An experimental study of the dc breakdown voltage for planar MEMS interdigitated aluminum electrodes with gaps ranging from 10 to 500 µm is presented. Unlike most research on the breakdown in MEMS electrodes that was performed at atmospheric pressure, this work focuses on the effect of gas pressure and gas type on breakdown voltage, because this is central for chip-scale plasma generation and for reliable operation in aerospace applications. The breakdown voltage is measured in helium, argon, and nitrogen atmospheres for pressures between 102 to 8.104 Pa (1 to 800 mbar). For higher values of the pressure P, or of the gap d (i.e., for high values of the Paschen reduced variable Pred=P·d), classical Paschen scaling is observed. For lower values of Pred, however, significant deviations are seen: the Vbd versus. Pd curve shows an extended flat region rather than a narrow dip. These differences cannot be attributed to field emission, but are due to the many length scales effectively present in a planar geometry (on-chip and even off-chip) that leads to the superposition of several Paschen curves. Guidelines are formulated for low-pressure operation of MEMS to avoid or encourage breakdown.
We present an experimental study of the DC breakdown voltage of MEMS interdigitated aluminum electrodes with gaps
ranging from 10 to 500 μm. Unlike most research on MEMS electrodes, that was done at atmospheric pressure, our work
has focused on the effect of gas pressure and gas type on the breakdown voltage. A main goal was to identify geometries
that favor the creation of low-voltage discharges. Helium, argon and nitrogen pressure was varied from 10<sup>2</sup> to 8.10<sup>4</sup> Pa (1
to 800 mbar). The breakdown voltage was plotted as a function of the Paschen reduced variable P<sub>red</sub> = p·g. For higher
values of pressure, p or gap, d (high P<sub>red</sub>), classical Paschen scaling was observed. For lower values of P<sub>red</sub> however,
significant deviations were seen, particularly at low pressures. We attribute these differences not to field emission, but to
the scale of the mean free path (which explains the higher than predicted voltages), and to the many length scales
effectively present in our planar geometry (on-chip and even off-chip, that lead to the superposition of several Paschen
curves). Guidelines are proposed for low-pressure operation of MEMS to avoid or to encourage breakdown.