Princeton Teams Advance Nanoimprint Understanding

With two recent papers in the journal Nanotechnology , researchers at Princeton University (Princeton, N.J.) have published important advances in nanoimprint lithography (NIL), taking the technique a couple of steps further toward feasibility as a next-generation lithography (NGL) option in mainstream semiconductor manufacturing.

The researchers in both cases were led by Stephen Y. Chou, Joseph C. Elgin Professor of Engineering and the head of Princeton's NanoStructure Laboratory. Chou invented NIL in the 1990s, and has since founded Nanonex Corp. (Monmouth Junction, N.J.) to commercialize the technique. In the Jan. 17 issue of Nanotechnology, Chou and others from Princeton and Nanonex revealed possible solutions to problems with air bubbles forming in resist liquids .¹ For a Feb. 14 publication in the same journal, Chou and other Princeton researchers described a technique to monitor NIL in situ and in real time.²

In the first paper, the researchers reported on their studies of air bubbles in dispensing-based NIL, and their impact on nanoimprint yield and throughput. They used real-time video observation to discover two means of air bubble formation — feature pinning and multi-droplet encircling — and studied the air absorption and air bubble shrinking under various conditions. Through their studies, the researchers concluded that air bubbles could potentially be controlled by such techniques as increasing the imprinting pressure and using liquids with a higher air solubility. In experiments and theoretical studies, both methods significantly increased the chances that the air bubbles would dissolve before the liquid resist hardened.

However, it may not be as simple as it sounds in terms of volume semiconductor production. A key benefit of dispensing-based nanoimprint, in which liquid droplets harden quickly after being pressed into a mold, is that it does not need to be done in a vacuum. But the bubbles may be difficult to overcome without a vacuum environment: “One of our key conclusions from the study, which has significant practical importance, is that although the air in a bubble can be completely dissolved in a resist liquid as long as the bubble is smaller than a certain size, the air absorption time might be too long for the dispensing-NIL operating in atmosphere or poor vacuum to have a necessary throughput in mass manufacturing,” the authors reported.

Schematic illustration of the RIMS setup. (Source: Princeton University)

In the more recent journal article, Princeton researchers described their method for gaining a better understanding of the polymers used in NIL by being able to monitor them in situ and in real time (within nanoseconds). Unlike today's optical lithography techniques, which shine light through a mask and demagnify its lines and trenches through a series of lenses, NIL physically deforms a resist thin film by using a 3-D mold to form a negative replica. Whether this is ultimately done by embossing a thermoplastic polymer at high temperatures or using a photocurable resist that is then cured with ultraviolet (UV) exposure, the polymer deformation plays a critical role in NIL.

The real-time imprint monitoring by scattering of light (RIMS) enables the researchers to detect the degree of resist deformation and the duration of resist penetration by a mold during the imprint process. In RIMS, a surface relief diffraction grating is created on a transparent imprint mold, and the intensity of light diffracted from the grating is continuously monitored during the imprint process. Before imprint, when there is no polymer in the trenches, the diffracted light intensity is stable. During imprint, when the trenches are being filled, the light intensity drops continuously until it reaches to near zero, indicating that the trenches are fully filled with polymer.

With this technique, the researchers have been able to directly study effects that have previously been studied only through simulations or indirect measurements — including the effects of processing parameters such as temperature, pressure, type and thickness of resist, and pre-NIL baking conditions. Their data shows that resist penetration is facilitated by increasing processing temperature, pressure and resist film thickness, with a longer pre-process resist baking step acting as a means to slow everything down.

A better understanding of the polymer flow during the nanoimprint process could help in the design and optimization of NIL tools and processes. RIMS measurements show not only how long an imprint takes to complete, but also how an imprint progresses with time and how it is affected by the differences in processing parameters.