The new release 5 of HyperChem, an outstanding computational chemistry program, offers many significant enhancements in visualization tools such as plotting electrostatic potentials on isoelectron density surfaces. By including Langevin dynamics and Monte Carlo simulations, the new release of HyperChem slightly extends the already significant array of quantum mechanical (ab initio and semi-empirical but not density functional) and classical mechanical (molecular mechanics and molecular dynamics) methods that were already featured in version 4.5. In terms of the number of such calculational methods, no other computational chemistry program (e.g., Alchemy 2000 [Tripos, Inc.], MacSpartan Plus/PC Spartan [Wavefunction, Inc.], CAChe [Oxford Molecular Group], and Chem3D [CambridgeSoft Corp.]) running on either a PC or a Macintosh rivals Hyperchem. Moreover, compared to programs such as MacSpartan Plus, HyperChem offers many more options within each calculational method. Release 5 of the program requires a PC running Windows 95 or Windows NT with at least 8 Mb of RAM (16 Mb recommended) and a CD-ROM drive.

Although the new Chemist's Development Kit (CDK) clearly makes HyperChem quite powerful and unique in terms of expandability, it would require more time than is typically available to an undergraduate student for him/her to implement it in a meaningful way for course work. A faculty member, however, might decide to use the CDK to develop a customized menu approach to complex calculations such as docking experiments for use in teaching biochemistry. Driving HyperChem with the spreadsheet program Microsoft Excel is more straightforward and may be used, for example, to perform batch calculations on a series of molecules.

Some relatively minor changes in version 5 are quite useful. The reporting of molecular and orbital symmetry is a welcome new feature. The menus have become more interactive; and, for example, the ability to set the quantum print level for a log file from the menu, rather than having to modify the chem.ini file, is appreciated. Below we suggest a few changes that should further enhance the usability of the program.

1. HyperChem should revert to a set of default settings (e.g., "show multiple bonds") each time the program is launched. This feature would eliminate confusion among multiple users of the program, especially when HyperChem is used in different courses.

2. The log file, containing extended information about a particular calculation, should be accessible from within HyperChem. Currently, an external word processor or a text editor must be used to view this information.

3. A display of the CPU time required for a calculation would facilitate comparison between various computational methods for the same task.

4. Although the new ray-tracing program produces excellent graphics, the saved files appear to retain only the framework, precluding the export of HyperChem-generated publication-quality 3-D graphics onto the WWW.

5. We find it still somewhat clumsy to draw transition metal and organometallic structures. A significant portion of the time, unless we are careful about the order in which we connect atoms, we generate meaningless structures.

It would have been beneficial if the manuals were a little clearer about the care that one must take in applying and interpreting the various calculations. We suspect that this shortcoming is common among similar programs, although there are several excellent optional texts accompanying MacSpartan/PCSpartan. The problem is that if one does not take the time to explore calculations in some detail and understand the limitations of the various methods, one arrives at conflicting and confusing results. For example, the HyperChem manual introduces frontier orbitals with a calculation of HOMO electron density for benzofuran (p 141 in the Computational Chemistry manual). The figure is generated from an extended Hückel calculation, and agrees with experiment. However, if one uses another more sophisticated semi-empirical method, PM3, the HOMO predicts a different carbon atom to be the site of nitration (electrophilic attack).

HyperChem's usefulness throughout the curriculum, from introductory chemistry through organic, biochemistry, inorganic, and physical chemistry is unsurpassed. For example, we have had students in our introductory chemistry course perform geometry optimization on a series of molecules to verify VSEPR rules, visualize and rotate three-dimensional representations of molecular orbitals, generate energy-level diagrams of molecular electronic states, and visualize the different vibrational motions of CO₂. HyperChem's new transition state search feature should be useful in organic, inorganic, and physical chemistry courses. As another example, Hyperchem affords the opportunity to explore electron correlation, a topic of interest to physical chemists, by providing two post Hartree-Fock methods: configuration interaction and MP2. The wide applicability of the program not only justifies the cost, but also provides a thread that ties many different courses together. Moreover, the program is sufficiently sophisticated to be a valuable research tool.

William F. Coleman
Christopher R. Arumainayagam
Department of Chemistry
Wellesley College
Wellesley, MA 02181