1、ST-ECF Instrument Science Report ACS 2001-009Selection of WavelengthCalibration Targets for theACS GrismA. Pasquali, N. Pirzkal, J.R. WalshNovember 14, 2001ABSTRACTWe discuss the criteria for the selection of wavelength calibration targets for the ACSgrism to determine its dispersion and zero point.
2、 The primary requirement is for compactemission line sources. SLIM 1.0 simulations indicate that Galactic Wolf-Rayet stars are tobe preferred to extragalactic Planetary Nebulae, since they can provide the wavelengthcalibration of both the 1stand 2ndgrism orders with exposures shorter than 1 minute a
3、ndare genuine point sources. Two selected targets, WR45 and WR96, were observed withNTT/EMMI at high spectral resolution (1.26 /pix) between 5000 and 1 m. Thesespectra will be used as templates to identify the emission lines in the SMOV observationsof the same objects using ACS, and to ultimately as
4、sess the on-orbit dispersion of thegrism with an accurancy of 5% of the nominal dispersion.IntroductionSimulations carried out for emission-line objects of varying size and position angle on thesky have shown (Pasquali et al. 2001) that the size of the source projected onto the disper-sion axis smea
5、rs out the spectral resolution of the grism, producing severe line blendingfor sizes larger than 2 pixels (0.1 in the case of the WFC) and a major axis PA 45owiththe image X axis. This result poses important restrains in the planning of wavelength cali-bration observations, such as those forseen for
6、 the SMOV tests and subsequent pipelinecalibration. Wavelength calibration sources should thus have minimal spatial extension toThe Space Telescope European Coordinating Facility. All Rights Reserved.allow sensitive determination of the dispersion and zero point in orbit.ST-ECF Instrument Science Re
7、port ACS 2001-009The selection criteriaTo ensure an accurate wavelength calibration of the ACS grism, the targets have to meetthe following requirements:- Their luminosity allows for reasonably short exposure times ( 1 orbit).- Their optical spectra present a significant number of emission lines.- N
8、o nebulosity is associated with the targets which would degrade the spectral resolutionof the instrument and increase the local background.- Spectrophotometric variability (either intrinsic or induced by an eclipsing companionand by orbital motion) is negligible so that emission features can be iden
9、tified at thesame wavelength at any observational date.- Targets should not lie in crowded field to avoid contamination by nearby spectra.- Targets should be visible as long as possible (likely CVZ) to allow repeated HST visits.In the past, Planetary Nebulae have well served calibration purposes bec
10、ause of their richemission-line spectrum and large population in the Galaxy and in the Local Group. Inwhat follows, we will show that they do not satisfy the above criteria and that anotherclass of sources has to be used to calibrate the ACS grism in wavelength.The traditional calibrators: Planetary
11、 NebulaeGalactic Planetary Nebulae (PNe) are rarely compact (full size 0.1) when imaged withHST and even when compact often have a low intensity halo. Therefore, the need for pre-serving the nominal spectral resolution of the ACS grism requires extragalactic PNe to beobserved.Ciardullo et al. (1989)
12、 and Jacoby it includes both the 1stand 2ndorder spec-tra of CJFN #1, #31 and JC #57.Figure 1: The 1stand 2ndorder grism simulated WFC spectra.The above spectra have been extracted with a box aperture 12 pixels (0.6) wide and arepresented in Figure 2, in units of count/pixel from the object position
13、 in the direct image.The background has not been subtracted; the subtraction however would not improve theS/N ratio since additional noise would be added.For the adopted exposure time, none of the sample PNe provide a sufficient number ofemission lines to wavelength calibrate the 2ndorders. As far a
14、s S/N is concerned, CJFN#31 and JC #57 are most suited in deriving the dispersion correction for the WFC grism1storder and especially its zero point, although this takes almost half a HST orbit.In the case of the HRC and accounting for the lower CCD QE and the higher spectral res-olution than for th
15、e WFC, SLIM 2400 s exposures have been generated in order toachieve the same S/N ratio as for the WFC grism. Background and read-out noise havebeen added to the simulated grism image. The stamp image of the 1stand 2ndorder spectraare shown in Figure 3.The 45otilt of the spectra requires an ad hoc ex
16、traction procedure, whereby the originalgrism image is rotated to align the dispersion axis with the image X axis and each spec-trum is extracted with a box aperture of 5 pixels (0.14) in the cross-dispersion direction.The resulting 1stand 2ndorder spectra are plotted in Figure 4 in units of count/p
17、ixel fromthe object position in the direct image.CJFN #1CJFN #31JC #573ST-ECF Instrument Science Report ACS 2001-009Figure 2: SLIM simulations of PNe in M31, as observed with the WFC grism in a 1200 sexposure. The 1st(in the left panels, at 40 /pix) and the 2ndorder spectra (in the rightpanels, at 2
18、0 /pix) have not been background-subtracted.Figure 3: the 1stand 2ndorder grism spectra simulated for the HRC.CJFN #1CJFN #31JC #574ST-ECF Instrument Science Report ACS 2001-009Similarly to the WFC grism, PNe CJFN #31 and JC #57 can be used to determine the dis-persion correction of the 1storder of
19、the HRC grism, but this still requires a HST orbitintegration. The 2ndorders are hardly detected.In conclusion, observations of extragalactic PNe turn out to be too time-consuming forroutine wavelength calibration and considering the number of orbits allocated to theSMOV grism tests. Moreover, extra
20、galactic PNe can only be used to wavelength calibratethe 1storders of the grism.Figure 4: SLIM simulations of PNe in M31, as observed with the HRC grism in a 2400 sexposure. The 1st(in the left panels, at 25 /pix) and the 2ndorder spectra (in the rightpanels, at 12 /pix) have not been background-sub
21、tracted.In addition, the requirement to have multiple wavelength calibration spectra at differentpositions in the field, to measure the variation of the dispersion solution as a function ofspatial position, cannot be met in one orbit for either the WFC or the HRC.5ST-ECF Instrument Science Report AC
22、S 2001-009An alternative choice: Galactic Wolf-Rayet StarsThe ACS spatial sampling and the spectral resolultion of the grism suggest that the wave-length calibration be performed with point source emission line objects like GalacticWolf-Rayet (WR) stars.The spectrum of WR stars is characterised by a
23、 number of emission lines of H, He, N andC originating in the stellar wind. The predominance of either N or C features determinesthe classification of WRs into either WN or WC spectral types, respectively.An example of these spectral classes is shown in Figure 5 (courtesy by P. Crowther).The absorpt
24、ion at 6250 is an artefact.Figure 5: The optical spectra of typical WN6 and WC8 Wolf-Rayet stars.WRs are known for their high velocity winds which considerably broaden their emissionfeatures. The line broadening produced by a wind speed of 2000 km s-1in the case of theWFC and HRC grism is listed in
25、Table 2, and has been computed with respect to the wave-length at which the 1stand 2ndorder responses peak, which represents a mean valuethroughout the spectral range of the grism. The spectral resolution is 40 /pix and 20 /pix for the WFC 1stand 2ndorder spectra, 25 /pix and 12 /pix for the HRC 1st
26、and 2ndorders, respectively.Table 2. The line broadening induced by a stellar wind at 2000 km s-1on the 1stand 2ndorder spectra of the WFC and HRC grism. The wind “effect” has been computed for thepeak wavelength of the 1stand 2ndorder responses and is expressed in units of and pix-els along the dis
27、persion axis.LinebroadeningWFC grism1storder=7653 WFC grism2ndorder=5997 HRC grism1storder=7005 HRC grism2ndorder=5999 in 51404740in pixels 1.3 1.9 1.9 3.36ST-ECF Instrument Science Report ACS 2001-009According to Table 2, the stellar wind is hardly resolved in the 1storder spectra of bothChannels,
28、and partially affects the 2ndorder spectra. van der Hucht (2001, The VIIthcata-logue of galactic Wolf-Rayet stars) lists a wind velocity range between 700 km s-1and3300 km s-1, with 19% of the sample having a wind velocity larger than 2100 km s-1. Themean wind velocity computed over the catalogue is
29、 (1730 +/- 700) km s-1.Following the selection criteria listed in page 2, two WR stars have been first selectedfrom the VIIthCatalogue by van der Hucht which are listed in Table 3.Table 3. Basic parameters for the selected Wolf-Rayet stars.Their finding charts follow in Figure 6; WR45 is on the left
30、 and WR96 on the right. Thefield of view is 3 x 3.Figure 6: The finding charts of WR45 (on the left) and WR96 (on the right). The fieldof view is 3 x 3 (images from the Digital Sky Survey).To overcome the lack of homogeneous spectra published for these objects over the grismwavelength range, we obse
31、rved WR45 and WR96 with EMMI mounted on the ESO NTTtelescope in June and August 2001. One hour of ESO Director General Discretinary Timewas allocated to acquire spectra in the following ranges:- 5000 - 7500 at 1.26 /pix (in REMD mode with Grating #8, c= 6200 );SpectraltypeRA (2000) DEC (2000) VWindv
32、elocityWR45 WC6 11 38 05.2 -62 16 01 14.802100 kms-1WR96 WC9 17 36 24.2 -32 54 29 14.141100 kms-17ST-ECF Instrument Science Report ACS 2001-0098- 7300 - 9750 at 1.26 /pix(in REMD mode with Grating #8, c= 8550 );- 4000 - 9000 at 2.7 /pix (in RILD mode with Grism #2).The high resolution spectra have b
33、een used to accurately identify the He and C emissionlines, while the low resolution data have been used to properly calibrate the spectra in flux.The spectra at high resolution are plotted in Figure 6 for both WR45 and WR96.SLIM 1.0 (Pirzkal et al. 2001) has been used to produce WFC and HRC grism i
34、mages,using the NTT observed spectra, which also include the background and read-out noise.The ESO spectra shown in Figure 7 have served as SLIM input spectra.The WFC 1storder spectra (extracted with IRAF) are plotted in Figure 8, left column, foran exposure time of 10 s, while the right column show
35、s the 2ndorders obtained with anintegration time of 60 s. They are in units of count/pixel from the object position in thedirect image. The spectra are background-subtracted.Both targets provide a large enough number of emission lines to calibrate in wavelengthboth the 1stand 2ndorders of the WFC gr
36、ism in exposures shorter than 1 minute.The 1stand 2ndorder spectra of WR45 and WR96 as observed with the HRC grism arefound in Figure 9. They are plotted in units of count/pixel from the object position in thedirect image. The 1storders have been obtained by integrating for 20 s, while the 2ndorders
37、pectra have an exposure time of 60 s. Again, the number of emission features is largeenough to allow an accurate wavelength calibration of both orders of the HRC grismwithin 1 minute exposure, thus allowing several dispersion measurements as a function ofposition on the chip with one orbit.ST-ECF In
38、strument Science Report ACS 2001-0099Figure 7: The high resolution (1.26 /pixel) spectra of WR45 and WR96 obtained withthe NTT/EMMI spectrograph.Figure 8: The SLIM output spectra for WR45 and WR96. First orders are plotted in theleft column and refer to an exposure time of 10 s. The second order spe
39、ctra are shown inthe right column and have been computed for an integration time of 60 s.ST-ECF Instrument Science Report ACS 2001-009Figure 9: The SLIM output spectra for the HRC grism. 1storders are plotted in the leftcolumn and refer to an exposure time of 20 s. The 2ndorder spectra are shown in
40、theright column and have been computed for an integration time of 60 s.As a counter-check, we have derived the dispersion correction of the 1stand 2ndorders ofthe grism for both the WFC and the HRC. Specifically, we have measured the position ofeach identified line in the SLIM output for WR45 and WR
41、96 in pixel scaled to the positionof the object in the direct image, and fitted pixel against wavelengths with POLYFIT inIRAF, assuming a first order polynomial. The results, dispersion and zero-point of eachgrism order in each channel are listed in Table 4 as averaged values over the two simu-lated
42、 WRs. The RMS of the fits is 5% of the derived dispersion.Table 4. The grism parameters: dispersion and zero-point, in the case of the WFC and theHRC. These values are averaged over the SLIM simulated spectra of WR45 and WR96.WFC1storderWFC2ndorderHRC1storderHRC2ndorderDispersion (/pix) 39.91 +- 0.0
43、3 19.95 +- 0.01 25.02 +- 0.02 12.45 +- 0.03Zero-point () 5219.04 +- 0.57 2606.56 +- 4.19 3273.41 +- 3.23 1629.51 +- 0.9010ST-ECF Instrument Science Report ACS 2001-009ConclusionsProper wavelength “calibrators” for the ACS grism have to be compact sources, in noncrowded fields. They should not show s
44、ignificant variability in line intensity and shouldbe bright enough to allow the calibration of the 1stand 2ndorder spectra of the WFC andHRC grism, using short exposure times such as those planned for the SMOV tests andanticipated for routine wavelength calibration.In the case of ACS optical spectr
45、oscopy (i.e. with the WFC and HRC grism), SLIM 1.0simulations have shown that extragalactic PNe (i.e. in M31), although compact enough notto be resolved by HST, are generally too faint to be calibration targets.Galactic WR stars, which satisfy the above mentioned requirements, are alternative wave-l
46、ength “calibrators”. Simulations of template WC spectra have shown that WC stars canbe observed in less than 1 minute and provide a large number of emission lines at goodS/N for the calibration of both 1stand 2ndgrism orders, for both the WFC and HRC. As aresult, two objects, WR45 and WR96, have bee
47、n selected as main wavelength calibratorsof the WFC and HRC grism.AcknowledgementsWe would like to acknowledge the ESO Director General Discretionary Time programmeand the NTT team who carried out the observations in servicing mode.ReferencesCiardullo, R., Jacoby, G.H., Ford, H.C., Neill, J.D., 1989
48、, ApJ, 339, 53Jacoby, G.H., Ciardullo, R., 1999, ApJ, 515, 169Pasquali, A., Pirzkal, N., Walsh, J.R., Hook, R.N., Freudling, W., Albrecht, R., Fosbury,R.A.E., 2001, “The Effective Spectral Resolution of the WFC and HRC Grism”, ST-ECFISR ACS 2001-002Pirzkal, N., Pasquali, A., Walsh, J.R., Hook, R.N., Freudling, W., Albrecht, R., Fosbury,R.A.E., 2001, “ACS Grism Simulations using SLIM 1.0”, ST-ECF ISR ACS 2001-001van der Hucht, K. A., 2001, “The VIIth catalogue of galactic Wolf-Rayet stars”, New AR,45, 13511
