A Cr-K emission line survey in young supernova remnants with Chandra
aa r X i v : . [ a s t r o - ph ] O c t draft version A Cr-K emission line survey in young supernova remnants with
Chandra
X.J. Yang , , H.Tsunemi , F.J. Lu , L. Chen ABSTRACT
We performed a Cr-K emission line survey in young supernova remnants(SNRs) with the
Chandra archival data. Our sample includes W49B, Cas A,Tycho and Kepler. We confirmed the existence of the Cr line in W49B anddiscovered this emission line in the other three SNRs. The line center energies,equivalent widths (EWs) and fluxes of the Cr lines are given. The Cr in Cas Ais in a high ionization state while that in Tycho and Kepler is in a much lowerone. We find a good positive correlation between Cr and Fe line center energies,suggesting a common origin of Cr and Fe in the nucleosynthesis, which is consis-tent with the theoretical predictions. We propose that the EW ratio between Crand Fe can be used as a supplementary constraint on the progenitors’ propertiesand the explosion mechanism.
Subject headings:
ISM: supernova remnants – ISM: individual: W49B, Cas A,Tycho, Kepler
1. Introduction
The X-ray emission of young supernova remnants (SNRs) is predominantly from theejecta heated by the reverse shock. Since the ejecta are metal abundant, their X-ray spectrausually show strong emission lines of heavy elements such as O, Ne, Mg, Si, S, Ar, Ca and Particle Astrophysics Center, Institute of High Energy Physics, Chinese Academy of Sciences, Beijing100049, P.R. China, [email protected], [email protected] Department of Astronomy, Beijing Normal University, Beijing 100875, P.R. China Department of Earth and Space Science, Graduate School of Science, Osaka University, Toyonaka, Osaka560-0043, Japan; [email protected] − ASCA ) (Hwang et al. 2000a)and
XMM-Newton (Miceli et al. 2006). Hwang et al. (2000a) proposed that Cr, Mn and Niwould be the most promising heavy atomic species for future detection, since they are themost abundant elements next to Fe with K-shell emission lines at energies above ∼ Chandra
X-ray observatory. Wedescribe the observations and data analyses in §
2. The results are shown in §
3, discussionin § §
5. All through this paper, the statistical uncertainties are given at90% confidence level.
2. Observations and Data analyses
SNRs are important observational targets for
Chandra . For example, as one of thelargest projects,
Chandra has observed the Cas A SNR for a few times with a total exposuretime of 1 Ms (Hwang et al. 2004). The Kepler SNR was also observed for 750 ks (Reynoldset al. 2007). In this paper, we collected almost all the available
Chandra
ACIS-S data ofthe four SNRs. The detailed observation information is listed in Table 1. The data wereprocessed using the CIAO software package (version 3.4). We created new level 2 event filesfor all the observations we used, applying gain map, time-dependent gain and charge transferinefficiency corrections with the latest released calibration files (CALDB version 3.4.3 and 3 –ATOMDB version 1.3.1). The only exception is that the charge transfer inefficiency cannot be corrected for all the Cas A data since they are acquired in GRADED mode. Fig. 1shows the X-ray images of the four SNRs. Since the Cr line is very weak, we extracted thespectrum of each SNR from almost the entire source region, as illustrated in Fig. 1. Thebackground spectra were extracted from the off-source regions.Cas A and Kepler have been observed for a few times spanning 4 to 5 months. Wetherefore carefully performed the analysis to eliminate the effect of any CCD performanceevolution with time. Taking Cas A as an example, we first created source and backgroundspectra as well as the corresponding RMF and ARF files for each of the 9 observations (in-cluding IDs 4634, 4635, 4636, 4637, 4638, 4639, 5196, 5319, 5320), and then combined themwith FTOOLs . This process has taken into account the degradation of the CCD perfor-mance during the observation span. We noticed that there is no clear difference betweenthe 9 RMF files, therefore the ARF files can be added with FTOOLs addarf . The totalphoton number of each source spectrum was taken as the corresponding adding weight. Weperformed a joint fit to the spectra of all the observations by using their respective RMFand ARF files, and found that the fitting results are consistent with those obtained by usingthe combined data and the added RMF and ARF. This demonstrates that the combiningprocess works well and suggests that the gain is properly calibrated from observation toobservation. The spectral fitting was done with XSPEC version 11.3.2 (Arnaud 1996).The 0.5 − − χ ± .
87 eV with the power law model and 10.7 ± .
95 eV with the bremsstrahlung.For the other three SNRs, the power law model also fits the continuum emission better thanthe bremsstrahlung model does. The latter model seems to overestimate the spectra at lowenergies, while underestimating at high energies. In this case, the power law model was used http://heasarc.gsfc.nasa.gov/docs/software/ftools/ − § ), the two lines below5.0 keV could be the He-like Ca-K emission, while the Ti-K emission could not be ruledout as it also emits around those energies. Unfortunately, Ti has not been considered inthe currently available plasma codes, thus we cannot get the emissivities of the Ti lines tojudge whether they represent these observed line features or not. Since the main topic ofthe current paper is Cr emission and the line candidates below 5.0 keV are beyond its scope,we only adopt data above 5.0 keV for the analyses of the continuum and line emission. Asan element of the Fe-group, Cr is closely related to Fe in its synthesis process (Woosley &Weaver 1994; Woosley & Weaver 1995; Thielemann et al. 1996). A comparative study ofthe Fe and Cr emission should be important. Therefore we take the 5.0 − − −
3. Results3.1. W49B
In the analysis of the
ASCA spectrum, Hwang et al. (2000a) found evidence of the He-like Cr and Mn lines near 5.69 and 6.18 keV, and the existence of these lines was confirmedby
XMM-Newton (Miceli et al. 2006). From the spectra (Fig. 2 (top left) and Fig. 4 (topleft)), we can see that the Cr and Mn lines are also detected with
Chandra . We fitted the5.0 − http://cxc.harvard.edu/atomdb χ , from 93.4 for d.o.f 68 to 185.8 (or 103.7) for d.o.f71. The Cr line emission is firmly detected in the Chandra spectrum, while the Mn line isless significant.Fig. 4 (top left) shows the fitting model and residuals, and Table 2 gives the best fitresults for the Cr and Fe lines, including line center energies and fluxes along with theiruncertainties. The EWs of the two lines, calculated from the XSPEC command eqwidth ,are given in the same table. From our fitting analyses, the Cr line center energy is E Cr =5.656 +0 . − . keV, and the flux is f Cr = (0.32 +0 . − . ) × − photons cm − s − , while for the Mnline, E Mn = 6.126 +0 . − . keV and f Mn = (0.10 ± . × − photons cm − s − . The fittedline parameters are generally in good agreement with those from ASCA and
XMM-Newton (c.f Table 3).
The 5.0 − ± . ± .
001 keV, with their widths 100 ± ± §
2. Adding the third Gaussian leads to amuch better fit (c.f. Fig. 3), with the reduced χ decreasing from 993.8/65 to 222.9/62 fordata in 5.0 − XMM-Newton (Willingale et al. 2002) and
Chandra (Yang et al. 2008)shows that different parts of this remnant are moving with different line of sight velocities,with typical value of ± − . Considering this, we selected two regions based on theDoppler shift map from the XMM-Newton observation (Willingale et al. 2002), as illustratedin Fig.1. The spectra of the two regions are given in Fig. 5. They were fitted with a powerlaw plus two Gaussian components, representing the continuum, Cr and Fe-K line emission.The Cr line is very significant in both spectra. Adding a Gaussian component for Cr leadsto the reduced χ decreasing from 440.0/62 to 197.4/59 and from 670.9/62 to 184.2/59 forthe blueshifted and redshifted regions respectively in 5.0 − ± .
001 keV and 6.669 ± .
001 keV, which are generallyconsistent with the line center energies of the two components fitted to the Fe-K line of thewhole remnant spectrum. This demonstrates that the Doppler shift dominates the structureof the Fe-K complex of Cas A. The centroid energies of the Cr line in the two regions are5.590 +0 . − . and 5.657 +0 . − . keV, with a separation similar to that of the two Fe-K lines. TheCr and Fe ejecta are thus probably moving with similar velocity. We noticed that the Fe-Kline from a small region of Cas A also cannot be well represented by one Gaussian component(c.f Fig. 5, residual distribution), suggesting small velocity difference even within such a scale(c.f Fig. 7 in Willingale et al. 2002).On the other hand, the existence of more than one strong Fe-K line that cannot beresolved with Chandra
ACIS could also contribute to the broadening of the Fe-K line in CasA. Various Fe-K lines around this energy have been clearly displayed in the spectra of thecataclysmic variables V1223 Sagittarii (Mukai et al. 2001) and U Geminorum (Szkody et al.2002) using data collected by the High Energy Transmission Grating onboard
Chandra . Thecentroid energy difference of the two Gaussian components in Cas A is about 60 eV, which iscomparable with the difference of the He-like triplet of Fe-K lines ( ∼
65 eV ). Higher energyresolution spectra from future missions will help us to resolve the Fe-K complex. The 5.0 − χ decreased from 36.6/21 to11.4/18 in 5.0 − χ from 47.6/24 to27.8/21. Finally, the 5.0 − χ mentioned above and the lower limits of the confidence ranges. From Table 2, we cansee that the centroid energies of the putative Cr lines in Tycho and Kepler are about 200 eVlower than those of W49B and Cas A. This will be further discussed in §
4. Discussion4.1. The ionization state and spatial correlations between Cr and Fe
The different line center energies of the Cr emission lines in these SNRs reflects thatthe ionization states of Cr in these SNRs are dissimilar. The Cr in W49B is in relativelyhigh ionization state since its line center energy is 5.656 +0 . − . keV (c.f Table 2). For Cas A,the centroid energy of the Cr line is 5.635 +0 . − . keV, which also suggests a high ionizationstate of Cr. The center energies of the putative Cr lines in Tycho and Kepler are about 200eV lower than those in W49B and Cas A. If these lines do come from Cr, it should be in arelatively low ionization state (c.f Table 4). We noticed that the Fe-K line center energies inTycho and Kepler are both around 6.44 keV. This implies low ionization states of Fe as well,as suggested by previous observations (Tsunemi et al. 1986; Hwang et al. 1998; Hwang etal. 2002; Kinugasa & Tsunemi 1999).In Fig. 6, we plot the Cr line center energy versus that of the Fe line in these four SNRs.The theoretical centroid energies of Cr and Fe K lines in different ionization states given inTable 4 are overplotted in the same figure. Obviously, there is a positive correlation betweenthe two line center energies both theoretically and observationally. We can conclude fromFig. 6 that the emission lines around 5.46 keV in Tycho and Kepler are from Cr. Meanwhile,the Cr in Cas A and W49B might be He/Li-like, while Ne-like or an even lower ionizationstate in Tycho and Kepler. Since the line center energy is closely related to the ionizationages of the emitting plasma, such a positive correlation implies that the ionization ages ofCr and Fe are closely related to each other.The above ionization state correlation suggests that the Cr and Fe ejecta are co-located,which is also supported by the spectra of the blueshift and redshift regions in Cas A. Asshown in § +0 . − . %, c.f. Table2) within the confidence range. This is further evidence that Cr and Fe are co-located.According to the nucleosynthesis theory (Woosley & Weaver 1994; Woosley et al. 1995), Cr is generated by explosive oxygen and silicon burning, while , Cr are the products of 8 – , Fe decay in explosive silicon burning. The most abundant Fe is Fe decayed from Ni,which is also produced mainly from explosive silicon burning. Among all the isotopes of Cr, Cr is the most abundant, and its production is generally at least one order of magnitudegreater than the others (Nomoto et al. 1984; Woosley et al. 1995). In this case, most of theCr would be located near and thus share a similar ionization time with Fe in SNRs. Cayrelet al. (2004) observed a number of so-called “first stars”, i.e. very metal-poor dwarfs andgiants. They found that the scatter of the Cr/Fe values of these stars is very small, indicatingthat the production of Fe and Cr are very closely linked. Our results are consistent withthe theoretical predictions and further strengthen the previous observational statement in adifferent way.
Many theoretical calculations of nucleosynthesis have included Cr for both Ia (Woosley& Weaver 1994; Iwamoto et al. 1999) and core-collapse SNe (Woosley & Weaver 1995;Thielemann et al. 1996; Maeda & Nomoto 2003). In order to compare with these models,we need the mass (or abundance) ratio between Cr and the other elements. Hwang et al.(2000a) interpolated the He α emissivity of Cr from those of Si, S, Ar, Ca and Fe, using theRaymond-Smith (RS) code for the 2 keV plasma in collisional ionization equilibrium (CIE),and then calculated the Cr abundance of W49B. They found that the Cr and Fe abundancesare consistent with a solar ratio, corresponding to an atomic number ratio 1 .
0% of Cr toFe (Anders & Grevesse 1989) and thus a mass ratio M Cr/F e ∼ . § Cr/F e ) would be a good representationof the mass ratio of these two elements. Considering this, we use the EW ratio for thediscussion below.From Table 2, we can see that EW
Cr/F e of these four SNRs differ from one another.This ratio might be used to constrain the properties of the corresponding SNe. Badenes et al.(2006) have made detailed comparisons of the X-ray spectra of the type Ia SNR Tycho withthe theoretical models. They found the one-dimensional delayed detonation model can wellreproduce its X-ray emission. From numerical calculations, the standard SNe Ia models,i.e carbon deflagration and Chandrasekar mass models (e.g W7, W70 etc, Nomoto et al.1984) often yield relatively small M Cr/F e ( < M Cr/F e (Travaglio et al. 2004; 2005).On the other hand, the delayed detonation models (WDD, CDD etc) produce much larger M Cr/F e ( > M Cr/F e decreases as the transition density increases (Nomoto et al.1997; Iwamoto et al. 1999). Therefore, our observational results of Tycho suggest that thereshould be a deflagration-detonation transition at some stage of Tycho’s SN explosion, whichfurther confirm the results of Badenes et al. (2006). Meanwhile, we favor a relatively smalltransition density, probably 1 . × g cm − (Nomoto et al. 1997; Iwamoto et al. 1999).This is also consistent with that suggested by Badenes et al. (2006, 2 . × g cm − ).Cas A has been identified as the remnant of a core-collapse SN. The recent study ofits progenitor implies the mass to be 15 − M ⊙ (Young et al. 2006). According to thecalculations of spherical models by Thielemann et al. (1996), a 20 M ⊙ progenitor givesa M Cr/F e of 1.2%, which is higher than our measured value. However the non-sphericalexplosion may lead to a smaller M Cr/F e (Maeda & Nomoto 2003). It has already beensuggested that the explosion of Cas A is asymmetric, based on the jet structure (Hwang etal. 2004) and the Doppler map (Willingale et al. 2002). Meanwhile, a bigger progenitormass would lead to a smaller mass cut and thus smaller M Cr/F e , since Cr is mainly producedin the incomplete Si-burning zone (Umeda & Nomoto 2002). Therefore, we support a higherprogenitor mass (Young et al. 2006) and the asymmetric explosion scenario for Cas A.The classifications of W49B and Kepler are not conclusive. For W49B, Hwang et al.(2000a) compared the relative abundances of Mg, S, Ar, Ca, Fe and Ni to Si, and suggestedthat W49B may have a type Ia progenitor. However they claimed that a low mass (13 − M ⊙ ) type II progenitor is also possible. The Chandra image of W49B shows a bipolar jet,which was taken as evidence for a gamma-ray burst (GRB) remnant . This was furthersupported by multi-band observations (Keohane et al. 2007). The nucleosynthesis calcu-lation for bipolar core-collapse SN explosions (Maeda & Nomoto 2003) generally predictsan EW Cr/F e of 1 . ± . Cr/F e would favor the carbon deflagration models (W7,W70 etc) rather than those involving detonation (Nomoto et al. 1997; Iwamoto et al. 1999;Travaglio et al. 2004; 2005).The above discussions are based on the overall mass ratio of Cr to Fe in an SNR. It ispossible that in these SNRs there are still a fraction of Fe and Cr that has not been overtakenby the reverse shock and so invisible in X-rays. However, according to our discussion in § Http://chandra.harvard.edu/press/04 releases/press 060204.html
10 –overall Cr to Fe mass properly. Therefore the main conclusions are reliable no matter whatfraction of Fe is observed.
5. Summary
We performed a Cr K line survey with the
Chandra data in young SNRs W49B, CasA, Tycho and W49B. We confirmed the Cr line in W49B, and gave a consistent flux andline center energy with respect to the previous results. Then we report, for the first time,the detection of Cr lines in Cas A, Tycho and Kepler. We conclude that Cr in Cas A isin a high ionization state similar to that of W49B, while Cr is in a low ionization statelow in Tycho and Kepler. We find that Cr and Fe have similar ionization states and areco-located in these four SNRs. The reason might be that Cr and Fe are synthesized bythe same process deep inside the progenitor. We propose that the EW ratio of Cr to Femight be used as an supplementary constraint on the properties of the SN explosions. Forthe type Ia SNR Tycho, EW
Cr/F e favors the delayed detonation model with relatively smalltransition density (1 . × g cm − ) from deflagration to detonation. This is consistentwith the model suggested by Badenes et al. (2006) from the comparison of Tycho’s X-rayspectra with theoretical calculations. The relatively small EW Cr/F e in Cas A and W49Bsuggests their asymmetric explosions, which is also consistent with the previous results. Ifwe adopt the type Ia origin for Kepler, its small EW
Cr/F e could be attributed to the carbondeflagration explosion.We are grateful to the anonymous referee for very helpful comments and suggestionsleading to significant improvements of the paper. The manuscript is read by Dr. E. Millerin MIT. This work is supported by the Nature Science Foundation of China through grants10533020, 10573017, 10778716 and by a Grant-in-Aid for Scientific Research by the Ministryof Education, Culture, Sports, Science and Technology, Japan(16002004).
REFERENCES
Anders, L., & Grevesse, N. 1989, Geochim. Cosmochim. Acta, 53, 197Arnaud, K. A. 1996, ASPC, 101, 17Badenes, C., Borkowski, K. J., Hughes, J. P., Hwang, U., & Bravo, E. 2006, ApJ, 645, 1373 11 –Cassam-Chena¨ı, G., Decourchelle, A., Ballet, J., Hwang, U., Hughes, J. P., Petre, R. 2004,A&A, 414, 545Cayrel, R., Depagne, E., Spite, M., Hill, V., Spite, F., Francois, P., Plez, B., Beers, T.,Primas, F., Andersen, J., et al. 2004, A&A, 416, 1117Hatsukade, I., Tsunemi, H., Yamashita, K., Koyama, K., Asaoka, Y., & Asaoka, I. 1990,PASJ, 42, 279Holt, S. S., Gotthelf, E. V., Tsunemi, H., & Negoro, H. 1994, PASJ, 46, L151Hwang, U., Decourchelle, A., Holt, S. S., Petre, R. 2002, ApJ, 581, 1101Hwang, U., & Gotthelf, E. V. 1997, ApJ, 475, 665Hwang, U., Hughes, J. P., Petre, R. 1998, ApJ, 497, 833Hwang, U., Laming, J. M., Badenes, C., et al. 2004, ApJ, 615, L117Hwang, U., Petre, R., Hughes, J. P. 2000a, ApJ, 532, 970Hwang, U., Holt, S. S., & Petre, R. 2000b, ApJ, 537, L119Hughes, J. P., Hayashi, I., Helfand, D., Hwang, U., Itoh, M., Kirshner, R., Koyama, K.,Markert, T., Tsunemi, H., & Woo, J. 1995, ApJ, 444, L81Hughes, J. P., Rakowski, C. E., Burrows, D. N., & Slane, P. O. 2000, ApJ, 528, L109Iwamoto, K., Brachwitz, F., Nomoto, K., Kishimoto, N., Umeda, H., Hix, W. R., & Thiele-mann, F-K 1999, ApJS, 125, 439Keohane, J. W., Reach, W. T., Rho, J., & Jarrett, T. H. 2007, ApJ, 654, 938Kinugasa, K., & Tsunemi, H. 1999, PASJ, 51, 239Maeda, K. & Nomoto, K. 2003, ApJ, 598, 1163Mewe, R., Gronenschild, E. H. B. M. & van den Oord, G. H. J. 1985, A&A, 62, 197Mukai, K., Kallman, T., Schlegel, E., Bruch, A., Handler, G. & Kemp, J. 2001, ASPC, 251,90Miceli, M., Decourchelle, A., Ballet, J., Bocchino, F., Hughes, J. P., Hwang, U., & Petre, R.2006, A&A, 453, 567 12 –Nomoto, K., Thielemann, F.K., & Yokoi, K. 1984, ApJ, 286, 644Nomoto, K., Iwamoto, K., Nakasato, N., Thielemann, F.-K., Brachwitz, F., Tsujimoto, T.,Kubo, Y., & Kishimoto, N. 1997, NuPhA, 621, 467Rakowski, C. E., Badenes, C., Gaensler, B. M., Gelfand, J. D., Hughes, J. P. & Slane, P. O.2006, ApJ, 646, 982Reynolds, S. P., Borkowski, K. J., Hwang, U., Hughes, J. P., Badenes, C., Laming, J. M., &Blondin, J. M. 2007, ApJ, 668, L135Szkody, P., Nishikida, K., Raymond, J. C., Seth, A., Hoard, D. W., Long, K. S., & Sion, E.M. 2002, ApJ, 574, 942Thielemann, F.-K, Nomoto, K., & Hashimoto, M. 1996, ApJ, 460, 408Travaglio, C., Hillebrandt, W., Reinecke, M., & Thielemann, F.-K. 2004, A&A, 425, 1029Travaglio, C., Hillebrandt, W., & Reinecke, M. 2005, A&A, 443, 1007Tsunemi, H., Yamashita, K., Masai, K., Hayakawa, S., & Koyama, K. 1986, ApJ, 306, 248Umeda, H., & Nomoto, K. 2002, ApJ, 565, 385Willingale, R., Bleeker, J. A. M., van der Heyden, K. J., Kaastra, J. S., & Vink, J. 2002,A&A, 381, 1039Woosley, S. E., & Weaver, T. A. 1994, ApJ, 423, 371Woosley, S. E., Langer, N., & Weaver, T. A. 1995, ApJ, 448, 315Woosley, S. E., & Weaver, T. A. 1995, ApJS, 101, 181Yang, X.J., Lu, F.J., & Chen, L. 2008, ChJAA, 8, 439Young, P. A., Fryer, C. L., Hungerford, A., Arnett, D., Rockefeller, G., Timmes, F. X., Voit,B., Meakin, C., Eriksen, K. A. 2006, ApJ, 640, 891
This preprint was prepared with the AAS L A TEX macros v5.2.
13 –Table 1: Information of
Chandra observations we usedTarget Obs ID t exp ( ks ) Obs-dateW49B 117 ∼
50 July, 2000Cas A VLP ∗ ∼ ∼ May, 2004Tycho 115 ∼
50 September, 2000Kepler LP † ∼
750 April ∼ August, 2006 ∗ Very Large Project, Observation IDs: 4634, 4635, 4636, 4637, 4638, 4639, 5196, 5319 and 5320. † arge Project, Observation IDs: 6714, 6715, 6716, 6717 and 6718. Table 2: The Cr and Fe line parameters of W49B, Cas A, Tycho and KeplerCr FeSNR line center EW flux line center EW flux EW
Cr/F e eV eV − photonscm s eV eV − photonscm s W49B 5656 +14 − +20 . − . +0 . − . +1 − +280 − +0 . − . +0 . − . %CasA 5635 +7 − +1 . − . +0 . − . +1 − +12 − +0 . − . +0 . − . %Tycho 5465 +45 − +40 . − . +0 . − . +6 − +52 − +0 . − . +3 . − . %Kepler 5469 +40 − +8 . − . +0 . − . +1 − +80 − +0 . − . +0 . − . %Table 3: The Cr, Mn and Fe line parameters of W49B from different instruments.Cr Mn FeInstrument line center flux line center flux line center fluxeV − photonscm s eV − photonscm s eV − photonscm s ASCA a +20 − +0 . − . +47 − +0 . − . +3 − ± . XMM-Newton b ±
10 0.25 ± .
04 6170 ±
50 0.10 ± . − − Chandra c +14 − +0 . − . +35 − ± .
05 6666 +1 − +0 . − . a Hwang et al. 2000a b Miceli et al. 2006 c this paper
14 –Table 4: The K shell line energies of Cr and Fe in different ionization states.element neutral a B-like b Be-like b Li-like b He-like b H-like c Cr † ‡ a K α line energy b ine energy with transition from level 2p to 1s c y α line energy † ‡ Mewe et al. 1985
15 –
Fig. 1.—
Chandra images of W49B, Cas A, Tycho and Kepler. The source spectral regionsare overplotted. In order to cover the whole Tycho SNR, we use multi-regions as shownin the stretch. The double-arrowed lines represent 1 arcmin for the corresponding image.The two boxes in the Cas A image represent the blueshift (southeast, blue one) and redshift(northwest, red one) dominated portions we selected to create the spectra as described in § Cas AKeplerTychoW49B
Fig. 2.— The 0.5 − − χ ) distributions are also plotted.The strength of all the Gaussian components are multiplied by a factor of 15, so as to beshown clearly. 18 – Tycho Cas AKeplerW49B
Fig. 4.— The 5.0 − §