Investigation of Fluorinated (Meth)Acrylate Monomers and

Jun 3, 2014 - The second method involves adding fluorinated surfactants to the UV-curable ... Heptadecafluoro-1,1,2,2-tetrahydrodecyl acrylate (CF3-te...
0 downloads 0 Views 3MB Size
Article pubs.acs.org/Langmuir

Investigation of Fluorinated (Meth)Acrylate Monomers and Macromonomers Suitable for a Hydroxy-Containing Acrylate Monomer in UV Nanoimprinting Shunya Ito,† Shu Kaneko,† Cheol Min Yun,† Kei Kobayashi,† and Masaru Nakagawa*,†,‡ †

Institute of Multidisciplinary Research for Advanced Materials, Tohoku University, 2-1-1 Katahira, Aoba-ku, Sendai 980-8577, Japan Japan Science and Technology Agency (JST), Core Research Evolutional Science and Technology (CREST), 7 Gobancho, Chiyoda-ku, Tokyo 102-0076, Japan



ABSTRACT: We investigated reactive fluorinated (meth)acrylate monomers and macromonomers that caused segregation at the cured resin surface of a viscous hydroxy-containing monomer, glycerol 1,3-diglycerolate diacrylate (GDD), and decreased the demolding energy in ultraviolet (UV) nanoimprinting with spincoated films under a condensable alternative chlorofluorocarbon gas atmosphere. The X-ray photoelectron spectroscopy and contact angle measurements used to determine the surface free energy suggested that a nonvolatile silicone-based methacrylate macromonomer with fluorinated alkyl groups segregated at the GDD-based cured resin surface and decreased the surface free energy, while fluorinated acrylate monomers hardly decreased the surface free energy because of their evaporation during the annealing of the spin-coated films. The average demolding energy of GDD-based cured resins with the macromonomer having fluorinated alkyl groups was smaller than that with the macromonomer having hydrocarbon alkyl groups. The fluorinated alkyl groups were responsible for decreasing the demolding energy rather than the polysiloxane main chains. We demonstrated that the GDD-based UV-curable resin with the fluorinated silicone-based macromonomer was suitable for step-and-repeat UV nanoimprinting with a bare silica mold, in addition to silica molds treated by chemical vapor surface modification with trifluoro-1,1,2,2-tetrahydropropyltrimethoxysilane (FAS3) and tridecafluoro-1,1,2,2tetrahydrooctyltrimethoxysilane (FAS13).

1. INTRODUCTION Ultraviolet nanoimprint lithography (UV-NIL) has received much attention because this new technology allows the fabrication of micrometer- and nanometer-scale structures that have small line edge roughness with sufficient throughput to meet industrial manufacturing needs.1−3 The fabrication of sub-10-nm device structures,4−6 moth-eye structures to decrease reflection,7,8 and 3D nanostructures9 have been demonstrated using ultraviolet (UV) nanoimprinting; the process comprises filling mold cavities with resin, curing by UV light exposure, and demolding. Applications of the UV-NIL technique have been demonstrated in static random access memory (SRAM),10 nanobiology,11,12 metamaterials,13 polarizers,14 and magnetic recording media15,16 and generally involve the subsequent removal of residual layers and dry etching of the substrates. Several types of defects have been observed to arise in the molded and cured resin patterns on substrates, including nonfill, particle, fringe, mouse-bite, and pull-out defects.2 Among these defects, mouse-bite and pull-out defects are thought to be generated during the demolding process owing to the adhesion of cured resin to the mold surface. Three main approaches have attempted to decrease the interactions between the mold and cured resin. The first approach is the modification of the mold surface with antisticking agents.17−20 © 2014 American Chemical Society

The antisticking agents form a release layer that decreases the surface free energy of the mold and reduces adhesive forces toward the cured resin. The second method involves adding fluorinated surfactants to the UV-curable resins. The fluorinated surfactants also decrease the surface free energy of the cured resin.21 The third approach employs both of the former methods in hopes that using antisticking agents on the mold surfaces in combination with fluorinated resin surfactants will promote a synergistic effect.22−24 Our group reported nonreactive fluorinated alcohol additives suitable for a UV-curable resin that was available for UV nanoimprinting under a condensable gas atmosphere of an alternative chlorofluorocarbon, 1,1,1,3,3-pentafluoropropane (PFP, HFC-245fa). The condensable gas eliminates nonfill defects generated by trapped gas bubbles because PFP condenses to the liquid state at an increased pressure of 0.15 MPa at room temperature.25 UV nanoimprinting in PFP has the unique advantages of rapid filling26 and the viscosity reduction27 of UV-curable resins, the reduction of demolding forces,28 and the resistance of resin component adsorption on fluorinated silica molds.29 We demonstrated that tridecafluoroReceived: April 28, 2014 Revised: June 2, 2014 Published: June 3, 2014 7127

dx.doi.org/10.1021/la501629n | Langmuir 2014, 30, 7127−7133

Langmuir

Article

1,1,2,2-tetrahydrooctan-1-ol functioned as the most effective additive to an acrylate-type UV-curable resin in air, while a longer fluoroalkyl chain made heptafluoro-1,1,2,2-tetrahydrodecan-1-ol the most effective additive in PFP.30 We further demonstrated that the Hansen’s solubility parameter was useful when identifying the base monomer suitable for a reactive fluorinated additive of hexadecafluoro-1,1,9-trihydrononyl acrylate. The reactive fluorinated acrylate caused segregation at the cured resin surface composed of glycerin 1,3dimethacrylate and a decrease in the surface free energy and demolding energy.31 We found that the addition of CF3terminated heptadecafluoro-1,1,2,2-tetrahydroacrylate to a UVcurable resin comprising a base monomer of glycerin 1,3dimethacrylate and a CHF2-terminated additive of hexadecafluoro-1,1,9-trihydrononyl acrylate allowed UV nanoimprinting with unmodified silica molds.32 The CHF2-terminated additive functioned as an auxiliary agent that improved the solubility of the CF3-terminated additive and caused the effective surface segregation of the fluorinated additives by the addition of a smaller amount of fluorine to the UV-curable resins. However, side effects have been found in recent studies of UV nanoimprinting under a PFP atmosphere, including the reduction of film thickness33 and resin pattern width34 after UV curing. We demonstrated a solution strategy.35 The absorption of PFP into the base monomers and UV-curable resins was confirmed by measuring the changes in weight upon exposure to a PFP atmosphere. The absorption of PFP is dependent on the monomer chemical structure. Using the solubility parameter calculated by the Hoy method, monomers with a solubility parameter of 20 (J cm−3)0.5 absorbed the most PFP, which caused a decrease in the transferred pattern thickness and a roughening of the outermost pattern surface. Before UV exposure, the UV-curable resin absorbed PFP, whereas after UV exposure the resin released PFP. We concluded that UV nanoimprinting under a PFP atmosphere requires the selection of base monomers with solubility parameters outside the range of 18−22 (J cm−3)0.5. In this study, we carried out step-and-repeat UV nanoimprinting under a PFP atmosphere using a viscous hydroxycontaining diacrylate monomer of glycerol 1,3-diglycerolate diacrylate (GDD) as a base monomer. The GDD monomer has a solubility parameter of 26.0 (J cm−3)0.5. We examined the quality of UV nanoimprinting when using a silica mold without and with chemical vapor surface modification by fluoroalkyltrimethoxysilanes. We also investigated fluorinated (meth)acrylate monomers and macromonomers that caused surface segregation in the viscous diacrylate monomer GDD and decreased the demolding energy in step-and-repeat UV nanoimprinting. The GDD-based cured resin films were examined by contact angle measurement to determine the surface free energies and by X-ray photoelectron spectroscopy (XPS) to determine the chemical composition at the cured resin surfaces. We selected suitable fluorinated additives based on their stability and average demolding energy values, which were monitored at every demolding process step in the stepand-repeat UV nanoimprinting up to 108 cycles.

Figure 1. Chemical structures of glycerol 1,3-diglycerolate diacrylate (base monomer, a), heptadecafluoro-1,1,2,2-tetrahydrodecyl acrylate (CF3-terminated acrylate, b), hexadecafluoro-1,1,9-trihydrononyl acrylate (CHF2-terminated acrylate, c), bis(n-butyl-terminated polytrifluoropropylmethylsiloxane)methylmethacryloxypropylsilane (MFSM15, fluorinated silicone-based macromonomer, d), and bis(n-butylterminated polydimethylsiloxane)methylmethacryloxypropylsilane symmetric macromer (MCS-M11, silicone-based macromonomer, e). Japan), and hexadecafluoro-1,1,9-trihydrononyl acrylate (CHF2terminated acrylate, c, Wako Chemical, Japan) were used as fluorinated monomer additives. Bis(n-butyl-terminated polytrifluoropropylmethylsiloxane)methylmethacryloxypropylsilane (fluorinated silicon-based macromonomer, MFS-M15, d, Gelest) and bis(n-buyl-terminated polydimethylsiloxane)methylmethacryloxypropylsilane symmetric macromer (silicon-based macromonomer, MCS-M11, e, Gelest) were used as macromonomer additives. The molecular weights of the macromonomers are in the range of 800−1000 g mol−1. All chemicals were used as received. Glass-transition temperatures of d and e were analyzed using a differential scanning calorimeter (DSC, SII Exstar X-DSC7000) over the temperature range of −50 to 160 °C. Five percent weight-loss temperatures of b, c, d, and e were measured by thermogravimetry with differential thermal analysis (TG/DTA) performed in air at a heating rate of 10 °C min−1 using a SII TG/ DTA7200. The chemical compositions of diluted UV-curable resin mixtures used for spin-coating are summarized in Table 1. A base resin, referred to as NL-SK1, comprising base diacrylate monomer a and a photoinitiator (2-methyl-1-[4-(methylthio)phenyl]-2-morpholio-1propanone, Irgacure 907, BASF Japan), was prepared by mixing using a magnetic stirrer. 1-Methoxy-2-propanol (PGME, SigmaAldrich) was added to NL-SK1 as a diluent solvent to adjust the thickness of the spin-coated film. An additive of the fluorinated (meth)acrylate monomers and macromonomers was mixed with a PGME solution of NL-SK1, abbreviated as a-1. The PGME solutions

2. EXPERIMENTAL SECTION 2.1. Preparation of UV-Curable Resins and Cured Films. Figure 1 shows the chemical structures of the agents used in this study. Glycerol 1,3-diglycerolate diacrylate (GDD, a, Sigma-Aldrich, Japan) was used as a base diacrylate monomer. Heptadecafluoro-1,1,2,2tetrahydrodecyl acrylate (CF3-terminated acrylate, b, Tokyo Chemical, 7128

dx.doi.org/10.1021/la501629n | Langmuir 2014, 30, 7127−7133

Langmuir

Article

calibrated by the C 1s peak (284.6 eV). The F/C and Si/C values were calculated using the areas of the respective peaks. The surface morphologies of the cured films were observed by using an atomic force microscope (AFM, SII S-image) in dynamic force mode with a cantilever (Olympus OMCL-AC200TS-R3). Both topographic and phase images were recorded. 2.3. Step-and-Repeat UV Nanoimprinting and Demolding Energy Monitoring. Step-and-repeat UV nanoimprinting was carried out under a PFP atmosphere using a UV nanoimprint stepper (Meisyo Kiko NM-0801) equipped with a load cell (Kyowa LUR-A-200SA1) for detecting respective stroke curves upon demolding. Release properties of the cured resin films were evaluated by measuring the demolding energies. The 6 in. silicon wafer substrates (Matsuzaki Seisakusyo) were cleaned using a UV/ozone cleaner for 20 min. The unmodified silica molds (10 × 10 × 0.6 mm3 silica mold, NTT-AT NIM-PH350, which contain line and space, hole, and pillar patterns with sizes ranging from 350 nm to 10 μm and a depth of 350 nm), were cleaned by exposure to vacuum UV light from a Xe excimer lamp emitting at 172 nm (Ushio UEM20-172) under a reduced pressure of 1.0 kPa for 15 min. To incorporate a release layer, the cleaned silica molds were further modified with a trimethoxysilane derivative by chemical vapor surface modification at 150 °C for 1 h.29 Tridecafluoro1,1,2,2-tetrahydrooctyltrimethoxysilane (FAS13, Gelest), trifluoro1,1,2,2-tetrahydropropyltrimethoxysilane (FAS3, Gelest), and 1,1,1,3,3,3-hexamethyldisilazane (HMDS, Tokyo Chemical) were used as surface modifiers. After 1-μm-thick UV-curable resin films were prepared on cleaned silicon wafers by spin-coating and annealing, an unmodified or modified silica mold was pressed onto the spincoated film at a speed of 2.0 mm min−1 under blowing PFP gas. The applied pressure was increased to 2.0 MPa in 10 s and then maintained at that pressure for another 10 s. The molded film was then cured by 365 nm UV light exposure at 100 mW cm−2 for 20 s. The mold was detached by moving the mold upward at a speed of 1.0 mm min−1. Thirty-six shots were performed on each 6 in. silicon wafer, and 108 shots were carried out separately for the four mold types. The respective demolding energies during the 108 cycles were determined from the areas of the respective stroke−force curves. The average demolding energies (Eave) were calculated for each shot except for the initial five shots on each wafer. The transfer patterns were observed by using an optical microscope (Olympus BX51) equipped with a lens (Olympus UPlanSApo 10 × /0.40) and a charge-coupled-device (CCD) camera (Hamamatsu ORCA-R2) and using a scanning electron microscope (SEM, Hitachi S-3000N).

Table 1. Chemical Compositions of UV-Curable Resin Mixtures Used for Spin-Coating weight percentage of components (wt %) in UV-curable resin additives UV-curable resin a-1 bc-1 bc-2 d-1 d-2 d-3 e-1 e-2 e-3

GDD a

Irgacure 907

PGME

b

c

d

e

19.4 19.1 18.9 19.4 19.4 19.3 19.4 19.4 19.3

0.6 0.6 0.6 0.6 0.6 0.6 0.6 0.6 0.6

80.0 78.9 78.0 80.0 79.9 79.8 80.0 79.9 79.8

0.0 1.0 1.5 0 0 0 0 0 0

0.0 0.4 1.0 0 0 0 0 0 0

0 0 0 0.02 0.10 0.20 0 0 0

0 0 0 0 0 0 0.02 0.10 0.20

with respective additives were named bc-1, bc-2, d-1, d-2, d-3, e-1, e-2, and e-3. The letters indicate the kind of fluorinated agents, and the numbers indicate the respective samples containing different amounts of fluorinated agents. The respective PGME solutions were spin-coated onto square pieces of silicon wafer (20 × 20 × 0.5 mm3) which were precleaned using a UV/ozone cleaner (Sen Light PL16-110) for 20 min. The spincoated films were annealed on a hot stage maintained at 80 °C for 2 min to evaporate PGME and were subsequently cured by exposure to UV light at an intensity of 100 mW cm−2 monitored at 365 nm for 20 s under a nitrogen atmosphere. The light intensity was measured using an optical power meter (Hamamatsu Photonics C-6080-03). A 200 W Hg−Xe lamp (San-ei Electric Supercure 203s) was used as the UV light source. The cured films, with thicknesses of approximately 1 μm, were then prepared for contact angle measurement and XPS analysis. The calculated chemical compositions of the cured resin films after the removal of PGME are summarized in Table 2.

Table 2. Calculated Chemical Compositions and Surface Free Energies of Cured Films after Solvent Removal weight percentage of components (wt %) additive cured film a-1 bc-1 bc-2 d-1 d-2 d-3 e-1 e-2 e-3

GDD a

Irgacure 907

b

c

d

e

surface free energy (mJ m−2)

96.9 90.4 85.7 96.8 96.4 95.9 96.8 96.4 95.9

3.1 2.9 2.7 3.1 3.1 3.1 3.1 3.1 3.1

0 4.8 7.0 0 0 0 0 0 0

0 1.9 4.6 0 0 0 0 0 0

0 0 0 0.1 0.5 1.0 0 0 0

0 0 0 0 0 0 0.1 0.5 1.0

57.6 58.0 57.9 25.1 24.5 29.6 26.5 29.5 27.7

3. RESULTS AND DISCUSSION 3.1. Changes to Surface Free Energies by the Incorporation of Additives. According to our recent study,32 CF3-terminated acrylate monomer b effectively assembled at the surface of the cured film comprising a glycerin 1,3-dimethacrylate (GDM: solubility parameter, 22.0 (J cm−3)0.5) base monomer and a Irgacure 907 photoinitiator in the presence of CHF2-terminated acrylate monomer c. CHF2terminated acrylate monomer c functioned as an auxiliary agent to improve the solubility of CF3-terminated acrylate monomer b in the hydroxy-containing base monomer and to cause the segregation of the fluoroalkyl moieties at the cured film surface at a small fluorine atomic percentage. The UV-curable resin (2.7 F at. %) comprising ternary monomers, namely, 83.0 wt % GDM, 7.5 wt % b, 4.3 wt % c, and 5.2 wt % Irgacure 907, formed a cured film showing a low surface free energy of 22.8 mJ m−2 and allowed step-and-repeat UV nanoimprinting using unmodified silica molds under a PFP atmosphere. We considered these experimental results and applied the combination of b and c to the similar hydroxy-containing diacrylate monomer of GDD with a solubility parameter of 26.0 (J cm−3)0.5.

2.2. Surface Analyses by Contact Angle Measurement, XPS, and AFM. Contact angles of two probe liquids, water and diiodomethane, were measured for all of the cured resin films by the sessile drop method (Kyowa Interface Science CA-X), and the surface free energies of the cured resin films were calculated according to the Owens−Wendt equation.36 Values reported by Tang et al.37 for the surface tension dispersion (γd) and polarity (γp) terms were used in the calculation as follows: γd = 21.8 mJ m−2 and γp = 51.0 mJ m−2 for water and γd = 48.5 mJ m−2 and γp = 2.3 mJ m−2 for diiodomethane. The fluorine to carbon (F/C) and silicon to carbon (Si/C) ratios at the cured film surfaces were determined from XPS measurements. The XPS spectra were obtained at takeoff angles of 15 and 90° using a ULVAC-PHI 5600 with a monochromatic Al Kα source operated at a voltage of 14 kV and a current of 200 mA. The binding energies were 7129

dx.doi.org/10.1021/la501629n | Langmuir 2014, 30, 7127−7133

Langmuir

Article

fluorinated acrylate monomers were evaporated during the annealing process. Other fluorinated additives with large molecular weights and nonvolatile properties were required for the highly viscous GDD that requires spin-coating followed by annealing. 3.2. Surface Segregation of Silicon-Based Macromonomers. It has been reported that diblock copolymers of poly(styrene)-block-poly(dimethylsiloxane)38 and polymethacrylate-block-poly(dimethylsiloxane)39 form microphase-separated nanostructures through self-assembly driven by the segregation force. Considering this, we selected methacrylatetype silicon-based macromonomers with and without fluorinated alkyl groups. Thermal analysis by DSC and TG/DTA revealed that the d and e macromonomers were amorphous and nonvolatile with glass-transition temperatures of