NRAMM Discussion Forum

Seems like a similar issue was observed for the Polara in James Conway’s lab. See last slide.

Just to clarify, the last slide was from data collected on a Krios/Falcon2 that is not in my lab (unfortunately). I have not yet run the same analysis on the Polara that is in my lab owing to it only recently being equipped for successful automated data collection.

The only times I have seen beam fluctuation is when the beam is splitting and reannealing. I've attached 3 images: 1st from a dataset that was collected on CCD and went to 4.5A, 2nd from a dataset on a DE20 that went to 3.9, and 3rd from a dataset that we're currently processing but has clear evidence of intensity fluctuations due to the tip splitting. We have had our current tip for almost 3 years and it has been splitting a lot lately. When we see evidence of splitting, we go through a reannealing procedure where we heat up the tip and increase the extraction voltage for a few days. That almost always fixes it, but it is highly variable how long it will stay fixed. We had a split a year ago that we reannealed and it lasted for almost a year. Lately, it's been splitting a lot and we're going to get a new tip.

And a further update... 3 August 2014.

FEI dropped the temperature of the FEG by 50 degrees. Looked stable in test. Then after running stable for about 4-5 days the beam on the Krios suddenly dropped by a factor of 5x. It stayed this away all night and then just as inexplicably recovered back to the original value. We know this is not a cameras related problem as both the K2 and the Ultrscan reported exactly the same drop and recovery. Since the recovery it has been stable for 2 days except for a small dip as shown below:

Here is our procedure used to document the collapse:

We did the monitoring by checking the mean intensity of the final exposure in Leginon run.
When it did collapse, the intensity change was quite characteristics. This one is the best example.
Ignore the first segment where we might have reduced the exposure time before the second segment,
and thus the pixel mean.

Usually the intensity mean decreases over time, with the slope increases as it goes, until it is half or
less. It then increases quickly, back to either the original value or some other value.

When it is fully collapsed, the apparent beam profile would change. We have a standard procedure now.
We set spot size to 3, take the magnification up to 195 kx, remove the objective aperture, and
use a large C2. On Krios, we use 150 um C2. and turn off C3 lens. When we condense the beam to cross-over,
we either see two lobes as if the FEG tip is split (see krios_tip_split.png), or two ellipse cross each other (see talos_crossover.png). We have
reproduced this kind of abnormal profile observation at the bottom of the ring collapse four times now and on two microscopes (Krios at 300 kV, Talos at 200 kV).