1: %\documentstyle[prb,preprint,aps]{revtex}
2: \documentstyle[prb,aps]{revtex}
3: %\documentstyle[eqsecnum,aps]{revtex}
4:
5: %\newcommand{\mwcd}[1]{\marginpar{\protect\scriptsize \protect\setlength
6: % {\baselineskip}{8pt} to editor- MS revision: #1}}
7: %\setlength {\marginparwidth}{7.0cm}
8:
9: \begin{document}
10: \draft
11:
12: \twocolumn[\hsize\textwidth\columnwidth\hsize\csname @twocolumnfalse\endcsname
13:
14: \title
15: {
16: The Equation of State and the Hugoniot of Laser Shock-Compressed Deuterium.
17: }
18: %
19: \author
20: {
21: M.W.C. Dharma-wardana\cite{byline1} and Fran\c{c}ois Perrot\cite{byline2}
22: }
23: %
24: \address
25: {
26: National Research Council, Ottawa,Canada. K1A 0R6
27: }
28: \date{-- Dec 2001}
29: \maketitle
30: %
31: \begin{abstract}
32: The equation of state and the shock Hugoniot of deuterium are
33: calculated using a first-principles approach,
34: for the conditions of the
35: recent shock experiments.
36: We use density-functional theory within a
37: a classical mapping of the quantum fluids
38: $\lbrack$Phys. Rev. Letters, {\bf 84}, 959 (2000)$\rbrack$.
39: The calculated
40: Hugoniot is close to the Path-Integral Monte Carlo(PIMC)
41: result. We also consider the {\it quasi-equilibrium} two-temperature
42: case where the Deuterons are hotter than
43: the electrons; the resulting quasi-equilibrium
44: Hugoniot mimics the laser-shock
45: data. The increased compressibility
46: arises from hot $D^+-e$
47: pairs occurring close to
48: the zero of the electron chemical potential.
49: \end{abstract}
50: \pacs{PACS Numbers: 62.50.+p, 02.70.Lq, 05.30.-d }
51: %\vspace{0.5in}
52: %\hspace{0.5in} see file /usr/people/chandre/tekstuff/hud/ms1.tex
53: %
54: %\hspace{0.5in} date: \today 3-12-2001
55: \vskip2pc]
56: \narrowtext
57: %
58:
59: Hydrogen and its isotopes have been extensively studied, yet
60: laser-shock experiments of Da Silva et al.\cite{dasilva},
61: Collins et al.\cite{collins}, and Mostovych et al.\cite{mosto}
62: produced unexpected disagreement
63: with the equation of state (EOS) of the
64: SESAME database.\cite{sesame}.
65: The disagreement with SESAME occurs
66: for temperatures T
67: with $\sim$ 0.8 eV $< T <$
68: $\sim$ 10 eV, and
69: for densities $1.8 < r_s < \sim 2$, where
70: the electron-density parameter
71: (in atomic units) is,
72: $r_s$ = $(3/4\pi\bar{n})^{1/3}$.
73: Here $\bar{n}$ is the electron-number density (au.).
74: The coupling constant
75: $\Gamma$=(potential energy)/(kinetic energy) ranges
76: from 1 to $\sim 30$.
77:
78: The electrons change from a degenerate liquid to
79: a classical system in the
80: anomaly-regime (AR),
81: while the $e-
82: D^+$ interaction is at the threshold of bound-state
83: formation.
84: Thus the AR poses a difficult, strongly-correlated
85: many-body problem of wide interest -
86: from astrophysics and fusion to materials science.
87: Hence a flurry of activity
88: has focused on the deuterium
89: EOS and its Hugoniot.\cite{hug}
90: These involve intuitive approaches
91: (called ``chemical models''),\cite{ross}
92: assuming the existence of molecules etc.,
93: and first-principles approaches, e.g., density-functional theory (DFT),
94: or quantum Monte Carlo (QMC).
95: Standard DFT
96: treats the ions as
97: an external potential, optimized via
98: Car-Parrinello type
99: techniques, while the
100: local-density-approximation
101: (LDA), or simpler
102: tight-binding approaches are used for the electrons.\cite{dftcalcs,lenox}
103: Path-integral
104: Monte Carlo (PIMC),
105: a finite-T QMC approach, has been used for this problem.\cite{pimc}
106:
107: PIMC and DFT calculations
108: yield shock Hugoniots close to SESAME,
109: showing no
110: strong softening.
111: Recent experiments from Sandia
112: are also close to the SESAME Hugoniot.\cite{knud}
113: However a high degree of agreement exists
114: in the laser-shock results.\cite{dbleshock}
115:
116: Our objective is to present new
117: first-principles results
118: for the EOS and Hugoniot of deuterium, using methods
119: quite independent of
120: previous methods. Our method
121: is computationally simple and uses
122: one-dimensional integral equations
123: based on DFT. We calculate
124: all the pair-distribution functions (PDFs),
125: $g_{ij}$, with
126: $i=$1,2,3; (or e$_\uparrow$, e$_\downarrow$, and D$^+$ nuclei)
127: of a three component
128: system.
129: Being based on the PDFs, it is manifestly non-local
130: and has no self-interaction errors.
131: It easily adapts to the
132: quasi-stationary two-temperature case with
133: $T_D\ne T_e$.
134: Consider the fluid with a deuterium nucleus at the origin,
135: and let the one-body densities of the electrons and $D^+$ be
136: $n(r)$ and $\rho(r)$.
137: Then $n(r)$
138: is really $n_{De}(r)=\bar{n}g_{De}(r)$. The free energy $F$
139: is a functional of the form
140: $F[n(r),\rho(r)]$. Thus, taking functional
141: derivatives, we have {\it two} coupled Kohn-Sham-Mermin equations:
142: \begin{eqnarray}
143: \delta F[n(r),\rho(r)]/\delta n(r)&=&0 \\
144: \delta F[n(r),\rho(r)]/\delta \rho(r)&=&0
145: \end{eqnarray}
146: As shown in
147: ref.~[\onlinecite{hyd0,ilciacco}]
148: Eq.(1) leads to
149: a (quantum) Kohn-Sham (K-S) equation for the electrons,
150: while Eq.(2),
151: a classical K-S
152: equation, becomes the Hyper-Netted-Chain (HNC)\cite{hncref}
153: equation for a specific
154: choice of the correlation potential (there being no exchange potential
155: in the classical system). Thus the K-S eigenfunctions, as well as the
156: $n(r)$, $\rho(r)$, in the Hydrogen fluid were calculated by
157: solving Eqs. (1-2).\cite{hyd0}
158: In Car-Parrinello approaches only
159: the electrons are treated in
160: DFT, while the N ions at sites $\vec{R_i}$ are explicitly treated.
161: Here both electrons and ions are treated using
162: their distributions.
163: Our hydrogen
164: calculations were later
165: confirmed by lengthy QMC.\cite{kwon}
166: However, while Eqs.(1-2) provide the $g_{ie}(r)$, $g_{ii}(r)$, i=ion,
167: and the K-S
168: states in the fluid,
169: the $g_{ee}(r)$ was available only in LDA.
170:
171: Recently, we showed how the electrons at the
172: physical temperature $T_e$
173: could be replaced by an equivalent classical system at
174: $T_{ee}=(T_e^2+T_{eq}^2)^{1/2}$,
175: such that
176: the quantum effects
177: are correctly incorporated.
178: A simple expression for the electron ``quantum temperature'' $T_{eq}$
179: as a function of $r_s$ was given
180: in ref.\onlinecite{prl}.
181: Application of the method, denoted the ``Classical mapping
182: of quantum systems
183: using the Hyper-Netted-Chain equation (CHNC)'', to
184: 3-D and 2-D uniform electron liquids at
185: $T=0$ and finite-$T$,
186: showed excellent agreement
187: of the $g_{ij}$,
188: energies, etc., with QMC results, for even very strongly
189: coupled situations.\cite{prl,prb00,2D}
190: Deuterons in a uniform neutralizing background
191: are mathematically identical to the electron system,
192: except for changes of scale,
193: (e.g, for deuterons, $r_{sD}=r_s(M_D/m_e)$ etc.,
194: $M_D$ is the deuteron mass and $m_e=1$).
195: Hence the $D^+$-quantum temperature $T_{Dq}$ is also available, and
196: is negligible in the
197: regime of interest; thus $D^+$ are treated as classical particles.
198:
199:
200: The mean
201: densities $\bar{\rho}$ and $\bar{n} $ are equal since
202: the nuclear charge $Z$ = 1.
203: Consider a fluid of total density $n_{tot}$, with
204: three species. Let
205: $x_i$ = $n_i/n_{tot}$, $n_{tot}=\bar{\rho}+\bar{n}$.
206: The physical temperature is $T$,
207: while the inverse temperature
208: of the electrons
209: is $1/\beta_{ee}$,
210: with $1/\beta_{ee}$= ${\surd{(T^2+T_q^2)}}$.
211: For $D^+$, the classical mapping $T_{DD}=1/\beta_{DD}$ is
212: taken to be $T$.
213: The inverse temperature
214: $\beta_{ij}$ between electrons
215: and deuterons is actually not needed since the {\it product},
216: $\beta_{eD}\phi_{eD}(r)$
217: is completely determined
218: by the quantum mechanics
219: of the problem (see below).
220: However, for the present application
221: it is the mean kinetic energy of
222: the $D^+-e$ pair, i.e., $T_{eD}=(T_{ee}+T_{DD})/2$, consistent with the
223: classical picture used here.
224:
225: The classical equations for the PDFs
226: and the Ornstein-Zernike(OZ) relations are:
227: \begin{eqnarray}
228: \label{hnc}
229: g_{ij}(r)&=&exp[-\beta_{ij}\phi_{ij}(r)
230: +h_{ij}(r)-c_{ij}(r) + B_{ij}(r)]\\
231: h_{ij}(r) &=& c_{ij}(r)+
232: \Sigma_s n_s\int d{\bf r}'h_{i,s}
233: (|{\bf r}-{\bf r}'|)c_{s,j}({\bf r}')
234: \label{oz}
235: \end{eqnarray}
236: Here $\phi_{ij}(r)$ is the pair potential between the
237: species $i,j$. For e-e (or $D^+-D^+$) this is
238: just the Coulomb potential $V_{cou}(r)$.
239: If the electron spins are parallel, the Pauli
240: principle prevents occupation of the same spatial orbital.
241: As before,\cite{prl}
242: we introduce a
243: ``Pauli exclusion potential'', ${\cal P}(r)$.
244: Thus $\phi_{ij}(r)$ becomes ${\cal P}(r)\delta_{ij}+V_{cou}(r)$,
245: when $i,j$ denote electrons.
246: The function $h(r) = g(r)-1$ is related to the
247: structure factor $S(k)$ by a Fourier transform.
248: The $c(r)$ is the ``direct correlation function (DCF)''
249: of the OZ equations.
250: The $B_{ij}(r)$ is
251: the ``bridge'' term due to certain cluster interactions.
252: If this is neglected,
253: Eqs.~\ref{hnc}-~\ref{oz} form a closed set defining the
254: HNC approximation. (In effect, the classical K-S equation
255: becomes the HNC equation if the correlation potential
256: is evaluated as a sum of hyper-netted-chain diagrams.)
257: The HNC is sufficient for the uniform 3DEG for a
258: range of $r_s$, up to $r_s=50$, as studied previously.\cite{prb00}
259: We neglect the bridge corrections in
260: this study of deuterium.
261:
262: The ${\cal P}(r)$ is defined as in ref.\onlinecite{prl} from the
263: zeroth-order PDFs of the parallel-spin electrons. Thus:
264: \begin{equation}
265: \beta{\cal{P}}(r)=h_{11}^0(r)-c_{11}^0(r)-ln[g_{11}^0(r)]
266: \end{equation}
267: where, e.g., $c^0_{11}(r)$ is the spin-$\|$ DCF
268: of the O-Z equation.
269: Only the
270: product $\beta{\cal{P}}(r)$ is needed.
271:
272: The Coulomb potential $V_{cou}(r)$
273: for two point-charge electrons is $1/r$.
274: However, an electron at
275: the temperature $T$ is
276: localized to within a thermal de Broglie length (dBL). Thus,
277: for the 3DEG we used a ``diffraction
278: corrected'' form:\cite{minoo},
279: %\begin{eqnarray}
280: \begin{equation}
281: \label{potd}
282: V^{ee}_{cou}(r)=(1/r)[1-e^{-rk_{ee}}]
283: \end{equation}
284: where
285: $ k_{ee}=(2\pi m^*T_{ee})^{1/2}$ as shown by Minoo et al.\cite{minoo}
286: Here $m^*$=1/2 is the reduced mass of the electron {\it pair}.
287: In Minoo et al, the physical temperature $T$ was
288: used and is
289: valid only at high $T$.
290: The use of $T_{ee}=(T^2+T_{eq}^2)^{1/2}$
291: instead of $T$ validates it down to $T=0$ as well.\cite{prl}
292:
293: Since the $D^+$ are classical particles,
294: the $D^+-D^+$ interaction is the
295: coulomb interaction $\phi_{33}(r)=1/r$.
296: The $e-D^+$ interaction is more tricky.
297: The
298: e-e interaction, eqn.~\ref{potd}, is based on the
299: quantum mechanical scattering of two electrons.
300: We determine the $e-D^+$ interaction
301: $\phi_{eD}(r)$, from the density profile $n_{De}(r)$
302: given by the K-S equation for
303: electrons interacting with {\it a single deuteron} at the origin.
304: We have discussed this in the
305: context of the ``neutral pseudo-atom'' DFT model (NPA-DFT) for solving
306: the K-S equations.\cite{ilciacco,perrot93}
307: This gives
308: the deuteron-electron PDF, i.e.,
309: $g_{De}(r)$ = $n_{De}(r)/\bar{n}$. This
310: $n_{De}(r)$ includes {\it both bound-state and continuum-state}
311: densities.
312: Applying the HNC
313: and the OZ equation to this system containing a {\it single}
314: deuteron, we have,
315: \begin{eqnarray}
316: \label{potDe}
317: -\beta_{De}\phi_{De}(r)&=& log[g_{De}(r)]-h_{De}(r)+c_{De}(r)\\
318: h_{De}(r)&=&c_{De}(r)+ \bar{n}\int d\vec{r'}c_{De}(\vec{r}-\vec{r'})h _{ee}(r')
319: \end{eqnarray}
320: The deuteron-deuteron DCF does not appear
321: as there is only a single $D^+$. Hence, knowing the $g_{De}(r)$
322: from the solution of the
323: Kohn-Sham equation for the single deuteron problem, we can obtain $c_{De}(r)$ in
324: terms of $h_{De}(r)$. Hence the e-D potential can be extracted.
325: This determines the product,
326: $\beta_{De}\phi_{De}(r)$, while $\beta_{De}$, and $\phi_{De}$ are not needed
327: individually.
328: However, on solving the K-S equation for the regime of interest,
329: {\it no atomic bound
330: states} are found; the effective ionic charge $Z-n_b$,
331: where $n_b$ is the number of bound electrons
332: per nucleus, remains unity.
333: This does not contradict
334: {\it transient} bound-states found in PIMC.\cite{lqdmets}
335: Hence, at least in this regime of $\bar{n}$ and $T$,
336: Kohn-Sham NPA-DFT is not needed;
337: We set:
338: \begin{eqnarray}
339: \phi_{De}(r)&=&-(1/r)[1-e^{-rk_{De}}]\\
340: k_{De}&=&(2\pi m_{e}T_{ee})^{1/2}\\
341: 1/\beta_{De}&=&(T_{DD}+T_{ee})/2
342: \end{eqnarray}
343: The first equation is just the $V_{cou}$ with the $r=0$
344: value set to an
345: inverse dBL for the D-e pair, $k_{De}$, as
346: in the e-e interaction.
347: The dBL $1/k_{De}$
348: contains only the electron contribution
349: since the D contribution is zero
350: for a classical
351: particle. Thus only the $T_{ee}$ appears in $k_{De}$
352: (the effective
353: mass of the D-e pair is $m_e$, since the deuteron mass $M_D >> m_e$).
354: The effective temperature $1/\beta_{De}$ for the D-e interaction is
355: the mean kinetic
356: energy of the pair.
357: However, the use of the K-S procedure for the
358: dimensionless potential $\beta_{De}\phi_{De}(r)$
359: becomes obligatory at lower densities ($r_s > 2$) and
360: temperatures ($T < 1.0$ eV) when
361: the bound-state population $n_b$ becomes nonzero.
362:
363: Using the above scheme, we solve the coupled set
364: of HNC equations to determine
365: the six PDFs of
366: the $e,D^+$ system.
367: The excess free-energy
368: $F_{exc}(r_s,T)$
369: is determined via a coupling-constant integration,
370: as in ref.~[\onlinecite{prb00}],
371: for a range of values of $r_s$
372: and $T$, and converted into the total free energy $F(r_s,T)$ by
373: adding on the
374: ideal electron and ion contributions $F_0^D,F_0^e$ (see Fig.1(a)).
375: The total pressure $P$ and the
376: total internal energy $E$
377: are obtained as usual by
378: $P=\partial F(r_s,T)/\partial V$, where $V$ is the volume, and
379: $E=\partial \beta F(r_s,T)/\partial \beta$, where $\beta=1/T$.
380: In the regime of interest, i.e, 1.8 $<r_s<$ 2.1, and
381: 0.8 eV $< T <$ 15,
382: we find that $F_{exc}(r_s,T)$ is
383: approximately linear, i.e., $F_{exc}(r_s,T)=M(T)r_s+C(T)$.
384: The $T$-dependence of $M(T)$, $C(T)$ is quite nonlinear.
385: Figure 1(b) shows that $M(T)$ changes character near
386: $T$ = 4 eV.
387:
388: Our $P$, $E$ results are compared with the PIMC
389: of Militzer and Ceperley (MC) in table I,
390: showing good agreement for $T > $ 2.75 eV. For lower $T$,
391: our pressures are smaller.
392:
393: The free-energy $F(r_s,T)$ is used to calculate
394: the deuterium Hugoniot for the
395: initial state, ($E_0,V_0,P_0$), with $T$ = 19.6 K and an
396: initial density $\rho_0$ = 0.171 g/cm$^3$.
397: The initial-state $E_0$ = -15.886 eV per atom, and $P_0=0$.
398: The CHNC result, and others are
399: shown in Fig.2.
400: The CHNC Hugoniot, similar to PIMC,
401: approaches SESAME
402: at high temperatures. A softening of the
403: Hugoniot around 2 Mbar, not seen in the PIMC curve is also noted.
404: This
405: appears near $\bar{n}$ and $T$
406: where the interacting electron chemical potential
407: $\mu_e(r_s,T)$ passes through zero.
408:
409: We turn to the HNC equations to consider
410: electrons and deuterons at two different temperatures,
411: $T_e$ and $T_D$. The
412: shock is launched from an Aluminum pusher; the ions are
413: initially much hotter than the electrons.
414: The
415: velocity measurements begin after about 3 ns
416: in the laser experiments, and after a longer time
417: in the Sandia work.
418: Landau-Spitzer theory would indicate that the
419: $D^+-e$ equilibriation
420: occurs well within nanosecond timescales.
421: However, the formation
422: of coupled modes in plasmas with $\Gamma > 1$
423: strongly quenches the ion-electron equilibriation process.\cite{elr}
424: Also, experimental evidence exists for this point of view.\cite{elrexp}
425: A compactly held screening-charge at each ion would
426: act like a neutral object which, while having a very hot
427: deuteron at the centre, would screen it
428: from the cooler outer electrons.
429: The effect could lower the
430: electron-ion relaxation by more than an order of magnitude.\cite{elr,elrexp}
431: In lieu of a systematic energy relaxation analysis,
432: here we
433: assume that the deuteron nuclei are about 5 eV hotter
434: than the nominal electron temperature ( i.e., $T_D = T_e+5$ eV),
435: except at the
436: lowest temperatures.
437: Using $T_e$, $T_D$ in the HNC equations as before,
438: we have calculated
439: a quasi-equilibrium $F_{exc}(r_s,T_e,T_D)$
440: and a shock Hugoniot (quasi-equilibrium concepts
441: are discussed
442: in ref.\onlinecite{elr}).
443: The resulting non-equilibrium
444: Hugoniot is given in Fig.2.
445:
446:
447: The higher $T_D$ makes
448: the deuteron-electron fluid
449: more compressible. This appears counter-intuitive if one
450: considers only the $D^+$ contribution.
451: The quasi free energy $F(r_s,T_e,T_D)$ consists of
452: $F_e$, $F_D$, and $F_{De}$. On setting $T_D > T_e$, the
453: $F_D$ term taken alone {\it reduces} the compressibility, but the total
454: compressibility
455: is increased by the major role of the pair-term $F_{De}$.
456: As seen in Fig.1~(a), the
457: fluid is in a regime close to the
458: $\mu_e= 0 $ transition.
459: Thus a higher $T_D$ increases $T_{De}$ and
460: reduces the electron degeneracy even more,
461: making it more compressible. This also reduces the screening
462: and makes the electrons interact more strongly with the nuclei.
463: When this effect is strong
464: enough to offset the reduction of the compressibility from the ideal
465: gas term of the hotter deuterons, a softening of the Hugoniot
466: could result. In Fig.2 we show a Hugoniot labeled
467: NEQ0 where the ideal term $P_0$ was computed just
468: as in the equilibrium Hugoniot, while in NEQ the full effect was included.
469: Thus we see that except for the lowest temperatures, the contribution of
470: the deuteron-electron pairs dominate.
471:
472:
473: Our explanation of the observed
474: laser-shock Hugoniot is incomplete until
475: detailed modeling of the passage to equilibrium
476: is achieved. Other factors like the planarity, constancy and
477: duration of the shockwave are also relevant.
478: However, the present
479: discussion strongly justifies the need
480: to include equilibriation effects as well.
481: The CHNC approach presented here is numerically and
482: computationally very simple and should be a handy tool
483: in such studies. In fact all the calculations presented
484: here were carried out using very modest computational
485: facilities.\cite{sgi}
486: Our computer codes may be remotely
487: accessed by interested researchers by visiting our website.\cite{web}
488:
489:
490: In conclusion, we present a parameter-free calculation
491: of the EOS of deuterium in the regime of densities and
492: temperatures addressed by recent laser and magnetic
493: shock experiments. Kohn-Sham calculations show the absence of
494: atomic bound states in this regime.
495: Hence a simple
496: classical mapping of the electron quantum fluid was used.
497: The calculated Hugoniot
498: is in good agreement with other first-principles calculations.
499: Calculations for the case with the ions hotter
500: than the electrons by about 5 eV are also presented. They
501: suggest the possibility
502: that the anomalous Hugoniots obtained
503: from laser experiments could arise from the slowness of the
504: system to equilibriate within the experimental time scales.
505:
506: %\newpage %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
507: %\twocolumn %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
508: %\narrowtext %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
509: %
510: \begin{references}
511: \bibitem[\dag]{byline1} electronic mail address: chandre@cm1.phy.nrc.ca
512: \bibitem[\ddag]{byline2}NRC visiting scientist.
513: %
514: \bibitem{dasilva}
515: L. B. Da Silva {\it et al}., Phys. Rev. Lett. {\bf 78}, 483 (1997).
516: %
517: \bibitem{collins}
518: G. W. Collins {\it et al}., Science. {\bf 281}, 1178 (1998).
519: \bibitem{mosto}
520: A. N. Mostovych {\it et al}., Phys. Rev. Lett. {\bf 85}, 3870 (2000).
521: \bibitem{sesame}
522: G. I. Kerley, {\it Molecular Based Study of Fluids } (ACS, Washington,
523: D. C., 1983), p 107.
524: %
525: \bibitem{hug}
526: Hugoniot is the locus of states ($P,\rho,T$) generated by shock compression
527: from a given initial state.
528: %
529: \bibitem{ross}
530: M. Ross, Phys. Rev. B {\bf 58}, 669 (1998); D. Beule {\it et al}.,
531: Phys. Rev. E, {\bf 63}, 60202(R), (2001); D. Saumon et al,
532: High Pressure Research, {\bf 16}, 331 (2000)
533: \bibitem{dftcalcs}
534: G. Galli et al, Phys. Rev. B {\bf 61}, 909 (2000)
535: \bibitem{lenox}
536: T. J. Lenosky {\it et al}., Phys. Reb. B {\bf 61}, 1 (2000)
537: \bibitem{pimc}
538: B. Militzer {\it et al}., Phys. Rev. Lett. {\bf 85}, 1890 (2000).
539: \bibitem{knud}
540: M. D. Knudson {\it et al}., Phys. Rev. Lett. {\bf 87}, 225501 (2001).
541: \bibitem{dbleshock}
542: B. Militzer {\it et al}., LANL preprint archive, cond-mat/0108076v1
543: \bibitem{hyd0}
544: M. W. C. Dharma-wardana and F. Perrot, Phys. Rev. A {\bf 26}, 2096 (1982)
545: \bibitem{ilciacco}
546: M. W. C. Dharma-wardana and Fran\c{c}ois Perrot, in {\it Density Functional
547: Theory}, Eds. E. K. U. Gross and R. M. Dreizler (Plenum, New York, 1995), p 625
548: %
549: \bibitem{hncref}
550: J. M. J. van Leeuwen, J. Gr\"oneveld, J. de Boer, Physica {\bf 25}, 792 (1959)
551: %
552: \bibitem{kwon}
553: I. Kwon {\it et al}., Phys. Rev. E {\bf 54}, 2844 (1996)
554: \bibitem{prl}
555: M. W. C. Dharma-wardana and F. Perrot, Phys. Rev. Lett. {\bf 84}, 959 (2000)
556: %
557: \bibitem{prb00}
558: Fran\c{c}ois Perrot and M. W. C. Dharma-wardana, Phys. Rev. B {\bf 62},
559: 16536 (2000)
560: \bibitem{2D}
561: Fran\c{c}ois Perrot and M. W. C. Dharma-wardana, Phys. Rev. Lett. {\bf 87},
562: 206404 (2001)
563: %
564: %
565: \bibitem{minoo}
566: H. Minoo et al., Phys. Rev. A {\bf 23}, 924 (1981)
567: %
568: \bibitem{perrot93}
569: Fran\c{c}ois Perrot, Rev. A {\bf 47},570 (1993)
570: %
571: \bibitem{lqdmets}
572: Similar transient bonding effects
573: are seen in, e.g., liquid-Si,Ge simulations
574: using Car-Parrinello methods, while parallel calculations by us
575: produce static PDFs
576: in excellent agreement with experiment. See:
577: M.W.C. Dharma-wardana and F. Perrot, Phys. Res. Lett. {\bf 65}, 76 (1990)
578: %
579: \bibitem{elr}
580: M. W. C. Dharma-wardana and Fran\c{c}ois Perrot, Phys. Rev. E {\bf 58},
581: 3705 (1998), erratum: {\bf 63},66901 (2001);
582: M. W. C. Dharma-wardana, Phys. Rev. E {\bf 64}, 35401(R) (2001)
583: \bibitem{elrexp}
584: D. Riley {\it et al}., Phys. Rev. Lett. {\bf 84}, 1704 (2000)
585: \bibitem{sgi}
586: A Silicon Graphics SGI02/MIPS R5000/180 MHz
587: workstation running IRIX 6.5 was used.
588: \bibitem{web}
589: http://nrcphy1.phy.nrc.ca/ims/qp/chandre/chnc/
590: \end{references}
591:
592:
593:
594: \newpage
595:
596: \begin{table}
597: \caption{
598: The total pressure $P$ (Mbar) and total energy $E$ (eV),
599: calculated using the Classical-map HNC (CHNC) approach, and the
600: path-integral Monte Carlo (PIMC)
601: approach of Militzer and Ceperley (MC) at $r_s$=2.0, i.e, at
602: a deuterium density of 0.6691 g/cm$^3$.
603: }
604: \begin{tabular}{ccccccccc}
605: $T(K)$&$F_{exc}$&$P$(CHNC)&$P$(MC)&$E$(CHNC)&$E$(MC)\\
606: \hline\\
607: 500 000& -5.35310 & 26.278 & 25.980 & 113.30 & 113.20 \\
608: 250 000& -2.14960 & 12.244 & 12.120 & 47.57 & 45.70 \\
609: 125 000& -0.99712 & 5.374 & 5.290 & 13.60 & 11.50 \\
610: 65 000& -0.64405 & 2.143 & 2.280 & -3.21 & -3.80 \\
611: 31 250& -0.57058 & 0.754 & 1.110 & -10.74 & -9.90 \\
612: 15 625& -0.57119 & 0.213 & 0.540 & -13.97 & -12.90 \\
613: 10 000& -0.57890 & 0.058 & 0.470 & -14.91 & -13.60 \\
614: \end{tabular}
615: \label{gamma0}
616: \end{table}
617: %
618: %
619: \begin{figure}
620: \caption
621: {(a) The temperature dependence of the excess free energy $F_{exc}(r_s,T)$,
622: $F_0^e((r_s,T)$, and the interacting chemical potential $\mu_e(r_s,T)$,
623: at $r_s$=1.86. Note the passage of $\mu$
624: through zero.
625: (b) The excess free energy in the range of the the shock experiments
626: can be fitted to the linear form $F_{exc}(r_s,T)=M(T)r_s+C(T)$;
627: the slope $M(T)$ and the intercept $C(T)$ are shown as a function of
628: $T$. Nore the change of character in $M(T)$ when
629: $\mu_e$ passes through zero.
630: }
631: \label{figFexc}
632: \end{figure}
633: %
634:
635: %
636:
637: \begin{figure}
638: \caption
639: { Comparison of the CHNC Hugoniot with experiment and other
640: theories. Two non-equilibrium Huoniots are also shown (see text).
641: Experiments 1, 2 and 3 refer to Da Silva {\it et al.},
642: Collins {\it et al.}, and Knudson {\it et al.}, respectively.
643: }
644: \label{fighu}
645: \end{figure}
646:
647: %
648:
649:
650: \end{document}
651:
652:
653:
654:
655:
656:
657: