**4.1 Step structure measurement**

#### **4.1.1 10 μm standard step height measurement**

A 10 μm standard step height fabricated by VLSI (9.976 μm+0.028 μm) was measured, as shown in figure 10, the mean height is 9.984 μm, while the standard deviation is 0.010 μm. The result comparison was presented in figure 11. We can clearly see the step height derived from WLPSI stays in the middle of all the results. Since the algorithms from the second group can achieve higher measurement resolution, the WLPSI can improve the measurement resolution compared with that of the algorithms from the first group.

Fig. 10. Measurement of the 10 μm standard step height

MEMS Characterization Based on Optical Measuring Methods 101

light signals. In this work, we repeated the measurements with WLPSI on the above 44 nm standard step height in figure 12. As illustrated in figure 13, it gave out correct results

(a) 3D structure (b) Profile

Fig. 14. Comparison of the results from the 44 nm standard step height calibrations

**Standard** 

**Mean height(nm)** 43.36 43.16 42.84

**deviation(nm)** 0.14 0.13 0.27

Figure 14 and table 1 show the 44 nm standard step height evaluations comparison between PSI, WLPSI and WLI. We can clearly see that the results from WLPSI ended up basically in the middle of these three methods and it is close to the results from PSI, which also shows the ability of WLPSI in improving the measurement accuracy compared with the traditional WLI.

Table 1. Measurement results of the 44 nm standard step height using different methods

**PSI WLPSI WLI** 

Fig. 13. Measurement of the 44 nm standard step height with WLPSI

without batwings.

Fig. 11. 10 μm standard step height measurement comparison

#### **4.1.2 44 nm standard step height measurement**

Many coherence peak detecting algorithms in white light interferometry perform the height evaluations through either recovering the envelop function or locating the centroid of the correlogram. It works quite well for most surfaces, however, when it comes to some step height alike structures that lower than the coherence length of the illuminator, overshoot will be observed at these discontinuities. This overshoot is also known as batwings, which comes out in form of the high frequency information in the measurement results. Normally, the larger magnification objective always has the higher cutoff spatial frequency, so it will also bring stronger batwings. The batwings is shown in figure 12, where a 44 nm standard step height manufactured by VLSI (43.2 nm±0.6 nm) was measured by Fourier transform.

(a) 3D structure of the batwings (b) Profile of the batwings

Fig. 12. Measurement of the 44 nm standard step height with batwings

During the vertical scanning, on condition that the step discontinuities is lower than the coherence length, the light diffracted by the top edge and the light reflected from the bottom will interfere with each other and then travel back to the interferometer. It is because of this diffraction that deforms the correlogram and produces batwings. However, the batwings never occurred in PSI measurements, which gives us a clue to find out the solution for this problem: intensities are much easier to be affected than the phase information in the white

Many coherence peak detecting algorithms in white light interferometry perform the height evaluations through either recovering the envelop function or locating the centroid of the correlogram. It works quite well for most surfaces, however, when it comes to some step height alike structures that lower than the coherence length of the illuminator, overshoot will be observed at these discontinuities. This overshoot is also known as batwings, which comes out in form of the high frequency information in the measurement results. Normally, the larger magnification objective always has the higher cutoff spatial frequency, so it will also bring stronger batwings. The batwings is shown in figure 12, where a 44 nm standard step height manufactured by VLSI (43.2 nm±0.6 nm) was measured by Fourier transform.

> (a) 3D structure of the batwings (b) Profile of the batwings

During the vertical scanning, on condition that the step discontinuities is lower than the coherence length, the light diffracted by the top edge and the light reflected from the bottom will interfere with each other and then travel back to the interferometer. It is because of this diffraction that deforms the correlogram and produces batwings. However, the batwings never occurred in PSI measurements, which gives us a clue to find out the solution for this problem: intensities are much easier to be affected than the phase information in the white

Fig. 12. Measurement of the 44 nm standard step height with batwings

Fig. 11. 10 μm standard step height measurement comparison

**4.1.2 44 nm standard step height measurement** 

light signals. In this work, we repeated the measurements with WLPSI on the above 44 nm standard step height in figure 12. As illustrated in figure 13, it gave out correct results without batwings.

Fig. 13. Measurement of the 44 nm standard step height with WLPSI

Fig. 14. Comparison of the results from the 44 nm standard step height calibrations


Table 1. Measurement results of the 44 nm standard step height using different methods

Figure 14 and table 1 show the 44 nm standard step height evaluations comparison between PSI, WLPSI and WLI. We can clearly see that the results from WLPSI ended up basically in the middle of these three methods and it is close to the results from PSI, which also shows the ability of WLPSI in improving the measurement accuracy compared with the traditional WLI.

MEMS Characterization Based on Optical Measuring Methods 103

With WLPSI, a micro resonator manufactured by Microelectronics Center of North Carolina (MCNC) was measured. With a 20× Mirau objective, the micro resonator was illuminated by a white light illuminator (central wavelength is 600 nm), the scanning range was configured to be 9 μm and the scanning step was 45 nm. The result is shown in figure 17. The height of

(a) 3D structure of the resonator (b) Comb-finger profile of the resonator

A 20× objective was used in the dynamic experiments. The stimulating signal was sine waveform with 10V amplitude, 20V offset voltage. The exposing time of CCD was 20ms for LED illumination and 100μs for LD illumination. The strobed pulse percent was 100%. The strobed phase number was 8 for in-plane motion measurement and 16 for out-of-plane

In order to get the resonant frequency of the tested device, the sweep-frequency measurement in a large range was firstly needed. The frequency increases in the logarithm mode, so the range of resonant frequency can be decided through one time frequency

**value(nm)** 

1051.8 1.43

**Standard deviation(nm)** 

**No. Thickness(nm) No. Thickness(nm) Mean** 

1 1053.1 6 1050.9

2 1053.1 7 1052.3 3 1050.0 8 1052.1 4 1048.8 9 1052.1 5 1050.9 10 1053.1

Table 2. Measurements on film thickness standard

the comb-finger profile is approximately 3.8 μm.

**4.4 Micro-resonator measurement** 

Fig. 17. Measurement of the resonator

**4.5 Dynamic characterization 4.5.1 In-plane motion measurement** 

motion measurement.

#### **4.2 Micro pressure sensor measurement**

Experiments were done on a micro pressure sensor using PSI under the air condition. From figure 15(a), a deformed membrane can be seen to show the air pressure. The deform value shows the change of the pressure. From figure 15(b), the height difference between the highest position and the lowest position is 1345.7 nm, which is matched with the designed value.

Fig. 15. Results of micro pressure sensor measurement

#### **4.3 Film structure measurement**

Combining with Otsu method from image segmentation technique, we measured a film thickness standard in WLI measurement (centre average thickness: 1052.2 nm±0.9 nm, refraction index: 1.46, model number: FTS4-10100, VLSI) which was calibrated by an ellipsometer in 632.8 nm wavelength. The system equipped 10× Mirau objective to perform a set of repeated measurements". The thickness and the surface topography were successfully extracted, which were shown in table 2 and figure 16. The *Ra* values of the upper and lower surfaces were 7.30 nm and 7.32 nm, respectively.

Fig. 16. Film thickness standard measurement


Table 2. Measurements on film thickness standard
