1: % File draft date: 9 Dec 2003
2: % \documentstyle[12pt,../aasms]{article}
3: %\documentstyle[12pt,emulateapj5]{article}
4: \documentclass[12pt,preprint,flushrt]{aastex}
5: %\doublespace
6:
7: \begin{document}
8: \newcommand{\etal}{{\it et al.}}
9: \include{idl}
10:
11: \title{Detection of Nitrogen and Neon in the X-ray Spectrum of GP Com
12: with {\it XMM/Newton}} \author{Tod E. Strohmayer} \affil{Laboratory
13: for High Energy Astrophysics, NASA's Goddard Space Flight Center,
14: Greenbelt, MD 20771; stroh@clarence.gsfc.nasa.gov}
15:
16: \begin{abstract}
17:
18: We report on X-ray spectroscopic observations with {\it XMM/Newton} of
19: the ultracompact, double white dwarf binary, GP Com. With the
20: Reflection Grating Spectrometers (RGS) we detect the L$\alpha$ and
21: L$\beta$ lines of hydrogen-like nitrogen (N VII) and neon (Ne X), as
22: well as the helium-like triplets (N VI and Ne IX) of these same
23: elements. All the emission lines are unresolved. These are the first
24: detections of X-ray emission lines from a double-degenerate, AM CVn
25: system. We detect the resonance (r) and intercombination (i) lines of
26: the N VI triplet, but not the forbidden (f) line. The implied line
27: ratios for N VI, $R = f/i < 0.3$, and $G = (f + i)/r \approx 1$,
28: combined with the strong resonance line are consistent with formation
29: in a dense, collision-dominated plasma. Both the RGS and EPIC/MOS
30: spectra are well fit by emission from an optically thin thermal plasma
31: with an emission measure (EM) $\propto ( kT / 6.5 \; {\rm keV})^{0.8}$
32: (model {\it cevmkl} in XSPEC). Helium, nitrogen, oxygen and neon are
33: required to adequately model the spectrum, however, the inclusion of
34: sulphur and iron further improves the fit, suggesting these elements
35: may also be present at low abundance. We confirm in the X-rays the
36: underabundance of both carbon and oxygen relative to nitrogen, first
37: deduced from optical spectroscopy by Marsh et al. The average X-ray
38: luminosity of $\approx 3 \times 10^{30}$ ergs s$^{-1}$ implies a mass
39: accretion rate $\dot m \approx 9 \times 10^{-13} \; M_{\odot}\; {\rm
40: yr}^{-1}$. The implied temperature and density of the emitting plasma,
41: combined with the presence of narrow emission lines and the low $\dot
42: m$ value, are consistent with production of the X-ray emission in an
43: optically thin boundary layer just above the surface of the white
44: dwarf.
45:
46: \end{abstract}
47:
48: \keywords{Binaries: general - Stars: individual (GP Com) - Stars:
49: white dwarfs - cataclysmic variables - X-rays: stars - X-rays:
50: binaries}
51:
52: \section{Introduction}
53:
54: The AM CVn stars are among the most compact binary systems
55: known. Their optical spectra are dominated by broad helium emission
56: lines originating in an accretion disk formed by mass transfer from a
57: degenerate helium dwarf onto the primary white dwarf. Their orbital
58: periods range from about 10 minutes to 1 hour. They are natural
59: laboratories for the study of poorly understood binary evolution
60: processes, such as common envelope evolution. Moreover, the absence
61: of hydrogen allows for the study of accretion disks dominated by
62: helium. They may also be a significant channel for the production of
63: Type Ia supernovae and neutron stars via accretion induced collapse
64: (for a review see Warner 1995).
65:
66: GP Com is one of the better studied, nearby AM CVn systems. It has a
67: 46.5 minute orbital period (Nather Robinson \& Stover 1981), resides
68: at a distance of $\approx 70$ pc (Thorstensen 2003), and has been
69: studied extensively in the optical and UV. Recent studies by Marsh
70: (1999) and Morales-Rueda et al.(2003) have used optical and UV+optical
71: spectroscopy to probe the dynamics of the system in great detail. In
72: addition to helium, nitrogen emission lines have been seen in the UV
73: (Lambert \& Slovak 1981; Marsh et al. 1995), and optical (Marsh, Horne
74: \& Rosen 1991). These observations identified a nitrogen
75: overabundance, relative to carbon and oxygen, as well as an
76: underabundance of heavy elements (for example, Ca, Si and Fe). The low
77: metallicity may be consistent with its suggested halo origin (see
78: Giclas, Burnham \& Thomas 1961). The evolution scenario favored by
79: Marsh et al. (1991) posits an initially low metallicity for the
80: progenitor stars. The CNO elements are subsequently produced within
81: the primary star and then mixed throughout the secondary during a
82: common envelope phase. The nitrogen abundance is then increased by
83: CNO-cycle hydrogen burning within the secondary (ie. the mass donor).
84: Finally, a number of ultracompact systems with neutron star primaries
85: also show apparent neon enrichment (see Juett \& Chakrabarty 2003;
86: Schulz et al. 2001).
87:
88: Accretion should make such systems soft X-ray sources, and indeed,
89: several have been detected in the X-ray band, including the prototype
90: AM CVn, CR Boo, and GP Com (van Teeseling \& Verbunt 1994; Ulla 1995;
91: and Eracleous, Halpern \& Patterson 1991). The X-ray flux from
92: accreting, non-magnetic white dwarfs is thought to originate in a
93: boundary layer that shocks and decelerates the Keplerian flow,
94: allowing it to eventually settle on the white dwarf. Early work
95: suggested that the boundary layer should be optically thick--radiating
96: in the soft X-ray band and extreme UV--at mass accretion rates above
97: $\dot m \approx 1.6 \times 10^{-10} M_{\odot} \; {\rm yr}^{-1}$
98: (Pringle \& Savonije 1979). At lower rates the boundary layer should
99: become more radially extended and optically thin, producing a harder
100: X-ray spectrum. Aspects of this basic scenario have been confirmed by
101: more recent calculations (see, for example, Patterson \& Raymond 1985;
102: Narayan \& Popham 1993; Popham \& Narayan 1995).
103:
104: To date, X-ray observations of most AM CVn systems have been made at
105: relatively low signal to noise levels and with modest spectral
106: resolution. For example, ROSAT observations of GP Com found the source
107: at a flux level $\approx 1.2 \times 10^{-11}$ ergs cm$^{-2}$ s$^{-1}$
108: in the 0.2 - 2 keV band, and confirmed that it is variable. These
109: observations also found that the 0.2 - 2 keV spectrum cannot be
110: described by a single temperature component (van Teeseling \& Verbunt
111: 1994), but detailed modelling was problematic due to the relatively
112: poor statistics and restricted bandpass.
113:
114: In this Letter we summarize results of recent, high signal to noise
115: X-ray spectral measurements of GP Com with {\it XMM/Newton}. These
116: data represent perhaps the best X-ray spectral measurements of an AM
117: CVn system to date, and also show the first X-ray emission lines from
118: such a system.
119:
120: \section{XMM Observations}
121:
122: XMM/Newton observed GP Com for 56.5 ksec beginning on January 3,
123: 2001. The EPIC observations were done in imaging mode. We used version
124: 5.4.1 of the XMM software (SAS) to analyze the data. GP Com was easily
125: detected at the few counts s$^{-1}$ level in the EPIC instruments. A
126: lightcurve combining data from the PN and MOS detectors is shown in
127: Figure 1, and confirms substantial X-ray variability, including
128: several 2 to 4 minute flares during which the X-ray flux increased by
129: factors of 2 - 3. Flaring has often been seen in the UV and optical,
130: particularly in the N V ($1240 \AA$) resonance line (Marsh et
131: al. 1995). We will present a detailed timing study of GP Com in a
132: subsequent paper. We extracted spectra and produced response files
133: for each of the instruments using the relevant SAS tools.
134: %For the purposes of this work we present results on only the total,
135: %phase-averaged spectrum.
136:
137: \subsection{Spectral Analysis: RGS and EPIC/MOS}
138:
139: The spectrum extracted from both RGS instruments is shown in Figure 2,
140: along with identifications for all the strongest line features. The
141: strongest detected feature is the Ly$\alpha$ line of hydrogenic
142: nitrogen (N VII) at 24.78 $\AA$, but other lines of nitrogen are also
143: present, including the Ly$\beta$ line as well as the helium-like
144: triplet of N VI at $\approx 29 \AA$. The only other signifcant line
145: features are due to neon. The hydrogenic Ly$\alpha$ and Ly$\beta$
146: lines as well as the helium-like triplet are detected. Particularly
147: striking is the absence of any strong carbon or oxygen lines in the
148: spectrum.
149:
150: We fit the observed line features using Lorentzian profiles with the
151: widths fixed at that of the RGS line spread function (LSF) for that
152: wavelength (see den Herder et al. 2001; Pollock et al. 2003). All the
153: hydrogenic lines are well fit with such a profile and are
154: unresolved. Table 1 summarizes the salient properties of all the
155: detected lines.
156:
157: As is well known, the helium-like triplets can be used as probes of
158: the plasma where the lines are produced (see Gabriel \& Jordan 1969;
159: Porquet \& Dubau 2000). The line flux ratios, $R = f/i$, and $G = (i +
160: f)/r$, of the resonance (r), intercombination (i) and forbidden (f)
161: transitions are sensitive to the electron density and temperature of
162: the plasma, respectively. For each triplet we fit a Lorentzian profile
163: at the rest wavelength of each transition. To determine the line
164: ratios, we fixed the wavelength and width of each line, but allowed
165: the line fluxes to vary. The signal to noise ratio in both triplets
166: is not particularly high, but the line ratios can at least be roughly
167: estimated. The nitrogen triplet is better resolved than that of
168: neon. The results of our fits are shown in Figure 3 (see also Table
169: 1). For the N VI triplet we detect the resonance and intercombination
170: lines, but interestingly, not the forbidden line. The line ratios for
171: N VI, $R = f/i < 0.3$, and $G = (f + i)/r \approx 1$, and the strong
172: resonance line are consistent with formation in a dense,
173: collision-dominated plasma (Porquet \& Dubau 2000). For nitrogen, a
174: strong UV radiation field can also decrease the strength of the
175: forbidden line, mimicking a density effect, however, the line ratios
176: from the neon triplet--which are much less sensitive to radiation
177: field effects--are also consistent with a plasma at or near the
178: critical density $n_c = 5.3 \times 10^{9}$ cm$^{-3}$ for nitrogen.
179:
180: We fit the EPIC/MOS and RGS spectra with several collisional
181: ionization equilibrium (CIE) plasma emission models. We found that
182: multi-temperature models fit the data significantly better than single
183: temperature models. For example, the multi-temperature and variable
184: abundance model {\it cevmkl} in XSPEC (version 11.3) provides an
185: excellent fit to the data from both the RGS and EPIC instruments (see
186: Mewe, Gronenschild \& van den Oord 1985; Mewe, Lemen \& van den Oord
187: 1986; Liedahl, Osterheld \& Goldstein 1995). This model uses a power
188: law in temperature for the emission measure distribution (EM), vis.,
189: ${\rm EM} \propto ( kT/ kT_{max} )^{\gamma}$, where $kT_{max}$ and
190: $\gamma$ are the maximum temperature and power law index,
191: respectively.
192:
193: There is some evidence for modest pile-up in the PN spectra, so for
194: the broad-band modeling we focused on the MOS data only. Figure 4
195: shows our best fit to the EPIC/MOS data using the {\it cevmkl} model.
196: The model fits the data well, with a minimum $\chi^2$ of 1,005 for
197: 1,045 degrees of freedom. Using the MOS data only, the best-fit
198: emission measure distribution is characterized by $kT_{max} = 6.3 \pm
199: 0.3$ keV, and $\gamma = 0.85 \pm 0.05$. The 0.3 - 10 keV fluxes are
200: $4.7$ and $5.3 \times 10^{-12}$ ergs cm$^{-2}$ s$^{-1}$ in the MOS1
201: and MOS2, respectively. Taking the mean, we arrive at an average
202: luminosity (at 70 pc) of $2.94 \times 10^{30}$ ergs s$^{-1}$ in the
203: same band.
204:
205: The elements strongly required by the fit are helium, nitrogen, oxygen
206: and neon, however, $\chi^2$ additionally improves by 13.35 and 23.9
207: with the inclusion of iron and sulphur, respectively, suggesting these
208: elements may also be present at low abundance. No other elements are
209: required to fit the spectrum. Using this model we determined the
210: abundances relative to the solar values tabulated by Anders \&
211: Grevesse (1989). Expressed as mass fractions we find; $X_{He} = 0.977
212: \pm 0.002$, $X_{N} = 1.7 \pm 0.1 \times 10^{-2}$, $X_{O} = 2.2 \pm 0.3
213: \times 10^{-3}$, $X_{Ne} = 3.7 \pm 0.2 \times 10^{-3}$, $X_{S} = 2.3
214: \pm 0.6 \times 10^{-4}$, and $X_{Fe} = 8 \pm 2 \times 10^{-5}$. In
215: addition, we placed limits ($3\sigma$) on several prominent elements
216: which were not detected; $X_{C} < 2 \times 10^{-3}$, $X_{Mg} < 1.6
217: \times 10^{-4}$ and $X_{Si} < 8 \times 10^{-5}$. Our limits on some of
218: the heavier metals (such as Si, Mg, and Fe) suggest underabundances
219: compared with solar values. This is qualitatively similar to the
220: abundances derived from optical spectroscopy (Marsh et al. 1991).
221:
222: \section{Discussion}
223:
224: The nitrogen lines in the X-ray band provide additional confirmation
225: of the overabundance of nitrogen relative to carbon and oxygen deduced
226: from the optical (Marsh et al. 1991). The high nitrogen abundance
227: likely reflects the operation of CNO-cycle thermonuclear processing in
228: the secondary (Marsh et al. 1991). We find a nitrogen to helium
229: ratio, by number, of $\approx 4.8 \times 10^{-3}$, which is moderately
230: higher than the value derived by Marsh et al. (1991) of $2.7 \times
231: 10^{-3}$ (see their Table 4). Our derived oxygen to helium number
232: ratio is, however, about a factor of 10 higher than that deduced from
233: the optical. Although systematic modelling effects could be present,
234: these results suggest that the X-ray emitting gas may have higher
235: oxygen abundance than the accretion disk (which produces the optical
236: emission lines). If the accretor were an oxygen-rich white dwarf, and
237: some of its matter is mixed with the accreted material in the
238: turbulent boundary layer, then this could perhaps increase the oxygen
239: abundance relative to the donor material.
240:
241: The neon lines provide the first definitive detection of an element
242: heavier than oxygen in GP Com. Our derived neon mass fraction
243: ($\approx 0.0037$) is a factor of 2 larger than the solar
244: value. Moreover, modelling of the optical spectra also suggests a neon
245: enrichment of about 1.3 compared to solar abundances (Marsh et al.
246: 1991). Neon is a by-product of helium burning under a wide range of
247: conditions (see, for example, Clayton 1983; Iben \& Truran 1983),
248: however, the low carbon abundance is likely incompatible with helium
249: burning having ocurred in the secondary. Therefore, this avenue for
250: the neon production appears to be a dead-end. Again, mixing of neon
251: from the accretor (which likely has undergone helium burning) into the
252: donor material in the boundary layer could perhaps increase the neon
253: abundance. It appears that gravitational settling and fractionation
254: of neon (in the form of $^{22}Ne$) are also not relevant, since
255: nitrogen is depleted in the production of this isotope, and we clearly
256: measure a large nitrogen abundance (see, for example, Bildsten \& Hall
257: 2001).
258:
259: Neon enhancements have been suggested for several ultracompact neutron
260: star binaries; 4U 1543-624 and 2S 0918-549 (Juett \& Chakrabarty
261: 2003), and 4U 1627-67 (Schulz et al. 2001). For 4U 1543-624, Juett \&
262: Chakrabarty (2003) deduced Ne/O $\approx 1.5$, by number, from studies
263: of absorption edges in {\it Chandra} high resolution spectra, however,
264: we note that these measurements can be influenced by the unknown
265: ionization structure in the circumbinary material. We find a similar
266: value of Ne/O = 1.4 (by number) for GP Com. Yungelson, Nelemans \& van
267: den Heuvel (2002) suggest fractionation of neon in the white dwarf
268: donor as a possible explanation for the neon enhancements in the
269: neutron star systems. This scenario again requires a helium burning
270: donor and thus seems not to work in the case of GP Com.
271:
272: The X-ray measurements provide a means to estimate the mass accretion
273: rate, $\dot m$, onto the white dwarf primary in GP Com. Assuming
274: matter falls from the inner Lagrange point onto the primary, and that
275: half the gravitational potential energy is dissipated in the accretion
276: disk, the X-ray luminosity, $L_x$ can be approximated as (see, for
277: example, Nelemans, Yungelson \& Portegies Zwart 2004),
278: \begin{equation}
279: L_x = \frac{1}{2}\frac{G M_1 \dot m}{R} \left (1 - \frac{R}{R_{L_1}}
280: \right ) \; ,
281: \end{equation}
282: where $M_1$, $R$, $R_{L_1}$, and $\dot m$ are the primary mass,
283: primary radius, the distance from the inner Lagrange point to the
284: center of the accretor, and the mass accretion rate,
285: respectively. Solving for $\dot m$, and inserting the measured X-ray
286: luminosity for $L_x$ gives,
287: \begin{equation}
288: \dot m = 3.48 \times 10^{-13} \frac{R_9}{(M_1/M_{\odot})} \left ( 1 -
289: \frac{R}{R_{L_1}} \right )^{-1} \; \; M_{\odot} \; {\rm yr}^{-1} \; ,
290: \end{equation}
291: where $R_9$ is the radius of the primary in units of $10^9$ cm, and
292: $R/R_{L_1} \approx 0.2$ for a system like GP Com. For a primary mass,
293: $M_1 = 0.5 M_{\odot}$ (Morales-Rueda et al. 2003) and $R_9 = 1$
294: (Zapolsky \& Salpeter (1969) we obtain $\dot m \approx 8.7 \times
295: 10^{-13} \; M_{\odot} \; {\rm yr}^{-1}$. If the accretor is a massive
296: O/Ne/Mg white dwarf (suggested by the possibility of oxygen and neon
297: mixing in the boundary layer), then the derived accretion rate drops
298: by about a factor of 6.
299:
300: This rate is substantially lower than the critical rate above which
301: one expects an optically thick boundary layer (Pringle \& Savonije
302: 1979; Narayan \& Popham 1993). In this regime the plasma should have a
303: maximum temperature of $\approx 10^8$ K, which is similar to the peak
304: temperature of $6.5$ keV ($7.5 \times 10^7$ K) deduced from our
305: spectral modelling. The presence of {\it narrow} emission lines also
306: supports spectral formation in a boundary layer. From the width of the
307: N VII Ly$\alpha$ line, we obtain an upper limit on the velocity
308: dispersion of the emitting gas of $\approx 250$ km s$^{-1}$
309: ($3\sigma$) (This includes bulk and thermal motions). This indicates
310: an origin close to the white dwarf and not, for example, further out
311: in the accretion disk, which would have much higher Keplerian
312: velocities. If the white dwarf accretor was synchronized with the
313: orbit, then its rotational velocity would be $\approx 23$ km
314: s$^{-1}$. To have a rotational speed equal to our 250 km s$^{-1}$
315: limit, a $10^{4}$ km white dwarf would require a spin period of
316: $\approx 250$ seconds. Indeed, an upper limit to the Doppler
317: (ie. thermal motion) width for the nitrogen line indicates a value
318: close to t250 km s$^{-1}$, which supports a slowly rotating, or
319: synchronized white dwarf. The narrowness of the X-ray lines also
320: appears consistent with the narrow ``central spike'' component of the
321: helium lines (Marsh 1999; Morales-Rueda et al. 2003), which almost
322: certainly originate on the accreting white dwarf.
323:
324: The XMM/Newton measurements provide a detailed look at the physics of
325: the boundary layer in GP Com. Detailed comparisons with theoretical
326: boundary layer models could provide interesting constraints on their
327: temperature, density and rotational profiles. Deeper X-ray spectra
328: would provide better temperature and density constraints from the
329: helium-like triplets, and likely more lines of less abundant species
330: could be detected.
331:
332: \acknowledgements
333:
334: The author thanks Richard Mushotzky, Martin Still, Craig Markwardt,
335: and the referee, Tom Marsh, for many useful comments and suggestions.
336:
337: \centerline{\bf References}
338:
339: \noindent{} Anders E. \& Grevesse N. 1989, Geochimica et Cosmochimica
340: Acta 53, 197.
341:
342: \noindent{} Bildsten, L. \& Hall, D. M. 2001, ApJ, 549, L219.
343:
344: \noindent{} Clayton, D. D. 1983, ``Principles of Stellar Evolution and
345: Nucleosynthesis,'' (Univ. of Chicago Press: Chicago).
346:
347: \noindent{} Eracleous, M., Halpern, J. \& Patterson, J. 1991, ApJ, 382, 290.
348:
349: \noindent{} den Herder, J. W. et al. 2001, A\&A, 365, L7.
350:
351: \noindent{} Gabriel, A. H. \& Jordan, C. 1969, MNRAS, 145, 241.
352:
353: \noindent{} Giclas, H. L., Burnham, R., Jr., \& Thomas, N. G. 1961, Lowell
354: Obs. Bull., 112, 61.
355:
356: \noindent{} Iben, I. \& Truran, J. W. 1978, ApJ, 220, 980.
357:
358: \noindent{} Juett, A. M. \& Chakrabarty, D. 2003, ApJ, 599, 498.
359:
360: \noindent{} Lambert, D. L. \& Slovak, M, H. 1981, PASP, 93, 477.
361:
362: \noindent{} Liedahl, D.A., Osterheld, A.L., and Goldstein, W.H. 1995,
363: ApJL, 438, 115
364:
365: \noindent{} Marsh, T. R., Wood, J. H., Horne, K. \& Lambert, D. 1995, MNRAS,
366: 274, 452.
367:
368: \noindent{} Marsh, T. R., Horne, K. \& Rosen, S. 1991, ApJ, 366, 535.
369:
370: \noindent{} Marsh, T. R. 1999, MNRAS, 304, 443.
371:
372: \noindent{} Mewe, R., Gronenschild, E.H.B.M., \& van den Oord, G.H.J. 1985,
373: A\&AS, 62, 197.
374:
375: \noindent{} Mewe, R., Lemen, J.R., \& van den Oord, G.H.J. 1986,
376: A\&AS, 65, 511.
377:
378: \noindent{} Morales-Rueda, L. et al. 2003, A\&A, 405 249,
379:
380: \noindent{} Narayan, R. \& Popham, R, G. 1993, Nature, 362, 820.
381:
382: \noindent{} Nather, R. E., Robinson, E. L. \& Stover, R. J. 1981, ApJ,
383: 244, 269.
384:
385: \noindent{} Nelemans, G., Yungelson, L.~R., \& Portegies Zwart, S.~F.\
386: 2004, MNRAS, 349, 181.
387:
388: \noindent{} Patterson, J. \& Raymond, J. C. 1985, ApJ, 292, 535.
389:
390: \noindent{} Pollock, A. M. T. et al. 2003, XMM Calibration Document,
391: XMM-SOC-CAL-TN-0030 Issue 2, ``Status of the RGS Calibration,''
392:
393: \noindent{} Popham, R. G. \& Narayan, R. 1995, ApJ, 442, 337.
394:
395: \noindent{} Porquet, D. \& Dubau, J. 2000, A\&AS, 143, 495.
396:
397: \noindent{} Pringle, J. E. \& Savonije, G. J. 1979, MNRAS, 187, 777.
398:
399: \noindent{} Schulz, N. S. et al. 2001, ApJ, 563, 941.
400:
401: \noindent{} Thorstensen, J. R. 2003, AJ, 126, 3017.
402:
403: \noindent{} Ulla, A. 1995, A\&A, 301, 469.
404:
405: \noindent{} van Teeseling, A. \& Verbunt, F. 1994, A\&A, 292, 519.
406:
407: \noindent{} Warner, B. 1995, {\it Cataclysmic Variable Stars}, Cambridge
408: Univ. Press, Cambridge UK.
409:
410: \noindent{} Yungelson, L. R., Nelemans, G. \& van den Heuvel,
411: E. P. J. 2002, A\&A, 388, 546.
412:
413: \noindent{} Zapolsky, H. S. \& Salpeter, E. E. 1969, ApJ, 158, 809.
414:
415: \pagebreak
416:
417: \begin{figure}
418: \begin{center}
419: \includegraphics[width=6in, height=6in]{f1.ps}
420: \end{center}
421: Figure 1: A lightcurve of a portion of the {\it XMM/Newton} EPIC
422: (PN+MOS) observation of GP Com. The binsize is 80 s. Significant
423: variability is evident on a range of timescales. A characteristic
424: error bar is also shown (upper left).
425: \end{figure}
426: \clearpage
427:
428: \begin{figure}
429: \begin{center}
430: \includegraphics[width=6in, height=6in]{f2.ps}
431: \end{center}
432: Figure 2: High resolution spectra of GP Com obtained with the RGS1
433: (bottom) and RGS2 (top) spectrometers in the $7 - 32 \; \AA$ band
434: (solid histogram). Identifications are given for the prominent
435: lines. The best fitting, variable emission measure plasma model
436: (cevmkl in XSPEC) is also shown (solid curve).
437: \end{figure}
438: \clearpage
439:
440: \begin{figure}
441: \begin{center}
442: \includegraphics[width=6in, height=5in]{f3.ps}
443: \end{center}
444: Figure 3: RGS spectra and best fitting Lorentzian line profiles for
445: the N VI (top) and Ne IX (bottom) helium-like triplets.
446: \end{figure}
447: \clearpage
448:
449: \begin{figure}
450: \begin{center}
451: \includegraphics[width=6in, height=6in,angle=-90]{f4.ps}
452: \end{center}
453: Figure 4: Phase-averaged spectrum of GP Com as observed in the
454: EPIC/MOS detectors (MOS1:triangles, MOS2:crosses). Shown are the data
455: and best fitting {\it cevmkl} model versus wavelength using XSPEC. The
456: fit includes contributions from helium, nitrogen, oxygen, neon,
457: sulphur and iron.
458: \end{figure}
459: \clearpage
460:
461: \begin{table*}
462: \begin{center}{Table 1: Emission Line Properties for GP Com}
463: \begin{tabular}{cccccrcl} \\
464: \tableline
465: \tableline
466: Species & & $\lambda$ ($\AA$) & & & & Flux & \\
467: \tableline
468: & r & i & f & & & ($10^{-4}$ cm$^{-2}$ s$^{-1}$) & \\
469: \tableline
470: N VII & & 24.781 & & & & 1.86 & \\
471: N VII & & 20.910 & & & & 0.28 & \\
472: N VI & 28.792 & 29.075 & 29.531 & & 0.34 & 0.22 & $0.07^1$ \\
473: Ne X & & 12.134 & & & & 0.54 & \\
474: Ne X & & 10.239 & & & & 0.16 & \\
475: Ne IX & 13.447 & 13.550 & 13.697 & & 0.20 & 0.11 & 0.09 \\
476: \tableline
477: \end{tabular}
478: \end{center}
479: $^1$ Not detected. $1\sigma$ upper limit.
480: \end{table*}
481:
482:
483: \clearpage
484:
485:
486:
487: %%%%%%%%%%%%%%%%%%%%%%%%%%%
488: %%%%% End of document %%%%%
489: %%%%%%%%%%%%%%%%%%%%%%%%%%%
490:
491: \end{document}
492: