Power Density Convertion to Light Intensity

ccd = charge coupled device

mW = miliwatts is a measurement of energy

nm = nanometer a measurement of length

http://www.horiba.com/scientific/products/raman-spectroscopy/raman-academy/raman-faqs/what-is-a-ccd-detector/

https://en.wikipedia.org/wiki/Wavelength

https://en.wikipedia.org/wiki/Power_%28physics%29

What is your ultimate goal?

Not that I know of....

...."This unit is the "spectral irradiance", explained as "radiative power per surface area and per wavelength interval".

There is no way to convert "counts" into this unit."...

https://en.wikipedia.org/wiki/Irradiance

https://www3.nd.edu/~sst/teaching/AME60634/lectures/AME60634_F13_lecture27.pdf

https://en.wikipedia.org/wiki/Radiative_transfer

https://www.researchgate.net/post/How_to_convert_counts_sec_to_photon_flux_or_A_W

Here's my understanding of what you're asking. Your lamp emits light with a spectrum that has energy in each nm wavelength defined by the theoretical spectrum curve. A portion of this energy passes through a slit to a grating and is focused on a linear CCD array.

You don't know this portion, so you don't know how many milliwatts of power hits the CCD array. So you cannot know the absolute response of the CCD. If your CCD is calibrated in wavelength (so you know the light wavelength for each cell), you should be able to determine the relative response curve of the detector by dividing the energy from each cell of the CCD with the corresponding value on the lamp spectrum curve.

There are a lot of unknowns here, so I couldn't give you an answer without more info.

Does the spectral sensitivity of the CCD roughly match the lamp spectrum? This is important because if the lamp emits mainly in the visible spectrum (400 to 700 nm) but the CCD is sensitive in the near IR (800 to 2000 nm) then you're not going to get useful data.

Are you using a grating or prism to separate the wavelengths of light from the lamp? Without a way to create a spectrum from the lamp you can only get a relative power response from the CCD (using it like a camera light meter), but to get spectral response you'll need a way to look at the lamp wavelength by wavelength. Typical lab spectrometers measure at 1 to 10 nm increments.

Do you have a wavelength calibration source? To calibrate a CCD, which you need to do to quantify the spectral response, you need a standard spectral source emitting narrow, well-known spectral lines. Without that, you don't know what part of the spectrum a given reading corresponds to. Typically a mercury-argon lamp is used since it emits spectral lines from the UV to the IR.

A second calibration is also needed: Intensity calibration. Since you know the spectrum of your lamp, it can be your intensity source to let you calibrate your CCD -- which is what you are asking. You will need to measure the lamp for a set length of time and then use 'counts', not 'counts per second' as your measure of spectral intensity at each wavelength band.

Here's an astronomy-related website for measuring spectra of lamps, stars, planets, etc. Perhaps you can get more guidance here.

https://www.rspec-astro.com/

You are calibrating the output of your CCD against the input. You need to use a known source to take measurements across the spectrum. Once you have the output of your CCD, you create a calibration table that will correlate energy in vs CCD counts out.

For a known source, deuterium arc lamps are a little heavy on the UV end with lower energy on the red end of visible. There also may be calibrated LED sources that give a better visible light output.

Keep in mind that you are calibrating the system. This will include the slit, the mirrors, the grating and the detector. If you use a fiber to bring the light into the system, you will be calibrating the system with that fiber and it will only be valid with that particular fiber. Changing it out will invalidate the calibration.

Also keep in mind that CCD's are integrating detectors. They integrate all energy that comes in over the scan time of the CCD. If you change that scan or integration time, that will affect the calibration as well.

I suggest that you read up on CCD detectors, spectrometers and spectrometer calibration. It's a complicated area, but it's really interesting. Ocean Optics (no implied endorsement) is a spectrometer manufacturer. Take a look at their website (www.oceanoptics.com) for their white papers and other knowledge base informational pages. Look around on the web, there is a lot of info out there.