Why the mask world is moving to curvilinear

If you’ve been to a lithography or photomask technology conference lately, you’ve likely noticed a trend: many papers and talks on curvilinear masks, curvilinear OPC, curvilinear ILT, curvilinear mask process correction (MPC), and curvilinear mask formats. The photomask industry is experiencing a fundamental shift from Manhattan masks to curvilinear masks.

Part 1.

By Leo Pang, D2S, Inc.

So, what is the motivation for this shift? After decades of Manhattan mask design, what are the benefits of moving to curvilinear? First, let’s look at some background on the technologies that set the stage for this shift to curvilinear masks.

ILT started the shift
Curvilinear masks started with inverse lithography technology (ILT). ILT treats mask optimization as an inverse problem, starting with the desired wafer target and calculating the mask that will produce the desired wafer target. Wafer scanners have band-limited optics which cannot produce 90-degree corners, so any 90-degree corners (especially the tip of the corner, which has infinite frequency) will be filtered by scanner optics, and so are not included in ILT solutions. For this reason, ILT mask patterns were naturally curvilinear.

Figure 1. Mask pattern created by this first ILT solution (left). “Manhattanized” version of the same pattern (center), Manhattanized ILT mask patterns without SRAFs. (right)[1]. Source: Luminescent/Synopsys.

At the 2006 SPIE Advanced Lithography Conference, Dan Abrams and I presented the first ILT paper from Luminescent, “Fast Inverse Lithography Technology.”[1] Luminescent and its partners published numerous papers that showed that ideal, fully curvilinear ILT mask patterns produced the largest process window[2-9].