CephC-LABS Question log

(Questions for march 15)

(the next week should always be on top)

Questions for March 8

LMR example question: Where did you guys find the graphic that shows the entire Ceph OB3 region? LMR example answer: In the Sargeant 1977 article. That's the discovery paper that named Ceph C, Ceph C. It's in the Box drive, in the papers directory, in the "maybe less relevant" subdirectory.

From Sam: Here are my questions...sort of a flurry all related to what the data we're going to be analyzing means...

So if a picture is just an array or grid of numbers that get assigned colors...What do those numbers represent or mean? Number of photons collected? The total energy of the photons collected during the collection window? Can multiple sources in the field of view send photons to a single pixel? How do we know which source sent the photons to each pixel?

I can learn more about HOW each pixel turns the collected photons into a number later, but I am just curious about WHAT the pixel does with the collected photons and what it outputs. In other words, what does the raw data we will be working with represent? What do the FITS file numbers mean? I imagine that a pixel collecting 5 high energy photons could feasibly give the same 'output value' as a pixel that collects 10 lower energy photons...assuming it is able to collect and count photons of different energies and filters weren't involved (even though they usually are). The material in the CCD probably only responds in certain ways to certain photon energies...hmm.

Response from Tom: The photons that are being collected have passed through a filter so they are of approximately the same energy-- the ones with too little or too much energy (they are the wrong "color") don't make it through the filter and so aren't detected.  This filter may be a physical filter in front of the detector or it could be a limit of the detector itself.

Response from Luisa (still stuck on iPhone): What the numebrs mean depends on how you detected the light and how you process the data before you display it. Spitzer and some Herschel data in MJy/sr.  Yes, multiple sources can send photons to the same pixel, and the only way we can distinguish them is to bring information from elsewhere into the thought process. re: raw data -- what I mean by raw data may not be what you mean. We aren't working with raw data. What the numbers mean depends on which instrument from what telescope, and what processing the image has undergone.  WISE data still in counts (close to raw) but calibration stuff in FITS header.  Spitzer definitely not raw; it's calibrated (processed such that absolute flux density in FITS image). That is in MJy/sr.  Herschel data also definitely not raw. From what I have read, some in MJy/sr, some in MJy/pix. I think. More when I get back to the office. The FITS files numbers all correspond to brightness in some units. Sometimes it is in counts (electrons), sometimes someone has converted those counts into absolute brightness in some other units. Exactly what it means (and how to convert it into units that we can use in SEDs and color-color/color-mag diagrams) depends on the instrument/telescope/processing. Tom is exactly right that filters impose a limit on what photons make it to the detector, AND the detector physics itself also cuts off some photons - some detectors aren't sensitive to any incident light beyond a certain frequency.

From Olivia: I would like Tom to share with us the logistics of bringing students into the project. What kind of information did you initially share with them? Was it all done during your Research class? 

Related to Sam's train of thought: energy per photon of IR is going to have less energy than photons of visible light, UV or X-rays - does this matter?

Response from Luisa (stuck on iPhone): energy of photon given by h*nu where h is Planck's constant and nu is frequency of light. So yes, bluer=more energy. If you want to make an X-ray detector you have to think (design, etc) differently than for FIR detector. This is why light detected very differently across EM spectrum.  

From Sam: I guess one of the next steps once I figure out what the pixels output and how they collect the photons is to be able to take that info and convert it into different units so that we can construct SEDs and other diagrams. I've worked on trying to understand some of the "Units" page on the wiki...http://coolwiki.ipac.caltech.edu/index.php/Units I got a little stumped about halfway down in the SEDs section...Could you elaborate on the equations in the big gray box that say "dF/dlambda = dF/dlambda*(dv/dv)...etc." and why you had to do that to do the conversion from Fv to Flambda? Others: In what situations is flux density is preferable to simply flux? For 'flux density'... why not divide J/s/m^2 by the photon energy rather than the frequency? Like this J/s/m^2/(J/photon) = photons/s/m^2 . This would tell you how many photons were hitting an area each second...right? aka..how "densely packed" in the photons were! I guess frequency and energy are proportional so it probably doesn't matter.

Response from Luisa: Just in explicit answer to your questions (but really, we will go over the complicated units in detail in July, so don’t stress if none of this makes sense). The sequence is (and I think you’re missing the ones with the *) : light hits detector, detector records information, *pipeline runs to convert the raw data into something physically reasonable and gives it to you as a FITS file, *make quantitative measures of the flux (e.g., do photometry), convert mags to flux densities (and flux densities to mags) as needed, add measurement to SEDs and color-color/color-mag diagrams. For "why lambda", look a bit down the page for "notes on plotting." For the derivation of this, with all the gory math, I can scan for you pages out of Rybicki & Lightman, “Radiative Processes in Astrophysics” but be aware this is a graduate textbook… I’m not expecting you to derive this. The "per Hertz" or "per length" in the energy density is really "per photon frequency" or "per photon wavelength." So it’s tightly tied to the photons. Flux is energy per unit area per unit time (flux per sec per square meter). Flux density is energy per unit area per unit time per photon. You need energy density to plot for a spectral energy distribution (otherwise it wouldn’t be an s*E*d). Energy density is lambdaFlambda or nuFnu