Gatan US 4000 @ 120 kV

Submitted by mcianfrocco on Mon, 2014-01-13 10:18

Forums

FindDQE

Hello,

I tried the Find DQE program this week on a beam stop image that I collected during a Leginon session on a Tecnai F20 operating at 120 kV. I'd like to confirm that I've run the program correctly:

Is it correct to assume that since I was using a dose of 25 e-/A2, the total number of electrons/pixel is 25?

I'm worried that something is wrong with my implementation of the routine since the calculated DQE appears to be 2 X better than the published DQE in your paper for the US 4000 @ 200 kV. Do you think this difference would be possible considering the camera response at 120 kV?

I'm attaching the calculated DQE for your consideration.

Thanks!
mike

rsruskin Mon, 2014-01-13 10:35

Hi Mike,

If your dose is 25 e/A2, your total number of electrons/pixel is probably not 25, unless your pixel size is exactly 1 A/pixel. Make sure you check to see what the conversion is between Angstroms and pixels at the magnification you used. (Even though the magnification used is not important for the function of FindDQE, it is important to the image collection software, which is showing your dose in terms of electrons/A2 rather than electrons/physical camera pixel).

Another thing you might want to check is whether there is a "Divide by 2" or "16 bit Divide by 2" option in the software you used to collect the image. If this option was checked, your DQE could be off by two-fold. This option is often present in the image-saving dialog. You can also check whether this happened by looking at the final counts in your saved image and seeing if they match the number of counts you actually expect from the given dose. The third parameter in your script can be set to "2" rather than "1" if you need to reverse the division by 2 applied by an image collection program.

I think that your overly high DQE has been affected by one or both of the above issues.

HTH,

Rachel

mcianfrocco Mon, 2014-01-13 13:10

Hi Axel & Rachel,

Thank you for the help - that fixed it! My electron dose was incorrect (as you both pointed out), instead it should've been 50 electrons/pixel for 25 e-/A2 with a pixel size of 1.37 A/pix.

At Nyquist 0.5, the DQE = 0.218 for our Gatan US 4000 detector, which is in the ballpark of your observed DQE = 0.20. It's a very easy program to use, thanks for developing it.

Mike

Axel Mon, 2014-01-13 10:36

Hi Mike,

Although I am not as much of an expert on the program's use as Rachel, I would venture a guess that your dose could be wrong. You should multiply the dose by your pixel size square to get the number of electrons per pixel (e.g. if your pixel size is 1.41 Angstrom, then you dose/pixel is 25*1.41*1.41 = 50 electrons).

The paper also does contain some data for the US4000 at 120 KeV (Table 2), so you should be able to get some data to guide your judgment, and decide whether all of your parameters are correct.

Best,

Axel