next up previous contents
Next: Auger Electron Spectroscope Up: Analysis Methods Previous: Electron Micro Probe

Transmission Electron Microscope (TEM)

 

In this chapter we will try to list all the information which is important for the simulation of eletron-target interactions and the generation of events caused by these interactions in a Transmission Electron Microscope (TEM).

We will begin this section with a description of a simple example. A more general explanation of the features and options will then be given at the end of this chapter (see General Description below). For a quick information on default values, see Table 4.3.

Example:

By means of a simple example, it will be demonstrated, step by step, how to perform an electron scattering simulation, starting from the generation of a geometry and an input file, to the visualization of the results and their interpretation.

The actual geometry and input files as read by SESAME during program execution are described here. These files might be also generated using the graphical user interface (GUI). For more information on the GUI see Chapter 5.

The geometry of this test example is shown in Figure 4.11.

 

To generate this geometry file (without using the GUI) you have to:

  1. Determine how many different regions your sample consists of. In this case there are three different regions, two are made of and the third one of Au. This is the only number you have to write in the first line of your geometry file.
  2. Give each vertex (point) of your geometry a consecutive number starting with one.
  3. Starting from region 1 you have to define each region by going from vertex to vertex around the regions in clockwise direction. For each region you have to write the number of vertices of this region and optionally, separated by a blank, the corresponding segment number on a separate line followed by all the numbers of the vertices, again each on a separate line (see copy of input file below).
  4. Enter the number of different vertices of the whole structure in a separate line.
  5. Write coordinate pairs (x,z) for each vertex in a separate line of the input file. Note: Units of coordinates are always Å!
  6. Save file (for instance: TEM_EXA1.GEO).
For a more accurate description of the geometry specification see Chapter 6.6. The geometry file should look like:
3                               number of regions
8 1                             number of vertices (1st region), segment
 1                              vertex number
 2                                   ''
 3                                   ''
 4                                   ''
 5                                   ''
 6                                   ''
 7                                   ''
 8                                   ''
5 2                             number of vertices (2nd region), segment
 6                              vertex number
 5                                   ''
 4                                   ''
 3                                   ''
 11                                  ''
7 3                             number of vertices (3rd region), segment
 2                              vertex number
 9                                   ''
 10                                  ''
 7                                   ''
 6                                   ''
 11                                  ''
 3                                   ''
11                              overall number of vertices
0.0,0.0                         coordinates (x,z) of first vertex
45000.0,0.0                     coordinates (x,z) of second vertex
45000.0,10000.0                 etc.
40000.0,12500.0
45000.0,15000.0
60000.0,15000.0
60000.0,20000.0
0.0,20000.0
100000.0,0.0
100000.0,20000.0
60000.0,10000.0

The next step is to write an input file. Following, you will find an example of an input file for this sample. This file might be generated either by means of the GUI or by using any text editor. You may also copy this file to your account from:

 
VMS 		  DISK$AXP_USER1:[SLED.NEW.SESAME.EXA]TEM_EXA1.INP

Unix tcad/sesame/exam/tem_exa1.inp

 &INIT   CUSTXT='THIS IS A FIRST TEM TEST EXAMPLE'
         JOB='TEM_1',TECHQU='TEM'
         SIGNAL='TE',UNITS='MICRON',&END
 &ETRNS  E0=100.,EF=500.,NETRN=1000,&END
 &EBEAM  ESTD=5,SCAN=F,SPSIZE=0.0035,DIVERG=1,&END
 &TARGET ATMAT='AL','AU','AL'
         DENS=2.7,19.3,2.7
         NREG=3,CGFILE='TEM_EXA1.GEO',&END
 &MODEL  MODR(1)='HYB',&END
 &MSETUP TAKOFF=37.5,XDEROT=0,EBTILT=0,EBROT=0,XESPOT=4.25,&END
 &OUTPUT CGRDFL='TEM_EXA1.GRD'
         SCREEN(1)='PET;,,600'
         SCREEN(2)='PEB,1'
         SCREEN(3)='BET,1'
         SCREEN(4)='TET,1'
         SCREEN(5)='ION,,3.8,6.2,0.8,1.7;,,600;AU,M'
         SCREEN(6)='XGD,,3.8,6.2,0.8,1.7;,,600;AU,MA'
         SCREEN(7)='XDD,,3.8,6.2,0.8,1.7;,,600;AU,MA;1'
         &END

For a detailed description of all namelists (input records) and their variables see Chapter 9. Here only short explanations of the variables used in this specific input deck will be given.

The input file should start with the &INIT namelist (see Chapter 9.1) although no obligatory variables exist for this namelist, i.e. all variables of this input record have default values.

CUSTXT is a user-defined text which should contain a short description of the example and which will appear in all generated output files. With the variable JOB a job name may be specified ( Note: No blanks are allowed!). The job name will appear in the names of generated output files if no file names for such files are explicitly specified in the input deck. The default value for the job name is the name of the input file without the extension, this would be TEM_EXA1 in this case.

TECHQU defines the measurement technique used. This variable guarantees the presetting of default values for many different variables appropriate for the specified technique. For the default settings for TECHQU=TEM see Table 4.2.

UNITS specifies the units of length used in the current input file. This variable is relevant for all length-dependent quantities (i.e. SPSIZE in the &EBEAM namelist (see Chapter 9.3), XESPOT and XRANGE in the &MSETUP namelist (see Chapter 9.6), and SCREEN, POST, PMSAV, and HIST in the &OUTPUT namelist (see Chapter 9.9) - the meaning of all these variables will be given in one of the next paragraphs). The default value is MICRON.

The next input record, &ETRNS (see Chapter 9.2), must be contained in each input file since the variables E0, the initial electron energy, and NETRN, the total number of electrons, are obligatory. For EF, the final electron energy, the default value is 100 eV ( Note: unit for E0 is keV, for EF it is eV!).

The namelist &EBEAM (see Chapter 9.3) describes the features of the electron beam. SCAN=F means that the beam should not be scanned (the electron impact position is kept at one location), ESTD determines the standard deviation of electron energy in eV (Gaussian distributed), SPSIZE describes the spot size of the electron beam at the electron impact position (see Chapter 6.1 and 9.3), and DIVERG specifies the beam divergence in degrees.

The specimen is described by the &TARGET namelist (see Chapter 6.5, 6.6 and 9.4). ATMAT is a string array containing the chemical abbreviations of all elements of the target and their relative fractions according to the different regions. The chemical abbreviations in one region as well as the strings describing different regions must be separated by commas ( Note: strings must be enclosed by single quotes!). Relative fractions of elements in one region must, if necessary at all, be specified right after the corresponding chemical abbreviation without a blank, comma or any other character inbetween. In this particular example, for instance, the first region is composed of pure Aluminum (Al), the second region is pure Gold (Au), and the third region is pure Aluminum (Al) again. For more information on specifying the target composition see Chapter 6.5.

DENS specifies the densities (in ) of the regions in the same sequence in which the regions are defined in the geometry file, i.e. the first real value in DENS corresponds to the first region, etc.. NREG determines the number of regions and CGFILE specifies the file name for the simulation geometry file ( Note: Do not forget the single quotes!). All of these variables are obligatory and must be specified in the input file.

In &MODEL (see Chapter 9.5) the physical model may be specified using the MODR variable in dependence of the region (index is the region number). The default is the single scattering model SIN (see also Chapter A.3). If a model is specified only for the first region then the same model will be applied to all other regions too. The only restrictions one may face are according to the direct simulation model, which is not available for every element and compound (see Chapter A.5).

In the &MSETUP namelist (see Chapter 9.6) the measurement setup is defined. TAKOFF is the x-ray detector take-off angle, XDEROT the the x-ray detector rotation angle, EBTILT the beam tilt angle, and EBROT the beam rotation angle (see Figure 6.2 and 6.3). All angles are measured in degrees. Actually, all detector angle variables are arrays which allow to specify more than one detector for one simulation. Up to five x-ray and electron detectors may be specified simultaneously. For more information on the specification of x-ray and electron detectors see Chapter 6.3 and 6.4, respectively.

XESPOT is the lateral electron beam impact position (see Figure 6.4) and is in units specified by the UNITS variable in the &INIT namelist, microns in this particular case.

The variables of the &OUTPUT namelist (see Chapter 9.9) control the output generated by SESAME. For a more detailed description refer also to Chapter 6.7.

If you never started SESAME before, then follow the instructions in Chapter 3 or 5, respectively. Otherwise simply start SESAME by executing the command "sesame" on your command line or invoke it from the VISTA TCAD framework (see Chapter 5).

For this particular input deck, one status window and four output windows will appear on your screen. The status window will contain information like electron energy, used CPU time, CPU time per electron, remaining CPU time, ionization sympols legend, and some more. The first output window will contain the electron beam ( PEB) and the trajectories of the primary electrons inside the sample ( PET), and outside the sample, the backscattered ( BET) and transmitted ( TET) electrons. The "1" after PEB, BET and TET in SCREEN(2), SCREEN(3), and SCREEN(4) means that the information specified by this variable should be directed to SCREEN(1).

On the second output window the ionizations of the gold (Au) M sub-shells will be visualized. The third and forth window contain the generated and detected gold x-ray distribution, respectively ( XGD stands for "generated x-ray distribution" and XDD for "detected x-ray distribution" (see Chapter 6.7 and 9.9)).

The integer after the third semicolon determines the x-ray detector to apply. Presently, up to 5 different detectors may be specified in the input deck. If no detector number is explicitly specified then detector number zero - the default detector - is to be used (see Chapter 6.3).

For XGD and XDD a grid is needed. If no grid is explicitly specified in the input deck (see Chapter 6.8) a grid will be automatically generated by SESAME during runtime. In this particular case we supply an external grid by specifying a grid file name using the CGRDFL variable in the &OUTPUT namelist (see Chapter 6.7 and 9.9). This grid file may be copied from:

 
VMS 		  DISK$AXP_USER1:[SLED.NEW.SESAME.EXA]TEM_EXA1.GRD

Unix tcad/sesame/exam/tem_exa1.grd

You may wish to modify the user defined grid. This can easily be done by editing this grid file, which contains the following lines, using any text editor:

  2
   40000.
    20
   60000.
  2
   10000.
    10
   15000.
This grid file specifies a rectangular grid in the region between 40000 and 60000 Å in lateral direction with 20 gridlines in-between these two values (i.e. altoghether 22 grid lines) and between 10000 and 15000 Å in vertical direction with 10 gridlines in-between (12 grid lines altogether).

To demonstrate how the x-ray intensity changes with detector position we modify the following line in the input deck:

 &MSETUP TAKOFF=25,XDEROT=0,EBTILT=0,EBROT=0,XESPOT=3,&END
TAKOFF is changed to 25 degrees, this means, that the path of the x-rays in the sample is now increased. As a consequence the absorption is also increased and the detected signal stems from a shallower region.

It is also possible to specify more than one detector in a single simulation. If you modify the following line in the input deck, for instance,

 &MSETUP TAKOFF=37.5,25,XDEROT=0,0,EBTILT=0,0,EBROT=0,0,XESPOT=3,&END
then you can apply either detectors to any of the modified distributions. To apply the second detector to the x-ray distributions specified above you have to change the following lines in your input deck:
         SCREEN(6)='XGD,,3.8,6.2,0.8,1.7;,,600;AU,MA;2'
         SCREEN(7)='XDD,,3.8,6.2,0.8,1.7;,,600;AU,MA;2'
The number after the third semicolon determines the x-ray detector number. In this particular case the second detector shall be applied. If no detector number is explicitly specified then detector number zero - the default detector - is to be used (see Chapter 6.3).

While the first window shows the whole simulation geometry, the second, third and fourth window visualize only the part of the geometry which lies between 3.8 and 6.2 microns in the lateral and between 0.8 and 1.7 microns in the vertical direction. The size of the windows is explicitly specified and is 600 pixels in lateral direction. Since no value for the vertical window size is specified this quantity is chosen so that the geometry is not distorted. For information on how to change the clipping visualized in the window and how to modify the window-size refer to Chapter 6.7 and 9.9. There you will also find more details about directing output to files using various formats.

So far, we have seen only electron trajectories, ionization and x-ray distributions. We shall now try, for instance, to visualize information more typical for TEM measurements. For that put the following two lines to the input deck ( &OUTPUT namlist) [all extensions to the input file described in this chapter are contained in the file TEM_EXA2.INP which also may be copied to your account]:

         SCREEN(8)='TEP'
         SCREEN(9)='TEE'
Running the extended input deck will bring two more windows to your screen. In one window you will see the polar angle distribution ( TEP) of transmitted electrons and in the other the energy distribution ( TEP) of transmitted electrons (see Chapter 6.7 and 9.9). In this case all backscattered electrons will be collected and visualized in a histogram in dependence on their polar angles and of their energies, respectivelyc since no detector number is explicitly specified. To take into account an electron detector, a detector number must be spcified after a third semicolon in the string of SCREEN(8) and SCREEN(9). For the specification of an electron detector see Chapter 6.4 and 9.8.

Next, we will change the input deck in that way that the sample is scanned. To do so modify the following lines in the input file:

 &EBEAM  ESTD=5,SCAN=T,SPSIZE=0.0035,DIVERG=1,&END
 &MSETUP TAKOFF=37.5,XDEROT=0,EBTILT=0,EBROT=0
         MAGNI=20000,XESPOT=5,XRANGE=6,&END
A magnification of 20kX is specified using the MAGNI variable of the &MSETUP namelist. The sample will be scanned over a range of 6 microns which is specified by XRANGE. XESPOT determines the center position of the scanning range (see Figure 6.4). In this particular case the sample is scanned between positions 2 and 8 microns (in the simulation geometry coordinate system). For further information on specification of scanning see Chapter 6.2 and 9.6.

Now, we would like to record the transmitted electrons in dependence of the scanning position. Therefore add the following line to your input file:

         SCREEN(10)='TEY'
An additional window will be opened up on the screen. This window shows a histogram containing the number of transmitted electrons in dependence of scan position. All transmitted electrons, independent of their energy and their direction, are counted since no detector number is explicitly specified. One can clearly see where the gold inclusion is located in the sample. By varying the initial electron energy it can be easily demonstrated how the detection of the gold inclusion changes (see Figure 4.15).

Furthermore, we will briefly show how one can write generated output to files. For that, put the following lines to the &OUTPUT namelist of the input file:

         PMSAV(1)='XGD;TEM1_XGD.SAV;AU,MA'
         PMSAV(2)='XDD;TEM1_XDD1.SAV;AU,MA;1'
         PMSAV(2)='XDD;TEM1_XDD2.SAV;AU,MA;2'
         HIST(1)='TEY;TEM1_TEY.H1D;;1'
         HIST(2)='BEY;TEM1_BEY.H1D;;1'
         HIST(3)='TEP,20,90,180;TEM1_TEP.H1D;;1'
         HIST(4)='TEE,15;TEM1_TEE.H2D;;1'
PMSAV() produces a POSTMINI SAV file (PROMIS format), one named " TEM1_XGD.SAV", the second " TEM1_XDD1.SAV", and the third " TEM1_XDD2.SAV" containing the generated and detected (for detector 1 and 2) characteristic gold x-ray distribution. These binary files may be visualized by POSTMINI. For further information on the use of PMSAV see Chapter 6.7 and 9.9. More information concerning POSTMINI can be found in Chapter 10.1. The two distributions generated by means of the first two lines given above are visualized in Figures 4.12 and 4.13 (sample not scanned!).

 

 

By means of the HIST() variable ASCII files are generated. The characters between the first and the second semicolon represent the file name. The first numbers in front of the first semicolon in TEP and in TEE define the histogram intervals. In this particular case, the polar angle distribution of transmitted electrons ( TEP) shall be stored in a 1-dimensional histogram with 20 intervals, and the energy distribution in one with 15 intervals. Since only the transmitted electrons should be collected, the range of angles must be restricted to the range between 90 and 180 degrees. For the energy distribution no range is explicitly specified which means that all electrons with energies between E0, the initial electron energy, and 100 eV are to be counted (default values).

Such ASCII files can be visualized by using different tools, for instance POSTMINI, GNUPLOT or MUFASplot (see Chapter 10.1, 10.3 and 10.4). For further postprocessing of data VIDE may be applied (see Chapter 10.2). Additional information on the use of HIST is provided in Chapter 6.7 and 9.9. Figure 4.14 shows the plot of the polar angle distribution of the transmitted electrons. In Figure 4.15 the yield of the transmitted electrons and in Figure 4.16 the yield of backscattered electrons, both in relation to the scan position, is visualized for four different energies (50, 75, 100 and 125 keV). POSTMINI was applied to generate the encapsulated PostScript files plotted in this manual.

 

 

 

Another possibility using the TEM in analytical mode, i.e. to record the generated characteristic x-rays, shall be demonstrated by adding the following two lines to the input deck:

         SCREEN(11)='XCY;;AU,MA'
         SCREEN(12)='XCY;;AL,KA'
Two additional windows will be opened on the screen. One will show the x-ray scan for the gold signal and the second that for the aluminium signal. Both linescans are plotted in Figure 4.17. It can be seen that also the Al signal may be used for the localization of the inclusion to some extent. But the Au signal provides much better information even clearly reflecting the asymmetry of the gold inclusion.

 

The last feature to be mentioned here is the possibility of splitting your input deck into up to 4 files. You have to provide a so called "master input file" which is the actual input file read by SESAME. In this input file you may specify three further input files by means of the FINSTR, the FDETCT, and the FMYDEF variables of the &INIT namelist (see Chapter 9.1) which SESAME includes in the master input file.

FINSTR specifies the file name for a file containing instrument-specific parameters, FDETCT one for detector-specific parameters, and FMYDEF one for user-specific parameters. But actually, any of these three files may contain all of the namelist records, except for the &INIT namelist ( Note: the &INIT namelist can only be specified in the master input file!). Important to know is that the specifications in the master input deck have the highest priority, followed by that in the FMYDEF file, then the FDETCT file, and last the FINSTR file. For instance, a take-off angle specified in the FDETCT file will be overridden if it is also specified in the FMYDEF or in the master input file.

For demonstration purposes we will split up our original input deck into different files. Then the master input deck may have the following form:

 &INIT   CUSTXT='THIS IS A FIRST TEM TEST EXAMPLE'
         JOB='TEM_1',TECHQU='TEM'
         SIGNAL='TE',UNITS='MICRON'
         FMYDEF='SAMPLE_3.INP'
         FDETCT='DETECT_3.DAT'
         FINSTR='INSTR_3.DAT',&END
 &ETRNS  E0=100.,EF=500.,NETRN=1000,&END
 &EBEAM  SCAN=F,&END
 &MODEL  MODR(1)='HYB',&END
 &OUTPUT CGRDFL='TEM_EXA1.GRD'
         SCREEN(1)='PET;,,600'
         SCREEN(2)='PEB,1'
         SCREEN(3)='BET,1'
         SCREEN(4)='TET,1'
         SCREEN(5)='ION,,3.8,6.2,0.8,1.7;,,600;AU,M'
         SCREEN(6)='XGD,,3.8,6.2,0.8,1.7;,,600;AU,MA'
         SCREEN(7)='XDD,,3.8,6.2,0.8,1.7;,,600;AU,MA'
         &END
The FMYDEF file may look like:
 &TARGET ATMAT='AL','AU','AL'
         DENS=2.7,19.3,2.7
         NREG=3,CGFILE='TEM_EXA1.GEO',&END
 &MSETUP XESPOT=4.25,&END
In FDETCT some detector specific parameters should be specified
 &MSETUP TAKOFF=37.5,XDEROT=0,&END
 &DETECT XFDEFF='DETECT_1.DAT',&END
and in FINSTR certain parameters describing a typical measurement setup for a specific instrument shall be defined:
 &EBEAM  ESTD=5,SPSIZE=0.0035,DIVERG=1,&END
 &MSETUP EBTILT=0,EBROT=0,&END
In this way, it is very easy to combine the specifications for different instruments, detectors and other specific parameters. Thus, if a measurement is repeated with a different instrument then it might be sufficient to change solely the file names specified by the FDETCT and FINSTR variable in the master input file. Once instrument- or detector-specific parameters are provided for each individual instrument (or instrument setting) simulations for different instruments/detectors can be performed by only making a very few modifications in the master input file. This feature also allows the users to specify their own set of default parameters. All in all, this option helps to keep the actual input file (master input file) short and clearly arranged.

Here, at the end of the example description it must be pointed out that only a relatively small part of all available features has been demonstrated. For more detailed information refer to Chapter 6 where SESAME features are described more comprehensively, and to Chapter 9 where all namelists and their variables are listed.

An example using a more sophisticated simulation geometry is described in Chapter gif.

General Description:

In the input file (see Chapter 9) the measurement technique can be specified by means of the TECHQU variable (see Chapter 9.1). If this variable is set to TEM then many input variables are set to default values which are appropriate for this analysis technique. The most important parameters and their default settings are listed in table 4.3. Thus, only those parameters with differing values must be explicitly specified in the input file.

  
Table 4.3: Default values for NAMELIST variables for TECHQU=TEM

As can be seen from the Table 4.3, the electron beam is, by default, not scanned, i.e. the beam impact position is kept constant during the whole simulation. The default electron beam impact location is the center position of the specified sample geometry. This position may be changed by explicitly specifying the electron beam impact position using the XESPOT variable in the &MSETUP namelist (see Chapters 6.1 and 9.6).

The default detector (detector number zero) for electrons is an ideal 2 detector. This means all transmitted (or backscattered) electrons are fully collected independent of their directions and their energies. In case of x-rays, the detector is located by default at a take-off angle of 30 degrees and a rotation angle of zero degrees (see Chapters 6.3 and 9.8). All x-rays leaving the sample in that direction are counted, i.e. the default detector efficiency is one, independent of the x-ray energy. More realistic detectors may be specified using the &DETECT namelist (see Chapter 6.3 and 6.4).

The default signal in TEM mode is TE (transmitted electrons). The default model is the hybrid model (see Chapter A.4). This model is used because it is faster by approximately a factor of two compared to the single scattering model. But in cases where ionizations and characteristic x-rays are calculated, the single scattering model yields much better statistics for the same number of simulated electrons, i.e. with respect of reaching the same accuracy the single scattering model is faster.

The single scattering model uses a strict continuous slowing-down approximation. Therefore, ionizations are not explicitly calculated but the ionization probabilities along the electron trajectory are evaluated. This is why no ionizations can be exactly localized in the sample and ionization events are therefore randomly distributed along the electron trajectory.

But generally, other signals as well, like backscattered electrons, may be used and specified. For all these signals which may be chosen in TEM mode (see Table 9.3) all three models may be applied (see Table 9.2).

Since both, the single scattering model ( SIN) and the hybrid model ( HYB) calculate energy loss by means of a continuous slowing-down aproach, the user must be aware that a simulated transmitted (and backscattered) electron spectrum cannot show the elastic peak (zero energy-loss electrons). Due to this approach, each electron penetrating the sample suffers a finite energy loss so that none of the electrons going through (and back out of) the sample (transmitted/backscattered) are able to retain their original energy. Therefore, if the elastic peak of transmitted and backscattered electrons should appear in the simulations then the direct simulation model must be applied. Presently, this is only possible for the elements and compounds listed in Table A.1.



next up previous contents
Next: Auger Electron Spectroscope Up: Analysis Methods Previous: Electron Micro Probe



Horst Wagner
Tue Mar 19 10:24:55 MET 1996