective etching of Group II-VI compound semiconductors.

For many years, sputtering and ion milling or etching have been employed as dry etching techniques to accomplish, among other things, etching of semiconductor layers. In general, processing is carried out by establishing a DC or rf generated plasma with an inert atmosphere, such as Ar. An example of such processes are disclosed in U.S. Pat. No. 4,622,094. These techniques generally involve the formation of a plasma and the physical removal of materials from the semiconductor surface due to bombardment of the surface with ions. However, the accuracy of these techniques has left much to be desired, particularly relative, for example, to improvements to pattern resolution, surface morphology, attained anisotropy, etching depth and reduced mask erosion.

On the other hand, in the case of such dry etching, for example, ion etching employing an inert gas medium, such as Ar, in order to enhance the etching rate to a level of practical utilization, it is necessary to increase the plasma discharge power. However, this, in turn, results in substantial damage to a semiconductor materials.

More recently, there has been an increased interest in reactive dry etching techniques because these techniques, as compared to the above mentioned previous techniques, promise better pattern resolution in submicron large scale integration providing a higher degree of circuit density with improvements in surface morphology, increased anisotropy, lower thermal stress due to lower temperature processing, higher plasma densities at lower pressures, enhanced etching rates, enhanced selectivity ratio, deeper etching capability and reduced mask erosion. Further, they eliminate the need for the above mentioned post etching treatment steps employed in wet etching and improved to a great degree the accuracy can be achieved in the resultant linewidth and pattern resolution. As a result, semiconductor structure yields may be made more uniform and higher. These techniques generally involve the chemical removal of materials from the semiconductor surface or a combination of chemical and physical removal from the semiconductor surface comprising atoms or molecules of etched materials and products of the reaction between surface molecules and the reactive gas species.

Dry etching techniques include reactive ion etching (RIE), ion beam assisted etching (IBAE) and hot jet etching (HJE), and reactive ion beam etching (RIBE), such as microwave plasma dry etching, each of which involves a chemically reactive vapor or gaseous species, for example, comprising a halogen, such as F.sub.2, Br.sub.2 or Cl.sub.2, in a vapor phase compound. In RIE, the sample or target to be etched is placed on a cathode in an electric field established between an anode and cathode in the presence of a selected flux of a chemically reactive species that reacts with atoms or molecules on the surface of the sample. The potential applied between the anode and cathode is sufficient to ionize atoms or molecules in the gas as well as produce radicals. The positively charged ions produced in the plasma are attracted to the cathode and upon impact physically remove or etch away material from the sample surface. The reactive species will also chemically react with atoms or molecules on the surface of the sample which are also removed by the incident ions on the surface of the sample. As an example of RIE, see U.S. Pat. No. 4,640,737.

In the case of reactive ion etching (RIE) employing a reactive gas, such as BCl.sub.3, damage to the semiconductor materials is, to a degree, less compared to that of ion etching. In any case, damage to the surface of semiconductor materials under this etching treatment is still major and not acceptable. In order to reduce the damage, the gas pressure of the etching system may be raised while lowering the discharge power. However, the ion sheath width and the mean free path of the ions and neutral particles become almost the same, causing the beam of ions to lose directionality thereby increasing the potentiality of isotropic etching to occur. Thus, RIE provides a significant drawback to dry etch processing particularly for Group II-VI compound semiconductors.

In IBAE, a combination of ions from an inert gas, e.g., Ar.sup.+, from an ion beam source and a flux of chemically active species, e.g., F or Cl, are directed to the sample and by control of the ion beam and the reactive species, a controlled anisotropic etching can be carried out. In the case of HJE, there is no ion beam employed and a flux of reactive radicals is formed and directed onto the sample. See, for example, the articles of M. W. Geis et al.: "A Novel Anisotropic Dry Etching Technique", Journal of Vacuum Science Technology, Vol. 19(4), pp. 1390-1393, Nov./Dec., 1981; "Hot-Jet Etching of Pb, GaAs, and Si", Journal of Vacuum Science Technology, Vol. B5(1), pp. 363-365, Jan./Feb., 1987; and "Summary Abstract: Etching With Directed Beams of Ions or Radicals", Journal of Vacuum Science Technology, Vol. A5(4), pp. 1928-1929, Jul./Aug. 1987. Also, see U.S. Pat. No. 4,874,459 relative to a modified IBAE method as well as a summary of other reactive dry method techniques mentioned in the background of this reference.

In RIBE, the source of ions (e.g., Cl.sup.+) and radicals (e.g., Cl*) is generally formed in and extracted out of a separate chamber and accelerated via an ion extraction grid or electrode into the etching chamber. See, for example, the article of K. Asakawa et al., "GaAs and GaAlAs Equi-Rate Etching Using a New Reactive Ion Beam Etching System", Japanese Journal of Applied Physics, Vol. 22(10), pp. L653-L655, October, 1983. Electron Cyclotron Resonance (ECR) microwave plasma source is employed which provides for higher efficiency in plasma generation and higher generation of reactive species achieving improved anisotropy and higher etching rates. Other example are found in U.S. Pat. Nos. 4,795,529; 4,778,561; 4,609,428; 4,859,908 and 4,734,157.

Thus, reactive dry etching processes generically provide a source of reactive species in the form of either reactive ions, e.g. Cl.sup.+, or reactive radicals, e.g. Cl*, or a combination of reactive ions and radicals forming a reactive flux, e.g. Cl.sup.+ and Cl*, or a source of reactive species assisted by other ions, e.g. Cl.sup.+ and/or Cl* in combination with Ar.sup.+, that are generated, focussed and/or accelerated to the sample target to provide a chemical action at the sample surface with surface molecules and sputter or otherwise remove reaction products from the sample surface via the outlet affluent.

It is of importance to note that all of the foregoing references relating to various dry etching techniques are methods that have specifically evolved for the purpose of etching Group III-V materials, e.g., GaAs and AlGaAs. The techniques have not been generally applied to Group II-VI compound semiconductors, such as ZnSe, ZnS, or ZnS.sub.x Se.sub.1-x, because the developed treatments, as reported in these references, have not been designed for these compounds and their attempted application according to their specific teachings would provide etching rate that are not of a practical level and would result in damage to the crystalline structure without good anisotropy. Further, problems persist in the utilization of these etching techniques of the prior art, particularly in the case of Group II-VI compound semiconductors wherein both selective wet etching and dry etching techniques have been used employing a mask comprising an insulation film, such as photoresist, SiO.sub.2, or the like. In particular, good anisotropy has not been obtained, particularly in connection with masking techniques, and good selectivity ratio has not been achieved.

It is an object of this invention to provide an improved method of reactive ion beam etching (RIBE).

It is another object of this invention to provide a reactive ion beam etching (RIBE) method particularly suitable for Group II-VI compound semiconductors.

It is another object of this invention to provide modified reactive ion etch methods from high density plasmas including a combination of reactive gases to form the reactive gas medium providing for high anisotropy, enhanced etching properties and much improved surface morphology, particularly for Group II-VI compound semiconductors.

It is a further object of this invention to manufacture semiconductor structures employing Group II-VI compound semiconductors with excellent reproducibility by providing etching methods of Group II-VI compound semiconductors with high reproducibility, excellent practical use with no or negligible damage to etched semiconductor materials and with a capability of producing anisotropically patterns previously not realizable in the prior art.

SUMMARY OF THE INVENTION

According to this invention, a microwave ECR plasma etching method and apparatus employes a combination reactive gas medium supplied to a microwave excitation ECR plasma chamber coupled to a treatment chamber containing a Group II-VI sample to be etched. A reactive gas plasma is formed containing reactive species in the plasma chamber and the Group II-VI sample is irradiated with a beam of reactive species formed from the reactive gas plasma. The reactive gas medium is selected from the group consisting of a hydrogen halogenate, a mixture of a halogen gas and an inert gas, a mixture of a halogen gas and hydrogen gas, a mixture of a halogen gas, an inert gas and hydrogen gas, and a mixture of a halogen gas and nitrogen gas.

The present invention demonstrates superior effects by providing ion and/or radical species produced from a plasma formed from a combination reactive gas medium containing at least one halogenate to etch Group II-VI compound semiconductors, such as ZnSe, ZnS, or ZnS.sub.x Se.sub.1-x, also recognizing the fact that these compound semiconductors. have higher ionicity than Group III-V compound semiconductors. Thus, in order to employ a RIBE method, it is necessary to take into consideration this ionicity if good anisotropy and minimal crystalline damage are to be attained. For example, if the particular disclosed treatment of Asakawa et al. for RIBE were to be utilized in connection with Group II-VI compound semiconductors, unacceptable levels of damage would generally occur.

We have discovered that since the ionicity of the constituents comprising Group II-VI compound semiconductors are high, they react much more readily with active halogen elements, such as Cl, compared to III-V compound semiconductors. As indicated from the work of Asakawa et al. and others, the etching mechanism of the RIBE method involves (1) the chemical adsorption of halogen ions or radicals to the surface of the sample to be dry etched, (2) chemical reaction with the sample constituents and (3) physical removal of reaction products from the sample surface constituents or materials. However, in the case of Group II-VI compound semiconductors, the removal of reaction products from the surface materials is by vapor pressure. As a result, the etching rate is large for a lower ion energy level, e.g., employing an acceleration or lead voltage in the range of 0 V to 600 V, preferably below 300 V, compared to ion energy levels employed in the case of Group III-V compound semiconductors. With the application of lower ion energy levels while achieving a comparable or acceptable etching rate for practical use relative to etching Group II-VI compound semiconductors, correspondingly, damage to the crystalline structure is minimized or negligible.

Further, the constituents of Group II-VI compound semiconductors form reaction products which, in most cases, produce two halogen reaction products. For example, in the case of RIBE dry etching of ZnSe with Cl ions, the reaction products are ZnCl.sub.x and SeCl.sub.x, such as, ZnCl.sub.2 and SeCl.sub.2. From a chemical point of view, these reaction products are not very active. Therefore, after they are formed at the sample surface and evaporate, they will be swiftly carried away due to the vapor pressure of the system and will not side etch the side walls of the masked pattern being etch in the sample because of their low reactive nature and their swift removal. As a result, clean anisotropic etching in the direction of the applied ion beam is achieved. Compared to the reaction products formed relative to Group III-V compound semiconductors employing the RIBE method and Cl ions, such as GaCl.sub.2, GaCl.sub.3 and the like, after these reaction products evaporate, they may still producing side etching of the side walls of the pattern being etch in the sample because of their high reactive nature prior to their removal from the immediate region of the sample surface. Thus, the application of the RIBE method to Group II-VI compound semiconductors under the proper controlled conditions provides for enhanced anisotropic etching compared to Group III-V compound semiconductors.

Employing reactive ion beam etching produced by microwave excitation and an ECR plasma of Group II-VI compound semiconductors, as compared to the wet etching technique or dry etching, such as, IRE or RIBE methods of the prior art, provides for superior reproducibility and controllability in the process. In particular, damage to semiconductor materials is dramatically reduced. Further, by controlling the ion beam and the shape of the etching mask, precision process of, for example, grooves with taper or vertical walls having perpendicular cross section or diagonal groove is fully possible. Furthermore, since etching rate can be controlled by choosing from various reactive gas and the stability of plasma can be achieved, etching depth can be controlled with good reproducibility. Also, good morphology, such as, extremely smooth etching side surfaces, and achievement of the exact, desired etching pattern of the mask pattern are obtainable.

Under the methods of this invention, the etching process can be carried out under low pressure and low density conditions thereby reducing to a minimum damage to the crystalline structure. Thus, reproducibility between batches in dry etching of Group II-VI compound semiconductors is dramatically increases, thereby rendering simplified mass production of semiconductor integrated circuits and devices.

Other objects and attainments together with a fuller understanding of the invention will become apparent and appreciated by referring to the following description and claims taken in conjunction with the accompanying drawings.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a schematic cross sectional illustration of a microwave plasma etching apparatus employed in connection with the dry etching of semiconductor structures employing a reactive gas medium.

FIGS. 2A and 2B are schematic cross sectional illustrations of a ZnSe semiconductor sample, respectively before and after dry etching, which is employed in several embodiments for the purpose of illustrating the application of the method of this invention as practiced relative to the apparatus shown in FIG. 1.

FIGS. 3A and 3B are graphic illustrations of the photoluminescence spectra for a ZnSe layer respectively before and after dry etching in accordance with the method of this invention.

FIG. 4 is a graphic illustration of the relationship of the etching rate relative to the gas pressure in the practice of the method comprising this invention.

FIG. 5 is a graphic illustration of the relationship of the etching rate relative to the microwave input power in the practice of the method comprising this invention.

FIG. 6 is a graphic illustration of the relationship of the etching rate relative to the lead voltage in the practice of the method comprising this invention.

FIG. 7 is a graphic illustration of the relationship of the etching rate as a function of the gas mixture ratio comprising a second embodiment of a reactive gas medium employed in the dry etching method of this invention.

FIG. 8 is a graphic illustration of the relationship of etching rate as a function of the gas mixture ratio comprising a third embodiment of a reactive gas medium employed in the dry etching method of this invention.

FIG. 9 is a graphic illustration of the relationship of etching rate as a function of the gas mixture ratio comprising a fourth embodiment of a reactive gas medium employed in the dry etching method of this invention.

FIG. 10 is a graphic illustration of the relationship of etching rate as a function of the gas mixture ratio comprising a fifth embodiment of a reactive gas medium employed in the dry etching method of this invention.

DESCRIPTION OF THE PREFERRED EMBODIMENTS

Reference is now made to FIG. 1 wherein there is shown a cross sectional schematic view of a microwave plasma etching apparatus 10 employed in connection with this invention. Apparatus 10 comprises an etching chamber 12 coupled to a sample preparation chamber 14 by means gate valve interlock 18. A highly pure, halogen gaseous element, e.g., Cl.sub.2, is employed in the practice of this invention and is highly reactive so that these two chambers 12 and 14 are maintained under separate environments and valve 18 is opened when passing a sample 28 to and from chamber 12 from and to preparation chamber 14 via sample transfer rod 15. Chamber 14 has its own UHV evacuation system 17. The air locked etching chamber 12 is maintained under high vacuum condition via an UHV evacuation system 13.

Etching chamber 12 is directly open to ECR plasma chamber 16 which produces a high intensity plasma which is excited by a combination microwave, introduced into chamber 16 via microwave waveguide 24 and quartz plate window (not shown), and a magnetic filed provided by cylindrical magnetic coil 20. Also, a highly pure halogen gas, such as Cl.sub.2, forming one component of the reactive gas medium is introduced via inlet 25. Another component of the reactive gas medium, such as, Ar, H or N, or combinations thereof may be introduced in inlet 27. For the purposes of initial explanation and background, however, reference will be made only to the use of Cl.sub.2 via inlet 25 and inlet 27 will be considered closed.

The microwave power from waveguide 24 creates discharges in the Cl.sub.2 gas in chamber 16 and a plasma is generated which contains electrons, ions and radicals. Electrons, generated by the microwave, repeatedly collide with gas molecules and atoms while performing cyclotron movement within chamber 16 caused by the symmetric magnetic field produced by coil 20. When the intensity of the generated magnetic field is, for example, 875 Gauss, the revolution cycle matches with the microwave frequency, for example, 2.45 GHz, and a cyclotron resonance phenominum is created and resonance is achieved in electron absorption of microwave energy. Therefore, discharge continues even when gas pressure is low, thereby providing a high plasma density and a long and useful life for the reactive gas medium. Also, because the electrons and ions are in cyclotronic motion and are gather toward the center of chamber 16 because of the established electric field distribution, any sputtering effect produced by the ions on the side walls of plasma chamber 12 is minimal and, thus, provides a pure plasma which is unidirectional and perpendicular to the planar extent of sample 28 in etching chamber 10.

The ions and radicals generated in plasma chamber 16 generally have linear motion while the electrons have cyclotron motion and are retained within plasma chamber 16. However, the ions and radicals are accelerated as a beam by extraction grid or electrode 26 toward sample 28 in a path perpendicular to the planar extent of sample 28 when applied acceleration or lead voltage is applied to electrode 26. The ions and radicals impinging sample 28 chemically react with the surface material of sample 28 and form volatile products which are removed from region of sample 28 by the sputtering action of the ions and evacuated from chamber 10 via evacuation system 13.

Sample 28 is supported on sample holder 30 in chamber 10. Holder 30 may be rotated 360.degree. about a vertical axis by manipulator 32 so that the direction of the beam emanating from chamber 16 can be made incident at an angle relative to the planar extent of sample 28. Also, holder 30 may be provided with a heater/cooling section (not shown) for the purpose of heating or cooling sample 28 to a controlled predetermined temperature.

FIG. 2A illustrates a cross sectional view of a sample 28 prior to etching in accordance with the method of this application. Sample 28 comprises substrate 34 of a Group II-VI compound semiconductor, e.g., ZnSe. While this example, as well as subsequently discussed samples, illustrates a Group II-VI compound semiconductor substrate, the sample may also be a Group II-VI compound semiconductor layer or a plurality of layers. A mask 36 comprising, for example, a positive type photoresist material, is formed on substrate 34 and a desired pattern 35 is formed in mask 36 employing conventional photolithography. Since mask 36 is formed by conventional photolithography, the cross sectional pattern of mask 36 will have a natural taper shape, as illustrated at 37.

A method of dry etching for sample 28 utilizing apparatus 10 was performed as follows. The results relative to this sample are referred in later discussions as the "original embodiment". Chlorine gas was employed as the reactive gas medium and this medium was 99.999% pure chlorine. The gas pressure of the system was 1.0.times.10.sup.-1 Pa, the microwave inlet power was 100 W, the lead voltage was 500 V, the sample temperature was 25.degree. C., and the direction of ion beam irradiation (indicated by arrows 39 in FIGS. 1 and 2B) was perpendicular relative to the planar extent of sample 28.

FIG. 2B illustrates a cross sectional view of sample 28 after etching. The etching rate of ZnSe was approximately 600 .ANG./min. while the etching rate of the positive type photoresist, comprising mask 36, was approximately 200 .ANG./min. This mask etching rate is applicable in the case where the mask had been post baked for 30 minutes at 120.degree. C. The etching rate varies to a certain degree depending upon the type of photoresist employed and the baking conditions, e.g., temperature and time, employed, particularly the temperature. A sufficiently good selectivity ratio is obtained if the photoresist baking temperature is in the range of about 120.degree. C. to 250.degree. C. Since the pattern of etching mask 36 has a taper shape 37 and sputtering effect of the beam also causes slight etching of etching mask 36. This is evident in FIG. 2B in that mask 36 has been reduced in thickness and its taper 37A is more pronounced compared to FIG. 2A. Also, the etched region of ZnSe in sample 28 is also taped as shown at 33. Thus, even where the ion beam is irradiated perpendicularly relative the planar extent of ZnSe sample 28, a perpendicular cross section cannot be achieved in the ZnSe material, i.e., anisotropic etching is not obtained in spite of the normal incident of ion beam 39 on sample 28. On the other hand, for practical use in etching applications, the etching rate achieved in this embodiment is sufficient. Further, the uniformity in distribution of the etching rate on the etched material across the extent of the sample surfaces in the case where the substrate sample size was 20 mm.times.20 mm was about or less than .+-.5%, and the surface morphology of the sample was about the same as the surface morphology before etching treatment.

FIG. 3A and FIG. 3B illustrate the comparison of photoluminescence spectra of ZnSe substrate 34 before and after the etching under the above mentioned conditions. FIG. 3A is the photoluminescent spectra before etching and FIG. 3B is the photoluminescent spectra after etching. The deep level luminescence 29 and the band edge luminescence 31 are illustrated in each of these figures. The lower the deep level 29, the less crystalline damage that occurs to the sample material. Since the ratio of relative intensity of luminescence of the band edge 31 relative to the deep level 29 is approximately 50 both before and after etching, it is clear that there is essentially no damage resulting to semiconductor substrate 34.

Examples representing the effective etching conditions for practical applications relative to this original embodiment for use in processing Group II-VI compound semiconductor structures are now described.

Qualitatively speaking, the etching rate increases as the gas pressure increases, as determined from experimental data illustrated in FIG. 4 wherein the lead voltage was maintained around 500 V and the microwave power was maintained around 80 W. However, when the gas pressure is too high, plasma discharge will not occur. Even when discharge occurs, such as at or above 1 Pa, the ion sheat width and the mean free path of the ions and radicals become substantially the same thereby causing the ion beam to lose its directionality. Therefore, these high pressure conditions are not suitable for etching treatment. When the gas pressure is too low, such as at or below 1.times.10.sup.-3 Pa, the etching rate progresses too slow, therefore, it is not suitable for practical applications.

Table 1 shows the relationship of the etching rate of ZnSc relative to the gas pressure when the microwave inlet power is 100 W, the lead or acceleration voltage is 500 V and pure chlorine gas is employed as the etching gas medium. As can be seen from Table 1, as the gas pressure becomes lower, so does the etching rate and lateral (side) or isotropic etching is negligible or nonexistent. The term, "slight", in Tables 1 and 4 may be expressed by a guideline indicative of the degree of lateral or side etching. If the mean depth of etching is A and the maximum depth of resultant lateral etching is B, then the degree of side etching is expressed as B/A. "Slight" lateral etching condition means that the ratio of B/A.ltoreq.0.1. There is no practical problems relative to the use of etched II-VI compound semiconductors in the manufacture of semiconductor devices or IC structures with a ratio of B/A.ltoreq.0.1.

As determined from experimental data illustrated in FIG. 5 wherein the gas pressure was maintained around 1.times.10.sup.-1 Pa and the lead voltage was maintained around 450 V, etching rate increases as the microwave inlet power increases because the plasma density becomes more intense as microwave excitation is increased. However, when the microwave inlet power is too high, the plasma temperature becomes too high deforming electrode 26. Also, temperature control of sample 28 becomes difficult because the sample temperature also rises due to the microwave power extending into etching chamber 12. Good etching results are, therefore, achieved with microwave power in the power range of 1 W to 1 KW.

TABLE 1______________________________________Gas Pressure Etching Rate Lateral Etching(Pa) (.ANG./min) Condition______________________________________5 .times. 10.sup.-3 100 None1 .times. 10.sup.-2 500 None5 .times. 10.sup.-2 650 None1 .times. 10.sup.-1 700 None5 .times. 10.sup.-1 800 Slight1.0 1,000 Slight______________________________________

Table 2 shows the dependency characteristics of the etching rate of ZnSe relative to the microwave inlet power when the etching medium comprises pure chlorine gas, the gas pressure is 1.times.10.sup.-1 Pa, and the lead voltage is 400 V. As can be seen in Table 2, the etching rate increases with the microwave inlet power.

TABLE 2______________________________________Microwave InletPower (W) Etching Rate (.ANG./min)______________________________________ 50 200100 700150 900200 1,000400 1,300600 1,9001,000 2,100______________________________________

As illustrated in FIG. 6 wherein the gas pressure was maintained around 1.times.10.sup.-1 Pa and the microwave power was maintained around 200 W, the etching rate increases as the lead voltage increases. However, if the voltage is too high, such as at or above 1 KV, physical sputtering of the sample becomes too strong causing substantial damage to the substrate crystalline structure. When lead voltage is 0 V, and the substrate temperature is approximately 200.degree. C., etching by radical species of Cl is accomplished and isotropic etching of the sample will occur.

Table 3 shows the dependency characteristics of the etching rate of ZnSe relative to the lead voltage when the etching gas is pure chlorine gas, the gas pressure is 1.times.10.sup.-1 Pa, and the microwave inlet power is 200 W. To be noted is that as the lead voltage increases, so does the etching rate and a point is reached, i.e., at about 700 V, when damage begins to occur to the crystalline structure of the sample.

TABLE 3______________________________________Lead Voltage (V) Etching Rate (.ANG./min) Substrate Damage______________________________________200 500 None300 700 None400 850 None500 1,000 None600 1,200 None700 1,300 Slight800 1,500 Slight1,000 1,600 Yes______________________________________

The foregoing original embodiment and discussed conditions for a optimum reactive dry etching method are principally the subject matter of patent application Ser. No. 07/484,786 and have been presented here as background and for the purposes of comparison with several embodiments of this application which are now to be discussed. In all subsequent embodiments, the same type of sample 28 disclosed in FIG. 2A was employed in apparatus 10.

A first embodiment is now explained wherein hydrogen halogenate was employed as an reactive gas medium. In this embodiment, etching was performed under conditions wherein HCl was employed as the reactive gas medium via inlet 25 of apparatus. The gas pressure of the system was 1.0.times.10.sup.-1 Pa, the microwave inlet power was 100 W, the lead voltage was 500 V, the sample temperature was maintained at 25.degree. C., and the ion beam irradiation direction was perpendicular relative to the planar extent of the sample substrate. The sample prepared was the same as sample 28 in FIG. 2A. The etching rate of ZnSe was found to be approximately 900 .ANG./min. The etching rate of positive type photoresist was approximately 280 .ANG./min when the post bake condition of the photoresist had been at 120.degree. C. for 30 minutes.

As recalled in connection with the original embodiment discussed employing only pure chlorine gas as the reactive gas medium but treated under identical conditions as the case of this first embodiment, the etching rate was 600 .ANG./min. However, the etching rate of ZnSe increased by 1.5 times when employing a HCl gas medium. Thus, when HCl gas is employed as the reactive gas medium, a lower plasma density and a lower acceleration or lead voltage are sufficient to provide a comparable etching rate of 600 .ANG./min, which lower values of applied energy further minimize damage to the semiconductor crystalline structure of the sample. Also, the photoluminescent spectra of this first sample embodiment was examined before and after dry etching and substantially no damage to the semiconductor material was observed due to the etching treatment and was the case for the photoluminescent spectra in FIGS. 3A and 3B relative to the original embodiment ZnSe sample.

An evaluation was also made of the contaminant content of ZnSe surface after etching where HCl was employed as the reactive gas medium. This evaluation was made employing the auger electron spectroscopy method and a comparison was made with the original embodiment method where Cl.sub.2 was employed as the reactive gas medium. It was discovered that the amount of molecular contaminants of oxygen and chlorine were less in the case where a HCl gas medium was employed compared to the case where a Cl.sub.2 gas medium was employed. The reason for this is believed to be that oxygen and chlorine contaminants remaining on the ZnSe surface during etching are removed by hydrogen radicals formed in the HCl plasma, which concurrently clean the surface of the sample of such contaminants.

The uniformity of the distribution of the etching rate when the substrate sample was 20 mm.times.20 mm was about or less than .+-.5%. Also, the surface morphology was substantially the same as the surface morphology before treatment.

Examples representing the effective etching conditions for practical applications relative to Group II-VI semiconductor compound processing in an HCl gas medium have been determined as follows. The etching rate increases as the gas pressure increases as shown from the results in Table 4 below. However, when gas pressure is too high, plasma discharge will not occur. When discharge occurs, such as at or above 1 Pa, the ion sheath width and the mean free path of ions and radicals become substantially the same and the ion beam loses directionality. Therefore, such high pressure conditions are not suitable for etching treatment. When gas pressure is too low, such as at or below 1.times.10.sup.-3 Pa, then, the etching rate becomes too low and, therefore, the etching treatment is not suitable for practical applications. Thus, these conditions are about the same as in the case of the original embodiment.

Table 4 shows the variation of the ZnSe etching rate relative to the gas pressure when the microwave inlet power is 100 W, the lead voltage is 500 V, and HCl gas is employed as etching gas medium. In comparison with these conditions for the original embodiment set forth in Table 1, it can be seen that corresponding higher etching rates are achieved in the first embodiment over the original embodiment for comparable pressure conditions with lateral (side) etching conditions about the same.

TABLE 4______________________________________Gas Pressure Etching Rate Lateral Etching(Pa) (.ANG./min) Condition______________________________________5 .times. 10.sup.-3 150 None1 .times. 10.sup.-2 700 None5 .times. 10.sup.-2 810 None1 .times. 10.sup.-1 900 None5 .times. 10.sup.-1 1,100 Slight1.0 1,300 Slight______________________________________

Also, the etching rate increases as the microwave inlet power increases because the plasma density increases as the microwave excitation becomes more intensified. However, when the microwave inlet power is too high, the plasma temperature increases causing electrode 26 to become deformed due to the intense heat. Further, the temperature control of the etching process becomes difficult because the substrate temperature is increased due to the microwave power extending into etching chamber 12 of apparatus 10. Good etching results were achieved with the microwave power in the range of 1 W to 1 KW.

Table 5 shows the dependency of the ZnSe etching rate relative to the microwave inlet power when the etching gas medium is HCl, the gas pressure is 1.times.10.sup.-1 Pa and the lead voltage is 400 V. As in the case of the original embodiment relative to Table 2 in comparison with the first embodiment relative to Table 5, the etching rate increases with the microwave inlet power. However, corresponding higher etching rates are achieved in the first embodiment over the original embodiment for comparable pressure conditions.

As the lead voltage increases, the etching rate increases. However, if the lead voltage is too high, such as at or above 1 KV, physical sputtering of the sample becomes too intensified causing substantial damage to its crystalline structure. When the lead voltage is 0 V, and the substrate temperature is approximately 200.degree. C., etching is brought about by radical species in the plasma and isotropic etching will occur.

TABLE 5______________________________________Microwave Inlet Etching RatePower (W) (.ANG./min)______________________________________ 50 380100 900150 1,200200 1,450400 1,700600 2,1001,000 2,300______________________________________

Table 6 shows the dependency characteristics of the etching rate relative to the lead voltage when the etching gas medium is HCl gas, the pressure is 1.times.10.sup.-1 Pa, and the microwave inlet power is 200 W. To be noted, in comparison with Table 3 for the original embodiment, the etching rate relative to the first embodiment increases at a faster rate with comparable increases in the lead voltages.

TABLE 6______________________________________Lead Voltage (V) Etching Rate (.ANG./min) Substrate Damage______________________________________200 550 None300 900 None400 1,450 None500 1,600 None600 1,900 None700 2,100 Slight800 2,150 Slight1,000 2,150 Yes______________________________________

Similar results are readily expected to be achieved when other hydrogen halogenates, such as HBr, HF or the like, are employed as etching gas.

In a second embodiment, a mixture comprising a halogen gas and an inactive gas was employed as the reactive gas medium. Etching was performed under conditions wherein a mixture of 99.999% pure chlorine and 99.9999% pure argon was employed as the reactive gas medium respectively via inlets 25 and 27 of apparatus 10. The gas pressure of the system was 1.0.times.10.sup.-1 Pa, the microwave inlet power was 100 W, the lead voltage was 500 V, the sample temperature was 25.degree. C., and the ion beam irradiation direction was perpendicular relative to the planar extent of sample 28. In this particular case, the Ar mixture ratio was Ar 25%. The etching rate of ZnSe was approximately 800 .ANG./min. The etching rate of the positive type photoresist was approximately 200 .ANG./min, when the post bake condition is set for 30 minutes at 120.degree. C. The uniformity of distribution of the etching rate across the substrate when the substrate was 20 mm.times.20 mm is about .+-.5 % or less. The surface morphology was substantially the same as the surface morphology before treatment.

FIG. 7 shows the relationship between the argon mixture ratio and the etching rate when employing chlorine gas and argon gas as a combination reactive gas medium. The highest etching rate was achieved when the argon mixture ratio was at 25%. This is because carbon and other foreign particles present on the etching surface were effectively removed by argon ion sputtering thereby aiding in the overall advancement and progression of the chemical reactive etching by the Cl ions. Further, it was found that the presence of argon gas in the dry etch mixture reduced the occurrence of abnormal discharges occurring in the plasma. Also, the reproducibility of etching rate was stable at .+-.2% between processed sample batches.

In a third embodiment, dry etching was performed under conditions wherein a mixture of 99.999% pure chlorine gas and 99.9999% pure hydrogen gas was employed as the reactive gas medium and were supplied respectively via inlets 25 and 27 of apparatus 10. The gas pressure of the system was 1.0.times.10.sup.-1 Pa, the microwave inlet power was 100 W, the lead voltage was 500 V, the sample temperature was 25.degree. C., and the ion beam irradiation direction was perpendicular relative to the planar extent of sample 28. In this case, the mixture ratio of hydrogen gas was 25%. The etching rate of ZnSe was approximately 750 .ANG./min. The etching rate of positive type photoresist was approximately 200 .ANG./min when the post bake condition were set for 30 minutes at 120.degree. C. The uniformity of distribution of the etching rate when the substrate was 20 mm.times.20 mm was .+-.5%. The surface morphology was substantially the same as the surface morphology before treatment.

FIG. 8 shows the relationship of the hydrogen gas mixture ratio and the etching rate when employing chlorine gas and hydrogen gas as a combination reactive gas medium. The highest etching rate was achieved when the hydrogen gas mixture ratio was at 25%. This is because foreign materials, such as carbon, present on the etching surface were effectively removed by the cleaning action of the hydrogen ions impinging on the surface of sample 28 thereby aiding in the overall advancement and progression of the chemical reactive etching by the Cl ions. Further, as in the previous case relative to argon gas, it was found that the presence of hydrogen gas in the dry etch mixture reduced the occurrence of abnormal discharges occurring in the plasma. Also, the reproducibility of etching rate was stable at .+-.1% between processed sample batches. Also, lateral or side etching is negligible. Further, due to the cleaning effect of the hydrogen ions, the etched side wall surfaces of sample 28 nearly have a mirror finish. Therefore, the employment of this reactive gas medium is very effective for semiconductor device processing applications.

In a fourth embodiment, dry etching was performed under conditions wherein a mixture of 99.999% pure chlorine gas, 99.9999% pure hydrogen gas and 99.9999% pure argon gas were employed as the reactive gas medium. The hydrogen and argon gas mediums were provided via inlet 27 of apparatus 10. The gas pressure of the system was 1.0.times.10.sup.-1 Pa, the microwave inlet power was 100 W, the lead voltage was 500 V, the sample temperature was 25.degree. C., and the ion beam irradiation direction was perpendicular relative to the planar extent of sample 28. In this particular embodiment, the mixture ratio of hydrogen gas 50% and argon gas 25%. The etching rate of ZnSe was approximately 900 .ANG./min. The etching speed of positive type photoresist was approximately 200 .ANG./min when the post bake condition is set for 30 minutes at 120.degree. C. The uniformity of distribution of the etching rate across the sample when the substrate was 20 mm.times.20 mm was about .+-.1% or less. The surface morphology was substantially the same as the surface morphology before treatment.

FIG. 9 shows the relationship of hydrogen gas mixture ratio and etching rate when mixture of chlorine gas, hydrogen gas, and argon gas was employed as the reactive gas medium. The highest etching rate was achieved when hydrogen gas mixture ratio was 50%. This is due to the effective cleaning action of hydrogen ions on the substrate surface and the effective removal of carbon and other such contaminants from the etching surface by the sputtering action of argon ions thereby aiding in the overall advancement and progression of the chemical reactive etching by the Cl ions. Further, it was found that the presence of argon gas in the dry etch mixture reduced the occurrence of abnormal discharges occurring in the plasma. Also, the reproducibility of etching rate was stable at .+-.1% between processed sample batches. Further, highly smooth etching side walls 33 were achieved when employing this gas medium, which was also true for the side walls 37A of sample patterned mask 36.

In a fifth embodiment, etching was performed under conditions wherein a mixture of 99.999% pure chlorine gas and a 99.9999% pure nitrogen gas was employed as the reactive gas medium. The gas pressure of the system was 1.0.times.10.sup.-1 Pa, the microwave inlet power was 100 W, lead voltage was 500V, the sample temperature was 25.degree. C., and the ion beam irradiation direction was perpendicular relative to the planar extent of sample 28. In this particular case, the mixture ratio of nitrogen gas was at 25%. The etching rate of ZnSe was approximately 900 .ANG./min. The etching rate of positive type photoresist was approximately 200 .ANG./min when the post bake condition is set for 30 minutes at 120.degree. C. The uniformity of the distribution of the etching rate across the sample when the substrate was 20 mm.times.20 mm was about .+-.5% or less. The surface morphology was substantially the same as the surface morphology before treatment.

FIG. 10 shows the relationship of nitrogen gas mixture ratio and etching rate wherein a mixture of chlorine gas and nitrogen gas was employed as the reactive gas medium. The highest etching rate was achieved when nitrogen gas mixture ratio was at 25%. This is because foreign materials, such as carbon, present on the etching surface were effectively removed due to the cleaning action of the nitrogen ions thereby aiding in the overall advancement and progression of the chemical reactive etching by the Cl ions. Further, it was found that the presence of nitrogen gas in the dry etch mixture reduced the likelihood of abnormal discharges occurring in the plasma. Also, the reproducibility of etching rate was stable at .+-.2% between processed sample batches.

While the invention has been described in conjunction with several specific embodiments, it is evident to those skilled in the art that many further alternatives, modifications and variations will be apparent in light of the foregoing description. Thus, the invention described herein is intended to embrace at such alternatives, modifications, applications and variations as fall within the spirit and scope of the appended claims.

Claims

1. In a method of reactive ion beam etching of Group II-VI compound semiconductors comprising the steps of

employing a combination reactive gas medium supplied to a microwave excitation ECR plasma chamber coupled to a treatment chamber containing a Group II-VI sample to be etched,

forming a reactive gas plasma containing reactive species in said plasma chamber and

irradiating said sample with a beam of reactive species formed from said reactive gas plasma, said combination reactive gas medium selected from the group consisting of a mixture of a halogen gas other than fluoride in combination with a gas selected from the group consisting of an inert gas, hydrogen gas, nitrogen gas, and a combination of an inert and hydrogen gas.

2. The method of reactive ion beam etching of Group II-VI compound semiconductors of claim 1 wherein said group selection is a combination of Cl.sub.2 and Ar.

3. The method of reactive ion beam etching of Group II-VI compound semiconductors of claim 1 wherein said group selection is a combination of Cl.sub.2, H.sub.2 and Ar.

4. The method of reactive ion beam etching of Group II-VI compound semiconductors of claim 1 wherein said group selection is a combination of Cl.sub.2 and N.sub.2.

5. In a method of reactive ion beam etching of Group II-VI compound semiconductors comprising the steps of

employing a combination reactive gas medium supplied to a microwave excitation ECR plasma chamber coupled to a treatment chamber containing a Group II-VI sample to be etched,

forming a reactive gas plasma containing reactive species in said plasma chamber and

irradiating said sample with a beam of reactive species formed from said reactive gas plasma, said combination reactive gas medium comprising a mixture of about 75% Cl.sub.2 and about 25% Ar.

6. In a method of reactive ion beam etching of Group II-VI compound semiconductors comprising the steps of

employing a combination reactive gas medium supplied to a microwave excitation ECR plasma chamber coupled to a treatment chamber containing a Group II-VI sample to be etched,

forming a reactive gas plasma containing reactive species in said plasma chamber and

irradiating said sample with a beam of reactive species formed from said reactive gas plasma, said combination reactive gas medium comprising a mixture of about 75% Cl.sub.2 and about 25% H.sub.2.

7. In a method of reactive ion beam etching of Group II-VI compound semiconductors comprising the steps of

employing a combination reactive gas medium supplied to a microwave excitation ECR plasma chamber coupled to a treatment chamber containing a Group II-VI sample to be etched,

forming a reactive gas plasma containing reactive species in said plasma chamber and

irradiating said sample with a beam of reactive species formed from said reactive gas plasma, said combination reactive gas medium comprising a mixture of about 25% Cl.sub.2, about 50% H.sub.2 and about 25% Ar.

8. In a method of reactive ion beam etching of Group II-VI compound semiconductors comprising the steps of

employing a combination reactive gas medium supplied to a microwave excitation ECR plasma chamber coupled to a treatment chamber containing a Group II-VI sample to be etched,

forming a reactive gas plasma containing reactive species in said plasma chamber and

irradiating said sample with a beam of reactive species formed from said reactive gas plasma, said combination reactive gas medium comprising a mixture of about 75% Cl.sub.2 and about 25% N.sub.2.