1: \documentclass{emulateapj}
2: \usepackage{apjfonts}
3:
4: \newcommand{\ddt}[1]{\frac{d #1}{d t}}
5: \newcommand{\sh}{\Sigma_{\mathrm{H}}}
6: \newcommand{\she}{\Sigma_{\mathrm{He}}}
7: \newcommand{\estarh}{E^{*}_{\mathrm{H}}}
8: \newcommand{\estarhe}{E^{*}_{\mathrm{He}}}
9: \newcommand{\estarb}{E^{*}_{\mathrm{rp}}}
10: \newcommand{\epsh}{\epsilon_{\mathrm{H}}}
11: \newcommand{\epsb}{\epsilon_{\mathrm{rp}}}
12: \newcommand{\epshe}{\epsilon_{\mathrm{He}}}
13: \newcommand{\xout}{X_{0}}
14: \newcommand{\yout}{Y_{0}}
15: \newcommand{\zout}{Z_{0}}
16: \newcommand{\lacc}{l_{\mathrm{acc}}}
17: \newcommand{\llacc}{\log(\lacc)}
18: \newcommand{\frp}{f_{\mathrm{rp}}}
19:
20: \begin{document}
21:
22: \title{On the Physics of Type I X-ray Bursts on Accreting Neutron
23: Stars at High Accretion Rates}
24: \shorttitle{Physics of X-ray Bursts}
25: \author{Randall L.\ Cooper and Ramesh Narayan}
26: \affil{Harvard-Smithsonian Center for Astrophysics, 60 Garden Street,
27: Cambridge, MA 02138}
28:
29: \email{rcooper@cfa.harvard.edu, rnarayan@cfa.harvard.edu}
30:
31: \begin{abstract}
32:
33: We investigate the effect of the hot CNO cycle breakout reaction
34: $^{15}$O($\alpha$,$\gamma$)$^{19}$Ne on the occurrence of type I X-ray
35: bursts on accreting neutron stars. For $\frp \lesssim 0.1$, where
36: $\frp$ is a dimensionless factor by which we multiply the
37: $^{15}$O($\alpha$,$\gamma$)$^{19}$Ne reaction rate of \citet{CF88},
38: our model predicts that bursts should occur only for accretion rates
39: $\dot{M}$ below a critical value $\approx 0.3 \dot{M}_{\mathrm{Edd}}$.
40: This agrees with observations. For larger values of $\frp$, including
41: the standard choice $\frp = 1$, the model switches to a new regime in
42: which bursts occur all the way up to $\dot{M} \approx
43: \dot{M}_{\mathrm{Edd}}$. Since the latter regime disagrees with
44: observations, we suggest that the true
45: $^{15}$O($\alpha$,$\gamma$)$^{19}$Ne reaction rate is lower than
46: usually assumed.
47:
48: \end{abstract}
49:
50: \keywords{dense matter --- nuclear reactions, nucleosynthesis,
51: abundances --- stars: neutron --- X-rays: binaries --- X-rays: bursts}
52:
53: \section{Introduction}
54:
55: Type I X-ray bursts are thermonuclear explosions that occur on
56: accreting neutron stars in low-mass X-ray binaries.
57: They are triggered by thermally unstable H or He burning
58: near the stellar surface \citep[for reviews, see][]{C04,SB03}. For
59: systems with accretion rates $\dot{M} \lesssim 0.1
60: \dot{M}_{\mathrm{Edd}}$, where $\dot{M}_{\mathrm{Edd}}$ denotes the
61: mass accretion rate at which the accretion luminosity is equal to the
62: Eddington limit, the basic physics of the burst onset is well
63: understood to be that of the thin shell thermal instability
64: \citep{SH65,HvH75}, and theoretical models have been rather successful
65: at reproducing the gross characteristics of burst observations
66: in this regime \citep[e.g.,][]{FHM81,FL87,CB00,NH03}.
67:
68: This {\em Letter} addresses a longstanding problem that afflicts
69: nearly all burst models when the accretion rate $\dot{M} \gtrsim 0.1
70: \dot{M}_{\mathrm{Edd}}$. Both simple one-zone burst models
71: \citep{FHM81,P83,B98,HCW05} and sophisticated time-dependent
72: multi-zone models \citep{AJ82,TWL96,FHLT03,HCW05} predict that bursts
73: should occur for all $\dot{M}$ up to $\approx \dot{M}_{\mathrm{Edd}}$.
74: Observations, however, indicate that bursts do not occur for $\dot{M}
75: \gtrsim 0.3 \dot{M}_{\mathrm{Edd}}$
76: \citep{vPCLJ79,vPPL88,Cetal03,RLCN06}. Furthermore, \citet{vPPL88}
77: found that, for $0.1 \lesssim \dot{M}/\dot{M}_{\mathrm{Edd}} \lesssim
78: 0.3$, a significant fraction of the accreted plasma burns stably
79: between consecutive bursts, leading to large values $\gtrsim 1000$ of
80: the parameter $\alpha$, the accretion energy released between
81: successive bursts divided by the nuclear energy released during a
82: burst. Most theoretical models, on the other hand, predict that
83: nearly all of the accreted matter burns unstably during bursts, giving
84: a nearly constant $\alpha < 100$ at all $\dot{M}$.
85:
86: \citet[][hereafter NH03]{NH03} developed a global linear stability
87: analysis of the accreted plasma on the surface of a neutron star.
88: They discovered a new regime of unstable nuclear burning for $0.1
89: \lesssim \dot{M}/\dot{M}_{\mathrm{Edd}} \lesssim 0.3$ that they
90: referred to as ``delayed mixed bursts.'' Their model reproduced both
91: the occurrence of considerable stable burning preceding a burst
92: (leading to values of $\alpha \gtrsim 1000$) and the absence of bursts
93: for $\dot{M} \gtrsim 0.3 \dot{M}_{\mathrm{Edd}}$, in agreement with
94: observations. However, the complexity of their model made it
95: difficult to understand the basic physics behind delayed mixed bursts
96: and to identify the reasons why their results differed from those of
97: other theoretical models at high accretion rates. To remedy this,
98: \citet[][hereafter CN06]{CN06} constructed a simple two-zone model
99: that helped elucidate the physics of delayed mixed bursts. They
100: showed that the competition between nuclear heating due to
101: triple-$\alpha$ reactions and hot CNO cycle H burning on the
102: one hand, and radiative cooling via photon diffusion and emission on
103: the other hand, drives an overstability that eventually triggers a
104: thin-shell thermal instability and hence a delayed mixed burst. They
105: asserted that H burning via the temperature-independent hot CNO
106: cycle, augmented by the extra seed nuclei produced from stable He
107: burning, significantly lowers the temperature sensitivity of the total
108: nuclear energy generation rate to such an extent as to suppress the
109: thin-shell thermal instability for $\dot{M} \gtrsim 0.3
110: \dot{M}_{\mathrm{Edd}}$. Therefore, H burning via the hot CNO
111: cycle is ultimately responsible for the lower critical $\dot{M}$ above
112: which bursts do not occur in nature.
113:
114: The above argument perhaps explains why one-zone models fail, since
115: the models generally focus only on He burning and make large
116: approximations with respect to H burning. But why do detailed
117: time-dependent multi-zone models with large reaction networks also
118: perform poorly in relation to observations? These models are much
119: more sophisticated than the models of NH03 and CN06 and consequently
120: ought to perform better, whereas in fact the latter models agree much
121: better with observations. CN06 hypothesized that the time-dependent
122: multi-zone burst models may have used too large a rate for the
123: experimentally poorly-constrained hot CNO cycle breakout reaction
124: $^{15}$O($\alpha$,$\gamma$)$^{19}$Ne
125: \citep{W69,WW81,LWFG86,WGS99,FGWD06}. We test this hypothesis. We
126: begin in \S \ref{themodel} with a description of the model, and we
127: present the results of the model in \S \ref{results}. We discuss the
128: results in \S \ref{discussion}, and we conclude in \S
129: \ref{conclusions}.
130:
131:
132: \section{The Model}\label{themodel}
133:
134: We use the general-relativistic global linear stability analysis of
135: \citet{CN05}, which is an expanded and improved version of the model
136: of NH03, to determine the stability of nuclear burning on accreting
137: neutron stars. We assume that matter accretes spherically onto a
138: neutron star of gravitational mass $M =1.4 M_{\odot}$ and areal radius
139: $R = 10.4$ km at a rate $\dot{M}$, where $\dot{M}$ is the rest mass
140: accreted per unit time as measured by an observer at infinity. We set
141: the composition of the accreted matter to be that of the Sun, such
142: that at the neutron star surface the H mass fraction
143: $X_{\mathrm{out}} = 0.7$, He mass fraction $Y_{\mathrm{out}} =
144: 0.28$, CNO mass fraction $Z_{\mathrm{CNO, out}} = 0.016$, and heavy
145: element fraction $Z_{\mathrm{out}} = 0.004$, where $Z$ refers to all
146: metals other than CNO. In this section, we describe the modifications
147: to the theoretical model.
148:
149: For temperatures $T \gtrsim 8 \times 10^{7}$ K, H burns
150: predominantly via the hot CNO cycle
151: $^{12}$C($p$,$\gamma$)$^{13}$N($p$,$\gamma$)$^{14}$O($\beta^{+}
152: \nu$)$^{14}$N($p$,$\gamma$)$^{15}$O($\beta^{+}
153: \nu$)$^{15}$N($p$,$\alpha$)$^{12}$C, the rate of which is determined
154: by the slow $\beta$-decays of $^{14}$O($t_{1/2}=70.6$ s) and
155: $^{15}$O($t_{1/2}=122$ s) \citep{HF65}. During the hot CNO cycle,
156: essentially all of the CNO ions are converted to $^{14}$O and
157: $^{15}$O. For the range of accretion rates we consider in this
158: investigation, the primary breakout reaction from the hot CNO cycle
159: into the rp-process of \citet{WW81} is
160: $^{15}$O($\alpha$,$\gamma$)$^{19}$Ne \citep{GWT95,Hetal96,SBCW99}. If
161: this breakout reaction occurs, the $^{19}$Ne($t_{1/2}=17.2$ s)
162: produced in this reaction can $\beta$-decay and return to the hot CNO
163: cycle via $^{19}$Ne($\beta^{+} \nu$)$^{19}$F($p$,
164: $\alpha$)$^{16}$O($p$, $\gamma$)$^{17}$F($p$,
165: $\gamma$)$^{18}$Ne($\beta^{+} \nu$)$^{18}$F($p$, $\alpha$)$^{15}$O,
166: the effect of which is only to expedite the hot CNO cycle by a factor
167: $\lesssim 1.6$. We ignore this effect in our model. If $^{19}$Ne
168: captures a proton, however, the resulting $^{20}$Na ion can never
169: return to the hot CNO cycle \citep{WW81}. The fate of a $^{19}$Ne ion
170: thus depends on whether or not it captures a proton before it can
171: $\beta$-decay. Therefore, the probability that a $^{19}$Ne ion is
172: removed from the hot CNO cycle is
173: \begin{equation}
174: R = \frac{\rho X \lambda_{p \gamma}(^{19}\mathrm{Ne})}{\rho X
175: \lambda_{p \gamma}(^{19}\mathrm{Ne}) +
176: \lambda_{\beta^{+}}(^{19}\mathrm{Ne})},
177: \end{equation}
178: where $\rho$ is the density, $X$ is the H mass fraction,
179: $\lambda_{p \gamma}(^{19}\mathrm{Ne})$ is the
180: $^{19}$Ne($p$,$\gamma$)$^{20}$Na reaction rate from \citet[][hereafter
181: CF88]{CF88} with electron screening from \citet{DGC73}, and
182: $\lambda_{\beta^{+}}(^{19}\mathrm{Ne}) = \ln 2 / t_{1/2}$ is the
183: $^{19}$Ne $\beta$-decay rate. If $^{19}$Ne captures a proton, further
184: proton captures will ensue until the nuclear flow reaches the first
185: waiting point of the rp-process, which we assume to be $^{24}$Si
186: \citep{WSC98}. The $^{15}$O($\alpha$,$\gamma$)$^{19}$Ne reaction is
187: the slowest reaction of this sequence, and so its rate governs the
188: total reaction rate of this flow. Therefore, we follow \citet{WGS99}
189: and approximate the total nuclear energy generation rate of the hot
190: CNO cycle breakout reaction sequence by
191: \begin{equation}
192: \epsb = 24 \estarb \left(\frac{Y}{4}\right)
193: \left(\frac{Z_{\mathrm{CNO}}}{15}\right) \rho R \lambda_{\alpha
194: \gamma}(^{15}\mathrm{O}) \times \frp,
195: \end{equation}
196: where $\estarb = Q_{\mathrm{rp}}/24 m_{\mathrm{p}} = 1.2 \times
197: 10^{18}$ $\mathrm{ergs\,g}^{-1}$, $Y$ is the He mass fraction,
198: $Z_{\mathrm{CNO}}$ is the mass fraction of CNO elements, and
199: $\lambda_{\alpha \gamma}(^{15}\mathrm{O})$ is the
200: $^{15}$O($\alpha$,$\gamma$)$^{19}$Ne reaction rate from CF88,
201: which is based on the rate derived by \citet{LWFG86}. For the
202: temperatures $T \lesssim 6 \times 10^{8}$ K we consider in this work,
203: the resonant contribution of the 4.03 MeV state in $^{19}$Ne dominates
204: the $^{15}$O($\alpha$,$\gamma$)$^{19}$Ne reaction rate \citep{LWFG86}.
205: The reaction rate of this contribution is directly
206: proportional to the $\alpha$ width $\Gamma_{\alpha}$. However,
207: $\Gamma_{\alpha}$ is very difficult to measure experimentally, and so
208: the true $\lambda_{\alpha \gamma}(^{15}\mathrm{O})$ is highly
209: uncertain \citep[for recent progress on experimental constraints, see,
210: e.g.,][]{Tetal05,Ketal06}. To account for this uncertainty, we
211: multiply the CF88 rate by a dimensionless free parameter
212: $\frp$ such that $\frp=1$ corresponds the the usual rate of CF88.
213: We omit the electron screening contribution to the
214: reaction rate since the enhancement factor itself depends on the
215: relatively unconstrained resonance strength \citep{SvH69,M77}.
216: Instead, we simply absorb this contribution into $\frp$. Equations
217: (6), (8), (9), and (10) of \citet{CN05} thus become
218: \begin{equation}
219: e^{-2 \Phi/c^{2}} \frac{{\partial}}{{\partial}\Sigma} \left (\frac{F
220: r^{2}}{R^{2}} e^{2 \Phi/c^{2}} \right ) = T \ddt{s} - (\epsh +
221: \epshe + \epsb + \epsilon_{\mathrm{N}} -
222: \epsilon_{\mathrm{\nu}}),
223: \end{equation}
224: \begin{equation}\label{dxdteqn}
225: \ddt{X} = - \frac{\epsh}{\estarh} - \frac{5}{24}\frac{\epsb}{\estarb},
226: \end{equation}
227: \begin{equation}
228: \ddt{Y} = \frac{\epsh}{\estarh} - \frac{\epshe}{\estarhe} -
229: \frac{4}{24} \frac{\epsb}{\estarb},
230: \end{equation}
231: \begin{equation}
232: \ddt{Z_{\mathrm{CNO}}} = \frac{\epshe}{\estarhe} - \frac{15}{24}
233: \frac{\epsb}{\estarb}.
234: \end{equation}
235: See \citet{CN05} for the definitions of the various symbols. Note
236: that the factor of $5$ in equation (\ref{dxdteqn}) accounts for the
237: five proton captures needed to reach $^{24}$Si. For a substantial
238: portion of the accretion rates we consider in this work, the nuclear
239: flow due to stable burning extends well beyond the relatively small
240: number of isotopes included in our limited network
241: \citep[e.g.,][]{SBCW99,FBLTW05}. However, the contributions from
242: H and He burning dominate the total nuclear energy
243: generation rate and set the thermal profile of the accreted layer, and
244: we calculate these contributions to high accuracy. Consequently, the
245: thermal profiles resulting from our network compare very well with
246: those of \citet{SBCW99} for the range of accretion rates in which our
247: calculations overlap. Furthermore, we are interested only in the
248: physics of the burst onset, for which only hot CNO cycle,
249: triple-$\alpha$, and breakout reactions are important. Thus, our
250: limited network should be adequate for our purposes.
251:
252: \section{Results}\label{results}
253:
254: Figure \ref{alphagraph} shows the values of the dimensionless quantity
255: $\alpha$ as a function of the Eddington-scaled accretion rate $\lacc
256: \equiv \dot{M} / \dot{M}_{\mathrm{Edd}}$ for five different values of
257: $\frp$, where $\alpha$ is defined as the accretion energy released
258: between successive bursts divided by the nuclear energy released
259: during a burst. We follow NH03 and call a type I X-ray burst
260: ``prompt'' if $\alpha \lesssim 100$ and ``delayed'' if $\alpha \gg
261: 100$. The case $\frp = 0$ omits the breakout reaction sequence
262: entirely, and so this calculation is representative of the results of
263: NH03 and CN06. The models gives prompt bursts for $\lacc \lesssim
264: 0.15$ and delayed bursts over the range $\lacc \approx 0.15$-$0.3$.
265: For low values of $\frp \lesssim 0.1$, the effects of breakout
266: reactions from the hot CNO cycle on the occurrence of type I X-ray
267: bursts are minor, and the models of NH03 and CN06 continue to provide
268: an adequate description of the nuclear physics that precedes bursts.
269: However, the situation changes for $\frp \gtrsim 0.1$. The critical
270: $\lacc$ above which delayed mixed bursts cease now significantly
271: decreases, and the range of accretion rates in which delayed mixed
272: bursts occur becomes notably truncated as well. More importantly, a
273: new burst regime appears at accretion rates close to
274: $\dot{M}_{\mathrm{Edd}}$, where the bursts have $\alpha < 100$ and are
275: hence prompt. This regime of bursts near the Eddington limit was
276: hypothesized by CN06, but it was not present in either the NH03 or
277: CN06 models due to the exclusion of hot CNO cycle breakout reactions.
278: The usual delayed mixed burst regime and this new prompt mixed burst
279: regime are separated by a short range of accretion rates in which
280: bursts do not occur. These results obtained with larger values of
281: $\frp$ agree rather well with previous theoretical models
282: \citep{FHM81,AJ82,P83,T85,TWL96,B98,FHLT03,HCW05}, most of which
283: predicted bursts to occur up to roughly the Eddington limit. Figure
284: \ref{alphagraph} thus explains the origin of the differences between
285: those models and the models of NH03 and CN06. If the breakout
286: reaction rate is large, say $\frp \gtrsim 0.1$, the results agree with
287: most published models (which include breakout reactions in full
288: strength), and if the rate is small, $\frp \lesssim 0.1$, the results
289: are similar to those obtained by NH03 and CN06 (who effectively set
290: $\frp = 0$). We reiterate that the latter results agree much better
291: with observations.
292:
293: \begin{figure}
294: \epsscale{1.25}
295: \plotone{f1.eps}
296: \caption{Plot shows $\alpha$-values of type I X-ray bursts as a
297: function of the Eddington-scaled accretion rate $\lacc \equiv
298: \dot{M}/\dot{M}_{\mathrm{Edd}}$ for five different values of $\frp$,
299: the dimensionless factor by which we multiply the
300: $^{15}$O($\alpha$,$\gamma$)$^{19}$Ne reaction rate from CF88.
301: }
302: \label{alphagraph}
303: \end{figure}
304:
305:
306: \section{Discussion}\label{discussion}
307:
308: For $\lacc \lesssim 0.1$, the temperatures reached during steady-state
309: nuclear burning in the accreted layer are too low for significant
310: leakage out of the hot CNO cycle via breakout reactions, and so these
311: reactions have a negligible effect on the onset of type I X-ray
312: bursts, regardless of the true $^{15}$O($\alpha$,$\gamma$)$^{19}$Ne
313: cross section. However, breakout reactions should affect the nuclear
314: flow during the burst itself, and hence they could influence the burst
315: lightcurve and possibly also the onset of subsequent bursts
316: \citep{FGWD06}. CN06 showed that the delayed mixed bursts of NH03
317: occur when $\sh$, the column depth at which H is depleted via
318: stable nuclear burning as measured from the stellar surface, is less
319: than but close to $\she$, the column depth at which He is depleted
320: via stable nuclear burning. The hot CNO cycle breakout reaction
321: sequence (i) eliminates seed nuclei from the hot CNO cycle which slows
322: H burning, thereby increasing $\sh$, and (ii) provides an
323: additional pathway by which He may burn, thereby decreasing
324: $\she$. Consequently, one expects the regime of delayed mixed bursts
325: to occur at lower $\lacc$ if breakout reactions are included.
326: Furthermore, CN06 showed that the interplay between H burning
327: via the hot CNO cycle and He burning via triple-$\alpha$ reactions
328: is integral to generating the oscillations that precede delayed mixed
329: bursts. Breakout reactions diminish this interplay by eliminating the
330: hot CNO cycle seed nuclei and should therefore reduce the range of accretion
331: rates over which delayed mixed bursts occur. Figure \ref{alphagraph}
332: illustrates both of these effects (compare $\frp=1$ with $\frp=0$).
333:
334: For accretion rates above those at which delayed mixed bursts occur,
335: NH03 found that nuclear burning is always stable and therefore bursts
336: do not occur, in accord with observations
337: \citep{vPCLJ79,vPPL88,Cetal03,RLCN06}. According to their model, for
338: $0.3 \lesssim \lacc \lesssim 1$, steady-state H burning via the
339: hot CNO cycle increases the effective radiative cooling rate and
340: thereby suppresses a He-triggered thin-shell thermal instability
341: (CN06). However, their model did not include breakout reactions. If
342: the $^{15}$O($\alpha$,$\gamma$)$^{19}$Ne breakout reaction rate is
343: significant in this regime, this reaction will suppress hot CNO cycle
344: H burning, and thus He burning via triple-$\alpha$
345: reactions will govern the total reaction rate, since the
346: triple-$\alpha$ reaction rate will be the slowest rate in the nuclear
347: flow at these high $\lacc$. The notion that the triple-$\alpha$
348: reaction rate is the slowest rate in the nuclear flow is precisely
349: what one-zone type I X-ray burst ignition models assume
350: \citep{FHM81,P83,B98,CB00,HCW05}, and we suggest that this is the
351: reason why the results of one-zone and time-dependent multi-zone
352: models agree so well. All of these models predict that nuclear
353: burning is thermally stable for temperatures $T \gtrsim 5 \times
354: 10^{8}$ K, which are reached only at accretion rates $\lacc \gtrsim 1$
355: \citep[e.g.,][]{SBCW99}. Figure \ref{alphagraph} illustrates that,
356: for relatively large values of $\frp$, there is a regime of prompt
357: bursts at accretion rates greater than the accretion rates at which
358: delayed mixed bursts occur, and the critical accretion rate above
359: which these prompt mixed bursts cease is roughly
360: $\dot{M}_{\mathrm{Edd}}$, in very good agreement with these other
361: models. Figure \ref{alphagraph} illustrates also that nuclear burning
362: is stable for the small range of $\lacc$ between these two regimes.
363: These accretion rates are high enough to suppress delayed mixed
364: bursts, but they are not high enough to cause a sufficient leakage out
365: of hot CNO cycle H burning and trigger a thermal instability.
366:
367: The model we present in this work suggests that, if the true
368: $^{15}$O($\alpha$,$\gamma$)$^{19}$Ne cross section is greater than
369: approximately $0.1$ of the CF88 rate, type I X-ray bursts should occur
370: in systems with accretion rates near the Eddington limit.
371: Observations indicate that this is not the case, since low-mass X-ray
372: binaries with $\lacc \gtrsim 0.3$ generally do not exhibit bursts.
373: This suggests that the true $^{15}$O($\alpha$,$\gamma$)$^{19}$Ne
374: reaction rate is less than that proposed by CF88. This conclusion is
375: complementary to that of \citet{FGWD06}, who were the first to propose
376: that the occurrence of type I X-ray bursts is sensitive to the
377: strength of the $^{15}$O($\alpha$,$\gamma$)$^{19}$Ne reaction rate,
378: and who found that the existence of bursts in systems with $\lacc
379: \approx 0.1$ suggests a lower bound on this rate. This lower bound
380: corresponds to $\frp \approx 0.05$ in our notation. We also note that
381: a relatively low $^{15}$O($\alpha$,$\gamma$)$^{19}$Ne reaction rate is
382: consistent with the existence of carbon-triggered superbursts, since a
383: low rate would increase the carbon yield resulting from stable nuclear
384: burning \citep{SBCW99,CMSN06,FGWD06}. The systems GX 17+2 and Cyg X-2
385: are exceptions to this empirical rule, however, for they exhibit
386: bursts at accretion rates near the Eddington limit
387: \citep[e.g.,][]{KG84,THKMM84,KHvdKLM02}. However, it is possible that
388: they show bursts for other reasons, e.g., they harbor mass donor stars
389: with H-deficient envelopes \citep{CMSN06}.
390:
391: Although the resonant contribution of the 4.03 MeV state dominates the
392: $^{15}$O($\alpha$,$\gamma$)$^{19}$Ne reaction rate for the
393: temperatures $T \lesssim 5 \times 10^{8}$ K at which helium burning
394: can possibly trigger a type I X-ray burst, other resonances such as
395: the 4.38 MeV state contribute as well. However, these resonances
396: contribute much less than a tenth of the 4.03 MeV contribution
397: \citep[e.g.,][]{Detal03}, whereas a reduction of the total
398: $^{15}$O($\alpha$,$\gamma$)$^{19}$Ne reaction rate by a factor of only
399: $\approx 10$ is needed to stabilize nuclear burning. Therefore, it is
400: the 4.03 MeV $\alpha$ width alone that determines the stability of
401: nuclear burning at high accretion rates.
402:
403: \section{Conclusions}\label{conclusions}
404:
405: Using the global linear stability analysis of \citet{CN05}, which is
406: an expanded and improved version of the model of NH03, we have
407: investigated the effects of the hot CNO cycle breakout reaction
408: $^{15}$O($\alpha$,$\gamma$)$^{19}$Ne on the occurrence of type I X-ray
409: bursts on accreting neutron stars at high accretion rates. For low
410: values of $\frp \lesssim 0.1$, where $\frp$ is a dimensionless factor
411: by which we multiply the $^{15}$O($\alpha$,$\gamma$)$^{19}$Ne reaction
412: rate of CF88, the hot CNO cycle breakout reaction slightly
413: lowers the critical accretion rate above which delayed mixed bursts
414: occur, but otherwise the breakout reaction has little effect on the
415: burst onset. The predictions of these models are in good agreement
416: with observations. For $\frp \gtrsim 0.1$, a new regime of prompt
417: mixed bursts appears at accretion rates above the rates at which
418: delayed mixed bursts occur, and the bursts survive up to roughly the
419: Eddington limit. The existence of this prompt mixed burst regime up
420: to the Eddington limit is consistent with nearly all previous
421: theoretical models. Our results support the hypothesis of CN06 that
422: the discrepancies between the results of the global linear stability
423: analysis of NH03 and the results of time-dependent multi-zone models
424: with large reaction networks may be caused by the latter models
425: assuming too large a strength for the
426: $^{15}$O($\alpha$,$\gamma$)$^{19}$Ne reaction rate at temperatures $T
427: \lesssim 6 \times 10^{8}$ K. The fact that observations agree much
428: better with the results of NH03 and CN06 implies that the true
429: reaction rate is lower than the rate assumed in these multi-zone
430: models. Specifically, we suggest that the true $\alpha$ width
431: $\Gamma_{\alpha}$ of the 4.03 MeV state in $^{19}$Ne is lower than the
432: $\alpha$ width proposed by \citet{LWFG86}. Calculations using
433: multi-zone models such as those of \citet{Wetal04} and \citet{FGWD06},
434: but with the $^{15}$O($\alpha$,$\gamma$)$^{19}$Ne reaction rate
435: lowered in strength, need to be carried out to either confirm or
436: refute this suggestion.
437:
438:
439:
440: \acknowledgments
441:
442: We thank the referee for several insightful comments and suggestions.
443: This work was supported by NASA grant NNG04GL38G.
444:
445: \begin{thebibliography}{48}
446: \expandafter\ifx\csname natexlab\endcsname\relax\def\natexlab#1{#1}\fi
447:
448: \bibitem[{{Ayasli} \& {Joss}(1982)}]{AJ82}
449: {Ayasli}, S. \& {Joss}, P.~C. 1982, \apj, 256, 637
450:
451: \bibitem[{{Bildsten}(1998)}]{B98}
452: {Bildsten}, L. 1998, in NATO ASIC Proc. 515: The Many Faces of Neutron Stars.,
453: 419
454:
455: \bibitem[{{Caughlan} \& {Fowler}(1988)}]{CF88}
456: {Caughlan}, G.~R. \& {Fowler}, W.~A. 1988, Atomic Data and Nuclear Data Tables,
457: 40, 283 (CF88)
458:
459: \bibitem[{{Cooper} {et~al.}(2006){Cooper}, {Mukhopadhyay}, {Steeghs}, \&
460: {Narayan}}]{CMSN06}
461: {Cooper}, R.~L., {Mukhopadhyay}, B., {Steeghs}, D., \& {Narayan}, R. 2006,
462: \apj, 642, 443
463:
464: \bibitem[{{Cooper} \& {Narayan}(2005)}]{CN05}
465: {Cooper}, R.~L. \& {Narayan}, R. 2005, \apj, 629, 422
466:
467: \bibitem[{{Cooper} \& {Narayan}(2006)}]{CN06}
468: ---. 2006, accepted by ApJ (astro-ph/0605001) (CN06)
469:
470: \bibitem[{{Cornelisse} {et~al.}(2003){Cornelisse}, {in't Zand}, {Verbunt},
471: {Kuulkers}, {Heise}, {den Hartog}, {Cocchi}, {Natalucci}, {Bazzano}, \&
472: {Ubertini}}]{Cetal03}
473: {Cornelisse}, R., {in't Zand}, J.~J.~M., {Verbunt}, F., {Kuulkers}, E.,
474: {Heise}, J., {den Hartog}, P.~R., {Cocchi}, M., {Natalucci}, L., {Bazzano},
475: A., \& {Ubertini}, P. 2003, \aap, 405, 1033
476:
477: \bibitem[{{Cumming}(2004)}]{C04}
478: {Cumming}, A. 2004, Nuclear Physics B Proceedings Supplements, 132, 435
479:
480: \bibitem[{{Cumming} \& {Bildsten}(2000)}]{CB00}
481: {Cumming}, A. \& {Bildsten}, L. 2000, \apj, 544, 453
482:
483: \bibitem[Davids et al.(2003)]{Detal03} Davids, B., et al.\
484: 2003, \prc, 67, 065808
485:
486: \bibitem[{{Dewitt} {et~al.}(1973){Dewitt}, {Graboske}, \& {Cooper}}]{DGC73}
487: {Dewitt}, H.~E., {Graboske}, H.~C., \& {Cooper}, M.~S. 1973, \apj, 181, 439
488:
489: \bibitem[{{Fisker} {et~al.}(2005){Fisker}, {Brown}, {Liebend{\" o}rfer},
490: {Thielemann}, \& {Wiescher}}]{FBLTW05}
491: {Fisker}, J.~L., {Brown}, E.~F., {Liebend{\" o}rfer}, M., {Thielemann}, F.-K.,
492: \& {Wiescher}, M. 2005, Nuclear Physics A, 752, 604
493:
494: \bibitem[{{Fisker} {et~al.}(2006){Fisker}, {Gorres}, {Wiescher}, \&
495: {Davids}}]{FGWD06}
496: {Fisker}, J.~L., {Gorres}, J., {Wiescher}, M., \& {Davids}, B. 2006, accepted
497: by ApJ (astro-ph/0410561)
498:
499: \bibitem[{{Fisker} {et~al.}(2003){Fisker}, {Hix}, {Liebend{\"o}rfer}, \&
500: {Thielemann}}]{FHLT03}
501: {Fisker}, J.~L., {Hix}, W.~R., {Liebend{\"o}rfer}, M., \& {Thielemann}, F.-K.
502: 2003, Nuclear Physics A, 718, 614
503:
504: \bibitem[{{Fujimoto} {et~al.}(1981){Fujimoto}, {Hanawa}, \& {Miyaji}}]{FHM81}
505: {Fujimoto}, M.~Y., {Hanawa}, T., \& {Miyaji}, S. 1981, \apj, 247, 267
506:
507: \bibitem[{{Fushiki} \& {Lamb}(1987)}]{FL87}
508: {Fushiki}, I. \& {Lamb}, D.~Q. 1987, \apjl, 323, L55
509:
510: \bibitem[{{G{\"o}rres} {et~al.}(1995){G{\"o}rres}, {Wiescher}, \&
511: {Thielemann}}]{GWT95}
512: {G{\"o}rres}, J., {Wiescher}, M., \& {Thielemann}, F.-K. 1995, \prc, 51, 392
513:
514: \bibitem[Hahn et al.(1996)]{Hetal96} Hahn, K.~I., et al.\ 1996,
515: \prc, 54, 1999
516:
517: \bibitem[{{Hansen} \& {van Horn}(1975)}]{HvH75}
518: {Hansen}, C.~J. \& {van Horn}, H.~M. 1975, \apj, 195, 735
519:
520: \bibitem[{{Heger} {et~al.}(2005){Heger}, {Cumming}, \& {Woosley}}]{HCW05}
521: {Heger}, A., {Cumming}, A., \& {Woosley}, S.~E. 2005, ApJ, submitted
522: (astro-ph/0511292)
523:
524: \bibitem[{{Hoyle} \& {Fowler}(1965)}]{HF65}
525: {Hoyle}, F. \& {Fowler}, W.~A. 1965, in Quasi-Stellar Sources and Gravitational
526: Collapse, 17
527:
528: \bibitem[{{Kahn} \& {Grindlay}(1984)}]{KG84}
529: {Kahn}, S.~M. \& {Grindlay}, J.~E. 1984, \apj, 281, 826
530:
531: \bibitem[Kanungo et al.(2006)]{Ketal06} Kanungo, R., et al.\
532: 2006, \prc, submitted (nucl-ex/0605033)
533:
534: \bibitem[{{Kuulkers} {et~al.}(2002){Kuulkers}, {Homan}, {van der Klis},
535: {Lewin}, \& {M{\' e}ndez}}]{KHvdKLM02}
536: {Kuulkers}, E., {Homan}, J., {van der Klis}, M., {Lewin}, W.~H.~G., \& {M{\'
537: e}ndez}, M. 2002, \aap, 382, 947
538:
539: \bibitem[{{Langanke} {et~al.}(1986){Langanke}, {Wiescher}, {Fowler}, \&
540: {Gorres}}]{LWFG86}
541: {Langanke}, K., {Wiescher}, M., {Fowler}, W.~A., \& {Gorres}, J. 1986, \apj,
542: 301, 629
543:
544: \bibitem[{{Mitler}(1977)}]{M77}
545: {Mitler}, H.~E. 1977, \apj, 212, 513
546:
547: \bibitem[{{Narayan} \& {Heyl}(2003)}]{NH03}
548: {Narayan}, R. \& {Heyl}, J.~S. 2003, \apj, 599, 419 (NH03)
549:
550: \bibitem[{{Paczy{\'n}ski}(1983)}]{P83}
551: {Paczy{\'n}ski}, B. 1983, \apj, 264, 282
552:
553: \bibitem[{{Remillard} {et~al.}(2006){Remillard}, {Lin}, {Cooper}, \&
554: {Narayan}}]{RLCN06}
555: {Remillard}, R.~A., {Lin}, D., {Cooper}, R.~L., \& {Narayan}, R. 2006, \apj, 646, 407
556:
557: \bibitem[{{Salpeter} \& {van Horn}(1969)}]{SvH69}
558: {Salpeter}, E.~E. \& {van Horn}, H.~M. 1969, \apj, 155, 183
559:
560: \bibitem[{{Schatz} {et~al.}(1999){Schatz}, {Bildsten}, {Cumming}, \&
561: {Wiescher}}]{SBCW99}
562: {Schatz}, H., {Bildsten}, L., {Cumming}, A., \& {Wiescher}, M. 1999, \apj, 524,
563: 1014
564:
565: \bibitem[{{Schwarzschild} \& {H{\"a}rm}(1965)}]{SH65}
566: {Schwarzschild}, M. \& {H{\"a}rm}, R. 1965, \apj, 142, 855
567:
568: \bibitem[{{Strohmayer} \& {Bildsten}(2006)}]{SB03}
569: {Strohmayer}, T. \& {Bildsten}, L. 2006, in Compact Stellar X-Ray Sources, ed.
570: W. H. G. Lewin and M. van der Klis (Cambridge: Cambridge Univ. Press)
571:
572: \bibitem[{{Taam}(1985)}]{T85}
573: {Taam}, R.~E. 1985, Annual Review of Nuclear and Particle Science, 35, 1
574:
575: \bibitem[{{Taam} {et~al.}(1996){Taam}, {Woosley}, \& {Lamb}}]{TWL96}
576: {Taam}, R.~E., {Woosley}, S.~E., \& {Lamb}, D.~Q. 1996, \apj, 459, 271
577:
578: \bibitem[{{Tan} {et~al.}(2005){Tan}, {G{\"o}rres}, {Daly}, {Couder}, {Couture},
579: {Lee}, {Stech}, {Strandberg}, {Ugalde}, \& {Wiescher}}]{Tetal05}
580: {Tan}, W.~P., {G{\"o}rres}, J., {Daly}, J., {Couder}, M., {Couture}, A., {Lee},
581: H.~Y., {Stech}, E., {Strandberg}, E., {Ugalde}, C., \& {Wiescher}, M. 2005,
582: \prc, 72, 041302
583:
584: \bibitem[{{Tawara} {et~al.}(1984){Tawara}, {Hirano}, {Kii}, {Matsuoka}, \&
585: {Murakami}}]{THKMM84}
586: {Tawara}, Y., {Hirano}, T., {Kii}, T., {Matsuoka}, M., \& {Murakami}, T. 1984,
587: \pasj, 36, 861
588:
589: \bibitem[{{van Paradijs} {et~al.}(1979){van Paradijs}, {Cominsky}, {Lewin}, \&
590: {Joss}}]{vPCLJ79}
591: {van Paradijs}, J., {Cominsky}, L., {Lewin}, W.~H.~G., \& {Joss}, P.~C. 1979,
592: \nat, 280, 375
593:
594: \bibitem[{{van Paradijs} {et~al.}(1988){van Paradijs}, {Penninx}, \&
595: {Lewin}}]{vPPL88}
596: {van Paradijs}, J., {Penninx}, W., \& {Lewin}, W.~H.~G. 1988, \mnras, 233, 437
597:
598: \bibitem[{{Wagoner}(1969)}]{W69}
599: {Wagoner}, R.~V. 1969, \apjs, 18, 247
600:
601: \bibitem[{{Wallace} \& {Woosley}(1981)}]{WW81}
602: {Wallace}, R.~K. \& {Woosley}, S.~E. 1981, \apjs, 45, 389
603:
604: \bibitem[{{Wiescher} {et~al.}(1999){Wiescher}, {G{\"o}rres}, \&
605: {Schatz}}]{WGS99}
606: {Wiescher}, M., {G{\"o}rres}, J., \& {Schatz}, H. 1999, Journal of Physics G
607: Nuclear Physics, 25, 133
608:
609: \bibitem[{{Wiescher} {et~al.}(1998){Wiescher}, {Schatz}, \&
610: {Champagne}}]{WSC98}
611: {Wiescher}, M., {Schatz}, H., \& {Champagne}, A.~E. 1998, Royal Society of
612: London Philosophical Transactions Series A, 356, 1
613:
614: \bibitem[{{Woosley} {et~al.}(2004){Woosley}, {Heger}, {Cumming}, {Hoffman},
615: {Pruet}, {Rauscher}, {Fisker}, {Schatz}, {Brown}, \& {Wiescher}}]{Wetal04}
616: {Woosley}, S.~E., {Heger}, A., {Cumming}, A., {Hoffman}, R.~D., {Pruet}, J.,
617: {Rauscher}, T., {Fisker}, J.~L., {Schatz}, H., {Brown}, B.~A., \& {Wiescher},
618: M. 2004, \apjs, 151, 75
619:
620: \end{thebibliography}
621:
622: \clearpage
623:
624: \end{document}
625:
626: