1: \documentclass{aastex}
2: \usepackage{emulateapj5}
3: %\usepackage{apjfonts}
4:
5: %\shorttitle{TURBULENCE IN INTERSTELLAR CLOUDS}
6: %\shortauthors{KOYAMA \& INUTSUKA}
7: \begin{document}
8:
9: \submitted{ApJL in press.}
10: %%\slugcomment{December 15, 2000}
11:
12: \title{An Origin of Supersonic Motions in Interstellar Clouds}
13:
14: \author{Hiroshi Koyama}
15: \affil{Astronomical Data Analysis Center, \\
16: National Astronomical Observatory, Mitaka, Tokyo 181-8588, Japan; \\
17: E-mail address: H.Koyama@nao.ac.jp }
18:
19: \and
20:
21: \author{Shu-ichiro Inutsuka}
22: \affil{Department of Physics, Kyoto University, Kyoto
23: 606-8502, Japan; \\
24: E-mail address: inutsuka@tap.scphys.kyoto-u.ac.jp }
25:
26:
27: \begin{abstract}
28: The propagation of a shock wave into an interstellar medium
29: is investigated
30: by two-dimensional numerical hydrodynamic calculation with cooling,
31: heating and
32: thermal conduction.
33: We present results of the high-resolution two-dimensional
34: calculations to follow the fragmentation due to the thermal
35: instability in a shock-compressed layer.
36: %We use realistic thermal conduction coefficient to resolve ``Field
37: % length''
38: %that is the critical wavelength for thermal instability.
39: We find that geometrically thin cooling layer behind the shock front
40: fragments into small cloudlets.
41: The cloudlets have supersonic velocity dispersion in the warm neutral medium
42: in which the fragments are embedded as cold condensations.
43: The fragments tend to coalesce and become larger clouds.
44: \end{abstract}
45:
46: \keywords{ISM: clouds
47: --- ISM: molecules
48: --- ISM: supernova remnants
49: --- shock waves
50: --- turbulence}
51:
52: \section{INTRODUCTION}
53: The interstellar medium (ISM) and cold gas clouds are characterized by a
54: clumpy substructure and a turbulent velocity field (Larson 1981).
55: The maintenance and dissipation processes of the turbulence are supposed
56: to be important in the theory of star formation (McKee 1989, Nakano
57: 1998, Elmegreen 1999).
58: The understanding of the origin of cold clouds and their internal
59: substructure has therefore fundamental importance for a consistent theory
60: of star formation and ISM.
61:
62: Some of the polarization maps and direct
63: measurements of field strength in some star forming regions suggested
64: the importance of MHD (Alfv\'{e}n waves) turbulence (Myers \&
65: Goodman 1988, Crutcher et al. 1993).
66: Recent simulations of MHD turbulence, however, suggest that it
67: dissipates rapidly (Gammie \& Ostriker 1996, Mac Low et al. 1998,
68: Mac Low 1999, Ostriker et al. 1999).
69: Possible sources of energy supply are winds and outflows from young
70: stellar objects (Franco \& Cox 1983, McKee 1989).
71: Note however that clumpy structures with supersonic
72: velocity dispersions are also observed
73: even in regions where star formation is inactive.
74: Thus, the origin of clumpy cloud
75: structure remains as an outstanding issue.
76:
77: We propose that the clumpiness in clouds arises naturally from their
78: formation through thermal instability which acts on timescales that
79: can be much shorter than the duration of the interstellar shocks (e.g.,
80: galactic spiral shocks and supernova remnants).
81: The basic properties of the thermal instability are studied in a
82: pioneering paper by Field (1965). Schwarz, McCray, and Stein
83: (1972) studied numerically the growth of condensation in cooling region
84: including the effects of ionization and recombination.
85: Hennebelle \& Perault (1999) studied the elementary condensation process
86: in turbulent flow in the restricted conditions of neutral atomic
87: gas in plane-parallel geometry. Burkert \& Lin (2000)
88: studied cooling and fragmentation of gas using simplified power-law
89: cooling function. Smith (1980) studied collisions between cold atomic clouds,
90: which produce thick layers of shock-heated atomic gas in which thin
91: sheets of cold molecular gas form by the thermal instability. Koyama \&
92: Inutsuka (2000, hereafter Paper I) have done one-dimensional
93: hydrodynamic calculations for the propagation of a strong shock wave
94: into warm neutral medium (WNM) and cold neutral medium (CNM) including
95: detailed thermal and chemical processes. They have shown that the post-shock
96: region collapses into a cold layer as a result of the thermal instability.
97: They expect that this layer will break up into very small cloudlets which
98: have different translational velocities.
99:
100: In this paper, we show that the fragmentation of the shock-compressed layer
101: indeed provides turbulent condensations,
102: by using two-dimensional hydrodynamic calculation with radiative cooling and
103: heating and thermal conduction.
104:
105: \section{NUMERICAL SIMULATIONS}
106: \subsection{Numerical Scheme}
107: The hydrodynamics module of our scheme is based on the second-order
108: Godunov method (van Leer 1979).
109: We solve following hydrodynamic equations.
110: \begin{equation}
111: \frac{\partial \rho}{\partial t}+\nabla\cdot(\rho {\bf v})=0,
112: \end{equation}
113: \begin{equation}
114: \frac{\partial {\bf v}}{\partial t}+{\bf v}\cdot\nabla{\bf v}=
115: -\frac{\nabla P}{\rho},
116: \end{equation}
117: \begin{equation}
118: \frac{\partial\rho e}{\partial t}
119: +\nabla\cdot(\rho e {\bf v})+P\nabla\cdot {\bf v}=
120: \frac{\rho}{m_{\rm H}}\Gamma-\left(\frac{\rho}{m_{\rm H}}\right)^2\Lambda(T)
121: +\nabla\cdot(K\nabla T),
122: \end{equation}
123: where $\rho$, $P$, ${\bf v}$ are the density, pressure, and velocity of
124: the gas, the specific internal energy $e=P/(\gamma-1)\rho$, with
125: $\gamma=5/3$.
126: $m_{\rm H}$ is the hydrogen mass in gram.
127: $\Gamma$ and $\Lambda$ are heating and cooling rate,
128: respectively.
129: For the coefficient of thermal conductivity we adopt $K=2.5\times
130: 10^3T^{1/2} {\rm erg \ cm^{-1}K^{-1}s^{-1}}$ (Parker 1953).
131: Effects of self-gravity and
132: magnetic field are not treated in this paper.
133:
134: In Paper I, we included the following processes in the hydrodynamic
135: calculations:
136: photoelectric heating from small grains and PAHs,
137: heating and ionization by cosmic rays and X-rays,
138: heating by H$_2$ formation and destruction,
139: atomic line cooling from hydrogen Ly$\alpha$,
140: \ion{C}{2}, \ion{O}{1}, \ion{Fe}{2}, and \ion{Si}{2},
141: rovibrational line cooling from H$_2$ and CO,
142: and atomic and molecular collisions with grains.
143: However the following analytic fitting function adopted here reproduces
144: the features relevant for our purpose:
145:
146: \begin{eqnarray}
147: \Lambda(T)/\Gamma&=&10^7\exp\left(-114800/(T+1000)\right) \nonumber \\
148: &&+14{\sqrt{T}}\exp\left(-92/T\right) {\rm cm^{3}}, \\
149: \Gamma&=&2\times 10^{-26} {\rm ergs/s},
150: \end{eqnarray}
151: where $T$ is in Kelvins.
152:
153: We use 2048 $\times$ 512 Cartesian grid points covering a 1.44 $\times$
154: 0.36 pc region so that the spatial resolution is 0.0007 pc = 140 AU.
155:
156: We have tested the linear growth of the thermal instability (Field
157: 1965). The test is performed with an eigenfunction by using 16 grid
158: points. The numerical results reproduce the linear growth rate with
159: relative errors of 0.05 \% when $\delta \rho/\rho <0.01$
160:
161: \subsection{Initial and Boundary Conditions}
162:
163: To investigate the shock propagation into WNM we consider a plane-parallel
164: shocked layer of gas.
165: We take the $y$-axis to be perpendicular to the layer.
166: An uniform flow approaches the layer from the upper side with a velocity
167: $V_{\rm y}=-26$ km/s and density $n=0.6 \ {\rm cm^{-3}}$ and temperature
168: $T=6000$ K.
169: We assume that the lower side of the
170: shocked layer is occupied by hot, tenuous gas (density $n=0.12 \ {\rm
171: cm^{-3}}$) with a high pressure
172: $P/k_{\rm B}=4\times 10^4 {\rm K cm^{-3}}$.
173:
174: We set up density fluctuations as follows:
175: \begin{equation}
176: \frac{\delta\rho(x,y)}{\rho_0}=
177: \frac{A}{k_{\rm max}}\sum_{{\rm i},{\rm j}=0}^{{\rm i}_{\rm max}}
178: \left({k_{\rm i}^2+k_{\rm j}^2}\right)^{\frac{n}{2}}
179: \sin(x k_{\rm i}+y k_{\rm j} + \theta(k_{\rm i},k_{\rm j})),
180: \end{equation}
181: where $k_{\rm i}=2\pi i/L$ is a wave number,
182: $\theta$ is a random phase, $n$ is a spectral index, and $L$ is a
183: calculation box size of $x$- direction.
184: We imposed out-going boundary condition for the $y$- direction, and
185: periodic boundary condition for the $x$- direction.
186:
187: \subsection{Results of Thermal Instability and Fragmentation}
188:
189: Figure \ref{fig1} shows density distribution at $t=$0.337, 1.06 Myr.
190: We used an initial fluctuation $A=0.05$, the power index $n=0$, and
191: ${\rm i}_{\rm max}=16$.
192: Thermally collapsed layer and its fragments are seen in the
193: shock-compressed layers.
194:
195:
196: Figure \ref{fig1}a shows the beginning of the
197: formation of cold thermally collapsed layer.
198: Density discontinuity at $y\sim 0.46$ pc corresponds to the shock front.
199: Density discontinuity at $y\sim -0.1$ pc is the contact surface between
200: the shock-compressed layer and the hot interior.
201: Behind the shock front, cooling dominates heating
202: and temperature decreases monotonically.
203: The ISM in the range of 300 -- 6000 K is thermally unstable.
204: Thus, thermally collapsing layer is subject to thermal instability.
205: The condensations continue to collapse and cool until the radiative cooling
206: balances the radiative heating, in such a way that approximate pressure
207: equilibrium is
208: maintained between condensations and the surrounding gas.
209: Temperature of these clumps finally reaches the value of thermal
210: equilibrium ($T=20$ K at $n=2000 {\rm cm^{-3}}$).
211: When the cooling balances the heating, the condensation ceases to
212: collapse but the
213: surrounding warm gas continues to cool radiatively and accrete onto the
214: cloud cores.
215:
216: The cold clumps have considerable translational velocity dispersion.
217: The typical velocity dispersion is about several km/s.
218: The linear analysis predicts that perturbation with sufficiently long
219: wavelength grows exponentially.
220: However, the velocity of the non-linearly developed perturbation
221: has the upper limit that is the sound speed of the warmer medium
222: ($\approx$ 10km/s),
223: because the driving force of the instability is the pressure of the
224: less dense warmer medium.
225: This non-linearly developed perturbation produces the translational
226: velocities of the cold clumps.
227: These velocity on the order of the sound speed of the WNM is
228: highly supersonic with respect to the sound speed of cold gas.
229: Thus we can understand why the supersonic velocity dispersion of
230: the cold medium is comparable to the sound speed of the WNM.
231:
232: The cold cloud cools the interface via thermal conduction.
233: The pressure of the warm medium between two
234: cold clumps becomes smaller, which in effect provides an attractive
235: force of two adjacent cold clumps. This attractive force is the driving
236: force of the coalescence of two cold clumps.
237: Thus, cold clumps coalesce into larger clouds.
238:
239: The sizes of these clumps are a few orders of magnitude smaller than the Jeans
240: length, $\lambda_{\rm J}=1.2 \ {\rm pc}\sqrt{(T/20{\rm K})(2000{\rm
241: cm^{-3}}/n) }$ .
242: Thus, each cloud is gravitationally stable.
243: The shock front itself remains stable at least in $10^6$ years in this
244: plane-parallel problem.
245:
246: For comparison, we changed the initial fluctuations in the range of the
247: amplitude (A=0.05--0.005), and the power index (n=0,$\pm$ 2), but the
248: resultant velocity dispersion defined by $\int \rho v_{rm x}^2\,dxdy/\int
249: \rho\,dxdy$ changed only by 10 \%. We also changed the spatial
250: resolution (512x128, 1024x256, and 2048x512) but the velocity dispersion
251: changed only by 10 \%. Thus, Figure \ref{fig1} shows a typical result of
252: this problem.
253: From these results we can demonstrate that the role of thermal instability
254: behind the shock front is one of the most important processes
255: in the evolution of interstellar shocked layers.
256:
257: \section{DISCUSSION}
258:
259: \subsection{Formation of Molecules}
260:
261: The main coolant in cold dense clouds is supposed to be CO (see, e.g.,
262: Neufeld, Lepp, \& Melnick 1995; Paper I).
263: H$_2$ enables the formation of CO and other molecules.
264: The abundance of H$_2$ in the CNM critically depends on the optical
265: depth and local density.
266: The typical column density of the layer becomes
267: $2\times 10^{20} {\rm cm^{-2}}$.
268: We calculated detailed thermal-chemical equilibrium state in the dense
269: region by adopting the values of density, temperature, and column
270: density from this dynamical simulation.
271: Twenty-five percent of the
272: hydrogen is H$_2$ and 0.03\% of the
273: carbon is CO in thermal and chemical equilibrium.
274:
275: We simulate the observation of molecular clumps at $t=1.06$ Myr
276: (Figure \ref{fig1}b) as the Position-Velocity (P-V)
277: diagram of the ${}^{12}$CO J = 1 -- 0 emission (Figure \ref{fig2}a).
278: We use the usual relationship $T_{\rm B}=\frac{1}{2}\lambda^2I/k_{\rm
279: B}$, with the specific intensity $I=n^2\Lambda L/(4\pi \Delta v)$, where
280: $n^2\Lambda$ is the cooling per unit volume, $L$ is the path length,
281: $\Delta v$ is the Doppler line width. We adopt $\Delta v$ to be the
282: sound velocity.
283: As shown in Figure \ref{fig2}a, the typical velocity dispersion in the
284: clumps is about a few km/s. The cloud to cloud velocity dispersion is
285: also a few km/s.
286: Figure \ref{fig2}b shows that the P-V diagram of hydrogen nuclei.
287: The velocity dispersion of the diffuse warm medium is about 10 km/s.
288:
289: \subsection{An Origin of ``Turbulence'' in Interstellar Clouds}
290: Diffuse ISM in the Galaxy is frequently compressed by supernova
291: explosions (McKee \& Ostriker 1977).
292: Therefore, the shock propagation into the ISM plays an important
293: role in the evolution of the Galactic ISM.
294: We have studied the shock propagation into WNM by two-dimensional
295: hydrodynamic calculations, and have shown that the thermally collapsed layer
296: breaks up into very small cloudlets.
297: Fragmentation of the thermally collapsing layer is a result
298: of the thermal instability.
299: The thermal instability produces many cloudlets which have different
300: translational velocities.
301: We expect that the Galactic ISM is occupied by these small
302: cloudlets which have supersonic velocity dispersion
303: , because the ISM is frequently compressed by supernova
304: explosions, stellar winds, spiral density waves, cloud-cloud
305: collisions, etc.
306: The \ion{H}{1} 21-cm observation maps show the existence of many
307: shell-like or filamentary structures in the Galaxy (Hartmann \& Burton
308: 1997). In addition, high resolution observations suggested that the clumpy
309: distribution is ubiquitous in the Galaxy (Heiles 1997).
310: If these structures are really the results of the shock waves, many
311: small molecular cloudlets should be hidden in the shock-compressed
312: layers. These cloudlets should have translational velocities owing to the
313: fragmentation of thermal instability.
314: Observational ``turbulence'' in the ISM should reflect these motions.
315: Thus, we propose that an origin of ``turbulence'' in the ISM is the
316: motion of this small cloudlet complex.
317:
318: In the radiative shocked layers, the gases lose thermal energy through
319: radiative cooling.
320: Thus, the initial kinetic energy of the pre-shock gas (in the comoving
321: frame of the post-shock gas) converted to the radiation energy, which
322: escapes from the system.
323: If, however, the post-shock gas becomes dynamically unstable
324: by the thermal instability as in this paper,
325: some portion of the thermal energy is transformed into
326: the kinetic energy of the translational motions of the cold cloudlets,
327: which does not easily escape from the system.
328: Therefore we can consider the origin of interstellar turbulence is
329: due to the conversion of the gas energy in supersonic motion.
330: In principle, the kinetic energy of the cloudlets can be lost
331: via coalescence of the cloudlets.
332: The damping rate of the velocity dispersion of the cloudlets
333: will play a key role in the evolution of this system,
334: and hence, should be studied in the subsequent paper.
335:
336: Expected effects of magnetic field modify the development of the thermal
337: instability described here.
338: The presence of magnetic field can suppress thermal conduction
339: efficiently, allowing the collapse of small scale structure by the thermal
340: instability.
341: The magnetic forces suppress the motion that is perpendicular to the
342: magnetic field lines, and hence suppress the growth of perturbations
343: whose wavevector is perpendicular to the magnetic field lines.
344: However,
345: the magnetic forces do not affect the motion that is parallel to the magnetic
346: field lines, and hence the growth of perturbations whose wavevector is
347: parallel to the magnetic field lines.
348: As a result, the dense sheets or filaments will form and
349: tend to align {\em perpendicular} to the magnetic field lines.
350: We need
351: three-dimensional calculation to analyze these effects of magnetic fields.
352:
353: Numerical computations were carried out on VPP300/16R and VPP5000/48 at the
354: Astronomical Data Analysis Center of the National Astronomical Observatory,
355: Japan.
356: %, which is an inter-university research institute of astronomy
357: %operated by Ministry of Education, Science, Culture, and Sports.
358:
359:
360:
361: \begin{references}
362: \reference{BL}
363: Burkert, A., \& Lin, D. N. C. 2000, \apj, 537, 270
364: \reference{CTGHKM}
365: Crutcher, R.M., Troland, T. H., Goodman, A. A., Heiles, C., Kazes, I. \&
366: Myers, P. C. 1993, \apj, 407, 175
367: \reference{EL}
368: Elmegreen, B. G. 1999, \apj, 527, 266
369: \reference{field1965}
370: Field, G. B. 1965, \apj, 142, 531
371: \reference{FC}
372: Franco, J., \& Cox, D. P. 1983, \apj, 273, 24
373: \reference{GO}
374: Gammie, C. F. \& Ostriker, E. C. 1996, \apj, 466, 814
375: \reference{HB}
376: Hartmann, D., \& Burton, W. B. 1997, Atlas of Galactic Neutral Hydrogen
377: (Cambridge: Cambridge Univ. Press)
378: \reference{heiles1997}
379: Heiles, C. 1997, \apj, 481, 193
380: \reference{HP}
381: Hennebelle, P. \& Perault, M. 1999, \aap, 351, 309
382: \reference{PAPER1}
383: Koyama, H., \& Inutsuka, S. 2000, \apj, 532, 980 (Paper I)
384: \reference{Larson}
385: Larson, R. B. 1981, MNRAS, 194, 809
386: \reference{MacLow1998}
387: Mac Low, M.-M., Klessen, R. S., Burkert, A.,\& Smith, M. D. 1998,
388: Phys. Rev. Lett., 80, 2754
389: \reference{MacLow1999}
390: Mac Low, M.-M. 1999, \apj, 524, 169
391: \reference{MG}
392: Myers, P. C. \& Goodman, A. A. 1988, \apj, 326, L27
393: \reference{McKee89}
394: McKee, C. F. 1989, \apj, 345, 782
395: \reference{MO}
396: McKee, C. F. \& Ostriker, J. P. 1977, \apj, 218, 148
397: \reference{Nakano}
398: Nakano, T. 1998, \apj, 494, 587
399: \reference{NLM}
400: Neufeld, D. A., Lepp, S. \& Melnick, G. J. 1995, \apjs, 100, 132
401: \reference{OGS}
402: Ostriker, E. C., Gammie, C. F. \& Stone, J. M. 1999, \apj, 513, 259
403: \reference{Parker1953}
404: Parker, E. N. 1953, \apj, 117, 431
405: \reference{schwarz1972}
406: Schwarz, J., McCray, R., \& Stein, R. 1972, \apj, 175, 673
407: \reference{smith1980}
408: Smith, J. 1980, \apj, 238, 842
409: \end{references}
410:
411:
412:
413: \begin{figure}
414: \figurenum{1}
415: \epsscale{0.3}
416: \plotone{fig1a2.ps}
417: \plotone{fig1b2.ps}
418:
419: \caption{Density distribution of the shock-compressed layers.
420: The figures show the snapshot at (a) $t=0.337$ Myr, (b) $t=1.06$ Myr,
421: respectively.}
422: \label{fig1}
423: \end{figure}
424:
425:
426:
427: \begin{figure}
428: \figurenum{2}
429: \epsscale{0.25}
430: \plotone{fig2a.ps}
431: \plotone{fig2b.ps}
432:
433: \caption{The panel a shows the position-velocity (P-V) diagram
434: of the ${}^{12}$CO J=1 -- 0 emission obtained from the calculation at
435: $t=1.06$ Myr (see Figure \ref{fig1}b ). We
436: assume 0.03 \% of the carbon is CO. The panel b shows the P-V
437: diagram of the hydrogen nuclei.}
438: %(\ion{H}{1}+H$_2$)
439: \label{fig2}
440: \end{figure}
441:
442:
443:
444: \end{document}
445: