Chances for a NanoImprint Lithography based fabrication technology

An important question in nanoimprint lithography is the matter of flow and dynamics of polymers. During imprint varying amounts of polymer have to be transported away from the intruding parts of the stamp as a function of stamp feature density and complexity. The matter is complicated since the system is not in thermal equilibrium and additionally also suffering from an applied, locally, very high pressure. The flow of polymer under these extreme situations in not previously examined. The knowledge gained may be of large importance for a successful production technology based on NIL. Albeit, a lot of work already performed a full description of resist flow and dynamics during imprint situation is not yet available.

We have developed a complete process technology allowing simultaneous printing of large and nanodefined structures combined with suitability for a lift-off process. The process technology has been implemented for fabrication of interdigitated nanobiosensors. A low cost method for fabrication of such nanosensors is a pre-requisite for a full-scale exploitation of such devices that so far has hindered its industrial employment. This situation is now changed. Description and background to the result It has been shown by Montelius et al. that IDA electrodes can be manufactured by nanoimprint lithography followed by UV-lithography for interconnections and bond pads [L. Montelius, B.Heidari, M. Graczyk. I. Maximov, E.-L. Sarwe, T.G.I. Ling: Nanoimprint and UV-lithography: mix & match process for fabrication of interdigitated nanobiosensors, Microelectronic Engineering, 53(1-4), (2000), pp.521-524]. To reduce production costs further it would be advantageous to print IDA, interconnections and bond pads in one step so that no subsequent lithographic methods are needed. In that case micrometer-sized bond-pads and counter electrodes as well as nanometer structured IDA¿s have to be print simultaneously in close proximity. It has been shown by Heyderman [L.J. Heyderman, H. Schift, C. David, J. Gobrecht, T. Schweizer: Flow behaviour of thin polymer films used for hot embossing lithography, Microelectronic Engineering, 54, (2000), pp. 229¿245.] and Scheer [H.-C. Scheer, H. Schulz, T. Hoffmann and C.M. Sotomayor Torres: Problems of the nanoimprinting technique for nanometer scale pattern definition, Journal of Vacuum Science & Technology B, 16(6), (1998), pp.3917-3921.] that micrometer structures are difficult to print due to resist flow problems. The main difficulty here is not the imprint itself it is the insufficient displacement of resist below large structures resulting in thicker residual polymer layers. This makes a subsequent pattern transfer of both micro- and nanostructures impossible. We circumvent this problem by designing micrometer structures in a mesh-like pattern with cavities large enough to capture the displaced imprint polymer. We present here a complete process from stamp production over imprint to metal lift-off.

Nanoimprint lithography (NIL) is a new technique having a large potential in many different areas of research as well as industry. The basics involve pressing a predefined stamp into a heated polymer layer situated on top a substrate material. The printing process may take only a few seconds. The printing method allows defining structures in the polymer layer with a sub 10 nm resolutions over a large area (presently up to 6 inch wafers have been processed within CHANIL). The technique fills in a needed process technology gap, which is a prerequisite for a continued and sustained growth of the modern society. We have studied all aspects of NIL from how to fabricate stamps to print and process the printed features. With regards to stamps we have studied fabrication of stamps at large and small scales, at a low or a high cost, how to make copies of stamps, how to coat stamps with molecular anti-stick layers etc. With regards to printing we have studied wafer scale printing as well as step&stamp printing, we have studied in detail the flow of polymer during printing, the antistick properties of the polymers, development of polymers especially suitable for NIL, developments of methods for assessing quality in NIL, developments of lift-off processes etc. We have also studied how imprint may damage substrate material properties. We have also investigated the performance of various applications fabricated with NIL. Also important etching procedures for transferring an imprinted NIL defined mask into an underlying substrate has been studied as well as methods to fabricate imprint stamps. Collaborations with companies outside the consortium exist. Obducat AB in Malmoe, Sweden has acquired several patents with regards to NIL technology. They produce NIL machines for research as well as for industrial needs. Another partner is Karl-Suss, which addresses the step&stamp printing procedure as a basis fro future NIL equipment. A new enterprise is looking into the possibility to implement the quality assurance system being developed within Chanil.

Anti-stick treatment of stamps for nanoimprint lithography made of silicon. This method relies on dry chemistry and it gives a silanised monolayer on top of a nanometer-structured stamp. The monolayer is covalently linked and thus strong and sturdy. It allows multiple printing using the stamp many times. It overcomes the limitations encountered using wet chemistry when the stamp has a nanoscale pattern on its surface. The result has been protected by patent and disseminated in several publications as well as at international and national conferences. The patent has been acquired by Obducat AB. The result is presently employed by many research groups as well as by Obducat AB, Malmoe, Sweden as a key technology of theirs. The expected benefits are large being a key component in nanoimprint lithography and it may be of profund future value for making of micro- and nanolectronic products. It may also be of large value for other applications requiring a non-sticky surface.

The process window for obtaining optimised imprinting results tends to be different for small and large features. Also, the time of one step&stamp cycle can be minimised by minimising the amount of displaced polymer during imprinting. A solution to these issues is to combine NIL with standard UV-lithography. Here, the technique of a repeated stamp&step procedure for NIL can be advantageous as compared to wafer scale NIL. We have modified a commercial flip-chip bonder into a step&stamp NIL tool having an alignment accuracy of 1,5 micrometer. The step&stamp process has been optimised and provides a possibility to function as a complete lithography method being able to align nanometer NIL made features with large previously UV-defined structures. The results obtained are now in collaboration with a commercial company exploited for possible use in a new NIL-machine. The expected benefits for step&stamp versus wafer scale NIL is the simple way to print only the nano-required parts with an integrated alignment possibility. Thus, it may serve primarliy the research community (provided that wafer scale NIL will be adequately solved for industrial fabrication).

Since the introduction of nanoimprint lithography (NIL) in 1995 [S.Y. Chou, P.R. Krauss, P.J. Renstrom, Appl. Phys. Lett. 67 (1995) 3114. ]a rapid development has followed. In addition to the original hot embossing technique, UV-based nanoimprinting [M. Bender, M. Otto, B. Hadam, B. Vratzov, B. Spangenberg, H. Kurz, Microelectron. Eng. 53 (2000) 233.] and step and repeat [T.C. Bailey et al., Microelectron Eng. 61¿62 (2002) 461] methods have been established. The interest in these techniques has been largely due to the fact that NIL combine the merits of electron beam lithography (EBL) with UV-lithography, i.e. the ability to make nanostructures with parallel processing. One key element to establish NIL as a viable research technology and production tool is the development of a stable and well-controlled method for lift-off. For all lift-off processes it is essential to create an undercut profile. In NIL this is challenging and several approaches has been published [A. Lebib et al., Microelectron Eng. 61¿62 (2002) 371 and S.Y. Chou, P.R. Krauss, W. Zhang, L. Guo, L. Zhuang, J. Vac. Sci. Technol. B15 (1997) 2897]. In this paper we report a bi-layer method utilizing differences in dissolution rates to create an evaporation-mask of the upper resist layer. The use of a liquid solvent minimizes the ashing time, thereby reducing the line widening effect. Furthermore, a two-step lift-off process makes the use of ultra sonic agitation redundant. The most important parameter in the imprint step is pattern transfer. In our case the stamp is in contact with PMMA, which works well with our anti-sticking protection. Furthermore, this resist has been proven, indirectly, to be able to replicate patterns down to and below 10 nm. When PMMA is heated above its glass transition temperature (T 5105 8C) it shows rubber like elastic dynamics. Depending on the size of the structures to be imprinted one need to adjust the temperature and time to optimize the flow of the resist, i.e. higher temperature and longer time for larger structures. After imprinting a residual layer of PMMA is present, covering the LOL and thus prohibiting the possibility of creating an undercut. It is the case even if the stamp is imprinted through the PMMA layer and into the LOL. To remove the residual resist oxygen plasma ashing is used. This will incur a widening of the imprinted features and consequently constituting a limiting factor to the minimal size of the metal structures one can deposit. The bi-layer lift-off process addresses this problem naturally, being essentially a wet process, and as such it drastically reduces the need for ashing. In order to dissolve the LOL layer it is preferable to work with a diluted solvent. The MF 319 is mixed with deionized water to give a dissolution time of around 1 min for a 50 nm thick layer. Slowing down the process will enhance control of the undercut and prevent over development, in which case the top layer may collapse or brake apart. The LOL is dissolved isotropically, hence creating a lower boundary on the minimum distance between lines. This is an inherent property of the wet etch process and a distance of at least 1.5 times the vertical resist thickness is needed between structures. By imprinting into the LOL the dissolution times can be reduced, and pattern density increased. The lift-off process is done in two steps. First a warm acetone bath is used to expand and dissolve the PMMA. This will create cracks in the metal layer and start an initial lift-off. The LOL is unaffected by the acetone and will remain to prevent metal flakes from adhering to the silicon substrate electrostatically. Following this warm Remover S-1165 will dissolve the LOL layer and remove the remaining metal. This two-step process makes the use of ultra sonic agitation redundant.

Anti-stick treatment of stamps for nanoimprint lithography made of Nickel. This method relies on dry chemistry and it gives a silanised monolayer on top of a nanometer-structured stamp of Nickel by employing an extra metal layer in a sandwich configuration allowing the silanisation procedure to take place. The monolayer is covalently linked and thus strong and sturdy. It allows multiple printing using the stamp many times. It overcomes the limitations encountered using wet chemistry when the stamp has a nanoscale pattern on its surface. It also overcomes the problem of covalently link an anti-stick layer on the reactive Nickel surface. The result has been protected by patent and disseminated in several publications as well as at international and national conferences. The patent has been acquired by Obducat AB. The result is presently employed by many research groups as well as by Obducat AB, Malmoe, Sweden as a key technology of theirs. The expected benefits are large being a key component in nano-imprint lithography and it may be of profound future value for making of micro- and nano-electronic products. It may also be of large value for other applications requiring a non-sticky surface.

Efforts have been made in recent years to develop nanoimprint lithography (NIL) as a viable technology for creating sub-100 nm patterns. In NIL nanometer-scale features of a stamp are embossed into a thin polymer layer heated above its glass transition temperature (Tg). Polymer and stamp are cooled to a temperature below Tg of the polymer, and the stamp is detached. Imprinted structures with 10 nm resolutions have been achieved this way. Most of the research for examining this new technology was performed with poly (methyl methacrylate) (PMMA). Task of micro resist technology GmbH within the EC-funded project CHANIL was to develop polymers dedicated for NIL. Imprint temperatures of 80-100 K above the Tg , i.e., 160-200 °C, are necessary to produce a sufficiently low polymer viscosity, when PMMA or other thermoplastic polymers are used. High imprint temperature can cause thermal stress and degradation in the polymer film and increases process time. It is detrimental to some substrates and potential applications and is very exacting to the imprinting equipment. In addition, plasma etch resistance, which is necessary for use of the imprinted patterns as an etch mask, is unsatisfactory for PMMA. Alternatives to PMMA were proposed to respond to these demands. Thermoplastic polymers for NIL with good thermal and high dry etch stability were developed [H. Schulz, H.-C. Scheer, T. Hoffmann, C.M. Sotomayor Torres, K. Pfeiffer, G. Bleidiessel, G. Gruetzner, Ch. Cardinaud, F. Gaboriau, M.-C. Peignon, J. Ahopelto, B. Heidari, J. Vac. Sci. Technol. B 18 (2000) 1861-1865]. The suitability of a plasma-etch resistant e-beam resist for NIL was also checked [C. Gourgon, C. Perret, G. Micouin, Microelectron. Eng. 61-62 (2002) 385-392]. High thermal stability of thermoplastic polymers like PMMA is inherently related to high Tg , and implicitly requires high imprint temperature. This led to the idea to use special precursors of cross-linked polymers for eliminating this requirement. Prepolymer solutions can be processed in the same way as thermoplastic polymer solutions. Imprinting temperatures below 100 °C can be applied, since the pre-polymers have low molecular weights and low Tg. The structures have to be stabilized by a cross-linking reaction during or after imprinting. We have developed two cross-linked polymer systems; thermally cross-linked polymers mr-I 9000 (micro resist technology GmbH) and photochemically cross-linked polymers mr-L 6000 (micro resist technology GmbH). Information on the type of the cross-linking polymerization and its impact on the behaviour of polymers during the imprint process is necessary for specifying appropriate processing conditions. When thermally cross-linked polymers are used, process time and temperature depend on the polymerization rate. Photochemically cross-linked polymers offer the possibility of short imprint time and low process temperature. Tg of the polymer governs temperature. Low Tg enables short heating and cooling cycles. The imprinted films are thermally and mechanically stable after cross-linking. That is why the patterns can be used not only as an etch mask, but also for permanent applications. For demonstrating their stability these materials have been used as stamps for nanoimprinting themselves. The scientific results were presented at the international conferences MNE 2000, 2001 and 2002, SPIE 2001 and 2003, NNT 2002 and trade shows Semicon Europa 2001, 2002 and 2003 in Munich, Productronica 2001 in Munich and are public. The polymers for NIL provided by micro resist technology GmbH are presented on the website of the company and have been an integral part of the company's marketing. Micro resist technology GmbH has gained a good reputation as a material supplier in the international community of nanoimprint lithography based on these activities and intends to continue the development in this field.

As a low-cost quality and routine control method for nm-scale patterns we will develop an optical measurement technique (UNI-DUI) based on adding a fluorescence label to the polymer(MRT). The total fluorescence intensity corresponds to the integral polymer adhesion as well as to the local polymer deposits. This diagnostic will give information on stamp condition, i.e. evolution during imprint sequences as in step&stamp, and as a rapid test of anti-sticking layers. We are just preparing the final steps for the implementation of scanning software that allows for automatic scanning of imprint samples. This is done in cooperation with a small start-up business.

Deliverables

Share this page

Download