November_EDFA_Digital

edfas.org 33 ELECTRONIC DEV ICE FA I LURE ANALYSIS | VOLUME 23 NO . 4 well (to push the ion into the trap center radially). Floating (unpinned) potentials can cause local fluctuations of the electric field thatmove the ion unfavorably, in some cases changing the ion characteristics such that quantumgates cannot be performed. Because every electrode on the ion trap functions as an antenna to define and shape the trapping fields, validat- ing their connection to the specific net to which they are assigned is critical. Validation occurs by testing each die after packaging. This requires measuring the resistance and capacitance of each net, comparing it to expected results, and thus identifying shorts. Unlikemany electronic devices, there is no redundancy in ion trap devices so an electrode connection failure is likely catastrophic for the device (in some cases a control electrode shorted to ground or another control electrode may still allow use of the device albeit with reduced functionality). A major cause of shorts is particles, and for this reason the devices are packaged in a class 100 cleanroomand then tested for shorts in the RF, control, and ground electrodes in one- to-one, dual-polarity fashion using an automated tester. For opens testing, the full circuit cannot be interro- gated in a standard end-to-end mode as probe contact damage to an electrode surface can change the electric fields defining the trap. Capacitance testing doesn’t work because the electrodes only contribute a small amount of parasitic capacitance relative to the total (nF-scale RF shunt plus parasitic) capacitance, so breaks near the electrode are basically undetectable. Consequently, an alternate non-contact method was employed, passive voltage contrast (PVC) using scanning electronmicroscopy (SEM). By grounding all pins of the packaged trap device, open electrodes are revealed by charge induced image contrast in the SEM. Floating electrodes image as bright when charged up by the incident primary electron beam, revealing a break in that circuit (see Fig. 4). Additional SEM analysis using energy-dispersive x-ray spectroscopy (EDS/EDXA) has been found to be a powerful tool for identifying and understanding the cause of elec- trical opens and via loss. An open in a control electrode net on a trap is most frequently caused by a failure at a material interface that leads to the subsequent (undesired) removal of a tungsten (W) electrical via at the end of the fabrication sequence. Using EDS for elemental mapping allows detecting the presence and absence of W vias immediately under the Al-Cu electrode, in the top IMD Fig. 4 SEM image showing a floating electrode. The oblong black regionbetween the electrodes is ahole through the trap chip used for loading ions into the trap. The bright control electrode has accumulated more charge than its neighbors creating a passive voltage contrast signal indicating that it is floating. Fig. 5 (a) SEM image of ion trap chip; the vertical rectangular features are the control electrodes. The trapping node runs horizontally through thedark region,which is a through-chip slot for laser access. TheRF electrodes are in thepenultimate metal level below the top metal control electrodes for this unique ion trap. (b) EDS elemental distribution map using W within the gray square from(a). The Lα1 signal fromtheWvias identify extant andmissing tungstenvertical interconnects in the top IMD. Electrical vias are missing for 13 of the 40 control electrodes in this trap. (a) (b)