Polymers for High Technology - American Chemical Society

Chapter 20 Effects of Additives on Positive Photoresist Development

Downloaded by GEORGE MASON UNIV on March 12, 2016 | http://pubs.acs.org Publication Date: August 26, 1987 | doi: 10.1021/bk-1987-0346.ch020

R. C. Daly, T. DoMinh, R. A. Arcus, and M. J. Hanrahan Eastman Kodak Company, Rochester, NY 14650 The addition of specialized small molecules to a polymer coating is the functional basis for most photoresists. Conventional positive-working photoresists function owing to the difference in solubility caused by the imagewise exposure of a small molecule naphthalene diazoquinone sulfonate ester (NDS). The presence of this small molecule dramatically inhibits the dissolution of the novolac binder while its photodecomposition accelerates the binder dissolution in aqueous base. An attempt to formulate a poly(4-hydroxy styrene)-based resist was less than completely successful because the difference in the rates of dissolution were too small to be used to give high contrast images. Other small molecules were added to the NDS/novolac resist and these were also found to have a profound effect upon the performance of the resist, particularly the development properties. When it was necessary to obtain higher dissolution rates, several triazoles and sulfonamides were found to improve the rate of development in the exposed areas without causing unacceptable thickness losses in the unexposed areas. Dyes incorporated to minimize problems of reflection and scattered light were also found to alter the dissolution behavior of the resist coating. Making a polymer relief image commonly requires two processes. First, there is a photochemical process which alters the solubility of the exposed areas relative to the unexposed areas. This is followed by the actual dissolution of the most soluble areas during development. Historically, studies of photoresists have emphasized the photochemical aspects of image formation rather than the dissolution process. The central theme of this paper is the very large effects on dissolution rate that can result from adding small molecules to the matrix. Three distinctly different types of compounds have been added to a polymeric binder and studied for photoresist applications. 0097-6156/87/0346-0237$06.00/0 © 1987 American Chemical Society Bowden and Turner; Polymers for High Technology ACS Symposium Series; American Chemical Society: Washington, DC, 1987.

238

POLYMERS FOR HIGH T E C H N O L O G Y


These are naphthalene diazoquinones as photoactive compounds, benzot r i a z o l e s as development enhancement agents, and dyes to reduce exposure from r e f l e c t e d l i g h t . The l a t t e r compounds have been added to r e s i s t formulations which already contain a photoactive compound i n addition to the polymer binder. The r e s i s t system that we have explored i s based upon the photochemistry of naphthalene diazoquinone i n a base-soluble polymer matrix. This type of system i s of interest because i t i s the basis for a l l commercial near-UV photoresists. The photochemistry i s o l d but recently has been studied i n d e t a i l (1). Reaction scheme 1 shows the series of reactions and rearrangements that occur a f t e r the photolysis of a diazoquinone i n the presence of water.

COgH

R Scheme 1 There are three major events i n the reaction sequence. F i r s t the naphthalene diazoquinone i s destroyed. Second, nitrogen gas i s released. Third, an a c i d i s produced. However, the photochemistry i t s e l f does not make a r e l i e f image. Rather i t i s used to modify the s o l u b i l i t y of the polymeric binder. The diazoquinone compounds used i n r e s i s t s are referred to as dissol u t i o n i n h i b i t o r s or photoactive components (PAC's). The addition of a diazoquinone molecule dramatically i n h i b i t s the d i s s o l u t i o n rate of a t h i n f i l m of a novolac r e s i n . Upon exposure, the d i s s o l u t i o n rate of the novolac based r e s i s t i s considerably f a s t e r than the rate for the novolac alone. The accelerated d i s s o l u t i o n rate may be caused by formation of acid and i t s subsequent i o n i z a t i o n during development or by enhanced d i f f u s i o n of the developer i n t o the coating because of changes caused by the formation and fate of the nitrogen (2). When working with t h i s type of r e s i s t to make VLSI devices there are several problems that come up, p a r t i c u l a r l y the need f o r improved thermal s t a b i l i t y from the r e s i s t images and a method t o control l i g h t that i s r e f l e c t e d or scattered i n a nonimagewise manner. The function of a development enhancement agent i s to increase the d i s solution rate of the photoresist so that polymers with better physic a l properties but slow dissolution rates may be used. The dyes

Bowden and Turner; Polymers for High Technology ACS Symposium Series; American Chemical Society: Washington, DC, 1987.

20.

DALY ET A L .

Effects of Additives

on Positive Photoresist

239

Development

mentioned above are being incorporated i n r e s i s t s i n hopes that they can absorb the unwanted l i g h t and help to produce higher resolution images·

Fesult?

Discussion


Table I contains d i s s o l u t i o n rate data for an m-cresol novolac, 1, and for the novolac containing 15 wt% of the diazoquinone 2. Using

Table I.

Development Rates

Polymeric binder Polymeric binder + 15 wt% 2 unexposed Polymeric binder + 15 wt% 2 exposed Discrimination PR exposed UK unexposed

Novolac, 1 μπι/min ~

pHOSt, 3 μπι/min

0.15 0.02 1.70 85

6.15 2.50 10.10 4

Note: The developer was Kodak micro positive developer 934 d i l u t e d to a 1.5 wt% l e v e l of (CH ) N0H and used at 20°C. 3

4

a 1.5% solution of tetramethyl ammonium hydroxide, Me^NOH, developer the novolac dissolves at a rate of 0.15 μπι/min. The addition of the photoactive model compound decreases t h i s rate to 0.02 μπι/min· The mechanism of t h i s dissolution i n h i b i t i o n i s not c l e a r l y understood. However, upon exposure the rate of d i s s o l u t i o n i s increased to 1.70 μπι/min, a rate substantially faster than the o r i g i n a l novolac. The developer discrimination between exposed and unexposed novolac coatings containing 2 i s above 80. It i s t h i s high discrimination that has allowed the naphthalene diazoquinone chemistry to dominate p o s i t i v e r e s i s t chemistry f o r almost 25 years. The e f f e c t of the PAC*s i s not l i m i t e d to novolac binders. Poly(4-hydroxy styrene), pHOSt, 3, has been suggested as a binder r e s i n f o r r e s i s t formulations. The primary advantage of a r e s i s t based upon poly(4-hydroxy styrene) i s the high glass t r a n s i t i o n temperature of the polymer (Tg = 187°C). A high Tg polymer has the potential to give a r e s i s t that i s better able to withstand the harsh environment of ion implantation and plasma etching. The disadvantage of a high Tg i s that a l l of the solvent cannot be removed without decomposing the PAC. The addition of a PAC to pHOSt causes



240

d i s s o l u t i o n i n h i b i t i o n , and a f t e r exposure the d i s s o l u t i o n i s a c c e l erated. However, the discrimination/dissolution rate r a t i o i s only 4 , a factor of 20 lower than for the novolac r e s i s t .

OH


3 S i g n i f i c a n t differences are observed i n the performance of novolac and poly(4-hydroxy styrene) as r e s i s t binders. Though the two polymers have equivalent a c t i v a t i o n energies of about 10 k c a l / mole f o r d i s s o l u t i o n at workable rates, the pHOSt i s s l i g h t l y more soluble i n strong base (requires pH>12 using Me^NOH) than i s novolac (requires pH>12.5 using Me^NOH)· pHOSt behaves as a stronger a c i d than does novolac. The plfOSt dissolves much faster i n aqueous base than does the novolac at approximately the same Mw of 10,000. Another major difference i s that the discrimination between exposed and unexposed coatings i s considerably less for the pHOSt r e s i s t . The rate of development i s always faster for the pHOSt but the discrimination between the exposed and the unexposed r e s i s t s i s smaller. Comparison of the dissolution rate curves f o r novolac and pHOSt i n Figure 1 gives a good view of the difference i n response of the two r e s i s t systems. It i s d i f f i c u l t to formulate the pHOSt into a simple r e s i s t because there i s too much thickness l o s t i n the unexposed areas when a developer i s used that i s appropriate to develop the exposed areas i n a reasonable time. However, there are compelling reasons f o r wanting to have a higher Tg r e s i s t that would be better able to withstand harsh t r e a t ment during the f a b r i c a t i o n of the device. One of the reasons that the Tg's of the novolac resins are generally low i s that they are quite low i n molecular weight (Mw between 2000 and 6000). This low molecular weight allows f o r easier synthesis and much faster development times but at a cost i n r e s i s t performance. Careful attention to the process of making the novolac can overcome the f i r s t problem but the r e s u l t i n g r e s i s t s made from high molecular weight resins (Mw*s greater than 12,000) dissolve too slowly i n the d i l u t e developers needed to prepare high resolution images. One way to speed the d i s s o l u t i o n i s to add other components to the r e s i s t . In p a r t i c u l a r the addition of benzotriazoles and sulfonamides to a r e s i s t formulation produces a r e s i s t that appears to be more sensitive (3-5). Since s e n s i t i v i t y i s a combination of photospeed and d i s s o l u t i o n , an increase i n s e n s i t i v i t y can imply either a lower exposure or a faster development from the o r i g i n a l exposure. The data i n Table II give a simple picture of the e f f e c t of several benzotriazole additives, 4 , on the d i s s o l u t i o n rate of a conventional


20.


DALY ET AL.


R

241

Development

H

3

4


Table I I . Additive

Time to clear (sec)

None 4 s R^fl^/R^"^!

4a s R^ C1* R 2 , R ^ H 4b: R , R , R = C 1 =

1

=

2

3

Enhanced Development

38.0 18.0 16.5 7.5

Thickness

loss (A)

400 100 100 100

Note: The developer was Kodak micro p o s i t i v e developer 934 d i l u t e d l s l with deionized water. Puddle development took place at 20°C for these 1.2-μπι coatings.

novolac/NDS r e s i s t formulation. Time-to-clear, the figure of merit i n Table I I , i s the development time needed to dissolve a l l of the r e s i s t i n an exposed area. In each case the additive was used at 12 wt% of the s o l i d s . Putting a benzotriazole i n t o the coating s i g n i f i c a n t l y increases development speed. Adding the electronwithdrawing chlorine substituents to the benzotriazole further hastens development. It i s assumed that i t i s the increased a c i d i t y of the N - H which causes t h i s e f f e c t . The benzotriazoles seem to only e f f e c t the d i s s o l u t i o n rate of the exposed areas; there i s no increase i n thickness loss i n the unexposed areas. Less PAC needs to be destroyed to give a developable r e s i s t image since the f i l m i s more soluble to begin with. Mechanistically, there are three ways that these molecules could e f f e c t the speed of the r e s i s t s they could 1) make the photochemistry more e f f e c t i v e , 2) chemically a s s i s t i n the production of a c i d or 3) simply a l t e r the d i s s o l u t i o n of the thin f i l m . It i s r e l a t i v e l y easy to follow the course of the photochemistry since the i n i t i a l photochemical event results i n the destruction of the quinone diazide chromophore. For s i m i l a r r e s i s t formulations with and without additives there was no change noted i n the rate of bleaching of the UV absorption of the quinone diazide. The p o s i t i o n , shape and extinction c o e f f i c i e n t of the absorption were not a l t e r e d by the a d d i t i v e s . With these experiments i n mind, i t i s very unlikely that the additives are involved i n the photochemistry. If the development enhancement agents react chemically with the photoproducts of the N D S , i t should be possible to postulate the reaction products and a mechanism for the enhanced response. Faster development might r e s u l t i f the benzotriazole competes e f f e c t i v e l y




242

Scheme 2 with any p a r a s i t i c side reactions of the indene ketene intermediate. Scheme 2 shows the desired reaction of the ketene along with one of several undesired reactions and a competing reaction of benzotriazole. The undesired reaction i s the formation of an ester with the novolac r e s i n . This type of reaction i s p a r t i c u l a r l y bad since the ester saponifies too slowly i n the developer to y i e l d a c i d during development. Since the photoactive component normally has several reactive groups the novolac r e s i n becomes crosslinked and less soluble due to t h i s reaction. If the benzotriazole reacts quickly enough to reduce the amount of ester formed and forms an amide product that i s very r e a d i l y hydrolized i n the developer, then there w i l l be a net gain i n the effectiveness of the photoactive component. It should be possible to see the chemical products i n the i n f r a r e d spectra of the t h i n f i l m . In order to look at the small quantities of material involved, i t was necessary to do i n s i t u exposures of the r e s i s t coatings on s i l i c o n wafers i n a Fourier transform i n f r a r e d spectrophotometer. The technique was capable of following the loss of quinone and the formation of ketene with considerable success. By purging the wafer i n the chamber f o r some time i n the presence of dry nitrogen, i t was possible to observe a stable ketene s i g n a l even hours a f t e r the exposure. While these experiments were not quantitative, they d i d give two pieces of



20.

DALY ET A L .



Development

243

information. F i r s t , the reaction of the ketene with novolac must be quite slow i n comparison with i t s reaction with water. Second, the s t a b i l i t y of the ketene signal was not effected by the additives. There was no i n d i c a t i o n of new species i n the FTIR spectra. The combination of these two phenomena makes i t highly u n l i k e l y that chemistry occurs between the reactive intermediates and the addit i v e s . Experiments i n the s o l i d state but without the novolac r e s i n indicated that the proposed amide intermediate could be observed for benzotriazole i n a glassy quinone diazide matrix. The increased rate of development for exposed coatings with additives i s , however, quite pronounced. This can be seen most e a s i l y i n the traces from a Perkin Elmer d i s s o l u t i o n rate monitor. The data i n Figure 2 show that at a l l exposure l e v e l s the r e s i s t containing a chlorobenzotriazole development enhancement agent requires a shorter development time than does the unaltered r e s i s t . The converse of t h i s process i s that the r e s i s t with the DEA requires shorter exposure times when the same development time and process are used. For example, only 25 mJ of exposure are required to give an image that clears at 30 sec for the r e s i s t with DEA while a 70 mJ exposure i s needed for the model r e s i s t i t s e l f . The contrast of the r e s i s t image i s also improved by the addition of a benzotriazole DEA. Using the r e l a t i v e l y long exposure times needed f o r quick development i n track equipment, the contrast i s always higher f o r the r e s i s t with DEA, Figure 3. Since the thickness loss i n the unexposed areas i s l i t t l e effected by the DEA, the improved contrast i s probably an a r t i f a c t of increased developer discrimination between exposed and unexposed r e s i s t . Another class of small molecules that may be added to p o s i t i v e photoresists are dyes. Control of the scattered and r e f l e c t e d l i g h t i n the r e s i s t coating i s the common reason for adding a dye to a r e s i s t formulation. I n i t i a l l y , the NDS photoactive component absorbs less than one half of the incoming l i g h t at 436 nm. This means that there w i l l be s i g n i f i c a n t amounts of l i g h t r e f l e c t e d from the substrate even as the exposure begins. During the exposure, the NDS absorbance i s bleached, allowing more l i g h t be be scattered and r e f l e c t e d . Since the d i f f e r e n t microelectronic substrates have d i f f e r e n t r e f l e c t i v i t i e s , the r e s i s t w i l l receive d i f f e r i n g exposures depending upon the substrate under the area being exposed. Several other f a c t o r s , including variations i n thickness, only serve to make the problem of obtaining uniformly exposed r e s i s t l i n e s more difficult. Several approaches to making images over r e f l e c t i v e topography use dyes that absorb at the exposure wavelength. Two of these processes involve putting a dye i n a separate layer under the photor e s i s t . The use of a t h i n dyed layer (an a n t i r e f l e c t i o n coating) under the r e s i s t keeps the r e f l e c t e d l i g h t from reentering the r e s i s t (fi). However, t h i s layer must be coated and removed, possibly requiring several extra steps, and care must be taken to see that the r e s i s t and the a n t i r e f l e c t i o n layer do not mix. Placing the dye i n a thick layer under the r e s i s t o f f e r s the advantage of addressing the topography of the device wafer surface (7). A thick planarizing layer i s coated to reduce the d i s c o n t i n u i t i e s of the substrate and to present a uniform surface for the r e s i s t coating. Both of these dyed layers suffer from the problem of how to transfer the image from the r e s i s t on top through the under layer. While these concepts



244

—ι

ι

c

2

°4

1 ε

Φ

S *

yιο y

ο

S Downloaded by GEORGE MASON UNIV on March 12, 2016 | http://pubs.acs.org Publication Date: August 26, 1987 | doi: 10.1021/bk-1987-0346.ch020

I

ι

1

c ο

-2

-

y

/

ο

-4 /

ό

/

/

;*77

r

-6 -0.7

/

/

/

/

/

/

/

/

2

-0.4 -0.1 0.2 0.5 0.8 Log developer concentration

I.I

Figure 1. Development occurred at 20°C with (CH ) NOH developers ( Δ-) novolac; ( o-) pHOSt. The numbers at the data points represent discrimination between the d i s s o l u t i o n rate f o r exposed and unexposed r e s i s t formulated from the bender with 15 wt% 2. 3

15

20 25 30 35 Time to cleor (sec)

40

4

50

Figure 2. The model r e s i s t , 13 wt% PAC i n novolac, was developed at 21°C with 1.5% (CH ) NOH developers • = no DEA, Δ = 13 wt% of 4b. 3

4



20.

DALY ET A L .

Effects of Additives on Positive Photoresist

Development

245

are workable they present formidable engineering d i f f i c u l t i e s . Also, since the dyes absorb r e f l e c t e d l i g h t which might otherwise enter the r e s i s t , extra exposure i s required to achieve the desired image when exposing r e s i s t on top of a dyed layer. However, s i g n i f i c a n t improvements i n the image uniformity over r e f l e c t i v e topography can be seen with either of these dyed processes. Putting a dye into the r e s i s t i s the most d i r e c t way to reduce the amount of r e f l e c t e d l i g h t (£). Here the dye i n the r e s i s t coating competes d i r e c t l y with the photoactive component f o r the incoming l i g h t , thereby causing an even larger speed decrease than seen i n the previous two cases. However, s i g n i f i c a n t performance improvements have been shown f o r dyed r e s i s t s , p a r t i c u l a r l y i n the area of reduced notching over A l . However, the dyes also a l t e r the dissolution properties of the r e s i s t (9). Figure 4 shows the absorption spectra of a cyanine dye, 5, which has been used i n a positive r e s i s t t o reduce the image

Bu Bu

+

P

F

Bu

e

Bu

7

0

iProOCCH

C

N

0

0

? \ Kh0^ ^~^-0- I 3 2

CH=

iProOCCH

2

w

NH

ch

0

d i s t o r t i o n caused by r e f l e c t i o n s . Important features of t h i s dye are that i t has a very high extinction c o e f f i c i e n t near 436 nm and that i t has very l i t t l e absorbance i n the mid- and deep-UV. When the dye i s added to Kodak micro positive r e s i s t 820 (820 r e s i s t ) , there i s a s i g n i f i c a n t improvement i n the linewidth control and also a large decrease i n the dissolution rate of the r e s i s t . Figure 5 shows the v a r i a t i o n of the dissolution rate as dye i s added. At a dye l e v e l of 0.25% the dissolution rate i s less than 60% of the undyed r e s i s t . Other dyes have been found that accelerate the d i s s o l u t i o n of the dyed coatings. The dyes that a i d d i s s o l u t i o n are generally those that have functional groups that dissolve r e a d i l y i n aqueous base, such as sulfonamides, phenols and acids. Figure 6 i s a plot of the dissolution behavior of a dissolution-accelerating chalcone sulfonamide dye, 6. The a c i d i c sulfonamide proton should ionize i n the basic developer and hasten the d i s s o l u t i o n of the exposed and already dissolving f i l m s . Those dyes that i n h i b i t d i s s o l u t i o n seem to be molecules that can react or complex with the acid product of the photochemistry. For example, the i o n i c ammonium function of the cyanine could interact with one or two of the indene carboxylic acids to form complexes that would be much less soluble i n the aqueous base developer.



246

2.0i

τ

1

1

1

1

1

1 Δ


Α 3 - ι.δ| ο Ε Ε

Δ

Γ Ο

Δ

Q

*Δ

& χ 1.01

00" 0

J

5

I

10

I

15

20

I

25

L

30

35

40

45 50

Time to clear (sec)

Figure 3. The model r e s i s t , 13 wt% PAC i n novolac, was developed at 21°C with 1.5% (CH ) NOH developer: π = no DEA, Δ = 13 wt% of 4b. 3

4

2.0

1.6

O.D.

W4l8nm(IV \

W2.9XI0

4

0.8

0.4

200

300

400

500

Wavelength (nm) Figure 4. UV absorption spectra of cyanine dye 5.


DALY ET A L .

Effects of Additives on Positive Photoresist

Development


1200

0.1

0.2

0.3

0.4

0.5

Dye cone. (%W) Figure 5 . Dissolution rate f o r 820 r e s i s t containing dye 5 i n 9 3 4 developer. 4000i

0

0.5

1.0

1.5

2.0

2.5

3.0

Dye conc.(%W) Figure 6. Dissolution rate f o r 820 r e s i s t containing dye 6 i n 9 3 4 developer.

American Chemical Society Library

1155 16thforSt.. Bowden and Turner; Polymers HighN.W. Technology ACS Symposium Series; American Chemical D.C. Society:20036 Washington, DC, 1987. Washington,


248

The best development discrimination occurs with the d i s s o l u t i o n i n h i b i t i n g dyes, while the higher speed occurs with the development accelerating dyes, Table I I I .

Table I I I . Performance of Dyed 820 Resist Control Dye cone. %W Dissolution rate A/min Relative rate Stepper exp. msec (436) Thickness loss Selectivity


o

None 1290 1.00 140 3% 28

Cyanine, 5 0.25% 810 0.63 180 1% 31

Chalcone, 6 1.0% 1910 1.48 160 5% 11

Summary

The purpose of t h i s work was to show several of the ways i n which small molecules e f f e c t the d i s s o l u t i o n of a polymer f i l m . This phenomena i s s i g n i f i c a n t for microresists because i t represents the basis f o r micro imaging and because several new ways of improving or a l t e r i n g the r e s i s t performance depend upon adding new chemicals t o the r e s i s t formulation. The intent of these materials may not be to a l t e r the r e s i s t d i s s o l u t i o n but the potential f o r change i s ever present. Experimental The novolac used i n t h i s study was prepared by the base catalyzed solution condensation of m-cresol with formaldehyde (Mn = 1800, Mw = 9300, Mw/Mn = 5.07). The poly (p-hydroxy styrene) was PHP-6817-24, obtained from Marusen O i l Co. (Mn = 3900, Mw = 10200, Mw/Mn = 2.62). Kodak micro positive developer 934 (934 developer) (predominantly tetramethyl ammonium hydroxide i n water) was used at various d i l u t i o n s with deionized water. M-cresyl naphthalene diazoquinone sulfonate was prepared by a pyridine catalyzed condensation of m-cresol with naphthalene diazoquinone s u l f o n y l chloride. The simple d i s s o l u t i o n rate results were obtained using a laser interferometer with a 15 mw/cm He-Ne laser at normal incidence to the wafer surface i n the agitated developer bath. The r e f l e c t e d beam was directed by a beam s p l i t t e r onto a photocell. The photocell output was fed through a Keithly series 500 interface into an IBM-PC. The more complex d i s s o l u t i o n data were c o l l e c t e d on a Perkin-Elmer d i s s o l u t i o n rate monitor using 934 developer at a 1:1 d i l u t i o n with deionized water at a temperature of 21°C. The stepped exposures were obtained using a c a l i b r a t e d multidensity chrome stepwedge. 2


20. DALY ET AL.

Effects of Additives on Positive Photoresist Develop


Acknowledgments The authors wish to thank Sandra Finn for FTIR studies and both Kathlene Hollis and Caroline Little for running many of the dissolution rate experiments. Literature Cited 1. 2.

Pacansky, J.; Lyerla, J. R. Polym. Eng. Sci. 1985, 20, 1049. Hinsberg, W. D.; Willson, C. G.; Kanazawa, Κ. K. SPIE Proc.

3. 4. 5. 6. 7. 8. 9.

Western Electric G.B. Patent 1 317 796, 1972. IBM French Patent 2 325 076, 1976. Eastman Kodak U.S. Patent 4 365 019, 1982. Brewer, T.; Colson, R.; Arnold, J. J. Appl. Photogr. Eng. 1981, 7, 184. O'Toole, M. M.; Liu, E. D.; Chang, M. S. SPIE Proc., 1981, 128, 275. Brown Α. V.; Arnold, W. H. SPIE Proc., 1985, 539. Bolsen, M.; Buhr, G.; Merren H. J.; van Werden, K. Solid State Technol. 1986, Mar, 83.

1985, 6, 593.

RECEIVED

June 15, 1987


Polymers for High Technology - American Chemical Society

Recommend Documents