ELSEVIER 13 September 1996 Chemical Physics Letters 259 (1996) 679-682 CHEMICAL PHYSICS LETTERS The total atomization energy and heat of formation of HCN(g) Jan M.L. Martin l The Institute of Chemistry, The Hebrew University of Jerusalem, Givat-Ram, Jerusalem 91904, Israel Received 27 June 1996; in final form 31 July 1996 Abstract An exhaustive ab initio convergence study, using coupled cluster methods and basis sets of up to [7s6p5d4f3g2h/5s4p3d- 2flg] quality, of the total atomization energy of HCN(g), both directly and via the isogyric reaction 2 HCN ¢, C2H2 + N2, leads to a best estimate of ~ D0=303.08+0.25 kcal/mol, in between thermochemical and spectroscopic experimental values. The best estimate for A/-/}'f = 31.04 4- 0.25 kcal/mol, or 129.9-1-1 kJ/mol. 1. Introduction Hydrogen cyanide (prussic acid, HCN) was pro- duced by Scheele in dilute solution in 1782, and by Gay-Lussac as the pure compound in 1815. HCN- based poisons may have been known and produced since Antiquity [ 1 ] : less morbid modern uses include the production of plastics (poly-methylmethacrylate) and antioxidants (EDTA), as well as the extraction of gold and silver using NaCN. In 1979, the world pro- duction was estimated to be 0.5-0.6 million ton per year. One would expect the heat of formation and total atomization energy of such a common molecule to be precisely known. Yet the JANAF tables [2] list a heat of formation of 32.4:t:2 kcal/mol, which largely derives from two experiments, one by Berthelot [ 3] from 1881, and one by Thomsen [4] from 1886. Ex- periments from 1951 by Horiuchi, Yano, and Kanai (HYK) [5] suggest a value of AH 298"15 = 31.354-0.6 IOn sabbatical from: Limburgs Universitair Centrum, Depart- ment SBG, UniversitaireCampus, 3590 Diepenbeek,Belgium and University of Antwerp (UIA), Institute for Materials Science, De- partmentof Chemistry,Universiteitsplein 1, 2610 Wilrijk, Belgium kcal/mol, corresponding (from the thermodynamic functions in Ref. [2] ) to A/-/~f = 31.44+0.6kcal/mol, or ~ Do = 302.8 4- 0.6 kcal/mol. Recently, Morley et al. [6], in their study of the dissociation dynamics of HCN, obtained Do (H-CN) = 43740 4-150 cm -1 . An exhaustive ab initio calibration study by Pradhan et al. [7] confirmed definitively that the revised value [8] of D0(CN) = 7.738-1-0.02 eV is the correct one. Together, they lead to a total atom- ization energy ~ Do = 303.5 4- 0.6 kcal/mol, where the error bar is the RMS (root mean square) of the experimental uncertainties for the two steps. The author has shown elsewhere [9] that ab ini- tio coupled cluster [ 10] calculations in basis sets of spdfgh quality, together with an extrapolation formula based on the known [ 11 ] (l + 1/2) -4 dependence of the electron-electron energy (l being the highest angu- lar momentum present) and inclusion of core correla- tion through an additivity approximation [ 12], lead to computed ~ De values correct to 0.3 kcal/mol, on av- erage. If a correction is introduced for triple bonds, the mean absolute error drops to 0.2 kcal/mol. It would appear that computational techniques of such accuracy are sufficient to resolve the issue of the total atomiza- 0009-2614/96/$12.00 Copyright @ 1996 Published by Elsevier Science B.V. All rights reserved. PII S0009-2614(96)00899-8