astro-ph0310558/ts.tex
1: \documentstyle[11pt,newpasp,epsf,twoside]{article}
2: \markboth{Tissera \& Scannapieco}{Cosmological chemical evolution}
3: \pagestyle{myheadings}
4: \nofiles
5: 
6: % Some definitions I use in these instructions.
7: 
8: \def\emphasize#1{{\sl#1\/}}
9: \def\arg#1{{\it#1\/}}
10: \let\prog=\arg
11: 
12: \def\edcomment#1{\iffalse\marginpar{\raggedright\sl#1\/}\else\relax\fi}
13: \marginparwidth 1.25in
14: \marginparsep .125in
15: \marginparpush .25in
16: \reversemarginpar
17: 
18: \begin{document}
19: \title{Chemical Evolution in Hierarchical Clustering Scenarios}
20:  \author{Patricia B. Tissera}
21: \affil{Institute of Astronomy and Space Physics, CONICET, Argentina}
22: \author{
23:  Cecilia Scannapieco 
24: }
25: \affil{Institute of Astronomy and Space Physics, CONICET, Argentina}
26: 
27: \begin{abstract}
28: We present first results of an implementation of chemical evolution
29: in a cosmological hydrodynamical code,
30:  focusing the analysis on the effects of cooling baryons
31: according to their metallicity. We found that simulations with primordial
32: cooling can underestimate the star formation rate from $z < 3$ and 
33:  by up to $\approx 20 \%$. We constructed simulated spectra by combining
34: the star formation and chemical history of galactic systems with
35: spectra synthesis models and assess the impact of chemical evolution
36: on the energy distribution.
37: 
38: \end{abstract}
39: 
40: 
41: 
42: \section{Introduction}
43: 
44: 
45: The current cosmological paradigm  predicts a 
46: hierachical building up of the structure. This scenario has largely supported
47: by observations in the local and high redshift Universe.
48: However, few but important descrinpances remain to be explanied
49: such as the inner distribution of dark matter in small haloes,
50: the origin of the high baryonic content and 
51: metallicity of the intercluster medium, etc (Shanks 2004). Most of these
52: open problems requires the comprehensive understanding of galaxy formation,
53: including all the involved scales. 
54: Numerical simulations have played an important role in understanding and
55: proving the connection between theory and observations at different
56: scales. However, the modelling of small scale physical processes such
57: as gas cooling, star formation and chemical evolution within
58: a cosmological framework presents different difficulties because of 
59: our poor understanding of the relevant physics
60: and numerical limitations. In particular, only recently chemical
61: evolution has started to be treated within cosmological 
62: hydrodynamical simulations (Yepes et al. 1998; Cen \& Ostriker 1999; 
63: Mosconi et al. 2001).
64:   
65: The dramatic building up of observational evidences on chemical properties of
66: the stellar population and interstellar medium 
67: in the Universe has proved to
68: have a significant  impact on the study of galaxy formation.
69: Chemical elements are produced in the stellar interior and ejected in
70: a pacific or violent way after characteristic timescales determined by
71: stellar evolution. Afterwards,  these elements can  be   mixed up by 
72:   mergers and interactions. It is also important to
73: notice that the  metal production is directly linked to the star formation
74: history which can be also affected by the  environment (Barton et
75: al. 2000;
76: Lambas et al. 2003; Balogh et al. 2004). As a consequence, 
77: it is expected that  chemical patterns store information of the
78: process of galaxy formation ( Freeman  \& Bland-Hawthorn 2002).
79:  
80: Another important motivation to include a chemical treatment in galaxy
81: formation models is related to  the fact that
82: baryons cool according to their metallicity. Sutherland \& Dopita (1993)
83: showed that a gas cloud with solar metallicity cools at
84: a rate of up to few  orders of magnitude 
85: more efficiently than a primordial gas cloud.
86: Hence, the enrichment of the interstellar medium can have
87: dynamical  consequences and affect the formation of the structure.
88: 
89: In this work, we present first results of detail treatement  of
90: chemical evolution in Gadget2 (Springel \& Hernquist 2002) based
91: on the implementation of Mosconi et al. (2001).
92: In Section 2 we briefly summarize the main important aspects of the
93: chemical code. Section 3 shows the first results and Section 4
94: presents the conclusions.
95: 
96: \section{The model}
97: 
98: The chemical model has been implemented  within 
99:  Gadget2 (Springel \& Hernquist
100: 2002) which is
101: based on a  modified version of SPH 
102: that conserves entropy (energy and angular momentum).
103: Stars are formed according to a probability function defined by
104: the relation between cooling and dynamical time (see also Lia et al. 2001).
105:  A  standard Salpeter Initial mass function has been assumed.
106: Supernovae 
107: II are considered to form from stars with $M > 8 M_{\odot}$
108: and to produce metals according to Woosley \& Weaver (1995) yields.
109: Supernova Ia are assumed to  originate from binary systems following
110: the  model of Thielemann et al. (1993). We have assumed a random
111: sampling of the lifetime of binary systems in the range 0.1-5 Gyrs.
112: 
113: Chemical elements are distributed within gas neighbouring particles by using a
114: spline kernel function.
115: In order to estimate the correct environment of stars when SN explode we
116: re-calculate the gaseous neighbours of the stellar
117:  particles each time elements
118: are released.
119: In this implementation we have not included the treatment of energy feedback.
120: 
121: %We performed tests of the model by running simulations of an isolated collapse
122: %of a gaseous sphere with $\lambda =0.1$ onto a NFW potential well.
123: %The disc has a total mass of $10^{9} {\rm M_{\odot}}$ and has been  resolved
124: %with 2500, 10000 and 40000 baryonic particles to test numerical effects.
125: %We run tests by letting the gas cool 
126: % according to primordial and metal-dependent
127: %cooling functions.
128:  
129: We  run cosmological tests of $\Lambda$-CDM cosmologies 
130:  consistent with the concordance
131: model ($\Omega_{\Lambda}=0.7$, $\Omega_{m}$=0.3, $H_o =70 \ {\rm km
132: s^{-1} Mpc^{-1}}$).
133:  The simulations correspond to a volume box of $10 h^{-1}$ Mpc 
134: ($h=0.7$)
135: with $2 \times 50^3$ particles. Three different runs
136: were performed using different cooling functions: solar, primordial and
137: metal-dependent. 
138: 
139: 
140: \section{Results}
141: 
142: The cosmological simulations are analyzed from a global point
143: of view and then, by identifying the gravitational bounded structures. 
144: In Figure 1 (left panel) 
145: we show the comoving star formation $\rho_{\rm SFR}$ and
146: metal production $\rho_{\rm Z}$ 
147: rates for the three runs with solar (solid lines),
148: primordial (dashed lines) and metal-dependent (dotted-dashed lines)
149: cooling functions.
150:  We can see that the largest
151: difference is found for the run with fixed solar cooling  and for 
152: $z \le 3$. 
153: The $\rho_{\rm SFR}$ of the metal-dependent cooling run
154:  can be up to  $\approx 50\%$ lower than that of the solar cooling test,
155: and up to $\approx 20\%$
156:  higher than the primordial  cooling run.
157: Similar relations  are found for the metal production
158: rate $\rho_{\rm Z}$ .
159:  Figure 1 (right panel)  shows the
160: fraction of gas with $T > 10^{5}$ K  for the three runs,
161:  with 
162: the largest
163: gas fraction associated to the primordial cooling run, as expected.
164: the difference between primordial and metal-dependent cooling are
165: localized in the densest regions.
166: 
167: 
168: At $z=0$ we identified galaxy-like objects (GLOs) at their virial radius by
169: combining the friends-of-friends algorithm with a density contrast method
170: in the three simulations (with solar, primordial and metal dependent cooling).
171:  We constructed the star formation history
172: of the  GLOs and analyzed the chemical properties of their stellar population
173: and interstellar medium.
174: In Figure 2a we show the SFR history of a typical GLO 
175: in the three runs. In the small box we have plotted the ratio
176: of the SFR of the solar and metal-dependent cooling to that of the 
177: primordial one.
178: We can see that the differences are important for $z <3$,
179: but do not present a regular behavior.
180:  This suggests that 
181: the non-linear evolution of the structure
182: determines  complicate star formation history and chemical patterns that can produce
183: non-linear responses to the characteristics of the cooling functions. Our model
184: will allow us to try to understand the link between
185: the dynamical evolution and the  star formation history and chemical patterns,
186: by studying the particular history of evolution of
187: GLOs in different environments.
188:  
189: \begin{figure}[t]
190: \plottwo{rates-3613.ps}{frac-glob.ps} 
191: \caption{Comoving star formation and metal production rate for
192: three different cosmological runs of the same initial condition, where the
193: gas has been allowed to cool according to solar  (solid lines),
194: primordial (dashed lines)
195: and metal dependent (dotted-dashed lines) cooling functions.
196:  As it can be seen in the right 
197: panels, the largest differences are found for $z <3$. We
198: found at most a difference of up to $\approx 20 \%$ between the comoving
199: star formation rates of primordial and metal
200: dependent cooling runs.
201: }
202:  \label{rates}            % for cross-references
203: \end{figure}
204: 
205: It was shown by Tissera, Lambas \& Abadi (1997) that the combination
206: of the history of star formation with models of population synthesis 
207:  models results in a powerful tool to cnonect theory with
208: observations. The treatment of the chemical enrichment in the simulations
209: improve the technique since each simulated stellar population has its age and
210: metallicity estimated consistently with the history of formation
211: in a hierarchical scenario.
212: With this new chemical code, we can also assess the effects of cooling 
213:  the gas according to its metallicity.
214: For this purpose, we constructed the synthetic spectral distribution 
215: of the GLOs but combining the SEDs of Bruzual \& Charlot (2003, BC03) 
216: according to the metallicity and age of stars.
217: We also estimated the simulated spectra (SS) of the GLOs by assuming
218: fixed metallicity at $Z=0.0001 Z_{\odot}$ and $Z=Z_{\odot}$ for the SEDs
219: of BC03.
220: In Figure 2b we show the SS for the GLOS of 
221: ~\ref{SED-glo1}a for the three runs.
222: Note that the SS has been calculated with the same SFR history
223: (~\ref{SED-glo1}a, solid line).
224: From this figure, we can see that assuming a low fixed  metallicity leads
225: to a very different energy distribution. For this GLO, the real SS is
226: more similar to that produced by assuming fixed solar metallicity.
227: 
228: We also analyzed the simulated spectra for the same GLO in the three
229: runs (solar, primordial and metal dependent cooling), and estimated
230: the SS by using the full information on age and metallicity.
231: Hence, in this case we will looking at the impact of different gas cooling
232: on the star formation and chemical history through the
233: differences in  the energy distribution.
234: %In this case, we show the SS of the whole stellar population as
235: %a function of redshift (note that we are distinguishing if the stars belong to
236: %the progenitor or to a substructure).
237: Figure 3 we show
238: the evolution of the SS for the  GLO of Figure 2  in the three runs.
239: As it can be seen, the GLOs in the solar cooling run exhibits the
240: more luminous spectra since because of the high efficient cooling, 
241: a large fraction of stars formed quickly before the gas has time to enriched.
242: In the case of the corresponding GLO in the runs with primoridal and
243: metal dependent cooling, the interstellar medium has time to get
244: enriched by previous star generations. Hence, in this case, the 
245: SS of the GLO metal-dependent cooling run is more similar to that
246: obtained from the primordial one.
247: 
248: 
249: 
250: 
251: \begin{figure}[t]
252: \plottwo{SFR-GLO7-16.ps}{SED-glo1-new.ps} 
253: \caption{Letf panel: Star formation rate history 
254: for a typical galaxy-like object  
255: three different cosmological runs of the same initial condition, where the
256: gas has been allowed to cool according to solar  (blue solid lines),
257: primordial ($Z=0.0001 Z_{\odot}$, green dashed lines)
258: and metal dependent (red dotted-dashed lines) cooling functions.
259: Right panel: Simulated spectra by using the star formation rate history
260: of the metal dependent cooling run and the SED of BC03 for
261: $Z=0.0001 Z_{\odot}$ (green lines), $Z=Z_{\odot}$ (blue lines)
262: and for the corresponding metallicities of the stars (red lines).
263: }
264:  \label{SED-glo1}            % for cross-references
265: \end{figure}
266: 
267: \begin{figure}[t]
268: %\plotone{CF-GLO1-evol.ps}
269: \plotfiddle{CF-GLO1-evol-new.ps}{2.4in}{0}{50}{40}{-160}{-85}
270: \caption{Simulated spectra of the GLO shown in Figure 2   as a function
271: of redshift for  the 
272: three different cosmological runs of the same initial condition, where the
273: gas has been allowed to cool according to solar  (blue solid lines),
274: primordial (green dashed lines)
275: and metal dependent (red dotted-dashed lines) cooling functions.
276: }
277:  \label{CF-glo1-evol}            % for cross-references
278: \end{figure}
279: 
280: 
281: 
282: \section{Conclusions}
283: 
284: We have incorporated a chemical evolution algorithm within
285: Gadget2 which allows the description of the enrichment of baryons
286: within a cosmological framework.  A first analysis shows
287: that in $\Lambda$-CDM scenarios,  the comoving star formation rate
288: could be underestimated by up to $ \approx 20\%$  
289: if the  primordial cooling function is adopted.
290: However, within individual galactic systems it is complicated
291: to disentangle 
292: the effects of cooling the gas consistently with
293: its metallicity owing to the non-linear growth of the structure
294: which affects the star formation,  chemical histories and the mixing of metals
295: in a non-linear way.
296:  First results show that the star formation history
297:  could be either
298: underestimated or overestimated by an order of magnitude depending
299: on the particular history path.
300: We also show the simulated spectra obtained by combining the age
301: and metallicity information of stars in the simulation with
302: the SEDs of Bruzual \& Charlot (2003).Future works will focus
303: on the study of  the link between the characteristic of the spectra and
304: history of evolution of the galactic systems.
305: 
306: 
307: %In Figure~\ref{frac} we show  a fraction of gas with $T >10^{5}$ K
308: %within the virial radius of GLOS as a function of the virial mass.
309: % As we can appreciate
310: %the solar cooling function is very efficient leaving  small fraction
311: %of warm gas. The primordial and metal-dependent cooling
312: %have higher and similar fractions, although the metal
313: %dependent run exhibits sistematically lower warm gas fractions.
314: %However, note that there is also a trend for the warm fraction 
315: %to increase with the virial mass and for the largest systems 
316: %(consistent with small groups), both runs agree
317: %very well reflecting the fact that metals have not escape to 
318: %intercluster medium (ICM) in our models. Multiphase-feedback implementations
319: %are needed to study the impact of the presence of metals in the
320: %ICM.
321: 
322: %\begin{figure}[t]
323: %\plotone{SFR-GLO7-16.ps} 
324: %\caption{ Star formation rate for a typical galaxy-like object in 
325: %a $\lambda$-CDM scenario where baryons have been let to cool according
326: %to  primordial (dashed lines),  metal dependent (dotted dashed lines)
327: %and solar (solid lines) cooling functions.
328: %In the small window we show the difference between the logarithm of
329: %the SFR of the solar (s)  and primordial (p)  cooling  runs respect to
330: %the metal dependent (m) one.}
331: % \label{sfr-glo1}            % for cross-references
332: %\end{figure}
333: 
334: 
335: 
336: 
337: \section*{Acknowledgments}
338: 
339: We thank Simon White and Volker Springel for useful discussions,
340: and V.Springel for making Gadget2 available for this work.
341: We acknowledge the support of Fundanci\'on Antorchas
342: and  DAAD that made this work possible. 
343: C. Scannapieco thanks the Alexander von Humboldt Foundation, the Federal Ministry
344: of Education and Research, and the Programme for Investment in the Future (ZIP)
345: of the German Government for partial support.
346: This work was partially funded by CONICET.
347: 
348: 
349: \begin{references}
350: 
351: \reference Balogh et al. 2004, submitted
352: 
353:  \reference Barton, E. J., Geller, M. J. \& Kenyon, S. J., 2000, ApJ, 530, 660
354: 
355: \reference Bruzual, G. \& Charlot, S., 2003, in preparation
356: 
357: \reference Freeman, K., \& Bland-Hawthorn, J., 2002, ARA\&A, 40, 487
358: 
359: \reference Lambas, D. G., Tissera, P. B., Alonso, M. S., \& 
360: Coldwell, G., 2003, MNRAS, in press
361: 
362: \reference Lia, C., Portinari, L,
363: \& Carrari, G. 2002, MNRAS, 339, 821
364: 
365: \reference Mosconi, M. B., Tissera, P. B., Lambas, D. G.,
366: \& Cora, S. A. 2001, MNRAS, 325, 34
367: 
368: \reference Springel, V., \& Hernquist, L., 2002, MNRAS, 333, 649
369: 
370: \reference Sutherland, R. S., Dopita, M. A., 1993, ApJS, 88, 253
371: 
372: \reference Thielemann, F. K., Nomoto, K., \& Hashimoto, M.  1993,
373:   Prantzos, N., Vangoni-Flam, E., Cass\'e N., eds.,
374: Origin and Evolution of the Elements,
375:  p.299
376: 
377: \reference Tissera, P. B., Lambas, D. G., Abadi, M., 1997, MNRAS,
378: 286, 384
379: 
380: \reference Yepes, G., Elizondo, D., \& Ascasibar, Y., 1998, Ap\&SS,
381: 263, 31
382: 
383: \reference Woosley, S. E., \& Weaver, T. A. 1995, ApJS, 101, 181
384: 
385: \end{references}
386: 
387: 
388: \end{document}
389: