When mirrors for the laser scattering diagnostic for large fusion devices n
eed to be inside the vacuum chamber, they are subjected to irradiation by m
ultiple high-energy laser pulses and bombardment by charge exchange atoms.
Both of these assaults are known to degrade and eventually damage metal las
er mirrors given sufficient time and flux. Our aim in this article is to us
e current data on these damage mechanisms to make design selections of meta
l mirror materials for application in fusion device diagnostics. We identif
y tradeoffs between low sputtering rates and multipulse laser damage resist
ance in candidate metals. The data for multipulse laser damage are incomple
te and extend to a maximum of only 10(4) shots for a few metals. However, t
here is a clear trend of decreasing laser-damage threshold with increasing
number of shots, and damage threshold fluences can fall to 10% of the singl
e-pulse damaging laser fluence. Further experiments up to 10(6) or 10(8) la
ser shots need to be conducted on the likely mirror candidate metals for us
e in new plasma devices. We define a figure-of-merit based on current laser
damage data and employ it in our analysis. Recent data on the sputtering y
ield and reflectance degradation of metal mirrors give a different priority
ranking of candidate metals. Overall, the preferred material selection dep
ends on the number of laser shots and the number of plasma pulses that the
mirror must endure before replacement is allowed. For example, we find that
for conditions typical of the LHD (10 s plasma pulses with a 10 Hz laser P
RF), Au, Ag, and Cu are candidate materials if mirrors are replaced after 1
0(3) plasma pulses; Au and Rh are candidates if the replacement interval is
10(4) pulses; and if the replacement interval is further increased to 10(5
) plasma pulses, then Mo is the candidate material. Other materials might a
lso be candidates but the data on them are still insufficient. (C) 1999 Ame
rican Institute of Physics. [S0034-6748(99)02104-8].