1: % Group addresses by affiliation; use superscriptaddress for long

2: % author lists, or if there are many overlapping affiliations.

3: % For Pdhys. Rev. appearance, change preprint to twocolumn.

4: % Choose pra, prb, prc, prd, pre, prl, prstab, or rmp for journal

5: %  Add 'draft' option to mark overfull boxes with black boxes

6: %  Add 'showpacs' option to make PACS codes appear

7: %  Add 'showkeys' option to make keywords appear

8: %\documentclass[aps,pre,preprint,superscriptaddress,showpacs,floatfix]{revtex4}

9: \documentclass[aps,prl,twocolumn,superscriptaddress,floatfix]{revtex4}

10: %\documentclass[aps,prl,preprint,superscriptaddress]{revtex4}

11: %\documentclass[aps,prl,twocolumn,groupedaddress]{revtex4}

12: \usepackage{latexsym,hyperref,color,amssymb,graphicx}

13: % You should use BibTeX and apsrev.bst for references

14: % Choosing a journal automatically selects the correct APS

15: % BibTeX style file (bst file), so only uncomment the line

16: % below if necessary.

17: %\bibliographystyle{h-physrev}

18: \newcommand{\beq}{\begin{equation}}

19: \newcommand{\eeq}{\end{equation}}

20: \newcommand{\bea}{\begin{eqnarray}}

21: \newcommand{\eea}{\end{eqnarray}}

22: \newcommand{\bean}{\begin{eqnarray*}}

23: \newcommand{\eean}{\end{eqnarray*}}

24: \newcommand{\bei}{\begin{itemize}}

25: \newcommand{\eei}{\end{itemize}}

26: \newcommand{\ben}{\begin{enumeration}}

27: \newcommand{\een}{\end{enumeration}}

28: \newcommand{\nn}{\nonumber}

29: \newcommand{\thalf}{\textstyle{ {1 \over 2} } }

30: \newcommand{\shalf}{\scriptstyle{\frac{1}{2}}}

31: \newcommand{\sshalf}{\scriptscriptstyle{\frac{1}{2}}}

32: \newcommand{\tcol}[2]{\textcolor{#1}{#2}}

33: \newcommand{\tcr}[1]{\textcolor{red}{#1}}

34: \newcommand{\tcb}[1]{\textcolor{blue}{#1}}

35: \newcommand{\tcg}[1]{\textcolor{green}{#1}}

36: \newcommand{\tcy}[1]{\textcolor{yellow}{#1}}

37: \newcommand{\tcc}[1]{\textcolor{cyan}{#1}}

38: \newcommand{\tcm}[1]{\textcolor{magenta}{#1}}

39: \definecolor{darkorange}{rgb}{.6,.2,.0}

40: \newcommand{\tcdo}[1]{\textcolor{darkorange}{#1}}

41: \definecolor{darkgreen}{rgb}{0.0,0.7,0.0}

42: \newcommand{\tcdg}[1]{\textcolor{darkgreen}{#1}}

43: \newcommand{\tcp}[5]{\definecolor{#1}{rgb}{#2,#3,#4}\textcolor{#1}{#5}}

44: \begin{document}

45: 

46: \title{Rydberg-London Potential for Diatomic Molecules

47: and Unbonded Atom Pairs}

48: 

49: \author{Kevin Cahill}

50: \email{cahill@unm.edu}

51: \affiliation{Department of Physics and Astronomy, 

52: University of New Mexico, Albuquerque, NM 87131}

53: \affiliation{Center for Molecular Modeling, 

54: Center for Information Technology, 

55: National Institutes of Health, 

56: Bethesda, Maryland 20892-5624}

57: \author{V.~Adrian Parsegian}

58: \email{aparsegi@helix.nih.gov}

59: \affiliation{National Institute of Child Health

60: and Human Development, National Institutes of Health, 

61: Bethesda, Maryland 20892-0924}

62: 

63: \date{\today}

64: 

65: \begin{abstract}

66: We propose and test a pair potential that

67: is accurate at all relevant distances

68: and simple enough 

69: for use in large-scale computer simulations.  

70: A combination of the Rydberg potential from spectroscopy 

71: and the London inverse-sixth-power energy, 

72: the proposed form fits 

73: spectroscopically determined potentials

74: better than the Morse, Varnshi,

75: and Hulburt-Hirschfelder potentials

76: and much better than the 

77: Lennard-Jones and harmonic potentials.

78: At long distances, it 

79: goes smoothly to the correct London force 

80: appropriate for gases and preserves 

81: van der Waals's ``continuity of the gas and liquid states," 

82: which is routinely violated by coefficients assigned 

83: to the Lennard-Jones 6-12 form.

84: \end{abstract}

85: 

86: %\pacs{82.35.Pq,34.20.Cf,77.84.Jd}

87: 

88: \maketitle

89: There are at least three classes of 

90: interatomic potentials.

91: Commercial codes

92: use the Lennard-Jones 

93: \beq

94: V_{LJ}(r) = |V(r_0)| \, \left[ \left( \frac{r_0}{r} \right)^{12}

95: - 2 \, \left( \frac{r_0}{r} \right)^6 \right]

96: \label{LJ}

97: \eeq

98: and harmonic 

99: \beq

100: V_H(r) = V(r_0) 

101: + \frac{(r - r_0)^2}{2} \, 

102: \frac{d^2 V(r_0)}{dr^2} 

103: \label{Ha}

104: \eeq

105: forms, which

106: are accurate

107: near the minimum at \(r=r_0\)\@.

108: But this first class of potentials may be too simple

109: for complex materials

110: away from from equilibrium.

111: \par

112: The Morse~\cite{Morse1929}

113: \beq

114: V_M(r) = |V(r_0)| \, \left[

115: \left( 1 - e^{-\kappa \, x} \right)^2

116: - 1 \right]

117: \label{Morse}

118: \eeq

119: (\(x = r - r_0\), \( \kappa = \sqrt{k_e/(2|V(r_0)|)} \)),

120: Varnshi~\cite{Varnshi1957}

121: \beq

122: V_V(r) = |V(r_0)| \, \left[

123: \left( 1 - \frac{r_0}{r} \, e^{- \beta \, (r^2 - r_0^2)} 

124: \right)^2

125: - 1 \right],

126: \label{Varnshi}

127: \eeq

128: (\( \beta = (\kappa r_0^2 -1)/(2r_0^2) \)), 

129: and Hulburt-Hirschfelder~\cite{Hulbert1961}

130: \beq

131: V_{HH}(r) = V_M(r) + |V(r_0)| \, 

132: c'\kappa^3x^3\,e^{-2\kappa x}(1+\kappa b'x)

133: \label{HH}

134: \eeq

135: (for \(b'\), \(c'\), see ref.\cite{Vanderslice1961})

136: potentials represent

137: a second class of potentials~\cite{Vanderslice1961}

138: accurate over a wider

139: range of distances.

140: \par

141: Quantum 

142: chemists~\cite{Scoles1984,Aziz1993,Meath2001,Tang2003,Varandas2004}

143: have derived a third class 

144: that reproduce spectroscopic

145: and thermodynamic data with impressive fidelity.

146: But the potentials of this class

147: involve many parameters and

148: may be too cumbersome for use in large-scale simulations.

149: \par

150: We propose and test a form

151: \beq

152: V(r) = a e^{-b \, r}\,( 1 - c \, r ) 

153: - \frac{d}{r^6 + e \, r^{-6}}

154: \label{phenpot}

155: \eeq

156: that is nearly as accurate as the class-3 potentials 

157: but simpler than many class-2 potentials.

158: It is a combination of the Rydberg formula 

159: used in spectroscopy 

160: and the London formula for pairs of atoms.

161: In Eq.(\ref{phenpot}), the terms involving  \(a\), \(b\), and \(c\) 

162: were proposed by Rydberg~\cite{Rydberg1931}

163: to incorporate spectroscopic data,

164: but were largely ignored until 

165: recently~\cite{Ferrante1991,Murrell1984}.

166: The constant \(d = C_6\)

167: is the coefficient of the London tail.

168: The new term \( e \, r^{-6} \) cures

169: the London singularity.

170: As \(r \to 0\),

171: \( V(r) \to a\), finite;

172: as \(r \to \infty\), \(V(r)\) 

173: approaches the London term,

174: \(V(r) \to - d/r^6 = -C_6/r^6\)\@.

175: In a perturbative analysis~\cite{Cahill2003},

176: the \(a, b, c\) terms 

177: arise in first order, and the \(d\) term

178: in second order. 

179: \par

180: \begin{figure}

181: \centering

182: \input RKRH2

183: \caption{The hybrid form \(V\)

184: (with \(a = 53.8\) eV,

185: \(b = 2.99\) \AA\(^{-1}\),

186: \(c = 2.453\) \AA\(^{-1}\),

187: \(d = C_6 = 3.884\) eV\,\AA\(^6\),

188: and \(e = 47.6\) \AA\(^{12}\))

189: (solid, red) 

190: fits the RKR spectral points for 

191: the ground state of

192: molecular hydrogen (pluses, blue)

193: and gives the correct London tail for \(r > 3\) \AA\@.

194: The Lennard-Jones \(V_{LJ}\) (dashes, green) and harmonic

195: \(V_H\) (dots, magenta) forms fit

196: only near the minimum.}

197: \label{rkr}

198: \end{figure}

199: \begin{figure}

200: \centering

201: \input H2REV

202: \caption{The hybrid form \(V\)

203: (Eq.(\ref{phenpot}) as in Fig.~\ref{rkr},

204: solid, red) 

205: fits the RKR spectral points for 

206: the ground state of

207: molecular hydrogen (pluses, blue).

208: The Morse \(V_{M}\) (Eq.(\ref{Morse}), dashes, green),

209: Varnshi \(V_V\) (Eq.(\ref{Varnshi}), dots, magenta), and 

210: Hulburt-Hirschfelder

211: \(V_{HH}\) (Eq.(\ref{HH}), dot-dash, cyan) 

212: are too low for

213: \(1.5 < r < 3\) \AA\@.}

214: \label{h2rev}

215: \end{figure}

216: \begin{figure}

217: \centering

218: \input RKRN2

219: \caption{For the ground state of molecular nitrogen,

220: the proposed potential \(V\) of Eq.(\ref{phenpot}) 

221: (with \(a = 3.25\) keV,

222: \(b = 4.30\) \AA\(^{-1}\),

223: \(c = 1.1906\) \AA\(^{-1}\),

224: \(d = 14.1\) eV\,\AA\(^6\), and

225: \(e = 27.1\) \AA\(^{12}\))

226: (solid, red) 

227: fits the RKR spectral points (pluses, blue)

228: and the FO points (x's, cyan)\@.

229: \(V_{LJ}\) and \(V_H\) as in Fig.~\ref{rkr}\@.}

230: \label{rkrn2}

231: \end{figure}

232: \begin{figure}

233: \centering

234: \input N2REV

235: \caption{For the ground state of molecular nitrogen,

236: the hybrid form \(V\) (Eq.(\ref{phenpot}),

237: as in Fig.~\ref{rkrn2}, solid, red) 

238: fits the RKR spectral points (pluses, blue)

239: and the FO points (x's, black)\@.

240: \(V_{M}\), \(V_V\), \&

241: \(V_{HH}\) as in Fig.~\ref{h2rev}.}

242: \label{n2rev}

243: \end{figure}

244: \begin{figure}

245: \centering

246: \input RKRO2

247: \caption{For the ground state of molecular oxygen,

248: the potential \(V\) of Eq.(\ref{phenpot}) 

249: (with \(a = 3.61\) keV,

250: \(b = 4.48\) \AA\(^{-1}\),

251: \(c = 1.05\) \AA\(^{-1}\),

252: \(d = 16.08\) eV\,\AA\(^6\), and

253: \(e = 58.4\) \AA\(^{12}\))

254: (solid, red) fits

255: the RKR points (pluses, blue)\@.

256: \(V_{LJ}\) and \(V_H\) as in Figs.~\ref{rkr} 

257: \& \ref{rkrn2}\@.}

258: \label{rkro2}

259: \end{figure}

260: \begin{figure}

261: \centering

262: \input O2REV

263: \caption{For the ground state of molecular oxygen,

264: the hybrid form \(V\) (Eq.(\ref{phenpot}),

265: as in Fig.~\ref{rkro2}, solid, red) 

266: fits the RKR spectral points (pluses, blue)\@.

267: \(V_{M}\), \(V_V\), \&

268: \(V_{HH}\) as in Fig.~\ref{h2rev}.}

269: \label{o2rev}

270: \end{figure}

271: \begin{figure}

272: \centering

273: \input AR2374

274: \caption{The potential \(V\) (Eq.(\ref{phenpot}) 

275: for the Ar-Ar ground state  

276: with \(a = 1720\) eV,

277: \(b = 2.6920\) \AA\(^{-1}\),

278: \(c = 0.2631\) \AA\(^{-1}\),

279: \(d = 37.943\) eV\,\AA\(^6\), and

280: \(e = 177588\) \AA\(^{12}\))

281: (solid, red) fits

282: the Ar\(_2\) Aziz potential (pluses, blue)

283: with the correct London tail.

284: When matched at the minimum,

285: the Lennard-Jones form \(V_{LJ}\) 

286: (Eq.(\ref{LJ}) with \(r_0 = 3.757\) \AA\ 

287: and \(V(r_0) = -0.01234\) eV) (dashes, green) 

288: is too low for \(r > 4\) \AA.}

289: \label{ar237}

290: \end{figure}

291: \begin{figure}

292: \centering

293: \input AR24

294: \caption{Positive potential \(V\) (Eq.(\ref{phenpot}) 

295: as in Fig.~\ref{ar237}),

296: Ar\(_2\) Aziz potential (pluses, blue),

297: L-J form \(V_{LJ}\) (Eq.(\ref{LJ})

298: as in Fig.~\ref{ar237})\@.}

299: \label{ar2}

300: \end{figure}

301: \begin{figure}

302: \centering

303: \input KR2374

304: \caption{The potential \(V\) (Eq.(\ref{phenpot})

305: for the Kr-Kr ground state  

306: with \(a = 2499\) eV,

307: \(b = 2.5249\) \AA\(^{-1}\),

308: \(c = 0.2466\) \AA\(^{-1}\),

309: \(d = 78.214\) eV\,\AA\(^6\), and

310: \(e = 199064\) \AA\(^{12}\))

311: (solid, red) fits

312: the Kr\(_2\) Aziz points (pluses, blue)

313: with the correct London tail.

314: When matched at the minimum,

315: the Lennard-Jones form \(V_{LJ}\) 

316: (Eq.(\ref{LJ}) with \(r_0 = 4.008\) \AA\ 

317: and \(V(r_0) = -0.017338\) eV) (dashes, green) 

318: is too low for \(r > 4\) \AA.}

319: \label{kr237}

320: \end{figure}

321: \begin{figure}

322: \centering

323: \input KR24

324: \caption{Positive potential \(V\) (Eq.(\ref{phenpot}) 

325: as in Fig.~\ref{kr237}),

326: Kr\(_2\) Aziz points (pluses, blue),

327: L-J form \(V_{LJ}\) (Eq.(\ref{LJ})

328: as in Fig.~\ref{kr237})\@.}

329: \label{kr2}

330: \end{figure}

331: \par

332: To test whether the 

333: hybrid \(V(r)\)

334: can represent covalent bonds

335: far from equilibrium,

336: we used Gnuplot~\cite{Gnuplot}

337: to fit Eq.(\ref{phenpot}) 

338: to empirical potentials for % Remark: we had 'of,' not 'for.'

339: molecular H\(_2\), N\(_2\), and O\(_2\)

340: obtained from spectroscopic 

341: data~\cite{Weissman1963,Krupenie1977,Krupenie1972}

342: by the RKR (Rydberg~\cite{Rydberg1931}, Klein~\cite{Klein1932},

343: Rees~\cite{Rees1947}) method,

344: setting \(d\) equal to the London values.

345: Figure~\ref{rkr} shows

346: that the hybrid potential of Eq.(\ref{phenpot}) (solid, red)

347: goes through the RKR points for H\(_2\) (pluses, blue)

348: from 0.5 to 4 \AA\@.

349: Fitted to the minimum,

350: the harmonic potential (\ref{Ha}) 

351: (dashes, green)

352: and the Lennard-Jones potential (\ref{LJ})

353: (dashes, blue)

354: are accurate only

355: near the minimum at 0.74 \AA\@.

356: Figures~\ref{rkrn2} \& \ref{rkro2}

357: show that the hybrid potential \(V(r)\)

358: fits the N\(_2\) and O\(_2\) 

359: RKR points~\cite{Krupenie1977,Krupenie1972}

360: and the N\(_2\) first-order (FO)

361: points~\cite{Ermler1982}

362: between 1 and 1.8 \AA\@.

363: The Lennard-Jones and harmonic potentials

364: of Eqs.(\ref{LJ}--\ref{Ha}) fit only near the minima.

365: \par

366: How does the hybrid form compare

367: with the class-2 potentials

368: of Eqs.(\ref{Morse}--\ref{HH})?

369: Figures~\ref{h2rev}, \ref{n2rev}, \&

370: \ref{o2rev} show that for 

371: \( 1.4 < r < 2 \) \AA\ 

372: the hybrid \( V(r) \)

373: is closer than (\ref{Morse}--\ref{HH})

374: to the RKR points.

375: A useful estimate of how well a 

376: particular potential \( V_P(r) \)

377: fits \(N\) data points \( V_D(r_i) \)

378: is the dimensionless error 

379: \beq

380: \delta = \frac{1}{|V_D(r_0)|} \,

381: \left[ \frac{1}{N} \,

382: \sum_{i = 1}^N 

383: \left( V_P(r_i) - V_D(r_i) \right)^2 \,

384: \right]^{1/2}.

385: \label{delta}

386: \eeq

387: For \(H_2\),

388: \(N_2\), and \(O_2\), the average

389: error \(\delta\) was 59.7 for \(V_H\),

390: 49.3 for \(V_{LJ}\),

391: 0.037 for \(V_M\),

392: 0.031 for \(V_V\), 0.021 for \(V_{HH}\),

393: and 0.0044 for the hybrid \(V\)\@.

394: The hybrid form \(V\) is five times more

395: accurate than the

396: class-2 potentials (\ref{Morse}--\ref{HH})

397: and four orders of magnitude more accurate 

398: than the class-1 potentials (\ref{LJ}--\ref{Ha})\@.

399: \par

400: Can \(V(r)\) also represent 

401: weak noncovalent bonds?

402: Using Gnuplot, we fitted 

403: \(a\), \(b\), \(c\), and \(e\) in Eq.(\ref{phenpot})

404: to Aziz's accurate HFDID1 potential for 

405: Ar\(_2\)~\cite{Aziz1993} and HFD-B potential

406: for Kr\(_2\)~\cite{Aziz1986}

407: and set \(d\) equal

408: to their London-tail coefficients.

409: Figures~\ref{ar237}--\ref{kr2} 

410: show that for \(2 \le r \le 7\)\AA,

411: the hybrid form

412: (\ref{phenpot}) (solid, red) 

413: fits the Aziz potentials for both Ar\(_2\) and

414: for Kr\(_2\) (pluses, blue)\@.

415: The Lennard-Jones \(V_{LJ}\) curves

416: (\ref{LJ}) (dashes, green) 

417: matched at the minima

418: are too deep for \( r > 4.5 \)~\AA\ 

419: (Figs.~\ref{ar237} \& \ref{kr237})

420: and too hard for \( r < 3 \)~\AA\ 

421: (Figs.~\ref{ar2} \& \ref{kr2})\@.

422: %Being accurate at all relevant distances,

423: The potential \(V(r)\) of Eq.(\ref{phenpot}) 

424: represents weak noncovalent bonds

425: better than \(V_{LJ}\)

426: (and \(V_H\))\@.

427: \par

428: Does it matter that \(V_{LJ}(r)\)

429: fails to fit 

430: the Ar--Ar and Kr--Kr interactions?

431: To find out, we used \(V\) and  \(V_{LJ}\)

432: to compute the dimensionless second virial coefficient \(B_2/r_0^3\)

433: of Ar and Kr at room temperature (\(kT = 0.025 \) eV)\@.

434: Here \(r_0\) is the minimum of the potential, and

435: the second virial coefficient \(B_2\) is the integral

436: over all space

437: \beq

438: B_2(T) = - {\thalf} \, \int \! d^3r \, 

439: \left( e^{-\beta V(r)} -1 \right) 

440: \label{B_2}

441: \eeq

442: in which \(\beta = 1/(kT)\)\@.

443: The hybrid potential \(V\) 

444: fitted to the curves of Figs.~\ref{ar237} -- \ref{kr2}

445: gives \(B_2/r_0^3 = -0.499\) for Ar and 

446: \(-1.35\) for Kr,

447: which respectively differ from 

448: the experimental~\cite{Lide1994}

449: values of \(-0.552\) and \(-1.41\) by 9.6\% and 4.1\%\@.

450: The Lennard-Jones potential \(V_{LJ}\)

451: fitted to \(r_0\) and \(V(r_0)\)

452: gives \(B_2/r_0^3 = -0.899\) for Ar and \(-1.92\) for Kr

453: (errors of 63\% and 37\%);

454: it also wags a long-range tail with  

455: London coefficients \(2 \, |V(r_0)| \, r_0^6 \)

456: that are too large by 83\% for Ar and by 84\% for Kr\@.

457: If the parameters \(r_0\) and \(V(r_0)\)

458: in Eq.(\ref{LJ}) for \(V_{LJ}\)

459: are chosen to give the correct 

460: second virial coefficient \(B_2/r_0^3\)

461: for a range of temperatures~\cite{Hill1986},

462: then \(V(r_0)\) 

463: is too shallow

464: by 16 \% for Ar and 15\% for Kr,

465: and the London coefficients of the 

466: long-range tail are too large by 70\% for Ar and by 64\% for Kr\@.

467: With only two parameters,

468: Lennard-Jones fits are procrustean.

469: \par

470: What about additivity (\cite{Axilrod1943,Muto1943,Cole1988})?

471: When three or more atoms interact,

472: their potential energy is not

473: the sum of the three (or more) pair potentials.

474: Is the accuracy of the hybrid form important

475: in the liquid phase

476: where additivity is only approximate?

477: \par

478: To test whether the lack of complete additivity

479: in the liquid phase obscures the 

480: advantages of the hybrid form \(V\) over the

481: Lennard-Jones potential \(V_{LJ}\),

482: we used both to compute  

483: the heats of vaporization \(\Delta_{\mathrm{vap}}H\)

484: of Ar and Kr at their

485: boiling points at atmospheric pressure.

486: In our Monte Carlo simulations,

487: we imposed periodic boundary conditions 

488: to reduce finite-size effects.

489: Our Monte Carlo code is 

490: available at 

491: \href{http://bio.phys.unm.edu/latentHeat/}

492: {bio.phys.unm.edu/latentHeat}.

493: We used it

494: to compute the potential energy \(U\) per

495: atom in the liquid and gas phases.

496: The latent heat of vaporization 

497: \( \Delta_{\mathrm{vap}}H \)

498: is the difference between the potential

499: energies \(U_{\textrm{gas}}\) and \(U_{\textrm{liquid}}\)

500: plus the work done in expanding by \(\Delta V\)

501: against the pressure \(p\)

502: of the atmosphere,

503: \(\Delta_{\mathrm{vap}}H = U_{\textrm{gas}} \,

504: - \, U_{\textrm{liquid}} \, + \, p \, \Delta V \)\@.

505: \par

506: The hybrid form \(V\) 

507: fitted to the curves 

508: of Figs.~\ref{ar237} -- \ref{kr2} gave

509: \(\Delta_{\mathrm{vap}}H = 0.0694\) eV (per atom) for Ar

510: and 0.0982 eV for Kr,

511: which differ from 

512: the experimental values~\cite{Lide1994} of 0.0666

513: and 0.0941 eV by 4.2\% and 4.4\%\@.  % CRC

514: %and 0.0932 eV by 5.0\% and 6.0\%\@. % NIST

515: In equivalent Monte Carlo simulations,

516: the Lennard-Jones potential \(V_{LJ}\) fitted

517: to \(r_0\) and \(V(r_0)\) gave

518: and \(\Delta_{\mathrm{vap}}H = 0.0787\) eV for Ar 

519: and 0.111 eV for Kr

520: (errors of 18\% and 18\%)\@.

521: So the errors due to a lack of additivity

522: are of the order of 4\%,

523: while those due to the Lennard-Jones potential

524: are about 18\%\@.

525: Even in the liquid phase,

526: limited additivity is less of a problem

527: than the defects of the Lennard-Jones potential.

528: \par

529: For a wide class of atom pairs,

530: the hybrid form 

531: (\ref{phenpot})

532: can reliably represent the best

533: spectroscopically determined

534: potentials over all relevant distance scales. 

535: It also yields accurate second virial coefficients

536: and heats of vaporization.

537: Its simplicity recommends it as a teaching tool

538: and as a practical form for computation.

539: Given the differences between it

540: and the 

541: Lennard-Jones, harmonic, Morse,

542: Varnshi, and Hulburt-Hirschfelder potentials, 

543: it would be worthwhile

544: to examine the consequences of

545: these differences in Monte Carlo searches

546: for low-energy states of biomolecules

547: and in numerical simulations

548: of phase transitions and reactions

549: far from equilibrium.

550: \begin{acknowledgments}

551: Thanks to S.\ Valone for conversations 

552: and for RKR data;

553: to S.\ Atlas, C.\ Beckel, 

554: B.\ Brooks, J.\ Cohen, K.\ Dill, D.\ Harries, G.\ Herling,

555: M.\ Hodoscek, R.\ Pastor, 

556: R.\ Podgornik, W.\ Saslow, and C.\ Schwieters for advice.

557: P.~J.\ Steinbach kindly hosted KC\ at NIH,

558: where we used the Biowulf computers.

559: \end{acknowledgments}

560: 

561: \bibliography{chem,cs,physics,vdw}

562: 

563: \end{document}

564: 

565: