Page Navigation

Rudolph AutoEL II Ellipsometer Users Manual

I. Capabilities of the Ellipsometer:

The ellipsometer can be used to find refractive index and thickness of a single thin, isotropic, homogeneous, non-absorbing film on a reflective substrate. The ellipsometer measures the change in state of polarized light upon reflection from a surface. The state of polarization is determined by the amplitude ratio of the parallel (p) and perpendicular (s) components of radiation, and the phase-shift difference between the two components. The fundamental equation of ellipsometry relates these parameters to a complex amplitude reflection ratio =_p/_s=tan()exp (i), where the reflection coeff icients rp and rs are the ratio of the reflected to incident electric field components for the parallel and perpendicular directions, respectively. The ellipsometer does not directly measure the index of refraction (n) or the thickness (t) of the layer of interest. A computer program must be used to model the sample geometry and find a solution of n and t that is consistent with the measured and values. The ellipsometer is capable of determining the and of many complex structures. However, our current software can only analyze the simple case of a single non-absorbing film on a reflective substrate. Therefore, not only must care be taken in making an accurate and measurement, but the sample being measured must have a close correspondence to the model being used in order to obtain realistic results. It is up to the user to check the results for physical validity.

For best results, do not use patterned samples. If a patterned sample must be used, an unpatterned area large enough to contain the ellipsometer's light spot must be available. The light beam's footprint on a sample is approximately 3 mm x 1.5 mm. A card or piece of paper may be used to center the beam on the feature to be characterized. Rough or non-uniform films will also affect the measured and values and could produce inaccurate results.

II. Ellipsometry Basics:

A schematic of the ellipsometer is given in Figure 1. The main components of the ellipsometer and their function are described below.

1. White Light Source - This light source puts out unpolarized, incoherent white light which contains multiple wavelengths of light.

2. Polarizer - After passing through the polarizer, the white light is linearly polarized. This linear polarization of the electric field may be decomposed into two orthogonal components, p and s.

3. Compensator - This component will change the phase of the p polarization by 90° with respect to the s polarized light, which produces circularly polarized light at the output of the compensator.

Note: The compensator gives a 90° phase shift between the polarizations at only one wavelength. Therefore, depending on the wavelength desired for measurement, the compensator knob will need to be adjusted to the correct colored dot as described by the configuration chart on the ellipsometer. This configuration chart is shown in Table 1.

4. Sample - As the circularly polarized reflects off the sample, the p and s components of the electric field will experience different amplitude and phase changes. In general the light reflected off the sample will be elliptically polarized.

5. Analyzer - This component is another linear polarizer used to separate the elliptically polarized light from the sample into its p and s components.

6. Filter - A filter is used to select the wavelength of interest from the white light continuum. A set of three bandpass filters enable the user to take measurements at = 405 nm, 632.8 nm, and 830 nm.

Note:

The filter works to pass only one wavelength for each measurement. If a different wavelength is needed for a measurement, then the filter will need to be adjusted to the correct colored dots on the filter knob as described by the configuration chart on the ellipsometer.

7. PMT (Photo Multiplier Tube) - This device is a sensitive light detector. Incident photons produce an electrical signal that is proportional to the number of photons. The gain of the detector can be varied by changing the bias voltage.

Note: The sensitivity of the PMT changes with wavelength. Therefore, the bias voltage applied to the PMT needs to be adjusted for each wavelength in order to equalize the signal level. The configuration chart on the ellipsometer specifies the bias voltage that should be used for a given wavelength.

III. Operating Procedure

A. Staring Up the System

This section outlines how to set up the system so that it is ready for calibration.

Turn on the ellipsometer using the key on lower left of the instrument. Also, make sure that the computer and monitor are on.
The ellipsometry software will automatically boot up on the computer. Program No. 1 is used by the ellipsometer (the others are inactive).
The cursor seen on the computer monitor can be moved around the screen by using space and backspace on the keyboard. As the cursor is moved to a new item on the monitor, a comment line at the bottom of the screen briefly describes each item.
A list of allowed commands can be seen at the bottom of the screen. A command is entered by typing the first letter of the command on the keyboard. Not all the available commands are displayed on one screen. To see the rest of the possible commands, type "r" for rest. (All entered commands in this paper will be in quotation marks. They are not part of the command, so don't type them).
Move the cursor to the filename on the screen, type "m" for modify and enter the filename of the program that you will use. For the present time, enter the filename "633NM," since this file will eventually be used to check the ellipsometer calibration. The filenames that may be entered are the available wavelengths of the ellipsometer: "405NM," "633NM," or "830NM." Then type "f" for file and "l" to load the necessary parameters used in the measurement. The default model assumes a single thin film on a silicon substrate. The silicon substrate's refractive index ns and extinction coefficient ks are displayed on the screen for the wavelength being tested. Compare the loaded ns and ks values for the silicon substrate with those listed in the configuration chart on the ellipsometer to be sure that the values were loaded correctly. Note: Degenerately doped silicon will have different optical indices than those for lightly doped silicon (which we use). Therefore, if a lightly doped silicon substrate is not used, the user must determine the proper ns and ks values for the substrate material at the wavelength being tested, and manually enter them.
For best results, let the computer iterate for a fixed n and a variable t in order to match the measured and values. To implement this, the computer should display the film refractive index nu as nu * 1.462 i. The letter "i" at the end means that n will be fixed during the iteration. If the "i" is not present, simply move the cursor to nu and enter the letter "i." The value entered for nu (e.g. 1.462) is just the starting point for the iteration. The program will converge on the actual value during the iteration.
Although it is usually not necessary, the iteration limits on n and t may also be set to help the software converge to a solution. In order to bound the thickness or refractive index solution, move the cursor to item Tu and/or XX. Type "m" and enter "y" for the parameter to be bound. These items will then appear as Tu= __ to __ or XX = __ to __. The iteration limits are set typing "m" for modify and the numerical values of the limits. If the actual n and t are not within the bound specified by the user then the software will not be able to converge to a solution, and the results will be displayed as *****.
The white light source needs to warm up for at least 10 minutes before continuing on to the initialization step. Make sure that the source is in the "RUN" mode for taking measurements.

B. Ellipsometer Initialization

This section lists the procedure necessary to initialize the ellipsometer. This is primarily an internal self-calibration which the ellipsometer runs in order to set its mechanical and electrical components.

Carefully place the Rudolph SiO₂ calibration wafer from its case and place it in the sample holder so that the light spot is centered on the rectangle in the middle of the wafer. This procedure is aided by wedging the wafer against the stops at the top right quadrant of the sample holder, and setting the red lines on the translation stages to zero.
Adjust the tilt of the sample holder to level the sample. This is accomplished by looking through the eyepiece and adjusting the three tilt adjusters at the bottom of the sample holder until the bright dot seen in the eyepiece is centered in the crosshairs. All adjusters should feel snug when the tilt alignment is complete. But be careful not to turn the adjusters too tightly.
Note: The lateral tilt adjusters work by moving opposing wedges. If one of the tilt adjusters becomes hard to move, then back off the opposing adjuster in order to move the initial adjuster further.
Looking at the entrance of the analyzer aperture, determine if the reflected light beam is centered on the aperture. If not, it will be necessary to adjust the height of the sample holder. This task is accomplished by first loosening the hexagonal screw on the bottom of the sample holder with the red handled allen wrench. Then adjust the height of the substrate holder with the big dial just below the sample holder until the reflected light beam is centered on the analyzer aperture. Re-tighten the hexagonal screw. Re-check the tilt alignment.
The eyepiece may gently be pulled toward the user (about 0.5 inches) to bring in line a microscope objective that can be used to visually inspect the wafer surface to be measured.
Follow the configuration chart on the ellipsometer to set up the instrument for initialization. This entails pulling the knob on the compensator out until the white dot is showing and lining up the silver ring so that it is just outside the compensator box.
Note: If the silver ring is not consistently lined up just outside the compensator box for each wavelength, then the measurements may differ each time.
Pull the knob on the filter out until the red dot is showing and line up the silver ring so that it is just outside the filter box. The same precautions apply as in step 1 in regard to lining up the ring.
Set the voltage for the PMT to 200 using the dial readout next to the PMT.
Press "CONT" on the ellipsometer to start the ellipsometer's initialization program. When the blue LED display shows the message "Down Comp - . . . ," the ellipsometer is calibrated.

C. Verifying Ellipsometer Calibration

Once the ellipsometer has been calibrated, its accuracy needs to be verified. This procedure utilizes a known standard against which the ellipsometer results are compared.

Carefully place the Rudolph SiO₂ calibration wafer on the sample holder and check the tilt and height of the wafer, as described in the previous section.
Note: The n and t values of the SiO₂ film at = 632.8 nm are printed on the wafer carrier.
Use the microscope option on the eyepiece to inspect the surface of the wafer for any dust or surface imperfections which may affect the measurement.
Follow the configuration chart on the ellipsometer to set the instrument for a measurement at = 632.8 nm. This involves pulling the knob on the compensator out until the red dot is showing and lining up the silver ring so that it is just outside the compensator box.
Then pull the knob on the filter out until the red dot is showing and line up the silver ring so that it is just outside the filter box.
Set the voltage for the PMT to 175 using the dial readout next to the PMT.
Set up the computer to interpret the measured data.
- Load the 633NM program and its silicon substrate values as explained previously in Section III A.
- Make sure that the computer will iterate a solution from a fixed n, as explained in Section III A.
- Check that the limits on n and t will allow the software to converge to a solution, as explained in Section IIIA.
- Check that angle of incidence is 70°. Do not attempt to change the angle of incidence of the ellipsometer!
- Verify that lambda is set to 6328.0 (wavelength entered in Angstroms). If not, type "m" and enter "6328."
- Set Print Results to yes if you want a printout of the results. Type "m" and then "y" to get a printout.
- Verify that the iteration limit is 0.001. At this value, the program will iterate on n until the results differ by no more than within 0.001. The program is said to have converged to a solution at this point.
- Set the number of orders to print. The thickness of the film can be determined to within an integer half wavelength (i.e. /2 + (m × 2), m=0,1,2,... ). Each of these solutions for t is equally possible. The only difference between them is an integer number of 2 phase shifts through the film given by the cycle thickness in the printed results.
Type "g" on the keyboard. This tells the computer to go and check for and values from the ellipsometer on the interface bus.
Press "CONT" on the ellipsometer to start a measurement of and . After approximately thirty seconds, the ellipsometer will print the measured and values on the blue LED display and send the data to the computer.
On the computer monitor, the fits for n and the multiple orders of possible t values will be shown. Check that the first order for tu and nu match the printed numbers on the carrier of the calibration wafer to within the specified tolerances. If the values match, then the calibration is complete. If the n and t are out of spec, consult the trouble shooting procedure in section IV. The user can return back to the model parameter screen by pressing the spacebar.
Note: After the ellipsometer is calibrated, the results should be valid for approximately 2 hours. If the ellipsometer sits for longer than this time, it should be re-calibrated.
After a successful calibration, the Rudolph calibration wafer should be carefully removed and placed upside down in its wafer carrier to prevent damage to the wafer.

D. Measuring Samples

This section describes the procedure used to measure samples which have a film with an unknown refractive index and or thickness. However, the user will usually have a rough idea of what the film's n and t values may be. This knowledge can be helpful in narrowing down the possible solutions or determining the validity of the result.

Mount the sample on the sample holder. A piece of cleanroom paper may be used to align the incident light beam to the desired location on the sample.
Check the sample alignment and inspect the surface as described in section III B 2-4.
Consult the configuration chart on the ellipsometer to determine correct detector bias voltage, compensator knob position, and filter knob position for the wavelength to be used in the measurement.
Note: The choice of which wavelength should be used for the measurement is somewhat dependent on the film to be measured. The shorter wavelengths can have better resolution, but may lie closer to absorption bands and may be too sensitive to film irregularities. The longer wavelengths attributes will be just the opposite. Usually an initial measurement at = 632.8 nm is a good compromise.
If the sample is to be measured at = 632.8 nm, then follow the same procedure as outlined in section III C 6. If a different wavelength is to be used, then enter the appropriate filename ("405NM" or "830NM") in lieu of "633NM." When the file has been loaded, confirm that the model parameters (e.g., ns and ks) are correct by comparing with the configuration chart.
Once the computer model is set up and the ellipsometer is configured, press "CONT" on the ellipsometer to start the measurement, and type "g" on the keyboard to ready the computer for receiving data from the ellipsometer.
It is recommended that the user repeat the measurement several times to determine the repeatability of the results. This can be accomplished by simply pressing "CONT" at the end of each measurement while the computer is left in the ready mode. The user can return back to the model parameter screen by pressing the spacebar.
After the ellipsometer has completed a measurement, the computer will solve for the refractive index and give the multiple orders of possible thickness values of the film. If the user has a rough idea of what the film thickness should be (i.e. from the film color, deposition rate, or Dektak measurement), then the correct thickness order can be easily identified.
If the film thickness is unknown, then another measurement will need to be make at a different wavelength. Although the film will have a different refractive index at this second wavelength, its thickness will be the same. Therefore, measurements taken at two or more wavelengths should have a single thickness value in common. Note: If no thickness correlation is found, then the film may be thicker than the printed thickness orders. The user can check this by specifying higher thickness orders for "tu" in the computer model.

E. Shutting Down the System

When all the necessary measurements on the ellipsometer are completed, or if the ellipsometer will be left idle for many hours, the system should be shut down. Turn the ellipsometer off, but leave the computer running with the screen turned down.

IV. Helpful Hints and Troubleshooting

A. Making a Successful Measurement

The following is a list of items which the user should consider in order to obtain good results.

Make measurements with the lights turned off to help reduce noise in your measurements.
Do not touch the table while the ellipsometer is collecting data.
Check for good SNR (signal to noise ratio).while the ellipsometer collects data. The analog voltage meter on the ellipsometer should swing up to 5-7 volts while collecting data. If the meter reads 10 volts then it is saturated.
Try and be consistent where you put the compensator and filter knob settings. If they are in different spots each time, the results may not be consistent.
Check that the measured value for is greater than 10°. If not, there will be significant fluctuations in the refractive index and thickness results. When this situation is encountered, it is best to take a measurement at a different wavelength.
Take multiple measurements at a given wavelength to determine the repeatability of the results.
If possible, measure the film thickness with the Dektak profilometer and correlate data with the ellipsometer's result.
If the computer cannot converge on a solution, then change the iteration limits for "nu" and "tu" until it does.
Make sure that the Rudolph calibration wafer and the user's sample are aligned before taking measurement.
Measure only thin, isotropic, homogeneous, non-absorbing films on reflective substrates.
If you are using a substrate other than silicon (i.e. degenerately doped silicon or GaAs), then the optical constants of the substrate need to be determined and entered in the computer model.

B. Troubleshooting a Bad Calibration

If the results of the Rudolph calibration wafer measurement are not within the specifications listed on the wafer carrier, check the following items.

Re-measure the calibration wafer several times to determine its repeatability.
Check that the measured value is greater than 10°. If not, then check to see that the settings (i.e. compensator knob, filter knob, or PMT voltage) are correct.
Make sure the ellipsometer has good SNR during the measurement. This can be verified by ensuring that the analog voltage meter on top of the ellipsometer swings up to 5-7 volts during the measurement. If not, then check for incorrect settings on the ellipsometer and that the light source is on.
If the calibration is still bad after checking items 1-3, then re-initialize the system and repeat the calibration verification measurement. The system can be re-initialed by using the red "RESET" button in the depressed cabinet next to the "HALT" button on the ellipsometer. See figure 1.

Figure 1: Diagram of Rudolph AutoEL II Ellipsometer.

Table 1: Configuration chart used for the Rudolph AutoEL II Ellipsometer.

Wavelength:	405.0 nm	632.8 nm	830.0 nm	Initialization
Compensator/Filter:	Blue	Red	Brown	White/Red
Voltage Dial:	400	175	325	200
Silicon Index: (Ns, Ks)	5.42, 0.329	3.858, 0.018	3.672, 0.005
Program File:	"405NM"	"633NM"	"830NM"

Northern California Nanotechnology Center