materials Article
An Investigation of Polyoxometalate Hybrid Materials as Patternable Dielectrics and Lithographic Resists Brandon Hardie and Mark Roll * Department of Chemical and Materials Engineering, University of Idaho, Moscow, ID 83844, USA;
[email protected] * Correspondence:
[email protected]; Tel.: +1-208-885-6196 Received: 1 November 2017; Accepted: 9 November 2017; Published: 15 November 2017
Abstract: Polyoxometalate (POM) hybrid materials have shown potential as spin-coatable, patternable dielectric thin-films and components for lithographic resists. In particular, the octamolybdate cluster has been shown to possess good spin-coating properties and the patterning capabilities of hybrid octamolybdate thin-films were explored using a combination of broadband UV and electron beam lithography (EBL) techiniques. Dielectric properties of these films were determined by ellipsometry, and octamolybdate clusters were subsequently investigated as negative resists in various blends for potential uses in next-generation photolithography, where contrast, sensitivity, and line edge roughness characteristics were determined. Preliminary evidence for the suppression of the diffusion of photo-generated acids is presented. Keywords: polyoxometalate; hybrid materials; dielectrics; photolithography
1. Introduction Applications utilizing polyoxometalate hybrid material chemistry have continued to make interesting strides over recent years. Ranging from catalysis to nanotechnology and medical science, these novel materials continue to show vast potential in many fields of study [1–4]. The work outlined within this article focuses primarily on polyoxometalates (POMs) as photo-active thin-films. The initial goal of this project was to assess the capabilities of POMs as photoresist components for next-generation photolithography. Spin-coatable hybrid materials, especially their inorganic component traits, have consistently been sought after for properties such as etch resistance and reduced pattern swelling [5,6]. Also of interest for next-generation photolithography, metal-containing resists more strongly absorb light at extreme ultra-violet (EUV; 13 nm) wavelengths [7]. While several POMs were explored over the duration of this project [8,9], the octamolybdate (Mo8 ) POM cluster, [α-Mo8 O26 ]4− [10] was directly spin-coatable and patternable under UV exposure, properties that best furthered the primary thin-film patterning objective. Multiple patterning methods were utilized, beginning with high exposure doses from a UV curing lamp, to more controlled, conventional instrumentation such as e-beam lithography (EBL) and a UV mask aligner. Mo8 films were found to act as negative-tone resists both alone and in resist blends. Lithographic properties such as contrast, sensitivity, and pattern line edge roughness (LER)/resolution were determined through this experimental effort. A combination of profilometry—for physical measurements of film thickness—and electron microscopy—for detailed imaging—were used to characterize these films. An additional topic of interest throughout the progress of this work was patternable high dielectric thin-films. Optical/dielectric properties were obtained by modeling these novel materials using ellipsometry data. The findings within this research have begun some foundations of novel POM applications and may be built upon in the near future to realize new potential applications. Materials 2017, 10, 1309; doi:10.3390/ma10111309
www.mdpi.com/journal/materials
Materials 2017, 10, 1309
2 of 12
research have begun some foundations of novel POM applications and may be built upon in the near Materials 2017, 10, 1309 2 of 12 future to realize new potential applications. 2. 2. Materials Materials and and Methods Methods 2.1. Octamolybdate Synthesis and Blends The synthesis of the Mo88 clusters clusters (Figure (Figure 1) 1) was a facile procedure accomplished according to literature [10]. A solution of 4.60 g (14.3 mmol) tetrabutylammonium bromide (Matrix Scientific) in (4.04 mmol) mmol) ammonium ammonium heptamolybdate 14 mL water was added dropwise to a solution of 5.00 g (4.04 tetrahydrate (Strem Chemicals) dissolved in ~20 mL water while stirring. A tetrahydrate Chemicals) dissolved ~20 A white precipitate was immediately formed formed and andthe thesolution solutionwas wasstirred stirred 5 minutes. The product was recovered as a forfor 5 min. The product was recovered as a colorless colorless powder, collected via filtration and rinsed with water three times (once with heated water powder, collected via filtration and rinsed with water three times (once with heated water to ensure to excesswere reactants wereProducts removed). Products were confirmed elemental analysis anyensure excessany reactants removed). were confirmed with elementalwith analysis (Appendix A.1) (Appendix A.1) using an elemental Exeter CE-440 elemental analyzer. Inadditional some cases,materials additional materials were using an Exeter CE-440 analyzer. In some cases, were blended in blended in with the Mo 8 materials to enhance the mechanical and/or lithographic properties of the with the Mo8 materials to enhance the mechanical and/or lithographic properties of the overall overall These materials included both photoacid generators (PAGs triphenylsulfonium system. system. These materials included both photoacid generators (PAGs triphenylsulfonium triflate triflate from Synquest Laboratories and N-hydroxynaphthalimide triflate from Sigma Aldrich; from Synquest Laboratories and N-hydroxynaphthalimide triflate from Sigma Aldrich; TPSTF and TPSTF and NHNT, respectively) as well as a molecular tetra-epoxide (tetraphenylolethane glycidyl NHNT, respectively) as well as a molecular tetra-epoxide (tetraphenylolethane glycidyl ether from ether Sigma Aldrich, 4-EP).were Materials were used as received. Sigmafrom Aldrich, 4-EP). Materials used as received.
Figure 1. Mo88 cluster cluster with with tetrabutylammonium tetrabutylammonium cation cation [11].
2.2. 2.2. Spin-Coating Spin-Coating Optimization Optimization While found that that acetonitrile acetonitrile While several several conventional conventional casting casting solvents solvents were were employed, employed, it it was was found (Fisher Chemical)was wasthethe most suitable casting solvent Mo8 materials, given the poor (Fisher Chemical) most suitable casting solvent for thefor Mothe 8 materials, given the poor solubility solubility in other solvents. Materials were coated on Si wafers using a Laurell Technologies WS-650 in other solvents. Materials were coated on Si wafers using a Laurell Technologies WS-650 Series Spin Series SpinatProcessor at the following settings: 100 rpm for (165 10 s,rpm/s) 615 rpm for Processor the following settings: 100 rpm (20 rpm/s) for(20 10 rpm/s) s, 615 rpm for(165 35 s,rpm/s) 1000 rpm 35 s, 1000 rpm (500 rpm/s) for 1 min. Physical blends were identified based on their solution (500 rpm/s) for 1 min. Physical blends were identified based on their solution compatibility with Mo8 compatibility 8 inspin acetonitrile. The same settings werefilms. used Solutions for these blended films. in acetonitrile.with The Mo same settings were used spin for these blended were typically Solutions were madewith at 5% wt. in systems acetonitrile, withfrom blended systems ranging from to 0–100% made at 5% wt. typically in acetonitrile, blended ranging 0–100% Mo8 ratios (relative 4-EP). Mo 8 ratios (relative to 4-EP). PAG contents in these systems were examined at both 5 and 10% mol PAG contents in these systems were examined at both 5 and 10% mol based on solids in the system. based on solids in the system. 2.3. Patterning and Development Methods 2.3. Patterning and Development Methods Several methods of patterning/lithography were used to examine the behavior of Mo8 materials. Several methods patterning/lithography examinetone, the behavior of Mo8 materials. For each technique, Moof8 alone and in resist blendswere actedused withtonegative where unexposed portions For each technique, Mo8 alone and ina resist blends actedstep. with Initially, negative an tone, where unexposed of material were removed following wet development industrial UV curing portions of material were removed following a wet development step. Initially, industrial UV lamp (Fusion UV medium pressure mercury lamp; ~200–400 nm wavelength) was an implemented for 2 curing lamp (Fusionranging UV medium pressuredoses mercury ~200–400 nm wavelength) was rapid film exposures from estimated up to lamp; 450 J/cm . Mo8 films were also studied implemented for rapid film exposureslithography ranging from estimated doses up tomethods 450 J/cm2utilized . Mo8 films were using more general, dose-controlled tools. The first of these electron also usingonmore general, lithography The first of these beamstudied lithography a Zeiss Supra dose-controlled 35 VP FEG Scanning Electrontools. Microscope. Films were methods typically 2 . Exposures utilized beam lithography Zeiss VP FEGrange Scanning Electron Microscope. Films exposedelectron at an accelerating voltageon ofa20 keVSupra over 35 a dosage of 5–2500 µC/cm 2. were accelerating voltage of nm) 20 keV over a dosage μC/cm usingtypically a Quintelexposed Q-4000 at UVanmask aligner (~365–412 gave feature sizes range as lowofas5–2500 ~50 µm, given ◦ mask availability. All patterning methods were followed by a post exposure bake (PEB) of ~90 C for
Materials 2017, 10, 1309
3 of 12
Materials 2017, 10, 1309
3 of 12
Exposures using a Quintel Q-4000 UV mask aligner (~365–412 nm) gave feature sizes as low as ~50 μm, given mask availability. All patterning methods were followed by a post exposure bake ~3(PEB) min. of The development of Mo only films was with agitation with for ~1–3 min ~90°C for ~3 min. The8 development of accomplished Mo8 only films wasminor accomplished minor inagitation a deionized water bath. For blended Mo films, a 30-s ethyl acetate bath was used, followed for ~1–3 minutes in a deionized water 8 bath. For blended Mo8 films, a 30-s ethyl acetate bathby anwas isopropyl alcohol by rinse to the final deionized used, followed an prior isopropyl alcohol rinse priorwater to thebath. final deionized water bath.
2.4. 2.4.Film FilmCharacterization Characterization Optical examinedusing usingaaJA JAWoollam WoollamVASE VASEellipsometer. ellipsometer. Refractive Opticaldata dataof ofthe theMo Mo88 films was examined Refractive index were obtained obtainedacross acrossaawavelength wavelengthrange range 300–1000 index(n) (n)and andextinction extinctioncoefficient coefficient (k) (k) data were ofof 300–1000 nmnm ◦ angle angleofofincidence. incidence. In In addition addition to at at varying atata a6565° to ellipsometry, ellipsometry,measurements measurementsofoffilm filmthicknesses thicknesses varying exposuredoses doseswere wereobtained obtained using using a Tencor Alpha use of of these exposure Alpha Step Step200 200Profilometer. Profilometer.The Theprimary primary use these physicalmeasurements measurements was was to to construct construct contrast over a range of of physical contrast curves curvesand andresist resistsensitivity sensitivitydata data over a range exposuredoses dosesfor forboth bothEBL EBLand and UV UV octamolybdate octamolybdate blends. exposure blends. 2.5.SEM SEMImaging Imaging 2.5. Utilizingaatypical typical accelerating accelerating voltage voltage of 8 films were Utilizing of 55 keV, keV, higher higher resolution resolutionimages imagesofofMo Mo were 8 films obtained.These Theseimages imagesallowed allowed a more in-depth look at the morphology resolution/line obtained. a more in-depth look at the filmfilm morphology andand resolution/line edge edge roughness of patterned Mo 8 films, as well as a top-down perspective of photoacid diffusion roughness of patterned Mo8 films, as well as a top-down perspective of photoacid diffusion within within the films.the films. Resultsand andDiscussion Discussion 3.3.Results 3.1.Spin-Coating Spin-CoatingOptimization Optimization 3.1. Aspreviously previously mentioned, mentioned, acetonitrile acetonitrile as film As as aa casting casting solvent solventprovided providedthe thebest bestuniform uniform film distribution (as seen in Figure 2), while also catering—solubility-wise—to all involved resist distribution (as seen in Figure 2), while catering—solubility-wise—to all involved resist components. Given above, a colorless Mo8 Mo raw8 product resulted in a blue components. Given the thesettings settingsoutlined outlined above, a colorless raw product resulted in color a blue due to thin-film interference, though other colors were observed over a range of spin settings and color due to thin-film interference, though other colors were observed over a range of spin settings and resist compositions. Octamolybdate POMs by themselves were particularly insoluble in many resist compositions. Octamolybdate POMs by themselves were particularly insoluble in many common common castingincluding solvents, PGMEA includingand PGMEA and cyclopentanone. Mo8 was also soluble in hot casting solvents, cyclopentanone. While MoWhile 8 was also soluble in hot propylene propylene carbonate wt. %), poor spin-coating results were obtained. While acetonitrile provided carbonate (
