Continued from The Flat Molecule Got a Second Look.
I opened my notebook after the team's 28 October update. Some calculations had finished, a preparation run had stopped, and several scientific results were still unpublished. I needed to know what each entry actually told us.
In my last story, the team had checked coronene, a small molecule used to model a carbon surface. The next pages concern ethanol and formaldehyde. Before turning to the new work, I also need to record a correction from earlier in October.
A comparison and a missed promise
On 6 October, the team completed a comparison with published data for ethanol near coronene. The rule for judging that comparison had been signed before the final energy calculations. The project had also promised to publish the range of agreement it would accept before making the comparison.
That advance publication did not happen. The study page recorded a correction on 7 October: the criterion had not been publicly preregistered. Signing a rule internally does not make it available for readers to check in advance. The numerical report and signed rule remain unpublished, with no publication date set.
Source: R02 correction, 7 October, and status update, 28 October 2026.
Five more places to start
Place one molecule near another, and there is more than one way to arrange them. A calculation begins with a particular arrangement. Starting somewhere else can lead to a different candidate structure, so one starting point does not tell us what the others would find.
The original plan included three positions and two forms of ethanol, giving six starting combinations. One was already represented by the first candidate. The team completed calculations for the remaining five on 7 October. Internal review and the project decision were recorded on 8 October.
This was a follow-up to the earlier comparison. The update records completion and internal review, but does not say what the five additional starts found. Their scientific outcome remains unpublished.
Source: R02 progress update, 28 October 2026.
A reference model for formaldehyde
Formaldehyde appears on the next page. The team completed structure optimizations on 9 October and single-point energy calculations on 10 October. Both stages were internally reviewed. These are diagnostic calculations on a small reference carbon model.
An optimization searches for an arrangement of atoms under the chosen computational model. A single-point calculation then evaluates the energy at a specified arrangement. Finishing these calculations does not by itself establish that a structure is a local minimum.
These are not results for target prefilter materials. The wider project aims to reduce interference from ethanol while preserving the formaldehyde signal. Whether a physical prefilter can do that will need experimental testing.
Source: R02 progress update, 28 October 2026.
The run that stopped
The latest preliminary run for the planned comparison of methods stopped on 27 October. A disk-monitoring safeguard reported a technical failure. The public record identifies a computing problem, without reporting a completed scientific comparison.
The project plans a separate comparison, using a more demanding, higher-level calculation to check the method used so far. Both would evaluate the same arrangements of atoms. This keeps the structures fixed while examining how the choice of calculation method changes the energies.
The preliminary stage runs quantum-chemistry calculations and measures the time and computing resources they require. Those measurements are needed to plan the comparison. The technical stop left this preparation unfinished, without resolving the scientific question.
Some eligible calculations completed in earlier preparatory runs may be used later, after the required checks. The record therefore needs to distinguish work that finished from work that stopped, even within the same preparation stage.
The next page
The next task is to resolve the technical interruption and complete the preliminary higher-level calculations needed to plan the comparison on the same structures. Whether and when a report of that comparison is published is a separate decision.