file: fluffium_notes.txt Make Fluffium addition to the contamination files. -------------------------------------------------- 12/17/03 -dd Summary: Take the current table files that define the ACIS OBF contamination (from Herman Marshall) and create new versions with: - Fluffium component added to the 'abs file - Total tau at ~0.7keV agrees with ECS - Time dependance based on Tennant ECS tau fit Deleiver these new table files to Dave Huenemoerder for creation of an 'N0003 contamination file. - - - Start with the two files given to Dave H by Herman (used to make the 'N0001 and 'N0002 contamination files; relevant portions of the files are included and annotated here): --------------------------------- acis_contam_abs_v1.tbl ----------- # Created by HLM on 12/23/02 # Data set was a TOO observation of Mk 421 # Date of observation was 10/26/02 (2002.82) # Transmission increases linearly from # none at 1998.75 to the tabulated values at 2002.82 # Transmissions due to each component # wavelength C-K O-K F-K 1.01250 0.00000 0.00000 0.00000 1.03750 0.00000 0.00000 0.00000 etc. 0.706 keV = 17.562 A: 17.5625 0.241522 0.0499158 0.0888599 etc. 0.683 keV = 18.153 A: 18.1625 0.261969 0.0539561 0.00315442 etc. 0.618 keV = 20.062 A: 20.0625 0.336250 0.0692934 0.00439601 etc. Deepest part of Carbon edge and its recovery below the edge: 43.0375 2.92814 0.0243070 0.0364817 43.0625 2.69728 0.0243425 0.0365337 43.0875 2.29158 0.0243781 0.0365857 43.1125 1.95310 0.0244135 0.0366376 43.1375 1.66263 0.0244490 0.0366896 43.1625 1.40830 0.0244845 0.0367416 43.1875 1.18203 0.0245201 0.0367936 43.2125 0.978268 0.0245556 0.0368457 43.2375 0.792970 0.0245911 0.0368981 43.2625 0.623016 0.0246267 0.0369507 43.2875 0.466103 0.0246622 0.0370033 43.3125 0.358841 0.0246978 0.0370560 43.3375 0.358841 0.0247333 0.0371087 43.3625 0.358841 0.0247690 0.0371614 43.3875 0.358841 0.0248046 0.0372141 43.4125 0.358841 0.0248405 0.0372668 43.4375 0.317557 0.0248765 0.0373195 43.4625 0.221848 0.0249124 0.0373723 43.4875 0.130423 0.0249484 0.0374251 Below the edge: 43.5125 0.0748720 0.0249844 0.0374778 43.5375 0.0750216 0.0250203 0.0375307 43.5625 0.0751714 0.0250563 0.0375835 etc. 56.3375 0.165903 0.0487016 0.0700750 ----- -------------------------------- acis_contam_tau_t_v2_2006.tbl ----------- # Author: Herman Marshall # Date: Sat Feb 1 10:45:14 2003 # Usage: gives multiplicative factors to use when scaling # the optical depths for each element in the ACIS # contaminant. The wavelength dependence of the # elemental opacities is given in a separate file. # # Date: 2003.11.06 dph added last line c2006 for isis bug work-around # (interpol()'s extrapolation was broken) # # time relative_thickness # C-K O-K F-K 1999.700 0.0000 0.0000 0.0000 1999.705 0.0110 0.0124 0.0102 etc. 2002.820 0.9846 1.1099 0.9157 etc. 2004.695 1.3945 1.5719 1.2968 2006.000 1.6798 1.8935 1.5621 ----- The Two-Level ("Fluffy") concept/model is described on: http://space.mit.edu/HETG/technotes/contam/twolevel.html and the ISIS routine available there is run to create an output list of the Fluffium transmission vs wavelength at ~2003.0 time period: isis> .load contam_twolevel.031202.i etc. Fluffium transmission (20% thick model) at t = 2002.97: 1.01 1.0000 2.21 1.0000 etc. (see acis_contam_fluf20thk.tbl for other values) 35.70 0.8885 36.90 0.8912 38.10 0.8957 39.29 0.9037 40.49 0.9173 41.69 0.9376 42.88 0.9627 ------------------ The previous three points have the largest changes in transmission point-to-point but even so the change is 2.5% in transmission (at t=2002.82) at most, so simple linear interpolation should be fine given all the other edge structure in this range. - The lower limit of the Carbon edge is here at ~ 43.50 A The values in this list at 42.88A and 44.08A have very similar transmission: IDL> print, exp(0.9627-0.9681) 0.994615 (0.5%) and so interpolation here will add no extra structure at the edge. ------------------ 44.08 0.9681 45.27 0.9656 46.47 0.9632 47.67 0.9609 48.86 0.9585 49.99 0.9564 Add a point by hand out to ~56A (44.08 to 49.99 is factor of 0.987914, so value at 49.99+(49.99-44.08) = 55.9 should be 0.987914 * 0.9564 = 0.944841 : 55.90 0.9448 and 49.99+2.0*(49.99-44.08) = 61.81 and (0.987914)^2 * 0.9564 = 0.933422 61.81 0.9334 ----- OK, put the above 'Fluffium' values into a table using an editor: acis_contam_fluf20thk.tbl and include some header information similar to the other files. Now read in this Fluffium file as well as the "abs" file and write out a new version which has an extra Fluffium column - these steps were done in IDL as documented in: contam_add_fluffium.pro Manually pre-pend header info to the IDL-written file and now we have the new file: acis_contam_abs_v2.tbl Which looks like: # Created by DD on 12/17/03 # # This file was created by adding a Fluffium column # to the existing file # acis_contam_abs_v1.tbl . # The Fluffium transmission was taken from the file: # acis_contam_fluf20thk.tbl and converted to optical depths. # # See also: fluffium_notes.txt and contam_add_fluffium.pro . # # Optical depths due to each component # # wavelength C-K O-K F-K Fluffium 1.01250 0.00000 0.00000 0.00000 1.00000e-05 1.03750 0.00000 0.00000 0.00000 1.00000e-05 etc. 17.5625 0.241522 0.0499158 0.0888599 0.0699981 etc. 18.1625 0.261969 0.0539561 0.00315442 0.0728326 etc. 20.0625 0.336250 0.0692934 0.00439601 0.0826150 etc. 56.3125 0.165716 0.0486399 0.0700007 0.0576293 56.3375 0.165903 0.0487016 0.0700750 0.0576806 - - - - - Optical depth at ~0.7 keV - - - - - - The Total optical depth "at ~0.7 keV" for ECS purposes is the weighted average of optical depths at the three wavelengths shown above with weights 0.48, 0.47, and 0.05 (from Alexey) For this new with-Fluffium material we have (at t = 2002.82 where the time file gives additional factors: 0.9846 1.1099 0.9157 and assuming a fluffium factor of 1.0) > print, TOTAL( [0.9846, 1.1099, 0.9157, 1.0] * [0.241522, 0.0499158, 0.0888599, 0.0699981] ) 0.444571 > print, TOTAL( [0.9846, 1.1099, 0.9157, 1.0] * [0.261969 , 0.0539561 , 0.00315442 , 0.0728326] ) 0.393542 > print, TOTAL( [0.9846, 1.1099, 0.9157, 1.0] * [0.336250 , 0.0692934 , 0.00439601 , 0.0826150] ) 0.494621 weighting these: 17.5625 : 0.444571 * 0.48 = 0.213394 18.1625 : 0.393542 * 0.47 = 0.184965 20.0625 : 0.494621 * 0.05 = 0.0247311 ---------- 0.423090 ( trans = 0.655020) How does this compare with the Tennant equation prediction? Using the code in contam_add_fluffium.pro Tennant at 2002.82 gives tau: > print, tau_ref 0.460100 (trans = 0.631221) and the ratio if transmissions is: > print, 0.655020/0.631221 1.03770 So the fluffium needs to absorb ~3.8% more... Adjust it's tau factor in the time file to do this, calculating: It's average value in first two points above is 0.5*(0.0699981+0.0728326) = 0.0714153 and transmission = 0.9311. Multiply this by: 0.631221/0.655020 = 0.963667 gives tansmission_desired = 0.9311 * 0.963667 = 0.897270 which has a tau of 0.108398 So, need a tau factor for Fluffium of 0.108398/0.0714153 which is about: 1.52 . [ Why is the factor different from 1.0 if I already set the Fluffium to be at the correct level? I think the answer is I wanted a factor of 0.92 between the two E0102 observation dates... But, set the Fluffium to give a total factor of 0.92 at 2002.97 and since the first observation was at a non-zero-contamination time I should have put a bigger factor in so the _difference_ between the times due to fluffium would be 0.92... One could re-do the Fluffium model to get this, but given its approximate nature anyway...] -------------------------------------- OK, now the time dependance.... We want a new version of acis_contam_tau_t_v2_2006.tbl which will have the 2002.82 line read: 2002.820 0.9846 1.1099 0.9157 1.5200 and which will be scaled from these values by the Tennant tau trend equation... So do that in contam_add_fluffium.pro and write out the new file: acis_contam_tau_t_v3.tbl and add header info to this IDL-created file... et voila. - - - - - NOTE added February 1, 2004: A draft contamination file, 'N9998 was created using the results above. After in-house testing the decision was made to adjust the fluffium component further for better agreement with data in the 0.3 to 0.5 keV range so the process was repeated to create the final N0003 product, see: fluffium_N9997.txt in this same directory. - - - - - - -