Low-noise spectroscopic ellipsometer

Optics: measuring and testing – By dispersed light spectroscopy – Utilizing a spectrometer

Reexamination Certificate

Rate now

  [ 0.00 ] – not rated yet Voters 0   Comments 0

Details

C356S451000, C356S317000, C250S225000, C250S576000, C315S292000

Reexamination Certificate

active

06791684

ABSTRACT:

GENERAL FIELD AND STATE OF THE ART
Background
The principles of a known state of the art ellipsometer are shown in FIG.
1
.
Such an ellipsometer conventionally comprises a light source
1
, a polarizer
2
, an analyzer
3
, and a detector
4
associated with a monochromator
5
.
Those various elements are placed in such a manner that light output by the source
1
passes thorough the polarizer
2
before reaching a sample E to be analyzed, and then after being reflected on the sample E, it passes through the analyzer
3
prior to reaching the detector
4
after passing through the monochromator
5
which is generally a photomultiplier.
One of the two elements constituted by the polarizer
2
and the analyzer
3
is a rotary element.
The output from the detector
4
is connected to processor means
6
which perform Fourier analysis on the modulated signal as measured by the detector
4
in order to determine information relating to the surface state of the sample E.
It is recalled that when light is reflected on a sample E, its polarization is modified and that an ellipsometer setup makes it possible to measure firstly the phase difference &Dgr; and secondly the ratio tan (&PSgr;) between the parallel and perpendicular polarization components of the beam as reflected on the sample.
By means of the monochromator
5
, it is possible to perform measurements at different wavelengths, thereby characterizing the optical properties of the material.
For a general presentation of spectroscopic ellipsometric techniques, reference can advantageously be made to U.S. Pat. No. 5,329,357 (Bernoux et al.) which relates specifically to the advantage of adding optical fibers to the setup.
The visible spectroscopic ellipsometers available on the market generally operate in a spectral range of 1 micrometer (&mgr;m) to 230 nanometers (nm), using a xenon arc source (selected for high radiant flux density or “irradiance”).
Nevertheless, ellipsometers have been proposed that are capable of operating over a broader spectral range than the above-mentioned ellipsometers and that include an additional source, such as a deuterium (D
2
) source that provides less of a point source, emitting in the range 130 nm to 700 nm at a power of 30 watts (W) to a few hundreds of watts or more.
The detectors that are used are generally detectors of the Si or Ge photodiode type or photomultipliers (generally multi-alkali photomultipliers), operating at ambient temperature.
They use very high quality optical systems, possessing polarization extinction coefficients of about 10
−5
, and very high transparency, even in the ultraviolet.
This makes it possible in the above-specified spectral range to determine the &PSgr; and &Dgr; coefficients with precision equal to or less than {fraction (1/1000)}
th
of a degree (°).
Furthermore, the processor means of most ellipsometers implement a simplified photon counting method, which method is known as the “Hadamard method”. That method consists in counting photons with a signal that is amplitude sampled over a very limited number of channels: eight counters or channels, for each period of rotation of the rotary element of the ellipsometer (a configuration with a rotary polarizer or analyzer (modulated polarization) and/or a rotating plate (phase modulation)).
Drawbacks of State of the Art Ellipsometers
Ellipsometers of the type described above present several limitations.
A first limitation is associated directly with fluctuations in the source, i.e. with its lack of stability, with this constraint being known as shot noise limitation (SNL).
Another limitation is associated with noise coming from ambient light and also referred to as “leakage noise”, which can in theory be eliminated by isolating the entire ellipsometer (and not only the photomultiplier) completely from ambient light, but which nevertheless remains a difficulty encountered by many ellipsometer manufacturers.
Another limitation lies in the dark current or intrinsic noise associated with the photomultiplier and its internal amplification system. This noise is commonly referred to as detector noise limitation (DNL). It should be observed that all of the frequencies corresponding to the bandwidth of the photomultiplier are generally present therein.
Thus, the Hadamard sums (as determined over quarter periods of the modulated signal) are calculated by taking account of a previously measured offset which corresponds to the leakage noise and to the DNL.
Nevertheless, although conventional ellipsometers correspond in practice to synchronous detection (in-phase frequency filtering of the signal modulation), the Hadamart method becomes difficult when the amplitude of the modulation is low.
For a signal modulated at &ohgr;, the amplitude of the spectrum component at 2&ohgr; in the signal is of the same order of magnitude as the amplitude of the noise (with this being true more particularly in the ultraviolet where counts of only 100 to 1000 counts per second (cps) are measured).
The signal components are thus “buried” in the noise level which itself corresponds to a superposition of the spectrum density of the source noise, shot noise when using a xenon arc, ambient light, and noise from the detector and its associated electronics.
Furthermore, with conventional ellipsometers, when it is desired to work at wavelengths shorter than 200 nm, the observed signal
oise ratio is unfavorable.
The only known way of eliminating the effects of the various sources of noise is to increase acquisition times.
Unfortunately, measurement is then subject to systematic error, in particular concerning wavelengths shorter than 200 nm. This means that equipment must be pre-calibrated in use.
Furthermore, it should also be observed that another problem encountered with ellipsometers that use additional sources to enlarge their operating range is the problem of their cost and of the power that must be supplied to them.
Under such conditions, it is practically impossible to envisage a system that is sufficiently compact for in situ measurement (integrated metrology) even in a photon-counting system as described above. The impossibility of having a measurement head internal to the metrological casing leads to a limitation due to the windows of the casing giving rise to birefringent effects that need to be corrected.
SUMMARY OF THE INVENTION
An object of the invention is to mitigate those drawbacks.
In particular, the invention provides an ellipsometer structure in which noise is minimized.
Techniques are known, in particular from the abstract of Japanese patent application No. 0907995, that consist in cooling photomultipliers in applications that are very different from ellipsometer applications.
Those cooling techniques are not intended in any way to reduce noise. They serve to lower detection limits as much as possible.
The invention proposes a spectroscopic ellipsometer comprising a light source emitting a light beam, a polarizer placed on the path of the light beam emitted by the light source, a sample support receiving the light beam output from the polarizer, a polarization analyzer for passing the beam reflected by the sample to be analyzed, a detection assembly which receives the beam from the analyzer and which comprises a monochromator and a photodetector, and signal processor means for processing the signal output from said detection assembly, and including counting electronics.
This ellipsometer presents the characteristic of comprising cooling means for keeping the detection assembly at a temperature lower than ambient temperature, in particular at a temperature of about −15° C., or lower.
Also advantageously, its source is a deuterium lamp preferably having a power of about 30 watts.
Other low noise sources can be envisaged, and in particular plasma lamp and halogen lamp sources.
Also advantageously, the counting electronics is suitable for performing amplitude sampling over a number of channels lying in the range 8 (Hadamard equivalent) up to 1024 (filtered Fourier), and particularly preferably about 1000 or more, in partic

LandOfFree

Say what you really think

Search LandOfFree.com for the USA inventors and patents. Rate them and share your experience with other people.

Rating

Low-noise spectroscopic ellipsometer does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Low-noise spectroscopic ellipsometer, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Low-noise spectroscopic ellipsometer will most certainly appreciate the feedback.

Rate now

     

Profile ID: LFUS-PAI-O-3226015

  Search
All data on this website is collected from public sources. Our data reflects the most accurate information available at the time of publication.