Logo of the Physikalisch-Technische Bundesanstalt
Production sequence of Si-spheres and interferometrical determination of the sphere volume

AFM Linewidth Metrology

Working Group 5.23

Measurements and calibration capabilities


Measurements and calibrations of e.g. feature width, sidewall angle, line edge roughness, sidewall profile etc. can be performed. Selected measurement examples are demonstrated.


Calibration of feature width

 A SCCDRM sample developed at the National Institute of Standards and Technology (NIST) of the United States has been measured using the 3D-AFM. The SCCDRM has a size of approximately 1 cm x 1 cm x 1 mm. The top surface of the standard is comprised of a number of patterns, many of them consisting of up to six line features of different widths ranging from 30 nm to 250 nm, as shown in figure 7 (a). The measurement area is precisely selected by using the alignment markers. The measurement of the SCCDRM with the 3D-AFM was performed in vertical oscillation mode using a flared AFM tip of type CDR120. The driving frequency was set slightly below the vertical resonance frequency of the cantilever, and the structure was probed at a velocity of 1000 nm/s. An overview image of five SCCDRM structures taken with the 3D-AFM is shown in figure 9(b), and a cross sectional profile is shown in figure 7(c). Note that the 6th structure of this target was not measured due to a defect in the feature.


Fig.7 Layout of linewidth features of a single crystal critical dimension reference material (SCCDRM) is shown in (a), an overview 3D-AFM image of five line features in (b), and a cross sectional profile (c). The shown profile includes the dilation effects of the tip width.


Individual structures of the SCCDRM sample were measured using the 3D probing strategy. The measurement area was set to be about 0.8 μm × 1.0 μm. The profile was measured with 50 points each in the left bottom, left sidewall, top, right sidewall and right bottom regions each, and with 5 points each in the left-bottom, left-top, right-top and right-bottom corner regions. One complete profile thus contains 270 measurement points. Sixteen profiles were acquired in one measurement run, which takes about 20 minutes. A 3D view of the linewidth feature “B” (see figure 7) is shown in figure 8(a), and a typical cross-sectional profile is shown in figure 8(b).  These plots represent raw data without smoothing or filtering. The inclination of the profile was linearly fitted and removed. To evaluate the middle CD value, segments of the profile at the left and right bottom regions and at the top region (indicated in red) were selected for calculating the structure height, H. It should be noted that the distance from the profile segments to the corners of the structure should be properly selected (one-third of the line width was used in this study) so that the corner rounding regions are excluded. To achieve better statistical performance, the portions of the left and right sidewalls within the z range of 0.3*H and 0.7*H are linearly fitted, as shown, respectively in figures 8(c) and (d). For purposes of the middle CD calculation, the sidewall edge positions are taken as the intersections points between the lines fitted to the sidewalls with a horizontal line at half height. The middle CD is then evaluated as the distance between these two points. The slopes of the fitted sidewall segments are taken as the measurement of feature sidewall angles.



Fig.8 3D view of the measured linewidth feature B is shown in (a), a cross-sectional profile of the feature is shown in (b), the left and right sidewalls of the feature are zoomed-in and shown in (c) and (d).


The calculated middle CD values from the 16 profiles of 10 repeated measurements are shown in figure 9. It can be seen that the line width roughness (LWR) of the structure is not negligible and exhibits a variation in CD values with profile location of up to 10 nm. Note, however, that the CD values measured at the same profile position repeat very well. The mean CD value for a measurement was calculated as the average result of the CD values from the 16 profiles. The standard deviation of the mean of the CD values from the 10 repeated measurements was 0.1 nm – indicated excellent static repeatability of the instrument.

Fig. 9 Measured middle CD values on different profiles in 10 repeated measurements using vertical oscillation mode. (The given CD values include the effective tip width.)


Calibration of sidewall angle

The left and right sidewall slopes calculated from the 16 profiles of 10 repeated measurements are shown, respectively, in figure 10 (a) and (b).  The slope values have a standard deviation of about 0.1°, which can be attributed to both the sidewall roughness and the measurement noise. The averaged values of the left and right sidewall slopes are 89.5° and 89.4°, respectively, indicating that the sidewalls of the SCCDRM features are nearly vertical.

Fig. 10 Measured left and right sidewall slopes of structure B  in 10 repeated measurements are shown in (a) and (b), respectively.


An investigation of the long term stability of the 3D-AFM was performed by measuring a PTB Cr on quartz photomask. A line feature with a nominal width of 300 nm was measured using vertical oscillation mode and the 3D probing strategy over an area of 750 nm × 1000 nm.  Prior to the first measurement, a new flared AFM probe (type CDR120) was mounted. All measurement data were recorded and evaluated so that the tip wear could be monitored from the very first measurement. The whole investigation lasted about 30 hours during a weekend.  A total of 197 measurements were performed with each taking about 9 minutes.  The measurement process was fully automated and did require any operator intervention once the measurement task was defined and initiated. The measured values of middle are plotted in figure 11.  It can be seen that the variation of the middle CD is less than 1 nm. A linear fit of the measured CD values, shown as the red line in figure 13, allows the rate of change in apparent width to be estimated as approximately 0.00033 nm per scan line. This result indicates that the 3D-AFM has a very high measurement stability and very low tip wear.

Fig.11 Long term stability in CD measurements using vertical oscillation mode. (The given CD values include the effective tip width.)


Calibration of line edge roughness and line width roughness

The 3D-AFM is capable of measuring the line width roughness (LWR). A preliminary measurement example is shown in figure 14. In this investigation, the line structure of the photomask was measured with 128 profiles. Two edge points at the middle height of the structure are evaluated from each profile as shown in figure 12 (a). The width variation of the structure at different x positions (profile location) is shown in figure 12 (b) for two repeated measurements.

Fig.12 Measurement of line width roughness (LWR), shown as the measurement strategy in (a) and the measured LWR of a line structure of a photomask in (b).


The 3D-AFM is able to probe surfaces in 3 dimensions (xyz) and measure surfaces in an arbitrary plane (xz, yz, xy etc.). This feature allows, for example, very convenient and direct measurements on sidewalls. Such a measurement is illustrated in figure 13. During this measurement, the vertical sidewall of the IVPS sample was measured in the xz plane using a flared AFM tip type CDR120 probing in the y direction. The 3D view of the measured (vertical) sidewall of the structure is shown in figure 13(a). A measured profile is shown in figure 13(b) which was taken in two successive measurements, showing very good measurement repeatability. Note that the apparent inclination of the profile is attributed to the misalignment of the structure along the x-axis.



Fig.13 3D view of the measure sidewall topography of the vertical surface of the IVPS sample in (a) and a profile taken in two successive measurements in (b).