Optimizing molecular geometries is an important and ubiquitous task in computational chemistry. Determining the lowest-energy geometry of a given molecular structure is one of the most fundamental questions that a researcher can ask about it: the ground-state geometry of a molecule is essentially the "shape" of a molecule, and informs how it will interact with its environment and all downstream targets. For complexes of two or more molecules, optimizing the geometry can be used to determine the strength and orientation of the intermolecular interaction, and for potentially reactive species, geometry optimization can also determine if and how a reaction will take place.

Rowan's software makes it easy to run high-quality geometry optimizations. We stream results to your browser while an optimization is running, so you don't have to wait to see what's happening, and our algorithms and software are so fast that it's frequently possible to watch optimizations occur in real-time. Here's a video of a geometry optimization run in Rowan, showing just how fast a calculation can be, and here's the calculation shown in the video:

With Rowan, you can conduct geometry optimizations with lots of different computational methods, all through the same interface:

- Density-functional theory, including modern range-separated hybrids and meta-GGA functionals, with a variety of different basis sets and solvent models
- "Composite"/"3c" quantum chemical methods from Stefan Grimme's lab that display excellent performance at a reduced computational cost
- Semiempirical methods from xTB, which can handle all sorts of tricky transition-metal and lanthanide-containing structures while running extremely quickly
- Machine-learned interatomic potentials that mimic the performance of quantum chemistry while running far faster

AIMNet2

GFN-FF

GFN0-xTB

GFN2-xTB

HF-3c

B97-3c

rÂ²SCAN-3c

Ï‰B97X-3c

B3LYP

M06-2X

Ï‰B97X-D3

Ï‰B97M-D3BJ

PBE

PBE0

rÂ²SCAN

TPSS

TPSSh

M06-L

M06

CAM-B3LYP

Ï‰B97X-V

Ï‰B97M-V

DSD-BLYP-D3BJ

HF

Behind the scenes, Rowan automatically adjusts the optimization parameters and coordinate system used to make geometry optimizations as robust and performant as possible. Rowan also supports the addition of arbitrary geometric constraints to all geometry optimizations, making it easy to e.g. assess the interaction energy of two species at exactly a 3.0 Ã… distance, or keep a specific bond angle frozen to where it is in the transition state while relaxing other coordinates.

Rowan can initialize a geometry optimization from a variety of input file formats, including crystal structure `.cif`

files and SMILES strings.