Functional Imaging with Chemically Amplified Resists - ACS

Chapter 10

Functional Imaging with Chemically Amplified Resists 1

Downloaded via UNIV OF CALIFORNIA SANTA BARBARA on July 9, 2018 at 13:13:48 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

Alexander M. Vekselman, Chunhao Zhang, and Graham D. Darling

Department of Chemistry, McGill University, 801 Sherbrooke West, Montreal, Province of Quebec H3A 2K6, Canada

The same dramatic photosensitivity shown by films of chemically amplified resists that permits their patterned removal ("relief development", typical of microlithography), can also instead allow their further imagewise chemical modification ("functional development"), such as through exposure-controlled sorption of various species from contacting solutions or vapors. For example, radiation-defined deprotection of nonpolar poly(di-t-butyl fumarate-co-styrene) produced a pattern of polar and reactive carboxylic acid and anhydride moieties. Conditions were found for only these exposed areas of the resist material to take up Ca(II), Ni(II), Co(II), Pb(II) or some ammonium ions from the corresponding aqueous solutions, without being dissolved. Several organic dyes were also placed into either exposed areas from water/alcohol solutions, or into unexposed areas from hexane/toluene solutions. Modes and mechanisms are discussed in terms of resist, solute and solvent properties. Since the early 1980's, sensitive Chemically Amplified (CA) resists such as based on P(p-TBOCST) (i.e. poly(p-i-butyloxycarbonyloxystyrene)) have been designed and tested to rapidly produce fine patterns of contrasting materials, such as in the manufacture of high-density DRAM devices and microprocessor chips by microlithography [1]. The traditional role for a resist in such applications has been to capture a projected radiation pattern as a "relief image" of removed and remaining areas of resist film (Figure 1), which is then translated into an underlying inorganic substrate through an etching process [2], However, microlithography-like techniques can also be used for making fine-scale optical waveguides, couplers, and recording media, or array sensors, displays and supports for solid-phase (biochemical analysis or synthesis, or other useful patterned structures. Here, it would be often better not to remove, from the intermediate "latent image", either the exposed or unexposed areas of resist material ("relief development"), but instead only to further modify one or the other ("functional development"). Possible physical, chemical or biological properties that could thus be 1

Corresponding author 0097-6156/95/0614-0149$12.00/0 © 1995 American Chemical Society

Reichmanis et al.; Microelectronics Technology ACS Symposium Series; American Chemical Society: Washington, DC, 1995.

MICROELECTRONICS TECHNOLOGY

Figure 1. Relief and functional imaging with resists.


10.

VEKSELMAN ET AL.

Imaging with Chemically Amplified Resists 151

altered in selected patterns include density, refractive index (n), color, fluorescence, hyperpolarizability, reflectivity, heat capacity, dielectric constant, adhesiveness, liquid crystallinity, electrical or thermal conductivity; together with wettability, permeability, sorption, catalytic activity or chemical reactivity towards various liquid or gas-phase species; along with dissolution and etch resistance that would be relevant to later relief development steps as well. Upon or following exposure, the polymers within CA or other resist materials can become altered in three general ways: i) polymerization or crosslinking, that increases average molecular weight; ii) scission or depolymerization, that decreases average molecular weight; iii) functional group transformation, that does not affect the lengths of polymer chains, but drastically alters their polarity, solubility and reactivity. This third mode, as exemplified in P(p-TBOCST) [3], is not only the most attractive for relief development because of the lack of swelling in undissolved material and the possibility of dual-tone imaging, but is also the most adaptable towards functional development. Indeed, both solubility and many other immediate or potential properties (including, tendency to further chemical modification) are critically dependent on the exposure-controlled functional composition of such resists, through highly cooperative behavior of functional groups. Unlike relief development, a functional development step by definition does not remove any of the solid polymer film, and can be a heterogeneous reaction that produces a stratified product. Patterned modification exclusively at a surface has been done to adsorb catalysts for further metallization [4,5], or biomolecules for DNA sequencing [6]; non-photochemical surface imaging has also been reported by a stamping technique [7]. In near-surface imaging, only the upper 10-100 nm of material is transformed due to low penetration of the initial imaging radiation (ex. low-energy electron or ion beams, or UV in relatively opaque materials [8]), or due to low diffusion of chemical species during a later functional development step (ex. in the "diffusion enhanced silylated resist" = DESIRE processes [9,10]). Modifying species can also enter and diffuse deeper into the bulk of the polymer. The rate and extent of such uptake can depend on the polymer-solution interface, or on bulk polymer-polymer (ex. crosslinking), polymer-solvent (ex. plasticization), and polymer-solute interactions, covalent or not, often involving polymer functional groups [5,11-15]. The understanding and design of functional development would benefit from the study of relevant chemical and physico-chemical phenomena, similar to what has been done on airborne contamination of resists [16,17], and miscibility of their components [18]. One general way of modifying a thin film of material is by controlled uptake of ions or molecules from a contacting liquid or gas. We discuss here the functional development, through exposure-controlled sorption of various inorganic and organic chemical species, of a model CA resist based on poly(di-f-butyl fumarate-ca-styrene) (PDBFS). Experimental Instruments. Instruments for spin-coating, photo-exposure, and baking were described in earlier publications [11]. The pH of aqueous solutions was measured with a Cole-Palmer Digi-Sense 5938-00 pH-meter with 5992-40 combination electrode, which had been calibrated using commercial buffer solutions (Caledon Inc.) of 0.05 M potassium hydrogen phthalate (pH 4.00) and 0.01 M sodium borate (pH 9.22). FTIR spectra were obtained with a Brucker IF-48 spectrophotometer with a microscope accessory. UV-VIS spectra were done with a Shimadzu Spectronic-210UV


152


spectrophotometer. Fluorescence measurements were made in a Spex model F112 spectrofluorimeter with a 450 W xenon lamp, with the emission detector at 90° to the incident beam for solution measurement, and at 22.5° for solid films. Oxygen Reactive Ion Etching (O2-RIE) was performed in a custom-built Large-volume Microwave Plasma (LMP) apparatus [19]. Chemicals. Hexane, toluene, methanol, n- and wu-propanol solvents were "all glass distilled" grade from OmniSolv Inc.; pyridine from BDH Inc.; ί-butanol from Baker Inc.; and the dye Rhodamine-6G (R6G) from Aldrich Inc.. The preparation of 4-(5-N,N-dimethylaminonaphthalenesulfonylamido)-l-methylpiperazine (DSMP) was previously described [11]. 1,10-Diaminodecane, 2- and 4-aminopyridine and 2,6diaminopyridine were obtainedfromAldrich Inc.; 30% poly(ethyleneimine)/H20 ( M = 10K g/mol) from PolyScience Inc.. Metal salts of >99% purity were used as received: CuCl2, N1CI2, C0CI2, AICI3, FeCfe and FeCbfromAlfa Inc.; Pb(OAc)2 and PdCl2 from Aldrich Inc. Each of these was dissolved in H2O to ca. 0.1 M , except for PdCl2 which was almost insoluble. Ca(OH>2 was prepared [20] from calcium oxide (Anachemia Inc.), and saturated solutions were freshly filtered before use. 0.1 M HCI/H2O and 5% NH3/H2O were prepared by dilution of the commercial concentrates (Caledon Inc. and BDH Inc., respectively). w

Preparation and Exposure of Resist. Di-i-butyl fumarate was prepared from fumaryl chloride and potassium t-butoxide, then polymerized with styrene in toluene solution with 2,2'-azobis(isobutyronitrile) (AIBN). The resulting PDBFS contained ca. 45 mol % of fumarate units, with M = 15K g/mol by GPC in chloroform, and [η] = 0.08 in chloroform [11]. Unless otherwise stated, samples for lithographic evaluation were prepared by dissolving 50 mg of PDBFS, together with 5 mg of either triphenylsulfonium or 4-(phenylthio)phenyldiphenylsulfonium hexafluoroantimonates [21] as Photo-Acid Generator (PAG), into 250-350 mg of propylene glycol methyl ether acetate,filteringthis solution (0.2 μπι) and spin-coating it at 1000-1200 rpm onto a silicon wafer, then performing a Post-Apply Bake (PAB) at 130 °C for 60 s, to give a 0.8-0.9 μπι film that was optically clear. This was later irradiated with 0-100 mJ/cm of deep- (254 nm) or mid-near-UV (ca. 20 % of 313 nm and 80 % of 365 nm) through an Optoline-Fluroware density photoresist step table REK/73 with resolution to 1 μπι, or while half-shading the sample with an opaque object, before being subjected to a Post-Exposure Bake (PEB) at 135 °C for 60 s. w

2

Functional Development with Metal Ions. 30 mL of solution containing metal ions was placed in a beaker with a pH electrode and continuous magnetic stirring. As needed, drops of 0.1 N HCI/H2O or 5% NH3/H2O were then added to adjust the pH of the medium. Silicon wafers containing irradiated-baked resist were introduced for 10-180 s, then quickly rinsed for 30 s with distilled water, and air-dried before FTIR and other evaluation. Reactive Ion Etching. Samples were prepared and exposed in mid-near-UV through the step tablet (0-100 mJ/cm ), dipped in saturated Ca(OH)2/H20 for 90-180 s, dried, then later subjected to O2-RIE in die LMP apparatus under conditions of 10 seem O2 flow, 50 mtorr pressure, 120 W power, and -500 V bias. 2

Refractometry. Refractive index η of a thin resist film on a slide of fused silica was measured by a known wave-guiding technique [22], in which a laser beam was


10.

VEKSELMAN ET AL.


coupled by a prism into the resist layer, then decoupled ca. 1 cm further on by another prism. The relationship between coupling and decoupling angles and output light intensity gives both average η and film thickness. As the technique requires multi-mode wave-guiding, relatively thick films (1.1-2.0 μπι) were prepared using a more concentrated resist solution and lower spin speed. Functional Development with Amines and Dyes. After standard processing (PAB, exposure and PEB), resist samples were immersed into 0.1-1 wt % solutions of amino-compounds or fluorescent dyes for 30 s, air-dried and evaluated. Modification from the gas phase was performed by suspending resist-coated wafers for a measured time over the liquid in a covered beaker partly filled with dilute aqueous ammonia or pure pyridine. Results and Discussion PoIy(Di-*-Butyl Fumarate-co-Styrene) Resist. The evolution of resists can be described as a progressive adjusting of materials properties towards simultaneously meeting all the requirements of the lithographic process, which are many and often conflicting: shelf-life vs. sensitivity, transparency at the short wavelengths needed for highest resolutions vs. plasma resistance requiring aromatic groups, ease of annealing vs. thermostability, etc. Many CA resists have been inspired by the use of protecting groups in synthetic organic chemistry, in which "capped" carbonate, ester or ether groups can undergo acid-catalyzed "deprotection" reactions to become much more polar carboxyl or hydroxyl groups. Carboxylic acids would be particularly attractive groups to permit both relief and functional development, via deprotonation and/or ion exchange. Such are generated from co/terpolymers of ί-butyl with other alkyl acrylates undergoing thermo-acidolytic deprotection with photogenerated acid, permitting their relief development with aqueous base [23]. These aliphatic polymers also possess excellent UV transparency down to 190 nm but, without silicon or other plasmaresistant moieties, their films cannot later withstand conditions of the substrate etching step. Alternatively, polymers of protected vinyl-benzoic acid sufficiently resist such etching, however, conjugation between the carbonyl and aromatic ring results in strong optical absorbance below 320 nm [24]. A further improvement has been the copolymer of r-butyl methacrylate with α-methylstyrene, that combined etch-resistance with acceptable deep-UV transparency [25], However, with only one ionizable group per 12 carbons, even the largely-deprotected matrix proved hard to dissolve in aqueous base without the addition of cosolvents such as wo-propanol. The idea to copolymerize di-f-butyl fumarate with styrene for new CA resists appeared almost simultaneously in a Japanese patent [26], and in reports from our own [11,27] and Crivello's groups [28] (similar structure with additional α-cyano fumarate group was reported even earlier by Ito et al. [29]). Within the resulting PDBFS copolymer, each fumarate contributes a non-conjugated, UV-transparent unit with two acid-labile functionalities, while the styrene unit confers processability and plasma resistance. Excellent photolithographic properties have been reported [11,26-28], along with modest deep-UV absorbency (0.17-0.18 μπι- ), elevated glass transition temperature (T ) both before (134 °C) and after deprotection (199 °C), and high thermostability (>300 °C). After deprotection, the copolymer could be removed easily by aqueous base alone, obviously due to the dianionic fumarate unit. According to developing conditions, PDBFS/onium salt combinations could give either positive- or negative-tone micron-scale relief images, with high photosensitivity of 14-40 mJ/cm 1

g

2


154


PDBFS ("ester")

Photo-Acid Generator (PAG)

ι

Ar S+-X 3

'J 0 ^ 0

"Deprotected PDBFS - acid"

^ F

(acidolysis)

\ >245°c\^ (thermolysis)

Ο^ΌΗ /

p

h

>130°C (dehydration)

ι -

^ . V " D e p r o t e c t e d PDBFS S) - anhydride"

0 =

Scheme I. Reactions in PDBFS/onium salt resist

0

10 20 30 40 Deep-UV dose, mJ/cm

50

2

Figure 2. Acid-catalyzed ester-to-acid conversion and developing of PDBFS resist vs. deep-UV dose (254 nm). Symbols represent normalized intensity of 1145 cm-1 ( Ο ) and 2977 cm-1 ( V ) peaks, as well as normalized thickness after developing in toluene ( • ) and aqueous tetramethylammonium hydroxide ( A ).


10. VEKSELMAN ET AL.

155 Imaging with Chemically Amplified Resists

[11]. Versions of higher fumarate content and a more lipophilic PAG showed even better sensitivity of 2-4 mJ/cm2 [28]. For this study, we chose both a high UV dose (100 mJ/cm2) and PAG loading (10 wt %), as well as ample PAB and PEB (130-135 °C for 60 s), so that results of the new methods of development would not be affected by slight variations in the other processing steps. Many general principles and development techniques worked out for PDBFS could be expected to apply to the other carboxyl-based resists [23-25] as well. PEB Reactions and Relief Development of PDBFS. Our previous study of f-butyl elimination from PDBFS showed clean thermolysis at 245 °C in the absence of strong acid, and at ca. 100 °C in its presence [11]. However, FTTR showed that, at higher PEB temperatures, the appearance of free -COOH moieties was often followed by their conversion to anhydride (Scheme I). Comparison with IR spectra of commercial poly(styrene-c0-maleic anhydride), and ease of dissolution in some solvents, suggested the occurrence of mostly intramolecular 5-member ring formation, without intermolecular anhydride crosslinking. It was roughly found that the ratios between ester, acid and anhydride groups could be controlled by UV dose (up to ca. 35 mJ/cm2) and PEB temperature (variations in PEB times beyond 30 seconds had no great effect). Similar ester -> acid -> anhydride transformations have have also been studied in other resists [29,30]. Carboxylic acids and their ester, anhydride and anion derivatives each show characteristic strong absorption peaks in the carbonyl (1850-1400 cm-1) and C - 0 (1000-1200 cm-l) regions of their FTTR spectra; O-H of the free acids also absorb above 2900 cm-1. The relative quantities of f-butyl ester groups in solid PDBFS films were measured by the intensities of the strong, sharp and isolated absorbance peaks from 0-CMe3 (1145 cm-1) and CMe2CH2-H (2977 cm-1) beyond a baseline defined by a spectrum of the same sample after obviously complete f-butyl cleavage. Comparison was also made between these peaks and that from RC6H4-H bend (703 cm-1) in the same spectrum, as a reference that remained unaffected by chemical changes elsewhere in the polymer. The step tablet conveniently provided a wide range of UV dosage onto different areas of the same resist film. The silicon wafer substrates were effectively transparent to mid-IR radiation, so that FTTR could be conveniently performed on differently-exposed areas of film using an IR microscope in transmittance mode. Whatever mode of its development, an effective resist must show high contrast for sharp demarcation between exposed and unexposed areas. For reasonable concentrations of PAG, the amount of strong acid produced in a CA resist increases only in gentie linear fashion with the UV dose supplied. Even after subsequent PEB treatment of a PDBFS/onium salt resist, the relationship between the content of its remaining ί-butyl ester groups, and its previous UV dosage, typically appeared as a still shallow sigmoid curve (Figure 2). The slow start here of ester cleavage at lower doses may be due to basic contaminants, such as perhaps this polymer's -CN end groups from the AIBN [31], that deactivated the first few units of acid that had been photogenerated. By 15 mJ/cm , increasing exposure had a stronger effect on further alteration of material, possibly because the new -COOH groups assisted in further tbutyl cleavage (autocatalysis). Beyond 30 mJ/cm , the curve leveled off again, perhaps because some ester groups were in regions of low onium salt or acid concentration, either through bulk heterogeneity (i.e. microphase separation), or uneven vertical distribution (i.e. relating to the less-soluble "skin" often seen in positive-tone relief development of CA resists [32]). 2

2


156


Even the steepest slope of this "% ester vs. dose" curve was rather low. Similar low contrast between more- and less-exposed areas would also be expected for material properties which vary rather linearly with material composition, such as density, refractive index, or dielectric constant. However, for other material properties or processes that depend more on composition being above or below a critical threshold, contrast may be much higher. Thus, film dissolution of PDBFS/onium salt resist varied between nil and complete within a range of 7 mJ/cm of UV energy and 40 mol % f-butyl ester content for negative-tone development with organic solvent, and 15 mJ/cm and 10 mol % for positive-tone with aqueous base (Figure 2). Indeed, the complete mechanism of film dissolution involves consecutive steps whose thermodynamics and/or kinetics can be highly cooperative: (i) wetting - solvent or other solution components must penetrate the liquid/solid interface; (ii) permeation - these must then diffuse into the film to surround a given chain; (iii) (optional) reaction polymer may have to be transformed, for example by deprotonation, to a more solventcompatible form; (iv) solvation - chains are completely separated from each other and transported away into the liquid phase. Even without eventual dissolution, the further alteration of a CA resist material during liquid functional development would also require at least the wetting then penetration of the resist, if not further reaction of its polymer, and thus would also be critically dependent on the composition of the latent image. Thus, even though a steadily increasing UV dose only gradually produces more acid and acid-cleaved polymer product, the onset of new or greatly altered materials properties upon functional development could also be remarkably sudden. This suggests that the same resist that shows high contrast in relief development would also be a good candidate for many forms of functional development as well. 2

2

Functional Development with Metal Ions: General. Metal cations can be captured and held inside a polymer material by ion-ion or ion-dipole interactions, or may simply remain after swelling the polymer with solution then evaporating the solvent. Since an ion-exchange mechanism seemed to promise the best rate and capacity for metals loading, we explored the formation of insoluble polymer-supported metal carboxylates from the carboxylic acid moieties of deprotected PDBFS. Other candidates for good metal binding would be other carboxyl-based resists [23-25], protected phenolic resins (Novolac or hydroxystyrene) with possible formation of metal phenolates, or polymers containing other photo- or photocatalyst-deprotectable chelating groups. Exposure-Controlled Sorption of Calcium Hydroxide. As expected, beyond a critical UV dose (30 mJ/cm ), aqueous solutions of either sodium or potassium hydroxides quickly dissolved exposure-deprotected areas of PDBFS resist, presumably due to formation of a water-soluble monocation-polycarboxylate polyelectrolyte. However, the same treated resist did not dissolve in even highly basic, saturated solutions of calcium hydroxide in water, though FTIR of the still-solid films (Figure 3) showed the appearance of peaks corresponding to carboxylate anion (ca. 1415 and 1565 cm-1, and asym COO") and water (3000-3500 cm-1) within the polymer matrix. Their increasing height with prolonged contact of Ca(OH)2/H20 correlated with further shrinking of remaining carboxylic acid and anhydride peaks (1700-1850 cm-1). The broad shapes of these peaks suggested the formation of a variety of carboxylate-water-metal ion complexes in different microenvironments. 2

δ ν Γ η


10. VEKSELMAN IT AL.


Since no other cation is present at such high pH, it is obvious that at least one C a is being taken up into the polymer for every two COO- being formed, to give molar proportions of styrene:fumarate:calcium of 55:45:45 (calculated 14 wt % calcium) from the completely deesterified PDBFS resist. Each divalent cation would thus ionically link two carboxylate groups, either intramolecularly (forming at least a 7membered ring - not particularly favored), or intermolecularly to crosslink the polymer and keep it from dissolving. The literature reports instances of similar polycarboxylic acids being precipitated through ionic crosslinking by polyvalent cations: depending on pH, gels formed this way contained tetra- or hexadentate complexes, with up to 1/3 of the carboxylic groups still remaining un-ionized [33,34]. In our thin films, rough kinetics of Ca(OH)2/H20 loading (see Figure 4) showed an initial delay without detectable formation of carboxylate anion, probably for the material surface to become sufficiently ionized and wetted, and possibly relating to the annoying "skin" often seen in positive-tone relief imaging [32]. Once begun, the reaction proceeded until, within a time comparable to or shorter than the induction period, all carboxylic acid and anhydride groups made available by UV/PEB had been converted to carboxylate, without hydrolysis of any more f-butyl ester groups. Even with this crosslinking species, the rate of material alteration was probably more limited by propagation of the reaction front than by ion movement into and through the nowhydrophilic polymer matrix, as has been suggested for other, even oligomeric crosslinkers entering similar maleic anhydride copolymers [12]. More complete initial deesterification shortened both these times and thus the total time for maximum functional development (180 s for 30 mJ/cm , to 10 s for 60 mJ/cm ), and also increased the ultimate ion-exchange capacity of the film. As with relief development of the same resist, this abrupt infiltration of solute beyond critical exposure and developing times benefited the contrast of the resulting functional images (maximum slopes of curves A to D, Figure 4). + +

2

2

Exposure-Controlled Sorption of Other Metal Ions. 0.1 M aqueous solutions of CuCl2, N1CI2, C0CI2 or Pb(OAc)2 did not wet a film of PDBFS resist, even after its complete photo-deprotection. These solutions were neutral or slighdy acidic (pH 7 to 4): it was expected that more basic conditions would be needed to deprotonate the polymer films for better wetting and influx of metal cations. Precipitation of metal oxides withrisingpH was often largely forestalled by using NH3 as base, since this is also a good ligand for many cations. Actually, aqueous ammonia alone was an effective solvent for fully-deprotected PDBFS (ex. for relief development). However, many of its mixtures with crosslinking polyvalent metal ions (likely containing NH3, NH4+ ( N H ) M \ Μ^ΟΗ-Οπι and OH- species) did not dissolve even fully-deprotected resist at higher pH, though FTTR peaks near 1550 and 1400 cm-1 showed increasing conversion to carboxylate anion with increasing contact time and pH (Figure 5). Ammonia or ammonium hydroxide probably forms ammonium carboxylate first, but then metal or metal-ammonia complex would displace the ammonium ion for a thermodynamically more favorable association (lattice energies for acetates: 725 kJ/mol for NH4+ but 2835 for Cu++ and 2247 for Pb++ [35]). With increasingly higher ammonium ion concentrations (as reflected by increasing pH), monocation carboxylate groups eventually predominate within the material, ultimately leading to its dissolution. Thus, at pH 6-7, N i and Cu * " began to penetrate exposed areas, which dissolved at pH 10-10.5. At pH 8 Ni " * or C u ammonia solution does not cause any IR-detectable alteration within unexposed areas. With Co * *, dissolution occurred at pH 9, and was accompanied by small transformation of unexposed areas of n4

3

m

+ +

4

4

4

4

+ +

4

4



158

1900

1700

1500

1300

1100

900

700

1

Wavenumber, cm"

Figure 3. IR spectra of PDBFS resist after UV exposure at different doses, and 20 s treatment with aquaeous saturated Ca(OH>2.

Figure 4. Conversion of -COOH into -COO* after treating the deprotected resist with aquaeous saturated Ca(OH)2 for 10 s (A), 20 s (Β), 30 s (C) and 180 s (D).


10.

VEKSELMAN ET AL.


resist (new FTIR peaks at 3500-3000, 1480 and 1380 cm-1). Pb++ only slowly penetrated exposed areas even up to the point of dissolution at pH 10, while clearly causing hydrolysis of f-butyl ester groups in unexposed areas at all pH (new FTTR peaks at 1540 and 1390 cm-1). Overall then, pH 8 seemed best for exposure-selective binding of these metal ions, though contrast was poor for P b . Brief attempts to introduce FTIR-detectable quantities of Fe ^, Fe , Al + or Pd++ failed, due to their precipitation in basic solutions. ++

4

3+

3

Metal Ion-Developed Resist: Reactive Ion Etching. Mainly silicon [10,1214], and some other elements that form refractive oxides (boron [36], tin [37], germanium [38], nickel and zirconium [5]), are used in microlithography in "dry development" processes, in which their compounds are selectively introduced either in exposed or unexposed areas of resist. The stable oxide layer produced under plasma bombardment protects polymer below. This technique gives excellent wall profile and high aspect (depth/width) ratio [9]. Preliminary tests showed that introducing of Ca++ into 0.8 μπι films of exposed resist did indeed reduce the rate of its O2-RIE plasma etching from 7.0 (ester without Ca) down to 1.5 nm/s (carboxylic Ca salt), to create 1 μπι-ΓεβοΚίΓίοη relief images. Sensitivities were 40 mJ/cm2 for 90 or 120 s of immersion in Ca(OH)2/H20, and 25 mJ/cm2for180 s, similar to sensitivities associated with wet development (Figure 4). Metal Ion-Developed Resist: Refractive Index. A waveguide consists of a channel of material with high n, surrounded by one with lower n. The value of η depends on material composition, being higher in materials containing more polarizable atoms (such as heavier atoms or ions) or groups (dipoles, delocalized charges). Exposure/PEB, then development with inorganic cations, each significantly and progressively increased the refractive index of PDBFS resist, as measured in different regions of the same film (Figure 6), through the formation of increasingly polarizable functional groups: RCOOR' < RCOOH (or (RCO^O) < (RCOO~) M . In principle, patterns of sharply- or gradually- varying refractive index could thus be generated, to 1 μπι resolution and below. n+

n

Functional Development with Amines. Amines would generally be expected to form ionic or covalent bonds with carboxylic acid and anhydride groups, and none with i-butyl esters. Many biomolecules contain such groups which could be used to immobilize them in exposed areas of PDBFS resist Functional development with volatile amines proved possible from the gas phase. As shown by FTIR (Figure 7), deprotected resist (curve B) easily absorbed vapours of ammonia within 120 s to form the same carboxylates seen in liquid development (curve E: 1708 cm-1 replaced by 3200,1560 and 1450 cm-1). Though pyridine was also taken up in these areas, reaching a maximum at one hour with no further change in next 2 hours, polymeric carboxyls were not deprotonated by this weaker base (curves C and D, respectively; note new pyridine peaks at 1596,1435,1010,750 and 700 cm-1, and no peaks of carboxylate). Depending on their concentrations, aqueous solutions of molecules with only one strongly basic amino group either did not wet the surface of photoexposed PDBFS resist, or dissolved it away completely [39]. Thus, ammonia, 2-amino- and 4-aminopyridine at concentrations above 0.5 wt % acted as relief developers. Carboxylic amides were not detectably formed under these conditions. However, di- and polyamines at concentrations above ca. 0.5 wt % were able to ionically crosslink the


160


1

J ό

ι 7

,

,

ι 8

ι 9

r

ι 10

ι 11

pH

Figure 5.Uptake of different metal ions by the resistfromaqueous ammonia solutions at different pH after 30 s. Hollow and filled dots represent uptake by -COOH and -COOR areas, respectively.

2+

Figure 6. Refractive indices of unexposed, deprotected and Ca -developped resist thin films, and possible patterned features.


10.

VEKSELMAN ET AL.


polymer as it was ionized, preventing its dissolution as they penetrated. Thus, FTIR spectra of deprotected resist treated with 2,6-diaminopyridine, 1,10-diaminodecane or poly(ethyleneimine) showed formation of carboxylate anion (same peaks as from metal polycarboxylates), and also amide (3100 and 1640-1660 cm-1), presumably by reaction with the minor anhydride component of the polymer, of diamine percolating in advance of water. No FTIR or other evidence was seen of any of these amines being absorbed by unexposed resist under these conditions, meaning good contrast for this kind of functional development. Functional Development with Dyes. Development methods based on weak solute-polymer interactions were investigated with selected UV-VIS absorbent and fluorescent dyes (Table I). Even a minimal uptake of such compounds would allow direct visualization of a binding pattern. Moreover, such dyes are often able to supply interesting information, through shifts in their spectra, about their general environment and specific interactions. Although some of the chosen dyes dissolved in water, no sorption was evident from purely aqueous solutions (Table I): drops rolled off both exposed and unexposed areas of resist, indicating its low wettability. However, the addition of different alcohols dramatically increased wettability of areas of exposed resist. Thus, solutions of methanoliwater above 1:10 v:v wetted these areas of film but, beyond 8-12:10 v:v, swelled and even dissolved them to form positive-tone relief images. No tbutanohwater mixture dissolved any portion of the resist material, and one of 5:10 v:v seemed a good general mixture to introduce dyes deep into deprotected resist (Figure 8, curves A-B), so that prolonged water rinsing could not entirely remove them (curve C). UV-VIS spectroscopies could follow the uptake of dye into the film (curves D, E), and also subsequent drying of the film (curve F). Though hexane wets the unexposed areas of resist, and not the exposed, it does not seem to penetrate beyond the surface (a dye applied from hexane was leached out later very quickly with the same solvent), and is a poor solvent for polar compounds such as R6G. Toluene is such a good solvent for neutral organic compounds that it acts as a negative-tone relief developer of the resist. Whereas 3-4:10 v:v toluene:hexane also dissolves unexposed areas of resist, a 1:10 ν:ν mixture only slightly swells them, though still enough to carry a dye deep into the film and trap it on drying so that hexane cannot wash it away later. More detail report on selective binding with organic molecules will be published elsewhere. Conclusion As the dimensions of electrical, optical, bio- or chemo- devices and their features continue to shrink, demand will continue to grow for micron-scale patternable systems. One way to create such structures is by relief imaging, and filling spaces with inert or contrasting material. Another way is to area-selectively modify resists to active materials with valuable properties. A variety of experimental approaches were shown for selective binding of organic and inorganic molecules into chemically amplified resists. Area-selective binding in positive or negative tone, from solution or vapor, and into bulk of a film or onto its surface, were successfully realized by choosing appropriate polymer structures, binding species, and solvents and other conditions.


162


τ

2000

1

1

1

1

1800 1600 1400 1200 1000 Wavenumber, cm

Γ

800

600

Figure 7. IR spectra of PDBFS based-resist, (A) before and ÇB) after deprotection, and after treatment with (C) vapours of ammonia for 2 min, or (D) pyridine for 60 min or (E) 180 min.

Table I. Functional development of PDFS resist withfluorescentdyes Effect on Effect on Dyes Solvent(s) v:v exposed resist unexposed resist R6G H 0 + (no swelling) MeOH:H 0 1:10 +++ (no swelling) iBuOH:H 0 5:10 +++ (partial swelling) MeOH:H20 10:10 (dissolves) MeOH:tfeO 20:10 DSMP H 0 +++ (partial swelling) MeOH:H 0 10:10 Toluene (dissolves) Hexane + (no swelling) Hexane:Toluene 1:10 +++ (slight swelling) (+++) extensive sorption; (+) slight sorption; (-) no wetting 2

2

2

2

2


10. VEKSELMAN ET AL.

Imaging with Chemically Amplified Resists

Absopbtion spectra ι Λ

163

A)-deprotected PDBFS B)- R6G in A * ^ \ C)-BinH O300s 2

Emission spectra (excitation at 355 nm). F Resist was in dye solution % D)-30s,E)-180s, F)- Ε with heating 130°C/300s 1

Excitation spectrum (emission at 580 nm). G)-sample Ε

300

400

500

600

700

Wavelength, nm Figure 8. UV-vis absorbance and fluorescence spectra of R6G in deprotected PDBFS resist.

Acknowledgments We thank L. Martinu and I. Sapieha (Ecole Polytechnique, Montreal) for plasma etching experiments, and T. Kanigan (McGill University, Montreal) for refractive index measurements. Literature Cited (1)

Reichmanis, E.; Houlihan, F. M.; Nalamasu, O.; Neenan, T. X . In: Polymers for Microelectronics; Thompson, L. F.; Willson, C. G.; Tagawa, S., Eds.; ACS Symp. Ser. 537; American Chemical Society: Washington, DC, USA, 1994, pp 2-24.


164 (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) (17) (18)

(19) (20) (21) (22) (23)

(24) (25)


Thompson, L. F. In: Introduction to microlithography; 2-nd ed.; Larry, F.; Thompson, C.; Grant, W.; Murrae, J. B., Eds.; American Chemical Society: Washington, DC, 1994, pp 1-18. Frechet, J. M . J.; Eichler, E.; Willson, C. G.; Ito, H. Polymer 1983, 24, 9951000. Dressick, W. G.; Dulcey, C. S.; Georger, J. H.; Calvert, J. M . Chem. Mater. 1993, 3, 148-150. Schilling, M . L., et al. Macromolecules 1995, 28, 110-115. Borman, S. Chemical and Engineering News 1994, 72(23), 24-25. Gorman, C. B.; Biebuyck, Η. Α.; Whitesides, G. M . Chem. Mater. 1995, 7, 252-254. Shirai, M.; Sumino, T.; Tsunooka, M. In: Polymeric Material for Microelectronic Application; Ito, H.; Tagawa, S.; Horie, K., Eds.; ACS Symp. Ser. 579; American Chemical Society: Washington, DC, 1994, pp 185-200. Baik, K.; Van den hove, L. In: Polymeric Material for Microelectronic Application; Ito, H.; Tagawa, S.; Horie, K., Eds.; ACS Symp. Ser. 579; American Chemical Society: Washington, DC, 1994, pp 201-218. Coopmans, F.; Roland, B. Proc. SPIE 1986, 633, 34-41. Zhang, C.; Darling, G. D.; Vekselman, A. M . Chem. Mater. 1995, 7, 850-855. Sebald, M . ; Ahne, H.; Leuschner, R.; Sezi, R. Polymers for Advanced Technologies 1994, 5, 41-48. MacDonald, S. Α.; Schlosser, H.; Ito, H.; Clecak, N. J.; Willson, C. G. Chem. Mater. 1991, 3, 435-442. MacDonald, S. Α.; Schlosser, H.; Clecak, N. J.; Willson, C. G. Chem. Mater. 1992, 4, 1364-1368. Allcock, H. R.; Nelson, C. J.; Coggio, W. D. Chem. Mater. 1994, 6, 516-524. Ito, H., et al. J. Photopolym. Sci. Technol 1993, 6, 547-562. Hinsberg, W. D.; MacDonald, S. Α.; Clecak, N. J.; Snyder, C. D. Chem. Mater 1994, 6, 481-488. Reichmanis, E.; Galvin, M . E.; Uhrich, K. E.; Mirau, P.; Heffner, S. A. In: Polymeric Material for Microelectronic Application; Ito, H.; Tagawa, S.; Horie, K., Eds.; ACS Symp. Ser. 579; American Chemical Society: Washington, DC, 1994, pp 52-69. Cutee, O. M.; Clemberg-Sapieha, J. E.; Martinu, L.; Wertheimer, M . R. Thin Solid Films 1990, 193/194, 155-163. Perrin, D. D.; Armarego, W. L. F. Purification of Laboratory Chemicals; 3-rd ed.; Pergamon Press: Oxford, England, 1988, p 318. Crivello, J. V. In: Initiators - Poly-reactions - Optical activity; Advances in polymer science 62; Spriner-Verlag: Berlin, 1982, pp 1-48. Swalen, J. D.; Tacke, M.; Santo, R.; Fisher, J. Optics Communications 1976, 18, 387-390. Allen, R. D.; Wallraff, G. M.; Hinsberg, W. D.; Simpson, L. L.; Kunz, R. R. In: Polymers for Microelectronics; Thompson, L. F.; Willson, C. G.; Tagawa, S., Eds.; ACS Symp. Ser. 537; American Chemical Society: Washington, DC, 1994, pp 165-177. Ito, H.; Willson, C. G.; Frechet, J. M .J.Proc. SPIE 1987, 771, 24-31. Ito, H.; Ueda, M.; Ebina, M. In: Polymers for Microlithography; Reichmanis, E.; MacDonald, S. Α.; Iwayanagi, T., Eds.; ACS Symp. Ser. 412; American Chemical Society: Washington, DC, USA, 1989, pp 57-73.


10.

VEKSELMAN ET AL.

Imaging with Chemically Amplified Resists

165

(26) Murata, M.; Yamachika, M.; Yumoto, Y.; Miura, T. German Patent DE 4229 816 A1, 1993 (27) Darling, G. D., et al. Proc. 3-rd Pacific Polymer Conference. Pol. Div. Royal Australian Chemical Institute: Gold Coast, Australia, 1993, 399-400. (28) Crivello, J. V.; Shim, S. Y. J. Polym. Sci. Part Α-Polymer Chemistry 1995, 33, 513-523. (29) Ito, H.; Padias, A. B.; Hall, H. K. J. J. Polym. Sci. Part A. Polym. Chem. 1989, 27, 2871-2881. (30) Ito, H.; Ueda, M . Macromolecules 1988, 21, 1475-1482. (31) Ito, H. Jpn. J. Appl. Phys., Part 1 1992, 31, 4273-4282. (32) MacDonald, S. Α., et al. Chem. Mater 1993, 5, 348-356. (33) Sileo, E. E.; Morando, P. J.; Baumgartner, E. C.; Blesa, M . A. Thermichimica Acta 1991, 184, 295-303. (34) Allan, J. R.; Bonner, J. G.; Gerrard, D. L.; Birnie, J. Thermochimica Acta 1991, 185, 295-302. (35) CRC Handbook of Chemistry and Physics; 68-th ed.; Weast, R. C., Ed.; CRC Press, Inc.: Boca Raton, FL, USA, 1987, p D-101. (36) Talor, G. N.; Stillwagon, L. E.; Venkatesan, T. J. Electrochem. Soc. 1984, 131, 1664-1670. (37) Nalamasu, O.; Baiocchi, F. Α.; Taylor, G. N. In: Polymers in microlithography; Reicmainis, E.; MacDonald, S. Α.; Iwayanagi, T., Eds.; ACS Symp. Ser. 412; American Chemical Society: Washington, DC, USA, 1989, pp 189-209. (38) Yoshida, Y . ; Fujioka, H.; Nakajima, H.; Kishimura, S.; Nagata, H. J. Photopolym. Sci. Technol. 1991, 4, 497-507. (39) Vekselman, A. M . ; Zhang, C.; Darling, G. D. Proc. 10th International Conference on Photopolymers. Willson, C. G., Ed. SPE: Ellenville, NY, USA, 1994, 116-127. RECEIVED July 20,

1995


Functional Imaging with Chemically Amplified Resists - ACS

Recommend Documents