An XMM-Newton Observation of the Seyfert 2 Galaxy NGC 6300. I. The Nucleus

a r X i v :a s t r o -p h /0305239v 1 14 M a y 2003Mon.Not.R.Astron.Soc.000,000–000(0000)Printed 2February 2008(MN L A T E X style file v2.2)XMM-Newton spectral properties of the Narrow-Line Seyfert 1galaxy IRAS 13224–3809Th.Boller,1Y .Tanaka,1,2A.Fabian,3W.N.Brandt,4L.Gallo,1N.Anabuki,2Y .Haba 2and S.Vaughan 31Max-Planck-Institut f¨u r extraterrestrische Physik,Postfach 1312,85741Garching,Germany 2Institut of Space and Astronautical Science,3-1-1Yoshinodai,Sagamihara,Kanagawa 22,Japan 3Institute of Astronomy,Madingley Road,Cambridge CB30HA4Department of Astronomy and Astrophysics,Pennsylvania State University,525Davey Lab,University Park,PA 16802,USAReceived 21March 2003;Accepted 13May 2003ABSTRACTWe report on the first XMM-Newton observation of the highly X-ray variable,radio-quiet,Narrow-Line Seyfert 1galaxy IRAS 13224–3809obtained during the guaranteed time pro-gramme with a 64ks exposure.The most remarkable spectral feature is a sharp drop,by a factor ∼5,in the spectrum at 8keV .This is a similar,but stronger,feature to that which we found in 1H 0707–495.Significant flattening of the hard X-ray spectrum occurs when the source flux decreases.The flattening of the spectrum can be modelled as an increase in the column density of the absorbing material and/or its covering fraction.Below ∼1.5keV the spectrum is dominated by a giant soft X-ray excess,and at around ∼1.2keV there is a sig-nificant absorption feature detected,most probably due to ionized Fe L absorption.The new X-ray spectral properties detected with XMM-Newton in IRAS 13224–3809support a partial covering interpretation,i.e.the presence of dense material inside the accretion-disc region par-tially obscuring the emission from the accretion disc.However,the sharpness of the feature,if due to photoelectric absorption,is surprising and may require an alternative explanation.One possibility which does fit the whole spectrum is that it is dominated by ionized reflec-tion rather than absorption.The unusual spectral properties detected with XMM-Newton from Narrow-Line Seyfert 1galaxies increase the known spectral complexity of active galactic nu-clei and should further stimulate a combined theoretical and observational effort to achieve a better understanding of the physics of the innermost regions of AGN.Key words:galaxies:active,AGN –galaxies:individual:IRAS 13224–3809–X-rays:galax-ies1INTRODUCTIONROSAT ,ASCA and XMM-Newton have shown many Narrow-Line Seyfert 1galaxies (hereafter NLS1s;see Osterbrock &Pogge 1985and Goodrich 1989)to have remarkable X-ray properties com-pared to Seyfert 1galaxies with broader Balmer lines.In X-rays NLS1s are generally characterized by strong soft X-ray excesses,steep hard power-law continua and extreme X-ray variability (e.g.Puchnarewicz et al.1992,Boller,Brandt &Fink 1996,Brandt &Boller 1998,Vaughan et al.1999,Brandt,Mathur &Elvis 1997).With the arrival of XMM-Newton spectra a new feature was added:a sharp spectral drop above 7keV without any significant Fe-K line emission in the NLS11H 0707–495(Boller et al.2002;Fabian et al.2002).In the case of 1H 0707–495two physical interpreta-tions were discussed:(i)a partial covering scenario (following Rees 1987;Celotti,Fabian &Rees 1992;Brandt &Gallagher 2000)in which the sharp drop is interpreted as Fe absorption in high column density clouds (Boller et al.2002)and (ii)a reflection-dominatedaccretion-disc spectrum (Fabian et al.2002).Both models provide acceptable spectral fits to the 1H 0707–495data.The very high iron overabundance in the Boller et al.(2002)interpretation was reduced to a more realistic value by the reanalysis of Tanaka et al.(2003).IRAS 13224–3809is one of the most interesting members of the NLS1class due to its remarkable X-ray variability,huge soft X-ray excess (Boller,Brandt &Fink 1996)and high optical Fe II to H βline ratio (Boller et al.1993).Indeed,IRAS 13224–3809is among the most X-ray variable Seyfert 1galaxies known.The first systematic monitoring in soft X-rays in 1996showed persistent,rapid,giant-amplitude count rate variability.Over the course of the observations the maximum observed amplitude of variability was of a factor of ∼60,and a variation by about a factor of 57in just two days was observed.The ionizing luminosity rises from about 1.5×1043erg s −1to about 8.3×1044erg s −1,roughly equivalent to a typical Seyfert 1like MCG–6-30-15abruptly increasing its soft2Th.Boller etal.Figure1.EPIC pn spectrum of IRAS13224–3809.A simple power-lawmodel,fitted only in the2–7keV range,has been extrapolated into the softand hard bands to illustrate the sharp spectral feature at∼8keV,and thesoft X-ray excess emission.The photon index in the2–7keV band is(1.9±0.1).X-ray luminosity to become almost as powerful as a quasar(Bolleret al1997).Due to its extreme X-ray spectral and timing properties,IRAS13224–3809was observed during the XMM-Newton guaranteedtime programme.In this paper we discuss the results obtained fromthe64ks observation.2X-RAY OBSERV ATIONS AND DATA ANALYSISIRAS13224–3809was observed with XMM-Newton(Jansen et al.2001)on2002January19during revolution0387for about64ks.During this time all instruments were functioning normally.TheEPIC pn camera(Str¨u der et al.2001)was operated in full-framemode,and the two MOS cameras(MOS1and MOS2;Turner etal.2001)were operated in large-window mode.All of the EPICcameras used the mediumfilter.The two Reflection Grating Spec-trometers(RGS1and RGS2;den Herder et al.2001)also gathereddata during this time.The Observation Data Files(ODFs)were pro-cessed to produce calibrated event lists using the XMM-Newton Sci-ence Analysis System(SAS v5.3).Unwanted hot,dead orflick-ering pixels were removed as were events due to electronic noise.Event energies were corrected for charge-transfer losses.The latestavailable calibrationfiles were used in the processing.Light curveswere extracted from these event lists to search for periods of highbackgroundflaring.A significant backgroundflare was detected inthe EPIC cameras approximately20ks into the observation andlasted for5ks.This segment was removed and ignored duringthe analysis.The total amount of good exposure time selected was55898s and58425s for the pn and MOS detectors,respectively.The source plus background photons were extracted from a circularregion with a radius of35arcsec,and the background was selectedfrom an off-source region and appropriately scaled to the sourceregion.Single and double events were selected for the pn detector,and single-quadruple events were selected for the MOS.High reso-lution spectra were obtained with the RGS.The RGS were operatedin standard Spectro+Q mode for a total exposure time of63963s.Thefirst-order RGS spectra were extracted using RGSPROC,andthe response matrices were generated using RGSRMFGEN.TheOptical Monitor collected data through the UVW2filter(1800–Figure2.Simple power-law plus edgefit to the2–12keV energy range.The edge parameters areτ=1.6+0.9−0.5and E=8.2±0.1keV.2250˚A)for about thefirst25ks of the observation.The OM dataare discussed in Gallo et al.(2003,in preparation).In the followinganalysis we use the EPIC pn data to constrain the sharp spectralfeature at8keV,as they contain the highest photon statistics.Theenergy range up to about12keV can be explored with EPIC pnwhile above that value the spectrum is background dominated.Thecombined MOS1and MOS2spectrum is affected by high back-ground at energies greater than6keV.Above8keV,we collect atotal of120source plus background counts,while the number ofbackground photons,normalized to the source cell size,is68.3X-RAY SPECTROSCOPY3.1Mean spectral properties3.1.1Discovery of a sharp spectral featureIn Fig.1we show the time-averaged spectrum of IRAS13224–3809.We havefitted the2–7keV energy band with a simple power-law,which is extrapolated into the soft and hard bands for illustra-tion purposes.The most obvious features are the strong soft X-rayexcess emission below about1.5keV and the presence of a sharpspectral drop at∼8keV.If the2–12keV spectrum isfitted by asimple power-law plus edge model,the edge energy is8.2±0.1keV(Fig.2).The energy of the edge is significantly different from thatof the neutral iron absorption edge at7.1keV.Theflux absorbedby the edge is4.6×10−14erg cm−2s−1.The edge cannot be con-strained by the MOS data as the spectrum is background dominatedabove about6keV.Some positive residuals can be seen at around6.8keV(in therest frame of the object).They are consistent with emission fromH and He-like iron with a total equivalent width of about200eV.The detection of this emission,judged from an F-test,is marginal(at about the2.5σconfidence level).The strength of the line,whencompared with theflux of photons which is absorbed and the likelyfluorescent yield(Krolik&Kallman1987),indicates that the ab-sorber covers a solid angleΩ/4π∼<0.2.XMM-Newton spectral properties of the Narrow-Line Seyfert 1galaxy IRAS 13224–38093Figure 3.Unfolded spectrum modelled with a partially covered power-law and a disc blackbody.The upper curve gives the absorption-corrected spec-tral model (see text for details).3.1.2A partial covering modelThe observed spectrum shows little low-energy absorption and is consistent with interstellar absorption only.On the other hand,there is what appears to be a deep absorption edge at 8.1−8.2keV which would correspond to the K-edge of Fe XVIII –XX .The sharpness of the edge requires that the ionization states of Fe are confined to a narrow range.The absence of strong (narrow)emission at ∼6.6keV is in contradiction with an explanation in terms of a distant reflector.The observed features can be explained in terms of ab-sorption if allowance is made for partial covering (i.e.a patchy ab-sorber).A partial-covering model was therefore fitted to the data.The model comprised a multi-component disc (MCD)model for the soft component,and a partially covered power-law for the hard continuum.In addition,this model required a further absorption feature around ∼1.2keV (see Section 3.1.3),which is included by subtracting a Gaussian line.A good fit was obtained (χ2/ν=510/471)using two absorbed power-laws,but a very steep slope,Γ≃3.4was required,which appears extremely high given the the present values found in AGN (c.f.Fig.1of Brandt,Mathur &Elvis 1997).Allowing for an exponential cut-off in the power-laws did not significantly improve the fit (χ2/ν=509/470)but gave a more reasonable spectral slope.This model is illustrated in Fig.3.The best-fit parameters were as follows:the MCD (colour)temperature kT =0.16keV ,Γ=2.0,E cut =4.4keV ,and the column densities (and covering fractions)of the line-of-sight ab-sorbers were N H ≈1.2×1022cm −2(0.06)and ≈1.5×1023cm −2(0.77).The Fe abundance was about 10×solar with a 90per cent confidence lower limit of 2×solar.The average effective slope of the cut-off power-law over 1−5keV is approximately Γ=2.5.The temperature and (absorption-corrected)luminosity of the disc require that its inner radius is only about one gravitational ra-dius for a 106M ⊙black hole.Assuming the soft X-ray excess is physically described by a MCD model then the object would be super-Eddington when bright.3.1.3Detection of Fe L absorptionA broad absorption feature at E =1.19±0.02keV ,with an in-trinsic width σ≈0.12keV ,is required in the above model with high significance (Fig.4).The equivalent width is approximatelyEW =120eV .The absorbed flux is 1.9+0.02−0.04×10−13erg cm −2Figure 4.Spectral fit to the soft energy spectrum obtained with a MCD model plus partially covered cut-off power-laws.Strong residuals are appar-ent around 1.2keV ,most possibly due to Fe-L absorption (light data points)Including a Gaussian line removes these residuals (dark data points).The significance of the detection is shown in the lower panel in terms of sigmas.s −1.The RGS1spectrum is consistent with,but does not constrain,the presence of an absorption feature (no data were collected from RGS2in that energy range).Other features in the RGS spectra can not be constrained with the available photon statistics.This deep absorption feature is similar to that reported by Leighly et al.(1997)from the ASCA spectrum.Those authors in-terpreted the feature in terms of resonant absorption oxygen from a relativistic outflow.A more prosaic explanation was put forward by Nicastro,Fiore &Matt (1999)who interpreted the feature as reso-nance absorption,primarily by L-shell iron.Their analysis gave a column density of the absorbing material of log N H =23.5and an ionization parameter of log U =1with an outflowing and dis-persion velocity of 1000km s −1(see also their Fig.4).Such an absorption model does not simply explain the large observed drop at 8keV .3.1.4An alternative discline explanationIn the case of 1H 0707–495(Boller et al.2002)we noted that the sharp drop at 7.1keV could be the blue wing of a strong,relativistically-blurred iron emission line.This would require some unusual geometry or situation in order that such a strong line is produced;the spectrum needs to be reflection-dominated.Fabian et al.(2002)discuss one possibility involving a corrugated inner accretion disc;an alternative would be gravitational light bending (e.g.Martocchia,Matt &Karas 2002;Fabian &Vaughan 2003).The spectrum of IRAS 13224–3809can also be fitted by such a model (see Fig 5).Fitting the 2–12keV spectrum with a power-law continuum plus a emission line with a Laor (1991)profile worked well with an emissivity index q =6.2±1.2,inner and outer radiiof r in =1.3+0.2−0.1and r out >180gravitational radii for the disc,and an inclination of i ≈60±3deg.The disc inclination and emis-sivity were covariant in the fitting,and a wide range of inclination angles was possible.The equivalent width of the line is unusually high at 5.6keV .To see whether such a strong line can be produced,we used a grid of ionized slab models produced by D.Ballantyne (using the code described by Ross &Fabian 1993).With the abun-dances of metals at 7times solar and Galactic absorption we obtain a fit over the entire range from 0.4−12keV with a χ2ν=1.3.Some sharp deviations at the 2–3σlevel occur at ∼6keV and again at ∼8keV .As with 1H 0707–495this model requires the X-ray spec-4Th.Boller etal.Figure5.Ratio of the data to a power-lawfitted over the1.3–3keV and8–12keV bands.The deviations can be seen as a very strong emissionline.Figure6.Broad-band(0.3−12keV)pn light curve of IRAS13224–3809.The bin size is200s.Zero seconds on the time axis marks the start ofthe observation at03:15:02on2002-01-19.The light curve is divided intotwo high-flux states,a medium state,and a low state.The marked intervalcontaining the backgroundflare was omitted.trum to be reflection-dominated.The sharp jump in the spectrum at∼1keV is due to strong line and recombination emission by Fe-L,O,and other elements.3.2Flux-dependent spectral propertiesThe0.3−12keV light curve of IRAS13224–3809shows strongand rapid variability withflux changes by about a factor of10(Fig.6).We have divided the light curve into four differentfluxstates:(i)two high states with count rates above2.5counts s−1(time intervals less than20ks and greater than30ks after the be-ginning of the observations);(ii)a medium state with count ratesbetween1.5and2.5counts s−1;and a(iii)low state below1.5counts s−1.Evidence in support of partial covering is obtained from acomparison of the spectra between a high-flux level and a low-flux level.The high-flux and low-flux spectra are constructed fromthose intervals were the source count rate is 2.5counts s−1and1.5counts s−1,respectively.The averagefluxes of the two aredifferent by a factor of three.These two spectra werefitted simul-taneously with the same spectral model as in section3.1.2(i.e.aMCD and a cut-off power-law model,and two absorbedpower-lawFigure7.Unfolded spectralfits to the high-and low-flux data(middle andlower curve,respectively).Note theflattening of the hard spectrum in thelowerflux state.Theflattening can be explained in the partial covering sce-nario by an increased column density and/or increased covering fractionof the absorbing material.The upper curve gives the absorption correctedspectral model.components).The Fe abundance wasfixed at3×solar.A goodfit(χ2/ν=667/631)was obtained,as shown in Fig.7.The bestfitting parameters were as follows:kT=0.16keV,Γ=2.0,E cut=4.0keV.The column densities(and covering fractions)of the absorbers for the high-flux spectrum were N H≈1.0(0.66)and≈20(0.53),and those for the low-flux spectrum were≈3.1(0.53)and≈23(0.87),where the column is in units of1022cm−2.Thus when theflux dropped the absorption increased.4DISCUSSIONIRAS13224–3809shows an unusual combination of spectral fea-tures:(i)a sharp spectral drop by a factor of∼5at∼8keV,(ii)marginal Fe K emission lines,(iii)aflattening of the spec-trum above3keV towards lowflux states,(iv)strong soft X-rayexcess emission,(v)a significant Fe-L absorption feature at around1.2keV.The most plausible explanation for the spectral features(i)to(iii)appears to be provided by a partial-covering scenario.Thesoft X-ray excess is most probably due to a higher temperatureof the accretion disc as often seen in NLS1s.Presently it is notclear whether the Fe-L absorption feature is physically connectedwith the partial covering scenario.The lack of strong emission linescan be explained if the absorber subtends only a small solid angleas seen from the central black hole.An immediate consequence,given the variability in the absorber,is that the absorber must belocated close to the central black hole,maybe even within the ac-cretion disc region as suggested by e.g Rees(1987),Celotti,Fabian&Rees(1992)and Brandt&Gallagher(2000).Theflattening ofthe spectrum provides strong evidence for partial covering,as anincreased absorbing column and/or a larger covering fraction is ex-pected when the sourceflux is low.Although the partial coverermodel explains the data well,the present photon statistics abovethe edge still do not allow us to constrain precisely the iron abun-dance,nor can we disentangle the unique contribution of the cov-ering fraction or the absorbing column.The sharpness of the feature,if due to a photoelectric edge,is surprising.At8.2keV the absorbing matter must be partiallyionized(the threshold energies for Fe XIX–XXIII are7.93,8.07,XMM-Newton spectral properties of the Narrow-Line Seyfert1galaxy IRAS13224–380958.21,8.35and8.49keV)in which case a range of ionization states is expected.These make the observed edge broad rather than sharp (e.g.Palmeri et al.2002).Alternatively the feature may be a neutral edge in approaching matter(say in a wind or jet along the line of sight,e.g.Chartas et al.2002).The velocity of the matter is then required to be0.15c.A further possibility if the absorption is by approaching matter is that the feature is the start of a trough due to resonance absorption(e.g.Krolik et al.1985,with Fe XXV/XXVI instead of O VIII).There is then a significant outflow of mass and kinetic energy from the object.The extreme spectral properties detected with XMM-Newton increase the known spectral complexity in active galactic nuclei and should further stimulate a combined theoretical and observational effort to improve our understanding of the physics of the innermost regions of AGN.To more precisely constrain the form of the sharp spectral feature in IRAS13224–3809and distinguish between dif-ferent models we need a much stronger signal above8keV.This may be possible if the source is observed during one of its giant amplitudefluctuations.ACKNOWLEDGMENTSBased on observations obtained with XMM-Newton,an ESA sci-ence mission with instruments and contributions directly funded by ESA Member States and the USA(NASA).WNB acknowledges support from NASA LTSA grant NAG5-13035. REFERENCESBoller Th.,Tr¨u mper J.,Molendi S.,Fink H.,Schaeidt S.,Caulet A.,Den-nefeld M.,1993,A&A279,53Boller Th.,Brandt W.N.,Fink H.,1996,A&A,305,53Boller Th.,Brandt W.N.,Fabian A.C.,Fink H.1997,MNRAS289,393 Boller Th.,Fabian A.C.,Sunyaev R.,Tr¨u mper,J.,Vaughan S.,BallantyneD.R.,Brandt W.N.,Keil R.,Iwasawa K.,2002,MNRAS329,1 Brandt W.N.,Boller Th.,1998,AN319,3,163Brandt W.N.,Mathur S.,Elvis M.,1997,MNRAS,285,25Brandt W.N.,Gallagher S.C.,2000,New Astronomy Reviews44,461 Celotti A.,Fabian A.C.,Rees M.J.,1992,MNRAS255,419Chartas G.,Brandt W.N.,Gallagher S.C.,Garmire,G.P.,2002,ApJ579, 169Fabian A.C.,Ballantyne D.R.,Merloni,A.,Vaughan S.,Iwasawa K., Boller Th.,2002,MNRAS331,35Fabian A. C.&Vaughan,S.,2003,accepted for MNRAS (astro-ph/0301588)Goodrich,R.W.,1989,ApJ342,234den Herder,J.W.;Brinkman,A.C.;Kahn,S.M.;Branduardi-Raymont,G.;Thomsen,K.;Aarts,H.;Audard,M.;Bixler,J.V.;den Boggende,A.J.;Cottam,J.et al.2001,A&A365,7Jansen F.,Lumb D.,Altieri B.,Clavel J.,Ehle M.,Erd C.,Gabriel C., Guainazzi M.,Gondoin P.,Much R.et al.,2001,A&A365,1 Krolik,J.H.,Kallman,T.R.,Fabian,A.C.,Rees,M.J.,1985,ApJ295, 104Laor,A.,1991,ApJ376,90Leighly K.,Mushotzky R.F.,Nandra K.,Forster K.,1997,ApJL489,25 Martocchia,A.,Matt,G.,Karas,V.,2002,A&A383,23Mason,K.O.,Breeveld A.,Much R.,Carter M.,Cordova F.A.,Cropper M.S.,Fordham J.,Huckle H.,Ho C.,Kawakami H.et al.2001,A&A365, 36Nicastro F.,Fiore F.,Matt G.,1999,ApJ517,108Osterbrock,D.E.&Pogge,R.W.,1985,ApJ297,166Palmeri,P.,Mendoza,C.,Kallman,T.R.,Bautista,M.A.,2002,ApJ577, L119Puchnarewicz E.M.,Mason K.O.,Cordova F.A.,Kartje J.,Branduardi-Raymont G.et al.1992,MNRAS,256,589Rees M.,1987,MNRAS228,47Ross,R.R.&Fabian,A.C.,1993,MNRAS261,74Str¨u der L.,Briel U.,Dennerl K.,Hartmann R.,Kendziorra E.,Meidinger N.,Pfeffermann E.,Reppin C.,Aschenbach B.,Bornemann W.et al., 2001,A&A36518Tanaka Y.,Boller Th.,Ueda,Y.,Keil R.,2003,submitted to MNRAS Turner M.J.L.,Abbey A.,Arnaud M.,Balasini M.Barbera M.,Belsole,E., Bennie P.J.,Bernard J.P.,Bignami G.F.,Boer M.et al.,2001,A&A 365,27Vaughan S.,Reeves J.,Warwick R.,Edelson R.,1999,MNRAS309,113 This paper has been typeset from a T E X/L A T E Xfile prepared by the author.。

a r X i v :a s t r o -p h /0503240v 1 10 M a r 2005XMM–Newton EPIC &OM Observations of Her X-1over the 35d Beat Period and anAnomalous Low StateS.Zane ∗,G.Ramsay ∗,Mario A.Jimenez-Garate †,Jan Willem den Herder ∗∗,Martin Still ‡,Patricia T.Boyd ‡and Charles J.Hailey §∗Mullard Space Science Laboratory,University College of London,Holmbury St Mary,Dorking,Surrey,RH56NT,UK†MIT Center for Space Research,77Massachusetts Avenue,Cambridge,MA 02139,USA ∗∗SRON,the National Institute for Space Research,Sorbonnelaan 2,3584CA Utrecht,TheNetherlands‡NASA/Goddard Space Flight Center,Code 662,Greenbelt,MD 20771§Columbia Astrophysics Laboratory,Columbia University,New York,NY 10027,USAAbstract.We present the results of a series of XMM–Newton EPIC and OM observations of Her X-1,spread over a wide range of the 35d precession period.We confirm that the spin modulation of the neutron star is weak or absent in the low state -in marked contrast to the main or short-on states.The strong fluorescence emission line at ∼6.4keV is detected in all observations (apart from one taken in the middle of eclipse),with higher line energy,width and normalisation during the main-on state.In addition,we report the detection of a second line near 7keV in 10of the 15observations taken during the low-intensity states of the system.We discuss these observations in the context of previous observations,investigate the origin of the soft and hard X-rays and consider the emission site of the 6.4keV and 7keV emission lines.INTRODUCTIONHer X-1is one of the best studied X-ray binaries in the sky.The binary system consists of a neutron star and an A/F secondary star.It has an orbital period of 1.7d and the neutron star spin period is ∼1.24s.It is one of only a few systems which shows a regular variation in X-rays,over a “beat”period of 35d,which is generally interpreted as the precession of an accretion disk that periodically obscures the neutron star beam.The cycle comprises:i)a 10d duration main-on state,ii)a fainter 5d duration short-on state,and iii)a period of lower emission in between.Exceptions to the normal 35d cycle has been observed only in four occasions:in 1983,1993,1999and in January 2004the turn-on of the source has not been observed.During these "anomalous low states"(ALS;the period of which ranges from several months to 1.5years)Her X-1appears as a relatively faint X-ray source,with a strength comparable to that of the standard low state.The event registered last year was only the fourth one that has been seen since the discovery of Her X-1in 1972,and the first one that could be observed with a high capability satellite as XMM–Newton .Her X-1has been observed by XMM–Newton on 15separate occasions outside the ALS,giving good coverage over the beat period.Moreover,it has been observed 10φ35=0.17φ35=0.26Phase Shift (degrees)-1.0φ35=0.60180270090180-0.50.00.5C r o s s C o r r e l a t i o nφ35=0.02FIGURE 1.Cross correlation of the (0.3-0.7keV)and (2-10keV)light curves at four different Φ35.times during the ALS.The analysis of datasets taken before the ALS has been presentedby Ref.[1],[2]and [3].Here we report our main findings,referring the interested reader to the above papers for more details,and we present a preliminary analysis of the ALS.TIMING ANALYSISWe first focussed on datasets taken before the source entered the anomalous low state.We performed a search for pulsations in all datasets,confirming that,in marked contrast to the main or short-on states,the spin modulation of the neutron star is weak or absent in the low state.During the states of higher intensity,we observe a substructure in the broad soft X-ray modulation below ∼1keV ,revealing the presence of separate peaks which reflect the structure seen at higher energies (see Ref.[3]).The soft and hard X-ray lightcurves of Her X-1are known to be shifted in phase (see Fig.2,central panel).Under the assumption that soft X-rays are due to the reprocessing of the pulsar beam by the inner edge of the disk,this is usually interpreted as evidence for a tilt angle in the disk[4],[5].In our XMM–Newton data,we find the first evidence for a substantial and systematic change in the phase difference along the beat cycle,which is predicted by precessing disk models[6](see Fig.1).THE FE K αLINEThe strong emission line at ∼6.4keV is detected in all our XMM–Newton observations,with larger broadening and normalization during the main-on (see Fig.2,left panel).The line centroids observed using the EPIC PN deviate by 4σfrom the 6.40keV neutral value:the Fe line emission originates in near neutral Fe (Fe XIV or colder)in the low and short-on state observations,whereas in the main-on the observed Fe K αcentroid energies (6.65±0.1keV and 6.50±0.02keV at Φ35=0.02and 0.17)correspond to Fe XX-Fe XXI [7].Possible reasons for this behaviour may be:1)an array of Fe K αfluorescence lines exists for a variety of charge states of Fe (anything up to Fe XXIII);2)Comptonization from a hot corona for a narrower range of charge states centered around Fe XX;3)Keplerian motion.The Keplerian velocity measured at Φ35=0.02and 0.17is ∼15500A S M C t /sK e VE W e VS i g m a e VN o r mK e VE W e V35day phase1.0•102.0•103.0•104.0•10C t /s1001502002500.3-0.7keVC t /s501001502002-6keV0.010 N o r m a l i s a t i o n0.0000.0020.0040.0060.008Spin Phase0E W0.00.51.0 1.52.0100200300400U V W 1 c t /sU V W 2 c t /sOrbital PhaseF e K a l p h a N o r mFIGURE 2.Left panel:Variation of the K αand Fe XXVI line parameters along the beat cycle.From the top:1)mean ASM light curve;2)-5)central energy,equivalent width (EW),width and normalization of the prominent K αFe emission line;6)-8)central energy,EW and normalization of the Fe line at ∼7keV .When the second feature is undetected,we show the 90%confidence level upper limits to the EW and normalization (“v”symbols).Central panel:The observation made at Φ35=0.02.From the top panel:the light curve in the 0.3-0.7keV and 2-6keV band (note the shift in phase between hard and soft X-ray emission);the Fe K αline normalisation and EW as a function of the spin phase.Right:The UVW1(top panel),UVW2(central panel)data and the Fe 6.4keV normalisation folded on the orbital period.The blue circles superimposed in the left and right panels are data taken during the 2004ALS.and ∼13000km/s,respectively.This gives a radial distance of ∼2−3×108cm,which is close to the magnetospheric radius for a magnetic field of ∼1012G.However,another possibility is that the region responsible for the Fe K αline emission is different for lines observed at different beat phases.In fact,while data taken during the main-on clearly indicate a correlation between the fluorescent Fe K αline and the soft X-ray emission (Fig.2,central panel),suggesting a common origin in the illuminated hot spot at the inner edge of the disk,the same is not explicitly evident in data taken during the low state.Instead,at such phases the Fe K αline is a factor >5weaker and is clearly modulated with the orbital period (see again Fig.2,right panel).The correlation between the fast rising UVW1flux and the Fe K αdetected outside the main-on point to79.06897.05.06.08.0 FIGURE3.The spectral region around6.6keV fromΦ35=0.02(top)andΦ35=0.79(bottom).In both panel a solid line shows the bestfit after the normalisation of the one or two Gaussian components have been set to zero.a common origin,possibly in the disk and/or illuminated companion.A fraction of the Fe Kαemission may arise from relatively cold material in a disk wind,such as commonly observed in cataclysmic variables[8].However,we do not detect the Fe Kαline during the middle of the eclipse,and the upper limit on the line flux is∼10or less of that measured outside the eclipse.Also,there is no Doppler signature of a wind in the HETG spectrum of the Fe Kαline[10].On the other hand, the data reported here suggest a complex origin for the overall emission of the Fe Kαline.To our knowledge,a complex of lines which include all ionization states from Fe XVIII to XXIV Fe Kαhas not been observed in any astrophysical source.This may still indicate an outflow of relatively cold gas or some complex dynamics in the disk/magnetosphere interface.Such phenomena should be time-dependent and may be monitored in the future using Astro-E2.THE∼7KEV FE LINEXMM–Newton data have revealed,for the veryfirst time for this source,the presence of a second Fe line at∼7keV.The feature is only detected during the low and short-on states,and over several beat phases(see Fig.3).Also,it has been confirmed by a Chandra HETGS observation of the source(the only one made during the low state) taken atΦ35=0.44−0.46[10].The feature cannot be produced byfluorescence,and it is more likely to be a Fe XXVI line originating in widely extended photo-ionized plasma.This is consistent with the fact that also RGS data taken during the low and short-on states show the presence of photo-ionized gas[2].Grating spectra exhibit several narrow recombination emission lines,the most prominent being C VI,N VI,N VII,O VII,O VIII and Ne IX.The line ratio G= (f+i)/r,as computed for all the helium-like ion complexes,is G≃4,which indicates that photoionization is the dominant mechanism.Moreover,RGS spectra shows two radiative recombination continua of O VII and N VII,consistent with a low temperature of the emitting plasma(30000K<T<60000K)[2].None of these features is detected during the main-on state.The recombination X-ray line emission are not likely to originate in HZ Her,due to the absence of UV induced photoexcitation signatures in the He-like triplets(observed with HETGS)[10].Instead,we propose that an extended,photoionized accretion disk atmosphere may be responsable for such features[2].The evidence for the disk identi-fication relies on the modeled structure and spectra from a photoionized disk,which is in agreement with the limit set spectroscopically on the density of the low-energy lines emitting region.This theoretical model has been developed by Ref.[9],who computed the spectra of the atmospheric layers of a Shakura-Sunyaev accretion disk,illuminated by a central X-ray continuum.They found that,under these conditions,the disk de-velops both an extended corona which is kept hot at the Compton temperature,and a more compact,colder,X-ray recombination-emitting atmospheric layer(see Figure9in Ref.[2]).Interestingly,wefind that the Fe XXVI line detected by XMM–Newton may be a signature of the hottest external layers of the disk corona,which are located above the recombination-emitting layers.Again,the computed values of the density agree with the constraints inferred from the7keV line parameters[3].In summary,most of the spectral lines discovered with EPIC and RGS can be associ-ated with the illuminated atmosphere/corona of the accretion disk,which explains why they are more prominent during the low states when the direct X-ray beam from the pul-sar is obscured by the accretion disk.Therefore,the variability of the Her X-1spectrum lends support to the precession of the accretion disk,strengthen the interpretation of the low state emission in term of an extended source and open the exciting possibility to monitor spectroscopically the different atmospheric components of the disk during the transition from the low to the high state.THE ANOMALOUS LOW STATEAnomalous low states are rare and peculiar events,during which the source resides in a deep low state.While the mechanism that forces state changes is almost certainly variations in accretion disk structure,the engine ultimately driving structural evolution remains unknown.Each past anomalous low state,including the most recent one of January2004[11],has been preceded by a period of enhanced spin-down,that has been interpreted in terms of an increasing torque leading to a reversal in the rotation of the inner disk.This also implies that the onset of an anomalous low state is accompanied by a large variation in the structure of the inner region of the accretion disk,that becomes increasingly warped.In order to search for residual evidence of a35d cycle,we compared the line emission detected during the ALS with that observed in the several XMM–Newton datasets we have accumulated during the standard35d cycle.As we can see from Fig.2,as far as the Fe complex is concerned,there is no an obvious difference in the spectral properties of the ALS and of the standard low states.The line features are consistent with being the same in these epoches and the orbital variation of the KαFe line shows the same correlation with the UVW1.This supports our scenario in which the Fe line emission of the low state originates in an extended component(disk atmosphere/corona and/orcompanion)instead that in the inner region of the disk.Moreover,it is consistent with the fact that at higher energies a significant Compton reflection component has been detected by RXTE in the spectrum of the ALS([12]).A detailed comparison of the recombination emission lines measured by RGS can shed more light on this issue.If our interpretation is correct,by measuring the line ratios during the anomalous low state allows to infer the ionization state of the plasma,column density and optical depth in the visible portion of the disk atmosphere/corona,ultimately constraining the accretion geometry.REFERENCES1.Ramsay,G.,et al.,MNRAS,337,1185,(2002).2.Jimenez-Garate,M.A.,et al.,ApJ,578,391(2002)3.Zane,S.,et al.,MNRAS,350,506(2004)4.Oosterbroek,T.,et al.,A&A,327,,215(1997)5.Oosterbroek,T.,et al.,A&A,353,,5755(2000)6.Gerend,B.,&Boynton,P.ApJ,209,,562(1976)7.Palmeri,P.,et al.,A&A,403,1175(2003)8.Drew,J.IAU colloquium197,Accretion Phenomena and Related Outflows,editeb by D.T.Wickra-masinghe,L.Ferrario,&G.V.Bicknell(ASP Conf.Ser.,121;San Francisco:ASP),1997,pp.4659.Jimenez-Garate,M.A.,et al.,ApJ,558,448(2001)10.Jimenez-Garate,M.A.,et al.,ApJ accepted(2005),astro-ph/041178011.Still,M.&Boyd,P.,ApJ,606,L135(2004)12.Still,M.&Boyd,P.,American Astronomical Society,HEAD meeting#8,#33.02,(2004)。

a r X i v :0712.3538v 1 [a s t r o -p h ] 20 D e c 2007Draft version February 2,2008Preprint typeset using L A T E X style emulateapj v.10/09/06COMPARING SUZAKU AND XMM-NEWTON OBSERVATIONS OF THE SOFT X-RAY BACKGROUND:EVIDENCE FOR SOLAR WIND CHARGE EXCHANGE EMISSIONDavid B.Henley and Robin L.SheltonDepartment of Physics and Astronomy,University of Georgia,Athens,GA 30602Draft version February 2,2008ABSTRACTWe present an analysis of a pair of Suzaku spectra of the soft X-ray background (SXRB),obtained from pointings on and offa nearby shadowing filament in the southern Galactic hemisphere.Because of the different Galactic column densities in the two pointing directions,the observed emission from the Galactic halo has a different shape in the two spectra.We make use of this difference when modeling the spectra to separate the absorbed halo emission from the unabsorbed foreground emission from the Local Bubble (LB).The temperatures and emission measures we obtain are significantly different from those determined from an earlier analysis of XMM-Newton spectra from the same pointing directions.We attribute this difference to the presence of previously unrecognized solar wind charge exchange (SWCX)contamination in the XMM-Newton spectra,possibly due to a localized enhancement in the solar wind moving across the line of sight.Contemporaneous solar wind data from ACE show nothing unusual during the course of the XMM-Newton observations.Our results therefore suggest that simply examining contemporaneous solar wind data might be inadequate for determining if a spectrum of the SXRB is contaminated by SWCX emission.If our Suzaku spectra are not badly contaminatedby SWCX emission,our best-fitting LB model gives a temperature of log(T LB /K)=5.98+0.03−0.04and a pressure of p LB /k =13,100–16,100cm −3K.These values are lower than those obtained from other recent observations of the LB,suggesting the LB may not be isothermal and may not be in pressure equilibrium.Our halo modeling,meanwhile,suggests that neon may be enhanced relative to oxygen and iron,possibly because oxygen and iron are partly in dust.Subject headings:Galaxy:halo—Sun:solar wind—X-rays:diffuse background—X-rays:ISM1.INTRODUCTIONThe diffuse soft X-ray background (SXRB),which is observed in all directions in the ∼0.1–2keV band,is composed of emission from several different components.For many years,the observed 1/4-keV emission was believed to originate from the Local Bubble (LB),a cavity in the local interstellar medium (ISM)of ∼100pc radius filled with ∼106K gas (Sanders et al.1977;Cox &Reynolds 1987;McCammon &Sanders 1990;Snowden et al.1990).However,the discovery of shadows in the 1/4-keV background with ROSAT showed that ∼50%of the 1/4keV emission originates from beyond the LB (Burrows &Mendenhall 1991;Snowden et al.1991).This more distant emission origi-nates from the Galactic halo,which contains hot gas with temperatures log(T halo /K)∼6.0–6.5(Snowden et al.1998;Kuntz &Snowden 2000;Smith et al.2007;Galeazzi et al.2007;Henley,Shelton,&Kuntz 2007).As the halo gas is hotter than the LB gas,it also emits at higher energies,up to ∼1keV.Above ∼1keV the X-ray background is extragalactic in origin,and is due to un-resolved active galactic nuclei (AGN;Mushotzky et al.2000).X-ray spectroscopy of the SXRB can,in principle,de-termine the thermal properties,ionization state,and chemical abundances of the hot gas in the Galaxy.These properties give clues to the origin of the hot gas,which is currently uncertain.However,to de-termine the physical properties of the hot Galactic gas,one must first decompose the SXRB into its con-Electronic address:dbh@stituents.This is achieved using a technique called “shadowing”,which makes use of the above-mentioned shadows cast in the SXRB by cool clouds of gas be-tween the Earth and the halo.Low-spectral-resolution ROSAT observations of the SXRB were decomposed into their foreground and background components by modeling the intensity variation due to the varying absorption column density on and around shadow-ing clouds (Burrows &Mendenhall 1991;Snowden et al.1991,2000;Snowden,McCammon,&Verter 1993;Kuntz,Snowden,&Verter 1997).The same tech-nique was used to decompose ROSAT All-Sky Survey data over large areas of the sky (Snowden et al.1998;Kuntz &Snowden 2000).With higher resolution spectra,such as those from the CCD cameras onboard XMM-Newton or Suzaku ,it is possible to decompose the SXRB into its foreground and background components spectroscopically.This is achieved using one spectrum toward a shadowing cloud,and one toward a pointing to the side of the cloud.The spectral shape of the absorbed background component (and hence of the overall spectrum)will differ between the two directions,because of the different absorbing col-umn densities.Therefore,by fitting a suitable multicom-ponent model simultaneously to the two spectra,one can separate out the foreground and background emis-sion components.Such a model will typically consist of an unabsorbed single-temperature (1T )thermal plasma model for the LB,an absorbed thermal plasma model for the Galactic halo,and an absorbed power-law for the ex-tragalactic background.The Galactic halo model could be a 1T model,a two-temperature (2T )model,or a dif-2HENLEY AND SHELTONferential emission measure (DEM)model (Galeazzi et al.2007;Henley et al.2007;S.J.Lei et al.,in preparation).Recent work has shown that there is an additional complication,as X-ray emission can originate within the solar system,via solar wind charge exchange (SWCX;Cox 1998;Cravens 2000).In this process,highly ion-ized species in the solar wind interact with neutral atoms within the solar system.An electron transfers from a neutral atom into an excited energy level of a solar wind ion,which then decays radiatively,emitting an X-ray photon.The neutral atoms may be in the outer reaches of the Earth’s atmosphere (giving rise to geo-coronal emission),or they may be in interstellar ma-terial flowing through the solar system (giving rise to heliospheric emission).It has been estimated that the heliospheric emission may contribute up to ∼50%of the observed soft X-ray flux (Cravens 2000).The geocoro-nal emission is typically an order of magnitude fainter,but during solar wind enhancements it can be of similar brightness to the heliospheric emission (Wargelin et al.2004).SWCX line emission has been observed with Chandra ,XMM-Newton ,and Suzaku (Wargelin et al.2004;Snowden,Collier,&Kuntz 2004;Fujimoto et al.2007).As the SWCX emission is time varying,it cannot easily be modeled out of a spectrum of the SXRB.If SWCX contamination is not taken into account,analyses of SXRB spectra will yield incorrect results for the LB and halo gas.This paper contains a demonstra-tion of this fact.Henley et al.(2007)analyzed a pair of XMM-Newton spectra of the SXRB using the pre-viously described shadowing technique.One spectrum was from a direction toward a nearby shadowing filament in the southern Galactic hemisphere (d =230±30pc;Penprase et al.1998),while the other was from a direc-tion ∼2◦away.The filament and the shadow it casts in the 1/4-keV background are shown in Figure 1.Contem-poraneous solar wind data from the Advanced Composi-tion Explorer (ACE )showed that the solar wind was steady during the XMM-Newton observations,without any flares or spikes.The proton flux was slightly lower than average,and the oxygen ion ratios were fairly typ-ical.These observations led Henley et al.(2007)to con-clude that their spectra were unlikely to be severely con-taminated by SWCX ing a 2T halo model,they obtained a LB temperature of log(T LB /K)=6.06and halo temperatures of log(T halo /K)=5.93and 6.43.The LB temperature and the hotter halo temperature are in good agreement with other recent measurements of the SXRB using XMM-Newton and Suzaku (Galeazzi et al.2007;Smith et al.2007),and with analysis of the ROSAT All-Sky Survey (Kuntz &Snowden 2000).We have obtained spectra of the SXRB from the same directions as Henley et al.’s (2007)XMM-Newton spectra with the X-ray Imaging Spectrometer (XIS;Koyama et al.2007)onboard the Suzaku X-ray obser-vatory (Mitsuda et al.2007).The XIS is an excellent tool for studying the SXRB,due to its low non-X-ray background and good spectral resolution.Our Suzaku pointing directions are shown in Figure 1.We analyze our Suzaku spectra using the same shadowing technique used by Henley et al.(2007).We find that there is poor agreement between the results of our Suzaku analysis and the results of the XMM-Newton analysis in Henley et al.(2007).We attribute this discrepancy to previously un-Fig. 1.—The shadowing filament used for our observations,shown in Galactic coordinates.Grayscale :ROSAT All-Sky Sur-vey R1+R2intensity (Snowden et al.1997).Contours :IRAS 100-micron intensity (Schlegel et al.1998).Yellow squares :Our Suzaku pointing directions.recognized SWCX contamination in the XMM-Newton spectra,which means that SWCX contamination can oc-cur at times when the solar wind flux measured by ACE is low and does not show flares.This paper is organized as follows.The Suzaku obser-vations and data reduction are described in §2.The anal-ysis of the spectra using multicomponent spectral models is described in §3.The discrepancy between the Suzaku results and the XMM-Newton results is discussed in §4.This discrepancy is due to the presence of an additional emission component in the XMM-Newton spectra,which we also describe in §4.In §5we measure the total intensi-ties of the oxygen lines in our Suzaku and XMM-Newton spectra.We concentrate on these lines because they are the brightest in our spectra,and are a major component of the 3/4-keV SXRB (McCammon et al.2002).In §6we present a simple model for estimating the intensity of the oxygen lines due to SWCX,which we compare with our observations.We discuss our results in §7,and con-clude with a summary in §8.Throughout this paper we quote 1σerrors.2.OBSERVATIONS AND DATA REDUCTIONBoth of our observations were carried out in early 2006March.The details of the observations are shown in Ta-ble 1.In the following,we just analyze data from the back-illuminated XIS1chip,as it is more sensitive at lower energies than the three front-illuminated chips.Our data were initially processed at NASA Goddard Space Flight Center (GSFC)using processing version 1.2.2.3.We have carried out further processing and fil-tering,using HEAsoft 1v6.1.2and CIAO 2v3.4.We first combined the data taken in the 3×3and 5×5observa-tion modes.We then selected events with grades 0,2,3,4,and 6,and cleaned the data using the standard data selection criteria given in the Suzaku Data Reduc-tion Guide 3.We excluded the times that Suzaku passed through the South Atlantic Anomaly (SAA),and also times up to 436s after passage through the SAA.We also excluded times when Suzaku ’s line of sight was el-evated less than 10◦above the Earth’s limb and/or was1/lheasoft 2/ciao3/docs/suzaku/analysis/abc/abc.htmlOBSERVING CHARGE EXCHANGE WITH SUZAKU AND XMM-NEWTON3TABLE1Details of Our Suzaku ObservationsObservation l b Start time End time Usable exposureID(deg)(deg)(UT)(UT)(ks)Offfilament501001010278.71−47.072006-03-0116:56:012006-03-0222:29:1455.6Onfilament501002010278.65−45.302006-03-0320:52:002006-03-0608:01:1969.0less than20◦from the bright-Earth terminator.Finally,we excluded times when the cut-offrigidity(COR)wasless than8GV.This is a stricter criterion than that inthe Data Reduction Guide,which recommends exclud-ing times with COR<6GV.However,the higher CORthreshold helps reduce the particle background,and forobservations of the SXRB one desires as low a particlebackground as possible.The COR threshold that weuse has been used for other Suzaku observations of theSXRB(Fujimoto et al.2007;Smith et al.2007).Finally,we binned the2.5–8.5keV data into256-s time bins,andused the CIAO script analyzepublic4HENLEY AND SHELTON34:00.03:33:00.032:00.031:00.015:00.020:00.025:00.0-63:30:00.035:00.040:00.0Right ascensionD e c l i n a t i o nOn filament22:00.021:00.03:20:00.019:00.015:00.0-62:20:00.025:00.030:00.035:00.0Right ascensionD e c l i n a t i o nOff filamentFig.2.—Cleaned and smoothed Suzaku XIS1images in the 0.3–5keV band for our on-filament (left )and off-filament (right )observations.The data have been binned up by a factor of 4in the detector’s x and y directions,and then smoothed with a Gaussian whose standard deviation is equal to 1.5times the binned pixel size.The particle background has not been subtracted from the data.The red circles outline the regions that were excluded from the analysis (see text for details).in our fit (Snowden et al.1997).The R1and R2count-rates help constrain the model al lower energies (below ∼0.3keV),while the higher channels overlap in energy with our Suzaku spectra.We extracted the ROSAT spec-tra from 0.5◦radius circles centered on our two Suzaku pointing directions using the HEASARC X-ray Back-ground Tool 8v2.3.The spectral analysis was carried out using XSPEC 9v11.3.2(Arnaud 1996).For the thermal plasma com-ponents,we used the Astrophysical Plasma Emission Code (APEC)v1.3.1(Smith et al.2001)for the Suzaku data and the ROSAT R4–7bands,and the Raymond-Smith code (Raymond &Smith 1977and updates)for the ROSAT R1–3bands.For a given model component (i.e.,the LB or one of the two halo components),the tem-perature and normalization of the ROSAT Raymond-Smith model are tied to those of the corresponding Suzaku APEC model.We chose to use the Raymond-Smith code for the lower-energy ROSAT channels be-cause APEC’s spectral calculations below 0.25keV are inaccurate,due to a lack of data on transitions from L-shell ions of Ne,Mg,Al,Si,S,Ar,and Ca 10.As the upper-limit of the ROSAT R1and R2bands is 0.284keV,and the R3band also includes such low-energy photons (Snowden et al.1997),APEC is not ideal for fitting to these energy bands.For the absorption,we used the XSPEC phabs model,which uses cross-sections from Ba l uci´n ska-Church &McCammon (1992),with an updated He cross-section from Yan,Sadeghpour,&Dalgarno (1998).Following Henley et al.(2007),we used the interstellar chemical abundance table from Wilms,Allen,&McCray (2000).For many astrophysically abundant elements,these abundances are lower than those in the widely used solar abundance table of Anders &Grevesse (1989).However,recently several elements’solar photospheric abundances have been revised downwards (Asplund et al.2005),and are in good agreement with the Wilms et al.(2000)8/cgi-bin/Tools/xraybg/xraybg.pl9/docs/xanadu/xspec/xspec1110/atomdb/issuesOBSERVING CHARGE EXCHANGE WITH SUZAKU AND XMM-NEWTON5-4-2 0 2 40.3135channel energy (keV)(d a t a -m o d e l )/σ0.001 0.010.11c o u n t s s -1k e V-1On filamentLocal Bubble Halo (cold)Halo (hot)ExtragalacticInstrumental lines TotalO V I I O V I I IN e I XM g X I ?A l KS i KA u M -4-2 0 2 40.3135channel energy (keV)(d a t a -m o d e l )/σ0.0010.010.11c o u n t s s -1k e V-1Off filamentLocal Bubble Halo (cold)Halo (hot)ExtragalacticInstrumental lines TotalN V I I ?O V I I O V I I IN e I XA l KS i KA uMFig.3.—Our observed on-filament (left )and off-filament (right )Suzaku spectra,with the best-fitting model obtained by fitting jointly to the Suzaku and ROSAT data (Model 1in Table 7).The gap in the Si K instrumental line is where channels 500–504have been removed from the data (see §2).10100100010-6 c o u n t s s -1 a r c m i n-2-4-2 0 2 4R1R2R3R4R5R6R7ROSAT band(d a t a - m o d e l ) / σFig. 4.—The on-filament (dashed )and off-filament (solid )ROSAT All-Sky Survey spectra,compared with Model 1from Ta-ble 7.For clarity the individual model components have not been plotted.CALDB.This model component attenuated the emis-sion from the LB,halo,and extragalactic background for the Suzaku spectra only.We adjusted the model oxygen abundance to give C /O =6(Koyama et al.2007),and set the abundances of all other elements to zero.The results of this model are given as Model 2in Table 7.Figures 5and 6show this model com-pared with the Suzaku and ROSAT spectra,respec-tively.One can see from Figure 6that the fit to the ROSAT data is greatly improved.The model implies a column density of carbon atoms,N C =(0.28±0.04)×1018cm −2,in addition to the amount of contamina-tion given by the CALDB contamination model,which is N C =3.1×1018cm −2at the center of the XIS1chip(from the CALDB file aecontami xrtxis08.html6HENLEY AND SHELTON-4-2 0 2 40.3135channel energy (keV)(d a t a -m o d e l )/σ0.0010.010.1c o u n t s s -1k e V-1On filamentLocal Bubble Halo (cold)Halo (hot)ExtragalacticInstrumental lines Total-4-2 0 2 40.3135channel energy (keV)(d a t a -m o d e l )/σ0.0010.010.1c o u n t s s -1k e V-1Off filamentLocal Bubble Halo (cold)Halo (hot)ExtragalacticInstrumental linesTotalFig. 5.—As Figure 3,but with a vphabs component included for the Suzaku spectra to model XIS contamination above that included in the CALDB (Model 2in Table 7;see §3.2for details).10 100100010-6 c o u n t s s -1 a r c m i n-2-4-2 0 2 4R1R2R3R4R5R6R7ROSAT band(d a t a - m o d e l ) / σFig.6.—As Figure 5,but for Model 2from Table 7.In Model 5we fit exactly the same model to the Suzaku data alone.Without the ROSAT data,we cannot con-strain the LB temperature T LB .We therefore fix it at the value determined in Model 2:T LB =105.98K.We also fix N C for the vphabs contamination component at the Model 2value:N C =0.28×1018cm −2.The best-fitting model parameters are in very good agree-ment with those obtained by fitting jointly to the Suzaku and ROSAT spectra (compare Models 2and 5).We can only get consistent results between the Suzaku -ROSAT joint fit and the fit to just the Suzaku data by using a two-temperature model for the halo.However,note that the Model 5LB emission measure is consistent (within its errorbar)with zero.This is not to say that our data imply that there is no LB at all:as noted above,we need a LB component and two halo components to get a good joint fit to the Suzaku and ROSAT data.Instead,the Model 5results imply that our Suzaku data are consis-tent with the LB not producing significant emission in the Suzaku band (i.e.,E 0.3keV).Also shown in Table 7are the results of fitting our model (without any LB component)to the on-and off-filament Suzaku spectra individually (Models 6and 7,respectively).As has already been noted,the normaliza-tion of the extragalactic background differs significantly between the two spectra.However,the plasma model pa-rameters for the individual spectra are in good agreement with each other.These results seem to justify allowing the extragalactic normalization to differ between the two spectra while keeping all other model parameters equal for the two spectra.3.3.Chemical Abundances in the HaloWe investigated the chemical abundances of the X-ray-emissive halo gas by repeating the above-described mod-eling,but allowing the abundances of certain elements in the halo components to vary.In particular we wished to investigate whether or not varying the neon and magne-sium halo abundances improved the fits to the Ne ix and Mg xi features noted above.We also allowed the abun-dance of iron to vary,as iron is an important contributor of halo line emission to the Suzaku band.For this investigation we just fit to the Suzaku data,fixing the LB temperature at log(T LB /K)=5.98,and fixing the carbon column density of the vphabs contam-ination component at N C =0.28×1018cm −2.As we could not accurately determine the level of the contin-uum due to hydrogen,we could not measure absolute abundances.Instead,we estimated the abundances rel-ative to oxygen by holding the oxygen abundance at its Wilms et al.(2000)value,and allowing the abundances of neon,magnesium,and iron to vary.Both halo com-ponents were constrained to have the same abundances.The best-fitting temperatures and emission measures of the various model components are presented as Model 8in Table 7,and the abundances are presented in Table 2.The best-fitting model parameters are not significantly affected by allowing certain elements’abun-dances to vary (compare Model 8with Model 5).Iron does not seem to be enhanced or depleted relative to oxy-gen in the halo.Neon and magnesium both appear to be enhanced in the halo relative to oxygen,which is what one would expect from Figures 3and 5,as the modelsOBSERVING CHARGE EXCHANGE WITH SUZAKU AND XMM-NEWTON7 TABLE2Halo AbundancesElement AbundanceO1(fixed)Ne a1.8±0.4Mg a4.6+3.5−2.8Fe1.2+0.4−0.5Note.—Abun-dances are relative tothe Wilms et al.(2000)interstellar abundances:Ne/O=0.178,Mg/O=0.051,Fe/O=0.055.a These enhanced abun-dances may be an arte-fact of SWCX contami-nation;see§7.2.shown in thosefigures underpredict the neon and mag-nesium emission.We discuss these results in§§7.2and§7.5.In§7.2we discuss the possibility that the enhanced neon and mag-nesium emission is in fact due to SWCX contamination of these lines,rather than being due to these elements being enhanced in the halo.On the other hand,in§7.5 we discuss the implications of neon really being enhanced in the halo with respect to oxygen and iron.PARING THE SUZAKU AND XMM-NEWTONSPECTRAFor comparison,Table7also contains the results of the analysis of the XMM-Newton spectra from the same ob-servation directions by Henley et al.(2007).The Model9 results are taken directly from their“standard”model. However,it should be noted that Henley et al.(2007) used APEC to model all of their data,whereas in the analysis described above we used the Raymond-Smith code to model the ROSAT R1–3bands.We have therefore reanalyzed the XMM-Newton+ROSAT spec-tra,this time using the Raymond-Smith code to model the ROSAT R1–3bands,and using APEC to model the ROSAT R4–7bands and the XMM-Newton spec-tra.This new analysis allows a fairer comparison of our XMM-Newton results with our Suzaku results.The XMM-Newton spectra we analyzed are identical to those used by Henley et al.(2007)–see that paper for details of the data reduction.We added a broken power-law to the model to take into account soft-proton contamination in the XMM-Newton spectra.This broken power-law was not folded through XMM-Newton’s effective area, and was allowed to differ for the on-and off-filament datasets(see Henley et al.2007).The presence of this contamination means we cannot independently constrain the normalization of the extragalactic background.We therefore freeze the on-and off-filament normalizations at the Suzaku-determined values.The results of this new analysis are presented as Model10in Table7.As can be seen,there is poor agreement between the best-fit parameters of the Suzaku +ROSAT model(Model2)and the XMM-Newton+ ROSAT model(Model10).We believe this discrepancy is due to an extra emission component in the XMM-Newton spectra.In Figure7we plot the differences between the XMM-Newton spectra and our best-fitting0.10.20.30.40.5data-model(ctss-1keV-1)On filament MOS 1MOS 2 00.10.20.30.40.50.5125channel energy (keV)Off filament MOS 1MOS 2Fig.7.—The excesses in our on-filament(top)and off-filament (bottom)XMM-Newton spectra over our best-fitting model to the Suzaku+ROSAT data(Table7,Model2).The gap in the data between1.4and1.9keV is where two bright instrumental lines have been removed.Suzaku+ROSAT model(Model2from Table7).To our best-fitting Suzaku+ROSAT model we have added a broken power-law to model the soft-proton contami-nation in the XMM-Newton spectra.The parameters of this broken power-law are frozen at the values de-termined from thefitting to the XMM-Newton spectra described in the previous paragraph.The on-filament XMM-Newton spectra show excess line emission at∼0.57 and∼0.65keV,most likely due to O vii and O viii,re-spectively.The features in the off-filament spectra are not as clear.However,there appears to be excess O vii emission in the MOS1spectrum,and excess emission at ∼0.7keV(of uncertain origin)and∼0.9keV(proba-bly Ne ix)in the MOS2spectrum.We can estimate the significance of the excess emission by calculatingχ2 for the XMM-Newton data compared with the Suzaku +ROSAT model.We concentrate on the excess oxy-gen emission and calculateχ2for the0.5–0.7keV energy range.Wefindχ2=106.28for24degrees of freedom for the on-filament spectra,andχ2=43.03for22degrees of freedom for the off-filament spectra.These correspond toχ2probabilities of2.5×10−12and0.0047,respectively, implying that the excesses are significant in both sets of spectra at the1%level.We measure the intensities of the extra oxygen emis-sion byfittingδ-functions at E=0.570keV and 0.654keV to the excess spectra in Figure7.Wefit theseδ-functions simultaneously to the on-and off-filament XMM-Newton excess spectra.The intensities of the excess oxygen emission in the XMM-Newton spec-tra over the best-fitting Suzaku+ROSAT model are 3.8±0.5L.U.(O vii)and1.4±0.3L.U.(O viii).We believe that this excess oxygen emission is due to SWCX contamination in our XMM-Newton spectra.As noted in the Introduction,this was not taken into ac-count in the original analysis of the XMM-Newton spec-tra.This is because the solar windflux was steady and slightly below average during the XMM-Newton obser-vations,leading Henley et al.(2007)to conclude that SWCX contamination was unlikely to be significant.We discuss the SWCX contamination in our spectra further in§6.5.MEASURING THE OXYGEN LINES8HENLEY AND SHELTONAs well as using the above-described method to sep-arate the LB emission from the halo emission,we mea-sured the total intensities of the O vii complex and O viii line at∼0.57and∼0.65keV in each spectrum.These lines are a major component of the SXRB,accounting for the majority of the observed ROSAT R4diffuse back-ground that is not due to resolved extragalactic discrete sources(McCammon et al.2002),and are easily the most prominent lines in our Suzaku spectra.To measure the oxygen line intensities,we used a model consisting of an absorbed power-law,an absorbed APEC model whose oxygen abundance is set to zero,and two δ-functions to model the oxygen lines.As in the pre-vious section,the power-law models the extragalactic background,and its photon index was frozen at1.46 (Chen et al.1997).The APEC model,meanwhile,mod-els the line emission from elements other than oxygen, and the thermal continuum emission.The absorbing columns used were the same as those used in the ear-lier Suzaku analysis.As with our earlier analysis,we multiplied the whole model by a vphabs component to model the contamination on the optical blockingfilter which is in addition to that included in the CALDB con-tamination model(see§3.2).Wefix the carbon column density of this component at0.28×1018cm−2(Table7, Model2).Wefit this model simultaneously to our on-and off-filament spectra.However,all the model param-eters were independent for the two directions,except for the oxygen line energies–these were free parameters in thefit,but were constrained to be the same in the on-and off-filament spectra.We used essentially the same method to measure the oxygen line intensities in our XMM-Newton spectra.However,we did not use a vphabs contamination model,and,as before,we added a broken power-law to model the soft-proton contami-nation.Table3gives the energies and total observed intensities of the O vii and O viii emission measured from our Suzaku and XMM-Newton spectra.We can use the difference in the absorbing column for the on-and off-filament directions to decompose the ob-served line intensities into foreground(LB+SWCX)and background(halo)intensities.If I fg and I halo are the in-trinsic foreground and halo line intensities,respectively, then the observed on-filament intensity I on is given byI on=I fg+e−τon I halo,(1) whereτon is the on-filament optical depth at the energy of the line.There is a similar expression for the observed off-filament intensity I off,involving the off-filament op-tical depthτoff.These expressions can be rearranged to giveI fg=eτon I on−eτoff I offe−τon−e−τoff.(3) For the purposes of this decomposition,we use the Ba l uci´n ska-Church&McCammon(1992)cross-sections (with an updated He cross-section;Yan et al.1998)with the Wilms et al.(2000)interstellar abundances.We use the cross-sections at the measured energies of the lines. For the Suzaku measurements,the cross-sections we use are7.17×10−22cm2for O vii(E=0.564keV)and 4.66×10−22cm2for O viii(E=0.658keV).For the XMM-Newton O vii emission we use a cross-section of 7.03×10−22cm2(E=0.568keV).We cannot decom-pose the XMM-Newton O viii emission because the on-filament O viii line is brighter than the off-filament line. This gives rise to a negative halo intensity,which is un-physical.The results of this decomposition are presented in Ta-ble4.Note that the foreground oxygen intensities mea-sured from the Suzaku spectra are consistent with zero. This is consistent with our earlierfinding that the Suzaku spectra are consistent with there being no local emis-sion in the Suzaku band(see§3.2).The difference be-tween the foreground O vii intensities measured from our XMM-Newton and Suzaku spectra is5.1±3.1L.U.. This is consistent with the O vii intensity measured from the excess XMM-Newton emission over the best-fitting Suzaku+ROSAT model(3.8±0.5L.U.;see§4).The halo O vii intensities measured from our XMM-Newton and Suzaku spectra are consistent with each other.This is as expected,as we would not expect the halo intensity to significantly change in∼4years.6.MODELING THE SOLAR WIND CHARGE EXCHANGEEMISSIONIn§§4and5we presented evidence that our XMM-Newton spectra contain an extra emission component,in addition to the components needed to explain the Suzaku spectra.In particular,the O vii and O viii emission are enhanced in the XMM-Newton spectra.We attribute this extra component to SWCX emission,as it seems unlikely to be due to a change in the Local Bubble or halo emission.This extra component helps explain why our XMM-Newton and Suzaku analyses give such different results in Table7.Previous observations of SWCX have found that in-creases in the SWCX emission are associated with en-hancements in the solar wind,as measured by ACE. These enhancements consist of an increase in the pro-tonflux,and may also include a shift in the ionization balance to higher ionization stages(Snowden et al.2004; Fujimoto et al.2007).In§6.1,we present a simple model for heliospheric and geocoronal SWCX emission,and use contemporaneous solar wind data from the ACE and WIND satellites to determine whether or not the ob-served enhancement of the oxygen lines in the XMM-Newton spectra is due to differences in the solar wind between our two sets of observations.In addition to the variability associated with solar wind enhancements,the heliospheric SWCX intensity is also expected to vary during the solar cycle,due to the dif-ferent states of the solar wind at solar maximum and solar minimum(Koutroumpa et al.2006).As our two sets of observations were taken∼4years apart,at differ-ent points in the solar cycle,in§6.2we examine whether the SWCX intensity variation during the solar cycle can explain our observations.6.1.A Simple Model for Heliospheric and GeocoronalSWCX Emission6.1.1.The BasicsA SWCX line from a X+n ion of element X results from a charge exchange interaction between a X+(n+1) ion in the solar wind and a neutral atom.The intensity of that line therefore depends on the density of X+(n+1)。

a r X i v :a s t r o -p h /0010612v 1 30 O c t 2000A&A manuscript no.(will be inserted by hand later)ASTRONOMYANDASTROPHY SICSSend offprint requests to :J.N ReevesKey words:galaxies:active –galaxies:starburst –X-rays:galaxies2M.J.L.Turner et al.:XMM-Newton First-Light Observations of the Hickson Galaxy Group16Fig.1.The smoothed colour X-ray image of HGC 16made with the EPIC MOS CCD imaging spectrometers on XMM-Newton.The spatial resolution is 6′′Full Width at Half Maximum,and 15′′Half Energy Width,limited by the mirrors.The energy band is 0.2-10keV and the energy resolution varies from 140eV FWHM at 6keV to 70eV at 500eV.In the colour image,red corresponds to 800eV and blue to >3keV.The physical scale across the image corresponds to 200kpc (using H 0=50km s −1Mpc −1).Notice the very hard (blue)nucleus of NGC 833and the soft (red)halo emission around the companion galaxy NGC 835.<2.5(de Carvalho &Coziol 1999).The EPIC instrument is able to observe directly the ionising continuum,and dis-tinguish clearly between optically thin thermal emission from a starburst,and non-thermal hard X-ray emission from an AGN.These observations can therefore be used to determine the nature of the ionising continuum,helping to clarify the relationship between mergers,the triggering of starbursts,and the creation and fueling of black holes.2.The XMM-Newton ObservationsThe XMM-Newton Observatory (Jansen et al.2001)has three X-ray telescopes of area ∼1500cm 2,with the three EPIC instruments at the foci;two of the EPIC imaging spectrometers use MOS CCDs (Turner et al.2001,Hol-land et al.1996)and one uses a PN CCD (Str¨u der et al.2001).The observations of the HCG-16galaxy group weretaken in orbit-23as part of the XMM-Newton EPIC first-light.Exposures of 50ksec were taken with EPIC (sensi-tive from 0.2to 10keV)and 1ksec exposures were taken in V (550nm)and UV (280nm)with the XMM-Newton Optical/UV Monitor (OM)telescope (Mason et al.2001).The EPIC data were processed using the pipeline scripts emchain (MOS)and epchain (PN).Screening was applied using the XMM-Newton SAS (Science Analy-sis Software).Hot and bad pixels and negative E3events were removed from the data to reduce the level of elec-tronic noise.A low energy cut of 200eV was applied to the data.The first 10ksec of data were also removed from the EPIC observation,as this contained a high count-rate background particle flare.The resultant exposure time for each of the detectors was ∼40ksec.Figure 1shows the resultant EPIC X-ray colour image of the centre of the HCG-16field.The hard,absorbed,M.J.L.Turner et al.:XMM-Newton First-Light Observations of the Hickson Galaxy Group 163Fig.2.The OM colour image of HCG 16derived from 1000second exposures in V and UV bands.A bright foreground star has been removed from the lower right of the image.Blue here represents the UV band.spectrum of the AGN in NGC 833shows up as a blue point source,and the soft starburst emission in the outer regions of NGC 835shows as a red halo;the other galaxies show extended X-ray discs.Figure 2also shows the V-UV colour image from the OM.The nuclear regions of NGC 835and NGC 838show up brightly in the ultraviolet,indicative of hot stars or gas associated with enhanced star formation.There are also bright UV knots in the outer regions of NGC 835showing enhanced star formation there.A close-up OM image of NGC 835is shown in figure 3.3.Spectral Analysis of the HCG-16Galaxies Since EPIC resolves the optical disks of the galaxies,the spectra were prepared from photons falling within a re-gion of interest based on the X-ray image.For NGC 833the X-ray source is point-like,while for NGC 835the core and the surrounding region (the red halo in figure 1)were analysed separately,the spectra of the remaining galax-ies were made using the entire X-ray disc.Background spectra were taken from source-free regions on the central EPIC-MOS and PN chips;the background spectra were normalised to the area of the source extraction regions.The background subtracted EPIC spectra were fitted,using xspec v11.0,with the latest response matrices pro-duced by the EPIC team;the systematic level of uncer-tainty is <5%.Finally spectra were binned to a minimum of 20counts per bin,in order to apply the χ2minimisa-tion technique.All subsequent errors are quoted to 90%confidence (∆χ2=4.6for 2interesting parameters).Val-Fig.3.A close-up,greyscale UV image from the Optical Monitor,showing the galaxies NGC 833(right)and NGC 835(left).Bright UV knots,corresponding to possible re-gions of star formation,are seen in the outer disk of NGC 835.ues of H 0=50km s −1Mpc −1and q 0=0.5have been assumed and all fit parameters are given in the rest-frame of the HCG-16system.We now present the individual EPIC spectra of the 4main Hickson-16galaxies.4M.J.L.Turner et al.:XMM-Newton First-Light Observations of the Hickson Galaxy Group 163.1.NGC 833Optical imaging data on NGC 833reveal a disturbed velocity field and pronounced misalignment of the kine-matic and stellar axes,indicativeof an ongoing inter-action (Mendes de Oliveira et al.1998).The emission lines present in the optical spectra (de Carvalho &Co-ziol 1999)indicate weak non-stellar LINER-2activity in the core;there is no optical evidence for current star for-mation ([NII]/H β∼unity).The EPIC image of this galaxy is point-like,much smaller than the stellar disc.The best-fitting EPIC X-ray spectrum (Figure 4)shows three distinct components,all required at >99.99%confidence.The most obvious is the peak at high energies from an obscured AGN;this emission is in the form of a power-law of index Γ=1.8±0.5,absorbed by material of column density equal to N H =2.4±0.4×1023cm −2.The second component is an un-absorbed power-law,re-sulting from radiation scattered into our line of sight,by thin,hot,plasma directly illuminated by the AGN.The third component is radiation from an optically-thin ther-mal plasma,with a temperature of kT =470eV.The improvement in the fit upon adding the thermal emission is ∆χ2=36.7.A summary of the fits to NGC 833(and the other 3galaxies)are given in table 1.This complex X-ray spectrum amply confirms the presence of an AGN in NGC 833of luminosity 1.4±0.6×1042erg s −1,it is,remarkably,the dominant source of power in the galaxy.In contrast,the thermal X-ray emission,is more than 100times weaker (8.9±3.0×1039erg s −1)and the FIR luminosity (Verdes-Montenegro et al.1998)is also very low (<3×1042erg s −1).3.2.NGC 835The adjacent galaxy,NGC 835is undergoing a gravita-tional interaction with its neighbour NGC 833,as evi-denced by the tidal tails in the optical image;and ap-parently contiguous stellar discs (Mendes de Oliveira et al.1998).The velocity field is normal,but there is emis-sion line evidence (de Carvalho &Coziol 1999)for LINER nuclear activity,and for current starburst activity in the outer regions;the knotted ring structure seen in the OM image supports this.The X-ray emission from NGC 835can be spatially separated into two areas,the core,and an outer region corresponding to the remainder of the stellar disc.The core has a very similar spectrum (Fig.5)to that of NGC 833.There are absorbed and scattered power-laws indicat-ing a heavily obscured AGN (N H =4.6±1.5×1023cm −2)of luminosity 1.2×1042erg s −1(0.5-10keV),but the soft thermal component is more luminous than NGC 833at 2.5±0.3×1040erg s −1;it is almost certainly from current starburst activity and the FIR luminosity (Verdes-Montenegro et al.1998)is 100times larger at 2.7×1044erg s −1.The spectrum at the periphery of NGCFig.4.The X-ray spectrum of the galaxy NGC 833.The most striking feature is the high-energy,absorbed power-law (at >3keV)that is the direct emission from the active black hole at the centre of the galaxy.There is also an un-absorbed power-law,resulting from radiation scattered into our line of sight,by material directly il-luminated by the AGN.These two spectral components together show the presence of an AGN of luminosity 1.4±0.6×1042erg s −1.There is also weak soft X-ray emis-sion from an optically thin plasma,perhaps originating from starburst activity.Fig.5.The spectrum of the companion galaxy,NGC 835,which also shows an obscured active nucleus together with strong thermal soft X-ray emission.The AGN in both NGC 833and NGC 835may have been triggered by mu-tual gravitational interaction.835is purely thermal,with a temperature of 300eV and a luminosity of 2.9±0.7×1040erg s −1,similar to that of the core.This is the X-ray emission from the starburst region including the ring structure and knots seen in the OM V-UV image.M.J.L.Turner et al.:XMM-Newton First-Light Observations of the Hickson Galaxy Group 165Fig.6.The EPIC-MOS spectrum of NGC 838.Only emis-sion from the starburst is present,with no detectable hard X-ray emission from a central AGN;the hard X-ray emis-sion could arise from unresolved X-ray binaries in the galaxy.To summarize,the EPIC data clearly show the pres-ence of an AGN,in both NGC 833and NGC 835,that coexists with present starburst activity in the core (and for NGC 835in the periphery)of the galaxies.The detec-tions of the obscured AGN and thermal starburst compo-nents in both galaxies are highly significant,at >99.99%confidence (see table 1).3.3.NGC 838NGC 838is an ongoing merger with strongstarburst activity.Optical data (Mendes de Oliveira et al.1998)show kinematic warping,and multiple velocity compo-nents in the ionised gas,and a double optical core (also see de Carvalho &Coziol 1999).The infrared luminos-ity is 3.3±1044erg s −1.The EPIC spectrum of NGC 838(Fig.6)shows purely thermal emission,the disc is re-solved in X-rays,but there is no separate sharp core in the X-ray image.The spectrum is fitted with a two tempera-ture thermal spectrum (kT =3.2keV and kT =590eV)and the luminosity is high at 1.9±0.3×1041erg s −1.This is all consistent with the optical data:the soft X-ray emission is from the ionised gas produced in the starburst while the hard thermal spectrum could be characteristic of unresolved X-ray binaries.There is no statistically sig-nificant scattered or obscured power law;the upper limit for the AGN luminosity is 5×1040erg s −1,assuming an absorbing column of 5×1023cm −2.3.4.NGC 839NGC 839may also be a recent merger (Mendes de Oliveira et al.1998,de Carvalho &Coziol 1999),it has a double nu-Fig.7.The X-ray spectrum of NGC 839.There is both starburst emission and emission from a low luminosity ob-scured AGN.Unusually the elemental abundance of NGC 839appears to be ∼5times solar.cleus in the optical,a FIR luminosity of 3.1×1044erg s −1,and a disturbed velocity field;optical lines indicate an ac-tive LINER-2nucleus (de Carvalho &Coziol 1999).In the soft X-ray EPIC spectrum (Fig.7)there is optically-thin thermal emission,similar to that of the other galaxies,of temperature kT =600eV,and luminosity 1.8±0.3×1040erg s −1;a typical indicator of a current starburst.The spectrum also shows an obscured AGN,as found in NGC 835&NGC 833;it is however much less luminous (8±3×1040erg s −1)for a column of N H =5×1023cm −2.Interestingly the abundances in NGC 839appear to be higher than solar.There are apparent weak Lyman-αlines of O,Mg and Si in the EPIC spectrum,although the sig-nificance of these features is low (at only 90%confidence).Fitting the soft X-ray spectrum with the mekal model does however yield an over-abundance of 5.2±2.0times the solar value.One interesting possibility is that the heavier elements have been enriched through the intense starburst activity in this galaxy.4.ConclusionsDirect X-ray spectroscopy is the best way to identify hid-den AGN in galaxies,and here the EPIC cameras on XMM-Newton have produced clear evidence for active,massive black holes in three out of four galaxies in HCG-16.The presence of a similar active nucleus in NGC 838is unlikely,unless it is very heavily absorbed.The na-ture of the ionising continuum in the four galaxies has been elucidated:there is thermal emission from starburst activity in three,(possibly four)of the galaxies,and in three of them there is a coexisting active black hole.While LINER-1galaxies with broad H βlines do harbour black holes (Terashima et al.1998,2000)this is the first direct6M.J.L.Turner et al.:XMM-Newton First-Light Observations of the Hickson Galaxy Group16Table1.X-ray spectralfits to the4HCG-16galaxies.a Temperature of thermal component in keV.b Column density of the absorbed power-law in units of1022cm−2.c Improvement in the spectralfit upon adding the obscured hard power-law.d Improvement in thefit from adding a soft,thermal(Mekal)component.e Best-fit reduced chi-squared.f Indicates that parameter isfixed.Note ABS.PL is the absorbed power-law,SCAT.PL is the scattered power-law;Γfor these2components have been tied.NGC833ABS.PL+SCAT.PL+MEKALΓ=1.8±0.524±492.70.47±0.1236.70.632 NGC835(centre)ABS.PL+SCAT.PL+MEKALΓ=2.25±0.2346±1558.40.51±0.07176.2 1.02 NGC835(diffuse)MEKAL×2kT=4keV f––0.31±0.0573.4 1.2 NGC838MEKAL×2kT=3.2±0.8––0.59±0.04209.5 1.21 NGC839ABS.PL+SCAT.PL+MEKALΓ=2.1±0.845±2012.00.63±0.1040.9 1.38。

a r X i v :a s t r o -p h /0401075v 1 7 J a n 2004Mem.S.A.It.Vol.,1cSAIt 2002MemoriedellaF.HaberlMax-Planck-Institut f¨u r extraterrestrische Physik,Giessenbachstraße,85748Garching,GermanyAbstract.As part of the guaranteed time,guest observer and calibration pro-grams,XMM-Newton extensively observed a group of six thermally emitting,isolated neutron stars which are neither connected with a supernova remnant nor show pulsed radio emission.The high statistical quality of the EPIC data allows a detailed and homogeneous analysis of their temporal and spectral properties.Four of the six sources are now well established as X-ray pulsars:The 11.37s period discovered in EPIC-pn data of RX J0806.4−4123was confirmed in a second XMM-Newton observation.In the case of the X-ray faintest of the six stars,RX J0420.0−5022the period marginally indicated in ROSAT data was not seen in the EPIC data,instead a 3.45s pulse period was clearly detected.Spectral variations with pulse phase were discovered for the known 8.39s pul-sar RX J0720.4−3125and also RBS1223.For the latter EPIC data revealed a double-peaked pulse profile for a neutron star spin period of 10.31s.The X-ray continuum spectra of all six objects are consistent with a Planckian energy distribution with black-body temperatures kT between 40eV and 100eV.EPIC data of the pulsars RBS1223and RX J0720.4−3125revealed a broad absorption feature in their spectra at energies of 100-300eV and ∼260eV,re-spectively.The depth of this feature varies with pulse phase and may be caused by cyclotron resonance scattering of protons or heavy ions in a strong magnetic field.In such a picture the inferred field strength exceeds 1013Gauss,in the case of RX J0720.4−3125consistent with the value estimated from its pulse pe-riod derivative.A similar absorption feature in the RGS and EPIC spectra of RX J1605.3+3249was reported recently.Key words.X-rays:stars –stars:neutron –stars:magnetic fields1.IntroductionPresently more than half a dozen ROSAT-discovered X-ray dim isolated neutron stars (XDINs,for recent reviews see Treves et al.2 F.Haberl:Radio-quiet and X-ray dim isolated neutronstarsFig.1.EPIC-pn light curves folded on the pulse period for the four pulsars among the XDINs.remnants.Some of them exhibit pulsationsin their X-rayflux indicating the neutronstar rotation period.Although there is littledoubt that the soft X-rays are of thermalorigin from the surface of an isolated neu-tron star,the details for the formation ofthe spectrum are not clear.XMM-Newtonobserved six XDINs as part of the guaran-teed time,guest observer and calibrationprograms.Here I summarize some of thefirst results of a homogeneous analysis oftheir temporal and spectral properties.2.X-ray pulsationsThe second brightest of the known XDINsis RX J0720.4−3125and was thefirstdiscovered as X-ray pulsar(Haberl et al.1997)with a period of8.39s.XMM-Newton observed RX J0720.4−3125sixtimes.As example the folded EPIC-pnlight curve from satellite revolution534is drawn in Fig.1which shows a pulsedfraction of about11%.In Chandra dataof RBS1223=1RXS J130848.6+212708Hambaryan et al.(2002)discovered pulsa-tions with a period of5.16s and XMM-Newton observations revealed that the trueneutron star spin period is more likely10.31s.This is supported by pulse phase de-pendent hardness ratio variations whichare different for the two intensity max-ima(Haberl et al.2003).RBS1223ex-hibits the deepest modulation of the knownXDINs in its double peaked pulse profileof18%(Fig.1).Thefirst XMM-Newtonobservation of RX J0806.4−4123revealeda6%modulation(Fig.1)with a periodof11.37s(Haberl&Zavlin2002)whichwas confirmed in a second observation(Haberl et al.2004a).Four XMM-Newtonobservations of RX J0420.0−5022did notconfirm the22.7s pulsations originally in-dicated in ROSAT data,but clearly re-F.Haberl:Radio-quiet and X-ray dim isolated neutron stars3veal a3.45s period.The pulsed fraction is about12%(Fig.1).In Table1the de-rived pulse periods for the four pulsars are summarized together with pulse period derivatives assuming linear period changes between multiple XMM-Newton observa-tions.In most cases the time base line is too short to determine precise˙P values, only for RX J0720.4−3125observations by different satellites cover already more than 10years(Zane et al.2002;Kaplan et al. 2002).3.X-ray spectraThe X-ray spectra of XDINs obtained by the ROSAT PSPC were all consistent with black-body emission little attenuated by interstellar absorption suggesting that the objects are close-by.To look for ab-sorption features which may be created by heavy chemical elements in the stel-lar atmosphere high resolution spectra of RX J1856.4−3754were obtained by the LETGS on Chandra(Burwitz et al.2001, 2003).Surprisingly,no significant narrow features were detected.A spectral analysis(in a homogenous way using the latest calibration data)of the EPIC-pn spectra,which are of unprece-dented statistical quality,shows that in sev-eral cases a black-body model yields un-satisfactoryfits.The strongest deviations are seen from RBS1223and Haberl et al. (2003)demonstrate that a non-magnetic atmosphere models can neither explain the observed spectrum.However,it was found that adding an absorption feature mod-eled by a broad(100eV)Gaussian line at an energy between100and300eV to the Planckian continuum yields accept-ablefits.Similarly,but at a higher en-ergy of450eV and therefore inside the sensitive band of the RGS instruments van Kerkwijk(2004)presented the detec-tion of a broad absorption feature in the spectra of RX J1605.3+3249.Haberl et al. (2004b)report a broad absorption fea-ture in the EPIC-pn spectra of the pul-sar RX J0720.4−3125at an energy of270eV andfind that the depth of the feature varies with pulse phase by a factor of∼2. Finally,also spectra of RX J0420.0−5022 indicate a possible absorption line at330 eV(Haberl et al.2004a).In Table1the spectral parameters in-ferred from thefits using the two mod-els(absorbed black-body and absorbed black-body with Gaussian-shaped absorp-tion line)are listed.Thefirst value given for column density N H and black-body tem-perature kT refers to the model without ab-sorption line and the second to the model with line added.4.DiscussionFour of the six X-ray-dim isolated neu-tron stars are now known as X-ray pul-sars with spin periods between3.45s and 11.37s.The fraction of pulsedflux in their folded X-ray light curves ranges between 6%and18%.RBS1223,RX J0720.4−3125 and RX J0420.0−5022show hardness ratio variations with pulse phase(Haberl et al. 2003,2004b,a)while for RX J0806.4−4123 the shallow modulation makes the signifi-cant detection of such an effect more diffi-cult(Haberl et al.2004a).Pulse phase re-solved spectra for thefirst three pulsars show that the temperature changes only marginally with pulse phase.Any model for the pulsed emission from this group of iso-lated neutron stars must be able to explain this behaviour.An important piece of information comes from the detection of broad ab-sorption features in the X-ray spectra of several XDINs.At least in the case of RX J0720.4−3125the depth of the feature varies strongly with pulse phase.A likely interpretation of these features is cyclotron resonance absorption which can be ex-pected in spectra from magnetized neutron stars withfield strength B in the range of 1010–1011G or2×1013–2×1014G if caused by electrons or protons,respectively.In the case of RX J0720.4−3125the measured˙P, if interpreted as magnetic dipole braking, rules out the lower range for B,leaving pro-4 F.Haberl:Radio-quiet and X-ray dim isolated neutron starsTable1.X-ray-dim isolated neutron stars observed by XMM-NewtonRX J0420.0−5022 3.453<9×10−12 1.0/2.045/45330?6? RX J0720.4−31258.391(3−6)×10−14 1.4/0.885/842705 RX J0806.4−412311.371<2×10−120.496−RBS122310.313<6×10−127.1/4.195/86100-3002-6 RX J1605.3+3249−−0.3/0.996/934509 RX J1856.5−3754−−0.960−1?F.Haberl:Radio-quiet and X-ray dim isolated neutron stars5 2002,MNRAS,334,345。

a r X i v :a s t r o -p h /0307223v 2 21 D e c 2005The Astrophysical Journal,acceptedPreprint typeset using L A T E X style emulateapj v.11/26/04A CLUSTER OF GALAXIES HIDING BEHIND M31:XMM-NEWTON OBSERVATIONS OF RXJ0046.4+4204Oleg Kotov 1,2,3,Sergey Trudolyubov 1,2,4,and W.Thomas Vestrand 1The Astrophysical Journal,acceptedABSTRACTWe report on our serendipitous discovery with the XMM-Newton Observatory of a luminous X-ray emitting cluster of galaxies that is located behind the Andromeda galaxy (M31).X-ray emission from the cluster was detected previously by ROSAT,and cataloged as RX J0046.4+4204,but it was not recognized as a galaxy cluster.The much greater sensitivity of our XMM-Newton observations revealed diffuse x-ray emission that extends at least 5′and has a surface brightness profile that is well fit by the α-βmodel with β=0.70±0.08,a core radius r c =56′′±16′′,and α=1.54±0.25.A joint global spectral fit of the EPIC/MOS1,MOS2,and PN observations with Mewe-Kaastra-Liedahl plasma emission model gives a cluster temperature of 5.5±0.5keV.The observed spectra also show high significance iron emission lines that yield a measured cluster redshift of z =0.290with a 2%accuracy.For a cosmological model with H 0=71km s −1Mpc −1,ΩM =0.3and ΩΛ=0.7we derive a bolometric luminosity of L x =(8.4±0.5)×1044erg/s.This discovery of a cluster behind M31demonstrates the utility of x-ray surveys for finding rich clusters of galaxies,even in directions of heavy optical extinction.Subject headings:galaxies:clusters:Intergalactic medium -X-rays:observation -Cosmology1.INTRODUCTIONGalaxy clusters are the largest gravitationally bound structures in the universe.The evolution of cluster num-ber density of a given mass is sensitive to specific cos-mological scenarios (e.g.Press &Schechter (1974)).So observations of galaxy clusters are an important tool for constraining fundamental cosmological parameters.Due to the fact that 15%of the total cluster mass (e.g.Evrard (1997))is in the form of hot diffuse plasma emit-ting at X-ray band via thermal bremsstrahlung (Sarazin 1988),galaxy clusters are among the most luminous ob-jects in X-ray band.It makes X-ray selection an efficient means for constructing samples of galaxy clusters (see review by Rosati et al.(2002)).X-ray selection has the advantage that the measurable X-ray luminosity and temperature are correlated with the cluster mass.Further,X-ray selection is useful for studying regions where optical searches are complicated because of dust extinction and heavy stellar confusion.X-rays are much less affected by extinction than optical photons and X-ray selection is almost free from source confusion prob-lems (Ebeling et al.2002).Conducting X-ray selection based on ROSAT data at low Galactic latitude,Ebeling et al.(2002)were able to detect 137galaxy clusters ,70%of which were new discoveries.With a new era of XMM-Newton and Chandra ob-servatories with their large effective areas and wide en-ergy ranges 0.3−10keV,the capability of X-ray selec-tion increased.During an XMM-Newton observation of the galactic supernova remnant G21.5-09located close to the Galactic Plane,Nevalainen et al.(2001)detected1NIS-2,Space and Remote Sciences Group,Los Alamos Na-tional Laboratory,Los Alamos,NM 875452Space Research Institute,Russian Academy of Sciences,Moscow,Russia3Harvard-Smithsonian Center for Astrophysics,60Garden St.,Cambridge,MA 021384Institute of Geophysics and Planetary Physics,University of California,Riverside,CA 92521a new galaxy ing only XMM-Newton data,they measured cluster redshift z =0.1to 1%precision that is especially important in regions with such strong optical source confusion,where the optical redshift mea-surements of galaxies are difficult .Here we present new XMM-Newton observations of the source RX J0046.4+4202that indicate it is a high red-shift cluster located behind M31.RX J0046.4+4204was detected during the first and the second deep ROSAT RSPC surveys of M31performed in June 1991and July/August 1992respectively (Supper et al.2001).Based on a comparison of the first and the second sur-veys,RX J0046.4+4204was classified as a potentially long term variable source.Our analysis of the data ob-tained with XMM-Newton revealed spatially extended emission,up to at least 5′,from RX J0046.4+4204.The observed spectra show iron emission lines that yield a measured redshift of z =0.290with a 2%accuracy.All these facts combined with optical image from Digitized Sky Survey allow us to conclude that RX J0046.4+4204is actually a distant galaxy cluster.In this paper,we assume the ΩM =0.3and ΩΛ=0.7cosmology with the Hubble constant of H 0=71km/s/Mpc.For the defined above cosmology and the measured redshift of z =0.290,the angular size of 1′corresponds the physical size of 257kpc.Statistical un-certainties are quoted at the 90%confidence level unless there is a statement saying otherwise.2.OBSERVATIONS AND DATA REDUCTIONIn the following analysis,we use the data from XMM-Newton observation of the XMM North 3Field of M31centered at RA =00h 46m 38s .00;Decl =+42◦16′20.0′′.Fig.1(Left )shows optical image of M31from Digitized Sky Survey with X MM-Newton FOV shown as a circle for M31North 3field.The XMM-Newton observation was performed on 2002June 29as a part of the Guar-anteed Time Program (PI:K.O.Mason).An analysis of the complete North 3field is presented in Trudolyubov2KOTOV,TRUDOLYUBOV,&VESTRAND42:35:59.959:59.941:23:59.947:59.940:11:59.9R.A. (J2000)D e c l . (J 2000)RXJ0046.4+4204XMM M31 North 347:45.616.80:46:48.019.250.445:21.625:11.920:59.916:47.942:12:35.908:23.904:11.9R.A. (J2000)D e c l . (J 2000)32541RXJ0046.4+4204Fig.1.—Left:Optical image of M31from Digitized Sky Survey with X MM-Newton FOV show as a circle for M31North 3field.Right:The combined MOS1-MOS2-PN vignetting-corrected image of the XMM North 3Field of M31in 0.8−2.5keV energy band,square root intensity scale.33.628.80:46:24.019.214.406:35.905:59.923.947.942:04:11.903:35.959.902:23.9R.A. (J2000)D e c l . (J 2000)Fig. 2.—The Palomar Digitized Sky Survey optical image with overlaid brightness contours made from 0.8-2.5keV band MOS2/PN image.The X-ray image was smoothed with a 4′′Gaus-sian kernel.The contour levels are based on the background noise going up in logarithmic steps.MOS1image was not used for con-struction of the contours because the center of the cluster falls at the edge of a CCD in MOS1camera.The angular size of the largest contour corresponds to ∼0.7Mpc at z =0.29.et al.(2005).In current analysis,we used the data from two EPIC-MOS detectors (Turner et al.2001)and the EPIC-PN detector (Str¨u der et al.2001).The EPIC data was re-duced using the standard XMM-Newton Science Analysis System (SAS v6.1.0)5.We used the calibration database with all updates available prior to January,2005.Only X-ray events corresponding to patterns 0-12for MOS de-tectors and patterns 0-4for PN detector were selected.All known bad pixels were excluded.5See http://xmm.vilspa.esa.es/external/xmm cal/sas.shtmlThe EPIC background is highly variable and only its quiescent component can be accurately modeled.To de-tect and exclude the periods of high flaring background,we produced the light curves for each EPIC detector showing count rate in the 2.0-15.0keV energy band from the whole field of view but with detected sources masked.The light curves were binned to 200s time resolution.We screened the EPIC data to recursively exclude time in-tervals with the deviation of the count rate exceeding a 2σthreshold from the average.Experiments with differ-ent choices of energy bands and flare detection thresh-olds have shown that our choice was close to optimal.The remaining good exposure time was ∼24ks for the EPIC-PN,∼41ks for the EPIC-MOS1,and ∼43ks for the EPIC-MOS2.To account for strong XMM mirror vignetting,we used an approach proposed by Arnaud et al.(2001).Each photon was assigned a weight proportional to inverse vi-gnetting and these weights were then used in computing images and spectra.This was done using the SAS tool evigweight .Background modeling in our analysis was implemented following the double-subtraction method of Arnaud et al.(2002).The first step of this method is to subtract the particle-induced background component.This compo-nent can be estimated from a set of XMM observations with the filter wheel closed (so called “closed data”).We compiled the closed dataset from public observations available in the XMM data archive;these data were re-duced following identical steps as the science observa-tions.The closed background was adjusted to the cluster observations using the observed flux in the 10–15keV band outside the field of view.The scaling factors are 1.04,1.05,and 1.01for MOS1,MOS2,and PN respec-tively.The second step is to determine the cosmic X-ray background (CXB)component.Its spatial distribution should be flat because vignetting correction is already ap-plied.Therefore,the CXB component can be measured in the source-free regions of the field of view at large radii from the cluster center (see §3.1and 3.2below).A GALAXY CLUSTER BEHIND M313TABLE1Results of Spatial binedEPIC-PN,MOS1and MOS2data,0.8−2.5keV energy range.Parameter errorsquoted are90%confidence limits.Parametersβfitα−βfitα···1.54±0.25β0.60±0.030.70±0.08r c(′′)20.2±2.856±16r c(kpc for z=0.290)87±12240±69χ2/dof156.7(119)139.4(118)Finally,we applied a correction for photons registered during the CCD readouts,so called out-of-time events, to the EPIC-PN data.3.RESULTSThe combined MOS1,MOS2,and PN image of the XMM North3Field of M31in the0.8−2.5keV en-ergy band,corrected for the effects of instrumental vi-gnetting,is shown in Figure1(Right).The raw image was convolved with a Gaussian function with spatial scale σ=4′′.We define the cluster center to be at the location of the X-ray surface brightness peak of the cluster emis-sion in the0.8−2.5keV energy band,α=00h46m24s.8δ=+42◦04′26′′(J2000),with an estimated uncertainty radius of6′′(90%CL),determined by the wavelet de-composition algorithm of Vikhlinin et al.(1998).The Palomar Digitized Sky Survey optical image shows no extended optical counterpart for RX J0046.4+4204(see Fig.2).During the following spectral and spatial analysis,we excluded all detectable point sources from the data.The sources were detected separately in the“optimal”0.3-3keV,“soft”0.3-0.8keV,and“hard”2.0-6.0keV energy bands using the task wvdecomp of the ZHTOOLS package 6.Detected point sources were masked with circles of 80%PSF power radii.3.1.Spatial AnalysisSpatial analysis of the cluster emission was performed in the0.8−2.5keV energy band using all detectors.Ex-periments with different energy bands showed that the signal to noise ratio was close to optimal for the chosen band.We used a pixel size of4′′in our spatial analysis. The image for each camera was corrected for vignetting. The PN image was corrected for out-of-time events.We subtracted the particle background component from the images as described in§2.We extracted the azimuthally averaged surface bright-ness profiles centered on the X-ray surface brightness peak,excluding the CCD gaps and circles around the point sources.The profiles were logarithmically binned with a step of∆r=0.1r.A logarithmic radial bin-ning approximately preserves the signal to noise ratio in annuli until the background becomes comparable to the signal.For the extracted profiles,the chosen step keeps the signal to noise ratio above3in annuli until∼5′.The obtained profiles were used to derive the parameters of 6See/∼alexey/zhtools/the spatial distribution of the ICM,clusterfluxes,and the level of the CXB component.The cluster surface brightness profiles are often mod-eled with the so calledβ-model,n2e∝(1+r2/r2c)−3βor S x∝(1+r2/r2c)−3β+0.5(Cavaliere&Fusco-Femiano 1976).However,this model poorly describes clusters with sharply peaked surface bigness profiles related to the radiative cooling of the ICM in the cluster centers. We used the simple modification of theβ-modeln2e∝(r/r c)−α4KOTOV,TRUDOLYUBOV,&VESTRAND1998).We estimated the count rate coming from the 70kpc central region using the bestfitα-βmodel.The total count rate was calculated integrating the observed surface brightness profile within1.4Mpc radius.We ob-tained f70kpc=18%.The combined MOS1-MOS2-PN surface brightness profile along with the bestfitα-βmodel are shown in Fig.3.From Fig.3one can see how the XMM PSF can flatten a peaked profile.For comparison,we alsofit the surface brightness profiles by the standardβ-model set-tingα=0.The obtained parameters are summarized in Table1.3.2.Global SpectrumFor our spectral analysis,we used the screened vignetting-corrected data in the0.5−10.0keV energy band from all cameras.The spectra of the cluster were extracted from a circular region with angular radius of 3.1′(0.8Mpc for z=0.290)for all EPIC data.All de-tected contaminating point-like sources were excluded from the source and background extraction regions.We subtracted the particle background component from the images as described in§2.To estimate the remaining CXB component,we extracted the spectra from a cir-cular region with angular radius of9.0′centered at the on-axis position,but a part of the region falling into the 9.65′(2.5Mpc for z=0.290)circle centered at the cluster center was excluded.The response matrices and effective areafiles were generated by the standard SAS tasks.Be-cause the data were previously vignetting corrected,the effective areafiles were created for the on-axis position using the routine arfgen.The response matrices were generated in the spectrum extraction region via rmfgen. The source spectra were binned to have at least30 counts in each spectral bin andfit in XSPEC11.3.0(Ar-naud1996)by the Mewe-Kaastra-Liedahl plasma emis-sion model(Mewe et al.1985).We used abundances from Anders&Grevesse(1989).Galactic photoelectric absorption was accounted for using the WABS model (Morrison&McCammon1983).The spectra from the EPIC-PN(3677counts),MOS1(1666counts)and MOS2 (2179counts)detectors werefitted both jointly and sep-arately.For the jointfits,only spectral model normaliza-tions were allowed to vary independently.The results of both joint and separate spectralfitting of the EPIC-PN, MOS1and MOS2data are summarized in Table2.We obtained a redshift value of0.290with2%accuracy. The redshift values estimated independently from EPIC-PN,MOS1and MOS2data are in good agreement within measurement errors.For the value of hydrogen column density we obtained N H=(2.2±0.20.2)×1021cm−2that is significantly above the Galactic hydrogen column in the direction of M31,∼7×1020cm−2(Dickey&Lockman 1990).We checked that the values of absorption obtained independently from EPIC-PN,MOS1and MOS2data are consistent within the measurement errors.The EPIC-PN,MOS1and MOS2spectra,along with the best-fit spectral models,are shown in Figure4.Note that separate spectralfitting of the EPIC-PN,MOS1and MOS2data gives consistent values of the model param-eters.3.3.Temperature profileTo construct temperature profile,we extracted indi-vidual spectra in5annuli:0−50kpc,50−100kpc, 100−200kpc,200−400kpc,and800−1600kpc,centered on the position of the X-ray surface brightness peak us-ing data from all cameras.After background subtraction, the number of counts(MOS1+MOS2+PN)in0.5-10kev energy band in1,2,3,4,and5annuli was632,1275, 2071,2069,1063,and459accordingly.The correction for the XMM PSF effect was done us-ing an approach of Pointecouteau et al.(2004).Using the best-fitα-βmodels of the cluster brightness and the XMM PSF calibration,we calculated the redistribution matrix,R ij,of each temperature to each annulus which represents relative contribution of emission from annulus i to the observedflux in annulus j.The model spectrum, S j,is then given byS j= R ij S(T i),(2) where T i is the temperature in annulus i and S(T i)is the mekal spectrum for this temperature.Fitting this model to the observed spectra in all annuli simultane-ously and treating all T i as free parameters gives the de-convolved temperature profile.Unfortunately,the statis-tic was poor to measure metallicity profile.So for all annuli wefixed the metallicity values at the best-fit val-ues obtained from the global spectrum.We checked that allowing the absorption and the metallicity to be freely fitted does not significantly change the result.The val-ues of redshift and absorption values were alsofixed at the best-fit values.Wefitted the observed temperature profiles by a3-D temperature model:T(r)=T0/(1.+(x/0.6)2)γ,(3) where x=r/r0.A similar model describes the tem-perature profiles for low redshift clusters(Vikhlinin et al.2005).For local clusters,r0scales with the clus-ter temperature as r0⋍0.50T1/2Mpc where T is in keV(see Fig.16in Vikhlinin et al.(2005)).For the best-fit temperature of T=5.5keV derived from the global spectrum,r0=1.17keV.Wefixed r0at the value suggested by the low redshift clusters only with an additional redshift scaling r0∝1/E(z),where E(z)=A GALAXY CLUSTER BEHIND M315Fig.3.—Left:Upper panel:The combined MOS1-MOS2-PN surface brightness profile of RX J0046.4+4204in the 0.8−2.5keV energy band.Black solid line shows α-βmodel best fit convolved with the XMM PSF.Black dashed line shows α-βmodel best fit.Black dotted line shows the XMM PSF.Black dotted-dashed line shows the level of the cosmic X-ray background component.Lower panel:The residual between the data and the best-fit model in terms of sigmas.Right:The same with the standard β-model bestfit.Fig. 4.—Left:Count spectra of RX J0046.4+4204taken with the XMM-Newton /EPIC-PN (black circles ),MOS1(blue triangles )and MOS2detectors (magenta squares ).The corresponding best-fit spectral models (absorbed red-shifted Mekal model)convolved with instrumental responses are shown as black ,red and blue solid lines.Right:Expanded view of the 4.0-7.0keV energy band.Upper,Middle and Lower panels show PN,MOS1and MOS2data.A red-shifted iron emission line feature is clearly evident.We applied Gaussian scatter to the observed tempera-ture profile within its uncertainties and fitted simulated profiles.The best-fit model uncertainties estimated as rms scatters in narrow radial bins from 1000simulations are shown in Fig.5..3.4.Mass and Luminosity measurementsAssuming hydrostatic equilibrium for the ICM,we can use the best fit temperature and density profiles to derive the total cluster masses:M (r )=−rT (r )d log r+d log T (r )6KOTOV,TRUDOLYUBOV,&VESTRANDTABLE 2Results of Spectral Analysis.EPIC-PN,MOS1and MOS2data,0.5−10.0keV energy range.Spectral extraction radius is 3.1′.Parameter errors quoted are 90%confidence limits .Parameters MOS1MOS2PNCombinedN H (1022cm −2)0.21±0.040.030.26±0.030.030.21±0.020.020.22±0.020.02kT (keV )6.4±1.41.15.3±0.80.75.1±0.70.55.5±0.50.5Z ⊙0.44±0.190.270.72±0.200.190.45±0.210.100.57±0.150.13z0.296±0.0110.0230.292±0.0090.0070.287±0.0080.0080.290±0.0050.005Fig.5.—Upper panel:Same as Fig.3(Left upper panel).Lowerpanel:The temperature profile of the cluster as a function of an-gular radius obtained from EPIC-PN/MOS1/MOS2data.Solid black circles show the deconvolved projected temperature profile.For comparison,open circles shows the raw measurement from the X-ray fit in the same annuli.The error bars correspond to 68%(1-σ)confidence limits.Solid line shows the best-fit projected tempera-ture profile and dashed lines correspond to its 68%CL uncertain-ties.1014M ⊙and the radius of r 500=1.09±0.08Mpc corre-sponding to the mean overdensity ∆=500relative to the critical density at the cluster redshift.The uncer-tainties on the total mass due to the temperature profile were calculated analogically to the uncertainties on the best-fit temperature profile model.We calculated the total mass for each simulated profile.Then the uncer-tainties were calculated as the boundaries of the region containing 90%of all realizations.The total mass un-certainties due to the error on the density gradient were calculated following Pratt &Arnaud (2002).The value of d log ρ(r )/d log r at r 500was considered as an indepen-dent parameter of the α-βmodel instead of β.We refit the surface brightness profile with the new parameter set and measured the uncertainties on d log ρ(r )/d log r at r 500.The final total mass uncertainties were calcu-lated by adding quadratically the total mass error due to the temperature profile and the density profile.The uncertainties on R 500are related to the total mass un-certainties as σM 500/M 500=3σr 500/r 500.We calculated the emission measure-weighted temper-atures (volume-averaged with weight w =ρ2g )of T emw =6.0±0.7within 70kpc <r <r 500.It is not a surprisethat obtained T emw is higher than the best-fit spectro-scopic temperature because we excluded the innermost bin from the temperature profile fit.The uncertainties for T emw were calculated from Monte-Carlo simulations in the similar way as for the temperature profile best-fit.We can compare the obtained M 500with predicted value from M −T relations.We will use as a low-redshift reference the M −T relation from Vikhlinin et al.(2005),which is similar to the M −T relation measured by Arnaud et al.(2005).The M −T relation derived by Vikhlinin et al.(2005)predicts M 500=6.1×1014M ⊙for a 6keV cluster.The self-similar theory predicts that for the same temperature the mass evolves as M ∆∝E (z )−1with z ,where E (z )=H (z )/H 0=(0.3(1+z )3+0.7)1/2(e.g.Bryan &Norman (1998)).Dividing the predicted mass by E (z =0.29)to place it at z =0.29,we obtain M 500=5.3×1014M ⊙.This value is in agreement with the derived M 500=(5.1±1.0)×1014M ⊙.We derived the unabsorbed bolometric luminosity ofL x =(8.4±0.5)×1044h −271erg s−1using the Mewe-Kaastra-Liedahl plasma emission model.We used T emw as the temperature parameter for the model.The model was normalized by the following count rate:we sub-tracted from the observed 0.8−2.5keV count rate cal-culated within 0<r <1.4Mpc the count rate calculated from the best-fit α-βmodel within r <70kpc,and mul-tiplied the result by 1.06(Markevitch 1998).The fact that our analysis is similar to the one done by Markevitch (1998)for low redshift clusters allows us to compare the obtained luminosity with the prediction from their L −T relation.The L −T relation derived by Markevitch (1998)predicts the luminosity of 6.2×1044h −271erg s−1for a 6keV cluster.It is well below the derived luminosity.However,correcting the predicted luminosity for the evolution in L −T relation,L z =(1+z )1.5L 0,reported by Vikhlinin et al.(2002),we obtain9.2×1044h −271erg s −1.This value is close to the observed one.4.DISCUSSION AND CONCLUSIONSOur deep XMM-Newton observations of M31have shown that RX J0046.4+4204is not located in that galaxy,but rather is actually a distant cluster of galax-ies.We found RX J0046.4+4204has spatially extended X-ray emission and that the spectrum clearly shows a red-shifted iron emission line.Straightforward fitting of the iron line yields a cluster redshift of Z=0.290with 2%accuracy and that the redshift values estimated in-dependently from EPIC-PN and MOS data are in good agreement within measurement errors.The large scale spatial distribution of RXA GALAXY CLUSTER BEHIND M317J0046.4+4204is well fit by the α-βmodel with β=0.70±0.8,a core radius r c =(56±16)′′or r c =(240±69)kpc for z =0.290,and α=1.54±0.25.The obtained values of βand r c are consistent with the parameters of typical clusters (Vikhlinin et al.1999).The derived αagrees with values measured for cooling flow clusters (Vikhlinin et al.2005).The spatially integrated X-ray continuum is well fit by red-shifted (z=0.290)Mewe-Kaastra-Liedahl plasma emission model with low energy photo-electric absorp-tion.The best fitting global model for the joint PN,MOS1,and MOS2measurements yields the parame-ters kT (keV )=5.5±0.5,fractional solar abundanceZ ⊙=0.57±0.150.13,,a redshift z =0.290±0.005,and a column depth n H =2.2(±0.20.2)×1021cm−2.This derived column depth is significantly larger than the Galactic hy-drogen column in the direction of M31,∼7×1020cm −2(Dickey &Lockman 1990).It is interesting to compare the value of absorption obtained for the cluster with absorption for the nearby X-ray sources.Fig 1(Right )shows five bright X-ray sources lying in the direct vicinity of RXJ0046.4+4204for which Trudolyubov et al.(2005)were able to measure column depth based on their X-ray spectra and to iden-tify some of them.XMMU J004540.5+420806(source #5in Fig.1)was identified as a foreground star with the column depth of 0.2×1021cm −2(two sigma upper limit).XMMU J004648.0+420851(source #3)was identified as a background radio source withthe column depth of (4.0±1.05.0)×1021cm−2.XMMU J004627.0+420151(source #1)was identified as a glob-ular cluster source in M31.The derived column depth for this source was (1.3±0.1)×1021cm −2.The nature of two last sources,XMMU J004611.5+420826(source #4)and XMMU J004703.6+420449(source #2),was unclear.Trudolyubov et al.(2005)proposed that these sources could be two AGN located in the back-ground of M31.The derived column depths of XMMU J004611.5+420826and XMMU J004703.6+420449were (2.5±0.3)×1021cm −2and (2.3±0.6)×1021cm −2,re-spectively.The absorption value of the background radio source is 2times higher than the cluster absorption and most likely to be intrinsic.The column depth of the glob-ular cluster candidate is smaller than that of the clus-ter.To explain this fact,it might be proposed that the globular cluster candidate is located in front of the disk of M31,while the cluster is obscured by the disk.On the other hand,XMMU J004611.5+420826and XMMU J004703.6+420449have column depths consistent with that of the cluster,suggesting that they could be also located in the background of M31,although it is unclear what fraction of the column depths is intrinsic.This in-terpretation is in general agreement with results of spec-tral fitting of a large sample of M31globular cluster X-ray sources (Trudolyubov &Priedhorsky 2004).Tru-dolyubov &Priedhorsky (2004)found that globular clus-ter sources located in front of M31disk have typical val-ues of absorbing column in the range of (0.5−1.5)×1021,while the sources located behind the disk or embedded into it show higher absorbing columns of (2−4)×1021cm −2.The extracted temperature profile corrected for the XMM PSF shows the central temperature decline that confirms the indication of cooling flow presence from the spatial ing the spatially resolve tem-perature profile we derived R 500=1.09±0.07Mpc,M 500=(5.3±1.0)×1014M ⊙,and T emw =6.0±0.7within R 500.The values of M 500corrected for the evolution and T emw are in agreement with local M −T relations.The study we have presented here shows the utility of sensitive X-ray observations for identifying and studying clusters of galaxies in directions where foreground confu-sion or heavy optical extinction makes optical selection complicated.We thank A.Vikhlinin for extensive discussions and helpful comments.Also we’d like to thank the referee for providing useful suggestions.This paper is based in part on observations obtained with XMM-Newton ,an ESA science mission with instruments and contributions directly funded by ESA Member States and the USA (NASA).This work was supported in part by NASA grant NAG5-12390and by Internal Laboratory Directed Research and Development funding at Los Alamos Na-tional Laboratory.REFERENCESArnaud,K.,1996,in Astronomical Data Analysis Software and Systems V,ASP Conference Series 101,ed.G.Jacoby &J.Barnes (San Francisco:ASP)17Anders,E.,&Grevesse,N.1989,Geochim.Cosmochim.Acta,53,197Arnaud,M.,Neumann,D.M.,Aghanim,N.,et al.,2001,A&A,365,L80Arnaud,M.,Majerowicz,S.,Lumb,D.,et al.,2002,A&A,390,27AArnaud,M.,Pointecouteau, E.,&Pratt,G.W.2005,A&A,submitted (astro-ph/0502210)Bryan,G.L.,&Norman,M.L.1998,ApJ,495,80Cavaliere,A.,Fusco-Femiano,R.,1976,A&A,49,137Dickey,J.M.&Lockman,F.J.1990,ARA&A,28,215Ebeling,H.,Mullis,C.R.,&Tully,R.B.2002,ApJ,580,774Evrard,A.E.1997,MNRAS,292,289Kirsch,M.,2005,EPIC Status of Calibration and Data Analyis,http://xmm.vilspa.esa.es/docs/documents/CAL-TN-0018-2-4.pdfMarkevitch,M.,1998,ApJ,504,27Mazzotta,P.,Rasia, E.,Moscardini,L.,&Tormen,G.2004,MNRAS,354,10Mewe,R.,Gronenschild,E.H.B.M.,&van den Oord,G.H.J.1985,A&AS,62,197Morrison,R.,&McCammon,D.1983,ApJ,270,119Nevalainen,J.,Lumb,D.,dos Santos,S.,Siddiqui,H.,Stewart,G.,and Parmar,2001,A.N.,A&A,374,66-72Peres,C.B.,Fabian,A.C.,Edge,A.C.,Allen,S.W.,Johnstone,R.M.,&White,D.A.1998,MNRAS,298,416Pointecouteau,E.,Arnaud,M.,Kaastra,J.,&de Plaa,J.2004,A&A,423,33Pratt,G.W.,Arnaud,M.,2002,A&A,394,375Press,W.H.,&Schechter,P.1974,ApJ,187,425Rosati,P.,Borgani,S.,&Norman,C.2002,ARA&A,40,539Sarazin,C.L.1988,Cambridge Astrophysics Series,Cambridge:Cambridge University Press,1988,Supper,R.,Hasinger,G.,Lewin,W.H.G.,et al.,2001,A&A,373,63Str¨u der,L.,et al.2001,A&A,365,L1Trudolyubov,S.,&Priedhorsky,W.2004,ApJ,616,821Trudolyubov,S.,Kotov,O.,Priedhorsky,W.,Cordova, F.,&Mason,K.2005,ApJ,634,314Turner,M.J.L.,et al.2001,A&A,365,L27Vikhlinin,A.,McNamara,B.R.,Forman,W.,Jones,C.,Quintana,H.,&Hornstrup,A.1998,ApJ,502,558Vikhlinin,A.,Forman,W.,&Jones,C.1999,ApJ,525,47Vikhlinin,A.,VanSpeybroeck,L.,Markevich,M.et al.,2002,ApJ 578,L107。

a r Xiv:as tr o-ph/49638v223Nov24The astrophysics of cataclysmic variables and related objects ASP Conference Series,Vol.?,2004J.M.Hameury &sota What can XMM-Newton tell us about the spin periods of Intermediate Polars?P.A.Evans &Coel Hellier Astrophysics Group,School of Chemistry and Physics,Keele University,Keele,Staffordshire,ST55BG Abstract.XMM-Newton ’s unprecedented combination of spectral resolution and high throughput allows us to perform the best phase-resolved X-ray analysis of intermediate polars to date.The Optical Monitor gives optical/UV photom-etry simultaneously with the X-ray data.We present a comprehensive study of X-ray spin pulses in IPs,giving spin-pulses and hardness ratios for every IP looked at with XMM-Newton to date.1.Introduction In an intermediate polar (IP)the white dwarf has a magnetic field strong enough to affect the accretion flow,but not strong enough to synchronise the white dwarf rotation with the binary period;the magnetic dipole is also inclined to the spin axis of the star.Generally accretion occurs from a magnetically truncated accretion disc;at some radius the disc material threads to the white dwarf’s field lines and is channeled towards the gravitationally preferable pole in large accretion curtains.Stand-offshocks form above each pole,resulting in hard X-ray emission.As the white dwarf rotates,our view of these regions changes and the position of the accretion curtains relative to our line of sight also varies,giving rise to spin-period modulations (e.g.Hellier,Cropper,&Mason 1991).To date eleven IPs have been observed with XMM-Newton .In this paper we present a compilation of X-ray spin-pulse profiles,hardness ratios,and UV data from XMM ’s Optical Monitor (OM)for these systems.2.AO Piscium and V1223SagittariiAO Psc and V1223Sgr show sinusoidal X-ray and UV modulations (Fig.1).The fact that their spectra get harder at pulse minimum suggests that the modulation is caused by absorption.These are thus good examples of the accretion curtain model for X-ray modulation in IPs (e.g.Hellier et al.1991).3.HT CamelopardarisHT Cam’s X-ray spin-pulse (Fig.2)appears sinusoidal with a flattened maxi-mum,and,judged from the softness ratio,appears to be energy independent.No pulse is detected in the UV,although the B -band shows sinusoidal variation.For more information,see de Martino et al.(2004b).12Evans and HellierFigure1.The spin pulses of AO Psc(left)and V1223Sgr(right).We showthe0.2–12keV X-ray data,the(0.2–4)/(6–12)keV softness ratio and(whereavailable)the UV(2050–2450˚A)data.4.RX J1548.2−4528The recently discovered IP RX J1548.2−4528shows a sinusoidal X-ray spin-pulse(Fig.2).Haberl,Motch&Zickgraf(2002)analysed this observation,and suggested that while there may be a small change in absorption with spin phase, the predominant source of modulation is a change in the visibility of the emission.5.EX HydraeEX Hya’s softness ratio has the same shape and phasing as the lightcurve (Fig.3),making it tempting to interpret this modulation as changing absorption in the accretion curtains(indeed,it was for this system that accretion curtains werefirst proposed as the cause of X-ray modulation;Rosen,Mason,&C´o rdova 1988).However,various X-ray studies(e.g.Allan,Hellier,&Beardmore1998) have suggested that the upper emitting pole is periodically occulted by the white dwarf,giving rise to the modulation.Since such occultation will affect the lower, cooler parts of the accretion column more than the higher,hotter ones,we see a deeper modulation at lower energies.6.AE AquariiAE Aqr differs from many IPs as it is a rapid rotator(P spin∼33s),and thus its magneticfield might be expelling material from the system like a propeller(e.g., Eracleous&Horne1996).It is unclear whether the X-ray emission arises fromXMM-Newton IP spin pulses3Figure2.As for Fig.1,but for HT Cam(left)and RX J1548.2−4528(right). The UV band for HT Cam is1800–2250˚A,and the B band(3900–4900˚A)is also shown.Figure3.As Fig.1,but for EX Hya(left)and AE Aqr(right).4Evans and HellierFigure4.Left:As for Fig.1,but for GK Per.Right:Spin-pulse profile ofV405Aur.The upper panel shows the0.2–0.7keV band and the lower panelshows the0.7–12keV band.accretion,as is usual in IPs,or further out in the magnetosphere,as suggested by Ikhsanov(2001).However,the XMM-Newton data shows that the pulse profile is sinusoidal and largely independent of energy.7.GK Perseii(in outburst)Watson,King,&Osborne(1985)suggested that the X-ray spin-pulse of GK Per in outburst was caused by increased absorption at spin minimum.Ishida et al.(1992)blamed changing absorption for the profile in quiescence as well,but suggested that the outburst profile may show evidence for occultation of the upper pole.Fig.4shows that the outburst softness ratio follows the lightcurve, making it more likely that absorption changes are responsible for this pulse,as Watson et al.(1985)claimed.Hellier,Harmer,&Beardmore(2004)agree with this,in analysis of an RXTE observation.8.V405AurigaeV405Aur differs from IPs such as AO Psc,first because it shows a soft blackbody component to its X-ray emission(Haberl et al.1994),and second because the blackbody is double-peaked on the spin period,while the harder emission is single-peaked but sawtoothed(Fig.4).Furthermore,de Martino et al.(2004a) and Evans&Hellier(2004)have shown that the absorption does not change with spin phase.Evans&Hellier(2004)suggest that if the angle between the magnetic and spin axes in this system is very high,then the double-peaked profile of the blackbody emission is explained.XMM-Newton IP spin pulses5Figure5.Left:as Fig.1,but for FO Aqr.Right:PQ Gem spin folds forthe blackbody(upper panel),hard X-ray(lower panel)and UV(2450–3200˚A)components.9.FO AquariiThe X-ray pulse profile of FO Aqr(Fig.5)shows X-ray minimum to occur after UV minimum.Many authors(e.g.Hellier1993;Beardmore et al.1998) have identified the‘notch’at phase0.7as the result of occultation of the upper accretion column,and the dip around phase0.5as arising from the accretion curtains intercepting our line of sight.Evans et al.(2004)note that,if this is the case,the upper pole will be pointed towards the observer nearly a quarter of a cycle before the accretion curtain dip.They thus suggest that the accretion curtains are twisted,explaining the lightcurve and softness ratio.10.PQ GeminorumPQ Gem shows a soft blackbody component as well as the hard,optically thin emission characteristic of IPs(Mason et al.1992).The spin-pulse profile(Fig.5) differs greatly from that of AO Psc,with UV maximum coinciding with X-ray minimum.Potter et al.(1997)and Mason(1997)have suggested that the blackbody modulation is caused by changing views of the accretion region as the white dwarf rotates.Maximum(phase0)occurs when the upper pole is towards us.Since this occurs after the dip around phase0.8,interpreted as absorption by the accretion curtains,Potter et al.(1997)and Mason(1997)suggest that PQ Gem accretes preferentially alongfield lines preceding the magnetic pole.6Evans and HellierFigure6.Power spectrum of V2400Oph.The spin(ω)and beat(ω−Ω)frequencies are marked.11.V2400OphiuchiV2400Oph is thought to be the only discless IP,since it shows a dominant X-ray modulation on the spin-orbit beat period(Buckley et al.1995;Hellier& Beardmore2002).We have recently discovered evidence for a spin pulse during an XMM-Newton observation(Fig.6),thefirst in the X-ray band.ReferencesAllan,A.,Hellier,C.,&Beardmore,A.P.1998,MNRAS,295,167Beardmore,A.P.,Mukai,K.,Norton,A.J.,Osborne,J.P.,Hellier,C.1998,MNRAS, 297,337Buckley, D.A.H.,Sekiguchi,K.,Motch, C.,O’Donoghue, D.,Chen, A.-L., Schwarzenberg-Czerny,A.,Pietsch,W.,Harrop-Allin,M.K.1995,MNRAS,275, 1028Eracleous,M.,&Horne,K.1996,ApJ,471,427Evans,P.A.,Hellier,C.,Ramsay,G.,Cropper,M.2004,MNRAS,349,715Evans,P.A.,&Hellier,C.2004,MNRAS,353,447Haberl,F.,Thorstensen,J.R.,Motch,C.,Schwarzenberg-Czerny,A.,Pakull,M.,Sham-brook,A.,Pietsch,W.1994,A&A,291,171Haberl,F.,Motch,C.,&Zickgraf,F.-J.2002,A&A,387,201Hellier,C.,Cropper,M.,&Mason,K.O.1991,MNRAS,248,233Hellier,C.,&Beardmore,A.P.,2002,MNRAS,331,407Hellier,C.,Harmer,S.,&Beardmore,A.P.2004,MNRAS,349,710Ikhsanov,N.R.2001,A&A,374,1030Ishida,M.,Sakao,T.,Makishima,K.,Ohashi,T.,Watson,M.G.,Norton,A.J.,Kawada, M.,Koyama,K.1992,MNRAS,254,647de Martino,D.,Matt,G.,Belloni,T.,Haberl,F.,Mukai,K.2004a,A&A,415,1009 de Martino,D.,Matt,G.,Mukai,K.,Bonnet-Bidaud,J.M.,G¨a nsicke,B.T.,Haberl,F.,Mouchet,M.,2004b,in Hameury,J.M.,&Lasota,J.P.,eds.,The Physics ofCataclysmic Variables and Related Objects,ASP Conf.Ser,in pressMason,K.O.,et al.1992,MNRAS,258,749Mason,K.O.1997,MNRAS,285,493Potter,S.B.,Cropper,M.,Mason,K.O.,Hough,J.H.,Bailey,J.A.1997,MNRAS,285, 82Rosen,S.R.,Mason,K.O.,&C´o rdova,F.A.1988,MNRAS,231,549Watson,M.G.,King,A.R.,&Osborne,J.P.,1985,MNRAS,212,917。

a r X i v :a s t r o -p h /0409627v 1 27 S e p 2004Astronomy &Astrophysics manuscript no.(will be inserted by hand later)XMM-Newton observation of the most X-ray-luminous galaxycluster RX J1347.5−1145Myriam Gitti and Sabine SchindlerInstitut f¨u r Astrophysik,Leopold-Franzens Universit¨a t Innsbruck,Technikerstraße 25,A-6020Innsbruck,Austria Received /AcceptedAbstract.We report on an XMM-Newton observation of RX J1347.5−1145(z =0.451),the most luminous X-ray cluster ofgalaxies currently known,with a luminosity L X =6.0±0.1×1045erg /s in the [2-10]keV energy band.We present the first temperature map of this cluster,which shows a complex structure.It identifies the cool core and a hot region at radii 50-200kpc to south-east of the main X-ray peak,at a position consistent with the subclump seen in the X-ray image.This structure is probably an indication of a submerger event.Excluding the data of the south-east quadrant,the cluster appears relatively relaxed and we estimate a total mass within 1.7Mpc of 2.0±0.4×1015M ⊙.We find that the overall temperature of the cluster is kT =10.0±0.3keV .The temperature profile shows a decline in the outer regions and a drop in the centre,indicating the presence of a cooling core which can be modelled by a cooling flow model with a minimum temperature ∼2keV and a very highmass accretion rate,˙M∼1900M ⊙/yr.We compare our results with previous observations from ROSAT ,ASCA and Chandra .Key words.Galaxies:clusters:particular:RX J1347.5−1145–X-ray:galaxies:clusters –cooling flows1.IntroductionIn this paper we present the first results from an XMM-Newtonobservation of RX J1347.5−1145,the most X-ray-luminous galaxy cluster known (Schindler et al.1995).This cluster has been detected in the ROSAT All-Sky Survey and further stud-ied with ROSAT HRI and ASCA (Schindler et al.1995,1997).It shows a very peaked X-ray emission profile and presents a strong cooling flow in its central region.Submm observations in its direction showed a very deep SZ decrement (Komatsu et al.1999,2001;Pointecouteau et al.1999,2001).Due to the presence of gravitational arcs,this cluster is also well suited for a comparison of lensing mass and X-ray mass.Optical stud-ies of weak lensing have been performed by Fischer &Tyson (1997)and Sahu et al.(1998).Recent Chandra observations (Allen et al.2002)discovered a region of relatively hot,bright X-ray emission,located approximately 20arsec to the south-east of the main X-ray peak at a position consistent with the region of enhanced SZ e ffect.This could be an indication for a subcluster merger,pointing to a complex dynamical evolution.A comparison of the XMM-Newton and SZ results,a more de-tailed analysis of the complex dynamical state of the cluster and a comparison of lensing mass and X-ray mass will be presented in a forthcoming paper (Gitti et al.in prep.).RX J1347.5−1145is at a redshift of 0.451.With H 0=70km s −1Mpc −1,and ΩM =1−ΩΛ=0.3,the luminosity distance is 2506Mpc and 1arcsec corresponds to 5.77kpc.2Gitti&Schindler.:XMM-Newton observation of the most X-ray-luminous galaxy cluster RX J1347.5−1145Fig.1.Total(MOS+pn)XMM-Newton EPIC mosaic image of RX J1347.5−1145in the[0.9-10]keV energy band.The im-age is corrected for vignetting and exposure and is adaptively smoothed.(SE)of the X-ray peak and,on large scale(∼80arcsec),the extension of the X-ray emission to the south,already revealed inprevious observations with Chandra(Allen et al.2002).In Fig.2we show an overlay of the VLT image of the central region of RX J1347.5−1145with the X-ray contours derived from Fig.1.Fig.2.VLT image of the central region of RX J1347.5−1145 (Erben et al.in prep.).Superposed are the([0.9-10]keV)XMM X-ray contours(levels:0.003,0.015,0.045,0.06,0.15,0.3,0.6, 1.5,3cts/s/arcmin2).The image is∼4.4×4.3arcmin2(North is up,East is left).We compute a background-subtracted vignetting-corrected radial surface brightness profile in the[0.3-2]keV energy band for each camera separately.The profiles for the three detectors are then added into a single profile,binned such that at least a sigma-to-noise ratio of3was reached.The cluster emission is detected up to1.7Mpc(∼5′).In Fig.3we show the X-ray surface brightness profiles for the disturbed SE quadrant com-pared to that from data excluding the SE quadrant.We note that the data excluding the SE quadrant(hereafter undisturbed clus-ter)appear regular,while those for the SE quadrant(containing the X-ray subclump)show a clear excess of emission between radii of∼100and300kpc relative to other directions.0.110.0010.010.1110Fig.3.Background subtracted,azimuthally averaged radial surface brightness profile for SE quadrant data in the[0.3-2] keV range.The dotted line shows the profile in other directions (undisturbed cluster),which appears relatively regular and re-laxed.An excess of emission in the SE quadrant between radii of∼20-50arcsec(100-300kpc)is visible.The surface brightness profile of the undisturbed cluster isfitted in the CIAO tool Sherpa with various models,which are convolved with the XMM-Newton PSF.A singleβ-model (Cavaliere&Fusco Femiano1976)is not a good description of the entire profile:afit to the outer regions(350kpc-1.7 Mpc)shows a strong excess in the centre when compared to the model.The peaked emission is a strong indication for a cooling core in this cluster.We found that for350kpc-1.7Mpc the data can be described by aβ-model with a core radius r c=367±3 kpc and a slope parameterβ=0.93±0.01,while for r<350kpc the data can be approximated by aβ-model with r c=40±0.2 kpc andβ=0.55±0.02(90%confidence levels).4.Temperature mapThe temperature image of the central cluster region shown in Fig.4is build from X-ray colours.Specifically,we produce the mosaics of MOS images in four different energy bands([0.3-1]keV,[1-2]keV,[2-4.5]keV and[4.5-8]keV),subtract the background and divide the resulting images by the exposure maps.A temperature is obtained byfitting the values in each pixel with a thermal plasma.In particular we note that the very central region appears cooler than the surrounding medium and the SE quadrant,which corresponds to the subclump seen in the X-ray image,is significantly hotter than the gas in other directions.5.Spectral analysisFor the spectral analysis we treat the SE quadrant containing the X-ray subclump separately from the rest of the cluster.The data for the undisturbed cluster are divided into the annular re-gions detailed in Table1.A single spectrum is extracted for each region and then regrouped to contain a minimum of25Gitti &Schindler.:XMM-Newton observation of the most X-ray-luminous galaxy cluster RX J1347.5−11453Fig.4.Temperature map obtained by using 4X-ray colours ([0.3-1],[1-2],[2-4.5],[4.5-8]keV)and estimating the ex-pected count rate with XSPEC for a thermal MEKAL model,with fixed Galactic absorption N H =4.85×1020cm −2and metallicity Z =0.3Z ⊙.Superposed are the X-ray contours.The features outside the last contours are not significant,as they are mainly due to noise fluctuations.counts per channel,thereby allowing χ2statistics to be used.The data from the three cameras were modelled simultaneously using the XSPEC code,version 11.3.0.Spectral fitting is per-formed in the [0.5-8]keV band.The spectra are modelled using a simple,single-temperature model (MEKAL plasma emission code in XSPEC)with the absorbing column density fixed to the Galactic value (N H =4.85×1020cm −2,Dickey &Lockman 1990).The free parameters in this model are the temperature kT ,metallicity Z (measured relative to the solar values)and normalization (emission measure).The best-fitting parameter values and 90%confidence lev-els derived from the fits to the annular spectra are summarized in Table 1.The projected temperature profile determined with this model is shown in Fig.5.The temperature rises from a mean value of 8.8±0.3keV within 115kpc to kT =11.1±0.4keV over the 0.1-0.5Mpc region,then declines down to a meanvalue of 6.0+2.6−1.6keV in the outer regions (1.0-1.7Mpc).In Fig.5we also show for comparison the projected temperature profile measured by Chandra (Allen et al.2002).We note that while the general trend observed by the two satellites is consistent,there are some discrepancies in the measurements of the ab-solute temperature values.The discrepancy between Chandra and XMM temperature profile has been found in other clusters of galaxies (e.g.A1835,Schmidt et al.2001,Majerowicz et al.2002),and can be partially due to the e ffect of the XMM PSF (see Markevitch 2002).A fit with the same model to the data for the SE quadrant between radii 50-200kpc yields a best-fitting temperature kT =13.3±1.0keV .In other directions,the mean value is kT =11.0+0.5−0.4keV .The metallicity profile derived with the single-temperature model is consistent with being constant,with an overall value of Z =0.26±0.04Z ⊙.However,as shown in Table 2,the structure in the innermostTable 1.The results from the spectral fitting in concentric annularregions (undisturbed cluster).Temperatures (kT )are in keV ,metallic-ities (Z )in solar units and [2-10]keV luminosities (L X )in units of 1044erg s −1.The total χ2values and numbers of degrees of freedom (DOF)in the fits are listed in column 5.Errors are 90%confidence levels (∆χ2=2.71)on a single parameter of interest.0-1158.9+0.3−0.30.34+0.05−0.0516.7982/880115-23010.7+0.7−0.60.26+0.08−0.0810.1696/664230-34511.9+1.6−1.30.16+0.14−0.155.44433/384345-52010.7+1.1−1.00.24+0.13−0.134.52350/341520-6909.0+1.4−1.10.16+0.18−0.161.92239/210690-10409.4+2.1−1.40.19+0.26−0.191.64293/2641040-17306.0+2.6−1.70.40+0.50−0.370.91593/4210-17309.4+0.3−0.30.26+0.04−0.0441.81957/14524Gitti&Schindler.:XMM-Newton observation of the most X-ray-luminous galaxy cluster RX J1347.5−1145(e.g.Kaastra et al.2004).Within1Mpc wefind a total mass of1.0±0.2×1015M⊙,in agreement with Chandra(Allen et al.2002)and weak lensing analysis(Fischer&Tyson1997) results and slightly higher than that derived by ROSAT/ASCA (Schindler et al.1997).In Fig.6we also show for comparison (dashed line)the mass profile derived by assuming a constant temperature of9.5keV.Fig.6.Solid line:Profile of the integrated total mass.Dashed line:Profile of the integrated total mass calculated assuming a constant temperature of9.5keV.Dotted line:Error on the mass calculation coming from the temperature measurement.7.Cooling core analysisWe accumulate the spectrum in the central30′′(∼175kpc)by excluding the data for the SE quadrant.We use three different spectral models.Model A is the MEKAL model already used in Sect.5.Model B includes a single temperature component plus an isobaric multi-phase component(MEKAL+MKCFLOW in XSPEC),where the minimum temperature,kT low,and the normalization of the multi-phase component,Norm low=˙M, are additional free parameters.Finally,in model C the con-stant pressure coolingflow is replaced by a second isother-mal emission component(MEKAL+MEKAL in XSPEC).As for model B,this model has2additional free parameters with respect to model A:the temperature,kT low,and the normal-ization,Norm low,of the second component.The results,sum-marized in Table2,show that the statistical improvements ob-tained by introducing an additional emission component(mod-els B or C)compared to the single-temperature model(model A)are significant at more than the99%level according to the F-test,although the temperature of the hot gas is unrealisti-cally high.With our data,however,we cannot distinguish be-tween the two multi-phase models.This means that the extra emission component can be equally well modelled either as a coolingflow or a second isothermal emission component.We note that thefit with the coolingflow model sets tight con-straints on the existence of a minimum temperature(∼2keV). The nominal mass deposition rate in this empirical model is Table 2.The best-fit parameter values and90%confi-dence limits from the spectral analysis in the central0-30′′region.Temperatures(kT)are in keV,metallicities(Z)as a fraction of the solar value and normalizations in units of10−14n e n p V/4πD A(1+z)2as done in XSPEC(for the MKCFLOW model the normalization is parameterized in terms of the mass deposition rate˙M,in M⊙yr−1).kT9.2+0.3−0.323.8+6.1−4.717.7+5.7−3.6 Z0.32+0.05−0.050.39+0.06−0.060.42+0.06−0.06 Norm0.00528+0.00009−0.000080.00053+0.00223−0.000530.00355+0.00051−0.00048 kT low—2.0+0.5−0.43.9+0.7−0.6 Norm low—˙M=1880+260−2100.00194+0.00057−0.00058χ2/DOF1048/10031011/10011010/1001Gitti&Schindler.:XMM-Newton observation of the most X-ray-luminous galaxy cluster RX J1347.5−11455 Arnaud M.,Majerowicz S.,Lumb D.et al.2002,A&A,390,27Cavaliere A.,&Fusco-Femiano R.1976,A&A,49,137Dickey J.M.,&Lockman F.J.1990,ARA&A,28,215Fischer P.,&Tyson J.A.1997,AJ,114,14Kaastra J.S.,Tamura T.,Peterson J.R.et al.2004,A&A,413,415Komatsu E.,Kitayama T.,Suto Y.et al.1999,ApJ,516,L1Komatsu E.,Matsuo H.,Kitayama T.et al.2001,PASJ,53,57Majerowicz S.,Neumann D.M.,&Reiprich T.H.2002,A&A,394,77Markevitch M.2002,astro-ph/0205333Pointecouteau E.,Giard M.,Benoit A.et al.1999,ApJ,519,L115Pointecouteau E.,Giard M.,Benoit A.et al.2001,ApJ,552,42Sahu,K.C.,Shaw R.A.,Kaiser M.E.et al.1998,ApJ,492,L125Schindler S.,Guzzo L.,Ebeling H.et al.1995,A&A,299,L9Schindler S.,Hattori M.,Neumann D.M.et al.1997,A&A317,646Schmidt R.W.,Allen S.W.,&Fabian A.C.2001,MNRAS,327,1057。

牛顿万有引力英语作文From the moment Isaac Newton first formulated the law of universal gravitation in the late 17th century, our understanding of the cosmos has been irrevocably altered. "What goes up must come down," is a common phrase that encapsulates the essence of this fundamental force, which not only governs the motion of celestial bodies but also the very fabric of our daily existence. The gravity that keeps ourfeet firmly on the ground is the same force that holds the moon in orbit around the Earth, and the Earth in orbit around the sun.Newton's insight was nothing short of revolutionary. It was a time when the heavens were still shrouded in mystery, and his law provided a unifying principle that could explain the motion of planets and the fall of an apple with the same mathematical precision. The gravity that Newton described is an attractive force that acts between all objects with mass, and its strength is directly proportional to the product of their masses and inversely proportional to the square of the distance between their centers.The implications of this law are vast and far-reaching. It has been instrumental in the development of modern astronomy and space exploration. Without it, we would not have been able to send satellites into orbit, nor would we have been able to land rovers on distant planets or plan interstellar missions. The law of universal gravitation isthe cornerstone of our understanding of the gravitational field, which in turn is crucial for the study of black holes, the expansion of the universe, and the very origins of the cosmos itself.Moreover, Newton's law has had a profound impact on our philosophical understanding of the universe. It suggests a universe that is governed by a set of consistent and discoverable laws, rather than one that is capricious or subject to the whims of the divine. This has fueled the scientific method and the quest for knowledge that continues to this day.In essence, Newton's law of universal gravitation is more than just a scientific theory; it is a testament to the power of human curiosity and the relentless pursuit of understanding. It reminds us that the world around us, from the smallest pebble to the most distant star, is bound by the same principles, and that by studying these principles, we can unlock the secrets of the universe.。

a r X i v :a s t r o -p h /0502362v 1 18 F eb 2005Submited September 30,2004,accepted for publication in the Astrophysical Journal on December 8,2004Preprint typeset using L A T E X style emulateapj v.11/12/01XMM-NEWTON AND GEMINI OBSERVATIONS OF EIGHT RASSCALS GALAXY GROUPS 1Andisheh MahdaviInstitute for Astronomy,University of HawaiiAlexis Finoguenov and Hans B ¨ohringer Max-Planck-Institut f¨u r Extraterrestrische PhysikMargaret J.GellerHarvard-Smithsonian Center for AstrophysicsandJ.Patrick HenryInstitute for Astronomy,University of HawaiiSubmited September 30,2004,accepted for publication in the Astrophysical Journal on December 8,2004ABSTRACTWe study the distribution of gas pressure and entropy in eight groups of galaxies belonging to the ROSAT All-Sky Survey /Center for Astrophysics Loose Systems (RASSCALS).We use archival and proprietary XMM-Newton observations,supplementing the X-ray data with redshifts derived from the literature;we also list 127new redshifts measured with the Gemini North telescope.The groups are morphologically heterogeneous in both the optical and the X-ray,and several suffer from superpositions with background galaxies or clusters of galaxies.Nevertheless,they show remarkable self-similarity in their azimuthally averaged entropy and temperature profiles.The entropy increases with radius;the behavior of the entropy profiles is consistent with an increasing broken power law with inner and outerslope 0.92+0.04−0.05and 0.42+0.05−0.04(68%confidence),respectively.There is no evidence of a central,isentropic core,and the entropy distribution in most of the groups is flatter at large radii than in the inner region,challenging earlier reports as well as theoretical models predicting large isentropic cores or asymptotic slopes of 1.1as r →∞.The pressure profiles are consistent with a self-similar decreasing broken powerlaw in radius;the inner and outer slopes are −0.78+0.04−0.03and −1.7+0.1−0.3,respectively.The results suggest that the larger scatter in the entropy distribution reflects the varied gasdynamical histories of the groups;the regularity and self-similarity of the pressure profiles is a sign of a similarity in the underlying dark matter distributions.1.introductionGroups are the intermediate structures between galax-ies and clusters of galaxies,and thus are crucial to an un-derstanding of how galaxies evolve in dense environments (Zwicky 1937;Zwicky &Humason 1960;Hickson 1982;Geller &Huchra 1983;Ramella,Geller,&Huchra 1989;White et al.1999).In the X-ray,groups and clusters are customarily distinguished by the temperature of the hot,gaseous medium that makes up ≈10%of their total mass (Ebeling,Voges,&B¨o hringer 1994;Ponman et al.1996;Mahdavi et al.2000;Mulchaey 2000).Clusters,with temperatures 2keV,contribute only about 3%to the matter density of the Universe;but groups,with X-ray temperatures in the range 0.5-2.0keV (corresponding to a mass range 1013–1014M ⊙),contribute about twice as much (Ikebe et al.2002;Reiprich &B¨o hringer 2002).Because of the cooler intracluster medium (ICM)temperature of the groups,the set of observable X-ray spectral lines is richer than it is in clusters,and allows a more accuratedetermination of elemental abundances.The cD galaxies in groups make up a larger fraction of the total baryonic mass of the system than do the cD galaxies in clusters (Helsdon et al.2001;Lin &Mohr 2004);hence the prop-erties of the gaseous medium are more directly affected by the formation and evolution of the central galaxy.Fi-nally,because the typical velocities of galaxies in groups are comparable to the galaxies’internal (stellar)velocity dispersions,galaxy interactions through the merging insta-bility are common and very effective (Diaferio et al.1993;Ramella et al.1994).Recent studies suggest that in the X-ray,the distinc-tion between groups and clusters goes beyond an arbitrary temperature or mass boundary.Differences in the physical state of the intragroup medium make it difficult to view the 2keV systems simply as smaller,“rescaled”ver-sions of the more massive clusters.Of particular interest are the heating of the ICM by nongravitational processes such as supernova explosions,stellar winds,AGN activity,and shocks resulting from interaction with the surround-1Based on observations obtained with XMM-Newton,an ESA science mission with instruments and contributions directly funded by ESA Member States and the USA (NASA).The XMM-Newton project is supported by the Bundesministerium f¨u r Bildung und Forschung/Deutsches Zentrum f¨u r Luft-und Raumfahrt (BMFT/DLR),the Max-Planck Society and the Heidenhain-Stiftung,and also by PPARC,CEA,CNES,and ASI.Also based on observations obtained at the Gemini Observatory,which is operated by the Association of Universities for Research in Astronomy,Inc.,under a cooperative agreement with the NSF on behalf of the Gemini partnership:the National Science Foundation (United States),the Particle Physics and Astronomy Research Council (United Kingdom),the National Research Council (Canada),CONICYT (Chile),the Australian Research Council (Australia),CNPq (Brazil),and CONICET (Argentina).12ing large scale structure(Ponman,Cannon,&Navarro 1999;Loewenstein2001;Tornatore et al.2003).Some-times called“preheating,”these processes can leave a dis-tinct mark on the entropy distribution of the gas in groups, to a degree not observable in more massive clusters.For this reason groups provide a fossil record of this energy production during cosmic structure formation and galaxy evolution.In this study we focus on the distribution of entropy and pressure in eight systems drawn from the RASSCALS sur-vey of nearby galaxy groups.In§2and§3we discuss the optical and X-ray data we have gathered for this study,as well as the data reduction procedures used in our analysis. In§4we describe our attempts tofit self-similar profiles to the entropy and the pressure distribution.In§5we describe each system in detail,In§6we summarize our conclusions.2.optical data2.1.Sample SelectionAt the focus of our study are groups with emission-weighted temperature∼0.5−2keV.The X-ray luminosity-temperature relation(Ikebe et al.2002)sug-gests that these groups should have L X 1043erg s−1in the0.1-2.4keV energy range.2To build our sample,we begin with the ROSAT All-Sky Survey-Center for Astrophysics Loose Systems(RASS-CALS),a statistically complete,magnitude-limited cat-alog of optically identified groups with0.01<z<0.04 (Mahdavi et al.2000).This catalog contains260groups, of which43have X-ray emission detected in the RASS. To conduct a more detailed study of the properties of these groups,we select X-ray emitting RASSCALS with L X<1043erg s−1that have been observed with the XMM-Newton observatory.The properties of the sample are shown in Table1.The membership of the original RASSCALS groups was established spectroscopically with a completeness limit m R≈14.4(Ramella et al.2002).All but one of the groups we have selected for analysis have also been the targets of deeper redshift surveys,by Mahdavi&Geller (2004)and Rines et al.(2003)(complete to m R≈15.4), by Zabludoff&Mulchaey(1998)(70%–95%complete to different magnitude limits depending on the group),or by Pinkney et al.(1993)(unknown completeness).None of the groups had significant overlap with the current release of the Sloan Digital Sky Survey(Abazajian et al.2004). In addition,we have performed optical observations which we describe below.2.2.New Optical SpectroscopyWe observed three groups,RGH80,HCG97,and NRGb184with the8m Gemini North telescope on Mauna Kea,Hawaii.The Gemini multi-object spectrograph, GMOS,was used to measure redshifts for galaxies as faint as m R≈20within6.5′of the center of each system.Thefirst step was the selection of targets for spec-troscopy.We used GMOS in queue imaging mode to ob-tain a2×2,13′×13′R-band mosaic around the central galaxy in each group;the length of each imaging expo-sure was10minutes.The SExtractor package(Bertin& Arnouts1996)then separated galaxies from stars,and cal-culated the magnitudes and half-light radii of the galax-ies.It would have taken an inordinate amount of time to measure redshifts for all m R<20galaxies in each image. Therefore,galaxies were sorted by their half-light radii, with the assumption that regardless of magnitude,galax-ies with larger half-light radii would be more likely to have low redshifts and therefore belong to the group.All three groups are at a small enough redshift that foreground con-tamination was not an issue with this selection procedure. In each piece of the mosaic,we selected the20galax-ies with the largest half-light radii and without previ-ously measured redshifts.Thus there were a total of80 targets for NRGb184and HCG97.For RGH80,we in-cluded an additional central mask,for a total of100tar-get galaxies.We designed one slit mask per image to be used with the GMOS B600grating;the slit size was0.5′′(RGH80and HCG97)or0.75′′(NRGb184)achievinga resolution of4−6˚A at4000˚A.The wavelength coverage was4000−6000˚A,with a shift of0−1000˚A in either direc-tion depending on the position of the galaxy on the focal plane.We used two30minute exposures per slit mask (a total of one hour),taking arc lamp exposures between each set of two exposures to obtain accurate wavelength calibrations.Because absolute photometric calibration is not required for redshift measurements,we did not apply flat-fielding orflux calibrations to the spectra.The Gemini GMOS package for IRAF3was used to cal-culate the wavelength solutions and to reduce the multi-object observations into one-dimensional spectra.The RVSAO package(Kurtz&Mink1998),incorporating the methods of Tonry&Davis(1979),allowed us to measure redshifts by maximizing the cross-correlation of the spec-tra with absorption-and emission-line templates.The es-timation of errors in the optical velocities is described in detail by Kurtz&Mink(1998),who determine the rela-tionship between the shape of the cross-correlation func-tion peak and the68%velocity confidence interval using galaxies with known velocities.Our results are in Table2.2.3.MembershipWe assemble a galaxy catalog by combining all the avail-able redshift surveys in the direction of each group,and restricting ourselves to galaxies within2Mpc of the group center.The mean redshift z and velocity dispersionσof each group are defined as follows(Danese,de Zotti,&di Tullio1980):z≡1(N−1)(1+z)2.(2)The subtraction of the root-mean-square velocity error in the second term is an attempt to remove the contribution of the measurement uncertaintiesǫi toσ.In practice,this correction may be inaccurate when the underlying veloc-ity distribution is nongaussian,the sample size is small2We assume H0=70km s−1Mpc−1,Ω0=0.3,andΩΛ=0.7throughout the paper.3For a description of the Gemini data reduction package see /sciops/data/dataSoftware.html.3(N 20),and the velocity dispersion is of the same or-der of magnitude asǫi.The inaccuracy is not a source of concern for our sample,but may be relevant for some extremely low-velocity dispersion groups withσ≈70km s−1(Mahdavi et al.1999,2000).To determine group membership,we then use the “sigma-clipping”(Zabludoff,Huchra,&Geller1990;Mah-davi&Geller2004)algorithm to reject outliers,or galaxies unlikely to be bound to the group.This algorithm consists entirely of making sure that no group member is separated from its nearest neighbor in velocity space by more than the velocity dispersionσof the group.3.xmm-newton observationsWe conducted XMM-Newton Observations of three of the eight groups,while the remainingfive are publicly available in the XMM-Newton Data Archive.Table3con-tains the details of the observations.The initial steps of the data reduction are similar to the procedure described in Zhang et al.(2004)and Finoguenov et al.(2003).Thefirst important aspect is the removal offlares,which can significantly enhance the detector background,severely limiting the detection of low surface brightness features.Thus,for the group analysis, usingflare free observing periods is critical.At energies above10keV the particle background dominates the de-tected counts;there is little X-ray emission from our tar-gets because(1)the telescope efficiency is quite low at energies>10keV,and(2)the temperature of the objects in our study is less than2keV.We use the10–15keV energy band(binned in100s intervals)to monitor the par-ticle background and to excise periods of high particleflux. In this screening process we use the settings FLAG=0and PATTERN<5for the pn detector on XMM-Newton.We reject time intervals affected byflares by excising periods where the detector count rate exceeds the mean quiescent rate by more than2σ.To produce the broad-band images and hardness ratio maps,we have also used the MOS1 and MOS2events with PATTERN<13and FLAG=0.To reduce the widths of the gaps in the pn broad band im-ages,we included photons near the pn-CCD borders,near bad pixels,and near offset columns.Data from the pn de-tector were used exclusively in the spectroscopic analysis described below.Because most of the observations analyzed here were performed using a short integration frame time for pn(Full Frame Mode),it is important to remove the out-of-time events(OOTE)for accurate imaging and spectral analy-sis.We used the standard product of the XMMSAS5.4 epchain task to produce the simulated OOTEfile for all the observations and scale it by the fraction of the OOTE expected for the frame exposure time,(Table3).The cosmic component of the background consists of emission from the Galaxy as well as the extragalactic Cos-mic X-ray Background(CXB).Observations of blankfields also contain both components.Provided that the expo-sures are done with the same instrumental set-up(e.g. with a particularfilter)the spectra of the CXB are the same for both the target and the blankfield.The Galac-tic component acts as an absorber and emitter,thus it is important to choose a similar absorbing column density for both the target and background data sets,so that the ex-pected background spectra are similar.In addition,there are variations in the Galactic emission on small scales. The vignetting correction is performed taking into ac-count the source extent,recent vignetting calibration (Lumb et al.2004),and the pn response matrices released under XMMSAS6.0.The residual systematic uncertainty of theflux is below4%for the pn(Lumb et al.2004).4 The analysis of each group consists of two steps:(1) construction of separate wavelet-decomposed maps of es-timated surface brightness in the0.5-2.0keV band(I)and of emission-weighted temperature(T I),which may be used to derive the entropy and pressure“integrated”along the line of sight;and(2)use of the wavelet-decomposed maps to identify contiguous regions from which we extract X-ray spectra for independent model-fitting.The background-subtracted wavelet maps are based on photon images corrected for instrumental effects.The sur-face brightness map is constructed using the technique described in Vikhlinin et al.(1998).A hardness map of the emission then results from dividing the wavelet-reconstructed images in the0.5–1and1–2keV bands. The hardness ratio is a monotonic and unique function of the emission-weighted temperature,as long as the group redshift is well known(true for our sample).The advan-tages of using wavelets include the ability to remove ad-ditional background by spatialfiltering and a control over the statistical significance of the detected structures.We use the“a trous”method of wavelet image reconstruction with scales from8′′to4′,applying a four sigma detection threshold and retaining emission to a1.7σdetection limit. Occasionally,the wavelet algorithm generates small scale discontinuities in the reconstructed image.We remedy this effect by applying additional smoothing before producing the hardness ratio maps.We also construct maps of the entropy and pressure“in-tegrated”along the line of sight,using the surface bright-ness image and the emission-weighted temperature maps. The use of the word“integrated”is not strict,because these quantities do not have the dimensions of entropy or pressure times distance,and they serve as heuristic aids only.In studies of the intracluster medium,the common definitions of entropy and pressure are kT/n2/3and nkT, respectively,where n is the gas density and kT is the tem-perature in units of keV(see§4for further details).Be-cause the emission measure is proportional to n2,we use √IT I.The quantities shown in these maps are not simply related to the physical,unprojected entropy and pressure,and they suffer from a number of degeneracies,with metallicity-density being the strongest(a significant fraction of the group emission results from line emission).Nevertheless, the maps indicate the regions of primary interest for fur-ther spectroscopic analysis,in which most of the degen-eracies are removed.In general,we expect the entropy to increase monotonically with radius(Metzler&Evrard 1994),while the pressure should decrease.We stress that the“integrated”pressure and entropy are used only as tools to guide further spectroscopic analysis.They are4For further details of XMM-Newton processing we refer the reader to http://wave.xray.mpe.mpg.de/xmm/cookbook/general.4not used in thefits we discuss in§4.The second,spectroscopic part of the analysis uses a maskfile,constructed from both hardness ratio and sur-face brightness analysis described above.Thefirst appli-cation of this technique is in Finoguenov et al.(2004).In our analysis we select regions with similar spectral prop-erties.We combine the regions so that counting statistics are not the limiting factor in our derivation of the group properties.We use the wavelet-based maps to identify re-gions with similar X-ray colors and intensity levels.To generate the maskfile for use in the further spectral anal-ysis,we sample the changes in the intensity and hardness ratio at the precision allowed by the statistics of the data. We then examine each of the isolated regions with approx-imately equal color and intensity,imposing the additional criterion that the reg ions should be larger than the PSF width(15′′)and contain more than300counts in the raw pn image.A sample maskfile is shown in Figure1.The spectral analysis was performed using single-temperature models and the APEC plasma code,fitting el-emental abundances from O to Ni as one group and assum-ing the photospheric solar abundance ratios of(Anders& Grevesse1989).It should be noted that evidence for mul-tiphase gas in groups of galaxies exists(Buote et al.2003),such that gas at each projected radius may vary in tem-perature by a factor of≈2;however,if the variations are due to a smooth radial trend in the plasma temperature (i.e.,a temperature profile),a series of single-temperature fits across the group should roughly reproduce the general shape of the underlying pressure and entropy distributions. In our analysis,we paid special attention to the issue of background estimation.We employ a double background subtraction technique,following Zhang et al.(2004).We used the region with radius12–16′and2–12keV band to estimate the quality of the background subtraction using blankfields.This region is free of significant X-ray emis-sion from the groups.Wefix the shape of the residual background component,and also add a0.2keV thermal component(APEC with solar element abundance)to ac-count for a possible variation in the Galactic foreground, allowing the normalizations of both components to befit. To estimate the true(as opposed to“integrated”)pres-sure nkT and entropy kT n−2/3in each region,we need measurements of both the gas density n and the temper-ature T.The spectralfit products,however,are kT and n2V,where V is the volume of the emitting region.We do not have an exact knowledge of V.Because the mask generation technique identifies interesting regions by their wavelet-decomposed properties—producing region maps like Figure1—a straightforward deprojection of the spec-tra is not possible.We therefore need to estimate the length of the column for each selected two-dimensional re-gion on the sky.We assume that the gas emits uniformly along the line of sight as shown in Figure2.With these approximations,the longest length through each volume is L=2Fig. 2.—Simplification of the group geometry.The emitting re-gion is to the left of the dashed line,and has a minimum and max-imum projected distance R1and R2,respectively,from the groupcenter.The distance along the line of sight is then25Fig. 3.—X-ray and optical images of the eight groups in our sample.(Top)Wavelet decomposition of the X-ray surface brightness in the0.5-2.0keV energy band.Each image is19′×19′;the grid cells are200kpc on a side.The small boxes show the area enlarged below. (Bottom)Palomar Observatory Sky Survey images of the central region of each group,with X-ray surface brightness contours superimposed from the above images.Each image is5.2′×5.2′in size;the grid cells are50kpc on a side.6Fig.4.—(Top)Emission-weighted temperature distribution of the sample from wavelet analysis.The range in temperatures shown is0.5 keV(white)to2.3keV(black),with a linear scale.Member galaxies are shown as crosses.Each grid cell is200kpc on a side.(Bottom) Distribution of the entropy“integrated”along the sight,estimated from wavelet analysis.The range in integrated entropy shown is10(white) to110(black)in arbitrary units,with a linear scale.The contours show lines of constant gas pressure“integrated”along the line of sight; the innermost contours always show higher pressure,with each additional contour showing a decrement in pressure from the previous one of a factor of1.5.7 gravitational collapse,processes such as supernova explo-sions make a contribution to the thermodynamic state ofthe gas.This contribution takes the form of a“floor”inthe inner entropy distribution of the intragroup mediumas reported in these studies.In other words,gas withhigher entropy than expected from pure gravitational col-lapse appears to be pooled at the center of groups ob-served by the ROSAT satellite.Ponman,Sanderson,&Finoguenov(2003)use a larger sample of66systems ob-served by ASCA to study this effect in greater detail.Theyfind evidence of an entropyfloor in these observations,with the notable exception that in lower temperature(lessmassive)systems the effect is much less pronounced thanin higher temperature(more massive)ones.The entropyprofiles of the less massive systems are better described bya simple power law.Further complicating the picture,Finoguenov et al.(2002),using ASCA,found that gas at r500(the radiuswithin which the mean matter density is501times the crit-ical density of the universe)also exhibits excess entropyrelative to the pure gravitational value.This outer en-tropy excess seems consistent with models in which shockheating and galactic winds are also a major contributor tothe groups’dynamical evolution(Dos Santos&Dor´e2002;Voit&Ponman2003).In addition,Ponman et al.(2003)found that the entropy profiles,once scaled to T0.65,arein agreement with each other,hinting at the possible uni-versality of the entropy enrichment phenomenon.SomeXMM-Newton observations are consistent with this result(Pratt&Arnaud2002).Next we revisit all of these results.4.1.Calculation of the ProfilesFigures5and6show the pressure and entropy as a func-tion of true(as opposed to projected)distance from thecenter for our sample.The true distances were estimatednot via deprojection but using the method described in§3above.We define the gas entropy similarly to the ear-lier discussions of the intragroup medium(Ponman et al.2003):S≡T n−28Fig.5.—Pressure distributions for the eight groups in our sample.(top)Single-power lawfit to the entire sample,with the slope contrained to be same for each group;the reducedχ2is3.32.(bottom)Broken power lawfit;the reducedχ2is1.18.9Fig.6.—Entropy distributions for the eight groups in our sample.(top)Single-power lawfit to the entire sample,with the slope contrained to be same for each group;the reducedχ2is2.42.(bottom)Broken power lawfit;the reducedχ2is1.50.10We test the self-similarity of the entropy and pressure profiles byfitting broken power laws of the formP(r)=P0 (r/100kpc)αp r<r p(r p/100kpc)αp−γp(r/100kpc)γp r>r p(4) Similarly,wefit entropy profiles of the formS(r)=S0 (r/100kpc)αs r<r s(r s/100kpc)αs−γs(r/100kpc)γs r>r s(5) For these self-similar models we conduct two separate analyses:a pressurefit and an entropyfit.The pressure fit has18free parameters:the slopesαp andγp common to all the groups,and eight different pairs(r p,P0)for each group.Similarly,the entropyfit has its own set of18pa-rameters.Minimizingχ2using the same technique as in the pre-vious section yields goodfits;these are shown in the bot-tom half of Figures5and6.In Table5,we estimate the goodness-of-fit(the probability of observingχ2>χ2min if the model is correct)as the definite integral of theχ2dis-tribution,q=Γ(ν/2,χ2/2)/Γ(ν/2),whereΓ(a,x)is the incomplete gamma function,Γ(a)=Γ(a,0)is the gamma function,andνis the number of degrees of freedom(i.e., the number of data points minus the number of free pa-rameters minus one).Afterfinding the minimum,we calculate errors on the best-fit parameters by using likelihood ratio tests(Lupton 1993).This method fully takes the correlations among the parameters into account.The test involves repeatedly comparing the overall minimumχ2withχ2minimized as-suming many differentfixed values for the parameter in question.The set of differences between the global min-imum and the constrained minima has aχ2distribution with one degree of freedom asymptotically,and this fact may be used to estimate confidence intervals.For each pa-rameter we re-minimizeχ2100times around the best-fit value;the error estimation procedure thus involved1800 re-minimizations in all.4.4.ResultsWefind that a broken power law provides a good de-scription of the combined pressure profiles of the eight groups(see Table5and Figure5).The pressure bro-ken power law has a shallow slope−0.78+0.04−0.03near thegroup center,steepening to−1.7+0.1−0.3at the outer edges;the goodness-of-fit is an acceptable q=0.15.The agree-ment is not perfect:NRGb184is entirely consistent with a single,rather than a broken power law,and SS2b153ex-hibits a somewhat steeper inner pressure profile than the other seven groups.Otherwise,the notion that the groups are self-similar systems differing only in a radius and ab-solute pressure scaling is supported by the XMM-Newton data.The entropy profiles(Figure6),however,make the pic-ture more complicated.The broken power lawfit to the combined data hasfit quality q=0.0095;the model is acceptable at3σ.Excluding the binary cluster Abell 194would bring the goodness-of-fit to a wholly acceptable q=0.136,and would not affect the best-fit slopes sub-stantially.The inner power law slope,0.92+0.04−0.05,is steeperthan the outer entropy slope0.42+0.05−0.04,meaning there is noevidence for an constant-entropy“floor”as close as10kpcfrom the center of each group.Rather,in contrast to pre-vious observations with ROSAT and ASCA,the entropyprofile shows a tendency to steepen towards the center ofeach group.There is an entropy deficit at larger radii rela-tive to the extrapolation one would obtain from the innerpower law,a new result not seen in data from the olderX-ray observatories.It is instructive to compare our data with theoreticalmodels.Tozzi&Norman(2001)make detailed predictionsfor the entropy profiles of groups and clusters of galaxiesin two different scenarios:with preheating(where a signif-icant initial entropy excess exists in the systems at t≈1Gyr),and without preheating(where the gas dynamics ischiefly determined by shocks and gravitational heating).In thefirst scenario,theyfind that more massive clus-ters develop isentropic cores within∼0.05r200,and thatless massive groups develop relatively larger cores(within∼0.5r200).By contrast,in the scenario without preheat-ing,groups and clusters both tend to develop power lawentropy distributions,with S(r)∝r1.1.To test these scenarios,we combine our entropy profilesin several ways(Figure7).In(a),we show all the groupsfor which the preferredfit is a broken power law—that is,the groups for which r s is neither a lower nor an upperlimit.Here the radii are scaled by r s,and the entropiesby S(r s).As might be expected,this combination of pro-files exhibits little scatter around the best-fit line,becausethe highest contributors toχ2—Abell194and SRGb119—are not included.In(b)we show the profiles for all eightgroups,scaled by r500and S ,the mean entropy withinr500as listed in Table4.Despite the difference in scal-ing,the data are still well described by our best-fit brokenpower law withfixed slopesαs=0.92andγs=0.42—infact,thefit quality is higher,because the additional uncer-tainty introduced through division by S decreases the re-ducedχ2.We compare these profiles with the predictionsof the Tozzi&Norman(2001)preheating model,where weassume that S =S∞/2,where S∞is the entropy of thelast accreted gas shell.The model is rejected,because itpredicts a large isentropic core that we do not observe andan outer profile that rises more quickly than ours.Vary-ing S /S∞arbitrarily in either direction does not changethis result.Finally,in(c)we show the eight groups again,with the radii scaled by r500,but with the entropies scaledby(T/keV)0.65,a relation Ponman et al.(2003)derivefor their sample of66virialized groups and clusters.Thisscaling agrees with ourαs=0.92andγs=0.42fits aswell,but is clearly inconsistent with both the preheatedand non-preheated Tozzi&Norman(2001)models—bothmodels are rejected with q<10−6.Thus there are two key differences between our resultsand the Tozzi&Norman(2001)models:(1)we do notobserve any isentropic cores,and(2)wefind an entropydecrement at radii 0.5r500with respect to the extrap-olation from data at radii 0.5r500.Neither of theseresults is unique to our sample:both effects are also clearin the high quality Chandra observation of the NGC1550group(Sun et al.2003);the lack of an isentropic coreis also clear in the Chandra and XMM-Newton observa-tions of ESO306170(Sun et al.2004)and in a few ofthe groups studied by Ponman et al.(2003).These stud-。

a r X i v :a s t r o -p h /0501377v 1 18 J a n 2005Merging clusters of galaxies observed withXMM-NewtonE.Belsole a J-L.Sauvageot b G.W.Pratt c and H.Bourdin da H.H.Wills Physics Laboratory -University of Bristol,Tyndall Avenue,Bristol BS81TL,UKb Service d’Astrophysique,CE-Saclay,L’orme des Merisiers,91191Gif sur Yvette Cedex,Francec MPE Garching,Giessenbachstraße,85748Garching,Germanyd Dipartimentodi Fisica,Universit`a degli studi di Roma Tor Vergata,Via della Ricerca Scientifica 1,00133Roma,Italy1IntroductionAs the largest assembled structures in the Universe,clusters of galaxies are commonly thought to form by gradual accretion of matter alongfilaments and by interaction and merging with previously formed structures.Hydrody-namical simulations(e.g.Rowley et al.2004and references therein)predict that merger events strongly affect the physical characteristics of the intra-cluster medium(ICM).In particular the temperature structure is thought to be an excellent indicator of the cluster dynamical state and formation his-tory.Spectro-imaging observations with XMM-Newton and Chandra allow the building of precise temperature maps,enabling deeper investigation of the dynamical processes of cluster formation and the effect of the mergers on the ICM(e.g.,Markevitch et al.,2000;Neumann et al.,2003;Henry et al.,2004; Krivonos et al.,2003;Vikhlinin&Markevitch,2003).With the aim of describing an evolutionary sequence of cluster formation,we have selected a small sample of galaxy clusters showing morphological evidence of ongoing merger activity.The sample was selected,on the basis of previous X-ray observations,from nearby clusters which entered thefield of view of XMM-Newton.Here we summarise results for A1750and A3921,presented in detail in Belsole et al.(2004,2005),and we describe preliminary,new results for A2065.2Observations and data analysisIn Table1we list the main physical characteristics of the three clusters,to-gether with basic observation information.The A1750and A3921observations were very little contaminated by soft protonflares(details of the data prepa-ration can be found in Belsole et al.2004,2005).Unfortunately,emission from soft protons dominates the whole observation of A2065.This required an ad-hoc treatment for these data,which allowed us to model the background. A2065will be re-observed with XMM-Newton,and thus here we discuss only preliminary results.The background estimates were obtained using a blank-sky observation con-sisting of several high-latitude pointings with sources removed(Lumb et al. (2002)),and source and background events were corrected for vignetting us-ing the eviweight task in the Science Analysis System(SAS),enabling us to use the on-axis response matrices and effective area.Table1Journal of observations.z0.0860.0720.096RA(J2000)13h30m52s15h22m42.6s22h49m38.6sDec(J2000)−01o50′27′′+27o43′′21′′−64o23′15′′Obs time(ks)302330N H(1020cm−2)a 2.39 2.95 2.94Global T(keV) 3.87(2.84)b 5.4 5.0(a)(b)(c)Fig.1.(a)XMM-Newton EPIC0.3-7.0keV energy band count image of A1750;(b)MOS1+MOS20.1-1.4keV energy band count image of A2065,the band chosen to enhance cluster emission above the strong particle background;(c) XMM-Newton EPIC0.3-10keV energy band count image of A3921.These are raw, non-background subtracted images.to A1750were taken into account simultaneously.We found that A1750does not show significant residuals above aβ-model,in the region between the two clusters,but there is some excess emission in their cores.On the other hand,A3921shows large residuals to the west(other than those(a)(b)Fig.2.Surface brightness maps of A1750(a)and A3921(b)in the soft energy band (0.3-2.0keV).Contours are the residuals above a2Dβ-model which was subtracted from the image.Thefirst contour is at1σ.in the centre of the main cluster),indicating that a significant substructure is present and it is related to one or both the brightest galaxies in that area.The two galaxies are at7arcmin(RA=22h48m49s,Dec=−64◦23′10′′(J2000))to the north-west(BG2)and at8arcmin(RA=22h49m04s,Dec=−64◦20′35′′(J2000))to the WNW(BG3)from the central brightest galaxy BG1(RA= 22h49m58s;Dec=−64◦25′46′′(J2000);see Belsole et al.2005for details).3.2Temperature distributionWe computed temperature maps for A1750and A3921by applying the multi-scale spectro-imagery algorithm described in(Bourdin et al.,2004).EMOS cameras only were used for this analysis.For A2065,we obtained an hardness ratio(HR)map once the particle contamination was modelled and subtracted from the images.This gives us qualitative,but significant results about tem-perature substructure in this cluster.The temperature maps are shown in Figure3.All three clusters show significant temperature variations,even in the cases where surface brightness substructures were not detected(A1750).As is con-firmed by extraction of spectra from discrete regions(Belsole et al.,2004), the region between A1750N and A1750C has a temperature which is signif-icantly(30%)hotter than the global temperature of either the two clusters. Additional temperature gradients are observed in A1750C,indicating that this(a)(b)(c)Fig.3.Temperature maps.The wavelet algorithm described in Bourdin et al.(2004) was used for A1750(a)and A3921(c).In the case of A2065(b)the image is an hardness ratio map(see text for details).cluster was already perturbed in the past.The temperature enhancement,coincident with the isophotal compression, observed in the A2065HR map points to a bow shock.We also detect a cool core,coincident with the peak of X-ray emission.The orientation of the striking hot temperature bar in A3921represents our strongest proof against the interpretation that this is a merger before close core passage.The temperature here varies between7and8keV,up to60% higher than the main cluster temperature.4DiscussionFor each cluster we summarise the results obtained in the previous sections.A1750(1)From the morphological analysis we found:off-centre cores and twist ofthe isophotes,but lack of substructure in the region between the two clusters.(2)the temperature increases weakly(∼30%),but significantly,in the regionbetween the two clusters(3)strong temperature variations are observed within the main cluster(A1750C)and we also measured a discontinuity in the gas density profile of order 20%.(4)we found a high entropy in the core of the two clusters(see Belsole et al.2004)The combination of these results lead us to interpret A1750as an ongoing merger between two clusters of similar masses,which have just started to in-teract at a real distance comparable to their virial radii.These two units will be in the compact phase(core passage)within1Gyr.However the temperature and density variations within A1750C itself are not explicable by the merger event occurring with A1750N and are intrinsic to this cluster.The most likely interpretation is that this is the signature of(a)previous merger(s).A1750is a good observational test case to be compared with numerical simulations, since it shows all of the signatures expected from gravitationally dominated processes.However,the fact that effects of a previous merger are observed strongly supports the necessity of taking into account time dependent quan-tities,such us the relaxation time,in using galaxy clusters for cosmological purposes.A2065The surface brightness shows an isophotal compression towards the south-east,at∼80arcsec from the X-ray peak.The feature is accompanied by a forward shock in the same axis(NW-SE);moreover a surviving cooling coreis detected.Despite the limited quality of the data,A2065shows clear signatures of be-ing an ongoing merger in the compact phase,when the detection of strong shocks is the most favourable.This is similar to what is observed in numer-ical simulations of head-on collisions of merging clusters.The fact that thecore of the main cluster is cool suggests that this is probably the remnantof a cooling core,and thus the colliding object was probably of smaller mass (e.g.G´o mez et al.,2002).The new XMM-Newton observations of this object should allow us to put better constraint on these preliminary results.A3921Observational merger evidence includes:(1)Two peaks in both the X-ray emission and the galaxy distribution(Ferrari et al.,2005);(2)The hot bar is oriented parallel to the line joining the subclusters,it isnot orthogonal,as in the case of A1750;(3)The central regions of the main cluster and the subcluster to the westare strongly perturbed;(4)There is an off-set between the galaxy distribution of the smaller sub-cluster and the secondary X-ray peak(Belsole et al.,2005;Ferrari et al.,2005).This evidence cannot be interpreted as the result of a pre-merger.We are seeing maybe the best example of an X-ray observed off-axis post-merger.The subcluster has come from somewhere in the SE and is currently exiting towards the NW.The two merging units have different masses,on a ratio of1to3(or1to5;see also Ferrari et al.2005).Comparison of these high qualityX-ray results with optical observations and numerical simulations yield an estimate of the age of the merger of order0.5Gyr after core passage.Off-axis mergers are more likely to occur than head-on ones,and they are more efficientin mixing the gas via turbulence.If anything,A3921needs even deeper study,and the combination of these XMM-Newton data with upcoming Chandra observations will give further elements to our interpretation.5ConclusionsThe XMM-Newton observations of this small sample of galaxy clusters has confirmed that the comparison of X-ray morphology and temperature is an excellent tool to understand the dynamical status of these objects.In the case of A1750and A3921,spectroscopy of discrete regions((Belsole et al., 2004,2005)has confirmed the significance of the temperature structure found with the multi-scale wavelet approach.The significance of the temperature structure in A2065will be investigated with the upcoming XMM-Newton ob-servation.For the better studied cases we have confirmed the previous inter-pretation of a recent merger(A1750),but have added new evidence suggesting that at least one of the subclusters is itself a merger remnant.Good quality X-ray data have allowed us to completely revolutionise the interpretation of the dynamical state of A3921.We can organise these clusters along an evolutionary path,where A1750rep-resents the beginning of a merging event,which in a timescale of order1.5 Gyr will be in the same state as A2065,when a bow shock is departing in the direction of motion after the two cores have collided,and the two(or maybe more)colliding objects are not physically separable anymore.Finally,A3921 represents the epoch when the secondary object has already passed the close core passage phase,it is exiting on the far side with respect to the direction of motion,and it will be accreted by the main cluster on a larger timescale (of order3-5Gyr).From so small a sample we cannot extract global conclusions on the popu-lation of merging galaxy clusters.However,this work supports the necessity of a wider investigation of the effect of the physics of mergers on the global characteristics of galaxy clusters,especially if these objects are to be used to derive cosmological parameters.ReferencesBeers,T.C.,Gebhardt,K.,Forman,W.,Huchra,J.P.,Jones,C.,1991,AJ, 102,1581-1609,A dynamical analysis of twelve clusters of galaxies Belsole,E.,Pratt,G.W.,Sauvageot,J-L.,Bourdin,H.,2004,A&A,415,821-838,An XMM-Newton observation of the dynamically active binary cluster A1750Belsole,E.,Sauvageot,J-L.,Pratt,G.W.,Bourdin,H.,2005,A&A,in press, preprint astro-ph/0409544,An XMM-Newton observation of A3921:an off-axis mergerBourdin,H.,Sauvageot,J.-L.,Slezak,E.,Bijaoui,A.,Teyssier,R.,2004,A&A, 414,429-443,Temperature map computation for X-ray clusters of galaxiesCavaliere,A.,&Fusco-Femiano,R.,1976,A&A,49,137-144,X-rays from hot plasma in clusters of galaxiesDickey,J.M.,&Lockman F.J.,1990,ARA&A,28,215-261,HI in the Galaxy Ferrari,C.,Benoist,C.,Maurogordato,S.,Cappi,A.,Slezak,E.,2005,A&A, accepted,preprint:astro-ph/0409072,Dynamical state and star formation properties of the merging galaxy cluster Abell3921G´o mez,P.L.,Loken,C.,Roettiger,K.,Burns,J.O.,2002,ApJ,569,122-133, Do Cooling Flows Survive Cluster Mergers?Henry,J.P.,Finoguenov,A.,Briel,U.,2004,ApJ,615,181-195,Widefield X-ray temperature,pressure and entropy maps of A754Krivonos,R.A.,Vikhlinin,A.A.,Markevitch,M.L.,Pavlinsky,M.N.,2003, AstL,29,425-428,A Possible Shock Wave in the Intergalactic Medium of the Cluster of Galaxies A754Lumb,D.H.,Warwick,R.S.,Page,M.,De Luca,A.,2002,A&A,389,93-105, X-ray background measurements with XMM-Newton EPICMarkevitch et al.,2000,ApJ,541,542-549,Chandra Observation of Abell 2142:Survival of Dense Subcluster Cores in a MergerNeumann,D.M.,Lumb,D.H.,Pratt,G.W.,Briel,U.,2003,A&A,400,811-821,The dynamical state of the Coma cluster sith XMM-Newton Rowley,D.R.,Thomas,P.A.,Kay,S.T.,2004,MNRAS,352,508-522,The merger history of clusters and its effect on the X-ray properties of the intr-acluster medium,Vikhlinin,A.A.,Markevitch,M.L.,2003,AstL,28,495-508,A Cold Front in the Galaxy Cluster A3667:Hydrodynamics,Heat Conduction and Magnetic Field in the Intergalactic Medium。

An international research team led by a researcher from the University of Vienna has for the first time directly detected stellar winds from three Sun-like stars by recording the X-ray emission from their astrospheres(星状体), and placed restrictions on the mass loss rate of the stars via their stellar winds.The researchers observed the spectral fingerprintsof the oxygen ions(离子) with XMM-Newton and were able to determine the quantity of oxygen and ultimately the total mass of stellar wind emitted by the stars. For the three stars with detected astrospheres, named 70 Ophiuchi, epsilon Eridani, and 61 Cygni, the researchers estimated their mass loss rates to be 66.5±11.1, 15.6±4.4, and 9.6±4.1 times the solar mass loss rate, respectively.人类首次探测到三颗类太阳恒星的恒星风●南昌民德学校　谢钰洪1. What’s worrying us when it comes to wireless devices?A. Shockingly wide applications.B. Unsustainable power support.C. Poor practical adaptability.D. Quick power consumption.2. What’s the main idea of Paragraph 3?A. The working principle of the PEC.B. The composition of the PEC.C. The performance of the new material.D. The complex design of the device.3. What’s the function of the electric double layer?A. To stimulate ions.B. To produce energy.C. To preserve electricity.D. To adjust the system.4. What can be inferred from what Roundy said?A. All your devices should be closely watched.B. Smart watcher’s sensors are better data senders.C. The new device is only workable to some sensors.D. The new device’s sensors can transmit big data.参考答案1. B 。

**Newton's Discovery of Universal Gravitation**In the annals of science, Isaac Newton's discovery of universal gravitation stands as a monumental leap forward in understanding the workings of our universe. This groundbreaking revelation not only explained the mysterious forces that govern planetary motion but also laid the foundation for subsequent advancements in physics, astronomy, and even mathematics.Born in 1643, Newton was a man of many talents. Initially, he pursued his studies in mathematics and physics, but his interest soon shifted towards the mysteries of nature. It was during this period of profound inquiry that he made his remarkable discovery.The story often told about Newton's discovery involves an apple falling from a tree. While sitting beneath an apple tree in his garden, the sight of the falling fruit led him to ponder the forces at work. This simple observation sparked a train of thought that ultimately led to the formulation of his theory of universal gravitation.According to Newton's theory, every object in the universe attracts every other object with a force that is directly proportional to the product of their masses and inversely proportional to the square of the distance between them. This force, which we now know as gravitational force, is what keeps our planets in orbit and causes objects to fall towards the ground.The beauty of Newton's theory lies in its simplicity and universal applicability. It not only explained the motion of planets and satellites but also accounted for the falling of objects on Earth. This universal nature of gravity is what makes it such a fundamental concept in physics.Moreover, Newton's discovery also had profound implications for mathematics. His laws of motion and theory of gravitation led to the development of calculus, a powerful tool that has revolutionized the field of mathematics.In conclusion, Isaac Newton's discovery of universal gravitation is a testament to the power of observation, inquiry, and reason. His theory revolutionized our understanding of the universe and laid the foundation for centuries of scientific progress.。

芜湖牛顿英语作文Newton was an extraordinary figure in the history of science, and his contributions to physics, mathematics, and astronomy have had a profound impact on our understanding of the natural world. Born in 1643 in Woolsthorpe, England, Newton's early life was marked by hardship and challenges. However, his remarkable intelligence and curiosity led him to become one of the most influential scientists of all time.Newton's most famous work, "Philosophiæ Naturalis Principia Mathematica," commonly known as the Principia, was published in 1687. In this work, Newton laid the foundation for classical mechanics, formulating his three laws of motion and the law of universal gravitation. These laws provided a comprehensive explanation of motion and gravity, revolutionizing the field of physics.One of Newton's most significant contributions to mathematics was his development of calculus. Although thereis some controversy surrounding the development of calculus and its relationship with Gottfried Wilhelm Leibniz, Newton's work laid the groundwork for this branch of mathematics, which is widely used today in various fields.In addition to his work in physics and mathematics, Newton made significant contributions to astronomy. Hebuilt the first practical reflecting telescope, now known as the Newtonian telescope, which allowed for significant advancements in the field of astronomy. Newton's work in optics also laid the foundation for the modern understanding of light and color.Newton's work was not without controversy, particularly his disputes with other scientists of his time, most notably Robert Hooke and Leibniz. However, hiscontributions to science are undeniable, and his legacy continues to inspire scientists and researchers around the world.In conclusion, Newton's contributions to science, mathematics, and astronomy are unparalleled. His laws ofmotion and universal gravitation laid the foundation for classical mechanics, while his development of calculus revolutionized mathematics. Newton's work in optics and astronomy also significantly advanced these fields. Despite facing numerous challenges in his life, Newton's intellect and curiosity drove him to become one of the greatest scientists of all time, leaving a lasting legacy that continues to influence our understanding of the natural world.。

Isaac Newton,a towering figure in the history of science,was born in1642in Woolsthorpe,England.His contributions to various fields,including mathematics, physics,and astronomy,have left an indelible mark on human knowledge.Newtons early life was marked by a thirst for knowledge.He attended the University of Cambridge,where he studied mathematics and physics.It was during this time that he began to develop his groundbreaking theories.One of Newtons most famous achievements is the formulation of the laws of motion. These laws describe the relationship between a body and the forces acting upon it,and its motion in response to those forces.The laws of motion laid the foundation for classical mechanics and have been instrumental in the development of engineering and technology.In addition to his work in physics,Newton made significant contributions to the field of mathematics.He developed calculus,a branch of mathematics that deals with rates of change and the accumulation of quantities.Calculus has become an essential tool in fields ranging from economics to engineering.Newtons work in astronomy is also noteworthy.He formulated the law of universal gravitation,which states that every particle in the universe attracts every other particle with a force that is proportional to the product of their masses and inversely proportional to the square of the distance between their centers.This law helped to explain the motion of celestial bodies and laid the groundwork for modern astronomy.Newtons work was not limited to theoretical science.He also made practical contributions,such as the development of the reflecting telescope.This invention improved the quality of astronomical observations and contributed to the advancement of the field.Despite his many accomplishments,Newton was a humble man who was deeply religious. He believed that his discoveries were a way to understand the workings of Gods creation. His life serves as an inspiration to those who seek to explore the mysteries of the universe and to push the boundaries of human knowledge.In conclusion,Isaac Newtons contributions to science have had a profound impact on our understanding of the world.His work in physics,mathematics,and astronomy has shaped the course of scientific inquiry and continues to influence the way we view the universe. His legacy is a testament to the power of human curiosity and the pursuit of knowledge.。

Isaac Newton,a name that resonates with the very fabric of scientific discovery and intellectual curiosity,was not just a figure from history but a beacon of human potential.His life and work continue to inspire generations of scholars and laymen alike,a testament to the power of the human mind to unravel the mysteries of the universe.Born in Woolsthorpe,England,on January4,1643,Newtons early life was marked by a quiet,almost solitary existence.His mother,a widow,had remarried,leaving young Isaac in the care of his grandmother.This period of isolation,however,did not deter him from his innate curiosity about the world around him.Instead,it may have fueled his desire to understand the natural phenomena that surrounded him.Newtons academic journey began at the University of Cambridge,where he was admitted as a student at Trinity College in1661.It was here that his genius truly began to flourish.He delved into the works of philosophers and mathematicians,expanding his knowledge and challenging the prevailing ideas of his time.His studies in mathematics,optics,and mechanics laid the groundwork for his future contributions to science.One of the most significant moments in Newtons life came in1666,a year now famously known as the annus mirabilis or miracle year.During this time,he formulated the law of universal gravitation,developed the three laws of motion,and made significant advancements in the field of calculus. These groundbreaking ideas were not immediately shared with the world, however.It was not until1687that Newton published his magnum opus, PhilosophiæNaturalis Principia Mathematica,where he detailed his laws ofmotion and universal gravitation.Newtons laws of motion have become the cornerstone of classical physics. His first law,often referred to as the law of inertia,states that an object at rest will remain at rest,and an object in motion will continue in motion with a constant velocity unless acted upon by an external force.The second law,which introduces the concept of force,states that the acceleration of an object is directly proportional to the net force acting upon it and inversely proportional to its mass.The third law,perhaps one of the most wellknown,posits that for every action,there is an equal and opposite reaction.In addition to his work in physics,Newton made significant contributions to the field of optics.He conducted a series of experiments with light and prisms,demonstrating that white light is composed of a spectrum of colors.This discovery challenged the prevailing theory of the time,which held that colors were created by the interaction of light with objects. Newtons work in this area led to the development of the reflecting telescope,which used mirrors instead of lenses to gather and focus light, thereby reducing chromatic aberration.Despite his monumental contributions to science,Newton was not without his flaws.His personal life was marked by periods of intense work, punctuated by episodes of mental instability.He was known to be somewhat reclusive and had a contentious relationship with other scientists of his time,including the famous rivalry with Robert Hooke and Gottfried Wilhelm Leibniz over the invention of calculus.Newtons influence,however,extends far beyond his lifetime.His ideas have shaped the way we understand the physical world and have laid the foundation for much of modern science.His work in physics,in particular, has been instrumental in the development of technologies that we take for granted today,from the construction of buildings to the launching of spacecraft.In conclusion,Isaac Newtons life and work serve as a powerful reminder of the potential of human intellect to explore and understand the world around us.His contributions to science have left an indelible mark on our understanding of the universe,and his legacy continues to inspire future generations to push the boundaries of knowledge and discovery.。

牛顿提出关于自然科学的研究准则1.牛顿提出了重要的三大运动定律。

Newton proposed his important three laws of motion.2.这些定律成为了后来科学研究的基石。

These laws became the cornerstone of later scientific research.3.牛顿的研究为物理学和天文学领域带来了巨大的革新。

Newton's research brought about great innovation in the fields of physics and astronomy.4.他的工作极大地影响了整个科学领域的发展。

His work greatly influenced the development of the entire scientific field.5.牛顿的研究准则强调了实验和观察在科学研究中的重要性。

Newton's research principles emphasized the importance of experiment and observation in scientific research.6.他的方法论注重逻辑思维和严密的推理。

His methodology emphasized logical thinking and rigorous reasoning.7.牛顿在他的著作《自然哲学的数学原理》中详细阐述了他的研究准则。

Newton detailed his research principles in his work "Mathematical Principles of Natural Philosophy".8.这些准则对今天的科学研究依然具有深远影响。

These principles still have a profound impact on today's scientific research.9.牛顿的研究准则强调了对自然现象的客观观察和测量。

英语写作常用名人例子（2）英语写作常用名人例子Jeremy LinJeremy Lin, one of the most famous Chinese American basketball players, can be a good case of how confidence works in the course of pursuing success. He was once a common and little-known player at Harvard University. When fell on his evil days and became homeless, he could only lived in his friends' home. Without confidence and diligence, how could he have distinguished himself from other excellent players and achieved unprecedented success in NBA ?作为最著名的华裔篮球运动员之一的林书豪就是证明自信在追求成功的过程中是如何起作用的最好的例子。

他曾经只是哈佛大学校队的一名普通运动员，落魄的时候，无家可归，甚至只能住在朋友的家里。

如果没有强大的自信心和自身的努力，他怎么可能在高手如林的美职篮赛场上脱颖而出并且最终取得史无前例的成功呢?NewtonNewton，one of the most influential scientists throughout the history of mankind, could be best case of how modesty works in the course of making further progress and achieving greater success. Just imagine, having discovered the Law of Momentum Conservation, how could he have succeeded in establishing the Three Laws of Motion if he had simply taken pride in his past achievement ?作为人类历史上最具有影响力的科学家之一的牛顿，是谦虚如何在获得持续进步和取得更大成就上发挥作用的最好的例子。