1: %\documentclass[12pt,preprint]{aastex}
2:
3: %\documentstyle[12pt,aasms4]{article}
4:
5: %\documentstyle[emulapj,draft]{article}
6:
7: \documentstyle[emulapj,epsfig]{article}
8:
9: %\documentstyle[emulapj,draft]{article}
10:
11:
12: %\documentclass{aastex}
13: %\usepackage{emulateapj5,epsfig}
14: %\input epsf.def
15: %\input{standard-defs.tex}
16:
17: \font\tenbg=cmmib10 at 10pt
18: \def \rvecmu{{\hbox{\tenbg\char'026}}}
19: \def \rvecphi{{\hbox{\tenbg\char'036}}}
20: \def \Omegabold {{\hbox {\tenbg\char'012}}}
21: \def \rvecOmega {{\hbox {\tenbg\char'012}}}
22:
23:
24: %\usepackage{emulateapj5}
25:
26: \begin{document}
27:
28: \bigskip
29:
30: %\magnification=\magstep1
31: %\baselineskip=12pt \vbadness=10000
32: %\font\one=cmbx10 scaled\magstep2
33: \font\two=cmbx10 scaled \magstep1
34:
35:
36: %\def\idem{\smallskip\noindent
37: %\hangindent 2 pc }
38:
39:
40:
41:
42: \title{\bf Oscillations of MHD shock waves on
43: the surfaces of T~Tauri stars}
44:
45:
46: \author{A.V. Koldoba}
47: \affil{Institute of Mathematical Modelling, Russian Academy of
48: Sciences, Moscow, Russia; ~ koldoba@rambler.ru}
49:
50: \author{G.V. Ustyugova} \affil{Keldysh Institute of Applied Mathematics, Russian Academy of
51: Sciences, Moscow, Russia; ~ ustyugg@rambler.ru}
52:
53: \author{M. M.~Romanova}
54: \affil{Department of Astronomy, Cornell University, Ithaca, NY
55: 14853-6801; ~ romanova@astro.cornell.edu}
56:
57: \author{R. V.E.~Lovelace}
58: \affil{Department of Astronomy and Applied and Engineering Physics,
59: Cornell University, Ithaca, NY 14853-6801; ~RVL1@cornell.edu }
60:
61:
62:
63:
64: \def\dst{\displaystyle} %math styles
65: \def\tst{\textstyle} %
66: \def\sst{\scriptstyle} %
67: \def\ssst{\scriptscriptstyle} %
68:
69: % New operators.
70: \def\const{{\rm const}} %definition of constants
71: \def\cos{\mathop{\rm cos}\nolimits}
72: \def\sin{\mathop{\rm sin}\nolimits}
73: \def\tg{\mathop{\rm tg}\nolimits}
74: \def\ctg{\mathop{\rm ctg}\nolimits}
75: \def\matharrow{\mathop{\protect{\relbar\joinrel\longrightarrow}}\limits}
76: %\def\matharrow{\mathop{\longrightarrow}\limits}
77:
78: %% Height of strings in the mathematical array:
79: \def\arh#1{\renewcommand{\arraystretch}{#1}}
80: \def\arhp{\renewcommand{\arraystretch}{1.5}}
81:
82: %% regular derivartives:
83: \def\prodi#1#2{\frac{d #1}{d #2}}
84: \def\prodii#1#2{\frac{d^2 #1}{d {#2}^2}}
85:
86: %% partial derivatives:
87: \def\parti#1#2{\frac{\partial #1}{\partial #2}}
88: \def\partii#1#2{\frac{\partial^2 #1}{\partial {#2}^2}}
89:
90: %\newcommand{\grad}{\mathop{\rm grad}\nolimits}
91: %\renewcommand{\sectionmark}[1]{}
92: %\renewcommand{\subsectionmark}[1]{}
93: %\renewcommand{\refname}{Reference List}
94: %\renewcommand{\figurename}{Fig.}
95:
96:
97:
98: \begin{abstract}
99:
100: This work treats the matter deceleration in a
101: magnetohydrodynamics
102: radiative shock wave at the surface of a star.
103: The problem is relevant
104: to classical T~Tauri stars where
105: infalling matter is channeled along the star's magnetic
106: field and stopped in the dense layers of photosphere.
107: A significant new aspect of the present work is that the magnetic
108: field has an arbitrary angle with
109: respect to the normal to the star's surface.
110: We consider the limit where
111: the magnetic field at the surface of the
112: star is not very strong in the sense that
113: the inflow is super Alfv\'enic.
114: In this limit the initial
115: deceleration and heating of plasma (at the entrance to the cooling
116: zone) occurs in a fast magnetohydrodynamic shock wave.
117: To calculate the intensity
118: of radiative losses we use ``real" and ``power-law"
119: radiative functions.
120: We determine the stability/instability of the
121: radiative shock wave as a function of parameters of the incoming
122: flow: velocity, strength of the magnetic field, and its inclination
123: to the surface of the star.
124: In a number of simulation runs with the
125: ``real" radiative function, we find a simple criterion for
126: stability of the radiative shock wave.
127: For a wide range
128: of parameters, the periods of oscillation of the shock wave are of
129: the order $0.02-0.2 ~{\rm s}$.
130:
131: \end{abstract}
132:
133: \begin{keywords} stars: magnetic fields --- stars: oscillations ---
134: MHD --- accretion --- shock waves --- instabilities
135: \end{keywords}
136:
137:
138:
139: \section{Overview of the Problem}
140:
141: In classical T Tauri stars (CTTSs) matter accretes from the disk to
142: a star through magnetospheric funnel streams (Camenzind 1990;
143: K\"onigl 1991; see also recent review by Bouvier et al. 2007).
144: Similar type accretion but at smaller scales is expected to a
145: magnetized white dwarf (e.g., Warner 1995) and magnetized neutron
146: stars (e.g., Ghosh \& Lamb 1979). In the funnel stream, matter is
147: lifted above the equatorial plane and falls down onto the star due
148: to the gravitational acceleration. Large-scale magnetospheric flow
149: has been recently investigated in 2D magnetohydrodynamic (MHD)
150: simulations (Romanova et al. 2002; Bessolaz et al. 2008) and in full
151: 3D MHD simulations (Romanova et al. 2003, 2004; Kulkarni \& Romanova
152: 2005). Many aspects of the global magnetospheric flow are now
153: understood. However, interaction of the funnel streams with the
154: surface of the star has not been adequately investigated.
155:
156:
157: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
158:
159: \begin{figure*}[t]
160: \epsscale{1.0} \plotone{f1.eps} \caption{Sketch of the geometry of
161: the flow.
162: Matter with density
163: $\rho_{\rm in}$ and pressure $p_{\rm in}$ flows towards the star
164: with velocity $v_{\rm in}$ along the magnetic field which has
165: strength $B_{\rm in}$ and is inclined relative to the normal to the
166: surface of the star $x$ at an angle $\chi$. Matter slows down in the
167: MHD shock wave close to the surface of the star and radiates and
168: cools down in the ``cooling zone". Matter may have different angle
169: relative to the field, but we consider the coordinate system in
170: which vectors velocity and magnetic field are parallel to each
171: other.} \label{Figure 1}
172: \end{figure*}
173:
174: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
175:
176: Theoretical models indicate that close to a star matter in the
177: funnel streams is accelerated to almost free-fall velocity, before
178: it hits the high-density layers of the stellar atmosphere. The
179: matter rapidly slows down, forming a shock wave close to the
180: stellar surface.
181: Most of the energy of the flow is radiated behind the shock
182: wave (e.g., Lamzin 1995, 1998; Muzerolle et al. 1998; Calvet \&
183: Gullbring 1998; Gullbring et al. 2000; Ardila \& Basri 2000).
184: In the case of CTTSs most of energy is radiated in the ultraviolet and soft
185: X-ray bands (e.g., Calvet \& Gullbring 1998; G\"unther \& Schmitt
186: 2008). The 3D MHD simulations of magnetospheric flow have show
187: that the hot spots are inhomogeneous and are expected to have higher
188: temperature in the central regions of spots compared to peripheral
189: regions (Romanova et al. 2004). This may have a number of important
190: consequences for investigation of hot spots including the dependence
191: of the filling factor on the wavelength.
192:
193:
194: If star's magnetic field of the star is {\it strong}, then close
195: to the stellar surface the magnetic energy-density is larger than
196: that the kinetic energy-density of the matter.
197: Consequently the matter is passively channeled along the field
198: lines. In this sub-Alfv\'enic regime a hydrodynamic approach is
199: usually adopted for modeling the shock waves (see \S 2).
200: The shock wave is found to be non-stationary. It {\it oscillates} with
201: a high frequency due to the competition
202: between accretion heating in the shock front and radiative cooling
203: behind the front (Langer, Chanmugan \& Shaviv 1981; Chevalier \&
204: Imamura 1982).
205: If the magnetic field is not very strong near the
206: surface of the star then the flow may be super-Alfv\'enic.
207: In this regime the orientation of the magnetic
208: field may influence stability of the shock.
209: Only the special case of the flow perpendicular to the
210: magnetic field has been considered so far.
211: It is know that this transverse magnetic field can suppress
212: instability of the shock wave (Smith 1989; Toth \& Draine 1993).
213: In this paper we investigate the stability of the radiative shock waves in
214: the super-Alfv\'enic regime for different orientations of the magnetic
215: field relative to stellar surface.
216: We consider small patch of the hot
217: spot and investigate the stability of the radiative MHD shock wave
218: for parameters typical for CTTSs.
219: Figure 1 shows a sketch of the considered geometry.
220:
221:
222: For CTTSs the expected periods of the oscillations of the shock are
223: very short. The periods vary between $0.02$ and $0.2$ seconds
224: depending on the parameters. Oscillations in this period range have
225: not been observed so far. Smith, Jones \& Clarke (1996) searched
226: for rapid photometric variability in several CTTSs in the range of
227: periods from minutes to hours and did not find oscillations. Much
228: higher temporal resolution is required to resolve the oscillations
229: discussed in this paper.
230:
231:
232: Section 2 of this paper discusses the earlier research on radiative
233: shocks.
234: Section 3 discusses the model and basic equations,
235: and Section 4 the dimensionless variables and scalings.
236: Section 5 comments on the dimensionless variables and the
237: scalings of different quantities, and also describes the stationary
238: structure of the shock. Section 6 discusses the methods used to
239: study the time-dependent shocks, and Section 7 gives our results.
240: Section 8 gives the conclusions of this work.
241:
242:
243:
244: \section{Earlier Research of Radiative Shocks and Radiative Cooling Function}
245:
246:
247: The stability of shock waves has been investigated by different groups
248: both analytically (linear analysis) and numerically. Most of the
249: investigations have been restricted to a {\it purely hydrodynamic
250: analysis}, because in many situations matter is channeled along the
251: field lines and the problem can be considered as non-magnetic.
252:
253: First results on interaction of the funnel streams with a star and
254: cooling in the radiative shock wave have been obtained in
255: application to accreting white dwarfs. Numerical modeling of
256: radiative shock wave at the surface of white dwarf led to discovery
257: of instability which is driven by alternation of accreting heating
258: in the shock wave and radiative cooling behind the shock wave and in
259: resulting oscillations of the position of the shock front
260: (Langer, Chanmugan \& Shaviv 1981). Linear analysis of
261: the stability of a one-dimensional radiative shock wave was done by
262: Chevalier \& Imamura (1982).
263: These authors assumed that the
264: radiative cooling from a unit volume is $\rho^2 T^{\alpha}$, and
265: they studied the stability as a function $\alpha$.
266: In a more general form the problem has been investigated by
267: Ramachandran \& Smith (2004).
268: This work assumed that the
269: radiative losses vary as $\rho^{\beta} T^{\alpha}$.
270: The boundaries of the stable and unstable regions were found
271: as well as the frequencies and growth rates of the lowest frequency
272: modes for different values of $\alpha,~ \beta$.
273: Ramachandran \& Smith (2006)
274: investigated the influence of the Mach number of the inflowing gas
275: to the stability of the radiative shock wave. In addition they
276: considered flow at different adiabatic indices $\gamma$ typical for
277: astrophysical applications and two types of the boundary conditions
278: at the ``wall" where the flow is stopped.
279: The influence of boundary conditions on the stability
280: of radiative shock waves has been investigated in detail by Saxton
281: (2002).
282:
283: A detailed investigation of the linear and nonlinear evolution of
284: one-dimensional radiative shock waves has been done by Mignone
285: (2005, see also Toth \& Draine 1993). The intensity of the radiative
286: losses was taken to be $\rho^2 T^{\alpha}$.
287: The stability has been investigated in the
288: linear approximation for the first eight low-frequency modes and it
289: has been established, that:
290: (1) The first eight modes are
291: stable for $\alpha > 0.92$; (2) The fundamental mode ($n=0$)
292: becomes unstable for $\alpha < 0.388$,
293: the $n=1$ mode for $\alpha < 0.782$;
294: (3) In the unstable regime the growth rate is larger for larger mode
295: numbers, but the growth rate of the higher-$n$ modes is not
296: very different from the growth rate of the $n=7$ mode;
297: (4) The normalized frequencies of the corresponding modes have
298: an approximately linear dependence on $n$, that is, $\omega_n(\alpha)
299: = \omega_0(\alpha) + n \Delta \omega(\alpha)$, and decrease
300: as $\alpha$ increases (excluding the fundamental mode $n=0$).
301: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
302:
303: \begin{figure*}[t]
304: \epsscale{0.9} \plotone{f2.eps} \caption{Cooling function
305: $\Lambda_{\rm RTV}(T)$ (thick solid line) and $\Lambda_{\rm GS}(T)$
306: (thin solid line) as a function of $T$ in K. The vertical dotted
307: line shows our illustrative value of velocity before the shock
308: $v_{in}\cos\chi = 1.3\times 10^7 {\rm cm~s^{-1}}$ and the
309: corresponding temperature behind the shock.} \label{Figure 2}
310: \end{figure*}
311:
312: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
313:
314: The stability of a radiative shock wave in the presence of a {\it
315: magnetic field} was investigated by Smith (1989).
316: A more detailed investigation of the shock stability
317: in presence of a magnetic field was done by Toth \& Draine (1993;
318: hereafter TD93).
319: For the situation considered, the matter flows onto the shock
320: wave perpendicular to its front and the magnetic field has only
321: component parallel to the front.
322: It was concluded that even a
323: modest magnetic field may lead to stabilization of a radiative
324: shock wave which is unstable in the hydrodynamic limit.
325: The higher the harmonic number, the larger the value of the magnetic field
326: which is needed for stabilization of the front for a fixed value of
327: $\alpha$.
328: First of all, the magnetic field stabilizes the
329: fundamental mode.
330: In particular, for $\alpha = -0.5$ the fundamental
331: mode is stabilized at $ M_A^{-1} = [( B_y/\sqrt{4 \pi\rho})/v]_{in}
332: = 0.15$, while the higher modes are stabilized at $M_A^{-1}
333: = 0.5$, where $v$ is velocity of the incoming flow and $B_y$ is
334: magnetic field parallel to the front.
335:
336: The subsequent investigation of the stability of the radiative
337: shocks with transversal magnetic field has been done by
338: Ramachandran \& Smith, (2005; hereafter RS05). They considered
339: different values of $\alpha,~ \beta$ in the dependence
340: of the radiative
341: losses, different values $\gamma$ for the adiabatic index, and also
342: different Mach numbers in the inflowing matter.
343: They obtained new
344: results, and also rederived accurately results by TD93 for the
345: case where $\gamma = 5/3$ and $\beta = 2$.
346:
347:
348:
349:
350: These earlier studies show that the stability of radiative
351: shock waves depends strongly on the
352: functional dependence of the radiative cooling.
353: In the present work we consider
354: $\beta = 2$ and a monatomic
355: gas with $\gamma = 5/3$.
356: We investigate the stability of
357: the radiative shock waves using a ``real"
358: radiative loss function.
359: We calculate this ``real"
360: function and approximate it with the power laws.
361: We assume that there
362: is collisional ionization equilibrium (CIE).
363: In this approximation photons
364: freely escape the plasma.
365: Thus, full
366: thermodynamic equilibrium is not established, and the Saha's formula
367: (for calculation of the degree of the ionization) is not applicable
368: (e.g., Spitzer 1968). The radiative losses under these conditions
369: have been calculated by a number of authors.
370: In these
371: calculations the abundances of the elements are assumed to be
372: Solar.
373: In the paper by Rosner, Tucker, \& Vaiana (1978; hereafter
374: RTV) based on the calculations of Raymond \& Smith (1977), the
375: radiative losses in the temperature interval $10^{4.3}{\rm K} < T <
376: 10^7 {\rm K}$ are $n_e n_H \Lambda(T)$, where $n_e, n_H$ are
377: electron density and hydrogen density (total), while the radiation
378: function $\Lambda(T)$ has been approximated by a multi-segmented
379: power-law (see below). Subsequently, $\Lambda(T)$ was been
380: calculated at the interval $10^{3.65}{\rm K} < T < 10^8 {\rm K}$
381: (Peres et al, 1982).
382: We refer to this function as the ``real"
383: cooling function and add a subscript $RTV$.
384: The dependence is
385: $$
386: \Lambda_{\rm RTV}(T) = \left\{
387: \begin{array}{lc}
388: (10^{-7.85} T)^{6.15} & 10^{3.9} {\rm K} < T < 10^{4.3} {\rm K}\\
389: 10^{-21.85} & 10^{4.3} {\rm K} < T < 10^{4.6} {\rm K}\\
390: 10^{-31} T^2 & 10^{4.6} {\rm K} < T < 10^{4.9} {\rm K}\\
391: 10^{-21.2} & 10^{4.9} {\rm K} < T < 10^{5.4} {\rm K}\\
392: 10^{-10.4} T^{-2} & 10^{5.4} {\rm K} < T < 10^{5.75} {\rm K}\\
393: 10^{-21.94} & 10^{5.75} {\rm K} < T < 10^{6.3} {\rm K}\\
394: 10^{-17.73} T^{-2/3} & 10^{6.3} {\rm K} < T < 10^{7} {\rm K}\\
395: 10^{-18.21} T^{-0.6} & 10^{7} {\rm K} < T < 10^{7.6} {\rm K}
396: \end{array}
397: \right.
398: $$
399:
400: The most recent results for radiative
401: losses in the CIE-approximation are
402: given by Gnat \& Sternberg (2007; hereafter GS07).
403: In this work it
404: was accepted that the relative abundances of the hydrogen and helium
405: are $n_{\rm He}/n_{\rm H} =1/12$. Abundances of other elements
406: (${\rm C, Ni, O, N, Mg, Si, S, Fe}$) are small and they do not give
407: a contribution to the total pressure. However, the intensity of the
408: radiative losses significantly depends on the relative abundance of
409: these elements. For the Solar abundance GS07 proposed an
410: approximation formula:
411: \begin{equation}
412: \Lambda_{\rm GS} = 2.3 \times 10^{-19} T^{-0.54}~~~~{\rm erg~
413: cm^3~s^{-1}}~, \label{1}
414: \end{equation}
415: in the temperature range $10^5K < T < 10^8 {\rm K}$.
416: The left
417: boundary of this interval corresponds approximately to the maximum
418: of the ``real" radiation function which is at $2.3 \times
419: 10^5 {\rm K}$.
420: For temperatures higher than $6 \times 10^7 {\rm K}$,
421: the dominant mechanism of radiative losses is bremsstrahlung
422: radiation with the temperature dependence $\Lambda \sim \sqrt{T}$.
423:
424:
425:
426: According to calculations of GS07 in
427: CIE-approximation, hydrogen is completely ionized for
428: temperature $T > 3 \times 10^4 {\rm K}$, and Helium for
429: $T > 2 \times 10^5 {\rm K}$.
430: Accepting their abundances discussed above we
431: obtain for $T > 2 \times 10^5 {\rm K}$, the electron density $n_e =
432: n_{\rm H} + 2 n_{\rm He}$, and an average mass per particle is
433: $0.6m_p$.
434: Thus, the model which we use (based on the
435: CIE-approximation) is applicable for $T > 2 \times 10^5 {\rm K}$.
436: At lower temperatures, the partial ionization
437: of Helium and the change of the
438: average mass per particle become significant.
439: At the temperature $T
440: = 3 \times 10^4 {\rm K}$ the fractions of the neutral and ionized
441: atoms of hydrogen and helium are: $x({\rm H})=3.6 \times
442: 10^{-3},~x({\rm H^+})=0.997,~x({\rm He})=0.4,~x({\rm
443: He^+})=0.6,~x({\rm He^{++}})=0$. Thus an average mass per particle
444: is $0.62 m_p$.
445: At a temperature $T =
446: 2 \times 10^4 {\rm K}$, the corresponding fractions are $x({\rm
447: H})=0.078,~x({\rm H^+})=0.922,~ x({\rm He})=0.993,~x({\rm
448: He^+})=0.007,~x({\rm He^{++}})=0$, where an average mass per
449: particle is $0.66 m_p$. We neglect this factor.
450:
451: Subsequently we assume that plasma is an ideal gas with equation of
452: state $p={\cal R} \rho T$ where ${\cal R} = k_B/(0.6 m_p) = 1.385
453: \times 10^8 {\rm erg~g^{-1}~K^{-1}}$ is gas constant.
454:
455: In the paper GS07 it is shown that for $T < 10^6 {\rm
456: K}$, the radiative losses, calculated in CIE-approximation, exceed
457: losses which occur in non-stationary plasma cooling.
458: This is connected with the fact that
459: in the first case the ionization level
460: of elements participating in the main radiative processes is lower
461: because in the non-stationary regime recombination lags the
462: cooling. However, we use the more detailed approximation mentioned
463: above after changing the normalization.
464:
465:
466:
467:
468:
469: Figure 2 shows the radiative cooling functions $\Lambda_{\rm RTV}(T)$
470: (thick solid line) and $\Lambda_{\rm GS}(T)$ (thin solid line).
471: The diagonal dashed line shows the dependence of the
472: upstream velocity normal to the shock ($v_{in}\cos\chi$)
473: on the temperature behind the shock (for $\gamma=5/3,
474: p_{\rm in} = 0$):
475: $ T_s = 3 (v_{\rm in} \cos \chi)^2/(16
476: {\cal R})$.
477: The horizontal dashed
478: line shows the velocity of the incoming flow $v_{\rm in}$ (for
479: illustration of the calculated results we take $v_{\rm in} \cos \chi
480: = 1.3 \times 10^7 {\rm cm~s^{-1}}$) and corresponding to this
481: velocity temperature behind the shock wave front.
482:
483: Section 2 of the paper discusses the model and basic equations,
484: and Section 3 the dimensionless variables and scalings.
485: Section 3 comments on the dimensionless variables and the
486: scalings of different quantities. Section 4 describes the
487: stationary structure of the shock. Section 5 discusses
488: the methods used to study the time-dependent shocks, and
489: Section 6 gives our results. Section 7 gives the conclusions
490: of this work.
491:
492:
493:
494: \section{Formulation of the Problem}
495:
496:
497: %\textbf{ We consider the interaction of accreting matter with the
498: %surface of a star in application to classical T~Tauri stars.
499: We investigate formation and evolution of the shock wave near the
500: surface of the star which forms as a result of disk accretion to a
501: star through a funnel flow. The accreting matter is sufficiently
502: ionized to satisfy the frozen-in condition so that matter of the
503: funnel streams is channeled by the magnetic field and close to the
504: star it flows along the field lines.
505:
506: Figure 1 shows the geometry.
507: Matter with velocity
508: $v_{\rm in}$, density $\rho_{\rm in}$, and pressure $p_{\rm in}$
509: flows towards the surface of the star along the magnetic field.
510: The magnetic field has a strength $B_{\rm in}$ and is
511: directed at an angle $\chi$ relative to the normal vector to the
512: surface of the star $\hat{\bf x}$. The $x$-axis is normal to the
513: surface of the star and the $y-$axis is tangential to its surface
514: and is directed such that the magnetic field is located in the
515: $(x,y)$ plane. We neglect small perturbations in $z-$direction
516: associated with propagation of the Alfv\'en waves. We consider that
517: the magnetic field has an arbitrary inclination angle $\chi$
518: relative to the normal
519: to the star's surface.
520: In general, the matter flow velocity is not
521: parallel to the magnetic field.
522: However, such parallel orientation can be obtained by transformation of
523: the coordinate system.
524:
525: Heating of matter occurs in the front of the MHD shock wave.
526: In the cooling zone behind the shock, matter radiates
527: energy, is decelerated, and become denser up to the moment, when the
528: radiative cooling stops.
529: Formally this happens when $T=0$.
530: The height of the radiative zone is small compared to either width
531: of the funnel stream or to the radius of the star.
532: Thus we can neglect the
533: inhomogeneity of the accretion flow in the $(y,z)$ directions. (See
534: Canalle et al. 2005 for a discussion of cases where the converging
535: of the field lines is important.)
536: We also neglect small
537: effects associated with acceleration by gravity of the star
538: (considered e.g., by Cropper et al. 1999) because the region we
539: consider is small.
540: Generally, the structure consisting of the shock wave and
541: the cooling zone is unstable to both longitudinal, $x$)
542: perturbations (which can be studied in one-dimensional approach),
543: and to transverse ($y,z$) perturbations (Bertschinger 1986; Imamura
544: et al. 1996).
545: The latter makes the problem more than
546: one-dimensional even if the flow is homogeneous.
547: In the present work
548: the spatial perturbations are not considered.
549:
550: We consider conditions where
551: the star's mass is $M_* =
552: 0.8M_{\sun}$ and its radius is $R_* = 2 R_{\sun}$.
553: In this case the
554: free-fall speed at the star's surface is $ v_{ff} = \sqrt{2GM_*/R_*}
555: = 4 \times 10^7 {\rm cm~s^{-1}}$.
556: If the temperature
557: of accreting matter is $10^4 {\rm K}$, then the sound speed is $2
558: \times 10^6 {\rm cm~s^{-1}}$ (for average mass per particle
559: $0.6m_p$).
560: Clearly, the accretion is strongly
561: supersonic; the sonic Mach number is $20$.
562: For a surface magnetic field
563: $B=10^3 {\rm G}$ and density of the inflowing matter $10^{-11} {\rm
564: g~cm^{-3}}$ (Romanova et al. 2002), the Alfv\'en velocity is $ c_A
565: = B/\sqrt{4 \pi \rho} = 10^8 {\rm cm~s^{-1}}$.
566: For these
567: parameters the flow is sub-Alfv\'enic.
568: However, for $B < 3\times 10^2 {\rm G}$ and the other parameters
569: the same the flow is super-Alfv\'enic
570: and perturbations from the shock wave cannot propagate up the
571: stream.
572: Strictly speaking we should compare the velocity of the
573: accreting matter with the fast magnetosonic velocity, but it does
574: not differ significantly from the Alfv\'en velocity because the sound speed
575: is small.
576: Thus, both cases are interesting: stability of the
577: radiative shock waves for sub-Alfv\'enic
578: and super-Alfv\'enic inflows.
579: However, we restrict the present study to
580: super-Alfv\'enic inflow.
581: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
582: \begin{figure*}[t]
583: \epsscale{1.1} \plotone{f3.eps} \caption{The position of the
584: boundary between stable and unstable radiative shock waves as a
585: function of parameters $(\sigma, \sin \chi)$ for a velocity of
586: accretion $v_{\rm in} \cos\chi = 1.3 \times 10^7 {\rm cm~s^{-1}}$.}
587: \label{Figure 3}
588: \end{figure*}
589: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
590:
591: We assume that the radiation zone behind the shock
592: is optically thin in the
593: direction perpendicular to the star's surface and to the front
594: of the shock wave.
595: The dominant radiation loss mechanisms
596: are determined by the ``thermodynamic" state of plasma.
597: We assume
598: that the state of the matter can be
599: described in terms of a temperature
600: which is the same for the ions and electrons.
601: For the considered conditions,
602: the radiative losses per unit volume is
603: $\rho^2 \Lambda(T)$, where $\rho$ is the plasma density and
604: $\Lambda(T)$ is the radiative function obtained from $\Lambda_{\rm
605: RTV}(T)$ or from $\Lambda_{\rm GS}(T)$ by renormalization.
606: We consider
607: that the accreting matter is an ideal gas with adiabatic index
608: $\gamma = 5/3$ both before the compression at the shock front and
609: in the radiation zone.
610:
611: In the absence of the magnetic field, and when the matter falls
612: perpendicular to the surface of the star, the temperature behind the
613: shock is $T = 10^5 v_7^2 ~{\rm K}$, where $ v_7 = v/(10^7 {\rm
614: cm~s^{-1}})$, and the average mass per particle is $0.6m_p$.
615: For velocities of the incoming matter $v = 4
616: \times 10^7 {\rm cm~s^{-1}}$, the temperature behind the shock
617: reaches $1.6 \times 10^6 {\rm K}$. At such temperature both
618: hydrogen and helium atoms are completely ionized according to the
619: CIE approximation.
620:
621: We consider the situation where the matter accretes to the star along
622: magnetic field lines inclined by an angle $\chi$ to the normal to
623: the star's surface.
624: The shock is parallel to the star's surface and it remains
625: parallel as it moves.
626: That is, we consider variations only in the $x-$direction.
627: The equations describing this situation are
628: following:
629: \begin{eqnarray}
630: \parti{\rho}{t} + \parti{(\rho v_x)}{x} = 0~,
631: \nonumber
632: \end{eqnarray}
633: \begin{eqnarray}
634: \parti{(\rho v_x)}{t} + \parti{}{x} \left( \rho v_x^2 + p +
635: \frac{B_y^2}{8\pi} \right) = 0~,
636: \nonumber
637: \end{eqnarray}
638: \begin{eqnarray}
639: \parti{(\rho v_y)}{t} + \parti{}{x} \left( \rho v_x v_y -
640: \frac{B_x B_y}{4\pi} \right) = 0~,
641: \nonumber
642: \end{eqnarray}
643: \begin{eqnarray}
644: B_x = \const~,
645: \nonumber
646: \end{eqnarray}
647: \begin{eqnarray}
648: \parti{B_y}{t} + \parti{}{x} \left( v_x B_y - v_y B_x \right) = 0~,
649: \nonumber
650: \end{eqnarray}
651: \begin{eqnarray}
652: \parti{}{t} \left( \frac{\rho v^2}{2} + \frac{p}{\gamma-1} +
653: \frac{B^2}{8\pi} \right)
654: \nonumber
655: \end{eqnarray}
656: \begin{eqnarray}
657: \quad+ \parti{}{x} \left[ v_x \left( \frac{\rho v^2}{2}+ \frac{\gamma
658: p}{\gamma-1} \right) + \frac{B_y}{4\pi} \big( v_x B_y - v_y B_x
659: \big) \right]
660: \nonumber
661: \end{eqnarray}
662: \begin{equation}
663: \quad \quad \quad= - \rho^2 \Lambda(T)~.
664: \end{equation}
665: Here $v^2 \equiv v_x^2 + v_y^2~,~B^2 \equiv B_x^2 + B_y^2$.
666: In the unperturbed flow upstream of the shock,
667: \begin{eqnarray}
668: B_x = B_{\rm in} \cos \chi~,~~~~B_y = B_{\rm in} \sin \chi~,
669: \nonumber
670: \end{eqnarray}
671: \begin{eqnarray}
672: v_x = -v_{\rm in} \cos \chi~,~~~~v_y = - v_{\rm in} \sin \chi~,
673: \nonumber
674: \end{eqnarray}
675: \begin{eqnarray}
676: \rho = \rho_{\rm in}~,~~~~p = p_{\rm in}~.
677: \nonumber
678: \end{eqnarray}
679: We consider that the pressure in the incoming flow is very small so
680: that it does not influence the MHD-shock wave and cooling zone.
681: Equivalently, the sonic Mach number is much larger than unity. As
682: discussed earlier, we choose a coordinate system in which the flow
683: velocity and magnetic field are both in the $(x,y)$ plane.
684:
685:
686:
687:
688: In contrast with the non-magnetic case
689: (TD93), the density $\rho$
690: does not increase unrestrictedly at the
691: right-hand boundary of the radiative zone and correspondingly the
692: velocity $v_x$ does not approach zero.
693: Here, the temperature $T =
694: p/({\cal R} \rho)$ may approach zero not because $\rho \to \inf$ but
695: because $p \to 0$.
696: At the same time the total, gas plus magnetic field pressure
697: remains constant.
698: In the cooling zone the accreting
699: matter does not reach zero flow speed.
700: Thus, the radiative shock provides
701: only part of the deceleration and absorption of
702: matter by the magnetized star.
703:
704:
705:
706:
707: Our treatment of the
708: stability of the radiative MHD
709: shock wave is different in essential
710: respects from that of TD93 and RS05.
711: In these
712: papers the magnetic field is transverse to the
713: flow and parallel to the star's surface, ${\bf B}=B_y\hat{\bf y}$.
714: We neglect the difference in cooling laws and adiabatic index
715: $\gamma$ (RS05) which are not significant.
716: To approach this case in our model,
717: we need to take $\chi \to \pi/2$, and
718: parameters in the incoming flow $v_x = -v_{\rm in} \cos \chi,~ B_y =
719: B_{\rm in} \sin \chi \to B_{\rm in}$.
720: The Alfv\'enic Mach number is
721: \begin{eqnarray}
722: M_A = \frac{v_{\rm in} \cos \chi}{B_{\rm in}/ \sqrt{4\pi
723: \rho_{\rm in}}} = \frac{\cos \chi}{\sigma}~.
724: %\nonumber
725: \end{eqnarray}
726: To have $M_A$ a fixed value as $\chi \to \pi/2$, we let
727: $\sigma \equiv \cos \chi /M_A$ and require
728: \begin{eqnarray}
729: B_x = B_{\rm in} \cos \chi \to 0~,\quad
730: v_x \to - M_A \frac{B_{\rm in}}{\sqrt{4\pi \rho_{\rm in}}}~,
731: \nonumber
732: \end{eqnarray}
733: \begin{eqnarray}
734: v_y = - M_A \tan \chi \frac{B_{\rm in}}{\sqrt{4\pi \rho_{\rm in}}}
735: \to \infty~.
736: \nonumber
737: \end{eqnarray}
738: A tangential velocity $v_y$ can be included in
739: the calculations of TD93 by a Galilean transformation
740: to another reference frame.
741: Thus, a radiative MHD shock wave with the
742: magnetic field parallel to the star's surface
743: and perpendicular to the flow corresponds to the
744: limit where $ \chi \to
745: \pi/2$ and $ \sigma = \cos \chi/M_A \to 0$.
746:
747:
748:
749: \section{Dimensionless Variables and Scalings}
750:
751: Consider firstly the case where the radiation function is a power law,
752: $\Lambda(T) = A ({\cal R}T)^{\alpha}$, where ${\cal R}$ is the gas
753: constant. The coefficient $A$ has dimension
754: $ {\rm cm}^{5-\alpha}{\rm g^{-1}}{\rm s}^{\alpha-3}$.
755: For estimates we assume that the hydrogen
756: and helium are completely ionized and that the average mass per particle
757: is $0.6m_p$.
758: In the paper GS07 the radiative energy
759: losses per unit of volume are given in the form of equation (1).
760: With these simplifications and renormalization,
761: \begin{eqnarray}
762: \Lambda_{\rm GS} = 1.37 \times 10^{32} ({\cal R} T/{\rm
763: cm^2~s^{-2}})^{-0.54} {\rm cm^5~g^{-1}~s^{-3}} ~. \nonumber
764: \end{eqnarray}
765: We rewrite our equations in dimensionless form choosing fiducial
766: dimensional values for the main variables and introducing
767: dimensionless variables $\tilde A=A/A_0$ for different variables A.
768: The fiducial values are taken to be:
769: For the velocity, $v_0 = v_{\rm in}\cos
770: \chi$; for the density, $\rho_0 = \rho_{\rm in}$; for the distance,
771: $v_0^{3-2\alpha}/A \rho_0$; for time, $v_0^{2-2\alpha}/(A \rho_0)$; for the
772: pressure, $\rho_0 v_0^2$; for the magnetic field, $v_0 \sqrt{\rho_0}$;
773: and for the temperature $v_0^2/{\cal R}$.
774: In dimensionless variables, the system
775: of equations has the same form. Subsequently we remove tilde signs
776: from dimensionless variables.
777: The formula for
778: the radiative losses now has the form $\rho^2 T^{\alpha}$, while in
779: the incoming flow $v_{\rm in} \cos \chi = -1, \rho_{\rm in} = 1,
780: B_{\rm in} = \sigma \sqrt{4 \pi},~ p_{\rm in} \ll 1$, where $\sigma$
781: is the inverse of the Alfv\'en Mach number.
782:
783:
784: It is clear that in power law case,
785: $\Lambda \sim T^\alpha$, the stability/instability of
786: the radiative shock in presence of a magnetic field does not
787: depend solely on $v_{\rm in}$ and $B_{\rm in}$, but instead on their
788: combination in the form $ (B_{\rm in}/\sqrt{4 \pi \rho_{\rm
789: in}})/v_{\rm in} = \sigma = M_A^{-1}$.
790: Furthermore, the stability
791: criterion does not depend on the coefficient $A$.
792: However, $A$ determines
793: the spatial scale of the radiative
794: zone and the temporal scale for
795: oscillations if they are present.
796:
797:
798: We chose the spatial scale is $ 10^7 {v_7^{4.08}}/(\rho_{-11}) {\rm
799: cm}$ and the time scale is $ {v_7^{3.08}}/(\rho_{-11}) {\rm s}$,
800: where $v_7$ is the velocity in units of $10^7 {\rm cm~s^{-1}}$ and
801: $\rho_{-11}$ is the density in units of $10^{-11} {\rm g~cm^{-3}}$.
802: In reality, both estimates are about three orders of magnitude
803: larger than the observed values. Such scales follow from the
804: solution of the equation for the stationary shock wave, where the
805: pressure goes to zero for $x>>1$ (TD93). The same is true for the
806: scale of time.
807: Thus, the height of the cooling zone is in fact $\sim 10^4$ cm,
808: and the periods of oscillation are several hundredths of a second.
809:
810:
811:
812:
813:
814:
815: We now consider the ``real" radiative function
816: (RTV) which is not a power law.
817: This is why strictly speaking the stability condition
818: of radiative shock waves depends not only on $\sigma$ but
819: also on both $v_{\rm in}$ and $B_{\rm in}$.
820: However, over a
821: sufficiently wide range of temperatures,
822: the radiative function can be roughly
823: approximated by a power law of the temperature.
824: This suggests that the stability condition
825: will depend mainly on $\sigma$ with a weaker
826: dependence on $v_{\rm in}$ and $B_{\rm in}$.
827:
828: \begin{table*}
829: \begin{center}
830: \begin{tabular}{|c|c|c|c|c|c|} \hline
831: $N $ &$ cooling $&$ v_{in} \cos \chi $&$ \sin \chi $&$ B_{in} $&$ stab / $\\
832: &$ function $&$ {\rm cm~s^{-1}} $&$ $&$ (\rm G) $&$ unstab $\\ \hline
833: $I $ &$ RTV $&$ 1.3 \times 10^7 $&$ 0.154 $&$ 40.8 $&$
834: unst $\\ \hline $II$ &$ GS $&$ 1.3 \times 10^7 $&$ 0.154
835: $&$ 40.8 $&$ unst $\\ \hline $III$&$ RTV $&$ 3 \times
836: 10^7 $&$ 0.154 $&$ 94.9 $&$ unst $\\ \hline $IV $&$ RTV
837: $&$ 1.3 \times 10^7 $&$ 0.368 $&$ 36.4 $&$ stab $\\ \hline
838: \end{tabular}
839: %\end{center}
840: \caption{The Table shows parameters used in our main runs.}
841: \end{center}
842: \end{table*}
843:
844: \section{Stationary structure}
845:
846: We are interested in the stability of the stationary flow consisting
847: of the MHD shock wave and the downstream radiative zone.
848: The stationary flow is described by
849: \begin{eqnarray}
850: \rho v_x = -j = - \rho_{\rm in} v_{\rm in} \cos \chi~,
851: \nonumber
852: \end{eqnarray}
853: \begin{eqnarray}
854: p + \rho v_x^2 +
855: \frac{B_y^2}{8\pi} = q_x = p_{\rm in} + \rho_{\rm in} v_{\rm in}^2
856: \cos^2 \chi + \frac{B_{\rm in}^2}{8\pi} \sin^2 \chi~,
857: \nonumber
858: \end{eqnarray}
859: \begin{eqnarray}
860: \rho v_x v_y - \frac{B_x B_y}{4\pi} = q_y = \left( \rho_{\rm in}
861: v_{\rm in}^2 - \frac{B_{\rm in}^2}{4\pi} \right) \sin \chi \cos
862: \chi~, \label{3}
863: \nonumber
864: \end{eqnarray}
865: \begin{eqnarray}
866: v_x B_y - v_y B_x = 0 ~,
867: \nonumber
868: \end{eqnarray}
869: \begin{equation}
870: j \frac{d}{dx}
871: \left( \frac{v^2}{2} + \frac{\gamma p/\rho}{\gamma-1} \right)
872: = \rho^2 \Lambda~.
873: \end{equation}
874: In the last equation there is no term describing the energy flux of
875: the electromagnetic field, because the Poynting flux in our
876: coordinate system is zero (${\bf E}=-{\bf v}\times {\bf B}/c =0$).
877:
878: The stationary structure of the shock wave and
879: radiative zone is described
880: by the following equation for specific volume $V$,
881: \begin{eqnarray}
882: \frac{dV}{dx} = \frac{\Lambda(T)}{j V^2 F(V)}~,
883: %\nonumber
884: \end{eqnarray}
885: where
886: \begin{eqnarray}
887: F(V) = j^2 V \left[ 1 + \frac{q_y^2 B_{x}^2/4\pi}
888: {(B_{x}^2/4 \pi - j^2V)^3} \right] +
889: %\nonumber
890: \end{eqnarray}
891: \begin{eqnarray}
892: \quad \quad \frac{\gamma}{\gamma+1} \left[ q_x - 2j^2V - \frac{q_y^2
893: B_x^2/8\pi}{(B_x^2/4 \pi - j^2V)^3} \left( j^2V +
894: \frac{B_x^2}{4\pi}\right) \right]~.
895: \nonumber
896: \end{eqnarray}
897: The other variables are determined from equations (\ref{3}).
898: In particular,
899: \begin{eqnarray}
900: v_x = -jV~,~~~~B_y = -\frac{B_x q_y}{B_x^2/4\pi - j^2V}~,
901: \nonumber
902: \end{eqnarray}
903: \begin{eqnarray}
904: v_y = \frac{B_y}{B_x}v_x = \frac{q_y jV}{B_x^2/4\pi - j^2V}~,
905: \nonumber
906: \end{eqnarray}
907: \begin{equation}
908: p = q_x- j^2V - \frac{q_y^2 B_x^2/8\pi}{(B_x^2/4\pi -
909: j^2V)^2}~,~~~~ T = \frac{pV}{{\cal R}}~.
910: \label{4}
911: \end{equation}
912:
913: As mentioned we investigate stability of the structure
914: ``fast MHD shock wave" plus ``cooling zone".
915: Note that as the angle $\chi$
916: between the flow velocity and the normal to the
917: shock is decreased,
918: the stationary MHD structure does not convert to the
919: hydrodynamic one.
920: In the limit $\chi \to 0$,
921: $ q_x \approx {B_x^2}/(4 \pi \sigma^2)$.
922: The pressure in the radiative zone is determined from the
923: relation
924: \begin{equation}
925: p \approx \frac{B_x^2}{4 \pi \sigma^2} - j^2 V - \frac{q_y^2
926: B_x^2/8\pi}{(B_x^2/4 \pi - j^2 V)^2}~,
927: \label{5}
928: \end{equation}
929: where $q_y \propto \sin \chi$ is small.
930: For the considered conditions
931: where $\sigma < 1$, the denominator of the
932: fraction of the last term goes to zero before the
933: $ {B_x^2}/({4 \pi \sigma^2}) - j^2V$ goes to zero.
934: Thus the
935: pressure goes to zero for $B_x^2/4\pi - j^2V \approx 0$.
936: That is, the
937: velocity of the flow at the exit from the radiative zone
938: approaches the Alfv\'en velocity and is $v_x \approx -B_x^2/(4\pi j)$.
939: Taking into account equation (\ref{4}), we obtain from
940: equation (\ref{5}),
941: \begin{eqnarray}
942: 0 \approx \frac{B_x^2}{4 \pi \sigma^2} - \frac{B_x^2}{4 \pi} -
943: \frac{B_x^2}{8 \pi} \left( \frac{v_y}{B_x^2/4\pi j} \right)^2~.
944: %\nonumber
945: \end{eqnarray}
946: >From this relation we obtain
947: \begin{eqnarray}
948: v_y \approx - B_x \sqrt{\frac{1-\sigma^2}{2 \pi \rho_{\rm
949: in}}}~,~~~~ B_y \approx \frac{B_x}{\sigma} \sqrt{2 (1-\sigma^2)}~.
950: %\nonumber
951: \end{eqnarray}
952: Thus, as $\chi \to 0$ the components of the velocity
953: and magnetic field parallel to the shock
954: approach finite values.
955: In contrast, in the gas dynamic case these
956: components approach zero.
957: The observed behavior
958: is similar to that known to occur
959: in MHD switch-on shock waves where
960: finite tangential velocity and magnetic
961: field components are generated, and
962: where the velocity of the flow behind the front is Alfv\'enic
963: (Smith, 1993).
964: Note that the parallel MHD shock wave becomes
965: non-evolutionary, and it is replaced by a switch-on
966: shock wave for $
967: [ {v_x}/(B_x/\sqrt{4 \pi \rho})]_{\rm in} < 2$ for
968: $\gamma = 5/3$ and $ p_{\rm in} = 0$.
969:
970:
971: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
972: \begin{figure*}[t]
973: \epsscale{1.} \plotone{f4.eps} \caption{Time dependence of the shock
974: front coordinate $x_s$ (top panel) and luminosity $J$ (bottom panel)
975: for case I (``real" cooling function, $v_{\rm in} \cos \chi = 1.3
976: \times 10^7 {\rm cm~s^{-1}})$. The shock front coordinate is in
977: units $\Delta$, luminosity in units of $J_{\rm in}$, and time is in
978: seconds.} \label{Figure 4}
979: \end{figure*}
980: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
981: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
982: \begin{figure*}[b]
983: \epsscale{1.} \plotone{f5.eps} \caption{Distribution of different
984: parameters along the $x-$axis: velocity component along the field
985: lines (top panel), density and temperature (middle panel), magnetic
986: field and plasma parameter $\beta$ (low panel). All variables are
987: shown in dimensionless form. In the incoming flow $v_x=-1$,
988: $\rho=1$, $T=0$ (almost zero), $B_x=0.48$, $B_y=0.076$. Matter
989: inflows from the right boundary, a star is at the left.}
990: \label{Figure 5}
991: \end{figure*}
992: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
993: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
994: \begin{figure*}[t]
995: \epsscale{1.} \plotone{f6.eps} \caption{Time dependence of shock
996: front coordinate $x_s$ (top panel) and luminosity $J$ (bottom panel)
997: for case II (power-law cooling function, $v_{\rm in} \cos \chi = 1.3
998: \times 10^7 {\rm cm~s^{-1}})$. The coordinate of the shock front
999: $x_s$ is in units $\Delta$, the luminosity is in units of $J_{\rm
1000: in}$, and the time is in seconds.} \label{Figure 6}
1001: \end{figure*}
1002: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1003: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1004: \begin{figure*}[b]
1005: \epsscale{1.} \plotone{f7.eps} \caption{Time dependence of the shock
1006: front coordinate $x_s$ (top panel) and luminosity $J$ (bottom panel)
1007: for case III (``real" cooling function, $v_{\rm in} \cos \chi = 3
1008: \times 10^7 {\rm cm~s^{-1}}$. The shock front coordinate is in units
1009: $\Delta$, the luminosity is in units of $J_{\rm in}$, and the time
1010: is in seconds.} \label{Figure 7}
1011: \end{figure*}
1012: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1013: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1014: \begin{figure*}[t]
1015: \epsscale{1.} \plotone{f8.eps} \caption{Fourier amplitudes of shock
1016: front velocity $v_s$ (top panel) and luminosity $J$ (bottom panel)
1017: for case I. Velocity is in $v_{\rm in} \cos \chi$, luminosity is in
1018: $J_{\rm in}$, frequency is in ${\rm s}^{-1}$.} \label{Figure 8}
1019: \end{figure*}
1020: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1021: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1022: \begin{figure*}[b]
1023: \epsscale{1.} \plotone{f9.eps} \caption{Fourier amplitudes of shock
1024: front velocity $v_s$ (top panel) and luminosity $J$ for case II
1025: (cooling function $\Lambda_{\rm GS}$, $v_{\rm in} \cos \chi = 1.3
1026: \times 10^7 {\rm cm s^{-1}})$). The velocity is in units of $v_{\rm
1027: in} \cos \chi$, the luminosity is in units of $J_{\rm in}$, and
1028: frequency is in ${\rm s}^{-1}$.} \label{Figure 9}
1029: \end{figure*}
1030: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1031:
1032: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1033: \begin{figure*}[t]
1034: \epsscale{1.} \plotone{f10.eps} \caption{Fourier amplitudes of shock
1035: front velocity $v_s$ and luminosity $J$ for case III (cooling
1036: function $\Lambda_{\rm RTV}$ $v_{\rm in} \cos \chi = 3 \times 10^7
1037: {\rm cm s^{-1}})$). The velocity is in units of
1038: $v_{\rm in} \cos \chi$, the luminosity is in units of
1039: $J_{\rm in}$, and the frequency is in ${\rm s}^{-1}$.} \label{Figure
1040: 10}
1041: \end{figure*}
1042: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1043:
1044: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1045: \begin{figure*}[b]
1046: \epsscale{1.} \plotone{f11.eps} \caption{Time dependence of
1047: luminosity $J$ for case IV (cooling function $\Lambda_{\rm RTV},
1048: v_{\rm in} \cos \chi = 1.3 \times 10^7 {\rm cm~s^{-1}})$. The
1049: luminosity is in units of $J_{\rm in}$, and the time is in
1050: seconds.} \label{Figure 11}
1051: \end{figure*}
1052: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1053: \section{Method}
1054:
1055:
1056: We study the stability/instability of radiative
1057: shock waves in the presence of the
1058: magnetic field by integrating the
1059: time-dependent one-dimensional MHD equations in a region
1060: containing the shock and the radiative cooling zone.
1061: We used an Eulerian
1062: variables.
1063: Consequently, the simulation region is chosen large
1064: enough to contain
1065: both the shock wave and the remote part of the
1066: radiative zone where energy losses are negligibly small.
1067: The location of the right-hand
1068: boundary of the simulation region is
1069: chosen so that
1070: the shock wave did not leave the region during
1071: oscillations.
1072: For the calculations of the MHD flows we used
1073: a high resolution Godunov type numerical scheme (see, e.g.,
1074: Kulikovskii, Pogorelov \& Semenov 2000).
1075:
1076: For initial conditions we take the stationary flow with the shock
1077: wave and the radiation zone. At the top of the simulation region
1078: (see Figure 1) we determined either parameters of unperturbed flow
1079: (density, velocity, etc. ) if the incoming flow is super-Alfv\'enic,
1080: or we had ``free" boundary conditions, if incoming flow is
1081: sub-Alfv\'enic.
1082:
1083:
1084: At the bottom of the simulation region (closer to stellar surface),
1085: we have a ``layer" with no radiative losses. At this boundary we
1086: fixed the longitudinal velocity at a small value corresponding to
1087: the stationary solution. All other variables at the boundary had the
1088: same value as in the previous cell. Simulations have shown that
1089: results of modeling are not sensitive to the boundary conditions on
1090: the transverse components of velocity and magnetic field. At the
1091: boundary cell the density varied around some average value and the
1092: mass did not accumulate there. For example in case I the sound speed
1093: in this cell has been $(1-3)\times 10^6 {\rm cm~s^{-1}}$, the size
1094: of the grid $=81 {\rm cm}$, the sound-crossing time $<10^{-4} {\rm
1095: s}$. We observed from simulations that variation of this boundary
1096: did not influence much to the oscillations.
1097:
1098: The spatial resolution has been chosen so that the radiation zone is
1099: covered by 200 cells. The size of the simulation region has been
1100: chosen such that during the oscillations the shock wave stayed
1101: inside the simulation region. Depending on the amplitude of
1102: oscillations the simulation region incorporated from $500$ to
1103: $1,500$ cells. The time-step has been chosen automatically such that
1104: the Courant number is $0.5$.
1105:
1106:
1107:
1108:
1109: \section{Results}
1110:
1111: We investigate the stability/instability
1112: of the radiative shock wave as a
1113: function of two main
1114: dimensionless parameters: the inverse
1115: Alfv\'en Mach number
1116: of the unperturbed upstream flow $\sigma =
1117: M_A^{-1}$ and the inclination angle of the flow, $\chi$,
1118: relative to the normal to the star's surface
1119: The limit $\sigma = 0$ corresponds to zero magnetic field.
1120:
1121:
1122: We performed a series of calculations
1123: in which we varied these two
1124: parameters.
1125: Figure 3 summarizes the results.
1126: In the plane $(\sigma, \sin \chi)$ markers show the
1127: parameters where calculations were done.
1128: The ``squares"
1129: show the parameters for which the radiative shock wave is stable
1130: and the ``triangles" show cases where it is unstable.
1131: The solid straight
1132: line with a dashed continuation is the stability/instability
1133: boundary.
1134: The condition for stability can be expressed as
1135: \begin{equation}
1136: 3.7 \sigma + 1.4 \sin
1137: \chi \ge 1~,
1138: \end{equation}
1139: for conditions where the magnetic
1140: field is not very weak, $\sigma \ge 0.02$.
1141:
1142: For weak magnetic fields ($\sigma \le 0.02$),
1143: the boundary of stability is located close to the vertical axis in
1144: Figure 3.
1145: This part of the boundary is shown as a
1146: dashed line.
1147: Thus, in the absence of a magnetic field, the radiative shock
1148: wave (for the same parameters of the incoming flow) is unstable.
1149: However, even a small magnetic field, in particular, inclined one,
1150: stabilizes the radiative shock wave.
1151: For $\sigma \ge 0.02~~(M_A \le 50)$,
1152: we find stability for angles $\chi > 45^\circ$.
1153: If the magnetic field is sufficiently
1154: strong $\sigma \ge 0.27~~(M_A \le 3.7)$,
1155: then the radiative shock wave is stable for all angles $\chi$.
1156:
1157: We emphasize again that in the limit of small
1158: inclination angles, $\chi \to 0$, our model
1159: does not correspond to a
1160: gas-dynamical flow with unperturbed magnetic field.
1161:
1162: To understand the temporal
1163: characteristics of unstable radiative shock
1164: waves, we performed calculations using
1165: the ``real'' cooling function $\Lambda_{RTV}$
1166: and for some cases with the
1167: function $\Lambda_{\rm GS}(T)$.
1168: In all calculations we followed the
1169: location of the shock wave $x_s(t)$ and the total intensity of radiation
1170: from the radiative zone per unit area $J(t) = \int dx \rho^2 \Lambda$.
1171:
1172:
1173:
1174:
1175: Figure 4 shows results of calculation of evolution of the radiative
1176: shock wave with radiation function $\Lambda_{\rm RTV}$ for the
1177: following parameters of the incoming flow: $v_{\rm in} = 1.3 \times
1178: 10^7 {\rm cm~s^{-1}},~\rho_{\rm in} = 10^{-11} {\rm g
1179: cm^{-3}},~B_{\rm in} = 40.8 {\rm G},~\sin \chi = 0.154$.
1180: Also, $\sigma = 0.13$.
1181: The width of the cooling zone in the stationary regime is $\Delta
1182: \approx 1.64 \times 10^4 {\rm cm}$, and the energy-density
1183: (excluding the much smaller thermal energy) is $J_{\rm in} = (1/2)
1184: \rho_{\rm in}~v_{\rm in}^3 \cos \chi = 1.07 \times 10^{10} {\rm
1185: erg~cm^{-2}~s^{-1}})$.
1186: We point out once again that the incoming flow
1187: velocity is along the magnetic field, so that there is no
1188: Poynting flux.
1189:
1190:
1191: For the mentioned parameters, the stationary shock is unstable
1192: and the shock and radiative zone oscillate.
1193: The position of the
1194: front of the shock wave oscillates with a period $\approx 0.025$ s.
1195: The amplitude of oscillations of the front is modulated and varies
1196: with period $\approx 0.3 $ s.
1197: The top panel of Figure 4 shows the
1198: position of the front as a function of time at the time-interval
1199: approximately equal to twice period of modulation.
1200: The bottom panel of Figure 4 shows
1201: the temporal variation of the radiation
1202: intensity from the radiation zone
1203: for the same time-interval as the top panel.
1204: The position of the
1205: shock front is normalized to the width of the stationary cooling zone
1206: $\Delta$.
1207: The intensity of radiation is normalized to the energy-density in the
1208: incoming flow $J_{\rm in}$. The time is in seconds.
1209:
1210:
1211:
1212:
1213:
1214: Figure 5 shows spatial distribution of the longitudinal
1215: ($x-$direction) velocity (top panel), density and temperature
1216: (middle panel), longitudinal magnetic field and plasma parameter
1217: $\beta=8\pi p/({B_x^2+B_y^2})$ (bottom panel) at a time
1218: corresponding to the maximum distance of the shock from the surface
1219: of the star.
1220: In stationary regime the shock wave is located at $x=0$,
1221: while the radiative zone is below this. See animations at
1222: http://www.astro.cornell.edu/us-rus/shock.htm
1223:
1224: Figure 6 shows results of calculation of the evolution of the
1225: radiative shock wave for the same parameters in the incoming flow
1226: but with the radiation function $\Lambda_{\rm GS}$ (case II).
1227: The width
1228: of the cooling zone in the stationary regime is $\Delta \approx 2.65
1229: \times 10^4 {\rm cm}$.
1230: This figure shows position of the front of the shock wave $x_s(t)$ (top
1231: panel) and intensity of radiation $J(t)$ (bottom panel).
1232: One can see
1233: that qualitatively the oscillations are similar.
1234: However,
1235: the main frequency of oscillations is different: $\omega_{\rm RTV}
1236: \approx 240 {\rm s^{-1}},$ whereas $ \omega_{\rm GS} \approx 140 {\rm
1237: s^{-1}}$.
1238: With the radiation function $\Lambda_{\rm
1239: GS}$, the modulation of the oscillation amplitude of the position of
1240: the front is small.
1241:
1242:
1243:
1244: Figure 7 shows results of calculation of evolution of the radiative
1245: shock wave with radiation function $\Lambda_{\rm RTV}$ for the
1246: following parameters of the incoming flow (case III): $v_{\rm in} =
1247: 3 \times 10^7 {\rm cm~s^{-1}},~\rho_{\rm in} = 10^{-11} {\rm g
1248: cm^{-3}},~B_{\rm in} = 94.9 {\rm G},~\sin \chi = 0.154$. Also,
1249: $\sigma = 0.13$. The width of the cooling zone in the stationary
1250: regime is $\Delta \approx 8.86\times10^5 {\rm cm}$, and the
1251: energy-density (excluding the small thermal energy) is $1.35 \times
1252: 10^{11} {\rm erg~cm^{-2}~s^{-1})}$.
1253: As in case I for these parameters the
1254: stationary shock is unstable and the shock oscillates.
1255: The position of the front of the shock wave
1256: oscillates with period $\approx 0.21 {\rm s}$.
1257: The top panel shows the
1258: position of the front as a function of time,
1259: and the bottom panel shows the
1260: variation of the radiation intensity
1261: for the same time interval as the top panel.
1262:
1263: In case III we increased both the velocity and magnetic field with
1264: the aim of checking our hypothesis, that the qualitative solution of
1265: the problem (while using the ``real" radiation function
1266: $\Lambda_{\rm RTV}$) is determined by the dimensionless parameters
1267: of the problem, $\sin \chi$, and the magnetization, $\sigma$, and
1268: not by dimensional values of velocity of the incoming flow and the
1269: magnetic field.
1270:
1271:
1272: Figure 8 shows results of the Fourier-analysis of the speed of the
1273: shock front and total luminosity per unit of area $v_s(\omega)$ and
1274: $J(\omega)$. One can see that when modelling with radiation function
1275: $\Lambda_{\rm RTV}$ (case I) then two nearby maxima of the Fourier
1276: amplitudes are observed at the frequencies $\omega_1 = 240 {\rm
1277: s^{-1}},~ \omega_2 = 260 {\rm s^{-1}}$.
1278: We suggest that the combination of these two frequencies
1279: gives the amplitude modulation evident in Figure 4.
1280:
1281:
1282: Figure 9 shows the Fourier amplitudes for case II where we use
1283: $\Lambda_{\rm GS}$.
1284: The Fourier-amplitudes for both, $v_s$, and $J$
1285: have sharp maxima at frequencies divisible by the main frequency.
1286:
1287: Figure 10 shows the Fourier amplitudes for case III where we use
1288: $\Lambda_{\rm RTV}$. The Fourier-amplitudes for both, $v_s$, and $J$
1289: have sharp maxima at frequencies divisible by the main frequency.
1290: The highest peak in Figures 8-10 corresponds to the main oscillation
1291: frequency of the shock.
1292: The other peaks correspond to higher harmonics due
1293: the oscillations being anharmonic.
1294:
1295:
1296:
1297: We also investigated the stable regime of the radiative shock wave
1298: (case IV). To study the damping of the oscillations we introduced a
1299: small ($10\%$) perturbation of the velocity in the upstream flow
1300: during a limited time. We used the ``real" radiative function, and
1301: parameters as shown in the Table for case IV. One can see that when
1302: perturbations reached the front of the shock wave, the shock wave
1303: began to oscillate. However, these oscillations damped during
1304: several periods and stationary flow was
1305: re-established. Figure 11 shows an example of such
1306: damping.
1307:
1308:
1309:
1310:
1311: \section{Conclusions}
1312:
1313: This work has studied the
1314: stability/instability of the radiative
1315: MHD shock waves in the funnel
1316: streams of classical T~Tauri stars.
1317: Matter flowing to the surface
1318: of the star along the magnetic field is decelerated in the
1319: radiative shock wave.
1320: The shock may be stable or unstable.
1321: In the case of instability the shock position and other
1322: variables oscillate and this can give observable
1323: short time-scale variability in the emitted radiation.
1324: A significant new aspect of the present work is
1325: that the magnetic field and the flow velocity parallel to it can
1326: have an arbitrary angle with respect to the normal to the star's
1327: surface.
1328:
1329: The shock wave has been modeled by solving the
1330: time-dependent MHD equations in one dimension (perpendicular
1331: to the star's surface) taking into account the radiative losses.
1332: For the radiative losses we used either the ``real"
1333: radiative function, approximated by segments of power laws
1334: (RTV) or by the power law function proposed by GS07.
1335:
1336: Results of modelling of the radiative shock waves show that there is
1337: a simple criterion of the shock stability: $3.7 \sigma + 1.4 \sin
1338: \chi > 1$. This is for the case where the inflow to the shock is
1339: $v_{\rm in} \cos \chi = 1.3 \times 10^7 {\rm cm~s^{-1}}$.
1340: We believe that this criterion will not
1341: change significantly at larger inflow velocities.
1342:
1343: Comparison of the simulation results with the ``real" (RTV) and power-law
1344: (GS) radiative functions shows that the qualitative results are
1345: similar. However, the periods of oscillations are significantly
1346: different.
1347:
1348: The periods of oscillations are of the order of hundredths of a
1349: second for $v_{\rm in} \cos \chi = (1.3-3.0) \times 10^7 {\rm
1350: cm~s^{-1}}$.
1351: This period is expected to increase with $v_{\rm
1352: in}$. We estimate that $ P \approx 6 \times 10^{-3} [(v_{\rm in}
1353: \cos \chi)/(10^7 {\rm cm~s^{-1}})]^3 {\rm s}$. The period of the
1354: oscillations varies, $P=0.02-0.2$, depending on parameters.
1355:
1356:
1357: Global three-dimensional simulations of magnetospheric accretion
1358: through the funnel streams have shown that hot spots on the surface
1359: of the star are {\it not homogeneous}: most of the kinetic energy
1360: flows in the central regions of the funnel stream so that the
1361: central regions of the spots are expected to be hotter (and also
1362: denser) compared to peripheral regions (Romanova et al. 2004;
1363: Kulkarni \& Romanova 2005). Romanova et al. (2004) have shown that
1364: the spots may have very small filling factor at highest density and
1365: temperature (less than $1\%$) and much larger filling factor at
1366: smaller densities and temperatures (see Fig. 3 of Romanova et al.
1367: 2004). This fact has been recently confirmed observationally by
1368: G\"unther et al. (2007) who have shown that the filling factor in
1369: X-ray is smaller compared with UV and optical bands which confirmed
1370: the inhomogeneity of the hot spots. Future research should be done
1371: for analysis of the stability of the global shock wave which would
1372: cover a significant part of the hot spot (not a small part as
1373: usually considered including this paper).
1374:
1375: If the magnetic field near the star has a complex geometry (e.g.
1376: Valenti \& Johns-Krull 2004; Gregory et al. 2006; Donati et al.
1377: 2007; Long, Romanova \& Lovelace 2007, 2008), then it is likely
1378: that some field lines are inclined to the surface of the star as
1379: considered in this paper. The present analysis is thus applicable
1380: to the stability/instability of the shocks. If the complex field
1381: has significant transverse component then it may suppress
1382: oscillations.
1383:
1384: Kravtsova \& Lamzin (private communication) searched for
1385: oscillations of the shock in RW Aur using Crimean Observatory
1386: facilities. They did not find oscillations, though their
1387: time-resolution was low ($\Delta P=0.5 - 1 {\rm s}$. They concluded
1388: that in this star there are no oscillations with periods $P > 2 {\rm
1389: s}$ with amplitudes $> 5\%$ above the noise level. Higher
1390: time-resolution observations in a larger sample of CTTSs are needed
1391: to obtain a conclusive answer.
1392: It would be useful to
1393: have high ({\rm ms}) time-resolution observations in the UV and
1394: X-ray bands in stars with high veiling, such as RW Aur and
1395: others, because these wavebands would correspond to oscillations of
1396: the central part of the hot spots (Romanova et al. 2004; G\"unter et
1397: al. 2007) which may go into global oscillation mode with higher
1398: probability compared to peripheral parts observed in the optical
1399: band.
1400: A search
1401: in the optical band may bring interesting results as well because we
1402: do not know the details of the interaction of the funnel stream
1403: with a star on a global scale of the size of hot spot. To understand
1404: such physics, observations of variability time-scales would be
1405: informative and may help to shed a light to process of the
1406: funnel-star interaction.
1407:
1408:
1409: \section*{Acknowledgements}
1410:
1411: Authors thank Dr. Lamzin and Dr. Beskin for helpful discussions and
1412: an anonymous referee for multiple questions and comments which
1413: improved the paper. This work has been partially supported by the
1414: NSF grants AST-0507760 and AST-0607135, and NASA grants NNG05GG77G
1415: and NAG5-13060. A.V.K. and G.V.U. were partially supported by RFBR
1416: grant 06-02-16608.
1417:
1418: \begin{thebibliography}{}
1419: \bibitem{} Ardila, D., \& Basri, C. 2000, ApJ, 539, 834
1420: \bibitem{} Bertschinger E. 1986, ApJ, 304, 154
1421: \bibitem{} Bessolaz, N., Zanni, C., Ferreira, J., Keppens, R., Bouvier,
1422: J. 2008, A\&A, 478, 155
1423: \bibitem{} Calvet, N., Gullbring, E. 1998, ApJ 509, 802
1424: \bibitem{} Canalle, J. B. G., Saxton, C. J., Wu, K.,
1425: Cropper, M., Ramsay, G. 2005, A\&A, 440, 185
1426: \bibitem{} Chevalier R.A., Imamura J.N. 1982, ApJ, 261, 543
1427: %\bibitem{} Gahm, G.F. 1990, IAUS, 137, 193G
1428: \bibitem{} Gnat O., Sternberg A. 2007, ApJ SS, 168, 213 (GS07)
1429: \bibitem{} Gregory, S.G., Jardine, M., Simpson,I.,
1430: Donati, J.-F. 2006, MNRAS 371, 999
1431: \bibitem{} Cropper et al 1999 MNRAS 306, 664
1432:
1433: \bibitem{} Gullbring, E., Calvet, N., Muzerolle, J., Hartmann, L.
1434: 2000, ApJ, 544, 927
1435: \bibitem{} G\"unther, H.M., Schmitt, J.H.M.M., Robrade, J., \& Liefke
1436: 2007, A\&A, 466, 1111
1437: \bibitem{} G\"unther, H.M., \& Schmitt, J.H.M.M., 2008, arXiv:0801.2273v2
1438: \bibitem{} Imamura, J.N., Aboasha, A., Wolff, M.T., Wood, K.S.
1439: 1996 ApJ 458, 327
1440: \bibitem{} Kulikovskii, A.G., Pogorelov, N.V., \& Semenov, A.Yu.
1441: 2000, {\it Mathematical Aspects of Numerical Solution of Hyperbolic
1442: Systems}, Publisher: Chapman \& Hall/CRC
1443: \bibitem{} Langer S., Chanmugan G., Shaviv G. 1981, ApJ, 245, L23
1444: \bibitem{} Lamzin, S.A. 1995, A\&A, 295, L20
1445: \bibitem{} Lamzin, S.A. 1998, Astron. Reports 42, 322
1446: \bibitem{} Long M., Romanova M.M., Lovelace R.V.E. 2005, ApJ, 634, 1214
1447: \bibitem{} Long, M., Romanova, M.M. \& Lovelace, R.V.E. 2007, MNRAS
1448: \bibitem{} Long, M., Romanova, M.M. \& Lovelace, R.V.E. 2008, MNRAS,
1449: arXiv:0802.2308
1450: \bibitem{} Mignone A. 2005, ApJ, 626, 373
1451: \bibitem{} Peres G., Rosner R., Serio S., Vaiana G.S. 1982, ApJ, 252, 791
1452: \bibitem{} Ramachandran B., Smith M.D. 2005, MNRAS, 357, 707
1453: \bibitem{} Ramachandran B., Smith M.D. 2005, MNRAS,362, 1353 (RS05)
1454: \bibitem{} Ramachandran B., Smith M.D. 2006, MNRAS, 366, 586
1455: \bibitem{} Raymond J.C., Smith B.W. 1977, ApJ SS,35, 419
1456: \bibitem{} Romanova M.M., Ustyugova G.V., Koldoba A.V., Lovelace R.V.E. 2002, ApJ, 578, 420
1457: \bibitem{} Romanova M.M., Ustyugova G.V., Koldoba A.V., Lovelace R.V.E. 2004, ApJ,
1458: 610, 920
1459: \bibitem{} Rosner R., Tucker W.H., Vaiana G.S. 1978, ApJ, 220, 643
1460: \bibitem{} Saxton, C.J. 1998, MNRAS 299, 862
1461: \bibitem{} Saxton, C.J. 2002, PASA, 19, 282
1462: \bibitem{} Smith, K.W., Jones, D.H.P., \& Clarke, C.J. 1996, MNRAS,
1463: 282, 167
1464: \bibitem{} Smith M.D. 1989, MNRAS, 238, 235
1465: \bibitem{} Smith M.D. 1993, AA, 272, 571
1466: \bibitem{} Smith, K.W., Jones, D.H.P., \& Clarke, C.J. 1996, MNRAS,
1467: 282, 167
1468: \bibitem{} Spitzer, L. 1968, {\it Diffuse Matter in Space},
1469: Interscience Pub.: New York, ch. 2
1470: \bibitem{} Strickland, R. \& Blondin, J.M. 1995, ApJ 449, 727
1471: \bibitem{} Sutherland, R.S., Bicknell, G.V., Dopita, M.A.
1472: 2003, ApJ 591, 238
1473: \bibitem{} Toth G., Draine B.T. 1993, ApJ, 413, 176 (TD93)
1474: \bibitem{} Valenti, J.A., \& Johns-Krull, C.M. 2004, Ap\& SS, 292,
1475: 619
1476:
1477: \end{thebibliography}
1478:
1479: \end{document}
1480: