FFT analysis of a spectrometer signal

[KevinKM]

Hi all,

I am acquiring spectrometer (Ocean optics hardware) signal via USB data acquisition into LabVIEW.

The signal is

Wavelength(nm)        Intensity

xyz                              xyz

xyz                              xyz

xyz                              xyz

xyz                              xyz

The set up is as follows. The light is incedent on the PET film which has a thin coating of dielectric material using Ocean optics spectrometer.

Now i need to analyse this signal to find the thickness of the dielectric material and frequency components using FFT.

Excuse me if i am not clear about anything.

[Norbert_B]

Are you talking about this?

Norbert

[ejkaiser]

Ocean Optics http://www.oceanoptics.com/products/nanocalc.asp offer this kind of software for their systems. I doubt if someone would just provide something like this for free. At least I wouldn't do it if I had such software.

You can try to research about the algorithms that are required (this is not LabView specific) and make something by yourself in LabView. It can certainly be done in LabView. However, in most cases it doesn't pay to make own software if it is commercially available.

Cheers

Edgar

[KevinKM]

Hi Norbert_B and ejkaiser,

Thanks for your reply.

My application is more or less the same. I am using Ocean Optics hardware to get the raw data(Intensity v/s Wavelength) into LabVIEW. I need to find the thickness of the ceramic film and the PET film. I am working on the code which has been previously developed by a Senior Developer. I am able to get the thickness of both the films using the FFT Analysis. The problem here is the resolution is little high. I need the resolution in range of 0.1nm.

Below is the data obtained after FFT Analysis and Mathematical computations to find the Thickness of both the films. Excuse me if i am not clear about anything.

I have shown a part of the data below

Wavelength                            Amplitude

150.2n	15.0889
300.4n	22.9708
450.7n	11.9293
600.9n	13.9295
751.1n	37.245
901.3n	81.2098
1.052u	122.347
1.202u	111.464
1.352u	48.7365
1.502u	27.3915
1.652u	19.6806

[KevinKM]

Due to some reasons, i am not able to share the code with you ppl. Apalogies on that.

[ejkaiser]

So, seems to be simple: you will need to go ahead refining your code. Good luck!

[KevinKM]

Hi ejkaiser,

Thanks for your reply.

Simple things are hard to figure out sometimes. I have been to trying to figure it out with no luck, can you please eloborate on how to go about it. Is there any specific method to increase the resolution?

I tried using this formula. Is this the right approach?

t=λ /n x 4

[ejkaiser]

It is not possible to increase the useful resolution of a given spectral dataset. You can of course fit in interpolated data between the measurements but this doesn't contain useful information. It will result in fake signals.

I have no idea what you expect from the formula.

[KevinKM]

Hi ejkaiser,

Sorry i was little bit confused with the data set i posted before. let me make it clear. 

The spectral data set (Raw Data) is as shown below.

The thickness data is shown below.  I need to increase the resolution of the thickness data. Excuse me if i am not clear about anything.

Raw data from Spectrometer

Wavelength (nm)	Intensity
344.16	1257.45
344.381	1257.45
344.603	1257.45
344.824	1273.84
345.045	1232.88
345.267	1671.14
345.488	1605.61
345.709	1634.28
continues uptil
1041.76	1693.67
1041.91	1621.99
1042.07	1638.37
1042.23	1716.2
1042.39	1708.01
Output after FFT Thichness (X-Axis) 150.2n Amplitude (Y-Axis) 15.0889 300.4n 22.9708 450.7n 11.9293 600.9n 13.9295 751.1n 37.245 901.3n 81.2098 1.052u 122.347 1.202u 111.464 1.352u 48.7365 1.502u 27.3915 1.652u value continues 19.6806

[johnsold]

Are you taking the FFT of the intensity values or the wavelength values?

You provided a relationship between thickness and wavelength. What is the formula for the amplitude of your outputs?  What does that amplitude represent physically? How does the thickness formula relate to the FFT?

Since you already have a spectrum, it is not clear to me why you need to take another FFT.  I am having trouble getting my mind around the concept of a spectrum of a spectrum and what that might mean.

In general to increase the resolution of the FFT you need to increase the size of the input data set.

Lynn