An internal reference method is used for the first time to clearly demonstrate the glass transition temperature (Tg) depression effect in 5 nm thick polystyrene films spin-cast on silicon wafers. Initially flat films exhibit depressed Tg at approximately 85 °C. Temperature-induced dewetting on hexamethyldisilazane-treated silicon substrates leads to formation of discontinuous films with average effective thickness of 15-30 nm. Dewetted films demonstrate Tg close to the bulk value (≈ 100 °C) and are used as internal references. Data both for continuous and discontinuous films are obtained in the same experimental run for the same sample, which allows direct comparison between datasets. Phase-modulated ellipsometry in vacuum is used to monitor glass transition. Both traditional linear temperature scan method and a novel temperature modulated technique have been employed in the measurements.

A strategy to replicate fingerprint patterns formed by the self-assembly of lamella-forming block copolymer (BCP) was investigated. To accomplish this, liquid conformal layers were placed between the surfaces of a "master" BCP film and a transparent "replica" substrate that solidified and covalently bonded to the BCP upon exposure to light The benzophenone-containing conformal layer enabled pattern replication over areas limited only by the size of the samples and exposure field. The replication step is light activated, occurs below the glass transition of the BCP, and takes less than 1 h. This demonstration used a poly(styrene-b-methyl methacrylate) BCP with a bulk domain periodicity of 42 nm, but it is possible that the chemistry may be generalized to many other BCPs. Control experiments conducted with alternative conformal layer compositions indicate that interfacial photosensitization of the BCP by excited benzophenone, followed by propagation to residual acrylate groups present in the conformal layer, is the primary mechanism by which pattern replication takes place.

A chemical pattern consisting of end-grafted polystyrene brushes (20 nm lines on a 40 nm pitch) on the native oxide of silicon wafers was defined by molecular transfer printing from assembled block co-polymer films. End-grafted hydroxyl-terminated poly(2-vinyl pyridine) brushes were selectively deposited in the interspatial regions. The poly(2-vinyl pyridine) regions selectively sequester acidic HAuCl4 from solution and form arrays of small Au nanoparticles upon exposure to oxygen plasma within the confines of the macromolecular brush layer. This print and fill process to pattern polymer brushes is a generalizable strategy to create functional chemical surface patterns.

Chemical patterns consisting of poly(2-vinyl pyridine) (P2VP) brushes in a background of a cross-linked polystyrene (PS) mat enabled the highly selective placement of citrate-stabilized Au nanoparticles (NPs) in arrays on surfaces. The cross-linked PS mat prevented the nonspecific binding of Au NPs, and the regions functionalized with P2VP brushes allowed the immobilization of the particles. Isolated chemical patterns of feature sizes from hundreds to tens of nanometers were prepared by standard lithographic techniques. The number of 13 nm Au NPs bound per feature increased linearly with increasing area of the patterns. This behavior is similar to previous reports using 40 nm particles or larger. Arrays of single NPs were obtained by reducing the dimensions of patterned P2VP brushes to below ~20 nm. To generate dense (center-to-center distance = 80 nm) linear chemical patterns for the placement of rows of single NPs, a block-copolymer (BCP)-assisted lithographic process was used. BCPs healed defects associated with the standard lithographic patterning of small dimensions at high densities and led to highly registered, linear, single NP arrays.

In this study, directed self-assembly of block copolymers (BCP) was investigated as a function of the dimensions of the alternating stripes on chemically nano-patterned substrates, at a constant pattern pitch. It was shown that alignment of the block copolymer with the chemical pattern depends on the width of the guiding stripe. Also, when the ratio of this parameter over the natural periodicity of the BCP (W/Lo) is around 0.6, a high degree of perfection is achieved. Finally, the formulation of the BCP was found to impact the process window, as three batches with the same Lo, presented good alignment at different conditions.

Directed Self-Assembly (DSA) of block copolymers is considered to be a potential lithographic solution to achieve higher feature densities than can be obtained by current lithographic techniques. However, it is still not well-established how amenable DSA of block copolymers is to an industrial fabrication environment in terms of defectivity and processing conditions. Beyond production-related challenges, precise manipulation of the geometrical and chemical properties over the substrate is essential to achieve high pattern fidelity upon the self-assembly process. Using our chemo-epitaxy DSA approach offers control over the surface properties of the slightly preferential brush material as well as those of the guiding structures. This allows for a detailed assessment of the critical material parameters for defect reduction. The precise control of environment afforded by industrial equipment allows for the selective analysis of material and process related boundary conditions and assessment of their effect on defect generation. In this study, the previously reported implementation of our feature multiplication process was used to investigate the origin of defects in terms of the geometry of the initial pre-patterns. Additionally, programmed defects were used to investigate the ability of the BCP to heal defects in the resist patterns and will aid to assess the capture capability of the inspection tool. Finally, the set-up of the infrastructure that will allow the study the generation of defects due to the interaction of the BCP with the boundary conditions has been accomplished at imec.

Directed Self-Assembly (DSA) is gaining momentum as a means for extending optical lithography past its current limits. There are many forms of the technology, and it can be used for creating both line/space and hole patterns.1-3 As with any new technology, adoption of DSA faces several key challenges. These include creation of a new materials infrastructure, fabrication of new processing hardware, and the development of implementable integrations. Above all else, determining the lowest possible defect density remains the industry's most critical concern. Over the past year, our team, working at IMEC, has explored various integrations for making 12-14nm half-pitch line/space arrays. Both grapho- and chemo-epitaxy implementations have been investigated in order to discern which offers the best path to high volume manufacturing. This paper will discuss the manufacturing readiness of the various implementations by comparing the process margin for different DSA processing steps and defect density for the entirety of the flow. As part of this work, we will describe our method for using programmed defectivity on reticle to elucidate the mechanisms that drive self-assembly defectivity on wafer.

The synergy of self- and directed-assembly processes and lithography provides intriguing avenues to fabricate translationally ordered nanoparticle arrangements, but currently lacks the robustness necessary to deliver complex spatial organization. Here, we demonstrate that interparticle spacing and local orientation of gold nanorods (AuNR) can be tuned by controlling the Debye length of AuNR in solution and the dimensions of a chemical contrast pattern. Electrostatic and hydrophobic selectivity for AuNR to absorb to patterned regions of poly(2-vinylpyridine) (P2VP) and polystyrene brushes and mats was demonstrated for AuNR functionalized with mercaptopropane sulfonate (MS) and poly(ethylene glycol), respectively. For P2VP patterns of stripes with widths comparable to the length of the AuNR, single- and double-column arrangements of AuNR oriented parallel and perpendicular to the P2VP line were obtained for MS-AuNR. Furthermore, the spacing of the assembled AuNR was uniform along the stripe and related to the ionic strength of the AuNR dispersion. The different AuNR arrangements are consistent with predictions based on maximization of packing of AuNR within the confined strip.

The implementation of our previously reported chemo-epitaxy method for directed self-assembly (DSA) of block copolymers (BCPs) on 300-mm wafers is described in detail. Some challenges to be addressed include edge bead removal control of the layers forming the exposure stack and uniformity of the deposited films across the wafer. With the fine tuning of the process conditions, this flow provides chemically nano-patterned substrates with well-defined geometry and chemistry. After a film of BCP is annealed on the chemical patterns, high degrees of perfection are achieved. A BCP with natural periodicity of 25 nm was assembled on 100-nm pitch prepatterns, obtaining 4X feature multiplication. Top-down scanning electron microscope images show a wide process window with depth of focus >200 nm and exposure latitude >40% for lines and spaces of 12.5-nm half-pitch. We provide a platform for future study of the origin of DSA generated defects and their relationship to process conditions and materials that are amenable to use by the semiconductor industry. (C) 2012 Society of Photo-Optical Instrumentation Engineers (SPIE). [DOI: 10.1117/1.JMM.11.3.031302]

Glass transition in thin (1-200 nm thick) spin-cast polystyrene films on silicon surfaces is probed by ellipsometry in a controlled vacuum environment. A temperature-modulated modification of the method is used alongside a traditional linear temperature scan. A clear glass transition is detected in films with thicknesses as low as 1-2 nm. The glass transition temperature (T-g) shows no substantial dependence on thickness for coatings greater than 20 nm. Thinner films demonstrate moderate T-g depression achieving 18 K for thicknesses 4-7 nm. Less than 4 nm thick samples are excluded from the T-g comparison due to significant thickness nonuniformity (surface roughness). The transition in 10-20 nm thick films demonstrates excessive broadening. For some samples, the broadened transition is clearly resolved into two separate transitions. The thickness dependence of the glass transition can be well described by a simple 2-layer model. It is also shown that T-g depression in 5 nm thick films is not sensitive to a wide range of experimental factors including molecular weight characteristics of the polymer, specifications of solvent used for spin casting, substrate composition, and pretreatment of the substrate surface.

Inexpensive, large area patterning of ex-situ synthesized metallic nanoparticles (NPs) at the nanoscale may enable many technologies including plasmonics, nanowire growth, and catalysis. Here, site-specific localization of Au NPs onto nanoscale chemical patterns of polymer brushes is investigated. In this approach, patterns of hydroxyl-terminated poly(styrene) brushes are transferred from poly(styrene-block-methyl methacrylate) (PS-b-PMMA) block copolymer films onto a replica substrate via molecular transfer printing, and the remaining areas are filled with hydroxyl-terminated poly(2-vinyl pyridine) (P2VP-OH) brushes. Citrate-stabilized Au NPs (13 nm) selectively bind to P2VP-OH functionalized regions and the quality of the resulting assemblies depends on high chemical contrast in the patterned brushes. Minimization of the interpenetration of P2VP-OH chains into PS brushes during processing is the key for achieving high chemical contrast. Large area hexagonal arrays of single Au NPs with a placement accuracy of 3.4 nm were obtained on patterns (similar to 20 nm spots, similar to 40 nm pitch) derived from self-assembled cylinder-forming PS-b-PMMA films. Linear arrays of Au NPs were generated on patterns (40 nm lines, 80nm pitch) derived from lamellae-forming PS-b-PMMA that had been directed to assemble on lithographically defined masters.

Chemically patterned surfaces comprised of polymer mats and brushes of well-defined chemistry were fabricated at the length scale of 10 nm. A key concept is the integration of new materials, cross-linked polymer mats, with traditional lithographic processing. Resist was patterned on top of cross-linked polystyrene mats. After etching, regions of the remaining mat with dimensions ranging from 10 to 35 nm were separated by interspatial openings to the underlying substrate. End-grafted polymer brushes, in this case hydroxyl-terminated poly(2-vinylpyridine) or polystyrene-poly(methyl methaoylate) random copolymer, were grafted into the exposed, interspatial regions from films spin-coated over the patterned mat. Both block copolymer wetting studies, with polystyrene-block-poly(methyl methacrylate) (PS-b-PMMA), and near-edge X-ray fine structure spectroscopy showed that with sufficient cross-linking the polymer mat chemistry was unaffected by the subsequent grafting of the polymer brush. The precise definition of both the chemistry and the geometry was demonstrated two sensitive application of nanoscale chmican patters: the site-specific immobilization of Au nanoparticles and the directed assembly of overlying PS-b-PMMA films.

This study modifies the authors’ previously reported directed self-assembly (DSA) process of polystyrene-block-poly(methyl methacrylate) (PS-b-PMMA) in order to meet the throughput and material-related requirements of a semiconductor manufacturing environment. It is demonstrated that all of the bottleneck steps in the authors’ DSA process, including the deposition of the cross-linkable mat and the deposition of the brush layer, can be done in minutes on a hot plate in an N2 atmosphere, which simulates the processing environment of a lithography track module. A 25-nm-pitch pattern resulting from a 4:1 density multiplication was demonstrated with a manufacturing-compatible organic solvent. A preliminary uniformity study on 300 mm wafers was also presented. The modified DSA process presents a viable solution to some of the anticipated throughput-related challenges to DSA commercialization and thus, brings integration of DSA within reach of the semiconductor manufacturing industry.