During tantalum oxide deposition I use single-wavelength optical control. I have two control points on the substrate. For this, I use an optical monitor at a single wavelength. In real-time it is necessary to calculate the difference between the two measurement points with high accuracy. how can you compare two points having data from single-wave control?
on the vertical axis - transmission %,
deposition time on the horizontal axis
Hello Petr Rozel ,
How far apart are the two control points on the substrate on which you are depositing the tantalum oxide? Do you have one single-wavelength monitoring system or two such systems? Or are you trying to monitor the two spatially separated control points with one single-wavelength monitoring system? I am asking because I am confused by the statement of your problem.
Regards,
Tom Cuff
Hello Thomas Cuff,
distance between points 20 mm. I have one control system with two channels that operate at the same wavelength. I need to compare thickness in two-point in real-time.
Best regards,
Peter
Hello Petr Rozel ,
So is the real problem that the control system cannot look at the two channels at the same time? Can you electronically switch between the two channels, and if so how fast can you do the electronic switching? Or is electronic switching not available, and the switching has to be done manually?
Regards,
Tom Cuff
Thomas Cuff
the system measures two points at the same time and gives a change in transmission for two points in real-time. the question is how to calculate the film thickness with high accuracy of 0.01%
Hi Petr Rozel ,
If you are trying to measure the thickness of an oxide film grown on tantalum, I am not sure how you are getting the transmission shown in your figure. The incident light beam would be reflected back from the tantalum surface after it had passed through the oxide film a minimum of two times so the thing you are measuring is reflectance, is it not? And depending on the angle of incidence of the momochromatic light there may be multiple reflections back to the detector. Is the monochromatic light normally incident, i.e., perpendicular to the surface, or is it at some other angle.
The figure you presented, does that represent one channel or the difference between two channels?
There are formulas for film thickness for reflected and transmitted monochromtic light, but they require knowing the three indexes of refraction, see the book by Heavens.
[1] O. S. Heavens; Optical Properties of Thin Solid Films; Dover Publications, Inc.; 1991; pp. 55-59.
Regards,
Tom
Hello Petr Rozel ,
Sorry, I should have looked at your RG page. You are depositing tantalum oxide on a transparent substrate by sputtering, chemical vapor deposition, etc., which is why you are presenting the transmission data in your figure. I am guessing that the figure in your question represents a single monochromatic channel's transmission versus deposition time plot, and that you are trying to control or measure the differential thickness to an accuracy of 0.01%. Do I finally have the corect picture in my mind? The noise in your plot looks like it might be problem, but it is hard to tell. Is this true? I am assuming that you are using a chopped light beam in each channel for stability. Is that true?
Regards,
Tom Cuff
Hi Thomas Cuff,
I know how to calculate the optical and physical magnitude of optical materials. The task is how to correctly and accurately compare two measurement points during the deposition process (PVD), having data from a single-wave control. a laser is used as a light source.
noise level is 0,02%. it's no problem for the measure.
Hello Petr Rozel ,
So you are looking for a signal processing method that can calculate the difference between two transmission curves in realtime? And since you are also using this difference signal to control the shut off of the deposition process, the signal processing method must also be able to predict one time step ahead in the presence of noise? In other words, the signal processing method must be able to predict (extraplolate) the future value of the difference in the transmission curves.
Regards,
Tom Cuff
Hello Thomas Cuff,
Yes, you are absolutely right. Now my system can predict a breakpoint, but cannot compare both signals
Best regards,
Petr
Hello Petr Rozel ,
I am confused. If you already have a signal processing method that works for one channel, then what prevents you from applying the same method to both channels or to the difference between both channels? There is something you are not telling us. I can understand your reluctance to reveal company technical information, but it will be very difficult for any reader of your question to help you without knowing more details than you have already provided. We are not mind readers, and we cannot look over your shoulder to see exactly what you are doing and how you are doing it.
Regards,
Tom Cuff
Hello Thomas Cuff,
I am using a turning point method to define a breakpoint. This method does not provide data on layer thickness. this is the complexity of the possibility of comparison.
I think that it is necessary to use some other method that can give accurate data on the thickness of the layer
Regards,
Petr
Hello Petr Rozel ,
So you are looking at the periodic variations (minimums and maximums, i.e., turning points) in transmissivity in terms of the quarter wavelength thickness of the dielectric layer. Are you looking at integer multiples of the quarter wavelength thickness or a fraction of one quarter wavelength? In other words, do you count the number of oscillations of the transmissivity curve or just try to estimate the first minimum as shown in your figure?
I can see why you are having difficulties calculating the actual thickness of the dielectric layer from a monochromatic light source measurement due to the state of aggregation of the dielectric layer and its heterogeneity as the layer is built up, and that this problem only gets worse with two points on the surface of the substrate. In addition, I am guessing that the thickness measurement must be done in realtime, i.e., that your deposition process is not stable enough to allow for offline determination of the thickness.
Do I finally understand your problem, or is there something else I need to know?
Regards,
Tom Cuff
- Badges
- Science topic
- Similar topics
- Agricultural Science
- Agronomy
- Crop Science