Well, Brian, it is good to know that optical micro-mirrors and electron micro-mirrors are slightly different, and this tiny detail is where a big devil is sitting and laughing in his beard. In the optical pixelated mirror the reflected visual light photons, being chemically not very active, practically do not induce any observable mirror surface modification, while the electrons, which are chemically very active, do modify the surface, and the change in surface material properties, and as a result in its electrical properties, becomes pretty quickly catastrophic for the pixel performance, namely, its reflection efficiency, and as a result, litho performance deteriorates to unacceptable level.
For example, it is well known from the CD SEM experience that it takes just seconds, in the slowest case, minutes, for the SEM FOV surface to become fully coated with carbonization surface film, which continues to grow with the exposure time changing even the formal size of the CD features.
Yes, the e-beam irradiation continuously changes the state the active DPG surface, making impossible to stabilize the response and the performance of the whole DPG-based REBL machine and technology. This is just one of many un-fixable problems in this technology.
The other one intrinsic conceptual killer is associated with un-avoidable e-impact induced device damage, which is expected to be especially significant for highly energetic DW EBL concept as the 70-100 KeV KLA-REBL is.
Also, Mark, I am afraid, there is a logical disconnect/gap in statements like "KLA-Tencor.... promises to lower the cost of ownership in lithography by eliminating the photomask in IC production."
Namely, it actually means, that patterning (using e-beam) a small size mask with only 1/4 of resolution power, which later will be used by in the stepper for multiple exposures, is, for some reason, much more expensive than using the same e-beam patterning directly for large numbers of production wafers.
My common sense logic refuses to absorb such an anti-fabrication logic: carving a stamp (mask) and using it to fabricate a large number of copies is a less economic manufacturing approach as compared to carving each workpiece individually.
Also, remember, each reticle is individually inspected for defects (bridging, gaps etc.)and, if necessary, repaired, which will be impossible to do with the production wafers. It is also protected by pellicle from particulate contamination. The DPG, oppositely, is a fully open system being attacked by numerous damaging factors (chemistry, particles, sputtering by highly energetic particle etc.)continuously deteriorating its patterning response.