SerialEM Note: Installation and Calibration¶

Author: Chen Xu
Contact: <chen.xu@umassmed.edu>
Date_Created: 2017-11-12
Last_Updated: 2022-10-06

Abstract

When I helped a few sites to install and calibrate SerialEM, I had impression that most new users felt this process was very hard. I felt the same way when I initially learned to install and calibate SerialEM by myself. I even got frustrated and had to call David for a few times. When I think back about all the troubles I had to install and calibrate SerialEM, I believe I would have an easier time if I had a brief guideline document for what steps to follow in order, and what to do in each step. The helpfile from SerialEM is very complete to provide almost all information needed, but it is perhasp a lot to read and not clear where to start for a beginner.

I wanted to list some steps here to guide you through this initial installation and calibration phase. It is like a crash list. For more detailed information, you should always find it from helpfile.

Installation¶

Here are steps to follow.

Step 1

Ask David for the initial system file. Normally, you would fill out a “questionnaire” available at the ftp server - http://bio3d.colorado.edu/ftp/SerialEM/questionnaire.txt and send it to David. David or Guenter Resch will then create a framework file on the same ftp server for you to download. This framework file is a zip file, you can download it to local like Desktop and unzip it by double clicking on the file. Beside a sub-folder “Admin” created under “C:\ProgramData\SerialEM”, the most important file in the framework is one initial system file called “SerialEMproperties.txt”. You must have this file to get started. Please refer to the SerialEM webpage for the latest information regarding this.

Note

If on Windows the Folder Option -> View -> “Hide extensions for known types” is enabled, the property filename shown will be “SerialEMproperties”. If you change it to “SerialEMproperties.txt”, it actually is “SerialEMproperties.txt.txt” which won’t be read in as a valid property file.

Step 2

Make sure your camera computer and microscope computers are on the same local network. For example, K2/3 computer can be configured to have a network interface with IP address 192.168.1.2, and FEI scope with 192.168.1.1. And they should be able to ping each other. - You might be confused by Gatan’s DM already being able to communicate with scope, as it can detect magnification change of scope. However, this DM connection to scope is usually via serial port by a direct serial cable. SerialEM uses standard TCP/IP to communicate to a remote computer and therefore requires a standard network setup in place. - If you configured the local network, you should have DM communicate to scope (via Gatan’s RemoteTEM) using TCP/IP too. It is a lot more robust. But this is not a concern for SerialEM operation here.

Step 3

Decide which computer to install SerialEM. In theory, you can install SerialEM on either computer - camera or microscope. For K2/3 camera, SerialEM should be normally installed on the K2/3 computer, as K2/3 image returning to SerialEM locally is usually faster than via network.

Step 4

Decide which type of executable to use. SerialEM builds for both 32 and 64-bit platforms. Unless you have to run it on a Windows XP, you should choose 64-bit.

Step 5

Download SerialEM software. You should start with the latest release version from ftp server at http://bio3d.colorado.edu/ftp/SerialEM/ and save it somewhere local like Desktop.

Step 6

Unzip the installation package file downloaded. You can double click on this file, it will unzip the program into C:\Program Files\SerialEM. The folder “SerialEM” will be created automatically if there isn’t one already. The new package content will be unzipped into a new sub-folder, e.g. SerialEM_3-6-13_64.

Step 7

Quit Gatan DM if it is running.

Step 8

Right click on a file called install.bat in the package folder C:\Program Files\SerialEM\SerialEM_3-6-13_64 and select ‘Run as Administrator’. This will copy some files into upper folder which is C:\Program Files\SerialEM, register DM plugin file and copy it to the Gatan plugin folder at C:\ProgramData\Gatan\Plugin.

Step 9

Manually copy a file called FEI-SEMServer.exe from C:\Program Files\SerialEM on K2/3 computer to C:\Program Files\SerialEM on scope computer. This is a bridging program to control scope by passing the scope function calls between SerialEM main program on remote K2/3 computer and the scope scripting interface. Run the program by double clicking on it(it needs to run or SerialEM cannot control scope). This is 32-bit application, runs on both 32 and 64-bit Windows platforms. So there is only one such executable to run on Windows 7, XP or 2000 Windows OS.

Step 10

On K2/3 computer, Edit SerialEMproperties.txt file in folder C:\ProgramData\SerialEM to have proper lines in general property area to define network properties.

#GatanServerIP 192.168.1.2
GatanServerIP 127.0.0.1
GatanServerPort 48890
SocketServerIP 1 192.168.1.1
SocketServerPort 1 48892

Step 11

On K2/3 computer where SerialEM is to be installed, define a system environment variable SERIALEMCCD_PORT with the value 48890 or other selected port number, as described in the section in helpfile.
If everything goes well, you should be able to start SerialEM and it should connect to “see” both scope and DM. Congratulations!

Calibration¶

Although most of calibration results will be written into another system file SerialEMcalibraions.txt when you save the calibrtion from Calibretion menu, there are a few places you need to manully edit the SerialEMproperties.txt to take in the calibration results. These include pixelsize and tilting axis angle - they are more like instrument parameters.

For pixelsize calibration, it is best to use standard 2160 line waffle grid. For all other calibration like Image Shift and Stage Shift, it would make things a lot easier to use a non-periodic sample. Please see the NOTE at the end of this document.

Tip

Since µP and nP modes give very different beams, it is required to perform calibration for both modes for some basic items:

Beam Crossover
Beam Intensity
Spot Intensity
Beam Shift
Autofocus
Electron Dose
Standard Focus

See https://bio3d.colorado.edu/SerialEM/betaHlp/html/setting_up_serialem.htm#nanoprobe

Step 0

Determine camera orientation configuration. Make sure the image orientation from camera shot agree with that of on large screen or FluCam. If it doesn’t, try to adjust the camera orientation of Gatan K2/3 camera from Camera - Configuration. You can use beamstop to help. You should add a property entry to reflect the DM configuration so SerialEM takes care of it even someone might have changed DM configuration.

DMRotationAndFlip 7

Step 1

Edit property file to define the camera configuration information about orientation determined by step 0. SerialEM will return to main display with proper orientation. This is initial starting point for all the calibrations.

RotationAndFlip 7

Step 2

SerialEM - Calibration - List Mag. Scope will go through all the mags and list them on log window, from lowest to highest. Check it with what are in SerialEMproperties.txt, update that if needed.

Step 3

Load standard waffle grating grid (TedPella Prod.# 607, http://www.tedpella.com/calibration_html/TEM_STEM_Test_Specimens.htm#_607).

Step 4

Start with lowest magnification above LM range. On Talos, it is 1250X. At close to Eucentricity, and clost to eucentric focus.

Step 5

Take a T shot with 2x binning on a K2/3 camera, make sure the counts are neither too low nor too high.

Step 6

Take a T shot, then Calibration - Pixel Size - Find Pixel Size. The log window shows both mag index and pixel size. Edit SerialEMproperties.txt to add a line like below in K2/3 camera property section.

# MagIndex  DeltaRotation (999 not measured)  SolvedRotation (999 not measured)   Pixel size (nm, 0 not measured)
RotationAndPixel 17 999 999 3.396

Here, 17 is mag index for 1250X, and 3.396 is pixel size in nm just calibrated.

Step 7

You might want to change to a grid without repeating features, please see “note” at the end of this document.
Calibration - Image & Stage Shift - IS from Scratch.

Step 8

Calibration - Image & Stage Shift - Stage Shift.

Step 9

Calibration - Administrator, turn it on.

Step 10

Calibration - Save Calibration.

Step 11

Take the tilting axis value (e.g. 86.1) from step 8 - stage shift calibration, edit it into the 2nd “999” in SerialEMproperties.txt like below.

RotationAndPixel 17 999 86.1 3.396

Note

The pixel size and tilting axis can just be done for a couple of switching mags such as the lowest M and the highest LM. SerialEM uses these couple of calibrations and all the Image Shift calibration to interpolate to obtain the pixelsizes and tilting axis angles for all other magnifications. This is very cute. HOWEVER, this is for a scope to have consistent image shift for all the mags. For some scopes this might not be true. Thus it is not a bad idea at all to just have pixelsize calibrated for all the important mags you will use.

When mag gets high, you can use montage overview images to calibrate pixelsize. With Image Shift already calibrated well, montaging using Image Shift can be pretty fast. So in the end you also have lines like below from menu Calibration - Pixel Size - List Relative Rotations

RotationAndPixel 18    0.22  999   2.602   # 1600, p=4164
RotationAndPixel 19   -0.29  999   2.006   # 2000, p=4012
RotationAndPixel 20   -1.29  999   1.571   # 2600, p=4084
RotationAndPixel 21   -0.25  999    1.22   # 3400, p=4149
RotationAndPixel 22    0.08  999   0.954   # 4300, p=4102
RotationAndPixel 24    999   999  0.6203   # 6700, p=4156
RotationAndPixel 25    0.02  999  0.4912   # 8500, p=4176
RotationAndPixel 26    0.58  999  0.3799   # 11000, p=4179
RotationAndPixel 27   -0.05  999  0.2992   # 13500, p=4039
RotationAndPixel 28   -0.58  999  0.2383   # 17500, p=4170
RotationAndPixel 29    3.37  999  0.1849   # 22000, p=4069
RotationAndPixel 30   -0.11  999  0.1456   # 28000, p=4077
RotationAndPixel 31   -0.09  999  0.1141   # 36000, p=4108
#RotationAndPixel 32   -0.06  999  0.09005   # 45000, p=4052
RotationAndPixel 32   -0.06  999 0.087       # from KK's pixel refine
RotationAndPixel 33    0.02  999 0.07141   # 57000, p=4071

Step 12

Increase Mag by 1 click and do Calibration - Image & Stage Shift - Image Shift

Step 13

Repeat above step to cover all the magnification till the highest to be used such as 100kX.

Step 14

Decrease Mag by 1 click and do Calibration - Image & Stage Shift - Image Shift

Step 15

Repeat above step to cover all magnification till the lowest to use like 46X.

Step 16

At about 20kX, do Autofocus calibration (only need to do at single mag).

Step 17

Beam Crossover calibration

Step 18

Start with most used spotsize like 7, do Beam Intensity calibration

Step 19

repeat Beam Intensity Calibration for all other spot sizes likely to be used: 3, 4, 5, 6, 8, 9.

Step 20

At one mag like 5000X, using spot size 9, do Beam Shift Calibration (only need to do at single mag).

Step 21

Usually, people use the lowest M mag for Low Dose View beam and with large defocus offset such as -200 or -300 microns. You need to the calibrate High-Defocus Mag for this View mag. This will make stage shifts still good for such large defocus, as they are interpolated for the defocus offset.

Note

Calibrations needed to be done for both µP and nP mode include: beam crossover, beam intensity, beam shift and autofocus.
Waffle grating grid is good and handy for pixel size calibration, butit is not ideal for Image Shift and Stage Shift calibrations, as the waffle pattern might screw up the correlation in the calibration procedures. I found the normal Quantifoil grid with some 10nm Au particles absorbed onto can be very good for normal calibration purpose. I glow discharge a Quantifoil grid and add 1 µl deca-gold solution on the grid and let it dry.
I found that standard PtIr grid for TFS to perform Thon Ring test also works very well for calibration purpose.
Most of SerialEM actions are cross-correlation based, including calibrating. Therefore, a clean and recent preparation of camera gain reference file is desired, because it will help to have less screw-up due to fixed noise pattern dominating the cross-correlation.