`
`J. M. Moran and D. Maydan
`
`Citation: Journal of Vacuum Science and Technology 16, 1620 (1979); doi: 10.1116/1.570256
`
`View online: http://dx.doi.org/10.1116/1 .570256
`
`\fiew Table of Contents: http://avs.scitation.orgltoc/jvs/16/6
`
`Published by the American Vacuum Society
`
`Page 1 of 6
`
`TSMC Exhibit 1045
`
`TSMC v. IP Bridge
`IPR2016—01379
`
`
`
`
`
`
`
`
`High resolution, steep profile resist patterns
`
`
`
`
`
`
`J. M. Moran and D. Maydan
`Bei'i' Laboratories Murray Hi”. New Jersey 07974
`
`
`
`
`
`
`
`(Received 13 June 1979; accepted 25 September 1979)
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`High resolution and steep profile patterns have been generated in a 2.6 pm thick organic
`layer which conforms to the steps on a wafer surface and is planar on its tep. This thick
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`organic layer (a photoresist in the present experiments) is covered with an intermediate layer
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`of SiOz and a top, thin layer of x-ray or photoresist. After exposure and development of the
`
`top resist layer,
`the intermediate layer is etched by CHF3 reactive ion etching. The thick
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`organic layer is then etched by 02 reactive ion etching. Submicron resolution with essentially
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`vertical walls in the thick organic material was achieved. The technique is also applicable to
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`It
`photo and electron lithography.
`reduces the need for thick resist patterns for the
`
`
`
`
`
`
`
`
`
`
`
`
`
`lithography step and, at the same time, ensures high resolution combined with good step
`coverage.
`
`
`
`
`
`
`PACS numbers: 07.68. + rn. 07.85. + n, 31.60.Dq
`
`
`
`
`
`
`
`
`I.
`
`
`INTRODUCTION
`
`
`
`
`
`
`
`
`
`
`
`
`With the advent of higher resolution lithographies, such as
`
`x rayL2 and electron beam, as well as increased resolution
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`irom optical lithography, there is a demand for using these
`capabilities for making VLSI devices. Often, the ability to
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`achieve good Iinewidth control, high resolution, and good step
`coverage tend to be mutually exclusive. Good step coverage
`
`
`
`
`
`
`
`
`
`requires thick resist while high resolution is more easily oh-
`
`
`
`
`
`
`
`
`
`
`
`
`FIG.
`] X-ray resist thickness variation as it step covers 1.0 pm of p—giass
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`flowed over a 1.0 pm] high poly-silicon feature. Resist varies from 0.8 pm on
`
`
`
`
`
`
`
`
`
`
`
`
`
`top of the feature to 1.7 pm in the valley between features [SEM profile
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`photo}.
`
`
`1620
`
`
`
`J. Vac. Sci. Technol.. 16(6), NovJDee. 1979
`
`
`
`
`
`
`
`
`
`Page 2 of 6
`
`
`
`
`
`
`
`
`
`
`
`
`tained in thin resista This is true for all resists, both positive
`and negative.
`
`
`With any resist, the ideal conditions for obtaining high
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`resolution and good linewidth control are a flat surface and
`a thin resist (0.3—0.4 pm). The flat surface ensures that the
`
`
`
`
`
`
`
`
`
`
`
`spun-on resist has very little variation in thickness and, as a
`
`
`
`
`
`
`
`
`
`
`
`result, there should be little variation in resist linewidth.
`
`
`
`
`
`
`
`
`
`However, resist linewidth variations will occur when lines
`
`
`
`
`
`
`
`
`traverse a step. Figure 1 illustrates the variation in resist
`
`
`
`
`
`
`
`
`
`
`BU22-5355ITBIDS1620-0580100
`(Q 1980 American Vacuum Society
`1520
`
`
`
`
`
`
`
`FIG, 2., Result of plasma etching p-glass when the resist is too thin on top of
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`[he feature. The holes in [he p-glass (the resist has been stripped away) in-
`
`
`
`
`
`
`
`
`
`
`
`
`
`dicalt‘ resist erosion on top of the feature. (SEM photo).
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`Page 2 of 6
`
`
`
`1521
`
`
`
`J. M. Moran and D. Maydan: High resolution. steep profile resist patterns
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`1621
`
`
`
`REACTIVE ION ETCH
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`40003 PHOTO.
`
`ELECTRON, 0n
`
`X—RAY RESIST
`
`
`2 - 3 pm
`
`THICK
`
`PHOTDRESIST
`
`
`
`5:02
`
`PHO T0RESiSI' VLSI SUBSTRATE
`
`SCHEMATIC PRESENTATiON OF THE VARIOUS STEPS
`REQUIRED TO DEFINE A STEEP PROFILE fiESIST
`
`
`
`
`
`PATTERN.
`
`
`
`
`
`
`
`
`
`
`
`
`FIG. 4. Processing sequence for three level. high resolution patterning.
`
`
`
`
`
`
`
`
`
`
`
`||. EXPERIMENT
`
`
`
`
`A 2.6 pm thick layer of photorcsist {HI’R—204—Htint
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`Chemical Co.) serving as the thick organic layer, was spun on
`a silicon wafer. The intermediate layer of 010 pm of silicon
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`Ethan—«i
`iii-3 BM W M 915
`
`FIG. 3. SEM profile of a very thick resist covering the same features as shown
`
`
`
`
`
`
`
`
`
`
`
`
`
`in Fig. 1. Note the flat surface due to the very thick resist spun-on the sur-
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`face.
`
`
`
`
`FIG. 5. Resulting pattern. which is all thick HPR resist, after using the triIeveI
`
`
`
`
`
`
`
`
`
`
`
`
`
`processing described in Fig. 4. Top pattern was done with x-ray resist. The
`
`
`
`
`
`
`
`
`
`
`
`
`walls are 90" and there is no Iinewidth loss (SEM photo}.
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`J. Vac. Sci. Technol.. Vol. 16, No. 6, NovJDec. 1979
`
`
`
`
`
`
`
`
`
`
`
`Page 3 of 6
`
` Illlilll
`
`
`
`
`
`
`
`
`
`
`
`
`PHOTOHESIST
`
`
`
`
`
`
`
`
`LSl SUBSTRATE
`
`
`
`
`—V
`
`thickness, of the spun-on resist, as it covers a 1.0 pm thick
`
`
`
`
`
`
`
`
`
`
`
`
`phosphosilicate glass flowed over a 1.0 pm thick polysilicon
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`step. The darker area is the resist which measures 0.8 pm thick
`
`on top of the feature and 1.7 run thick in the valley between
`
`
`
`
`
`
`
`
`
`
`
`
`
`features.
`
`Patterning a wafer with resist thickness variations as shown
`
`
`
`
`
`
`
`
`in Fig. 1 will result in both a difference in feature size as well
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`as a very thin resist on top of the step. In the case of a chemical
`etch, the thin resist does not present a problem. However, if
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`plasma etching is used, the resist is relied on as a mask. For a
`thin resist, the result of plasma etching the phosphosilicate
`
`
`
`
`
`
`
`
`
`glass is shown in Fig. 2. The horizontal lines are p-glass which
`
`
`
`
`
`
`
`
`
`
`
`
`has been etched. The slant line and vertical line in the center
`
`
`
`
`
`
`
`
`
`
`
`
`of the photo are polysilicon steps. Note that there are two
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`round holes in the p-glass where it goes over the step. This is
`due to erosion of the thin resist on the step during plasma
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`etching of the p~glass The resist has been stripped completely
`awayr to reveal this problem.
`
`
`
`
`
`
`
`
`
`
`
`
`Presented hereris a method for generating high resolution,
`steep profile resist patterns by first preparing a flatter surface
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`on the wafer. This is done hy spinning a thick organic layer
`of resist which conforms with the wafer surface and is planar
`
`
`
`
`
`
`
`
`
`
`
`on its top. Figure 3 illustrates this with a 2.6 pm thick layer
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`of HPR—ZOS spun down on the same topography shown in
`Fig. l. The curved, lighter area is the p-glass and the rough
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`darker area on top is the thick resist. The thick layer is then
`patterned using an intermediate masking layer of plasma
`
`
`
`
`
`
`
`
`deposited Si02 and a thin top layer of resist. The result is that
`
`
`
`
`
`
`
`
`
`
`
`
`a very thick resist material can be patterned with submicrori
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`resolution and steep sidewalls comparable with those in pos-
`itive photoresist.
`
`
`
`Page 3 of 6
`
`
`
`1822
`
`
`
`J. M. Moran and D. Maydan: ngh resolution, steep profile resist patterns
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`1622
`
`
`
`TABLE I. High resolution. steep profile resist patterns.
`
`
`
`
`
`
`
`
`
`
`A dva ntages
`
`
`
`
`
`
`
`
`
`
`
`“O‘V‘PP‘E‘JF‘
`
`
`
`Planar surface for resist patterning
`
`
`
`
`Excellent step coverage
`
`
`
`Good iinewidth control
`
`
`
`Thinner resist can be used for better resolution
`
`
`
`
`
`
`
`
`Eliminates standing waves and scattering in photolithogra phy
`
`
`
`
`
`
`Reduces proximity effects in electron lithography
`
`
`
`
`
`
`Minimal resist erosion during substrate etch by plasma or ions
`
`
`
`
`
`
`
`
`
`Disadvantages
`
`
`
`
`it In iui
`
`
`
`
`
`I. Requires extra processing steps
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`dioxide was plasma deposited, at 200°C, on the photoresist
`
`
`
`
`
`
`
`
`
`and then a 1.0 pm thick layer of chlorine based negative x—ray
`
`
`
`
`
`
`
`
`
`
`
`resist4 was deposited on top of the oxide. A schematic pref
`
`
`
`
`
`
`
`
`
`
`sentation of the processing sequence is shown in Fig. 4.
`The top layer of x-ray resist was exposed and developed to
`
`
`
`
`
`
`
`
`
`
`
`
`a final thickness of 0.45 pm using an x-ray exposure tool.”
`
`
`
`
`
`
`
`
`
`
`With the x-ray resist as a mask, the Sl02 was reactive ion
`
`
`
`
`
`
`
`
`
`
`
`
`etched with a CHFS gas. The pattern was then transferred into
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`the thick organic (resist) layer using reactive ion etching, with
`pure 02 gas forming the plasma and the SiOg acting as the
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`mask. The rf power density was 0.30 VV/ctn2 and the time
`
`
`
`
`
`
`
`
`
`
`required to etch the resist was 20 min. Figure 5 shows the
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`resultant pattern which is 2.5 gm high and has a trench width
`
`
`
`
`
`
`
`
`
`
`
`of 1.5 ,um. The photos are taken with a scanning electron
`
`microscope at a very steep angle so as to clearly Show the wall
`
`
`
`
`
`
`
`
`
`
`
`
`structure of the resist. Note that the walls are perpendicular
`
`
`
`
`
`
`
`
`
`
`
`UlllrtluF—l
`H-iEilifl
`
`FIG. 6. A snhmicron window, in 2.1 pm thick resist using the trilevel tech—
`
`
`
`
`
`
`
`
`
`
`
`
`
`nique and x-rsy patterning (SEM photo).
`
`
`
`
`
`
`
`FIG. '1'. Thick resist pattern obtained using optical patterning on a Perkin—
`
`
`
`
`
`
`
`
`
`
`Eirner projection printer for the top layer. (SEM photo).
`
`
`
`
`
`
`
`
`
`
`FIG. 8. Suhrnicron trench for a Ca As gate obtained usng it Kaslrr [Oil step
`
`
`
`
`
`
`
`
`
`
`
`
`
`and repeat camera for the top layer. (SEM photo]
`
`
`
`
`
`
`
`
`
`
`
`
`J. Vac. Sci. Technol.. Vol. 16, No. 6, NovJDec. 1979
`
`
`
`
`
`
`
`
`
`
`
`Page 4 of 6
`
`Page 4 of 6
`
`
`
`1623
`
`
`
`J. M. Moran and D. Maydan: High resolution, steep profile resist patterns
`
`
`
`
`
`
`
`
`
`
`
`
`
`1623
`
`
`
`
`
`FIG. 9 Lines and spaces obtained using the trilevel technique with a negative
`
`
`
`
`
`
`
`
`
`
`
`electron resist (CMC) for the top patterning layer {SEM photo)
`
`
`
`
`
`
`
`
`
`
`
`
`
`and there is very little undercut. The oxide is still on top of the
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`organic (resist) layer and its thickness loss during the reactive
`
`
`
`
`
`
`
`
`
`
`ion etch was less than 0.02 ,urn. Table I lists the advantages and
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`disadvantages of this technique herein denoted as the tri—level
`technique.
`
`
`
`
`
`
`
`
`
`
`Figure 6 is an SEM photograph of the high resolution ca—
`
`
`
`
`
`
`
`
`
`pability of the combination of the tri-lcvcl structure and a high
`
`resolution x~ray resist“ A 05 .um window has been transferred
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`to a 2.1 ,urn thick photoresisl layer with no perceptable un—
`
`dercut apparent
`
`
`
`
`
`
`Ii. DISCUSSION
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`The tri-level technique for obtaining high resolution fea—
`tures in thick resist can be applied to other lithographies be~
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`sides x-ray lithography since there are advantages for optical
`
`
`
`
`
`and electron-beam lithography as well.
`
`
`
`
`
`
`
`
`
`In the case of optical lithography, the thick underlying layer
`
`
`
`
`
`
`
`
`of resist and the intermediate layer of Si02 significantly re—
`
`duce reflection from the wafer surface and, as a result, reduce
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`standing wave problems. The flat surface of the thick organic
`
`
`
`
`
`
`
`
`
`
`
`layer keeps scattering down and the top resist can be made
`
`
`
`
`
`
`
`
`
`
`thin for high resolution. There is also the advantage of having
`the same surface to coat and expose on for all process levels
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`which means a fixed exposure time, developing time, and a
`consistent surface for the top resist to be coated on every time.
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`Figures 7 and 8 are SEM photos of results obtained from an
`
`
`
`
`
`
`
`
`
`
`optically exposed top resist layer. Figure 7 is a 2.0 our wide
`
`
`
`
`
`
`
`
`
`
`
`feature, 2.6 pm high, that is routinely achieved on a device
`
`
`
`
`
`
`
`
`wafer using a Perkin—Elmer projection printer and PIPE—204
`J. Vac. Sci. Technol., Vol. 16, No. 6, NovJDec. 1979
`
`
`
`
`
`
`
`
`
`
`
`
`Page 5 of 6
`
`WWW
`ire-HM J’EIII 3.45
`
`
`
`
`
`Whirl—4
`lit-i PM :5 Hi
`
`r33
`
`(b)
`
`in) to ,urn lines and spaces in thick resist, Iismg x ray tor the top
`FIG. 10.
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`layer The lines are traversing a 1.0 [all] high. periodic grating to demonstrate
`
`
`
`
`
`
`
`
`
`
`
`
`good step cow-rage with no linewidth variation over steps (HEM photo). [bi
`
`
`
`
`
`
`
`
`
`
`
`
`Profile- nl the wall of resist [or one of the I 0 am lines showing conformal
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`coating of the. wafer topology at the resist bottom and a flat surface at its top
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`[SEM photo).
`
`
`
`Page 5 of 6
`
`
`
`1624
`
`
`
`.i. M. Moran and D. Maydan: High resolution, steep profile resist patterns
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`1324
`
`
`
`both 1 pm and 2 pm lines and spaCes traversing a i ,um high
`
`
`
`
`
`
`
`
`
`
`
`
`p—glass step. The light lines running vertically are the linal
`
`
`
`
`
`
`
`
`
`
`tri-level resist lines initially patterned with a 0.45 pm of s—ray
`
`
`
`
`
`
`
`
`
`
`resist. The horizontal lines are the p—glass—covered 1 pm high
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`polysilicon steps. Figure lOibi is also an SEM photo, at a very
`shallow angle, of the broadside of the 1 pm lines and spaces
`
`
`
`
`
`
`
`
`
`
`
`
`demonstrating the flatness of the top surface and the con-
`
`
`
`
`
`
`
`
`
`formal coating of the step.
`
`
`
`
`
`Finally. we would like to indicate the usefulness of this
`
`
`
`
`
`
`
`
`
`
`technique for patterning device wafers. This technique has
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`been used for patterning silicon nitride, polysilicon, phos-
`phosilicate. glass and aluminum. As an example, Fig. 11 is an
`
`
`
`
`
`
`
`
`
`
`SEM photo of 0.65 pm thick polysilicon that was reactive ion
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`etched using the tri—leyel structure initially patterned with
`x-ray lithography as a mask. There is seen to be very little
`
`
`
`
`
`
`
`
`
`
`
`((0.1 p.111) linewidth loss in the. etched polysilicon.
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`III. CONCLUSION
`
`
`
`
`
`
`
`
`
`The steep- profile three-layer pattern generation technique
`(tn-level) presents many potential advantages and applications
`
`
`
`
`
`
`to all phases of lithography For optical lithography, it can
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`allow preparation of fine feature devices using present optical
`exposure systems. For xvray and electron lithography, sub-
`
`
`
`
`
`
`
`micron resolution features with better linewidth control and
`
`
`
`
`
`
`
`good step coverage are possible.
`
`
`
`
`
`
`
`
`
`
`
`
`
`ACKNOWLEDGMENTS
`
`
`
`The authors wish to thank D. B. Fraser. C. W, Van Hise,
`
`
`
`
`
`
`
`
`
`
`
`
`
`C. J. Mogab, F. B. Alexander. B. A. Dein, J. M. Zimmer, C.
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`C. Chang, D. N. K. Wang. R. B. Marcus, S. E. Haszko, T. A.
`
`
`Shankoli, and W. T. Willenbrocl: for their resistance in var-
`
`
`
`
`
`
`
`
`
`ious phases of the project. They would also like to thank R. F.
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`W, Pease, H.
`Levinstein, and M. P. Lepselter for their
`
`
`
`
`
`
`
`
`
`
`continuous support of this project. and to acknowledge the
`
`
`
`
`
`
`
`
`
`
`
`photolithogmphic support provided by R. C. Brandes, J. H.
`
`
`
`
`
`
`
`
`
`Velebir, J12, R. E. Kerwin, and
`V. DiLorenzo.
`
`
`
`
`
`
`
`
`
`
`
`Pu. Maldonado. M. Moran, 5. ti, Somek h, and
`lD. Mayilan, (i. A (.quiiin,
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`C. N. Taylor, Conference on Micrnlilhography, Paris. June. 21724 (t. Tit,
`
`
`
`
`
`
`
`
`
`
`pp.
`IQFi—IQQ.
`
`
`‘r'. Lou, and C.
`2L). Maydan, G. A. Coquin, j. R. Maldonado, S. Somekh, I).
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`N. Taylor, iFIFIFI Trans. Electron Devices Elli—22, 429—433 [1975]
`
`
`
`
`
`
`
`
`
`5As an example see Ll. Weissmanlel. M. Host, 0. Fietller, }. EsIer. H. [itc-
`
`
`
`
`
`
`
`
`
`
`
`
`
`gengack and T. Horn, Proc iith International Vacuum Congress (197-1).
`
`
`
`
`
`
`
`
`
`
`pp 439-442.
`
`
`‘J. M. \Jloran and G. 5i. Taylor, Filter-nth Symposium on Electron [on and
`
`
`
`
`
`
`
`
`
`
`
`
`Photon Beam Technology (May, HITS}.
`
`
`
`
`
`3L. F. Thompson, E. M Doerrtes. and 1.. D. Yau. Electrochem. Son. [to he
`
`
`
`
`
`
`
`
`
`
`
`
`
`published}.
`
`5M. linwdell. L. F Thompson, and]. P. Ballulityne, j Var: Sci, Tet-hoot. [2.
`
`
`
`
`
`
`
`
`
`
`
`1294 {1975)
`
`
`7L. F. Thompson and M. Bow-den, Electrochan Soc. IEU, FREE [1973}.
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`n a: so
`
`FIG 11 Application oi the trilr‘vel resisl to reactive ion etching of 0.65 ,am
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`of polysilicon on a rlevict= water. The thick resist is still on and is used as the
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`etch musk.
`
`
`
`
`
`
`
`
`
`photoresist. The thin 0.1 pm thick SiOg layer can just be seen
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`at the top of the feature. Figure 8 is a submicron trench that
`was patterned using a Kasper 10:1 step and repeat camera.
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`For electron-beam lithography, backseattering from a
`substrate covered by 0.1 urn of SiOg on top of 2.6 pm of
`
`
`
`
`
`
`
`
`
`
`
`
`polymer is less than that from an Si or SiOg/Si substrate be-
`
`
`
`
`
`
`
`
`
`
`
`cause of the lower atomic number material. Thus, proximity
`
`
`
`
`
`
`
`
`
`effects for thick resists (>05 um) are reduced and better
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`linewidth control has been observed with both negative and
`positive resists. in addition, high speed electron resists, such
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`as CMC5 or P8515"? can be used for patterning of waters since
`the top layer of resist need not be thick. Figure 9 shows an
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`SEM photo of a tri—level pattern in which 0.5 pm thick CMC
`
`was used to generate the initial pattern. The lines are 0.75 urn
`
`
`
`
`
`
`
`
`
`
`
`
`
`wide and the space is 1.5 pm.
`
`
`
`
`
`
`
`In the case of x-ray lithography, the resist currently being
`
`
`
`
`
`
`
`
`
`
`used ior direct water exposure is a negative, high speed resist.
`
`
`
`
`
`
`
`
`
`
`
`However,
`1 pm resolution is only possible4 when the final
`
`
`
`
`
`
`
`
`
`
`resist thickness is 504 pm. This precludes using it for direct
`
`
`
`
`
`
`
`
`
`
`wafer exposure since it does not have good plasma resistance.
`
`
`
`
`
`
`
`
`
`
`Howewer, the tri-level technique permits the use of thin resists,
`
`
`
`
`
`
`
`
`
`
`since there are no steps. Figure lOia) shows an SEM photo of
`
`
`
`
`
`
`
`
`
`
`
`
`
`J. Vac. Sci. Technol.. Vol. 16. No. 6. NovJDec. 1979
`
`
`
`
`
`
`
`
`
`
`
`Page 6 of 6
`
`Page 6 of 6
`
`