A question about the IR absorption modi of diiron nonacarbonyl on chemistry.se found two answers. One of them provided an illustration with animated vibrations.
In the 1990s, the Gasteiger group offered with TeleSpek the interested the possibility to submit organic structures, e.g., as a SMILES string for a simulation/prediction of likely IR spectra. At present, this service appears as no longer active:
(a couple of screen photos taken from the project's page, tab «how to use»).
The question to the readers of mattermodelling.se: Do you know about a similar, yet ongoing service which is able to provide such a (I guess, by today's standards much more affordable) quick, installation free prediction of IR modi (incl. animation), which would be instructive, e.g. for teaching/self study? With this purpose in mind, the implementation searched needn't be at high level of theory.
A first idea I had was MolCalc by Jensen and Kromann. Likely to preserve resources, the submission of the SMILES string O=C1[Fe]2(=C=O)(=C=O)(=C=O)C(=O)[Fe]1(=C=O)(=C=O)(=C=O)C2=O
Wikipedia uses to describe the complex is deflected with «Stop Casper. Max ten heavy atoms.» when aiming to enter the calculation of properties. (As far as I recall, the service by Gasteiger et al sent an e-mail about a successful computation, even if this meant multiple hours of work.)