0704.3166/ms.tex
1: \documentclass[12pt,preprint]{aastex}
2: %% manuscript produces a one-column, double-spaced document:
3: %% preprint2 produces a double-column, single-spaced document:
4: %\documentclass[preprint2]{aastex}
5: %\documentclass{emulateapj} %for page limit
6: %\documentclass[12pt,manuscript]{aastex} %for submission
7: %\usepackage{apjfonts}
8: 
9: \usepackage{graphics}
10: 
11: \shorttitle{XMM-Newton observations of Nova Sgr 1998}
12: \shortauthors{Hernanz and Sala}
13: 
14: \begin{document}
15: \title{XMM-Newton observations of Nova Sgr 1998}
16: 
17: \author{M. Hernanz}
18: 
19: \affil{Institut de Ci\`encies de l'Espai (CSIC-IEEC), 
20: Campus Universitat Aut\`onoma de Barcelona,
21: Facultat de Ci\`encies, Torre C5 - parell - 2a planta,
22: E-08193 Bellaterra (Barcelona), Spain; hernanz@ieec.uab.es} 
23: 
24: \and 
25: 
26: \author{G. Sala}
27: \affil{Max-Planck-Institut f\"ur extraterrestrische Physik,
28: Postfach 1312, D-85741, Garching, Germany and  
29: Institut de Ci\`encies de l'Espai (CSIC-IEEC); gsala@mpe.mpg.de}
30: 
31: 
32: \begin{abstract}
33: We report on X-ray observations of Nova Sagittarius 1998 (V4633 
34: Sgr), performed with XMM-Newton at three different epochs, 934, 1083 and 1265 
35: days after discovery. The nova was 
36: detected with the EPIC cameras at all three epochs, with emission 
37: spanning the whole energy range from 0.2 to 10 keV. 
38: 
39: The X-ray spectra do not change significantly at the different epochs, 
40: and are well fitted for the first and third observations 
41: with a multi-temperature optically thin thermal plasma, 
42: while lower statistics in the second observations lead to a poorer fit. 
43: The thermal plasma emission is most probably originated in 
44: the shock heated ejecta, with 
45: chemical composition similar to that of a CO nova. However, we can not completely 
46: rule out reestablished accretion as the origin of the emission.
47: We also obtain upper limits for the temperature and luminosity of a 
48: potential white dwarf atmospheric component, and conclude that hydrogen burning 
49: had already turned-off by the time of our observations.
50: \end{abstract}
51: 
52: \keywords{stars: individual (V4633 Sagittarius) --- stars: novae,
53: cataclysmic variables --- stars: white
54: dwarfs --- X-rays: binaries --- X-rays: individual (V4633
55: Sagittarius)}
56: 
57: \section{Introduction}
58: Nova Sgr 1998 (V4633 Sgr) was discovered by \cite{Lil98} on 1998 March 22.3~UT, 
59: with magnitude 7.8, and it was confirmed spectroscopically by \cite{DV98}
60: two days later, with relatively low expansion velocities and presence of 
61: iron, thus indicating that the nova belonged to the Fe~II class \citep{Wil92}. 
62: The nova light curve \citep{LJ99} indicated that it was moderately fast, with 
63: t$_2\approx$28~days and t$_3\approx$55~days. 
64: No indication of optical polarization was found in 
65: further observations by \cite{IKA00}. \cite{Lip01} performed 
66: optical observations during the period 1998--2000, which revealed two 
67: photometric periodicities; the shorter and constant period of 3.014~h was 
68: interpreted as the orbital period, whereas the longer and slightly 
69: variable period of 3.08~h could be interpreted as the spin period of the 
70: white dwarf, in a nearly synchronous magnetic system. Near-infrared 
71: spectroscopy, in the range 0.8 to 2.5~$\mu$m, obtained 525 and 850~days after 
72: peak brightness by \cite{Lyn01}, revealed some symmetric lines with widths 
73: (FWHM) of 1800~km/s. There was no evidence of dust formation and the 
74: shell was hydrogen-deficient. The reddening was very uncertain 
75: (E(B-V)=$0.3\pm0.2$, yielding A$_V=0.9\pm0.6$ for R=3.1), thus leading to a 
76: very uncertain distance, d$\sim$9~kpc (d=8.9$\pm$2.5~kpc, \cite{Lip01}). 
77: 
78: Nova Sgr 1998 was one of the targets included in our monitoring programme 
79: of recent galactic novae in X-rays, during the first cycle 
80: of the XMM-Newton satellite. The main goal of our observations was to 
81: determine the turn-off times of hydrogen burning. Novae are the 
82: consequence of explosive hydrogen burning, through a 
83: thermonuclear runaway, on top of an accreting white dwarf in a close binary 
84: system, of the cataclysmic variable type.
85: It is theoretically predicted that 
86: novae return to hydrostatic equilibrium after the ejection of a fraction of 
87: the accreted envelope. X-ray emission in post-outburst novae can arise 
88: through three different mechanisms. First, as the envelope mass is 
89: depleted, the photospheric radius decreases at constant bolometric luminosity 
90: (close to the Eddington value) with an increasing effective temperature. 
91: This leads to a hardening of the spectrum from optical through 
92: ultraviolet, extreme ultraviolet and finally soft X-rays, with 
93: the post-outburst nova emitting as a supersoft source with a 
94: hot white dwarf atmosphere spectrum.
95: The duration of this soft X-ray emission is related to 
96: the nuclear burning timescale of the remaining H-rich envelope 
97: (see, for instance, Table~1 in \cite{Geh98}) and depends 
98: among other factors on the white dwarf mass \citep{TT98,SH05a,SH05b}. 
99: The second site of X-ray emission in post-outburst novae is the ejected shell.
100: Internal or external shocks can heat the expanding gas up to temperatures of a few keV,
101: leading to the emission of X-rays with a thermal plasma spectrum. Finally, 
102: when accretion is reestablished in the cataclysmic variable, 
103: the accretion flow is responsible for the emission of X-rays, also with 
104: a thermal plasma spectrum. 
105: 
106: A systematic search for X-ray emission of classical novae
107: was performed by \cite{Ori01a} in the ROSAT archival data, 
108: which contained observations of 108 classical and recurrent novae, 
109: 39 of which were less than 10 years old. 
110: Contrary to expectations, just a few novae were detected in X-rays, and
111: only three of them with a soft X-ray spectrum: GQ~Mus (Nova~Mus~1983, 
112: already discovered by EXOSAT in 1983, \cite{Oge84}), 
113: V1974~Cyg (Nova~Cyg~1992, \cite{Kra96,Bal98}), and 
114: Nova~LMC~1995 \citep{OG99}. GQ~Mus is renowned for having the longest soft X-ray 
115: emission phase (around 9 years; \cite{Oge93,Sha95,BK01}).
116: Other novae were detected in X-rays with ROSAT, but without any 
117: soft component: V351~Pup (Nova~Pup~1991, \cite{Ori96}), and V838~Her (Nova~Her~1991,
118: \cite{Llo92,Obr94}). 
119: Finally, only 11 out of the 81 galactic quiescent novae observed by 
120: ROSAT were detected in hard X-rays. For all these non soft X-ray sources, 
121: the poor spectral resolution and/or limited spectral range 
122: of the detectors left the origin of the emission 
123: (either shocked ejecta or reestablished accretion) unclear.
124: 
125: After the ROSAT era, observations by Beppo-SAX and Chandra revealed new  
126: interesting results. A soft component was reported for
127: V382~Vel (Nova~Vel~1999) which lasted for 7.5--8 months and contained
128: a wealth of emission lines never previously detected in novae \citep{Ori02,Bur02}.
129: In the case of V1494~Aql (Nova~Aql~1999~No.2), a soft component appeared 6--10 months after the 
130: explosion, with a puzzling light curve including a short burst 
131: and oscillations \citep{Dra03}. Observations of the recurrent nova CI~Aql 
132: by \cite{Gre02} showed a soft component, but its spectrum implied an emitting 
133: radius of only 40~km for this source, which is too small for a white dwarf.
134: In addition, supersoft X-ray emission 6 months after outburst, with strong 
135: temporal variation was reported by \cite{Nes03} and \cite{Pet05} 
136: for V4743~Sgr (Nova~Sgr~2002~No.3).
137: 
138: XMM-Newton also contributed with some more observations of novae.
139: X-ray emission with a soft spectrum was detected from
140: Nova~LMC~1995 by \cite{Ori03} more than 5.5 years after the explosion.
141: This component was not detected in later observations by \cite{Ori04} in 2003. 
142: X-rays were also detected from Nova~LMC~2000 by \cite{Gre03} on days 17 and 51 after outburst, 
143: with a thermal plasma emission related to the ejecta and fading away 
144: on day 294. Finally, in the case of V2487~Oph (Nova~Oph~1998), \cite{HS02} 
145: detected X-ray emission with a broad thermal plasma spectrum extending up to $\sim 10$~keV and
146: a fluorescent Fe K$_\alpha$ line, clearly indicating the reestablishement of accretion in a 
147: magnetic cataclysmic variable.
148: 
149: X-ray observations of classical novae during their post-outburst
150: stages provide crucial information about the nova phenomenon 
151: \citep{Ori99,Ori04,Kra02}: soft X-rays yield a unique insight 
152: into the remaining nuclear burning on top of the white dwarf, 
153: while hard X-rays reflect the shocks in the ejecta in the post-outburst novae, 
154: and the properties of the accretion flow in the ``quiescent novae''. 
155: In view of the scarcity of objects observed and the diversity of 
156: behaviours detected, only the monitoring of as many novae as possible, 
157: with large sensitivity and spectral resolution (as those offered 
158: by XMM-Newton and Chandra) can help to understand and solve these issues. 
159: The observations of Nova Sgr 1998 with XMM-Newton reported in 
160: this paper contribute to achieve this goal. 
161: In \S2 we describe the observations, 
162: and in \S3 the spectral analysis of the data is explained in detail. 
163: Discussion follows in \S4 and a summary in \S5.
164: 
165: \section{Observations}
166: The X-Ray Multi-Mirror Mission, XMM-Newton, is an X-ray 
167: astrophysics observatory (see \cite{Jan01} for a general description)
168: including three X-ray telescopes, each with an 
169: European Photon Imaging Camera (EPIC) at its focal plane, and an optical/UV 
170: telescope, OM \citep{Mas01}. The detectors of two of the EPIC cameras use 
171: MOS CCDs \citep{Tur01}, 
172: whereas the third one uses pn CCDs \citep{Str01}. In addition, two 
173: Reflection Grating Spectrometers \citep{Her01} for high resolution 
174: X-ray spectroscopy are 
175: mounted in the light path of the two EPIC MOS cameras.   
176: 
177: We observed Nova Sgr 1998 at three epochs, with intervals of 6 months 
178: between observations: October 11, 2000, March 9, 2001 
179: and September 7, 
180: 2001 (i.e., 934, 1083 and 1265 days after outburst; see Table \ref{tab:log}).
181: Exposure times ranged from 6~ks to 9~ks. 
182: While the source was clearly detected with the EPIC MOS1, MOS2 and pn cameras, 
183: the signal to noise ratio was too small to detect it with the RGS 
184: instruments.
185: 
186: Data were reduced using the XMM-Newton Science Analysis System 
187: (SAS~6.5.0). Standard procedures described in the SAS documentation 
188: \citep{sas05,abc04} were applied. 
189: New calibration files affecting the soft spectrum 
190: of observations of the first XMM-Newton cycle 
191: were released in January 2006 \citep{Kir06}, 
192: and the pipeline products of some exposures of our observations were not available in the
193: archive. Therefore, we reprocessed the original Observation Data Files (ODFs) 
194: for all exposures using the most recent calibrations. 
195: Only events with pattern smaller than 12~(MOS) and 4~(pn) 
196: and with pulse height (PI) in the range 0.2--12~keV~(MOS) and
197: 0.2--15~keV~(pn) were selected for spectral analysis. 
198: Flaring background periods affecting some EPIC-pn exposures were 
199: filtered by creating good time interval (GTI) filter tables that 
200: were used for further event select procedures. 
201: Source photons were extracted from a circle of 35'', 
202: and the background was extracted 
203: from a region of the same area, close to the source but free from source 
204: photons, and keeping the same distance to the readout node (RAWY) as the 
205: source region \citep{Kir06}. The statistics of our observations were not 
206: good enough to perform timing analysis, which would have been interesting to 
207: determine the presence of some periodicity, as observed in the optical by \cite{Lip01}.
208: 
209: 
210: \section{Spectral analysis}
211: The X-Ray Spectral Fitting Package (XSPEC~11.3, \cite{xsp03}) was used 
212: for spectral analysis. For each observation, the spectra of the three EPIC 
213: cameras (pn, MOS1 and MOS2) were simultaneously fitted. 
214: An absorbing column density compatible with or smaller than the interstellar 
215: value was obtained for all spectral models.
216: A column density smaller than the interstellar N$_{\rm H}$ 
217: is clearly unphysical, and the spectral fits never indicate a larger value.
218: Therefore, we froze it in all our fits to N$_{\rm H}$= 1.6$\times 10^{21}$cm$^{-2}$,
219: which is the average interstellar column density in the direction of Nova~Sgr~1998
220: \citep{DL90}. This value is also consistent with the N$_{\rm H}$ derived from 
221: the measured extinction \citep{Lyn01}, through the empirical relationship between 
222: interstellar X-ray absorption and optical extinction \citep{Gor75}.
223: 
224: 
225: \subsection{Spectral models}
226: 
227: A first look at the data (Fig.~\ref{fig:spectra})
228: clearly shows that the emission spans the whole energy range 
229: of the EPIC instruments, $0.2-10$~keV, with a spectrum harder than 
230: that produced by residual H-burning on top of the white dwarf. 
231: This immediately indicates that the emission must be dominated by either the 
232: hot ejected shell, or by a reestablished accretion flow.
233: We thus fit the EPIC spectra with a multi-temperature 
234: optically thin thermal plasma, simulated with a 
235: Raymond-Smith model \citep{RS77} in XSPEC. 
236: The main processes included in this model are bremsstrahlung, 
237: recombination continua and line emission. More sophisticated models 
238: such as MEKAL \citep{MKL95} are senseless for our data, since these 
239: models require a larger number of parameters whereas we have a low number 
240: of counts. We checked, however, that the results obtained with MEKAL and 
241: Raymond-Smith models are similar. 
242: A single temperature thermal plasma model is completely unable to fit the data, 
243: because the emission spans a broad energy range. 
244: At least two components at different temperatures are needed to obtain a reasonable fit,
245: with the best fit being obtained for a three-temperature model.
246: 
247: Since the thermal plasma emission can arise either in the ejecta or in the 
248: accretion flow, two kinds of chemical composition are possible:
249: solar abundances (for an accretion flow), or metal enriched (for 
250: the nova ejecta). In this latter case, the underlying white dwarf of the nova can have 
251: two different compositions: carbon-oxygen (CO) for masses $\le 1.15$M$_\odot$, 
252: or oxygen-neon (ONe) for masses up to the Chandrasekhar limit. 
253: Here we use realistic compositions of the ejecta from \cite{JH98}. 
254: We tested their 14 ejecta compositions, 
255: corresponding to CO and ONe nova models with masses ranging 
256: from 0.8 to 1.15 M$_\odot$ and 1.0 to 1.35 M$_\odot$, respectively, and 
257: degrees of mixing between accreted mass and underlying core 
258: from 25 to 75 \%. In general, the best fit is obtained for 
259: low-mass CO novae (see below for details).
260: 
261: We also considered the possible contribution of residual H-burning on the 
262: post-outburst white dwarf (again CO or ONe).
263: We used white dwarf atmosphere emission models, gently 
264: provided by Jim MacDonald \citep{MV91}, and 
265: built tables to be read as external models in XSPEC.
266: Previous studies have shown the importance of including 
267: white dwarf atmosphere models instead of simple blackbodies 
268: for a correct spectral analysis of 
269: novae or supersoft X-ray sources \citep{Bal98,BK01,OG99}. 
270: 
271: 
272: \subsection{Results}
273: 
274: A three temperature (3T) optically thin thermal plasma model provides a good fit for the 
275: three EPIC cameras spectra of the first observation. The fit is only slightly poorer
276: for the third observation, while the poor statistics lead to a much worse 
277: fit at the second epoch. However, no other model provides a better fit and
278: the physical origin of the emission is very likely to be the same 
279: in all our observations. Therefore, we fit the spectra at the three epochs with 
280: the same 3T plasma model.
281: 
282: Figure~\ref{fig:spectra} (upper panel) shows Nova~Sgr~1998 spectra in October 2000, 
283: 934~days (2.6~years) after the explosion, with the best fit model and the 
284: residuals in units of $\chi$. 
285: The 3T plasma model provides a good fit to the EPIC MOS1, MOS2 and pn data, 
286: both simultaneously and individually.
287: The 3T model was used only after checking 
288: that a two temperature plasma model (which would be preferred because of its simplicity) 
289: results in a worse fit (see Table~\ref{tab:models_2T}), 
290: leaving significant residuals, in particular around 1~keV. 
291: We also 
292: attempted to fit the spectra with a cooling flow model, 
293: which is a good representation of a multi-temperature plasma in 
294: accretion disks in old novae \citep{Muketal03}. However, the spectrum is not well 
295: fitted with this model: with T$_{\rm low}$=0.08~keV, T$_{\rm high}$=3~keV 
296: and normalization constant $10^{-9}$M$_\odot$/yr, the reduced $\chi^2$ is 2.9;
297: therefore ºwe should rule out the cooling flow model 
298: for Nova~Sgr~1998.
299: 
300: Table~\ref{tab:models_3T} shows the 
301: best-fit parameters for the 3T plasma models both with solar and  
302: with CO nova abundances. In both cases, the three temperatures are
303: around 0.1, 0.8 and $> 5$ keV (probably indicating a continuous distribution 
304: between the two extremes). The main difference between the results with 
305: different abundances lies in the emission measures 
306: (EM, defined as ${\rm \int n_e n_i dV \simeq \int n_e^2 dV}$, with 
307: n$_{\rm e}$ and n$_{\rm i}$ the electronic and ionic densities, respectively).
308: The EM for the coolest component in the solar case is extremely large (by itself and 
309: when compared with the EM of the intermediate and high temperature components). 
310: But this problem disappears when a CO plasma is considered, because the emission 
311: measure for the coolest component decreases by a factor of 100. 
312: The difference is due to the strong emission 
313: lines at energies at 0.1-1~keV for the C, N and O enriched abundances. 
314: Confidence contours for the parameters of the three components are displayed 
315: in Figure~\ref{fig:contorns}.
316: 
317: %The corresponding luminosities are 
318: %L${\rm _{unabs.}}$ (0.2-10.0 keV) = (3-13)$\times 10^{33}\times ({\rm d/9kpc})^2$ 
319: %erg/s and L${\rm _{unabs.}}$ (0.2-10.0 keV) = 
320: %(2-8)$\times 10^{33}\times ({\rm d/9kpc})^2$ erg/s, for solar and CO cases, 
321: %respectively.
322: 
323: The presence of a soft X-ray atmospheric component in the spectrum of Nova Sgr 1998 
324: can not be ruled out, but only upper limits can be established (Fig.~\ref{fig:contatm}).
325: They include a broad confidence region of the T$_{\rm eff}$-L$_{\rm bol}$ parameters,
326: but taking into account the path of a post-outburst nova on the T$_{\rm eff}$-L$_{\rm bol}$ diagram, 
327: we can conclude that hydrogen burning had already turned off by the time of our first observation.
328: The initial evolution of a post-outburst nova follows a constant bolometric luminosity path  
329: (close to Eddington limit, i.e., $\sim 10^{38}$erg/s for a 1M$_{\odot}$ white dwarf) with increasing 
330: effective temperature \citep{KH94,SH05b}.
331: If Nova Sgr 1998 had been in this first phase with high luminosity 
332: at the time of our observation, 2.7~years after the outburst, 
333: it would have been at high temperature. 
334: But for a high luminosity, the upper limits for our first observation 
335: are compatible only with low temperatures, which would indicate an extended 
336: envelope. This is highly unlikely for a nova 2.6~years after the outburst.
337: Therefore we conclude that hydrogen burning had already turned-off 
338: by the time of our first observation.
339: The upper limits are however compatible with a cooling white dwarf: 
340: taking $10^8$cm as a hard lower limit for the radius of the degenerate star,  
341: Fig.~\ref{fig:contatm} indicates that it was colder than $3\times10^5$K. 
342: 
343: The second observation (3.0~years after the explosion, 
344: middle panel in Fig~\ref{fig:spectra}) provided lower 
345: signal to noise data (due to a smaller exposure time, 
346: $\sim 30\%$ less than in the first observation, Table~\ref{tab:log}). 
347: For completeness, we include a fit similar to that used for the first and third observations. 
348: Despite the large error ranges, the best-fit parameters  
349: (Table \ref{tab:models_3T}) indicate a slight softening of the spectrum, 
350: with lower temperatures for the first and third component.
351: 
352: The fit results for the third observation (3.5~years after the explosion, 
353: lower panel in Fig~\ref{fig:spectra}), 
354: are similar to the first one. However, 
355: the emission measures for the intermediate and highest 
356: temperature components are a bit smaller, whereas that associated with the coolest 
357: component is larger than one year before. This evolution is compatible with some 
358: cooling of the plasma between the first and third epochs. 
359:  
360: %The unabsorbed flux (in the observed range 0.2-10.0 keV) is 
361: %(2-33)$\times 10^{-13}$photons/cm$^2$/s (solar) and 
362: %(2-6)$\times 10^{-13}$photons/cm$^2$/s (CO) (see Tables \ref{tab:models_solar} 
363: %and \ref{tab:models_co1}). 
364: %The corresponding luminosities are 
365: %L${\rm _{unabs.}}$ (0.2-10.0 keV) = (2-32)$\times 10^{33}\times ({\rm d/9kpc})^2$ 
366: %erg/s and L${\rm _{unabs.}}$ (0.2-10.0 keV) = 
367: %(2-6)$\times 10^{33}\times ({\rm d/9kpc})^2$ erg/s, for solar and CO cases, 
368: %respectively.
369: 
370: \section{Discussion}
371: 
372: 
373: The optically thin thermal plasma emission may originate either 
374: in the shocked expanding nova shell or in the accretion flow. 
375: The problem of distinguishing between these two possibilities has already
376: been faced in the case of other sources:
377: Nova~Pup~1991 \citep{Ori96}, Nova~Her~1991 \citep{Llo92,Obr94}, 
378: Nova~Cyg~1992 \citep{Kra96,Bal98}, Nova~Vel~1999 \citep{Ori01b,MI01}. 
379: We discuss below some properties of the models which could help to disentangle the 
380: origin of the emission: luminosity and emission measure (i.e., value and 
381: temporal evolution of electronic density).
382: 
383: The unabsorbed X-ray (0.2--10 keV) luminosity for Nova~Sgr~1998
384: is in the range $10^{33}-10^{34}$~erg/s in all cases, for a distance of 9~kpc \citep{Lip01}. 
385: Distances to novae are quite uncertain, and Nova~Sgr~1998 is 
386: not an exception (d=8.9$\pm$2.5~kpc), but a factor of $\le 2$ in distance only 
387: changes the luminosity by a factor of $\le 4$.
388: This luminosity is too large to originate from the accretion
389: onto a non-magnetic white dwarf (typically $10^{29}-10^{32}$~erg/s, \cite{Kuu03}), 
390: or in a polar cataclysmic variable (with luminosities $\sim10^{32}$~erg/s, \cite{Muk03,Ram04}).
391: However, it could be related to accretion in an intermediate polar, 
392: with luminosities $\sim10^{33}$~erg/s \citep{Muk03}. In fact, optical observations suggested 
393: that Nova~Sgr~1998 is a nearly synchronous
394: magnetic system \citep{Lip01}, and photometric
395: observations of other post-novae indicate that
396: some are intermediate polars (for instance, V697~Sco, \cite{WW02}).
397: 
398: The X-ray luminosity of Nova~Sgr~1998 is also compatible with 
399: the observed values for nova ejecta. For example, 
400: Nova~Her~1991 emitted $5\times10^{33}-10^{35}$~erg/s only five days after outburst,
401: fading afterwards extremely quickly \citep{Llo92,Obr94};
402: Nova~Cyg~1992 ejecta X-ray emission reached $(0.8-2.0)\times 10^{34}$~erg/s 
403: around 150~days after outburst \citep{Bal98}; 
404: Nova~Pup~1991 emitted around $7.5\times 10^{33}$~erg/s 16~months 
405: after outburst \citep{Ori96}; and Nova~Vel~1999 also reached  
406: L$>10^{34}$~erg/s in the band above 0.8~keV 15~days after maximum 
407: (the shell luminosity was even larger, since emission lines detected below 0.8~keV 
408: also originated in the ejecta),
409: but it decreased to a few $10^{33}$~erg/s 5.5~months later \citep{Ori01b,Ori02}. 
410: In all those previous cases the ejecta cooled much faster than in Nova Sgr 1998. 
411: However, the expected cooling time for the electronic densities, n$_{\rm e}$, in nova ejecta 
412: (in general smaller than $10^6\rm{cm}^{-3}$) are larger than 10~years \citep{Bre77,Oge87}. 
413: Therefore, X-ray emission from shock heated ejecta in Nova~Sgr~1998 3.5~years after the 
414: outburst would be fully compatible with theoretical predictions of cooling times.
415: 
416: For a shock-heated plasma, temperatures are related to velocity as kT$\propto\rm{v}^{2}$
417: \citep{Bre77}. For the expansion velocities observed by \cite{Lyn01} 
418: (1800~km/s, on days 525 and 850 after outburst),
419: the plasma temperature would be $\sim$ 3 keV, which lies
420: within the range of temperatures of our best-fit thermal plasma models. 
421: Coronal lines, which are a good signature of shocks in the plasma, 
422: were also present in those observations, supporting the presence of a 
423: shocked ejecta in Nova~Sgr~1998. 
424: Assuming an expansion at constant velocity,
425: the emission volume at the time of our observations
426: would be increasing from $1.3\times 10^{49}\,\rm{cm}^{3}$ (for day~934) 
427: to $3.2\times 10^{49}\,\rm{cm}^{3}$ (day 1265).
428: With these volumes and the emission measures for CO models 
429: from Table~\ref{tab:models_3T}, electronic densities, $\rm{n}_{\rm{e}}$, in the shell would 
430: evolve from $1.5\times10^3\,\rm{cm}^{-3}$ (day~934) 
431: to $1.3\times10^3\,\rm{cm}^{-3}$ (day~1265) for the coolest component. Similarly,
432: $\rm{n}_{\rm{e}}$ decreases from $10^3\,\rm{cm}^{-3}$ to 500~cm$^{-3}$, and 
433: from $2\times10^3\,\rm{cm}^{-3}$ to $10^3\,\rm{cm}^{-3}$ 
434: for the intermediate and high temperature components, respectively. 
435: Since the spherical geometry gives the maximum emitting volume, these $\rm{n}_{\rm{e}}$
436: are lower limits. The small $\rm{n}_{\rm{e}}$ agree with the 
437: thinning of the ejecta found between days 525 and 850 by \cite{Lyn01}.
438: 
439: 
440: The electronic density in a spherical shell with its outer radius expanding 
441: at constant velocity evolves with time as t$^{-3}$, while in an expanding shell 
442: of constant thickness $\rm{n}_{\rm{e}}$ evolves as t$^{-2}$ \citep{Lyk03}. 
443: With the first expansion picture, a decrease of $\rm{n}_{\rm{e}}$ 
444: by a factor $\sim2.5=1/0.4$ is expected between
445: the first and the third observations, while the decay would
446: be by a factor of $\sim1.8=1/0.5$ in the second case. 
447: For Nova Sgr 1998, the $\rm{n}_{\rm{e}}$ derived above for the CO model 
448: changes by a factor of 0.8 for the coolest thermal plasma component.
449: For the intermediate and hottest components $\rm{n}_{\rm{e}}$ decreases 
450: by 0.5, the expected factor for an expanding shell of constant thickness.
451: Therefore, the $\rm{n}_{\rm{e}}$ evolution also supports the expanding ejecta as the 
452: possible site of the observed X-ray emission.
453: However, the large errors associated with n$_{\rm e}$ 
454: (in particular for the lowest temperature component) 
455: prevent a clear conclusion about the shell geometry.
456: 
457: 
458: \section{Summary}
459: 
460: The X-ray spectrum of Nova~Sgr~1998 as observed by XMM-Newton between 2.6 and 3.5 years 
461: after outburst is dominated by thermal plasma emission.
462: The best-fit is obtained with a three-component thermal plasma model, with temperatures
463: between 0.1 and 40~keV. Thermal plasma models with
464: different compositions (solar, and several abundances from  
465: realistic nova ejecta) have been tried. The best-fit is obtained with either the abundances of a 
466: CO nova shell (for 0.8~M$_\odot$ and 25\% mixing), 
467: or with solar abundances. 
468: The first case would correspond to emission from the expanding nova shell, 
469: while the second would be associated with reestablished accretion 
470: in the cataclysmic variable. 
471: 
472: The X-ray luminosity, the values of the plasma temperatures, 
473: the chemical composition, and the distribution and 
474: evolution of emission measures point to shock heated ejecta as the most likely 
475: origin for the plasma emission, 
476: rather than resumed accretion in the cataclysmic variable.
477: Also the inability to fit the data with a cooling flow model casts doubts on  
478: accretion as the origin of the X-ray emission.
479: Finally, the spectra of Nova~Sgr~1998 do
480: not show any fluorescence iron line, contrary to what was observed for
481: V2478~Oph \citep{HS02}, which clearly indicated that accretion 
482: was reestablished in the cataclysmic variable.
483: 
484: Regarding the presence of residual hydrogen burning on the white dwarf surface, 
485: our observations indicate that the post-outburst nova envelope had already 
486: turned off its hydrogen burning already 2.6~years after optical maximum.
487: 
488: \acknowledgments
489: We thank Jim MacDonald for kindly providing his white dwarf atmosphere 
490: models, and the referee for many helpful comments, which have considerably
491: improved this paper.
492: This work is based on observations obtained with XMM-Newton, an ESA science
493: mission with instruments and contributions directly funded by ESA Member
494: States and NASA. 
495: The XMM-Newton project is supported by the
496: Bundesministerium f\"ur Wirtschaft und Technologie/Deutsches Zentrum
497: f\"ur Luft- und Raumfahrt (BMWI/DLR, FKZ 50 OX 0001), the Max-Planck
498: Society and the Heidenhain-Stiftung.
499: This research has made use of the SIMBAD database,
500: operated at CDS, Strasbourg, France, and has been partially funded by the 
501: Spanish Ministry for Education and Science (MEC) project AYA2004-06290-C02-01 
502: and by FEDER. G.S. is supported through a postdoctoral fellowship from the 
503: MEC.  
504: 
505: \begin{thebibliography}{}
506: %\bibitem[Aizu(1973]{Aiz73}Aizu, K. 1973, Progr. Theor. Phys., 49, 1184
507: \bibitem[Arnaud \& Dorman(2003)]{xsp03} Arnaud, K., Dorman, B. 2003, 
508: XSPEC User's Guide for Version 11.3.x, (HEASARC, NASA/GSFC, Greenbelt, MD)
509: \bibitem[Balman et al.(1995)]{Bal95}Balman, S., Orio, M.
510: \& \"Ogelman, H. 1995, ApJ, 449, L47
511: \bibitem[Balman et al.(1998)]{Bal98}Balman, S., Krautter, J. 
512: \& \"Ogelman, H. 1998, ApJ, 499, 395
513: \bibitem[Balman \& Krautter(2001)]{BK01}Balman, S. 
514: \& Krautter, J. 2001, MNRAS, 326, 1441
515: \bibitem[Balman \& \"Ogelman(1999)]{Bal99}Balman, S. \& \"Ogelman, H. 
516: 1999, ApJ, 518, L111
517: \bibitem[Brecher et al.(1977)]{Bre77}Brecher, K., Ingham, W.H., Morrison, P. 
518: 1977, ApJ, 213, 492
519: \bibitem[Burwitz et al.(2002)]{Bur02}Burwitz, V, Starrfield, S., 
520: Krautter, J., Ness, J.U. 2002 in Classical
521: Nova Explosions, eds. M. Hernanz \& J. Jos\'e, AIP Conference Proceedings
522: 637, p.386
523: \bibitem[Della Valle et al.(1998)]{DV98}
524: Della Valle M., Pizzella A., Bernardi M., Jones A.F., Kiss L., Hornoch K., 
525: Schmeer P., Trigo J.M. 1998, IAUC 6848
526: \bibitem[den Herder et al.(2001)]{Her01}den Herder, J.W. et al. 2001, 
527: A\&A, 365, L7
528: \bibitem[Dickey \& Lockman(1990)]{DL90}Dickey, J.M., Lockman, F.J. 1990, 
529: ARAR, 28, 215
530: \bibitem[Drake et al.(2003)]{Dra03}Drake, J., Wagner, R. M., 
531: Starrfield, S., Butt, Y., Krautter, J.,
532: Bond, H. E., Della Valle, M., Gehrz, R. D., Woodward, C. E., Evans,
533: A., Orio, M., Hauschildt, P., Hernanz, M., Mukai, K., Truran, J. W.
534: 2003, ApJ, 584, 448
535: %\bibitem[Frank et al.(1988)]{FKR88}Frank, J., King, A.R., Raine, D.J. 1988 
536: %in Accretion power in astrophysics, Cambridge University Press.
537: %\bibitem[Fujimoto \& Ishida(1997)]{FI97}Fujimoto, R., Ishida, M. 1997, 
538: %ApJ, 474, 774
539: \bibitem[Gehrz et al.(1998)]{Geh98}Gehrz, R.D., Truran, J.W., Williams, R.E., 
540: Starrfield, S. 1998, PASP, 110, 3
541: \bibitem[Gorenstein(1975)]{Gor75}Gorenstein, P. 1975, ApJ, 198, 95
542: \bibitem[Greiner \& Di Stefano(2002)]{Gre02}Greiner, J. \& Di Stefano, R. 2002,
543: ApJ, 578, L59
544: \bibitem[Greiner et al.(2003)]{Gre03}Greiner, J., Orio, M. \& Schartel, N. 2003, 
545: A\&A, 405, 703
546: 
547: \bibitem[Hernanz \& Sala(2002)]{HS02}Hernanz, M. \& Sala, 
548: G. 2002, Science, 298, 393.
549: \bibitem[Ikeda, Kawabata \& Akitaya(2000)]{IKA00}Ikeda, Y., Kawabata, K.S., 
550: Akitaya, H. 2000, A\&A, 355, 256
551: \bibitem[Jansen et al.(2001)]{Jan01}Jansen, F. et al. 2001, A\&A, 365, L1
552: \bibitem[Jos\'e \& Hernanz(1998)]{JH98}Jos\'e, J., Hernanz, M. 1998, 
553: ApJ, 494, 680
554: \bibitem[Kato \& Hachisu(1994)]{KH94}Kato, M., Hachisu, I. 1994, ApJ, 437, 
555: 802
556: \bibitem[Kirsch(2006)]{Kir06}
557: Kirsch, M. 2006, EPIC status of calibration and data analysis, 
558: XMM-SOC-CAL-TN-0018 (ESA-ESAC, XMM-Newton SOC)
559: \bibitem[Krautter(2002)]{Kra02}Krautter, J. 2002, in Classical
560: Nova Explosions, eds. M. Hernanz \& J. Jos\'e, AIP Conference Proceedings
561: 637, p.345
562: \bibitem[Krautter et al.(1996)]{Kra96}Krautter, J., \"Ogelman, 
563: H., Starrfield, S., Wichmann, R. \& Pfeffermann. E. 1996, ApJ, 456, 788
564: \bibitem[Kuulkers et al.(2003)]{Kuu03}Kuulkers, E., Norton, A., Schwope, A., 
565: Warner, B. 2003, in Compact stellar X-ray sources, eds. W.H.G. Lewin, 
566: M. van der Klis, Cambridge University Press (in press; astro-ph/0302351)
567: \bibitem[Liller(1998)]{Lil98}Liller, W. 1998, IAUC 6846
568: \bibitem[Liller \& Jones(1999)]{LJ99}Liller, W., Jones, A.F. 1999, IBVS, 
569: 4664, 1
570: \bibitem[Lipkin et al.(2001)]{Lip01}Lipkin, Y., Leibowitz, E.M., Retter, A., 
571: Shemmer, O. 2001, MNRAS, 328, 1169
572: \bibitem[Lloyd et al.(1992)]{Llo92}Lloyd, H. M., O'Brien, T. J., 
573: Bode, M. F., Predehl, P., Schmitt, J.H.M.M., Trümper, J., Watson, M.G. 
574: \& Pounds, K. A. 1992, Nature, 356, 222
575: \bibitem[Loiseau(2005)]{sas05}Loiseau, N. 2005, Users' Guide to the
576: XMM-Newton Science Analysis System, Issue 3.2, (ESA/XMM-Newton Science 
577: Operations Centre, VILSPA, Madrid)
578: \bibitem[Lyke et al.(2003)]{Lyk03}
579: Lyke, J.E., Koenig, X. P., Barlow, M. J., Gehrz, R. D., Woodward, C.
580: E., Starrfield, S., P\'equignot, D., Evans, A., Salama, A.,
581: Gonz\'alez-Riestra, R., Greenhouse, M.A., Hjellming, R. M., Jones, T.J., 
582: Krautter, J., \"Ogelman, H. B., Wagner, R. M., Lumsden, S. L., Williams, R. E.
583: 2003, AJ, 126, 993
584: \bibitem[Lynch et al.(2001)]{Lyn01}Lynch, D.K., Rudy, R.J., Venturini, C.C.,
585: Makuz, S. 2001, AJ, 122, 2013
586: \bibitem[MacDonald \& Vennes(1991)]{MV91}Mac Donald, J. \& 
587: Vennes, S., 1991, ApJ, 373, L51
588: \bibitem[Mason et al.(2001)]{Mas01}Mason, K.O. et al. 2001, A\&A, 365, L36
589: \bibitem[Mewe et al.(1995)]{MKL95}Mewe, R., Kaastra, J.S., Liedhal, D.A., 
590: 1995, Legacy 6, 16 (available at 
591: http://heasarc.gsfc.nasa.gov/docs/journal/meka6.html)
592: \bibitem[Mukai \& Ishida(2001)]{MI01}Mukai, K., Ishida, M. 
593: 2001, ApJ, 551, 1024
594: \bibitem[Mukai (2003)]{Muk03}Mukai, K. 2003, AdSpR, 32, 2067
595: \bibitem[Mukai et al.(2003)]{Muketal03}
596: Mukai, K., Kinkhabwala, A., Peterson, J.R., Kahn, S.M. \& 
597: Paerels, F. 2003, ApJ, 586, 77
598: \bibitem[Ness et al.(2003)]{Nes03}Ness, J.-U., Starrfield, S., Burwitz, V., 
599: Wichmann, R., Hauschildt, P., Drake, J.J., Wagner, R.M., Bond, H.E., 
600: Krautter, J., Orio, M., Hernanz, M., Gehrz, R. D., Woodward, C. E., 
601: Butt, Y., Mukai, K., Balman, S., Truran, J. W. 2003, ApJ, 594, L127
602: \bibitem[O'Brien et al.(1994)]{Obr94}O'Brien, T.J.O., Lloyd, H.M., Bode, M.F. 
603: 1994, MNRAS, 271, 155
604: \bibitem[\"Ogelman(1984)]{Oge84}\"Ogelman, H., Beuermann, K. \& Krautter, 
605: J. 1984, ApJ, 287, L31
606: \bibitem[\"Ogelman(1987)]{Oge87}\"Ogelman, H., Krautter, J. \& 
607: Beuermann, K. 1987, A\&A, 177, 110
608: \bibitem[\"Ogelman et al.(1993)]{Oge93}\"Ogelman, H., Orio, M., 
609: Krautter, J. \& Starrfield, S. 1993, Nature,
610: 361, 331
611: \bibitem[Orio(1999)]{Ori99}Orio, M. 1999, Physics Reports, 311, 
612: 419
613: \bibitem[Orio et al.(1996)]{Ori96}Orio, M., Balman, S., Della 
614: Valle, M., Gallagher, J. \& \"Ogelman,
615: H. 1996, ApJ, 466, 410
616: \bibitem[Orio \& Greiner(1999)]{OG99}Orio, M. \& Greiner, J. 
617: 1999, A\&A, 344, L13
618: \bibitem[Orio et al.(2001a)]{Ori01a}Orio, M., Covington, J. \& 
619: \"Ogelman, H. 2001a, A\&A, 373, 542
620: \bibitem[Orio et al.(2001b)]{Ori01b}Orio, M., Parmar, A.N., Benjamin, R., 
621: Amati, L., Frontera, F., Greiner, J., \"Ogelman, H., Mineo, T., 
622: Starrfield, S., Trussoni, E. 2001b, MNRAS, 326, L13
623: \bibitem[Orio et al.(2002)]{Ori02}Orio, M., Parmar, A.N., Greiner, J., 
624: \"Ogelman, H., Starrfield, S., Trussoni, E. 2002, MNRAS, 333, L11
625: \bibitem[Orio et al.(2003)]{Ori03}Orio, M., Hartmann, E., Still, M., 
626: Greiner, J. 2003, ApJ, 594, 435
627: \bibitem[Orio (2004)]{Ori04}Orio, M., 2004, RMXAA, 20, 182
628: \bibitem[Petz et al.(2005)]{Pet05}Petz, A., Hauschildt, P.H., Ness, J.-U., 
629: Starrfield, S. 2005, A \& A, 431, 321
630: \bibitem[Ramsay \& Cropper (2004)]{Ram04}Ramsay, G. \& Cropper, M. 2004, MNRAS, 347, 497
631: \bibitem[Raymond \& Smith (1977)]{RS77}Raymond, J. \& Smith, B. 1977, 
632: ApJ Suppl, 35, 419
633: \bibitem[Sala \& Hernanz (2005a)]{SH05a} Sala, G. \& Hernanz, M. 2005a, A\&A,
634: 439, 1057.
635: \bibitem[Sala \& Hernanz (2005b)]{SH05b} Sala, G. \& Hernanz, M. 2005b, A\&A,
636: 439, 1061
637: \bibitem[Shanley et al.(1995)]{Sha95}Shanley, L., \"Ogelman, H., 
638: Gallagher, J. S., Orio, M. \& Krautter, J. 1995, ApJ, 438, L95 
639: \bibitem[Snowden et al.(2004)]{abc04}Snowden, S., Immler, S., Arida, M., 
640: Perry, B., Still, M., Harrus, I., 2004, An Introduction to XMM-NEWTON 
641: Data Analysis, 
642: Version 2.01 (NASA/GSFC XMM-Newton Guest Observer Facility, Greenbelt, MD) 
643: \bibitem[Starrfield(1989)]{Star89}Starrfield, S. 1989, in Classical 
644: Novae, Wiley, New York, p. 39
645: \bibitem[Str\"uder et al.(2001)]{Str01}Str\"uder, L.. et al. 2001, 
646: A\&A, 365, L18
647: \bibitem[Tuchman \& Truran(1998)]{TT98}Tuchman, Y. \&Truran, J.W. 1998, 
648: ApJ, 503, 381 
649: \bibitem[Turner et al.(2001)]{Tur01}Turner, M.J.L. et al. 2001, A\&A, 365, L27
650: \bibitem[Warner \& Woudt(2002)]{WW02}Warner, B., Woudt, P. 2002, PASP, 114,
651: 1222
652: \bibitem[Williams(1992)]{Wil92}Williams, R. 1992, AJ, 104, 725
653: 
654: \end{thebibliography}
655: 
656: \clearpage
657: 
658: % Tables
659: 
660: \begin{deluxetable}{rrccrccrcc} 
661: \tabletypesize{\scriptsize}
662: \tablecolumns{10} 
663: \tablewidth{0pt} 
664: \tablecaption{Observation log.
665: \label{tab:log}}
666: \tablehead{ 
667: \colhead{} & \colhead{} & 
668: \multicolumn{2}{c}{EPIC pn} & \colhead{} & 
669: \multicolumn{2}{c}{EPIC MOS1} & \colhead{} &
670: \multicolumn{2}{c}{EPIC MOS2} \\ 
671: \cline{3-4} \cline{6-7} \cline{9-10} \\ 
672: \colhead{Observation}  & \colhead{Time after}  & 
673: \colhead{Exposure}     & \colhead{Count Rate} & \colhead {} &
674: \colhead{Exposure}     & \colhead{Count Rate} & \colhead {} &
675: \colhead{Exposure}     & \colhead{Count Rate}\\ 
676: \colhead{Date (orbit)} & \colhead{Discovery}&
677: \colhead{Time (s)}     & \colhead{($10^{-2}\rm cts$ $\rm s^{-1}$)} 
678:                        & \colhead {} &
679: \colhead{Time (s)}     & \colhead{($10^{-2}\rm cts$ $\rm s^{-1}$)} 
680:                        & \colhead {} &
681: \colhead{Time (s)}     & \colhead{($10^{-2}\rm cts$ $\rm s^{-1}$)}}
682: \startdata
683: Oct.11, 2000 (154) & 934 d, 2.6 yr&
684: 6213 & $10.1\pm0.5$ & &
685: 9048 & $2.4\pm0.2$ & &
686: 9051 & $2.7\pm0.2$ \\ 
687: 
688: Mar. 9, 2001 (229) & 1083 d, 3.0 yr&
689: 3058 & $8.6\pm0.6$ & &
690: 6575 & $2.1\pm0.2$ & &
691: 6575 & $2.2\pm0.2$ \\ 
692: 
693: Sep. 7, 2001 (320) & 1265 d, 3.5 yr&
694: 5883 & $8.2\pm0.4$ & &
695: 9086 & $2.0\pm0.2$ & &
696: 9086 & $1.9\pm0.2$ \\
697: 
698: \enddata
699: \end{deluxetable} 
700: 
701: \begin{deluxetable}{lccc}
702: \tabletypesize{\scriptsize}
703: \tablecolumns{4} 
704: \tablewidth{0pt} 
705: \tablecaption{Parameters of the fit for each one of the three 
706: observations of Nova Sgr 1998, with a two temperatures thermal plasma 
707: (Raymond-Smith), with solar and CO1 abundances.
708: \label{tab:models_2T}}
709: \tablehead{ 
710: \colhead{} & \colhead{$1^{\rm st}$ Observation} & 
711: \colhead{$2^{\rm nd}$ Observation} & \colhead{$3^{\rm rd}$ Observation}\\
712: \cline{1-4}\\
713: \multicolumn{4}{c}{Solar abundances}}
714: \startdata
715: ${\rm kT_{RS1} (keV)}$  
716:                         & $0.12-0.20$
717:                         & $0.03-0.06$
718:                         & $0.06-0.12$\\
719: 
720: %${\rm K_{1}}$
721: %                       & $(0.7-2.8)\times 10^{-4}$
722: %                       & $0.02-2$
723: %                       & $(3-118)\times 10^{-4}$ \\
724: 
725: ${\rm EM_{RS1} (\times 10^{57} cm^{-3})}$ 
726:                         & $0.07-0.27$ 
727:                         & $20-2000$ 
728:                         & $0.3-11.3$\\
729: 
730: ${\rm kT_{RS2} (keV)}$  
731:                         & $4-11$
732:                         & $2-16$
733:                         & $2-6$   \\
734: 
735: %${\rm K_{2}}$
736: %                       & $(1-2)\times 10^{-4}$
737: %                       & $(0.6-1.0)\times 10^{-4}$
738: %                       & $(0.8-1.2)\times 10^{-4}$ \\
739: 
740: 
741: ${\rm EM_{RS2} (\times 10^{55} cm^{-3})}$  
742:                         & $12-15$
743:                         & $6-10$
744:                         & $8-12$\\
745: 
746: 
747: F$_{\rm unabs,0.2-10.0\, \rm keV} (\times 10^{-13} \rm erg\, cm^{-2}\, s^{-1})$ 
748:                         & $4-7$ 
749:                         & $2-210$ 
750:                         & $2-30$ \\
751: 
752: L$_{\rm unabs,0.2-10.0\, \rm keV} (\times 10^{33} \rm erg\, s^{-1})$ 
753:                         & $4-7$ 
754:                         & $2-203$ 
755:                         & $2-29$ \\
756: 
757: $\chi^{2}_{\nu}$        
758:                         & 1.55
759:                         & 2.07
760:                         & 1.66 \\
761: 
762: \cutinhead{CO abundances}
763: 
764: ${\rm kT_{RS1} (keV)}$  
765:                         & $0.09-0.19$ 
766:                         & $0.03-0.07$ 
767:                         & $0.07-0.15$ \\
768: 
769: %${\rm K_{1}}$
770: %                       & $(0.3-1.8)\times 10^{-5}$
771: %                       & $10^{-4}-0.3$
772: %                       & $(0.4-4.5)\times 10^{-5}$ \\
773: 
774: ${\rm EM_{RS1} (\times 10^{55} cm^{-3})}$ 
775:                         & $0.3-1.7$ 
776:                         & $10-3\times10^{4}$ 
777:                         & $0.4-4.4$  \\
778: 
779: ${\rm kT_{RS2} (keV)}$  
780:                         & $3-8$ 
781:                         & $1-6$ 
782:                         & $2-6$    \\
783: 
784: %${\rm K_{2}}$
785: %                       & $(5-6)\times 10^{-5}$ 
786: %                       & $(1-4)\times 10^{-5}$ 
787: %                       & $(3-5)\times 10^{-5}$ \\
788: 
789: 
790: ${\rm EM_{RS2} (\times 10^{55} cm^{-3})}$  
791:                         & $5-6$
792:                         & $1-4$
793:                         & $2-5$   \\
794: 
795: 
796: F$_{\rm unabs,0.2-10.0\, \rm keV} (\times 10^{-13} \rm erg\, cm^{-2}\, s^{-1})$ 
797:                         & $4-6$ 
798:                         & $1-80$ 
799:                         & $2-6$ \\
800: 
801: L$_{\rm unabs,0.2-10.0\, \rm keV} (\times 10^{33} \rm erg\, s^{-1})$ 
802:                         & $4-6$ 
803:                         & $1-77$ 
804:                         & $2-6$ \\
805: 
806: $\chi^{2}_{\nu}$        
807:                         & 1.37
808:                         & 2.01
809:                         & 1.57 \\
810: \enddata
811: \tablecomments{The best-fit models shown fit simultaneously the data 
812: from the three EPIC cameras.
813: All the limits are $3\sigma$.
814: The emission measures and luminosities are 
815: given for a distance d=9kpc; a factor 
816: ${\rm (d/9 kpc)^2}$, with d the distance in kpc, affects these magnitudes.
817: The absorption ${\rm N_H}$ is $1.6 \times 10^{21} {\rm cm^{-2}}$ everywhere.}
818: \end{deluxetable} 
819: 
820: 
821: \begin{deluxetable}{lccc}
822: \tabletypesize{\scriptsize}
823: \tablecolumns{4} 
824: \tablewidth{0pt} 
825: \tablecaption{Parameters of the best-fit model for each one of the three 
826: observations of Nova Sgr 1998: three temperatures thermal plasma 
827: (Raymond-Smith), with solar and CO1 abundances. 
828: \label{tab:models_3T}}
829: \tablehead{ 
830: \colhead{} & \colhead{$1^{\rm st}$ Observation} & 
831: \colhead{$2^{\rm nd}$ Observation} & \colhead{$3^{\rm rd}$ Observation}\\
832: \cline{1-4}\\
833: \multicolumn{4}{c}{Solar abundances}}
834: \startdata
835: 
836: ${\rm kT_{RS1} (keV)}$  
837:                         & $0.07-0.16$  
838:                         & $0.03-0.05$
839:                         & $0.06-0.10$\\
840: 
841: 
842: ${\rm EM_{RS1} (\times 10^{57} cm^{-3})}$ 
843:                         & $0.1-2.4$ 
844:                         & $50-5000$ 
845:                         & $0.5-6.4$\\
846: 
847: ${\rm kT_{RS2} (keV)}$  
848:                         & $0.6-1.0$
849:                         & $0.2-0.7$
850:                         & $0.2-1.0$   \\
851: 
852: ${\rm EM_{RS2} (\times 10^{55} cm^{-3})}$  
853:                         & $0.9-3.7$
854:                         & $0.5-3.9$
855:                         & $0.4-2.0$\\
856: 
857: 
858: ${\rm kT_{RS3} (keV)}$  
859:                         & $\geq5$
860:                         & $\geq2$
861:                         & $\geq3$   \\
862: 
863: ${\rm EM_{RS3} (\times 10^{55} cm^{-3})}$  
864:                         & $8-14$
865:                         & $5-10$
866:                         & $6-10$\\
867: 
868: 
869: F$_{\rm unabs,0.2-10.0\, \rm keV} (\times 10^{-13} \rm erg\, cm^{-2}\, s^{-1})$ 
870:                         & $3-13$ 
871:                         & $2-640$ 
872:                         & $2-33$ \\
873: 
874: L$_{\rm unabs,0.2-10.0\, \rm keV} (\times 10^{33} \rm erg\, s^{-1})$ 
875:                         & $3-13$ 
876:                         & $2-620$ 
877:                         & $2-32$ \\
878: 
879: $\chi^{2}_{\nu}$        
880:                         & 1.16
881:                         & 1.40
882:                         & 1.22 \\
883: 
884: \cutinhead{CO abundances}
885: 
886: ${\rm kT_{RS1} (keV)}$  
887:                         & $0.07-0.15$
888:                         & $0.02-0.07$
889:                         & $0.05-0.13$\\
890: 
891: 
892: ${\rm EM_{RS1} (\times 10^{55} cm^{-3})}$ 
893:                         & $0.3-5.8$ 
894:                         & $14-500000$ 
895:                         & $0.5-11.0$\\
896: 
897: ${\rm kT_{RS2} (keV)}$  
898:                         & $0.6-1.1$
899:                         & $0.4-1.2$
900:                         & $0.4-1.1$   \\
901: 
902: ${\rm EM_{RS2} (\times 10^{55} cm^{-3})}$  
903:                         & $0.6-2.0$
904:                         & $0.4-1.9$
905:                         & $0.2-1.7$\\
906: 
907: 
908: ${\rm kT_{RS3} (keV)}$  
909:                         & $\geq5$
910:                         & $\geq2$
911:                         & $\geq3$   \\
912: 
913: ${\rm EM_{RS3} (\times 10^{55} cm^{-3})}$  
914:                         & $3-7$
915:                         & $1-5$
916:                         & $2-5$\\
917: 
918: 
919: F$_{\rm unabs,0.2-10.0\, \rm keV} (\times 10^{-13} \rm erg\, cm^{-2}\, s^{-1})$ 
920:                         & $2-8$ 
921:                         & $1-5500$ 
922:                         & $2-6$ \\
923: 
924: L$_{\rm unabs,0.2-10.0\, \rm keV} (\times 10^{33} \rm erg\, s^{-1})$ 
925:                         & $2-8$ 
926:                         & $1-5300$ 
927:                         & $2-6$ \\
928: 
929: $\chi^{2}_{\nu}$        
930:                         & 0.92
931:                         & 1.61
932:                         & 1.25 \\
933: \enddata 
934: \tablecomments{Same comments as in table \ref{tab:models_2T}.}
935: \end{deluxetable} 
936: 
937:  
938: 
939: %% End of tables
940: 
941: \clearpage
942: 
943: %% Figures
944: 
945: \begin{figure}
946: \begin{center}
947: \includegraphics[width=0.7\textwidth]{fig1.eps}
948: \caption{Nova Sgr 1998 EPIC MOS1 (black), MOS2 (red) and pn (green) observed spectra, with the best fit 
949: model (with CO abundances) and the residuals. From top to low panels: first,
950: second and third observations. 
951: \label{fig:spectra}}
952: \end{center}
953: \end{figure}
954: 
955: \begin{figure}
956: \begin{center}
957: \includegraphics[width=0.5\textwidth]{fig2.eps}
958: \caption{Confidence contours (1-innermost-, 2, and 3 -outermost- $\sigma$)
959: for the temperature and normalization constant 
960: ($\rm norm_{RS}=10^{-14}$EM$/(4\pi$D$^{2})$, 
961: where EM is the emission measure and D the distance in cm)
962: of the low (upper panel), intermediate (middle panel) and high (lower panel) 
963: temperature thermal plasma components, for the models with
964: CO abundances corresponding to the first observation of Nova Sgr 1998. 
965: \label{fig:contorns}}
966: \end{center}
967: \end{figure}
968: 
969: 
970: 
971: \begin{figure}
972: \begin{center}
973: \includegraphics[width=0.8\textwidth]{fig3.eps}
974: %{sgr98_1a_contatmos_atm3rs_co1_paper.ps}
975: %\includegraphics[totalheight=6.0cm]
976: \caption{Upper limits
977: (from left to right, 1, 2, and 3 $\sigma$ confidence) 
978: for the effective temperature (T$_{\rm atmos}$) and normalization
979: (norm=L$_{39}$/D$^2_{10}$, where L$_{39}$ is the luminosity in units of 
980: $10^{39}$erg/s and D$_{10}$ the 
981: distance to the source in units of 10 kpc) of the CO white dwarf atmosphere 
982: component (first observation). The thermal plasma model has three components 
983: and abundances corresponding to CO novae ejecta, as in previous figures.
984: The luminosity of the nova during the constant bolometric luminosity phase, 
985: L$\sim$$10^{38}$ erg/s (dashed line), as well as the locii of constant 
986: white dwarf radius equal to $10^8$ and $10^9$ cm (dotted lines) are 
987: also displayed.
988: \label{fig:contatm}}
989: \end{center}
990: \end{figure}
991: 
992: 
993: 
994: 
995: 
996: \end{document}
997: