Simulations of nonuniform embossing: The effect of asymmetric neighbor cavities on polymer flow during nanoimprint lithography

Abstract

This paper presents continuum simulations of viscous polymer flow during nanoimprint lithography (NIL) for embossing tools having irregular spacings and sizes. Simulations varied non-uniform embossing tool geometry to distinguish geometric quantities governing cavity filling order, polymer peak deformation, and global mold filling times. A characteristic NIL velocity predicts cavity filling order. In general, small cavities fill more quickly than large cavities, while cavity spacing modulates polymer deformation mode. Individual cavity size, not total filling volume, dominates replication time, with large differences in individual cavity size resulting in non-uniform, squeeze flow filling. High density features can be modeled as a solid indenter in squeeze flow to accurately predict polymer flow and allow for optimization of wafer-scale replication. The present simulations make it possible to design imprint templates capable of distributing pressure evenly across the mold surface and facilitating symmetric polymer flow over large areas to prevent mold deformation and non-uniform residual layer thickness.

Design of Patterned Surfaces with Selective Wetting Using Nanoimprint Lithography

“…For an array of square pillars r in Equation (1) and w in Equation (2) are given by the following expressions: r ¼ 1 þ 4w(ha À1 ) and w ¼ (ba À1 þ 1) À2 , where a is the pillar width, b the pillar spacing, and h is the pillar height. [22] Our experiments confirm that OTS coated nanopillars prefer the Cassie condition ( Figure 4e) and excel as a superhydrophobic surface with 1638 water contact angle, that is in good agreement with predicted range 157-1628 for a Cassie state assuming pillar width of 100 nm, height of 400 nm and spacing range of 200-300 nm ( Figure 5). According to our experience and the prediction of the Cassie-Wenzel model, to achieve the greatest water contact angle of nanostructured surface with given surface energy, the height/width ratio (ha À1 ) of nanostructure should be larger than 3, and the spacing/width (ba À1 ) ratio should range from 2 to 4.…”

“…[21] This contradiction may be caused by the difference of nanopillar height, perhaps, in case of 1 the imprinted nanopillars can be a bit longer due to extremely low viscosity (Table 1). [22] Wetting Properties react with the -OH functionalities forming an organized SAM. The nanopillars coated with OTS are shown in Figure 3d.…”

Nanoimprint Lithography: Methods and Material Requirements

“…Several recent studies by Rowland et al have investigated the impact of polymer material properties, mold geometry, and process conditions on polymer deformation, and further studied the impact of the polymer deformation mode on the replication time. [90][91][92] They also found that when polymer flows vertically into an open cavity during imprinting, the polymer can deform as a single peak centered in the cavity, or as a dual peak, where each peak remains close to the vertical sidewalls, depending upon the geometry. The ratio of cavity width to film thickness modulates single versus dual peak cavity filling, regardless of the absolute size of the features and the pressure or temperature applied during imprinting.…”

“…In the case of the rightmost geometry, the polymer is squeezed beneath the mold, and this squeezed flow significantly limits the time required to fill this cavity. [92] The ability to displace a viscous polymer and fill cavities can be demonstrated in another example, in which NIL is used to form an optical waveguide directly from a very thin polymer layer. The device is a polymer-based optical microring resonator, [68] which can find applications in optical communication, label-free biosensing, [93] and high-frequency ultrasound detection.…”

Simulation and Mitigation of Pattern and Process Dependencies in Nanoimprint Lithography

“…Feature-scale simulation using continuum material models There have been several attempts to describe TNIL through the finite-element modeling of very simple geometries [18][19][20][21][22]. The embossed material layer has variously been described as a Newtonian liquid [18][19][20], [23], as a shear-thinning liquid [18], and with linear viscoelastic [21], [24] models. Surface tension effects have generally been regarded as negligible in the simulation of TNIL, where the polymer's viscosity during imprinting is usually at least 10 3 Pa.s and bulk forces are considered to dominate over surface forces.…”