Project Topics
Get Complete Project Material File(s) Now! »
Nanofabrication techniques and Nano Imprint Lithography (NIL)
From the early ages, technics to cut, sculpt, etch, mold, assemble pieces of matter have been developed and constantly optimized to satisfy the growing demand for functional materials. Since the inception of nanotechnology, these operations have to be mastered at the nanoscale. For these tasks, many top-down and bottom-up methods exist. However, they do not simultaneously fulfill all the necessary criteria of performance such as spatial resolution, pattern complexity, hierarchy, scalability, dimensionality, cost-effectiveness, a span of processable materials. Thus, motivations to optimize them and develop new ones persist as a flourishing domain of research and development.
Many materials exhibiting various intrinsic properties (mechanical, chemical, electrical, optical, thermal, etc.) are exploited in numberless functions once nanostructured onto a surface. Amongst them, metal oxides are extremely valuable for their extreme chemical, mechanical, and thermal stability and range of physical-chemical properties. Thanks to its hardness, chemical inertness, transparency, and low background fluorescence, glass is one of the preferred choices for micro-and nano-fluidics device fabrication. In photonics, metasurfaces require optical properties that are found in dielectrics such as SiO2 often combined with high refractive index dielectric TiO2 or (plasmonic) gold[1]. For nano-electronics, silica remains one of the key dielectric materials. Micro-mechanical scaffolds or persistent data storage systems need extremely stable materials such as TiO2[2].
For planar nanofabrication, either bottom-up (self-assembly) or top-down (lithography, etching) methods exist. First, bottom-up nano-structuration approaches based on micellar block copolymers self-assembly have shown great successes during the last decade[3]. When combined with sol-gel chemistry and liquid deposition processes, periodically organized mesoporous inorganic and hybrid thin films were obtained through the so-called micellar templating method[4] or Evaporation-Induced Self-Assembly (EISA)[5] (see Appendix 1). A large variety of templating agents can be coupled with inorganic polymerization reactions for the design of periodically organized nanostructured hybrid phases. Careful control over the physical-chemical parameters leads to the elaboration of metal-oxide nanopatterns[6],[7]. A plethora of combinations have been reported in the past 20 years and are available in many metal-oxides with different structures such as lamellar, 2D hexagonal, or 3D (fcc, bcc, gyroid, grid-like, worm-like, etc.). Achievable periods and/or feature dimensions ranging from several nanometers to several tens of nanometers. However, the patterns adopt a periodical arrangement and exhibit many defects, no spatial control (position and direction), and poor long-range ordering limiting their applications.
Amongst the top technological achievements of the last century in nanofabrication, photon-assisted processing methods are with no doubt the more technologically advanced ones. Photolithography combined with a dry or wet etching of the hard metal oxides is today available with acceptable performances in terms of quality, versatility, resolution, throughput, cost, and achievable morphologies[8–11]. Still, these methods remain cumbersome and are constrained by a limited choice of required photosensitive selective masks containing complex organic radical or cationic non-linear absorbers (photo-initiators) which are expensive, often colored, poorly soluble in photopolymers, and could be toxic[12]. Photolithography faces an intrinsic critical limitation in resolution dictated by the wavelength of incident writing light waves. High-energy ultraviolet radiation (deep-UV) is used nowadays in production lines to elaborate features smaller than 40 nm[13]. For this kind of lithographic method, hard-masks are needed for replication. They are generally fabricated using Electron Beam Lithography (EBL) in photosensitive materials[14]. Another top-down approach is direct writing with a focused ions beam (FIB)[15] where accelerated ions sputter locally on the surface of the target material. These two methods provide a high lateral resolution but they are time-consuming, hardly-scalable rendering them cost-effective only for the elaboration of masters and masks.
Subtractive techniques involving laser/material interactions sharing similar non-linear optical effects but at much higher powers, are also available and are used to etch dense metal oxide surfaces. For such inorganic materials, these effects are initiated by multiphoton absorption and tunneling ionization, promoting electrons to the conduction band. It is then followed by an avalanche ionization when a certain degree of ionization is reached. At such a high ionization density, the free-electron plasma absorbs the laser radiation and transfers the energy to the lattice by non-radiative relaxation, inducing non-reversible breakdowns in the material. This approach has the advantage of ablating hard transparent materials such as glass at the nanoscale[16]. Advanced laser technologies, such as ultrafast Bessel beams, have also been used to attain processing accuracy beyond the diffraction limit, well into the nanoscale domain[17]. In both examples, the high anisotropy of the channel (depth-to-width) is due to the typical anisotropy of the voxel beam and leads to poor axial resolution with respect to the lateral one. Indeed, diffraction and nonlinear self-focusing effects result in a non-punctual focalization, limiting the process to anisotropic motifs. This may be improved by shaping the beam, which remains a complex task to achieve[18].
Moreover, soft lithography (as a non-photolithographic strategy) has been extensively developed since the 19th [19]. Soft lithography englobes different techniques depending on the mold used in a stamping process or filled as a template. Nano Imprint Lithography (NIL) is one of the soft lithography methods, applicable to a large number of materials, appealing thanks to its high performance associated with ease of implementation. The first work on NIL was developed by Stephen Chou and coworkers in the 1990s when its team reported 20 nm patterns, initially etched in a silicon mold, transferred in a polymeric resist[20],[21]. This experiment was performed by strongly pressing the mold (P=50-100 bars) while heating at 100 °C to deform and shape the thermoplastic polymer. This study opened the way to soft NIL at the sub-100 nm scale. Imprinting of sol-gel resists was first demonstrated by Lukosz and coworkers in the 1980s at high pressure (P=280-500 bars) through hot embossing[22]. Even if the pressure used to imprint the sol-gel resist was rapidly reduced, it represented one of the major drawbacks of this technique. Indeed, silicon substrates used in microelectronics can easily break if the applied pressure is not perfectly homogeneous over the wafer surface. The use of hard and brittle molds (mostly silicon or fused silica) presented several downsides: (i) resist often sticks to the mold surface due to its high surface energy (ii) long range deformation can easily break the mold (iii) solvent and air bubbles cannot escape once trapped under the mold (iv) rigid molds cannot accommodate any curvature or defects. Due to all those limitations, costly hard molds could be used only a very limited number of times. The emergence of soft-NIL (Figure 1-1), by Whitesides and coworkers, tackled all those problems by the introduction of polydimethylsiloxane (PDMS) elastomeric molds[19]. Soft molds have low surface energy (ΥPDMS= 21 mJ.m-2 compared to ΥSi= 52 mJ.m-2), are permeable to solvent vapors, and can deform to follow the features of the substrate. PDMS is less expensive with respect to hard molds and can be used dozen times. Soft-NIL allowed to drastically reduce the pressure (P < 5 bars) and temperature applied to emboss the resist layer. With the huge improvements made in the past decades, Soft-NIL is currently performed at ambient pressure and room temperature and is now reaching sub-10 nm resolution[21] on sol-gel resists when performed with optimized PDMS molds.
PDMS stamps are generally fabricated from an initial master as a mold, stamp, or template. From a single master, generally made by an expensive and low throughput technique – such as EBL, FIB, or laser ablation on a silicon wafer – it is possible to easily fabricate many inexpensive PDMS molds bearing the same features. The fabrication of the molds is performed by using a prepolymer – generally composed of vinyl functionalized linear di-methyl-siloxane and a silicon-hydride functionalized linear di-methyl-siloxane – mixed with a catalyst – generally a platinum compound – and poured onto a prepared master as a pre-patterned surface. The cross-linking reaction, called hydrosilylation, is catalyzed by platinum and involves the addition of Si-H bonds across unsaturated bonds.
The transition of the prepolymer from a viscous liquid to a solid elastomer takes usually several hours at room temperature. It is generally accelerated by heating at a mild temperature for instance 70 °C. The hydrolyzation cross-linking route is generally preferred as no by-products are formed and it offers a comfortable time working window before extended cross-linking occurs [23,24]. Once fully cross-linked, the resulting mold is easily demolded from the master thanks to its elasticity (Figure 1-2). The mold perfectly replicates the initial master features (relative shrinkage between master and mold can be neglected). Many molds can be fabricated from a single master. However, with extensive use, defects will start to appear on the master owing to the local adhesion of PDMS during mold release. For this reason, the surface of the master (for instance often made of silicon) is functionalized (e.g. with fluoroalkyl silanes) to decrease the surface energy, reducing the interactions with the PDMS chains.
When patterns with features smaller than 100 nm are needed, the excessive softness of PDMS can induce deformations of the mold shape while printing. This phenomenon has been widely studied both theoretically and experimentally[25,26]. As first demonstrated by Delamarche and coworkers[25] only a fraction of the structures accessible by nanofabrication in hard materials will give stable structures in PDMS stamps. If the aspect ratio (AR = height/width) of the motifs is too low, insufficient relief exists on the surface of the stamp to withstand the compressive forces due to the adhesion between stamp and substrate. Thus, no patterns are achievable. Moreover, if the AR is too high, structures can collapse under their weight[25],[26]. High AR feature causes loss of structural integrity of the feature in a mold. During the stamping process, capillary forces can be strong enough to deform the stamp features and contact two neighboring structures. This is the case for densely packed structures, long grating lines, or tightly spaced pillars. These structures are prone to pairing due to their high surface-to-volume ratio. Once collapsed it is hard to restore the PDMS features. Different ways of restoring the structures have been explored such as washing with low surface tension solvents or with water-containing surfactants but with little success[25],[27]. The easiest way to prevent the deformation of small patterns is to increase the rigidity of the PDMS material. Such PDMS, called hard-PDMS (h-PDMS), has been firstly developed by Schmid and coworkers, in 2000[28]. Hard-PDMS uses branched siloxanes, shorter cross-linkers, and a higher concentration of reactive groups compared to standard PDMS to increase the cross-linking density and the rigidity between the cross-linking points. The different concentrations in reactants and their ratios have been extensively studied[27] and commercial products are now available. PDMS hardness is estimated through its Young’s modulus that is about 20-30 MPa for conventional PDMS whereas it goes to 60 to 70 MPa for h-PDMS. Other materials than PDMS can be used, such as siloxane-composites[29] or thermo-plastic polymers[30]. However, they miss the intrinsic advantages of PDMS. Similarly, using pure h-PDMS has its limitations: (i) a high-modulus polymer can increase the stress during the curing step, (ii) difficult rendering to separate it from the master, (iii) h-PDMS has low flexibility, (iv) making conformal contact on rough surfaces rather difficult. The increased rigidity leads also to a poor accommodation to inhomogeneous stress during imprinting that might lead to plastic deformation or cracking of the stamp or of the substrate. In order to increase the rigidity and, at the same time, keep the advantages of low-modulus PDMS, hybrid stamps have been developed with a thin h-PDMS layer of few microns that contains the patterns and a thick layer of standard PDMS of about 0.5 mm (Figure 1-2). In addition, to minimize in-plane deformation, for easier handling and automatization, this assembly (standard PDMS + hard PDMS) can then be backed by a thin glass plate.
Figure 1-2. a) Schematic representation of the mold preparation. The precursor is poured onto the hydrophobized surface of the master and cured before demolding. b) Picture of a silicon master and its negative replica into a flexible mold.
Applying soft-NIL processes to directly imprint sol-gel metal oxides is highly interesting. However, it is quasi-unachievable due to two limitations: the impossibility to deform a brittle ceramic film and the difficulty to reach a plastic state at moderate temperature. A promising soft-NIL method on metal oxides exploits as-deposited sol-gel layers (xerogels). Xerogels are poorly cross-linked networks of sol-gel, organometallic precursors, or hydrolyzed intermediates. Such gels remain flexible and will stiffen from advancing polycondensation during drying and aging of the films and finally with thermal annealing to form the metal oxides[7,31,32]. The properties of xerogels can be exploited in soft-NIL at two conditions: (i) that imprinting takes place when the layer is in a gel-like phase and (ii) the demolding is made after sufficient solidification. Both these controls are challenging in sol-gel since the as-deposited layers undergo more or less fast cross-linking upon drying and mild thermal activation. These phenomena are generally accompanied by the release of water or solvent molecules that are trapped within the materials underneath the mold, affecting the replica morphology, density, and shrinkage factor. Eventually, the demolded material is converted into a purely inorganic and rigid metal oxide after calcination.
Sol-gel approaches rely on the use of polymeric or organic stabilizing agents in order to adjust the rheology, the precursor reactivity and achieve a good resolution and a large geometrical aspect ratio of the final printed 3D structures. The main drawback associated with this approach is the severe shrinkage following the evaporation of organic moieties during the annealing step (up to 60-80% vol)[33–36]. Sol-gel formulations containing metal oxide nanoparticles (NPs) can be directly imprinted resulting in a reduced shrinkage[36] and materials having lower density (resulting from the presence of pores associated with the NP packing). A compromise thus needs to be found between lower shrinkage and denser material. Of course, stabilizing-agent-free xerogels can be used allowing for a reduced shrinkage factor. However, the very high viscosity of the intermediate xerogel will prevent efficient mass transfer within the sharp features of the mold.
Figure 1-3. Scheme presenting the Degassed Assisted Patterning (DAP) process applied to the nano impression of sol-gel layers[61].
To promote the filling of the mold features, a technique called Degassed Assisted Patterning (DAP), taking advantage of the properties of PDMS to enhance its permittivity towards solvents has been developed. Historically, Degassed Assisted Patterning was used in microfluidics as a pump-free manipulation. First reported by Hosokawa and coworkers in 2004[37], it allowed to pre-store the pumping energy in the degassed PDMS in order to obtain a liquid flow in the PDMS channels during the experiment without the use of external pumping. Indeed, the solubility of a gas in PDMS follows Henry’s law: the concentration of gas dissolved is proportional to the partial pressure of the gas around the mold.
When a PDMS is degassed and brought back to the atmosphere, it equilibrates by absorbing air[38]. Therefore, by using degassed PDMS molds, a better filling of the cavities can be achieved. The patterning process can thus be broken down into four steps (Figure 1-3): (i) the mold is first pumped under vacuum (P < 10 mbars) for a few minutes; (ii) straight after breaking the vacuum, the mold is deposited on as-prepared xerogels; (iii) Air trapped in the mold cavities is easily pumped through the mold, promoting the filling of the cavities; (iv) Solvents, generally trapped within the material underneath the mold, are pumped out improving the filling of the PDMS pattern cavities. To complete the imprinting step, consolidation of the xerogel is achieved through UV or thermal curing before demolding. Final thermal treatment is often required to achieve the pure metal oxide material.
Nano Imprint Lithography (NIL) Methods
UV-NIL
UV-NIL method is one of the most developed and used in recent years. From its outset till today, it has been adapted to a plethora of different materials, including organic, hybrid organic/inorganic[34,39–41], and inorganic metals and metal oxide materials from sol-gel solutions and/or nanoparticles (NPs) dispersions[42–46]. A prerequisite is that the imprinted materials (precursors) need to absorb UV radiation. UV-NIL processes involve several steps, starting first with liquid deposition (e.g. spin, dip, ink, spray, etc.) of the “resin” solution, followed by imprinting with a UV resistant and transparent mold stamp, composed of typically Polydimethylsiloxane (PDMS), poly-urethane-acrylate (PAU)[34], Quartz, or Perfluoropolyether (PFPE), so that curing through UV exposure takes place directly underneath the mold. Once the material is cross-linked enough, the mold is delicately peeled off to prevent damaging the replica. UV-NIL enables working at room temperature with low or no external pressure. UV-NIL has been used to replicate the patterns hybrid resins (TiO2-SiO2-PMMA) with relatively good aspect ratio patterns (AR = 1.2)[39]. Besides, using photosensitive titanium demonstrated the no-need photo-initiator in the solution, since the cross-section of Ti centers was sufficient to trigger the cross liking of the network. Patterns were thus obtained faithfully from PDMS and PUA mold[34]. Furthermore, using organic resins for a large proportion of the material composed of ZnO induced the role of UV-Binder for the UV-NIL method[40,41].
UV-NIL often does not require any temperature cycling and leads to relatively high throughput on substrates compatible with UV curing. In addition, UV-NIL has been demonstrated to be a high-volume and cost-effective patterning technique, able to achieve sub-10 nm resolution.
Step-and-Flash Imprint Lithography (S-FIL)[42],[43] is another way to replicate patterns from the mold by UV-NIL to prepare organic polymer patterns. Contrary to common UV-NIL, S-FIL relies on photopolymerization chemistry of low viscosity resins, enabling to fill the mold cavities by the capillary forces with low mechanical pressure at room temperature. This S-FIL UV-NIL method was exploited by Ganesam and coworkers in 2012[42] to elaborate TiO2 nanostructures. The imprinted patterns were heat-treated finally at 450 °C for 1 h in the air to remove the organics and convert the inorganic parts into Anatase. The decomposition and the thermal treatments induce high shrinkage due to the high proportion of organic stabilizers and fluidizers required in this process.
Therefore, the impact of chemical formulation on the shrinkage and crack formation during densification of the replica through UV-NIL was determined in 2014[47]. Three formulations (spherical, branched and pure binder) were studied by mixing the two sol-gel sources of SiO2 in different proportions. After curing the pattern by UV irradiation for 60 seconds under the PDMS mold, thermal annealing was applied between 200 and 500 °C for 1 h. When the annealing temperature increases from 200 to 500 °C, features lose 64% in volume, which can be considerably reduced when using high concentrations of dense silica sources. Besides, less shrinkage is accompanied by fewer cracks, suggesting that the presence of NPs allowed for better stress relaxation during densification.
Whatever the imprinting method, one has to find a compromise between allowing a faithful infiltration of the mold cavities with a fluid enough resin and with minimal deformation, while preventing as much as possible the shrinkage upon consolidation of the hard-inorganic final patterns. Organic fluidizers and binders are generally used to facilitate the replication, a minimal proportion of them is required to reduce as much as possible the related shrinkage. Lower shrinkage is obtained when solid inorganic nano-building blocks are used since they require fewer binders and fluidizers for the process as a result of their lower surface interacting with the medium. The issue of using these NPs lays in the need for longer treatment for consolidation, the creation of residual inter-particle porosity, and the formation of a rough surface associated with the size of the particles[47].
Thermal NIL
The thermal NIL (also known as hot embossing) process is the earliest type of NIL introduced by S.Y. Chou[20]. T-NIL is one of the most attractive NIL techniques due to its simplicity and capability to replicate nanostructures into various types of materials onto various thermally resistant substrates. Mold stamps can be PDMS, Nikel, Quartz, PFPE, PUA, etc. In typical T-NIL processes, a stamp mold is brought in contact with the resin, the whole system is heated up either to soften a thermoplastic resin (that will regain its rigidity after cooling) or to crosslink a thermally curable fluidic resin. T-NIL is categorized into three different methods: Pressure thermal NIL, Direct thermal NIL, and environmental control thermal NIL. They differ by the underlying phenomena used to fill up the mold cavities.
In Pressure T-NIL, a slight mechanical pressure helps the viscous resin to infiltrate the mold. Inorganic metal oxides, hybrid organic/inorganic resins, and dispersed nanoparticles[50] replicated patterns were approached by the pressure T-NIL. Transfer patterning of inorganic materials was reported by Kang and coworkers in 2008[51]. They deposited by electron beam evaporation a 40 nm thick layer of metals on a patterned PDMS pressed (< 40 kPa) and thermally cross-linked at 80 °C to deposit electrode wires. More recently, TiO2 Mie resonators on Si and glass substrates from a colloidal dispersion of anatase and rutile TiO2 NPs[52] are obtained with slight pressure and crosslinked at 100 °C[53]. Moreover, the T-NIL process has been studied lately, with different metal oxides (Al2O3, Ga2O3, In2O3, Y2O3, B2O3, TiO2, SnO2, ZrO2, GeO2, HfO2, Nb2O5, Ta2O5, V2O5, and WO2) from hybrid formulations[43][54]. In this method, PDMS mold was brought in contact with film metal oxide under 30 bar and at 110 °C to 130 °C for 15 min. After demolding, patterns were annealed between 300 and 475 °C for 1-2 h. As expected, more or less high shrinkage accompanies the final annealing for each metal oxide. The reason for this variation of the shrinkage is likely attributed to the difference in organic content used for each metal oxide. Latterly, replication of materials such as ZnO[54] on large surfaces was demonstrated the relatively high throughput and faithfully results, using sol-gel via pressure T-NIL. Additionally, replication of the 3D patterns can be conceived by two steps of hot embossing. Hence, in order to obtain higher aspect ratio patterns, two steps rotational method by pressure T-NIL is demonstrated[33],[55]. In this method, the PDMS stamp was brought into contact with film and both were heated at 110 °C for 240 s under < 0.2 MPa pressure then PDMS was demolded and rotated at 90° for the second imprinting to take place in the same conditions. Finally, the sample was annealed at 550 °C for 15 min to yield amorphous SiO2. The squares patterns were gained at a high aspect ratio around
5[33],[55].
Another method in T-NIL is direct T-NIL, which was developed after pressure T-NIL. The development of this technique in comparison with Pressure T-NIL is just the question of the stamp-filled cavity. In pressure T-NIL, the presence of slight pressure led the mold cavities to be filled easily and get into contact with the thin film. Whereas, in direct T-NIL, degassed molds are used. The pumping helps mold cavities to be filled (as mentioned in the Generalities, Figure 1-3) which technically replaces the extra pressure. Hence, developing the direct T-NIL was carried on by demonstrating the possibility to imprint a sol-gel solution such as zirconate titanate[56], TiO2 sol-gel, and TiO2 nanoparticles[36]. The method was improved by optimization the time and temperature during NIL. Therefore, TiO2/NPs replicated patterns with AR = 1 obtained at 120 °C for 15s demonstrated as a convenient and time-consuming sol-gel mingled with direct T-NIL[36].
The third method of thermal NIL, environmental control T-NIL, is based on adjusting the vapor pressure of the solvent (mainly water or ethanol) in the chamber to stabilize the resin content in a solvent at a given value defined by adsorption/desorption equilibrium, altering then, the viscosity for infiltration. This latter method can be used in combination with pumping. An alternative approach to pressing the mold on the resin soft layer to trigger cavity infiltration is to maintain the resin as liquid as possible with a volatile compound (the solvent phase) and pump it through the mold. Being the PDMS molds slightly microporous, a pre-degassing of the latter before the application is sufficient to allow this pumping (as mentioned in the Generalities part). Using a volatile solvent compound as a fluidizer requires preventing total evaporation from the resin by setting a proper environmental control in the NIL chamber before applying the degassed mold. Dalstein and coworkers[57] used such degassed PDMS molds to imprint TiO2 from a simple sol-gel. After deposition of the sol-gel layer on Si or glass substrates, the degassed mold was gently applied on it and maintained at 110 °C for 5 min before demolding. Annealing at 450 °C for 5 min ensures the complete densification and crystallization of TiO2 into anatase[57]. The 100 nm patterns had an aspect ratio of ~ 1[57], however, deposition of a too thick initial layer led to the presence of a 200 nm-thick residual layer. Later on, employing this method patterned surfaces for controlled wetting[58] and 250 nm TiO2 dielectric Mie resonator pillar arrays (aspect ratio ~ 1)[59] were elaborated with a residual layer as thin as ~ 10 nm by applying a thinner initial sol-gel layer (see Figure 1-5).
Figure 1-5. SEM image of a) Master, b) Replicated TiO2, c) Cross-section nanopatterns TiO2 indication the thin residual layer[58].
The imprinting was performed in a clean room, where temperature and humidity were kept constant. Imprinting outside the clean room was accompanied by a loss of reproducibility attributed to changes in the atmospheric humidity. A full study of the influence of humidity on replication quality was then performed with various metal oxides from metal chlorine, water-ethanol solutions by Bottein and coworkers[60] for the first time. The faithfulness of the replication of sol-gel systems (TiO2, mesoporous SiO2, Methyl-SiO2, Al2O3, and Y-ZrO2) can be optimized by simply adjusting relative humidity within a specific span during the NIL process that depends on the external temperature. Straight after deposition by dip-coating, the sol-gel layers were imprinted with a degassed PDMS mold at 22 °C in different relative humidity ranging from 20 to 85% depending on the film mobility (e.g. SiO2 for 70% and TiO2 for 50% at 22 °C)[60]. Then, the sample is consolidated for 5 min at 120 °C and final annealing at 450 °C for 10 min after demolding. Figure 1-5 shows that for the TiO2 system, a low humidity led to no, or low aspect-ratio, replicas, while at high humidity a dramatic loss of homogeneity is obtained. The best replication (homogeneous and high aspect ratio) was achieved with ~ 50% RH at 22 °C for TiO2. Other systems did not show the same dependency and a systematic investigation of the quantity of water up-taken into the as-prepared sol-gel initial layers with respect to RH was undertaken by environmental ellipsometry (see Figure 1-6). The deduction was that even if the optimal RH is different for each metal oxide system, the optimal water content of 30% volume in the films is common to all the cases, which is related to the difference in hygroscopic property of the as-prepared films. Furthermore, in absence of a stabilizer, the sol-gel tends to spontaneously cross-link with time, which needs to be taken into account[60].
Figure 1-6. (a) Water uptake of various as-prepared xerogels concerning relative humidity measured by in situ time-resolved Ellipsometry. (b) Optical microscopy dark-field images of TiO2 replicas imprinted at 45%, 50%, 70%, and 80% RH, replicated[61].
Relative humidity during NIL determines the optimal replication and ideal morphology of the nanopatterns, regarding xerogel viscosity. Figure 1-7 illustrates the ideal replication region for TiO2 nanopatterns[60].
Environmental control T-NIL was then exploited by Naffouti and coworkers[61] and Checcucci and coworkers[62] in 2019 to prepare high aspect ratio pillar arrays as optical metasurfaces. They implemented the fabrication over a large area of arrays of titania-based Mie resonators with high throughput and at low-temperature providing structural colors (e.g. for displays), dynamic and reversible color change, and refractive index sensing[62].
Table of contents :
CHAPTER 1: NANO-IMPRINT LITHOGRAPHY PROCESSING OF INORGANIC-BASED MATERIALS
1. Generalities
1.1. Nanofabrication techniques and Nano Imprint Lithography (NIL)
2. Nano Imprint Lithography (NIL) Methods
2.1. UV-NIL
2.2. Thermal NIL
2.3. Direct printing
3. Applications
3.1. Photonic Crystals
3.2. Solar cells
3.4. Sensors
3.5. Anti-reflection coatings
5. References
CHAPTER 2: METHYLATED SILICA SURFACES HAVING TAPERED NIPPLE-DIMPLE NANO PILLAR MORPHOLOGIES AS ROBUST, BROAD-ANGLE, AND BROADBAND ANTI-REFLECTION COATINGS
1. Introduction
2. Experimental part
2.1. Thin-film Methylated silica (Si4O7Me2) preparation
2.2. Preparation of the nanopatterned ARC morphologies by environmental control T-NIL
3. Optical performance
3.1. Abrasion resistance
3.2. Thermal resistance
3.3. Chemical resistance
4. Conclusion
5. References
CHAPTER 3: SOL-GEL TIO2 NANOIMPRINTED METASURFACE COMBINED TO HYBRID SILICA SENSITIVE LAYERS WITH HIGH REFRACTIVE INDEX SENSITIVITY AND SELECTIVITY FOR GAS SENSING
1. Introduction
2. Experimental section
2.1. Preparation of hybrid-silica (H-SiO2) sensitive plain coatings
2.2. Investigation of gas adsorption in the plain sensitive layer
2.3. Preparation of H-SiO2/TiO2 nanostructured composite system by environmental control T-NIL 3. Specular reflection of nanostructured sensitive layer composite systems
CHAPTER 4: 2D AND 3D (NANOSTRUCTURED PATTERNS) MULTI-STACK TIO2 COMBINED WITH LOW REFRACTIVE INDEX MESOPOROUS METHYLATED SILICA THIN FILM BY SOL GEL CHEMISTRY AND NANO IMPRINT LITHOGRAPHY
1. Introduction
2. Experimental section
2.1. Investigation of multi-stack thin films of mesoporous methylated silica and dense titania
2.2. Fabrication of 3D double-stack super-imposed nanopatterns by replication of TiO2 nanostructures combined with meso-MS thin layer by pressure and environmental control T-NIL
3. Optical investigation
