TY - JOUR
T1 - Subpixel alignment for scanning-beam lithography using one-dimensional, phase-based mark detection
AU - Krishnamurthy, A. V.
AU - Namepalli, R. V.
AU - Hastings, J. T.
PY - 2005/11
Y1 - 2005/11
N2 - A phase-based alignment technique for scanning electron-beam lithography is proposed and evaluated. The approach uses a grid alignment mark rotated with respect to the beam deflection axes and requires only a single, one-dimensional (1D) line scan to perform subpixel registration in both the x and y directions. This 1D, phase-based method promises subpixel alignment without interpolation, faster mark signal acquisition, and more computationally efficient detection. Theoretical limits for alignment accuracy and a simple phase-estimation algorithm that approaches these limits are described. Monte Carlo simulations indicate that nanometer-level registration precision can be achieved with modest grid periods and signal to noise ratios (SNRs). Experimental measurements of registration variance as a function of SNR agree well with both theory and simulation. Registration errors of 3 nm (1-standard deviation) are observed using a 1 μm grid period, 49 nm pixel size, and the highest observed SNR of 5.9.
AB - A phase-based alignment technique for scanning electron-beam lithography is proposed and evaluated. The approach uses a grid alignment mark rotated with respect to the beam deflection axes and requires only a single, one-dimensional (1D) line scan to perform subpixel registration in both the x and y directions. This 1D, phase-based method promises subpixel alignment without interpolation, faster mark signal acquisition, and more computationally efficient detection. Theoretical limits for alignment accuracy and a simple phase-estimation algorithm that approaches these limits are described. Monte Carlo simulations indicate that nanometer-level registration precision can be achieved with modest grid periods and signal to noise ratios (SNRs). Experimental measurements of registration variance as a function of SNR agree well with both theory and simulation. Registration errors of 3 nm (1-standard deviation) are observed using a 1 μm grid period, 49 nm pixel size, and the highest observed SNR of 5.9.
UR - http://www.scopus.com/inward/record.url?scp=29044444414&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=29044444414&partnerID=8YFLogxK
U2 - 10.1116/1.2062547
DO - 10.1116/1.2062547
M3 - Article
AN - SCOPUS:29044444414
SN - 1071-1023
VL - 23
SP - 3037
EP - 3042
JO - Journal of Vacuum Science and Technology B: Microelectronics and Nanometer Structures
JF - Journal of Vacuum Science and Technology B: Microelectronics and Nanometer Structures
IS - 6
ER -