Single component developer for use with ghost exposure

Abstract

A method for developing copolymer photosensitive resists wherein a single solvent is used in conjunction with a puddle develop tool. The copolymer resist is ZEP 7000 and the developer is ethyl 3-ethoxy propionate (EEP).

Description

FIELD OF THE INVENTION

[0002] This invention relates to the making of photolithography masks for semiconductor manufacturing and particularly to semiconductor technologies having minimum dimensions of 0.25 microns and less. More particularly, the invention relates to methods of proximity effect correction of photographic masks using a single component developer.

BACKGROUND OF THE INVENTION

[0003] In the manufacture of semiconductor devices, the ever present need to continue increasing of the density of image segments continually causes the seeking of new and more efficient techniques for processing semiconductor devices. The manufacture of semiconductor masks is no exception and the move to sub-0.25 micron technologies is imminent.

[0004] The instant invention relates to chain scission copolymer photo resists of the type formed of 1:1 copolymerization of -chloromethacrylate and -methyl styrene, available as ZEP-nnn resists from Nippon Zen, Japan. These resists are sensitive to light and electron beam exposure which makes them particularly useful in mask manufacture. Various compositions of resist are available based on their molecular weight. For example, ZEP-520, having a molecular weight of about 50,000, or ZEP-7000, having a molecular weight of about 333,000. Once exposed, light sensitive portions of the copolymer are rendered soluble in organic solvents to produce good relief images.

[0005] A typical application would to apply the resist in a solvent to a substrate, bake to remove the solvents, expose the resist via electron beam and then develop using a commercially supplied developer comprising one of the following developers, mixtures of diethyl ketone and diethyl malonate and single component solvents such as tolulene, xylene and alkyl esters of acetic acid such as amyl acetate or hexyl acetate. Following development, coated substrates are rinsed in a solvent such as 2-propanol.

[0006] The use of electron beams for lithography suffers from electron scattering in the substrate. This scattering is termed proximity effect. Many algorithms have been developed to correct for proximity effects. For raster scan electron beam lithography systems, GHOST proximity effect correction (PEC) is typically employed. GHOST PEC was developed in the early 1980's (G. Owen, et al, “Proximity Effect Correction for Electron Beam Lithography by Equalization of Background Dose,” J. Appl. Phys., 54(6), 1983). GHOST PEC consists of an additional exposure using the reverse polarity of the original pattern to be imaged into the substrate.

[0007] Typical problems with the use of GHOST correction are the increased process bias due to the additional GHOST exposure, the increased susceptibility to pinholes due to thinning of the resist film in the areas of the pattern where GHOST correction is applied, and potential degrade in critical dimension (CD) uniformity due to a reduction in resist contrast. To compensate for the increased process bias colder develop temperatures (16 deg C.) and shorter develop times have been employed (see “A 180 nm mask fabrication process using ZEP 7000, multipass gray, GHOST, and dry etch for Mebes 5000”, M. Lu, et al, SPIE Proc. 3376, Sept. 1998) using commercially available developers. In order to compensate for the additional resist loss in the areas which are processed using GHOST PEC, thicker resists are used (4000 ang used in previously quoted paper).

SUMMARY OF THE INVENTION

[0008] When previously practiced methods were attempted to be implemented in the manufacture of semiconductor masks of the sub-0.25 micron range, non uniform characteristics of the resulting photomasks were observed. This non uniformity can exist in both non-GHOST and GHOST processes. This was particularly true for the CD variation across a mask blank. In addition, for electron beam lithography forward scattering increases with resist thickness. So compensating for resist loss when undertaking GHOST PEC degrades resolution.

[0009] It has been discovered that the evaporative characteristics of the developer environment varied across the surface of the wafer, leading to non-uniform development due to varying concentration of reagents in the developer solution.

[0010] It is an object of the instant invention to provide an more uniform environment for the development of copolymer resists in order to provide more uniform and reproducible results in the critical dimensions in both non-GHOST and GHOST processes.

[0011] In accordance with the invention an active developer which comprises a substantially nonvolatile single component solvent is used for developing copolymer resists on semiconductor mask blanks. The preferred developer is ethyl 3-ethoxy propionate (EEP).

[0012] These and other objects of the invention will become more apparent when viewed in light of the following more particular description of the preferred embodiment of the invention.

BRIEF DESCRIPTION OF THE DRAWINGS

[0013] FIGS. 1 a and 1b are bar charts illustrating the critical dimension uniformity for various temperatures, develop times and doses.

DESCRIPTION OF THE PREFERRED EMBODIMENT

[0014] A mask blank coated with 3,000 A of ZEP 7000, an electron beam photoresist comprising a 1:1 copolymer of -chloromethacrylate and -methyl styrene, was exposed using an electron beam tool (accelerating voltage of 10-75 keV) at a dose ranging from 8-70 μC./cm2. The exposure dose used for this resist depends on the accelerating voltage of the electron beam tool. For 10 keV, the dose is 8-12 μC/cm2. For 75 keV, the dose is 60-70 μC/cm2. After electron beam exposure, the ZEP resist on the mask blank is developed in a single substrate spin tool using a solvent develop chemistry. The develop process was performed using either a puddle process, a spray process, or a combination puddle/spray process. The preferred tool for implementing the development process is a multiple puddle process in which developer is dispensed to cover the mask blank in sufficient quantity to allow puddle etching, see U.S. Pat. No. 3,953,265 to Hood issued Apr. 27, 1976.

[0015] During the develop process, a solvent develop chemistry is dispensed on the mask either via a spray or stream and is followed by a solvent rinse and high speed spin dry.

[0016] The ZEP resist manufacturer (Nippon Zen Inc.) recommended solvent develop chemistry for ZEP, consisting of a mixture of diethyl ketone and diethyl malonate, be used. Nippon Zen sells this developer premixed under the Trade name ZED 500. The use of the ZED 500 developer resulted in a Critical Dimension (CD) non-uniformity across the mask of 42 nm (3 sigma). This is an unacceptable result for fabrication of advanced photomasks. CD uniformity of 30 nm (3 sigma) or smaller is required for photomasks to be used in fabricating mask for technologies having 0.25 um and smaller device geometries. The 42 nm CD non-uniformity had a very consistent radial error characteristic seen across the 130 mm×130 mm quality area of the plate. The images at the outer edges of the plate were consistently smaller than the images in the center of the plate.

[0017] Experiments were performed to try to understand the main source of the high CD non-uniformity, which led to the following previously unrecognized results and conclusions: The temperature of the actual developer puddle was measured on the mask lank, and it was found that the puddle was 2 degrees Centigrade (°C.) cooler at the edges of the plate compared to the center of the plate. This 2 degree difference was found to be due to evaporative cooling of the developer that was occurring at the edges of puddle near the edge of the plate. Other work showed that the active developer component in the ZED 500 was the diethyl ketone which is quite volatile and has a low boiling point (102° C.). It was hypothesized that the highly volatile diethyl ketone component was evaporating at the edge of the plate (causing the 2° C. center to edge temperature gradient). The lower developer temperature combined with the decreased concentration of the active developer component at the edge of the plate caused a slower develop rate on images at the edge of the plate and consequently a smaller image size at the edge of the plate compared to images at the center of them plate.

[0018] Based on this information, a new single component higher boiling point and lower volatility develop chemistry was proposed. After evaluating a number of organic solvents, a puddle develop process using ethyl 3-ethoxy propionate (EEP) (CAS Registry Number 763-69-9) was defined that resulted in CD uniformity of 9 nm (3 sigma) across the plate. EEP has a boiling point of 166° C.

[0019] Many ZEP-coated plates have since been processed with this new EEP develop chemistry with CD uniformities consistently being achieved in the 8-15 nm range. In addition to giving significantly improved CD uniformity performance, the EEP developer causes less thinning of the unexposed ZEP 7000 resist which consequently results in a higher contrast process compared to the ZED 500 develop chemistry.

[0020] Based on the success of the single component developer of the invention, it can be said that any two component solvent develop system in which one component is more volatile than the other will suffer evaporative cooling and concentration changes due to this difference, resulting in CD non-uniformity issues of the same or similar magnitude as observed by the inventors.

[0021] Accordingly, by selecting a single evaporative component having a substantially high boiling point, one can ensure that the develop environment is controlled sufficiently to provide good CD control across a mask blank.

[0022] Additional single solvent developers were evaluated and, although each met some of the preferred characteristics exhibited by EEP, each had other drawbacks which rendered them unacceptable for use in a manufacturing environment.

[0023] For example, cellosolve acetate proved to provide acceptable physical results but because it is a teretogen, is believed to be unsafe for routine use in manufacturing.

[0024] Carbitol proved to be unacceptable due to its tendency to cause swelling, and thus distort the images, in the undeveloped patterns. Butyl carbitol proved to be unacceptable since it failed to develop exposed resist at all.

[0025] Methyl 3-methoxypropionate, a lower alkyl of EEP, proved to be unsuitable as it removed unexposed resist at a high rate.

[0026] The specific optimum developer for higher molecular weight copolymer resists might be other single solvent materials which will remain stable in the developing environment proposed.

[0027] When using the single component developer in GHOST PEC the process discussed above needs to be slightly altered. The preferred mask process is to use a ZEP-7000 resist and the single component developer at a colder temperature and reduced develop time for use with GHOST PEC in electron beam lithography. This process shows reduced loss of resist thickness after the develop step when using GHOST. By reducing the resist lost in the areas of the pattern where GHOST PEC occurs, thinner resist films can be used without encountering clear defects or pinholes. CD mean-to-target is achieved while using the same process bias as the non-GHOST process. CD uniformity achieved is also equivalent to the non-GHOST process.

[0028] The process uses a single solvent developer (ethyl ethoxy proprionate) at 17 to 18 degree C. with 3000 angstrom of ZEP-7000 resist. Other ZEP 7000 users use 4000 angstroms of resist when using GHOST due to resist loss in the develop step, as well as resist loss during etching. The resist loss at develop is limited to 220 angstroms compared to 500 angstroms for standard processing. Typical resolution improvement from the use of a 4000 angstrom vs. 3000 angstrom resist film at 10 kV accelerating voltage is approximately 70 nm (see “System architecture choices for an advanced mask writer (100-130 nm)”, V. Chakarian, et al, FIG. #2, SPIE 3873, Sept. 1999).

[0029] An AME-500 develop tool is used and the mini-environment temperature to a setup temperature equal to the developer temperature. If plate was not cooled towards developer temperature CD uniformity is degraded. FIGS. 1a and 1b are bar charts illustrating the critical dimension uniformity for various temperatures, develop times and doses. As illustrated 12 to 15 nm 3 sigma uniformity is typical for the conditions of 17 to 18 degrees Centigrade (°C.), dose of 9.5 to 11 uC/cm2 at 10 kV, and 3000 angstrom ZEP resist thickness.

[0030] While the invention has been described in terms of a specific embodiment, it will be understood by those skilled in the photolithography art that other single components of low volatility can be used to develop other copolymer resist systems.

Claims

1. A method of developing exposed regions of a copolymer photosensitive resist for proximity effect corrections including the steps of: exposing portions of the copolymer resist to an reverse image pattern at a reduced dose; developing the exposed portions of the resist by providing an atmosphere of a single solvent having low volatility.

2. The method of claim 1 wherein the copolymer photosensitive resist is a copolymer of -chloromethacrylate and -methylstyene with a thickness 3000 angstroms.

3. The method of claim 2 wherein the single solvent is ethyl 3-ethoxy propionate at 17 to 18° C.

4. The method of claim 1 wherein the copolymer resist is formed on the surface of a photolithography mask blank.

5. The method of claim 4 wherein the step of developing is carried out in a puddle etch environment at a setup temperature of 17 to 18° C.

6. The method of claim 2 wherein the molecular weight of the copolymer is in excess of about 100,000.

7. The method of claim 2 wherein the step of exposing is accomplished with an exposure dose of between 9.5 to 11 uC/cm2 at 10 kV.

8. The method of claim 5 wherein the puddle develop times are between 36 to 46 seconds.

9. A method of developing exposed regions of a copolymer photosensitive resist for proximity effect corrections including the steps of: exposing portions of the copolymer resist to an reverse image pattern at a reduced dose wherein the copolymer photosensitive resist is a copolymer of -chloromethacrylate and -methylstyene with a thickness 3000 angstroms; developing the exposed portions of the resist by providing an atmosphere of a single solvent having low volatility single solvent is ethyl 3-ethoxy propionate at 17 to 18° C.

10. The method of claim 9 wherein the copolymer resist is formed on the surface of a photolithography mask blank.

11. The method of claim 9 wherein the molecular weight of the copolymer is in excess of about 100,000.

12. The method of claim 11 wherein the step of exposing is accomplished with an exposure dose of between 9.5 to 11 uC/cm2.

13. The method of claim 12 wherein the puddle develop times are between 36 to 46 seconds.

14. A method of developing exposed regions of a copolymer photosensitive resist for proximity effect corrections including the steps of: exposing portions of the copolymer resist to an reverse image pattern at a reduced dose wherein the copolymer photosensitive resist is a copolymer of -chloromethacrylate and -methylstyene with a thickness 3000 angstroms; developing the exposed portions of the resist be providing an atmosphere of a single solvent having low volatility single solvent is ethyl 3-ethoxy propionate at 17 to 18° C. and is carried out in a puddle etch environment at a setup temperature of 17 to 18° C.