1: %% The command below calls the preprint style
2: %% which will produce a one-column, single-spaced document.
3: %% Examples of commands for other substyles follow. Use
4: %% whichever is most appropriate for your purposes.
5: %%
6: %\documentclass[12pt,referee]{aastex}
7:
8: %% manuscript produces a one-column, double-spaced document:
9:
10: %\documentclass[manuscript]{aastex}
11:
12: %% preprint2 produces a double-column, single-spaced document:
13:
14: \documentclass[12pt,preprint]{aastex}
15:
16: %% Sometimes a paper's abstract is too long to fit on the
17: %% title page in preprint2 mode. When that is the case,
18: %% use the longabstract style option.
19:
20: %% \documentclass[preprint2,longabstract]{aastex}
21:
22: %% If you want to create your own macros, you can do so
23: %% using \newcommand. Your macros should appear before
24: %% the \begin{document} command.
25: %%
26: %% If you are submitting to a journal that translates manuscripts
27: %% into SGML, you need to follow certain guidelines when preparing
28: %% your macros. See the AASTeX v5.x Author Guide
29: %% for information.
30:
31: \def\FLASH {{\sc flash}}
32: \def\PARAMESH {{\sc paramesh}}
33: %\newcommand{\orlando}{\textbf}
34: \newcommand{\orlando}{}
35: %\newcommand{\orlandobis}{\textbf}
36: \newcommand{\orlandobis}{}
37:
38: \newcommand{\vdag}{(v)^\dagger}
39: \newcommand{\myemail}{skywalker@galaxy.far.far.away}
40:
41: %% You can insert a short comment on the title page using the command below.
42:
43: %\slugcomment{Not to appear in Nonlearned J., 45.}
44:
45: %% If you wish, you may supply running head information, although
46: %% this information may be modified by the editorial offices.
47: %% The left head contains a list of authors,
48: %% usually a maximum of three (otherwise use et al.). The right
49: %% head is a modified title of up to roughly 44 characters.
50: %% Running heads will not print in the manuscript style.
51:
52: \shorttitle{Non-equilibrium of ionization in nanoflaring loops}
53: \shortauthors{Reale \& Orlando}
54:
55: %% This is the end of the preamble. Indicate the beginning of the
56: %% paper itself with \begin{document}.
57:
58: \begin{document}
59:
60: %% LaTeX will automatically break titles if they run longer than
61: %% one line. However, you may use \\ to force a line break if
62: %% you desire.
63:
64: \title{Non-equilibrium of Ionization and the Detection of Hot
65: Plasma in Nanoflare-heated Coronal Loops}
66:
67: \author{Fabio Reale\altaffilmark{1}}
68: \affil{Dipartimento di Scienze Fisiche \& Astronomiche, Universit\`a di
69: Palermo, Sezione di Astronomia, Piazza del Parlamento 1, 90134 Palermo,
70: Italy}
71: \author{Salvatore Orlando}
72: \affil{INAF - Osservatorio Astronomico di Palermo ``G.S.
73: Vaiana'', Piazza del Parlamento 1, 90134 Palermo, Italy}
74:
75: % Notice that each of these authors has alternate affiliations, which
76: % are identified by the \altaffilmark after each name. The actual alternate
77: % affiliation information is typeset in footnotes at the bottom of the
78: % first page, and the text itself is specified in \altaffiltext commands.
79: % There is a separate \altaffiltext for each alternate affiliation
80: % indicated above.
81:
82: \altaffiltext{1}{INAF - Osservatorio Astronomico di Palermo ``G.S.
83: Vaiana'', Piazza del Parlamento 1, 90134 Palermo, Italy}
84:
85: % The abstract environment prints out the receipt and acceptance dates
86: % if they are relevant for the journal style. For the aasms style, they
87: % will print out as horizontal rules for the editorial staff to type
88: % on, so long as the author does not include \received and \accepted
89: % commands. This should not be done, since \received and \accepted dates
90: % are not known to the author.
91:
92: \begin{abstract}
93: Impulsive nanoflares are expected to transiently heat the plasma confined
94: in coronal loops to temperatures of the order of 10 MK. Such hot plasma
95: is hardly detected in quiet and active regions, outside flares. During
96: rapid and short heat pulses in rarified loops the plasma can be highly out of
97: equilibrium of ionization.
98: Here we investigate the effects of
99: the non-equilibrium of ionization (NEI) on the detection of hot
100: plasma in coronal loops.
101: Time-dependent loop hydrodynamic simulations
102: are specifically devoted to this task, including saturated thermal
103: conduction, and coupled to the detailed solution of the equations of
104: ionization rate for several abundant elements. In our simulations,
105: initially cool and rarified magnetic flux tubes are heated to 10 MK
106: by nanoflares deposited either at the footpoints or at the loop apex.
107: We test for different pulse durations, and find that, due to NEI effects,
108: the loop plasma may never be detected at temperatures above $\sim 5$ MK for
109: heat pulses shorter than about 1 min.
110: We discuss some implications in the framework of
111: multi-stranded nanoflare-heated coronal loops.
112: \end{abstract}
113:
114: % The different journals have different requirements for keywords. The
115: % keywords.apj file, found on aas.org in the pubs/aastex-misc directory,
116: % contains a list of keywords used with the ApJ and Letters. These are
117: % usually assigned by the editor, but authors may include them in their
118: % manuscripts if they wish.
119:
120: \keywords{Sun: corona - Sun: X-rays}
121:
122: % That's it for the front matter. On to the main body of the paper.
123: % We'll only put in tutorial remarks at the beginning of each section
124: % so you can see entire sections together.
125:
126: %________________________________________________________________
127: \section{Introduction}
128: \label{sec1}
129:
130: Nanoflares -- small scale highly transient heating episodes --
131: are among the main candidates as source of coronal loop heating
132: (e.g., \citealt{1988ApJ...330..474P}; \citealt{1994ApJ...422..381C};
133: \citealt{2006SoPh..234...41K} and references therein). The conjecture is still under debate
134: because nanoflares have been hardly detected so far. There are
135: many possible reasons for this difficult detection. For instance,
136: a very frequent occurrence may inhibit the resolution of the single
137: event. Also the efficient thermal conduction and low emission measure
138: during the pulses can reduce and delay the signatures of the heating
139: (e.g., \citealt{1987ApJ...312..895P}; \citealt{1995A&A...299..225R}). Also very small
140: pulses distributed in a very finely structured loop may be difficult
141: to detect. One of the main arguments invoked as crucial evidence of
142: nanoflare heating is the detection of high temperature ($10$ MK)
143: plasma components in observations of coronal loops (\citealt{1997ApJ...478..799C}; \citealt{2006SoPh..234...41K}), outside of
144: proper flares. The presence of such hot component is often predicted
145: by hydrodynamic modeling of coronal loops heated by transient pulses
146: (\citealt{2005ApJ...628.1023P}; \citealt{2006ApJ...647.1452P}; \citealt{2004ApJ...605..911C}). In particular any
147: heat spike able to bring the loop to the observed brightness should
148: be so intense as to heat the plasma to temperatures of the order of
149: $10$ MK at least for a transient time interval. In the hypothesis of a
150: finely structured loop where a whole distribution of heat pulses occur
151: continuously, such plasma might be detectable, at least as a hot tail in
152: the emission measure distribution. The evidence of hot plasma has been
153: difficult so far. For instance, a dominant plasma component at about 3 MK
154: is shown by recent thermal maps of active regions obtained from wide-band
155: multi-filter imaging observations (\citealt{2007Sci...318.1582R}) with
156: the X-Ray Telescope (XRT, \citealt{2007SoPh..243...63G}) on board the
157: Hinode mission (\citealt{2007SoPh..243....3K}).
158:
159: There are possible explanations also for the difficult detection of hot plasma. The most immediate one
160: is linked to the inertia of the plasma dynamics. A heat pulse deposited
161: in a coronal loop drives evaporation of chromospheric plasma. The loop
162: is filled with hot and dense plasma which makes it bright in the X-rays.
163: For short heat pulses, the plasma may evaporate from the
164: chromosphere on time scales longer than the heat duration. Therefore
165: the heated strand may only later acquire enough emission measure to become
166: visible, while it is cooling. However, in short loops ($\sim 10^9$
167: cm) the sound crossing time is of the order of one minute (e.g.,
168: \citealt{2007A&A...471..271R}) and we may expect to detect hot spots
169: within less than half a minute.
170:
171: There is another less obvious effect which may make the detection of hot
172: plasma harder: the time lag of the plasma to change its ionization
173: from a cool to a hot state.
174: An impulsive energy input drives plasma thermal and dynamic changes on
175: relatively
176: short timescales. Electron
177: excitation, de-excitation, ionization and recombination processes of the
178: ion species have
179: other timescales. If the timescale, for instance,
180: of the temperature evolution is much shorter than the ionization and
181: recombination timescales, the degree of ionization can be very
182: different from the equilibrium conditions corresponding to the local
183: electron temperature (\citealt{1977ApJ...217..621S}; \citealt{1999A&A...346.1003O}; \citealt{2006SoPh..234...41K}). So, during a fast
184: temperature increase, the plasma ions can be at a lower ionization state
185: than the equilibrium state corresponding to the instantaneous temperature.
186: Such non-equilibrium of ionization (NEI) effects may become important in
187: the interpretation of what we observe. If the heat pulse is sufficiently
188: short, the ions and their emitted radiation may not even have enough
189: time to ``sense" the hot temperature status, they would adjust to the
190: temperature variations deep in the cooling phase, and we would detect
191: radiation from cooler ion conditions at any time.
192:
193: NEI effects in nanoflare-heated loops have already been investigated in
194: the past. \cite{1982ApJ...255..783M} found that NEI can significantly alter
195: the relative ionic abundances in the quiet corona. \cite{1989SoPh..122..245G}
196: examine the effect of NEI on the observability of coronal variations.
197: \cite{2003A&A...407.1127B}
198: found that NEI can modify considerably the radiative losses
199: function. \cite{2003A&A...411..605M} investigated the role of NEI in the
200: transition region brightenings driven by loop condensations.
201: \cite{2006A&A...458..987B} model nanoflare heating in coronal loops
202: and remark the importance of NEI effects, the presence of hot plasma with
203: low emission measure, and Doppler-shifts as possible diagnostics of the heating.
204:
205: Here we investigate the effect of NEI during nanoflaring activity on
206: the detectability of hot plasma in coronal loops. We will
207: model a coronal loop strand heated by nanoflares of different durations,
208: compute the corresponding evolution of the relevant ion species, and
209: compare it with the evolution in full ionization equilibrium.
210: We will evaluate the effects on the expected temperature distribution of the
211: loop emission measure, which will tell us about the existence of
212: significantly emitting hot plasma components.
213: This will allow us to put constraints on the nanoflare characteristics
214: which lead to the observability of hot plasma and indirectly also
215: on the fine loop structuring.
216: In Sec.\ref{sec:model} the modeling is described, in Sec.\ref{sec:result} the results are illustrated and discussed in Sec.\ref{sec:discus}.
217:
218:
219: \section{Modeling}
220: \label{sec:model}
221:
222: %\subsection{The Model}
223:
224: Our model is set up to explore the conditions for plasma confined in a
225: coronal magnetic flux tube and heated to $10$ MK by nanoflares to be
226: detectable as hot plasma considering the effects of NEI.
227: The coronal loop contains low $\beta$ plasma, and, as customary for
228: standard loop models, we assume that plasma moves and transports energy
229: only along the magnetic field lines, so that a one-dimensional description
230: is adequate (e.g., \citealt{1982ApJ...252..791P}). The model takes into
231: account the gravity stratification, the thermal conduction (including
232: the effects of heat flux saturation), the radiative losses, an external
233: heating input, and the NEI effects. The following fluid equations
234: of mass, momentum, and energy conservation are solved, considering
235: only the relevant components along the loop magnetic field lines:
236:
237: \begin{equation}
238: \frac{\partial \rho}{\partial t} + \nabla \cdot \rho \mbox{\bf v} = 0
239: \label{eq1}
240: \end{equation}
241:
242: \begin{equation}
243: \frac{\partial \rho \mbox{\bf v}}{\partial t} +\nabla \cdot \rho
244: \mbox{\bf vv} + \nabla P = \rho\mbox{\bf g}
245: \end{equation}
246:
247: \begin{eqnarray}
248: \lefteqn{\frac{\partial \rho E}{\partial t} +\nabla\cdot (\rho
249: E+P)\mbox{\bf v} =} \nonumber \\
250: & \rho \mbox{\bf v}\cdot \mbox{\bf g} -\nabla\cdot q
251: + Q(s, t)-n_e n_H \Lambda(T)
252: \end{eqnarray}
253:
254: \[
255: \mbox{where \hspace{0.5cm}} E = \epsilon +\frac{1}{2} |\mbox{\bf
256: v}|^2~,
257: \]
258:
259: \noindent
260: is the total gas energy (internal energy, $\epsilon$, and kinetic
261: energy), $t$ is the time,
262: $s$ is the coordinate along the loop,
263: $\rho = \mu m_H n_{\rm H}$ is the mass density,
264: $\mu = 1.26$ is the mean atomic mass (assuming solar abundances),
265: $m_H$ is the mass of the hydrogen atom, $n_{\rm H}$ is the hydrogen
266: number density, $n_{\rm e}$ is the electron number density, {\bf v}
267: the plasma flow speed, $P$ the pressure, {\bf g} the gravity,
268: $T$ the temperature, $q$ the conductive flux,
269: $Q(s,t)$ a function describing the transient input heating, $\Lambda(T)$
270: is the radiative losses per unit emission measure (e.g. \citealt{rs77};
271: \citealt{mgv85}; \citealt{2000adnx.conf..161K}). Here we consider the
272: radiative losses function in equilibrium of ionization. Although this
273: assumption makes our
274: description not entirely self-consistent, it does not affect
275: significantly our results,
276: because in the critical evolution phases the energy losses are dominated
277: by transport by thermal conduction (see Sec.\ref{sec:sim}). We use the ideal gas
278: law, $P=(\gamma-1) \rho \epsilon$, where $\gamma=5/3$ is the ratio of
279: specific heats.
280:
281: \orlando{The set of the continuity equations for each ion species is
282: expressed as:}
283:
284: \begin{equation}
285: \frac{\partial n_i^Z}{\partial t} + \nabla \cdot n_i^Z \mbox{\bf v} =
286: R_i^Z ~~~~~~~~\begin{array}{l}(Z = 1, ..., N_{elem})\\\\
287: (i = 1, ..., N_{ion}^Z) \end{array}
288: \label{eq4}
289: \end{equation}
290:
291: \noindent
292: \[
293: \mbox{where}~~~R_i^Z = n_e [n_{i+1}^Z\alpha_{i+1}^Z + n_{i-1}^Z
294: S_{i-1}^Z - n_i^Z(\alpha_i^Z+S_i^Z)]~,
295: \]
296:
297: \noindent
298: \orlando{$n_{i}^Z$ is the number density of the $i$-th ion of the element $Z$,
299: $N_{elem}$ is the number of elements, $N_{ion}^Z$ the number of ionization
300: states of element $Z$, $\alpha_i^Z$ are the collisional and dielectronic
301: recombination coefficients, and $S_i^Z$ the collisional ionization
302: coefficients (\citealt{summers74}).}
303:
304: Given the importance of rapid transients, fast dynamics and steep
305: thermal gradients in this work, we consider both the classical and saturated
306: conduction regimes.
307: To allow for a smooth transition between them,
308: we follow \citet{1993ApJ...404..625D} and define
309: the conductive flux as
310:
311: \begin{equation}
312: q = \left(\frac{1}{q_{\rm spi}}+\frac{1}{q_{\rm sat}}\right)^{-1}~.
313: \end{equation}
314:
315: \noindent
316: Here $q_{\rm spi}$ represents the classical conductive flux
317: (\citealt{spi62})
318:
319: \begin{equation}
320: q_{\rm spi} = -\kappa(T)\nabla T
321: \label{spit_eq}
322: \end{equation}
323:
324: \noindent
325: where the thermal conductivity is $\kappa(T) = 9.2\times 10^{-7}
326: T^{5/2}$
327: erg s$^{-1}$ K$^{-1}$ cm$^{-1}$. The saturated flux,
328: $q_{\rm sat}$, is (\citealt{cm77})
329:
330: \begin{equation}
331: q_{\rm sat} = -\mbox{sign}\left(\nabla T\right)~ 5\phi \rho c_{\rm
332: s}^3,
333: \label{therm}
334: \end{equation}
335:
336: \noindent
337: where $c_{\rm s}$ is the isothermal sound speed, and $\phi$ is a correction
338: factor
339: of the order of unity. We set $\phi = 1$ according to the values
340: suggested for the coronal plasma (\citealt{1984ApJ...277..605G};
341: \citealt{1989ApJ...336..979B}, \citealt{2002A&A...392..735F}, and
342: references therein).
343:
344: The transient input heating is described empirically as a separate
345: function of space
346: and time:
347:
348: \begin{equation}
349: Q(s,t) = H_0 \times g(s) \times f(t)
350: \end{equation}
351:
352: \noindent
353: where $g(s)$ is a Gaussian function:
354:
355: \begin{equation}
356: g(s) = \exp[-(s-s_0)^2/2\sigma^2]~,
357: \end{equation}
358:
359: \noindent
360: and $f(t)$ is a pulse function:
361:
362: \begin{equation}
363: f(t) = \left\{\begin{array}{ll}
364: 0, & t \leq 0 \\ \\
365: 1, & 0 <t\leq t_H \\ \\
366:
367: 0, & t > t_H
368: \end{array}\right.
369: \end{equation}
370:
371: The calculations described in this paper were performed using the 1-D version
372: of the \FLASH\
373: code (\citealt{for00}), an adaptive mesh refinement multiphysics code. For
374: the present application, the code has been extended by additional
375: computational modules to handle the plasma thermal conduction
376: (see \citealt{2005A&A...444..505O} for details of the
377: implementation), the NEI effects, the radiative losses, and the
378: heating function. The implementation of the description of the ionization
379: balance in the \FLASH\ code is described in the Appendix.
380:
381: %\subsection{Initial and Boundary Conditions}
382:
383: We consider an initially cool and rarified semicircular
384: magnetic flux tube (loop or loop strand)
385: with half-length $L = 3 \times 10^9$ cm, and uniform cross-section
386: area. The area appears only as a multiplicative factor in this description
387: and the size is typical of active region loops. The loop is initially
388: at equilibrium according to loop scaling laws (\citealt{1978ApJ...220..643R})
389: and to hydrostatic conditions
390: (\citealt{1981ApJ...243..288S})
391: assuming it lies on a plane vertical to the solar
392: surface.
393: The initial base pressure is $p_0 = 0.055$
394: dyn/cm$^2$, corresponding to a maximum temperature $T_0 = 0.79$ MK, at
395: the loop apex. The corona is linked with a steep transition region to
396: an isothermal chromosphere uniformly at $T_c = 20000$ K and $0.5 \times
397: 10^9$ cm thick. We assume that the loop is symmetric with respect
398: to the vertical axis across the apex, and therefore we simulate only half of
399: the loop, with a total extension of
400: $3.5 \times 10^9$ cm.
401: \orlando{All the ion species are assumed initially in
402: equilibrium of ionization in the whole computational domain.}
403:
404: \orlando{At the coarsest resolution, the adaptive mesh algorithm used
405: in the \FLASH\ code (\PARAMESH; \citealt{mom00}) uniformly covers the
406: computational domain with a mesh of $16$ blocks, each with $8$ cells.
407: We allow for 3 levels of refinement, with resolution increasing
408: twice at each refinement level. The refinement criterion adopted
409: (\citealt{loehner}) follows the changes in density and temperature. This
410: grid configuration yields an effective resolution of $\approx 3.4\times
411: 10^6$ cm at the finest level, corresponding to an equivalent uniform
412: mesh of $1024$ grid points. We use fixed boundary conditions at $s=0$
413: and reflecting boundary conditions at $s=s_{\rm max}$ (consistent with
414: the adopted symmetry).}
415:
416:
417: \section{Results}
418: \label{sec:result}
419:
420:
421: %\subsection{Simulation Strategy}
422: \subsection{The Simulations}
423: \label{sec:sim}
424:
425: In the loop outlined above we inject one nanoflare as intense as to heat the plasma
426: and keep it at $\approx 10$ MK, if the heating were steady. The choice of
427: the parameters is dictated by our scope of exploring the influence
428: of NEI on the detectable thermal conditions. In this perspective,
429: the pulse duration becomes the critical parameter: if the heat pulse
430: is short enough, the plasma may not have enough time to adjust to
431: ionization conditions appropriate of 10 MK, before the heating is
432: off. Therefore, we consider longer and longer heat pulse durations with a
433: logarithmic sampling, i.e. $t_H = 5$ s, 30 s, 180 s. For each of these
434: durations we consider two possible locations of the pulse depositions:
435: at the footpoints, namely $s_0 = 8 \times 10^8$ cm from the base of the
436: chromosphere (i.e. $3 \times 10^8$ cm from the base of the corona) and at
437: the apex, namely $s_0 = 3.5 \times 10^9$ cm. The spatial width of the
438: pulse is smaller at the footpoints ($\sigma = 10^8$ cm) and larger at
439: the apex (\orlando{$\sigma = 5\times 10^8$ cm}). To have the same total energy
440: deposition rate, the maximum volume deposition
441: rates are \orlando{$H_0 = 1.5$} erg cm$^{-3}$ s$^{-1}$ and \orlando{$H_0 =
442: 0.2$} erg cm$^{-3}$ s$^{-1}$, respectively.
443:
444: We come out with a total of six simulations. They are carried out for
445: a time which covers the pulse duration and a couple of loop cooling
446: times:
447:
448: \begin{equation}
449: \tau_{dec} = 120 \frac{L_9}{\sqrt{T_7}} \approx 360 ~~ \rm s
450: \end{equation}
451:
452: \noindent
453: where $L_9 = 3$ is $L$ in units of $10^9$ cm and $T_7 \approx 1$ is the
454: loop maximum temperature in units of $10$ MK (\citealt{1991A&A...241..197S}).
455: All simulations then span a time interval of about \orlando{800} s.
456:
457: %\subsection{Reference Simulation}
458:
459: The evolution of nanoflaring plasma confined in coronal loops is
460: well-known from previous work (\citealt{1993pssc.symp..151P};\citealt{2002ApJ...579L..41W};\citealt{2003ApJ...593.1174W};\citealt{2005ApJ...628.1023P};\citealt{2005ApJ...622..695T})
461: and is on a smaller scale similar to that
462: of properly flaring loops (e.g. \citealt{1980SoPh...68..351N};\citealt{1982ApJ...252..791P};\citealt{1983ApJ...265.1103D};\citealt{1984ApJ...279..896N};\citealt{1985ApJ...289..414F};\citealt{1987ApJ...312..895P};\citealt{1995A&A...299..225R}).
463: As additional feature, our simulations include the effect of saturated
464: thermal conduction, which might be important here since we study the
465: evolution on small time scales (see also \citealt{2003SPD....34.1006K}).
466: In the following, we just draw a basic outline of the results adapted to our simulations.
467:
468: Let's
469: consider the simulation with $t_H = 5$ s and heat pulses deposited at the
470: footpoints (Fig.~\ref{fig0}). The sudden heat deposition determines a
471: local increase of temperature (to $\sim 10$ MK) and pressure (to $\sim
472: 1$ dyn/cm$^2$). A fast thermal front propagates upwards along the
473: loop. The cool chromosphere is heated and expands upwards with a strong
474: evaporation front. By the time the thermal front has reached
475: $s\sim 1.5 \times 10^9$ cm and the evaporation front $s\sim 6\times10^8$
476: cm, the heat pulse is already over. Due to efficient thermal conduction,
477: the plasma then immediately begins to cool, already during the propagation of
478: the thermal front, and the maximum temperature rapidly decreases to $\sim
479: 2.5$ MK in less than a minute. The thermal front reaches the apex, and
480: therefore the loop thermalizes, in $\sim 30$ s. The impulsive evaporation
481: front moves at velocity of about 300 km/s (Fig.\ref{fig1}D)
482: and fills the loop in $\sim 75$
483: s. The density keeps on increasing throughout the loop with more gentle
484: fronts (the peak velocity rapidly decreases to less than 200 km/s) for
485: further tens of seconds. Meanwhile the plasma accumulates at the apex
486: reaching a density of about $10^{9}$ cm$^{-3}$. The compression heats
487: the plasma again above 3 MK at the apex. Then the plasma begins to drain
488: and the density to decrease (not shown), following the radiation cooling
489: time (\citealt{2004ApJ...605..911C}; \citealt{2007A&A...471..271R}).
490:
491: \begin{figure*}[!t]
492: \centering \includegraphics[width=15cm]{fig1.ps}
493: \caption{Evolution of the plasma temperature (A), density (B), pressure
494: (C), and velocity (D) distributions along the loop from the chromosphere
495: to the coronal loop apex, sampled every 5 s from 0 to
496: 100 s (thinner and thinner lines), for heat pulse location at the loop base and heat pulse
497: durations $t_H = 5$ s.}
498: \label{fig0}
499: \end{figure*}
500:
501: The overall evolution does not change much in the other
502: simulations. Longer heat pulses drive more and longer plasma
503: evaporation. Pulses deposited at the apex produce downward thermal fronts
504: but as soon as the thermal front hits the chromosphere, it drives an
505: evaporation front similar to that driven by the pulses at the footpoints.
506:
507: \subsection{Non-equilibrium of Ionization}
508: %\subsection{Ionization and Recombination Timescales}
509:
510: The importance of highly transient processes such as NEI is basically
511: dictated by their timescales related to the timescales of the dynamics
512: and heating/cooling driven by the nanoflares. Fig.~\ref{fig2} shows the
513: combined ionization/recombination timescale for the {\it i}-th ion species
514: of the element $Z$, derived from Eq.(\ref{eq4}) as
515:
516: \begin{equation}
517: \tau_{\rm i}^Z = \frac{1}{n_{\rm e}(\alpha_{\rm i}^Z+S_{\rm i}^Z)}~,
518: \end{equation}
519:
520: \begin{figure}[!t]
521: \centering \includegraphics[width=8.5cm]{fig2.ps}
522: \caption{Ionization/recombination timescale (in log scale)
523: for various ion species (dots) vs the ionization state for the labeled
524: elements, computed for a temperature $10$ MK and a density $10^8$
525: cm$^{-3}$. The dashed horizontal line marks 100 s.}
526: \label{fig2}
527: \end{figure}
528:
529: \noindent
530: for twelve important
531: elements, computed for a temperature $10$ MK and a density $10^8$
532: cm$^{-3}$. \orlandobis{Analogous timescales have been provided, in
533: tabular form, by \citet{1989SoPh..122..245G} in the study of observable
534: variability of spectral lines in soft X-ray and XUV regions of the
535: solar corona.}
536:
537: Clearly, $\tau_{\rm i}^Z$ increases with the ion species
538: and becomes larger than 100 s for the highest ion species of C, N, O,
539: Ne, Mg, Si, S, Ar and Ca. More species are involved in high ionization
540: times for Fe and Ni. The modeling shows that the plasma
541: immediately cools down considerably as soon as the heat pulse stops.
542: As a consequence, from
543: Fig.~\ref{fig2} we expect that the high ion species will
544: have no time to adjust to high temperature status if the heat pulse
545: lasts significantly less than 100 s.
546:
547: \orlandobis{The ionization and recombination timescale is dominated by
548: either ionization or recombination rate, depending on the sign of the
549: temperature jump
550: and on the ionization state of the element considered: in general,
551: $\tau_{\rm i}^Z$ is dominated by the ionization rate for the lowest ion
552: species and by the recombination rate for the highest ion species. The
553: ionization state for which ionization and recombination rates are
554: comparable depends on the temperature: the higher the temperature,
555: the higher is this ionization state. In the case considered here,
556: $\tau_{\rm i}^Z$ is dominated by the ionization rate for most of the ion
557: populations (see also \citealt{2006ApJ...647.1452P}) and by recombination
558: rate for the top ionization states (e.g. Si XV, S XVI, Ca 20,
559: Fe XXVI). As shown later, we found the effective ionization
560: temperature generally lower than the electron temperature, even when
561: the plasma is cooling; as a consequence, in the case of nanoflares
562: discussed here, $\tau_{\rm i}^Z$ is generally dominated by the ionization
563: rate for most of the ion populations.}
564:
565:
566: %\subsection{T/EM/NEI Evolution}
567:
568: \orlando{As an example, Fig.~\ref{fig1} shows the distribution of
569: population fractions of Fe along the loop derived assuming equilibrium
570: ionization (upper panels) and considering the deviations from equilibrium
571: ionization (lower panels) for the simulation
572: with heat pulses deposited at the footpoints
573: lasting $t_H = 5$ s
574: (see also Fig.~\ref{fig0}). During
575: the early phase of the evolution, the deviations from equilibrium of
576: ionization are very large at the loop footpoints due to the sudden
577: local increase
578: of temperature to $\sim 10$ MK (see left panels of Fig.~\ref{fig1}).
579: Whereas significant populations are expected above Fe XX low in the
580: loop (below $1.5 \times 10^9$ cm, reddish lines in the upper left panel),
581: no such highly ionized Fe species appears when NEI is taken into account.
582: As
583: expected, the high ion species are not able to adjust to the plasma
584: temperature before the heat pulse is over.
585: Then the plasma cools
586: down reducing the deviations from equilibrium ionization.
587: The deviations are still significant later at $t = 30$ s, when the loop
588: has thermalized (see Fig.~\ref{fig0}) and the maximum temperature is $\sim
589: 3$ MK. At this time, no highly ionized Fe species is present both
590: considering and not considering NEI, because of the effective cooling.
591: }
592:
593: \begin{figure*}[!t]
594: \centering \includegraphics[width=16cm]{fig3.ps}
595: \caption{Distributions of Fe population fractions (color-coded on the
596: right axis of the upper right panel) along the
597: loop from the chromosphere to the coronal loop apex, assuming
598: equilibrium ionization (upper panels) or considering the deviations
599: from equilibrium ionization (lower panels), at the labeled
600: times (5 s and 30 s), for the simulation with heat pulses located at the loop base and heat pulse
601: duration $t_H = 5$ s.}
602: \label{fig1}
603: \end{figure*}
604:
605: This time lag of the evolution of the ion species has important
606: implications on the temperature diagnostics of the plasma from its
607: emission. To address this issue, from the modeled population fractions,
608: we derive the temperature which best matches the actual ionization state
609: of the ion species for each of the elements in our simulations:
610: \orlandobis{operatively this is found as the one of the equilibrium temperatures with the
611: most similar three most populated ionized states.
612: It is worth noting that
613: at least two population fractions are necessary to find a
614: unique value of temperature and that this ``NEI" temperature should not
615: be intended as an exact temperature but as the temperature which best describes
616: the ionization state.} This is done at each time and position along
617: the loop. Fig.~\ref{fig3} shows the evolution (on a logarithmic time
618: scale) of the plasma maximum temperature and of the maximum NEI temperature
619: for five representative elements (C, O, Mg,
620: S, Fe). In the same figure the evolution of the coronal emission measure
621: at the maximum plasma temperature is also shown for comparison. The ion
622: species adjust to a hotter status very gradually, on a time scale of about
623: 100 s (Fig.~\ref{fig2}). Since this time is slightly longer than the
624: time taken by the emission measure to increase significantly (as shown
625: by the bottom panels), we expect observable effects in the X-ray band.
626: For $t_H = 5$ s, despite the maximum electron temperature is above 10 MK,
627: all the ion species are never represented by a temperature larger than
628: about 3 MK all over the event, for any pulse location. For $t_H = 30$
629: s, Fe (and Ni) reaches a temperature about $7-8$ MK for pulses located
630: at the footpoints and about 6 MK for pulse deposited at the apex. This
631: occurs for a very limited time range (between 100 and 200 s). The other
632: elements mostly stay at temperatures well below 5 MK. The results change
633: quite substantially for $t_H = 180$ s. Although with delay ($\ge 100$
634: s), here many elements have enough time to adjust to ionization stages
635: typical of higher temperatures, between 7 MK and 10 MK, for both heat
636: pulse locations. The high ionization state is maintained for longer time
637: intervals (more than 200 s).
638:
639: \begin{figure*}[!t]
640: \centering \includegraphics[width=13cm]{fig4.ps}
641: \caption{\orlando{Evolution of the maximum electron temperature (solid
642: lines) and of the maximum ``NEI" temperature (see text for definition) for
643: the labelled elements (dashed and dotted lines).
644: The figure shows the results for
645: heat pulse location at the loop base (left panels) and at the apex (right
646: panels)
647: and for heat pulse durations $t_H = 5$ s, 30 s, and 180 s.
648: Bottom panels: evolution of the coronal emission
649: measure at the maximum plasma temperature. The time is in log scale to
650: zoom up the fast initial evolution.
651: }}
652: \label{fig3}
653: \end{figure*}
654:
655:
656: %\subsection{Distribution of Emission Measure vs Temperature}
657:
658: The effect of long-lasting NEI is illustrated synthetically in Figs.
659: \ref{fig4}-\ref{fig6} which shows the total loop emission measure
660: distribution with temperature, EM($T$), averaged over different time
661: intervals. \orlando{The figure compares EM($T$) obtained either taking
662: or not taking the deviations from equilibrium ionization into account;
663: the former is obtained using the average ``NEI" temperature among those of
664: all the elements considered, the latter using directly the electron
665: temperature. }
666:
667: \begin{figure*}[!t]
668: \centering \includegraphics[width=14cm]{fig5.ps}
669: \caption{\orlando{Distributions of emission measure vs temperature
670: averaged over different time intervals obtained either taking (thick
671: lines) or not taking (shaded areas) the deviations from equilibrium
672: ionization into account (see text) for heat pulse location at the loop
673: base (left) or at the apex (right) and heat pulse duration $t_H = 5$
674: s.}}
675: \label{fig4}
676: \end{figure*}
677:
678: \begin{figure*}[!t]
679: \centering \includegraphics[width=14cm]{fig6.ps}
680: \caption{\orlando{Same as Fig. \ref{fig4}, for heat pulse duration $t_H =
681: 30$ s.}}
682: \label{fig5}
683: \end{figure*}
684:
685: \begin{figure*}[!t]
686: \centering \includegraphics[width=14cm]{fig7.ps}
687: \caption{\orlando{Same as Fig. \ref{fig4}, for heat pulse duration $t_H =
688: 180$ s.}}
689: \label{fig6}
690: \end{figure*}
691:
692: Averaging over the heat pulse duration only, we can see that EM($T$)
693: with and without NEI are invariably very different. The shorter the
694: pulse duration, the more the hot components are missing from the NEI
695: EM($T$) distributions. For $t_H = 5$ the hottest components including
696: NEI are only at about 1 MK, which increases to 3 MK for $t_H = 30$
697: s. The difference is less considerable for $t_H = 180$ s, but even in
698: this case the hot components are hardly above 5 MK, against 10 MK
699: without NEI. The differences are unsurprisingly reduced when we average over
700: longer times, which should approach a situation closer to the realistic
701: observations. Over a few durations of the heat pulses the differences are
702: still significant for the shortest heat pulses: for $t_H = 5$ s,
703: with NEI we do not find
704: components above 2 MK, whereas without NEI there are components almost to
705: 10 MK. For $t_H = 30$ s the two EM($T$) distributions are more similar,
706: but with NEI we do not find components hotter than 6 MK. For the longest
707: pulse duration, the EM(T) distributions almost coincide.
708:
709: \section{Discussion and conclusions}
710: \label{sec:discus}
711:
712: The target of this work is to explore the effect of non-equilibrium
713: of ionization (NEI) in
714: nanoflare-heated loops.
715: The importance of NEI effects on coronal observations had been already
716: pointed out and evaluated by \cite{1989SoPh..122..245G}.
717: That study was devoted more specifically
718: to the diagnostics of rapid variability. Here, we focus on the effects on
719: the detection of hot plasma in nano-flaring loops.
720: To this
721: purpose we set up hydrodynamic simulations of plasma confined in
722: an active region loop heated to $\sim 10$ MK by more or less long
723: nanoflares. As new achievements specifically set up for this study,
724: we have included the effect of the saturated thermal conduction in the
725: modeling (\citealt{2003SPD....34.1006K}), and computed the evolution of the ion population fractions
726: of several important elements driven by a short heat pulse. Each
727: ion species takes a characteristic time to reach equilibrium conditions
728: for a certain plasma temperature and density. If the plasma conditions
729: vary on very small time scales, the ions are unable to adjust to
730: the new and rapidly changing conditions and the emitted spectrum during the
731: plasma variations can become very different from the one expected in
732: equilibrium conditions. If the variation is a sudden heating followed by
733: a sudden cooling, the emission may never or only partially adapt to the
734: transient hot conditions. To evaluate the importance of this effect and
735: its dependence on the heat pulse parameters, we have computed the
736: actual temperature of the plasma as it would appear in the radiation spectrum
737: according to the effective ionization state.
738:
739: We find that the effects are significant if the heat pulses last less than
740: a minute or so. For durations of a few seconds the spectra will never show
741: temperatures higher than 2-3 MK even though the electron temperature
742: overcomes 10 MK. This effect is little dependent on the location of
743: the heat pulse, except for details.
744: These results can be considered valid in wide
745: generality in spite of some limitations of our approach. Our assumption
746: of pulses leading to 10 MK is, of course, arbitrary, although this has
747: been taken as possible typical temperature in other works (\citealt{2006ApJ...647.1452P}). Conclusions are certainly valid for less intense pulses,
748: which heat less the plasma. We do not expect dramatic differences also for
749: more intense pulses (e.g. 20 MK), because the thermal conduction becomes
750: even more efficient and the plasma initial cooling (after the end of the
751: pulse) will be faster. Also the details of the shape of the heat pulse
752: should not have much influence, provided that the overall duration is
753: the same. In general, of course, more gentle heating should make plasma
754: reach equilibrium conditions more easily, while the opposite occurs for
755: more spiky pulses. More important is instead our assumption of a single
756: pulse inside a given thread. With it, we are excluding a repetition of
757: the pulse inside the same thread, or, at least, a very long repetition
758: time, so long as to have a rarified loop ($\sim 10^8$ cm$^{-3}$) again
759: before the new ignition. More frequent pulses would be released in a denser
760: plasma which is faster to adjust to ionization equilibrium, and in which,
761: therefore, we expect less NEI effects. The ionization times in fact scale
762: inversely with plasma density. Heat pulses released in a $\sim 10^9$
763: cm$^{-3}$ dense plasma, i.e. about ten times denser than our initially
764: rarified flux tube, would excite highly ionized element species about
765: ten times more rapidly, with typical times of about ten seconds
766: (Fig.~\ref{fig2}).
767:
768: In the range $1-10$ MK the plasma emits radiation mostly in the X-ray
769: band and the spectra are dominated by the emission lines produced by
770: the recombination and deexcitation of highly ionized atoms (e.g.,
771: \citealt{1971ApJ...168..283T}, \citealt{1976ApJS...30..397K}). For
772: instance, hundreds of lines of at least ten Fe ion species populate
773: the spectra of plasma in that range of temperature. For this reason,
774: the ionization status of the emitting ions becomes extremely important
775: in the thermal ``appearance'' of the plasma, which we detect through
776: its radiation with remote sensing, while the continuum electron
777: bremsstrahlung emission -- which adjusts immediately to thermal and
778: dynamic plasma variations -- becomes more important at much higher
779: temperatures. Our results therefore apply both to the analysis of single
780: lines through high resolution spectroscopy, and to the diagnostics
781: from wideband multi-filter instruments in the X-ray and EUV bands,
782: e.g. Hinode/XRT or the Atmospheric Imaging Assembly (AIA) on board
783: the forthcoming Solar Dynamic Observatory.
784:
785: According to these considerations, a hard detection of hot plasma may
786: point to a scenario of coronal loops heated by single short and relatively
787: intense nanoflares. Very short durations and very long repetition times
788: naturally imply that the pulses should be deposited always in different
789: strands and that, therefore, the strands involved should be a very large
790: number, i.e. the loop must have a very fine transversal structure.
791: A scenario of loops heated by short nanoflares has some important
792: implications. Hot plasma may be difficult to detect not because of
793: a limitation of the telescopes, detectors and filters, but due to an
794: intrinsic property of the plasma emission. The critical parameter
795: for detectability of hot plasma is the duration of the heat pulse and
796: $\sim 1$ minute is the critical duration value, a useful indication for
797: detailed models of nanoflaring mechanisms. The fine temporal structure
798: would also automatically imply a very fine spatial structure, difficult
799: to resolve.
800:
801:
802: %Let's assume that an active region coronal loop "lives" and
803: %is steady for instance, for more than an hour, e.g., 5000 s. A pulse able
804: %to heat the plasma at 10 MK injects about 50 times the heating rate
805: %sufficient to heat the plasma at 3 MK (\citealt{1978ApJ...220..643R}). On average,
806: %therefore a 5 s pulse would inject the same amount of energy as a 250
807: %s 3 MK pulse. This means that about 20 heat pulses should be enough to
808: %keep the loop at 3 MK on average. However 20 pulses lasting 5 s combine
809: %to a total of 100 s of heating versus 5000 s of total loop life: most of the
810: %time the strands are cooling, but each for too long a time to have 3 MK
811: %on average. Unless the 5 s heat pulses are deposited in much thinner
812: %strands and are therefore much more in number, distributed with random phases.
813: %To have the loop life totally covered by heat pulses we need
814: %1000 pulses. Each pulse is therefore to be deposited in 1/50 of the loop
815: %cross-section area, i.e. 2\% of the loop is continuously heated.
816:
817: This picture is fully compatible with the current scenario
818: of nanoflaring multi-stranded coronal loops. For instance, it
819: naturally involves the hot-underdense/cool-overdense loop cycle
820: (\citealt{2002ApJ...579L..41W},\citealt{2004ApJ...605..911C},
821: \citealt{2006SoPh..234...41K}). Less trivial is the effect on the
822: diagnostics of the detailed thermal structure along and across the loops
823: (e.g., \citealt{2004ApJ...605..911C}, \citealt{2005ApJ...633..499A},
824: \citealt{2006A&A...449.1177R}, \citealt{2007ApJ...658L.119S}), which we
825: defer to later work.
826:
827: In conclusion, we have investigated in detail the effect of the
828: non-equilibrium of ionization driven by nanoflares in finely structured
829: coronal loops on the detection of hot plasma in the loops, and provided
830: constraints on the heat pulse duration which may make or make not the
831: detection difficult. Further more detailed information
832: and diagnostics are expected from high resolution spectroscopy such
833: as that from EUV Imaging Spectrometer on board the Hinode mission or
834: from the Extreme Ultraviolet Variablity Experiment (EVE) on board the
835: forthcoming Solar Dynamic Observatory.
836:
837: \bigskip
838: \acknowledgements{We thank G. Peres and the referee, J. Klimchuk,
839: for suggestions.
840: The software used in this work was
841: in part developed by the DOE-supported ASC/Alliance Center for
842: Astrophysical Thermonuclear Flashes at the University of Chicago,
843: using modules for non-equilibrium ionization, thermal conduction,
844: and optically thin radiation built at the Osservatorio Astronomico
845: di Palermo. The simulations were executed at the SCAN facility of the
846: Osservatorio Astronomico di Palermo. This work was supported by
847: Ministero
848: dell'Istruzione, dell'Universit\`a e della Ricerca, by Istituto
849: Nazionale
850: di Astrofisica, and by Agenzia Spaziale Italiana
851: (ASI), contract I/015/07/0. FR acknowledges support
852: from the International Space Science Institute in the framework of
853: an international working team.}
854:
855: \appendix
856:
857: \section{Non-equilibrium ionization in the FLASH code}
858:
859: The non-equilibrium ionization is added to the FLASH code using a
860: method of time splitting between the hydrodynamic and the NEI numerical
861: modules. A fractional step method is required to integrate the equations
862: and in particular to decouple the NEI solver from the hydrodynamic solver.
863: For each timestep, the homogeneous hydrodynamic transport equations given
864: by Eqs. \ref{eq1}--\ref{eq4} are solved using the FLASH hydrodynamic
865: solver with R = 0. After each transport step, the stiff system of
866: ordinary differential equations for the NEI problem with the form:
867:
868: \begin{equation}
869: \frac{\partial n_i^Z}{\partial t} = R_i^Z ~~~~~~~~ (i = 1, ...,
870: N_{spec})
871: \end{equation}
872:
873: \noindent
874: are integrated. This step incorporates the reactive source terms.
875: Within each grid cell, the above equations can be solved separately with
876: a standard ODE method. Since this system is stiff, it is solved using
877: the Bader-Deuflhard time integration solver\footnote{The variable-order
878: Bader-Deuflhard routine used here is a combination of the routine METANI
879: given by \cite{bader} and the routine STIFBS given by \cite{press}
880: (see also \citealt{1999ApJS..124..241T}).} with the MA28 sparse matrix
881: package. \cite{1999ApJS..124..241T} has shown that these two algorithms
882: together provide the best balance of accuracy and overall efficiency.
883: Note that the source term in the NEI module is adequate to solve the
884: problem for optically thin plasma in the coronal approximation; just
885: collisional ionization, auto-ionization, radiative recombination, and
886: dielectronic recombination are considered.
887:
888: \bibliographystyle{apj}
889: \bibliography{references}
890:
891: \clearpage
892:
893: %\figcaption[fig1.ps]{TBD}
894:
895:
896: \clearpage
897:
898: \clearpage
899:
900: %\begin{figure*}
901: % \centering
902: % \epsscale{0.9}
903: % \plotone{fig1.ps}
904: %\end{figure*}
905:
906: \end{document}
907:
908:
909:
910: Results and Discussion
911: - Outline of loop evolution for a sample case (5 sec)
912: - NEI vs equilibrium effects
913: - Different durations
914: - Estimation of NEI times
915: - Meaning of different durations: to have the same energy budget enough
916: to heat a typical active region loop, short duration means more
917: filamented loop
918: Implications:
919:
920: - for short heat pulses hot plasma is never detectable
921:
922: - a reference duration time scale is about 30 s
923:
924: - this might explain the non-evidence of hot plasma
925:
926: - constraint on nanoflare models: filamented + short duration
927:
928: - Conclusion
929: