International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B5. Istanbul 2004 
  
methods. Additionally, the pyramidal structure can be tilted 
in the SEM for calibration purposes with still all of 
nanomarkers visible to the electron beam. Also, due to the 
slope steps of the calibration object, AFM measurements are 
possible and provide the spatial information of the reference 
points that is needed for the calibration of scanning electron 
microscopes. 
4. MEASUREMENT RESULTS 
4.1 AFM measurement results 
High precision AFM measurements in non-contact mode 
were done at the PTB (Physikalisch-Technische 
Bundesanstalt, Braunschweig, Germany). The instrument 
used was a modificated SIS-AFM Nanostation III (SIS, 
Herzogenrath, Germany) with strain gauges in z-direction 
and lateral capacitive sensors to guarantee lateral high- 
precision measurements. Although the device is not 
approved for metrological measurements, results within 1% 
of uncertainty in z-direction can be expected. 
Alternatively, the nanomarker coordinates were measured 
with a “normal”, commercially available Veeco Explorer 
2000 (Veeco, Woodburry, USA) AFM in contact mode. 
Nanomarker coordinates of all measurements were detected 
using a geometric search routine, sensitive to sudden 
changes in altitude on smooth topographies. Then, the high 
precision SIS-AFM data were compared with the raw and 
corrected data of the Explorer AFM (Table 2 and Figure 5). 
Determination of the coordinates of the nanomarkers 
depends on the accuracy of the sensor as well as on the 
accuracy of the analysis used. From the analysis, a lateral 
mean point error of 0.9 Pixel has been evaluated, 
corresponding to a relative error of 0.0009 in a 1000 pixel 
scan. The relative vertical error is about 0.002. Therefore, the 
sensor is the limiting factor. This can be clearly seen, when 
comparing the accuracy of the high-precision SIS-AFM with 
a commercially available AFM, e.g. the Veeco Explorer we 
used in our first approaches for reference point 
determination. Further improvements will be possible by 
measuring with an  interferometrically controlled, 
metrological AFM (MAFM). 
  
  
  
| 
i AFM Sa = relative Sensor ! rel. S, + Nanomarker 
| Sensor | Accuracy | determination error | 
I Is IT | 
se TS jE 
| Veeco 0.013 
| calibrated | — ]- 
  
  
  
Table 2. The accuracy of the AFM measurements and 
nanomarker detection. 
5. CALIBRATION RESULTS 
With the nanomarker coordinates determined by high- 
precision AFM, we were able to calibrate a high-resolution 
field-emission SEM, the XL30 FEG as well as a XL30 ESEM 
under “wet mode” (1 Torr water vapour pressure) conditions. 
Calibration of the XL30 FEG was performed with 10 images 
tilted by steps of 5 degrees. Calibration of the XL30 ESEM 
was done with 5 images and arbitrary tilt steps. Tables 3 and 
4 show the calibrated magnification factor (m), the mean 
lateral (mxo, myo) and the mean tilt angle error calculated 
(mq, mk, mo). 
5.1 SEM calibration results 
ETT NEUE EET NIE 
| scale (m) [0.094 [pixel/nm] 
| mean (mxo, Myo) [nm] | 13.03, 13.15 [nm] —— | 
| mean (me, mx, me) | 0.781, 0.804, 0.248 [deg] | 
Table 3. Calibration results of the XL30 FEG scanning 
electron microscope. 
5.2 ESEM calibration results 
[Sensor 
scale (me 
| mean (mxo, Myo) . |l! 
| mean (mg, mx, mow) | 0. 
[0.1133 [pixelnm] — 
| 10.228, 10.075[nm] 
6 
   
  
  
— 
43, 0.654, 0.18 [deg] UT 
i 
Á i 
e 
Table 4. Calibration results of the XL30 ESEM scanning 
electron microscope. 
5.3 Spatial intersection and triangulation 
Results of the XL30 FEG calibration were tested by applying 
spatial intersection or triangulation formulas to the 
nanomarker image coordinates. 
Section y-z (Nanomarkers 13 - 18) 
NS ee ere re YY = 
| | * Spatial Intersection - 
calibrated values 
4- Triangulation - calibrated 
values 
20005600 1 re — © © Triangulation - REM settings 
-T VV 
T + © SIS-AFM 
| ^ | A 
1500.0000 + ed e i A Veeco AFM 
T | 
& i ; | 
| m 
1000.0000 + - + } MM iE 
| A | = 
500.0000 + 
i o2 | i | i sd 
NOS 
-5500.0000 -4500.0000 -3500.0000 ~260C.0000 -1500 0000 -500.0000 500.0006 
x [nm] 
Figure 5. Section in y-z of nanomarkers 13-18 from AFM 
measurements and from applying spatial intersection or 
triangulation formulas. 
The nanomarker coordinates in Figure 5 were calculated 
either from the calibrated values or from the microscope and 
tilting stage settings. Spatial nanomarker data from SIS- 
AFM and Veeco AFM are shown for comparison. In lateral 
direction, we found a good match of the calculated 
coordinates of the REM data with the AFM measurement. 
However, in vertical direction errors up to 5% did occur. 
Comparison of SIS-AFM and Explorer AFM data showed 
great discrepancy, even with calibrated Explorer data. 
Therefore, the importance of using high-precision AFM for 
accurate 3D micro-measurements is clearly underlined. 
5.4 Correlative investigations 
The calibration object has been tested preliminarily for other 
3D micro-measurement methods, e.g. CLSM (Leica, 
Bensheim, Germany) and laser profilometry (Nanofokus, 
Oberhausen, Germany). Resolution of both measurement 
methods is insufficient to visualize the nanomarkers, but 
profile plots of the CLSM and profilometer measurements 
were compared to the original high-precision AFM 
measurement (Figure 6). 
     
  
  
   
    
   
    
  
    
   
  
   
  
   
   
  
  
  
  
  
   
   
  
  
   
   
  
  
  
  
  
  
  
  
  
     
  
  
   
  
     
    
  
   
   
   
   
  
  
    
  
    
     
   
  
   
   
     
z [nm] 
Fi 
Sii 
th