astro-ph0512259/ms.tex
1: %\documentclass[manuscript]{aastex}
2: \documentclass[8pt,preprint2]{aastex}
3: \newcommand{\vdag}{(v)^\dagger}
4: \newcommand{\myemail}{skywalker@galaxy.far.far.away}
5: \usepackage{natbib}
6: \slugcomment{LA-UR-05-7544}
7: \shorttitle{H2 in white dwarf atmospheres}
8: \shortauthors{Kowalski}
9: 
10: \def\Teff{T_{\rm eff}}
11: \def\sss{\scriptscriptstyle}
12: \def\wig#1{\mathrel{\hbox{\hbox to 0pt{%
13:           \lower.5ex\hbox{$\sim$}\hss}\raise.4ex\hbox{$#1$}}}}
14: 
15: %********************************************
16: \begin{document}
17: 
18: \title{On the Dissociation Equilibrium of $\rm H_2$ in Very Cool, Helium-Rich White Dwarf Atmospheres}
19: \author{Piotr M. Kowalski 
20: }
21: 
22: \affil{Department of Physics and Astronomy, Vanderbilt University, Nashville, TN 
23: 37235-1807}
24: \affil{Los Alamos National Laboratory, MS F699, Los Alamos, NM 87545}
25: \email{kowalski@lanl.gov}
26: 
27: \begin{abstract}
28: 
29: We investigate the dissociation equilibrium of $\rm H_2$ in very cool, helium-rich white dwarf atmospheres.
30: We present the solution of the non-ideal chemical equilibrium for the dissociation of molecular hydrogen in a medium of dense helium.
31: We find that at the photosphere of cool white dwarfs of $T_{\rm eff}\rm=4000 \, K$, the non-ideality results in an increase 
32: of the mole fraction of molecular hydrogen by up to a factor of $\sim 10$, compared to the equilibrium value for the ideal gas. 
33: This increases the $\rm H_{2}-He$ CIA opacity by an order of magnitude and will affect the determination of the abundance of hydrogen
34: in very cool, helium-rich white dwarfs.
35: \end{abstract}
36: 
37: \keywords{dense matter -- stars -- atmospheres -- stars: white dwarfs}
38: 
39: 
40: \section{Introduction}
41: 
42: Several very cool white dwarfs with suspected $T_{\rm eff}\rm \wig< 4500 \, K$ have been discovered recently \citep{FAR,KM,G,OPP,Harris01,HOD,Ibata00,HR2}. 
43: Most of them are thought to posses helium-rich atmospheres with an very high $\rm He/ H \wig> 10^{3}$ ratio \citep{BAL,KM,G,BL,Bergeron01,OPP,HOD}.
44: In most cases, however, current atmosphere models fail to reproduce the observed spectra and photometry  
45: of these peculiar stars.  The reason, and there may be more than one, for this shortcoming of the models
46: is presently unknown. However, current models predict extreme physical atmospheric 
47: conditions for such stars, reaching densities of up to $\rm 2-3 \ g/cm^{3}$.  Under these conditions, the mostly ideal gas constitutive
48: physics used in published atmosphere models is demonstrably inadequate. A careful look at the dense matter
49: effects on the equation of state, chemistry, opacities, and radiative transfer is necessary to compute physically realistic models of these stars.
50: Several of these effects have been studied previously, such as refractive radiative transfer \citep{KS04}, the effects of fluid
51: correlations on $\rm He^-$ free-free and He Rayleigh scattering \citep{KSM05,IRS}, and the ionization of warm, dense helium \citep{KSM05,BSW}.
52: In this contribution we present an additional correction that arises in the dense fluid: The solution for the dissociation of molecular 
53: hydrogen in dense fluid helium, in the limit $\rm He/H>>1$.  
54: The relative importance of these corrections varies considerably, even more so when they are combined.
55: As several more dense matter effects remain unexplored, it is premature to ponder their implications for the analysis of the coolest
56: white dwarfs known and whether they will result in models that reproduce the data.  Nonetheless, incorporating adequate constitutive 
57: physics in atmosphere models is a necessary step to reach a proper understanding of these peculiar stars.
58: 
59: We introduce non-ideal effects into the equilibrium dissociation of molecular hydrogen
60: through a modification of the chemical potentials of $\rm  H \, \textrm{\scriptsize{I}}$ and $\rm H_2$ (section 2). 
61: We find that the strong interactions in the dense, helium-rich atmosphere 
62: results in a significant decrease in the dissociation fraction of molecular hydrogen, with a corresponding change in the 
63: $\rm H_{2}-He$ Collision-Induced Absorption (CIA) opacity, which is linear function of $n_{\rm H_{2}}$. 
64: In section 3, we illustrate the impact of the interactions on the $\rm H_{2}/H \,\textrm{\scriptsize{I}}$ ratio on a sequence of white dwarf atmosphere models
65: with $T_{\rm eff}\rm=4000 \, K$, a gravity of $\rm log \ \it g \rm = 8$ (cgs), and a homogeneous composition of ${\rm He/H} =10^{2},10^{4}, \ \rm and \ 10^{6}$,
66: where He/H is the number abundance ratio.
67: 
68: 
69: \section{The dissociation equilibrium of molecular hydrogen in a dense fluid}
70: \subsection{Theoretical approach}
71: The condition for chemical equilibrium (at a given density and temperature) for the dissociation reaction:
72: \begin{equation} \rm H_{2} \rightleftharpoons 2H \,\textrm{\scriptsize{I}} \label{1}\end{equation}
73: is given by \citep{CG}
74: \begin{equation} \mu_{\rm H_2}\rm -2 \it \mu_{\rm H \,\textrm{\tiny{I}}}\rm = 0,\end{equation}
75: where $\mu_{i}$ is a chemical potential of the species $i$ expressed as
76: \begin{equation} 
77: \mu_i= E_{0,i}+k_{B}T\ln\frac{n_{i}h^{3}}{Z_i(2\pi m_i k_B T)^{3/2}} + \mu_i^{non-id}. \label{91} 
78: \end{equation}
79: In the above equation, $k_{B}$ is the Boltzmann constant, $h$ the Planck constant, $T$ the temperature, 
80: $E_{0,i}$ the ground state energy, $n_{i}$ the number density, $Z_i$ the unperturbed internal partition function, and $m_{i}$ is the mass.
81: The first two terms on the r.h.s. of equation (\ref{91}) represent the ideal contributions of translational and internal degrees of freedom 
82: and $\mu_{i}^{non-id}$ is the non-ideal contribution to the chemical potential arising from the interparticle interactions in the fluid.
83: Setting $\mu_{i}^{non-id}\rm =0$ we recover the standard Saha equation for dissociation of molecular hydrogen:
84: \begin{equation} \beta^{id}=\frac{n_{\rm H_{2}}}{n_{\rm H \,\textrm{\tiny{I}}}^{2}}=\frac{Z_{\rm H_{2}}}{Z^{2}_{\rm H\,\textrm{\tiny{I}}}}\left[\frac{m_{\rm H_{2}}h^{2}}{2\pi m_{\rm H\,\textrm{\tiny{I}}}^{2} k_{B}T }\right]^{3/2}e^{D_{0}/{k_{B}T}} \label{3} \end{equation}
85: where $D_{0}=2E_{0,\rm H\,\textrm{\tiny{I}}}-E_{0,\rm H_{2}}=\rm 4.478 \, eV$ is the dissociation energy of the isolated hydrogen molecule.
86: Even for trace species, like $\rm H\,\textrm{\scriptsize{I}}$ or $\rm H_{2}$ in dense helium, the $\mu_{i}^{non-id}$ 
87: which arise from interactions with the atoms are not negligible and in principle should be comparable in magnitude to $\mu_{He}^{non-id}$. 
88:  
89: 
90: If we define the quantity $\Delta I$ as
91: \begin{equation} \Delta I = \mu^{non-id}_{\rm H_{2}}-2\mu^{non-id}_{\rm H\,\textrm{\tiny{I}}}, \label{10}\end{equation}
92: the non-ideal equilibrium equation can be written in the following form
93: \begin{eqnarray} \beta = 
94: \frac{n_{\rm H_{2}}}{n_{\rm H\,\textrm{\tiny{I}}}^{2}}
95: = \frac{Z_{\rm H_{2}}}{Z^{2}_{\rm H\,\textrm{\tiny{I}}}}\left[\frac{m_{\rm H_{2}}h^{2}}{2\pi 
96: m_{\rm H\,\textrm{\tiny{I}}}^{2} k_{B}T}\right]^{3/2}e^{(D_{0}-\Delta I)/{k_{B}T}} \nonumber 
97: \end{eqnarray}
98: \begin{equation} 
99:  =  \beta^{id}e^{-\Delta I/k_{B}T}. \label{6}
100: \end{equation}
101: 
102: Comparing (\ref{3}) with (\ref{6}), we see that the non-ideal effects on the dissociation equilibrium can be \it interpreted \rm as a change in the dissociation energy by a value of $\Delta I$. 
103: For simplicity, we will follow this interpretation hereafter. We emphasize that this description of the non-ideal contribution to 
104: the chemical equilibrium (Eq. 6) is {\it identical} to the occupation probability formalism of \citet{HM} if there
105: is only one bound state in the partition function (e.g. low temperature $\rm H \, \textrm{\scriptsize{I}}$) or if
106: the $\mu_{i,j}^{non-id}$ is the same for all bound states $j$ of the species $i$
107: ($\mu_{i,j}^{non-id}=\partial f(V,T,{n_{i,j}})/\partial n_{i,j}$ remains constant in Eq. (2.17) of \citet{HM}).
108: In both cases, the occupation probability can be factored out of the partition function 
109: and written as $e^{-\Delta I/k_BT}$ (Eq. 6).
110: 
111: 
112: 
113: In the atmosphere of cool white dwarfs, hydrogen exist mostly as $\rm H_2$ and $\rm H\,\textrm{\scriptsize{I}}$, 
114: and the $\rm H\,\textrm{\scriptsize{I}}/H_2$ ratio is governed by reaction (\ref{1}) only. For a given temperature, density $\rho$, and the atmosphere composition $y=\rm He/H$, 
115: the number densities of $\rm H_2$ and $\rm H\,\textrm{\scriptsize{I}}$ are:
116: \begin{equation} n_{\rm H\,\textrm{\tiny{I}}}=\frac{-1+\sqrt{1+8\beta n_{\rm tot}}}{4\beta} \label{12},\end{equation}
117: \begin{equation} n_{\rm H_{2}}=\it \beta n^{\rm 2}_{\rm H\,\textrm{\tiny{I}}} \label{13},\end{equation}
118: where
119: \begin{equation} n_{\rm tot}=n_{\rm H\,\textrm{\tiny{I}}}+ 2 n_{\rm H_{2}}=\frac{\rho}{m_{\rm H}+{y}m_{\rm He}}. \label{14}\end{equation}
120: refers to the hydrogen species only. 
121: 
122: \begin{figure*}[t]
123: \centering
124: \epsscale{1.75}
125: \plotone{f1.eps} 
126: \caption{The $\rm H-He$ and $\rm H_2-He$ pair interaction potentials.
127: The dashed line represents the radius (horizontal axis) of a sphere whose volume equals the effective volume of two H atoms 
128: colliding with a kinetic energy of ${3\over 2}k_BT$ (see text).}
129: \end{figure*}
130: 
131: 
132: \subsection{Computation of the
133: $\rm H_2$ dissociation equilibrium}
134: The non-ideal contributions to the chemical potentials of $\rm H\,\textrm{\scriptsize{I}}$ and $\rm H_2$ were obtained through the numerical solution of the Ornstein-Zernike equation 
135: in the Percus-Yevick (PY) approximation \citep{M}. For the 
136: $\rm H - He$ interaction we use the pair potential of \citet{SH}, and for the $\rm H_2 - He$ interaction the pair potential given by \citet{RE} (Figure 1). 
137: Both potentials are from \it ab initio \rm quantum mechanical calculations and are in good agreement with the independent calculations of \citet{TY} for $\rm H-He$ 
138: and of \citet{TAO} and \citet{SG} for $\rm H_2-He$. 
139: As we consider a helium-dominated mixture ($\rm He/H\wig>10^{2}$), the $\rm H-H$, $\rm H_2 - H$, and $\rm H_2 - H_2$ interactions can be neglected.  
140: 
141: High pressure experiments have shown that \it ab initio \rm pair potentials are too repulsive to describe dense systems where N-body effects become important \citep{Nellis,Ross}.
142: The softening of the pair potentials at high densities can only be quantified experimentally or, alternatively, estimated with N-body quantum mechanical calculations.
143: Since neither are available for mixtures of trace hydrogen in helium, we resort to \it ab initio \rm potentials. The net effect on the dissociation equilibrium depends
144: on the \it relative \rm softening of the potentials (Eq. 5) and is therefore less sensitive to N-body effects than the individual potentials.
145: We also calculated the chemical potentials in 
146: the Hypernetted Chain approximation and found them to agree within $5\%$ with the PY values up to $\rm 2 \, g/cm^{3}$. 
147: Since the PY approximation is better suited for short range potentials such as the ones we use here, 
148: we estimate that our PY calculations are reliable up to at least $\rm 2 \, g/cm^{3}$.
149: 
150: \begin{figure*}[t]
151: \centering
152: \epsscale{1.75}
153: \plotone{f2.eps} 
154: \caption{(a) Variation of the effective dissociation 
155: energy $D_{0}-\Delta I$ as a function of the density of helium for temperatures of
156: $\rm 3000\, K$ (solid line), $\rm 4000\, K$ (dotted line), and $\rm 5000\, K$ (dashed line). (b) The density $\rho$ at the photosphere 
157: of cool white dwarf atmosphere models of $T_{\rm eff}\rm=4000\, K$, $\rm log \ \it g \rm = 8$ (cgs), 
158: and various $\rm He/H$ compositions.}
159: \end{figure*}
160: 
161: 
162: For the internal partition functions $Z_i$, we use expressions for the electronic ground state of the unperturbed hydrogen molecule accounting 
163: for the vibrational/rotational excitations \citep{HH}, 
164: and set $Z_{\rm H \,\textrm{\tiny{I}}}=2$ for hydrogen atom. This approximation is justified as the electronic excitation energies of both species are large 
165: and for temperatures of a few thousands degrees the populations of the electronic excited levels are extremely small. 
166: However, there is significant thermal excitation of the rotational and vibrational levels of $\rm H_2$ 
167: and the effect of the dense medium on $Z_{\rm H_2}$ must be considered.  Since the molecule
168: does not have spherical symmetry, it is possible that its rotation modes will be hindered by
169: interactions with neighbors at very high density.  Furthermore, the energies of the vibrational
170: levels, which are spaced by a mere $0.54\,\rm eV$, could also be significantly shifted by these 
171: interactions. Both of these
172: effects are discussed by \citet{SCH} and can be neglected under the
173: present conditions.  The use of the unperturbed rotational and vibrational levels of $\rm H_2$ is thus
174: justified in the calculation of $Z_{\rm H_2}$.
175: 
176: \begin{deluxetable}{crrrrrrrrrrr}[hb]
177: %\footnotesize
178: \tablecaption{The change in the dissociation energy 
179: $\Delta I$ (eV) for representative densities and temperatures.}
180: \tablewidth{0pt}
181: \tablehead{& \colhead{
182: $\rho$ $\rm (g/cm^{3})$}   
183: && \colhead{$T \rm (K) : \ \ \ \ 2000\rm$} 
184: & \colhead{$4000\rm$} 
185: & \colhead{$6000\rm$} & \colhead{$8000\rm$}}
186: \startdata
187: 
188:  &0.001 &&$-$0.001 &$-$0.001 &$-$0.001 &$-$0.001 \\
189:  &0.010 &&$-$0.007 &$-$0.010 &$-$0.011 &$-$0.012 \\
190:  &0.100 &&$-$0.083 &$-$0.108 &$-$0.120 &$-$0.127 \\
191:  &0.250 &&$-$0.242 &$-$0.297 &$-$0.323 &$-$0.334 \\
192:  &0.500 &&$-$0.579 &$-$0.668 &$-$0.697 &$-$0.702 \\
193:  &0.750 &&$-$1.006 &$-$1.075 &$-$1.088 &$-$1.070 \\
194:  &1.000 &&$-$1.471 &$-$1.494 &$-$1.469 &$-$1.414 \\
195:  &1.500 &&$-$2.011 &$-$2.094 &$-$1.963 &$-$1.825 \\
196:  &2.000 &&$-$2.189 &$-$2.387 &$-$2.192 &$-$1.988 \\
197: 
198: \enddata
199: \end{deluxetable}
200: 
201: 
202: \begin{figure*}[t]
203: \centering
204: \epsscale{1.75}
205: \plotone{f3.eps} 
206: \caption{The density profiles of white dwarf atmosphere models of 
207: $T_{\rm eff}\rm=4000\,K$, $\rm log \ \it g \rm = 8$ (cgs), 
208: and $\rm He/H$ composition of $\rm He/H=10^6,10^4 \ and \ 10^2$ (from top to bottom, respectively).}
209: \end{figure*}
210: 
211: 
212: 
213: More importantly,  the excited rotational and vibrational 
214: levels of the $\rm H_2$ molecule may be 
215: differentially affected by the interactions.  This relates to the general problem of the
216: cutoff of the sum over states
217: in the internal partition function and is the subject of an extensive 
218: literature. This truncation of the partition function sum is
219: often described by an occupation probability formalism \citep{HM} and in the case of interactions 
220: between neutral particles, is generally described in the
221: context of the volume occupied by each bound state.  In the case of a diatomic molecule, the volume
222: occupied increases with the rotational quantum number (centrifugal stretching) and with
223: the vibrational quantum number (larger amplitude of vibration).  The first effect is very small 
224: and can be safely neglected.  We estimate the vibrational stretching by considering the change in 
225: the equilibrium internuclear separation  of the molecule with the vibrational quantum number.
226: Because of the anharmonicity of the potential, higher vibrational levels correspond to larger
227: equilibrium separations and a larger average molecular volume.
228: Based on the potential curve of the $\rm H_2$ molecule \citep{KW} we find that
229: for $\rm T < 6000\,K$ and densities of up to $\rm 2\,g/cm^3$, the vibrational excitation is
230: largely limited to the lowest five levels and the partition function is reduced by a few percent
231: when the larger volume of the excited states is taken into account.
232: This is a small effect which can be neglected in view of the other uncertainties in the model
233: such as the $\rm N$-body effects on the interaction potentials.  Thus, the use of the internal partition 
234: function of the isolated $\rm H_2$ molecule, with all states being affected identically by the interactions
235: with He (Eq. 6) is a very good approximation.
236: 
237: 
238: 
239: \section{Results and Discussion}  
240: \subsection{The H$_2$ dissociation equilibrium in fluid helium}
241: In general, the $\Delta I$ defined by (\ref{10}) is a function of the density of helium and the temperature (Fig. 2a), but is independent of $\rm He/H$
242: ratio if $\rm H$ is a trace species ($\rm He/H \wig>10^2$). For the conditions 
243: in white dwarf atmospheres, the temperature dependence is weak. The change in the dissociation energy $\Delta I$ is negative, making $\rm H_2$ more stable in dense helium.
244: This may be qualitatively understood by comparing the effective volume occupied by one $\rm H_2$ molecule to that of two $\rm H$ atoms in He. 
245: Lets assume that these effective volumes are the spaces around each of the particles where the energy of their interaction with $\rm He$ atoms is greater than 
246: the average thermal kinetic energy of the particles ${3\over 2}k_BT$. This radius is the classical distance of closest approach in a collision.
247: On Figure 1, we plotted the radius of a sphere whose volume equals 
248: the effective volume of two hydrogen atoms, as a function of $k_BT$ (dashed curve). The corresponding radius for the hydrogen molecule is represented by the
249: $\rm H_2-He$ potential (solid curve). 
250: The effective volume of $\rm H_2$ is smaller than that of two $\rm H$ atoms for $k_BT \wig<1\,$eV.
251: Since the exclusion of a greater volume results in a decrease of the entropy of He, $\rm H_2$ is more stable in dense helium than two H atoms. 
252: In dense hydrogen, we have the opposite situation where H$_2$ is less stable at high density (leading to pressure dissociation
253: of H$_2$) because H$_2$--H$_2$ is more repulsive than the sum of the H--H and H--H$_2$ interactions \citep{SCH}.
254: 
255: \begin{figure*}
256: \centering
257: \epsscale{1.75}
258: \plotone{f4.eps} 
259: \caption{Left panels: The number density of atomic (solid) and molecular (dotted) hydrogen with (thick lines) and without (thin lines)
260: the non-ideal chemical equilibrium along the atmosphere profiles of Figure 3.
261: Right panels: Corresponding contributions to the photospheric opacity: $\rm H_2-He$ CIA (solid),  
262: $\rm He^{-}$ free-free of Iglesias, Rogers, \& Saumon (2002) (long dashed), $\rm H^-$ bound-free 
263: and free-free (dotted) and Rayleigh scattering (short dashed).}
264: \end{figure*}
265: 
266: Numerical values for $\Delta I$ obtained from our model are given in Table 1. 
267: Together with relations (\ref{6})--(\ref{14}) this table allows for an easy implementation of the non-ideal abundances of atomic and molecular hydrogen 
268: into white dwarf atmosphere codes. 
269: 
270: 
271: 
272: The non-ideal recombination of H becomes significant when $-\Delta I/k_BT\wig> 1$.  For an atmosphere with $T_{\rm eff}=4000\,$K,
273: this will occur at the photosphere if  $\rho \wig>0.3\,$g/cm$^3$, a density easily achieved in He-rich models. 
274: As $D_0-\Delta I$ increases with increasing helium density, a lower dissociation fraction will result. 
275: A strong deviation from the ideal-gas abundance of $\rm H_2$ is expected.   
276: 
277: \subsection{The $\rm H_2$ dissociation equilibrium in He-rich white dwarf atmospheres}
278: We have applied the non-ideal correction to the $\rm H_2$ dissociation equilibrium to very cool white dwarf atmospheres. 
279: To show the effect, we solved for the chemistry and opacity along fixed $T-\rho$ white dwarf atmosphere profiles of $T_{\rm eff}=4000\,\rm K$, 
280: a gravity of $\rm log \ \it g \rm = 8$ (cgs), and various $\rm He/H$
281: ratios. Figure 3 shows the atmospheric density profiles. These profiles were obtained with a code that solves for static, 
282: plane-parallel, and LTE white dwarf atmospheres in thermal and hydrostatic equilibrium, accounting for refraction \citep{KS04}. 
283: The resulting new abundances of $\rm H\,\textrm{\scriptsize{I}}$ and $\rm H_2$ in our models are given on Figure 4 (left panels). As expected, the amount of molecular hydrogen 
284: increases significantly throughout the atmosphere for $\rm He/H\wig>10^{3}$, and $\rm H_2$ can become the dominant hydrogen species at the photosphere. 
285: 
286: The most important sources of opacity in helium-rich white dwarf atmospheres are $\rm He^{-}$ free-free, Rayleigh scattering, and $\rm H_{2}-He$ CIA 
287: \citep{HN}. Several physical and chemical effects that alter these opacity sources have already been discussed: the change in the 
288: free-free opacity and the Rayleigh scattering of helium, as caused by the strong correlations in the dense fluid \citep{IRS}, the change in the number density 
289: of free electrons \citep{KSM05,BSW}, the presence of heavy elements \citep{Bergeron01}, and the formation of trace species like $\rm He_{2}^{+}$ \citep{ML} and $\rm HeH^{+}$ \citep{HR}.
290: The effect on non-ideal chemistry on the $\rm H_2$ dissociation on the opacity is shown in the right hand panels of Figure 4. The increase in 
291: number density of molecular hydrogen results in an increase of 
292: the $\rm H_2-He$ CIA opacity. This follows from the linear dependence of $\rm H_2-He$ CIA opacity on $n_{\rm H_2}$. 
293: On the other hand, the $\rm H^-$ bound-free and free-free opacities are reduced, due to the decrease in the abundance of atomic hydrogen.    
294: Based on the atmospheric structures used here, the effects of the non-ideal chemistry of H$_2$ in dense helium are maximal for He/H $\sim 10^3$.  This
295: arises from a competition between the need for a high He/H ratio to increase the non-ideal effects by increasing the density at the photosphere (Fig. 2b)
296: and the need for a high enough hydrogen content in the atmosphere so that H$_2$ (or H) contributes to the total opacity (Fig. 4, right hand side).
297: 
298: \begin{figure*}[t]
299: \centering
300: \epsscale{1.75}
301: \plotone{f5.eps} 
302: \caption{The synthetic spectrum for the white dwarf atmosphere models of 
303: $T_{\rm eff}\rm=4000\,K$, $\rm log \ \it g \rm = 8$ (cgs), 
304: and $\rm He/H=10^{3}$ with (thick line) and without (thin line) the non-ideal dissociation equilibrium of $\rm H_2$.}
305: \end{figure*}
306: 
307: 
308: The impact of the non-ideal dissociation equilibrium of $\rm H_{2}$ on the synthetic spectrum of a He-rich white dwarf model is shown in Figure 5.
309: For a white dwarf atmosphere model of $T_{\rm eff}\rm=4000K$, $\rm log \it \ g\rm=8$, and $\rm He/H=10^{3}$, the rise in abundance 
310: of $\rm H_2$ increases significantly the opacity in the infrared, causing a redistribution of the flux toward 
311: shorter wavelengths. This effect on the spectrum of cool white dwarfs is largest 
312: for $T_{\rm eff}\rm =4000K-4500K$, where $\rm H_2$ is partially dissociated. At lower effective temperatures hydrogen exists mostly 
313: in molecular form and the effect of the non-ideal dissociation equilibrium 
314: on the spectrum vanishes at $T_{\rm eff}\rm \sim 3000K$. 
315: 
316: The determination of the $\rm He/H$ composition of very cool He-rich white dwarfs depends mostly on the relative $\rm He-H_2$ CIA, and $\rm He^-$ free-free 
317: opacities (Fig. 4, right hand side), Since the non-ideal recombination increases the CIA opacity, this effect, taken by itself, will result in a higher 
318: value for the $\rm He/H$ ratio inferred from a given stellar spectrum. On the other hand, current calculations 
319: for the ionization fraction of dense helium and the low-frequency behavior of the $\rm He^{-}$ free-free opacity
320: are uncertain and the $\rm He^{-}$ free-free opacity could be underestimated by as much as $\sim 2-4$ orders of magnitude \citep{KSM05}. It is therefore too early to
321: draw conclusions about the atmospheric composition of the coolest white dwarfs. 
322: However, we emphasize that, as the chemical equilibrium (\ref{6}) is not affected by the number of 
323: free electrons in the atmosphere, our solution for the non-ideal abundance of $\rm H_2$ will not be affected by weak 
324: ionization of $\rm He$, and the results are limited only by the validity of the $\rm H_2-He$ and $\rm H-He$ interaction potentials.
325: 
326: 
327: 
328: The non-ideal dissociation equilibrium in pure hydrogen atmospheres is discussed in \citet{SJ}. As hydrogen is much more opaque than helium, 
329: the density at the photospheres of these stars
330: is much smaller than for helium-rich atmosphere stars of the same effective temperature. The non-ideal effects are therefore weaker and become important
331: only in stars of lower effective temperatures and/or higher gravity, where high densities are achieved. \citet{SJ} show that the non-ideality matters in pure-hydrogen atmospheres 
332: with $\rm T_{eff}\wig<2500K$. White dwarfs with hydrogen-rich atmospheres of such low effective temperatures have not yet been identified.    
333: 
334:        
335: \section{Conclusions}
336: Recent discoveries of a number of cool white dwarfs with peculiar spectral energy distribution represent a challenge in the modeling of very cool white dwarf atmospheres, 
337: as the spectra of these stars cannot be fitted with existing models. We believe that the main reason of this shortcoming of the models is 
338: the poorly explored, extreme physical regime found inside these atmospheres. We presented a correction to the abundance of $\rm H_2$ that arises 
339: from the strong interaction of hydrogen molecules and atoms in a dense, fluid helium medium. We have found that in the dense, helium-rich, cool white dwarf 
340: atmospheres the formation of $\rm H_2$ is more favorable than in the ideal gas description. 
341: For white dwarfs of $T_{\rm eff}\rm=4000 \, K$ and $\rm He/H\wig>10^{3}$ the abundance of molecular hydrogen increases by an order of magnitude, with a corresponding increase in 
342: the $\rm H_2-He$ CIA opacity by the same factor. This improvement is a new, significant effect that must be included in realistic modeling 
343: of very cool white dwarf atmospheres. 
344: 
345: We have shown that the non-ideal effects affect strongly the abundances of even trace species in the atmosphere. This strongly suggest the
346: necessity of revising the abundances of other trace species with significant opacity for similar effects. We expect that the study of their abundances 
347: and absorption processes in fluid helium will significantly improve our understanding of the atmospheric physics, composition, and evolution of the oldest and coolest white dwarfs.
348: 
349: I thank D. Saumon for useful discussions and the referee, P. Bergeron, for suggestions that improved the clarity of the manuscript.
350: This research was supported by the United States Department of Energy under contract W-7405-ENG-36.
351: %\clearpage
352: 
353: \begin{thebibliography}{}
354: \bibitem[Bergeron, 2001]{Bergeron01} Bergeron, P. 2001, \apj, 558, 369
355: \bibitem[Bergeron \& Leggett, 2002]{BL} Bergeron, P., \& Leggett, S. K. 2002, \apj, 580, 1070
356: \bibitem[Bergeron et al., 2005]{BAL} Bergeron, P., Ruiz, M. T., Hamuy, M., Leggett, S. K., Currie, M. J., Lajoie, C. P., \& Dufour, P. 2005, \apj, 625, 838
357: \bibitem[Bergeron et al., 1995]{BSW} Bergeron, P., Saumon, D., \& Wesemael, F. 1995, \apj, 443, 764 
358: \bibitem[Cox \& Giuli, 1968]{CG} Cox, J. P., \& Giuli, R. T. 1968, Principles of stellar structure, Vol. 1 
359:          (New York: Gordon and Breach), Chap. 2.
360: \bibitem[Farihi, 2005]{FAR} Farihi, J. 2005, \aj, 129, 2382	 
361: \bibitem[Gates et al., 2004]{G} Gates, E et al. 2004, \apjl, 612, L129
362: \bibitem[Hummer \& Mihalas, 1988]{HM} Hummer, D. G., \& Mihalas, D. 1988, \apj, 331, 794
363: \bibitem[Hansen, 2004]{HN} Hansen, B. 2004, Physics Reports, 399, 1
364: \bibitem[Harris et al., 1999]{HR2} Harris, H. C., Dahn, C. C., Vrba, F. J., Henden, A. A., Liebert, J., Schmidt, G. D., \& Reid, I. N. 1999, \apj, 524, 1000   
365: \bibitem[Harris et al., 2001]{Harris01} Harris, H. C. et al., 2001, \apj, 549, L109   
366: \bibitem[Harris et al., 2004]{HR} Harris, G. J., Lynas-Gray, A. E., Miller, S., \& Tennyson, J. 2004, \apjl, 617, L143
367: \bibitem[Hodgkin et al., 2000]{HOD} Hodgkin, S. T., Oppenheimer, B. R., Hambly, N. C., Jameson, R. F., Smartt, S. J., \& Steele, I. A. 2000, Nature, 403, 57
368: \bibitem[Huber \& Herzberg, 1979]{HH} Huber, K. P. \& Herzberg, G. Molecular Spectra and Molecular Structure. IV. Constants of Diatomic Molecules (Princeton 1979: Von Nostrand)
369: \bibitem[Ibata et al., 2000]{Ibata00} Ibata, R., Irwin, M., Bienaym\'e, O., Scholz, R., \& Guibert, J. 2000, \apj, 532, L41
370: \bibitem[Iglesias et al., 2002]{IRS} Iglesias, C. A., Rogers, F. J., \& Saumon, D. 2002, \apjl, 569, L111
371: \bibitem[Kilic et al., 2005]{KM} Kilic, M., Munn, J. A., Harris, H. C., Liebert, J., von Hippel, T., Williams, K. A., 
372: Metcalfe, T. S.,Winget, D. E., \& Levine, S. E. 2005, preprint (astro-ph/0503601)
373: \bibitem[Kolos \& Wolniewicz, 1965]{KW} Kolos, W., \& Wolniewicz, L. 1965, J. Chem. Phys., 43, 7
374: \bibitem[Kowalski \& Saumon, 2004]{KS04} Kowalski, P. M., \& Saumon, D. 2004, \apj, 607, 970
375: \bibitem[Kowalski et al., 2005]{KSM05} Kowalski, P. M., Saumon, D., \& Mazevet, S. 2005, ASP Conf. Ser. 334, White Dwarfs, ed. Koester, D., \& Moehler, S. (San Francisco: ASP), 203 
376: \bibitem[Malo et al., 1999]{ML} Malo, A., Wesemael, F., \& Bergeron, P. 1999, \apj, 322, 891
377: \bibitem[Martynov, 1992]{M} Martynov, G. A. 1992, Fundamental theory of liquids (Bristol: Adam Hilger), Chap. 5, 6 and 10
378: \bibitem[Nellis et al., 1984]{Nellis} Nellis, W. J., Holmes, N. C., Mitchell, A. C., Trainor, R. J., Governo, G. K., Ross, M., \& Young, D. A. 1984, Phys. Rev. Lett., 53, 1248
379: \bibitem[Oppenheimer et al., 2001]{OPP} Oppenheimer, B. R., Hambly, N. C., Digby, A. P., Hodgkin, S. T., \& Saumon, D. 2001, Science, 292, 698
380: \bibitem[Ree, 1983]{RE} Ree, F. H. 1983, J. Chem. Phys., 78, 409
381: \bibitem[Ross et al., 1983]{Ross} Ross, M., Ree, F. H., \& Young, D. A. 1983, J. Chem. Phys.,  79, 1487
382: \bibitem[Saumon \& Chabrier, 1991]{SCH} Saumon, D., \& Chabrier, G. 1991, Phys. Rev. A, 44, 5122  
383: \bibitem[Saumon \& Jacobson, 1999]{SJ} Saumon, D., \& Jacobson, 1999, \apjl, 511, L107
384: \bibitem[Shafer \& Gordon, 1976]{SG} Shafer, R., \& Gordon, R. G. 1973, J. Chem. Phys., 58, 5422
385: \bibitem[Shalabi et al., 1998]{SH} Shalabi, A. S., Eid, Kh. M., Kamel, M. A., \& El-Barbary, A. A. 1998, Physics Letters A, 239, 87
386: \bibitem[Tang \& Yang, 1990]{TY} Tang, K. T., \& Yang, X. D. 1990, Phys. Rev. A, 42, 311
387: \bibitem[Tao, 1993]{TAO} Tao, F. 1993, J. Chem. Phys., 100, 4942
388: \end{thebibliography}
389: \end{document}
390: