Shared Copy Of Preproduction NbTi Vertical Flex Images

Flex cables/Panels back from BEI, 8/20/2018

Overview

Short vertical flex cables. We suspect the 3 cables in the right bag passed electrical checks. The 6 cables on the right side of the left bag appear to be the cables identified as bad during the July 2018 BEI visit. There are 5 white slips, all face down, but on at least one you can read the word "Failed". We suspect the 5 cables on the left side failed electrical checks - it was noted during the July 2018 visit that there were some high resistance shorts that BEI was trying to address through a cleaning process. Electrical connectivity needs to be verified.

Photos of the long panels. Two panels were etched. The third panel was laminated but had no/minimal etching (there is a faint image of the circuit on the panel, which could either be photoresist or a very slight etch).

Side view of the HEMT region, showing severe buckling in this region of the panel

Closeup Photos Taken Using ppareg / B33 Wire Bonder Microscope

Photos of a delaminated trace that was deliberately bent over to allow view of underside of trace.

Another photo of delaminated traces

Photos of the transition between plated and unplated regions. Note the significant narrowing of the unplated traces. Some of the thinner unplated traces delaminated here.

More photos in regions near the plated / unplated transition.

Closeup of the narrow "chassis" trace, which had significant delamination in many cables. There appears to be no good reason for this trace being so narrow.

Miscellaneous Photos

Photos of wire bonder inspection setup

Paul's theory of what we are seeing

Closeup of unetched panel(?)

SEM Images: N1 Side 1 - 4K stage, HEMT

SEM samples taken 8/23/18 from panel labeled 'N' 8/22/2018

The regions in between the NbTi appearing bright are areas that collect charge (acrylic/adhesive?).

Getting near to this area with the ion beam, especially under high mag, causes distortions and eventually a complete image loss. Trying to minimize this before using EDX analysis.

Several higher mag images of boundary region. All but one of the traces look similar to these where contact with the Cu is lost.

Only one trace on this side of the sample made some sort of contact with the Cu :

Several perspectives on a Cu trace that has lifted up/ left behind NbTi

The material folded back on itself is Cu. NbTi underneath the circular Cu pad also seems to have been etched to a similar width as those traces underneath the other Cu traces

Attempted to peel back one of the copper traces with tweezers and could see a thinner NbTi trace adhered to underside. From email by Jon: "the trace was adhered pretty strongly to the substrate."

N1 Side 1 - 4K stage, HEMT - Dispersive X-Ray Spectroscopy Data:

Site 1

Electron image:

Individual element images:

Visible structures:

Noise/Contamination? :

High reading for Beryllium?

Site 2

Electron Image:

Individual element images:

Visible Structures:

Noise/Contamination?:

Site 3

Electron Image:

Individual element images:

Visible Structures:

Noise/Contamination?:

Site 4

Revisiting the lifted trace from N1 sample. The trace was gently bent flat to fully expose the underside:

c161 c160

Electron Image:

Visible Structures:

Cobalt seems to be present mainly in the Copper and not NbTi. Nitrogen and Silicon are more concentrated on/in NbTi.

Greater counts would probably confirm this.

Beryllium image, see detailed spectrum below more more info on this reading:

Closer Look at Spectra:

Actual spectral line for Be is _{100 eV (red line). Detector is "seeing events" with energies close to but not really that of Be (0-50 eV)}, so the software chooses to plot these as Be in above image. So this is probably just some sort of noise on detector, thermal, EM? In any event, EDX is known to be inaccurate for atoms lighter than Carbon. Edit: Reading through the data sheet for the X-ray detector (Oxford x-max^n 20) it claims that this is one of the few detectors that can resolve Beryllium accurately.

The Cobalt present in the Copper appears to be in trace amounts. The lower energy emissions are effectively lost in the tail of the Copper data. The higher energy emissions are completely lost in the noise.