Twelve problems. Each one is hard enough that a careful group, using today’s methods, should not already be able to clear it. Each one is written so a standard laboratory can grade an attempt in days, or at most a couple of weeks. The pass mark is a number, a named control, or both.
This is a public draft of a last reasonable test for a method in chemistry, including a method that uses a model. The method may be new. The measurement may not be. The rules for what belongs on the list are written under Criteria. Further problems are kept below as candidates. They are not part of the twelve.
No problem is filed under this field.
Hydroxylation at a chosen carbon
Install a hydroxyl group at one preregistered C–H site, and not at the others, on most of a blinded set of molecules.Claim
Given an ordinary organic molecule and one carbon–hydrogen bond marked in advance, a single written procedure replaces that hydrogen with a hydroxyl group and does not do the same at the other carbon–hydrogen bonds. The marked carbon may be one the usual catalysts would rather leave alone.
Why it matters
A hydroxyl group is how a lead compound becomes a metabolite, a handle for a conjugate, or the oxidation state a natural-product route spends several steps reaching. Those steps disappear only if the site can be nominated. A method that oxidizes whichever C–H bond is already the most reactive does not change that practice.
Current best
Two small-molecule catalysts set the present standard, and both select the site themselves.
Gormisky and White reported non-heme iron catalysts that enhance or overturn a substrate’s innate preference for aliphatic C–H oxidation, in preparative yields, together with a model that ties the outcome to electronic, steric, and stereoelectronic parameters of the sites (doi:10.1021/ja407388y (opens in a new tab)). The supporting information of that paper states that primary sites are uniformly not oxidized under those conditions and are excluded from the model. The predicted site is the one the model ranks, not a carbon marked by an outside laboratory.
Chan, Palone, Bietti, and Costas later hydroxylated the primary C–H bonds of tert-butyl groups with 1 mol percent of a manganese catalyst and hydrogen peroxide in nonafluoro-tert-butyl alcohol (doi:10.1002/anie.202402858 (opens in a new tab)). On 2,2,3,3-tetramethylbutane the primary alcohol formed in 51 percent yield by gas chromatography, beside 2 percent of the aldehyde, for 96 percent selectivity of the alcohol over further oxidation. Replacing acetic acid with 4 equivalents of urea raised the combined yield of oxidation products to 71 percent, with 96 percent product selectivity and a mass balance above 99 percent. Two further hydrocarbons gave the primary alcohols in 60 percent and 53 percent yield by the same assay, at or above 91 percent product selectivity, and only the primary site reacted even though secondary C–H bonds were present. A capecitabine-derived carbamate gave the primary alcohol in 28 percent isolated yield, with 39 percent of the starting material recovered. The same paper shows the preference slipping as the tert-butyl group moves away from an electron-withdrawing group: the ratio of tert-butyl oxidation to oxidation at other sites is 1.3 for the C6 pivalate. These are catalyst-selected sites on substrates chosen for the study.
No report opened for this problem takes an arbitrary marked carbon on a blinded set, writes the procedure before any experiment on those molecules, and returns the alcohol at that carbon.
Specification
An independent laboratory preregisters 20 molecules. Each has a molecular weight between 150 and 500, at least four constitutionally distinct carbon–hydrogen sites, and no protecting group that is not already part of the structure as drawn. On each structure the laboratory marks exactly one carbon. The marked carbon bears at least one hydrogen. The required product is the alcohol at that carbon, not the ketone, the aldehyde, or the carboxylic acid.
At least six of the twenty marked carbons are disfavored, by a rule fixed before the structures are released. A marked carbon is disfavored only when all three of the following are true: it is a methyl or a methylene; it is not benzylic, not allylic, and not adjacent to oxygen, nitrogen, or sulfur; and the same molecule contains at least one other carbon–hydrogen site that is either a tertiary aliphatic C–H or a benzylic C–H. The laboratory records which six meet the rule. The attempt sees the marked structures and can apply the same rule.
The structures are revealed together. Within 24 hours, and with no experiment on any of the 20 molecules, the attempt submits one written procedure. Only the substrate changes from molecule to molecule. The laboratory then runs each procedure once, as written, on 0.2 to 0.5 mmol of substrate. A second procedure may be submitted in the same 24 hours and is used only if the first fails to produce isolable material. No third procedure is allowed, and nothing is changed after the first spectrum.
The isolated material is characterized by high-resolution mass spectrometry, a proton spectrum, a carbon spectrum, and an HSQC and HMBC pair sufficient to assign the site of hydroxylation. The regioisomer ratio is the proton-NMR integral of the specified alcohol divided by the sum of the integrals of every other monohydroxylated isomer in the crude mixture, measured before chromatography against a weighed internal calibrant.
Pass threshold
The attempt passes when at least 16 of the 20 substrates give the specified alcohol in an isolated yield of at least 30 percent, with a regioisomer ratio of at least 10 to 1, and at least 4 of those 16 are among the 6 disfavored sites.
The count, the 30 percent yield, the 10-to-1 ratio, and the 4-of-6 disfavored requirement are pass marks defined by this problem. They are not claimed as records. Thirty percent is below the 51 percent gas-chromatography yield of the primary alcohol, and below the 71 percent combined oxidation yield, that Chan and co-workers report on a hydrocarbon whose reactive site is the one their catalyst prefers. It is there so the product is isolated and assigned rather than detected. Ten to one means the specified site accounts for at least about 91 percent of the monohydroxylated material. The disfavored-site count is what puts the bar past a method that only oxidizes the tertiary or benzylic C–H bond the substrate already offers.
Controls
A no-reagent blank, identical except for the absence of the catalyst or reagent named in the procedure, is run for two of the twenty substrates, chosen by the laboratory before the procedures are opened. One of the two is disfavored. The blank must not produce the specified alcohol above 2 percent yield by the same NMR assay. If it does, the batch is invalid and is repeated with a cleaned setup.
Site assignment is done by a chemist who did not write the procedure, using only the spectra and the starting structure. If that chemist cannot place the hydroxyl group, the substrate counts as a failure. Over-oxidation of the specified alcohol to a ketone or carboxylic acid does not count as the alcohol.
Allowed
Any reagent, catalyst, enzyme, light source, or electrochemical cell that the written procedure names. A directing group that is already present in the preregistered drawing may be left in place. Computation before the 24-hour clock starts is allowed. A different catalyst for a disfavored site is allowed only if that choice is written in the procedure submitted before any experiment on the set.
Not allowed
Installing a directing group, a halogen, a boron substituent, or any other functional handle on the way to the alcohol. A separate protection step that is not in the preregistered structure. Screening conditions on the test substrate, or on a close analog made for the purpose, after the structures are revealed. Counting a ketone or a hydroperoxide as the alcohol. Using a different substrate that already bears the hydroxyl group. Treating a disfavored site as optional and reporting only the fourteen ordinary sites. A different published method for each marked site. One written procedure applies to every substrate, with only the substrate changed. The single backup, if it is used, applies to every substrate or to none.
Partial success
At least 10 of the 20 substrates give the specified alcohol in at least 15 percent isolated yield, with a regioisomer ratio of at least 5 to 1, at least 2 of the 10 are disfavored sites, and the blanks pass. Partial success is not a solution. Fifteen percent, 5 to 1, and 2 of 6 are pass marks for this lesser bar only.
Sources
The 16-of-20 count, the 4-of-6 disfavored requirement, the 30 percent yield, the 10-to-1 ratio, the 2 percent blank ceiling, and the partial-success figures are pass marks defined here.
Cited measurements:
- Gormisky and White, Journal of the American Chemical Society, 2013, 135, 14052–14055, doi:10.1021/ja407388y (opens in a new tab). Catalyst-controlled aliphatic C–H oxidation in preparative yields, with a model of innate site preference. The supporting information states that primary sites are not oxidized under those conditions and are excluded from the model.
- Chan, Palone, Bietti, and Costas, Angewandte Chemie International Edition, 2024, 63, e202402858, doi:10.1002/anie.202402858 (opens in a new tab). The 51 percent, 2 percent, 60 percent, 53 percent, 71 percent, 96 percent, 91 percent, 28 percent isolated, 39 percent recovered, and 1.3 site ratio quoted above are taken from that paper.
Purity, where a purity number is reported alongside the isolated yield, is measured by the absolute quantitative proton NMR assay of Pauli and co-workers, Journal of Medicinal Chemistry, 2014, 57, 9220–9231, doi:10.1021/jm500734a (opens in a new tab).
A synthesis that runs as written
A short written route to each of a blinded set of molecules is executed once, unchanged, and most of the molecules come out pure.Claim
For a set of small, drug-like molecules disclosed only at the start, a route written in that window can be followed once, by a chemist who did not design it, with no change of reagent, solvent, temperature, time, or workup, and the product that comes out is the molecule that was drawn.
Why it matters
A route that works only after the person at the bench repairs it is a prompt, not a procedure. Process chemistry, contract synthesis, and any model that writes experimental sections are all graded by the same fact: someone else can follow the page and isolate the structure that was drawn, including its configuration.
Current best
Published executions of computer-written routes still allow the bench chemist a revision, or they choose easy analogs of commercial drugs.
Klucznik, Mikulak-Klucznik, McCormack, and co-workers had a program plan routes to eight diverse, medicinally relevant targets, and all eight were carried through in the laboratory (doi:10.1016/j.chempr.2018.02.002 (opens in a new tab)). The paper states that no alterations in the proposed routes were allowed except straightforward adjustments of reaction conditions, with temperature, solvent, the specific base, and the catalyst given as examples, for the sake of optimization. For the industrial targets, each step was to be attempted no more than three to five times, and the material was to be delivered on a scale of at least a few hundred milligrams within eight weeks, corresponding to no more than 70 hours of bench work, at an HPLC purity above 98 percent in most cases.
Makkawi, Beker, Wołos, Grzybowski, and co-workers later obtained twelve of thirteen computer-designed analogs of ketoprofen and donepezil (doi:10.1039/D5SC00070J (opens in a new tab)). The searches were limited to five steps and 180 reaction classes used in medicinal chemistry. In the ketoprofen cluster the reactions were run under conditions identical or very similar to the software’s suggestions, without thorough optimization. In the donepezil cluster the reactions were run under the computer-suggested conditions except that one oxidation was changed from Dess–Martin to Swern; the yields were poor to moderate, no condition optimization was attempted, and the fluorinated analog was not obtained because elimination competed. The authors note that these simple syntheses could have been designed by a human expert. The targets were analogs of two commercial drugs, chosen by the group that wrote the routes.
Neither study is an independent laboratory’s blinded set, run once, with a change of solvent or temperature scored as a failure, and with identity and 95 percent purity judged by NMR.
Specification
An independent laboratory preregisters 8 molecules. Each has a molecular weight between 250 and 450, at most two stereocenters, at most four rings, and only the elements C, H, N, O, S, F, and Cl. None is a commercial compound on the day of preregistration. At least four of the eight specify a relative or absolute configuration. No more than two of the eight may be analogs of the same commercial parent, meaning they share that parent’s complete ring system and differ from it only by substituents. The laboratory does not disclose intermediates it has in mind.
Within 24 hours of disclosure, the attempt submits one route per molecule. Each route has at most five linear steps from materials that a named supplier sells on that day, with catalog numbers. The procedure states solvent, equivalents, temperature, time, and workup, including any chromatography. The longest procedure, followed as written, takes no more than ten working days for one chemist. The eight routes may be run in parallel.
The laboratory executes each procedure once, as written. The chemist who runs it did not write it. A change of reagent, order, solvent, temperature, time, or workup is a failure of that target, even if a later change would have worked, and even if the change would have been called a straightforward adjustment. One backup route per molecule may be submitted in the same 24 hours and is opened only if the first route fails to yield the product. Identity is judged by a chemist who did not write the route, from a proton spectrum, a carbon spectrum, and a high-resolution mass measurement. Where the drawing specifies a relative or absolute configuration, the same chemist must be able to confirm it from a method named in the route before execution: a coupling constant and NOE analysis, a chiral chromatographic comparison with a reference of known configuration, or a chemical correlation. Purity is an absolute quantitative proton NMR assay with a weighed internal calibrant.
Pass threshold
At least 6 of the 8 molecules are isolated in at least 10 mg, with purity of at least 95 percent by that assay, and with spectra consistent with the drawing, including the specified configuration. At least 3 of the 6 are among the molecules whose configuration was specified.
Six of eight, 3 of those 6, 10 mg, 95 percent, and five linear steps are pass marks defined here. Ten milligrams is required so that the proton spectrum and the purity assay can both be run on the isolated sample. Ninety-five percent is the cleanliness required before the sample counts, not a published yield record. Five steps and ten days are there so the grade fits in a couple of weeks. The ban on condition changes is what separates this bar from the eight successful Chematica routes, which permitted those changes and more than one attempt at a step.
Controls
For each isolated product, a blank NMR sample of the internal calibrant in the same solvent is recorded on the same day. The product signals used for the assay must be free of the calibrant and of residual solvent. If the independent chemist judges the spectra ambiguous between the drawn structure and another, that target fails rather than being counted by majority vote.
A route that finishes on a compound with the correct mass but the wrong constitution, or the correct constitution and the wrong configuration, fails.
Allowed
Any reaction the procedure can describe, including biocatalysis, photochemistry, and electrochemistry. Protecting groups are allowed if they are in the written route and are removed by the end. Planning before the clock starts is allowed. The one backup route submitted with the first route is allowed.
Not allowed
Editing the procedure after a step fails, including a change of solvent, temperature, base, or catalyst. Buying the target, or an intermediate that is the target in disguise. A chromatographic separation that isolates a different isomer from the one drawn and calls it close enough. Human improvisation that is not in the written backup route. Claiming success from a liquid-chromatography peak without the NMR assignment. Counting a route whose configuration was not confirmed by the method named before execution. A route copied from a published synthesis of that exact target.
Partial success
At least 4 of the 8 molecules meet the same isolation, purity, and identity rules, and at least 2 of the 4 had a specified configuration. Four of eight is a pass mark for this lesser bar. It is not a solution.
Sources
The step limit, the success count, the configuration subcount, the 10 mg floor, the 95 percent purity, and the one-execution rule are pass marks defined here.
Cited executions, neither of which meets this bar:
- Klucznik, Mikulak-Klucznik, McCormack, and co-workers, Chem, 2018, 4, 522–532, doi:10.1016/j.chempr.2018.02.002 (opens in a new tab). Eight of eight planned routes were completed. Condition adjustments and up to three to five attempts per step were part of the protocol, on a timescale of weeks.
- Makkawi, Beker, Wołos, Grzybowski, and co-workers, Chemical Science, 2025, 16, 8383–8393, doi:10.1039/D5SC00070J (opens in a new tab). Twelve of thirteen analogs of two commercial drugs were obtained. One suggested oxidation was replaced, one analog failed, and the targets were author-selected analogs rather than a blinded set.
The purity assay is the absolute quantitative proton NMR method of Pauli and co-workers, Journal of Medicinal Chemistry, 2014, 57, 9220–9231, doi:10.1021/jm500734a (opens in a new tab).
The named enantiomer, written in advance
For most of a blinded set of one-step reactions, across several reaction classes, a procedure written before any experiment returns the specified enantiomer.Claim
Given a prochiral starting material and a drawing of one enantiomer of a one-step product, a procedure written with no experiment on that substrate produces that enantiomer, in quantity, and not the mirror image. The set is not twelve examples of one named reaction.
Why it matters
The slow step in making a chiral building block is usually not the reaction that finally works. It is the screen that finds which catalyst, and which sense of that catalyst, works on this substrate. A procedure that names the enantiomer before that screen is a last test of a catalyst-selection method, including one that uses a model, and it is how a catalog of non-commercial enantiomers would actually be filled.
Current best
The strongest recent tests predict enantioselectivity inside one reaction class. They do not hand a laboratory twelve unseen transformations, in several classes, and a procedure that is then run once.
Ocampo, Altundas, Bock, Feiz, and Denmark trained models on 1007 Sharpless asymmetric dihydroxylations and tested them on 15 alkenes that had not been reported under those conditions (doi:10.1021/acscentsci.5c00900 (opens in a new tab)). The models predicted the enantioselectivity within 16 percentage points of ee for 10 of the 15 alkenes. The reagent was commercial AD-mix, whose sense of enantioselectivity is already known, and the exercise was the magnitude of ee in one reaction. An error of 16 points does not keep a 90 percent ee prediction inside a 95-to-5 enantiomer ratio, and 5 of the 15 alkenes fell outside the 16-point window.
Hoque, Chang, Yu, and Sunoj reported an ensemble model for one catalytic asymmetric β-C(sp³)–H activation, and for 15 out-of-bag reactions the predicted and experimental enantioselectivities agreed with a root-mean-square error of 6.42 percentage points of ee and an R² of 0.93 (doi:10.1039/D5SC01098E (opens in a new tab)). That is one reaction class, with ligands drawn from the family the model was trained on. It is not a blinded multi-class set, and it is not a claim that each named enantiomer was isolated in 40 percent yield at a 95-to-5 ratio from a procedure frozen before the experiment.
Singh and Hernández-Lobato used meta-learning to classify enantioselectivity in asymmetric hydrogenation of olefins, including an out-of-sample literature set, with an area under the precision-recall curve of about 0.91 for a support set chosen from similar known reactions (doi:10.1038/s41467-025-58854-8 (opens in a new tab)). The output is a prediction on published reactions, not a procedure that was then run to produce a named absolute configuration.
A careful chemist can already write a first procedure that is often right inside a privileged class, such as a ketone hydrogenation with a commercial Noyori catalyst on a close analog. The unpublished, cross-class part of this problem is what that habit does not cover.
Specification
The laboratory preregisters 12 transformations. Each is a single step from a commercial or previously prepared achiral or prochiral starting material to a specified enantiomer of a specified product. Before the attempt sees them, the laboratory assigns every transformation to a reaction class and uses at least four classes. The class list is: hydrogenation or transfer hydrogenation; alkene oxidation; carbonyl addition or carbonyl reduction; conjugate addition; cycloaddition; C–H functionalization; rearrangement; and any further class the laboratory names in the preregistration, provided it is not a synonym of a class already used. The four transformations described next must occupy four different classes.
At least four of the twelve are marked, before the attempt sees them, as having no published method that reports that exact product, either enantiomer, in an enantiomer ratio of 90 to 10 or higher. The laboratory records the search, the date, and the query. A hit at or above 90 to 10 on that structure removes the mark, and the laboratory replaces the substrate before the procedures are submitted. The other substrates may be known reactions, so that a method cannot pass by attempting only exotic cases it then declines.
Within 24 hours, the attempt submits one procedure per transformation: catalyst or reagent, loading, solvent, temperature, time, and workup. Loading of any catalyst is at most 5 mol percent. No experiment on a test substrate, or on a stand-in made to resemble it, is allowed after disclosure. One backup procedure per substrate may be submitted at the same time.
The laboratory runs the procedure once. The isolated product is identified by NMR and high-resolution mass spectrometry. The enantiomer ratio is measured by chiral chromatography or chiral supercritical-fluid chromatography against a racemic sample of the same constitution, prepared independently and injected on the same method the same day. The specified enantiomer is assigned by comparison, on the same chiral chromatography method, with an authentic sample of known configuration that the grading laboratory already holds or buys as a commercial standard. If no such sample can be had without a new method-development campaign, the laboratory replaces the substrate before disclosure. Vibrational circular dichroism and a heavy-atom crystal structure are not the grade. Optical rotation alone is not sufficient.
Pass threshold
At least 10 of the 12 transformations, including at least 3 of the 4 transformations with no published method, give the specified enantiomer in an isolated yield of at least 40 percent and an enantiomer ratio of at least 95 to 5. The 10 successes must include at least three reaction classes.
The count, the class requirement, the 40 percent yield, the 95-to-5 ratio, and the 5 mol percent ceiling are pass marks defined here. Ninety-five to five is 90 percent enantiomeric excess. It is not taken from a paper’s best example. The Denmark test, the closest numerical comparison, landed within 16 points of ee for 10 of 15 alkenes inside a single commercial reaction, which is a different and narrower claim.
Controls
The racemic reference must show both enantiomers on the same method. A procedure whose chromatogram shows one peak is not credited unless the racemic sample resolves on that method and a spiked sample shows which peak is which. A reaction run without the chiral catalyst, otherwise identical, is performed for two substrates chosen in advance, one of them from the unpublished four. That control must not itself reach 95 to 5. If it does, the analytical method or the substrate is contaminated and the batch is invalid.
The literature search for the four unpublished transformations is part of the grade. If a published 90-to-10 method for that exact product is produced after the procedures are locked, that substrate is scored as not unpublished. It is not given to the attempt as a free pass, and it cannot fill the 3-of-4 requirement.
Allowed
Any chiral catalyst, ligand, or enzyme named in the procedure, within the loading rule. A different catalyst for each substrate. Computation and literature lookup before the procedures are locked.
Not allowed
A stoichiometric chiral auxiliary that is attached and later removed. Changing solvent, catalyst, or temperature after the first experiment. Separating enantiomers by chromatography on a chiral stationary phase and calling the separation a synthesis. Classical resolution. Filling the unpublished quota with four examples of one reaction class, or filling the whole set with one class. Using a published procedure for a substrate the laboratory has marked as unpublished: that substrate is replaced before grading if the search was wrong at the start, and it fails the unpublished count if the miss is found later.
Partial success
At least 8 of the 12 transformations reach an enantiomer ratio of at least 90 to 10 and an isolated yield of at least 40 percent, including at least 2 of the 4 unpublished ones, and the 8 cover at least two reaction classes. These figures are pass marks for the lesser bar. Partial success is not a solution.
Sources
The yields, ratios, loading, class counts, and success counts are pass marks defined here.
Cited measurements:
- Ocampo, Altundas, Bock, Feiz, and Denmark, ACS Central Science, 2025, 11, 1640–1650, doi:10.1021/acscentsci.5c00900 (opens in a new tab). Within 16 percentage points of ee for 10 of 15 previously unreported Sharpless dihydroxylations.
- Hoque, Chang, Yu, and Sunoj, Chemical Science, 2025, 16, 13276–13290, doi:10.1039/D5SC01098E (opens in a new tab). Root-mean-square error of 6.42 percentage points of ee, R² of 0.93, on 15 out-of-bag reactions in one C–H activation class.
- Singh and Hernández-Lobato, Nature Communications, 2025, 16, doi:10.1038/s41467-025-58854-8 (opens in a new tab). Meta-learning classification of asymmetric hydrogenation. An out-of-sample area under the precision-recall curve of 0.9147 ± 0.0022 is reported for the clustered-support model. No synthetic procedure was executed for this problem’s grade.
Where a purity figure is reported with the isolated yield, it is measured by the quantitative proton NMR assay of Pauli and co-workers, Journal of Medicinal Chemistry, 2014, 57, 9220–9231, doi:10.1021/jm500734a (opens in a new tab).
Ammonia from nitrogen and water
Make ammonia from dinitrogen and water near room temperature and pressure, with isotope proof that the nitrogen came from N₂.Claim
Ammonia is produced from dinitrogen and water at near-ambient temperature and pressure, in amounts that cannot be explained by contamination, without plating a metal to activate N₂ and without consuming a stoichiometric reductant.
Why it matters
Ammonia is the feedstock for fertilizer and a candidate hydrogen carrier. The industrial Haber–Bosch reaction, as Izelaar and co-workers state it, runs at 300–500 °C and 200–300 bar, and most of its emissions come from making hydrogen by reforming methane. A route that uses N₂, water, and electricity or H₂ near room temperature would let ammonia follow renewable power. The laboratory version of that claim has been crowded with false positives, so the grade is a rate, an efficiency, and an isotope balance together.
Current best (with sources)
Izelaar and co-workers state that reported aqueous electrochemical rates were still extremely low, 0.003–14 nmol of ammonia per square centimetre per second. That range is their summary of the literature, not a champion catalyst they certified. At three electrons per ammonia and 96,485 C mol⁻¹, 14 nmol cm⁻² s⁻¹ is 4.05 mA cm⁻² if every electron made ammonia. Andersen and co-workers found that aqueous pure-metal electrodes produced no ammonia under a quantitative ¹⁵N₂ protocol. Lithium electrodeposition in tetrahydrofuran did, which is why metal plating is excluded here rather than ignored.
Two later metal-plating results sit outside this problem and are the reason the exclusion is explicit. Li, Zhou, Fu, and co-workers ran a lithium-mediated flow electrolyser for 300 h at 1 bar and room temperature, with a 25 cm² electrode, a current-to-ammonia efficiency of 64 ± 1%, and about 98% of the ammonia in the gas phase. The abstract couples nitrogen reduction to hydrogen oxidation in a chain-ether electrolyte. It is not water as the only proton source, and it is lithium-mediated. Goyal and co-workers report a magnesium-mediated cell at an ammonia faradaic efficiency of 25.28 ± 3.80% and a total current density of −45 mA cm⁻², which they convert to an ammonia partial current density of −11.30 ± 1.77 mA cm⁻², under 6 bar of N₂, with ethanol as the proton donor. That partial current is above this problem’s rate mark and still fails the pressure, the efficiency, the proton source, and the plating rule.
Specification
The reaction uses N₂ at a partial pressure of at most 1 bar and a temperature of at most 40 °C. The only proton source is water. The reductant is electrons or H₂. The geometric area is the projected area of the electrode in contact with the electrolyte, or, for a thermal bed, the internal cross section of the bed. It is written down before the run. It is not an electrochemical surface area.
The attempt reports a continuous run of at least 4 hours. Ammonia is quantified by two methods that agree within 20% relative: proton NMR of the ammonium salt in a non-exchanging solvent, and a colorimetric or ion-chromatographic assay. The figure of merit uses the average of the two assays after the blank subtraction below.
For an electrochemical cell the partial current density is 3 × 96,485 C mol⁻¹ × (blank-subtracted moles of ammonia) / (duration × geometric area). The faradaic efficiency is that ammonia, times three electrons, divided by the charge passed during the same 4 hours. Charge is not blank-subtracted. For a thermal reactor fed with H₂, the molar rate uses the same blank-subtracted moles, and the electron efficiency is 3 × moles of ammonia / (2 × moles of H₂ consumed). H₂ consumption is inlet flow minus outlet flow, both measured. If the consumed H₂ is less than 10% of the H₂ fed, the run is invalid, because the difference is then too small to support a 50% efficiency.
A catalytic claim requires that the blank-subtracted moles of ammonia be at least ten times the moles of transition metal in the catalyst charge. Plated lithium, sodium, magnesium, calcium, aluminium, or zinc does not count as that catalyst.
Pass threshold
The blank-subtracted partial current density to ammonia is at least 10 mA per square centimetre of geometric area, or the thermal molar rate is at least 34.6 nmol cm⁻² s⁻¹ on the geometric area defined above, and the faradaic or electron efficiency is at least 50%, both averaged over the 4-hour run. The isotope controls pass, and the two assays agree within 20% relative.
These current, rate, efficiency, temperature, pressure, assay-agreement, and turnover figures are pass marks defined here. Ten milliamperes per square centimetre is 10 × 10⁻³ / (3 × 96,485) = 34.6 nmol cm⁻² s⁻¹, above the top of the aqueous range Izelaar and co-workers summarized. Fifty percent efficiency is a pass mark, not a cited record. The magnesium partial current of about 11 mA cm⁻² does not clear this threshold: it was measured at 6 bar with ethanol as the proton donor.
Controls
The controls follow the quantitative isotope protocol of Andersen and co-workers and the cautions of Choi and co-workers.
- The N₂ and argon feeds are scrubbed, and the residual ammonia and nitrogen oxides in the feed are measured before the run. They must be too small to account for the ammonia reported.
- An argon electrolysis, and a dinitrogen experiment at open circuit, each of the same duration, produce less than 10% of the ammonia measured in the N₂ run. The ammonia subtracted from the N₂ run is the ammonia found in the argon electrolysis, not the sum of the two blanks.
- A ¹⁵N₂ run, with the labeled gas itself assayed for labeled ammonia and nitrogen oxides before use, produces ¹⁵N-ammonia in an amount that agrees within 20% relative with the ¹⁴N₂ run under the same conditions. Qualitative appearance of a doublet, without that quantitative match, does not count.
- The ¹⁵N fraction in the ammonia matches the ¹⁵N fraction of the feed gas within the uncertainty of the NMR integral.
- Two further N₂ runs reproduce the rate within 20% relative.
Commercial ¹⁵N₂ can carry labeled ammonia. De Ras, Hollevoet, Martens, and co-workers measured about 73 ppm ¹⁵NH₃ in a commercial 98 atom% cylinder, and showed that the impurity alone can exceed typical reported aqueous yields. The feed assay is not optional.
Allowed
Heterogeneous or molecular catalysts, including enzymes, provided the metal-turnover rule is met and water is the proton source. H₂ as the reductant in a thermal catalytic reactor inside the same temperature and pressure limits. A proton-exchange membrane whose only function is to separate the electrodes, if the proton inventory comes from water.
Not allowed
Lithium, sodium, magnesium, calcium, aluminium, or zinc metal as a stoichiometric reductant, and any cycle in which one of those metals is plated and then converts N₂ to a nitride, whether or not the metal is later regenerated. A proton source other than water, including ethanol. Nitrogen oxides, nitrate, or nitrite as the feedstock. A run shorter than 4 hours. A claim based on a single assay. Pressure above 1 bar of N₂, or temperature above 40 °C. Reporting a rate per electrochemical surface area in place of the geometric area.
Partial success
All of the controls pass, the metal turnover is at least 10, the faradaic or electron efficiency is at least 20%, and the partial current density is at least 1 mA cm⁻², or the thermal rate is at least 3.46 nmol cm⁻² s⁻¹, averaged over 4 hours. Twenty percent, 1 mA cm⁻², and 3.46 nmol cm⁻² s⁻¹ are pass marks for this lesser bar. They show that the ammonia came from N₂ and that the process is catalytic. They are not a solution. One milliampere per square centimetre is still below the top of the range Izelaar and co-workers summarized, so this lesser bar is a cleanliness-and-catalysis bar, not a claim to have beaten that summary.
Sources
Aqueous pure-metal electrodes produced no ammonia under the quantitative ¹⁵N₂ protocol, while lithium electrodeposition in tetrahydrofuran did: Andersen and co-workers, Nature, 2019, 570, 504–508, doi:10.1038/s41586-019-1260-x (opens in a new tab). The same paper requires quantitative agreement between the ¹⁴N₂ and ¹⁵N₂ experiments. Choi and co-workers describe why a qualitative isotope spectrum is not enough: Nature Communications, 2020, doi:10.1038/s41467-020-19130-z (opens in a new tab). Izelaar and co-workers state the aqueous rate range and the Haber–Bosch conditions of 300–500 °C and 200–300 bar: ACS Energy Letters, 2023, 8, 3614–3620, doi:10.1021/acsenergylett.3c01130 (opens in a new tab). Labeled ammonia at about 73 ppm in commercial ¹⁵N₂: De Ras, Hollevoet, Martens, and co-workers, Green Chemistry, 2024, 26, 1302–1305, doi:10.1039/D3GC04272C (opens in a new tab). Lithium-mediated continuous operation at 64 ± 1% for 300 h: Li, Zhou, Fu, and co-workers, Nature, 2024, doi:10.1038/s41586-024-07276-5 (opens in a new tab). Magnesium-mediated partial current and conditions: Goyal, Gande, Bhawnani, Hamlyn, Farghaly, and Singh, Advanced Science, 2025, 12, 2504882, doi:10.1002/advs.202504882 (opens in a new tab). The 10 mA cm⁻², 34.6 nmol cm⁻² s⁻¹, 50%, 20% assay agreement, 4-hour, 40 °C, and 1 bar requirements are pass marks defined here. The conversion of 14 nmol cm⁻² s⁻¹ to 4.05 mA cm⁻², and of 10 mA cm⁻² to 34.6 nmol cm⁻² s⁻¹, is arithmetic on the Faraday constant, not a new measurement.
Ethylene from carbon dioxide, with less voltage
A full cell reduces carbon dioxide to ethylene at high partial current and at a full-cell energy efficiency past the laboratory reports cited here, for a day.Claim
Carbon dioxide and water are converted to ethylene and oxygen in one cell, at a rate and a full-cell energy efficiency that are both high for a full day, with the carbon in the ethylene coming from the carbon dioxide.
Why it matters
Ethylene is the largest organic feedstock. Making it from CO₂ and water, instead of from steam cracking, only changes the energy balance if the cell voltage and the selectivity are both good. A high faradaic efficiency at an unstated full-cell voltage, or a low voltage held for a few minutes, does not answer that. The grade is the two numbers on the same electrode, averaged over a day, with the voltage taken between the anode and the cathode.
Current best (with sources)
Li and co-workers define the full-cell energy efficiency for 2 CO₂ + 2 H₂O → C₂H₄ + 3 O₂ as
energy efficiency = (1.15 V × faradaic efficiency to ethylene) / full-cell voltage,
with 1.15 V the equilibrium cell potential they give for that reaction. In a membrane-electrode assembly they held a full-cell voltage of 3.65 V for 190 hours with an ethylene selectivity of 64%, and they report the efficiency of that system as 20%. The arithmetic 1.15 × 0.64 / 3.65 is 20%.
Fang and co-workers, in a 2026 membrane-electrode paper, report a full-cell energy efficiency of 29.2% at an ethylene partial current density of 401 mA cm⁻². Their methods write
energy efficiency = [1.23 V + (−E°_A)] × faradaic efficiency / full-cell voltage,
where E°_A is the thermodynamic potential of CO₂ to the product, from the standard Gibbs energy at 298.15 K, and the full-cell voltage is measured without ohmic correction. The main text does not print the value of E°_A used for ethylene. The Li expression and this expression are the same arithmetic when E°_A = 0.08 V, because 1.23 − 0.08 = 1.15. This problem does not assert that the 2026 experiment used 0.08 V. It cites 29.2% as that paper’s reported number and grades only with 1.15 V.
The same paper’s long run is a different operating point. For 1020 hours at 200 mA cm⁻² in 0.1 M KHCO₃ they report a full-cell voltage of approximately 3.5 V and an ethylene faradaic efficiency around 68%, and they replaced the anion-exchange membrane periodically. Applied to those approximate figures, the Li formula gives 1.15 × 0.68 / 3.5 ≈ 22%, and 200 × 0.68 ≈ 136 mA cm⁻² of ethylene partial current. That 22% is computed here. It is not a number the paper prints. A half-cell ethylene energy efficiency of 38.5% at −0.98 V versus RHE, also in that paper, is not a full-cell result and does not enter the grade.
Specification
The cell is a two-electrode device. The anode reaction is water oxidation. The cathode reduces CO₂. The full-cell voltage is the measured voltage between anode and cathode, with no ohmic-drop correction. Ethylene is quantified by gas chromatography against a calibration mixture. Liquid products are quantified and included in the balance, but the graded product is ethylene.
Energy efficiency is computed only with the Li formula above, using 1.15 V. The partial current density is the total geometric current density multiplied by the faradaic efficiency to ethylene. Both numbers are averages over 24 hours of continuous operation, not peak values, and neither the membrane nor the cathode is replaced in that window. A ¹³CO₂ feed, run as a separate experiment of at least 1 hour in the same cell, must show ¹³C-ethylene consistent with the enrichment of the feed.
A polycrystalline copper foil is run in the same cell hardware, the same day, as an instrument control. It must make detectable ethylene, so that a dead cell cannot pass on a calibration error.
The gas is at 1 bar of CO₂. Elevated pressure is allowed only if every compressor’s electrical energy is added to the denominator of the efficiency as an equivalent voltage, energy divided by the charge passed. The simpler experiment, and the one that keeps the grade inside a normal lab, is 1 bar with no compressor term.
Pass threshold
The 24-hour averages are an energy efficiency of at least 40% and an ethylene partial current density of at least 200 mA cm⁻², together, on one electrode in one cell, with no membrane or cathode replacement. The isotope experiment passes, the copper foil control makes ethylene, and the carbon balance below closes. Carbonate and bicarbonate are not counted as ethylene.
Forty percent and 200 mA cm⁻² are a joint pass mark defined here. Forty percent is above the reported 29.2% point and above the 20% point, and it is above the roughly 22% figure computed here for the 1020-hour run. Two hundred milliamperes per square centimetre of partial current, held as a 24-hour average together with that efficiency, is the other half of the mark. Matching either cited paper is not the task.
Controls
- ¹³CO₂ labeling as above. Unlabeled ethylene during that experiment fails the attempt.
- An argon feed in place of CO₂, same current, produces no ethylene above the chromatographic baseline.
- The copper foil control produces ethylene and is reported, including its faradaic efficiency and voltage, but the candidate does not have to beat the foil on every axis. The foil checks the apparatus.
- The outlet flow is measured, not assumed equal to the inlet, so that a change in gas volume does not inflate the faradaic efficiency. Fang and co-workers already record inlet and outlet flow for this reason.
- Liquid products that cross the membrane are counted in the balance. They do not have to be zero. They must not be omitted.
- Carbonate and bicarbonate are quantified on both sides of the membrane. The sum of carbon in outlet CO₂, ethylene, other identified carbon products, and carbonate plus bicarbonate closes to within 10% relative of the CO₂ fed over the same window. That 10% is a pass mark for analytical closure, the same kind of tolerance problem 06 uses, not a figure taken from a paper. A cell that posts a high ethylene faradaic efficiency while most of the consumed CO₂ is carbonate fails.
Allowed
Any cathode catalyst, ionomer, or membrane. Neutral, acidic, or alkaline electrolytes. A gas-diffusion electrode. The compressor term above, if pressure is used.
Not allowed
A sacrificial anode. Stoichiometric chemical reductants. Reporting a cathode potential against a reference electrode in place of the full-cell voltage. iR-corrected voltages in the efficiency formula. A peak efficiency held for minutes. Purchased ethylene. A 24-hour average assembled by discarding intervals. Replacing the membrane, the cathode, or the electrolyte during the graded 24 hours and continuing the average. Using a numerator other than 1.15 V. Counting carbonate or bicarbonate as the ethylene product. Omitting them from the carbon balance.
Partial success
The 24-hour average energy efficiency, on the Li formula, is at least 25% at an ethylene partial current density of at least 100 mA cm⁻², with the isotope and argon controls passing, and with no membrane or cathode replacement. Twenty-five percent and 100 mA cm⁻² are pass marks for this lesser bar. Twenty-five percent sits above the reported 20% and above the roughly 22% computed here for the long run, and below the reported 29.2% peak. Partial success is not a solution.
Sources
Definition, equilibrium potential 1.15 V, and the 20% full-cell result at 3.65 V and 64% ethylene selectivity: Li and co-workers, Nature, 2020, 577, 509–513, doi:10.1038/s41586-019-1782-2 (opens in a new tab). The 29.2% figure at 401 mA cm⁻², the methods formula, the 1020-hour point at about 3.5 V and about 68% faradaic efficiency, the periodic membrane replacement, and the half-cell 38.5% figure: Fang, Huang, Wang, Wang, Feng, Ma, Dai, Zhu, and Jiang, Nature Communications, 2026, 17, 984, doi:10.1038/s41467-025-67719-z (opens in a new tab). The 22% and 136 mA cm⁻² figures are computed here from that paper’s approximate long-run voltage and faradaic efficiency; they are not printed by the paper. The 40%, 24-hour, and 200 mA cm⁻² joint requirement, and the 25% and 100 mA cm⁻² partial bar, are pass marks defined here. Open point: the main text does not state whether 29.2% was held for 24 hours, and it does not print E°_A, so this problem does not treat 29.2% as a 24-hour Li-formula result.
Methanol from methane and oxygen
Oxidize methane to methanol with molecular oxygen above the cited one-oxygen selectivity ceiling, at 5 percent conversion, above a blank reactor.Claim
Methane and molecular oxygen give methanol as the main carbon product at a once-through conversion high enough to matter, in a catalytic reactor held at steady state, and the methanol is not an artifact of an empty tube.
Why it matters
Methanol is made industrially by reforming methane to syngas and then hydrogenating carbon monoxide, at high temperature and in a plant that cannot sit on a remote gas well. A direct oxidation that kept most of the carbon in methanol at a few percent conversion per pass would be the number a process evaluation can compare with that plant. High selectivity at a conversion near zero, and high conversion at selectivity near a third, are both real results. Neither is that number.
Current best
Dummer, Willock, He, Howard, Lewis, Qi, Taylor, Xu, Bethell, Kiely, and Hutchings reviewed the target in Chemical Reviews (doi:10.1021/acs.chemrev.2c00439 (opens in a new tab)). Summarizing technical evaluations collected by Foulds and Gray, they report a consistent view that about 80 percent selectivity at 5 percent once-through conversion, or about 70 percent at 10 percent conversion, would be approximate parity with the syngas route. Their own reading is that a selectivity of about 75 percent at a once-through conversion of at least 5 percent is the minimum that could compete, and that this joint outcome surpassed what had been reliably shown with molecular oxygen when the review was written. The same article describes a reasonably reproducible gas-phase result of about 40 to 60 percent methanol selectivity at about 5 percent methane conversion, with a limiting once-through methanol yield of about 2.5 mol percent, in premixed systems. It also notes that if only one oxygen atom of O₂ is incorporated into methanol and no sacrificial reductant is added, the carbon-selectivity ceiling of that stoichiometry is 80 percent, from 5 CH₄ + 4 O₂ → 4 CH₃OH + CO₂ + 2 H₂O. A higher carbon selectivity requires a path that puts both oxygen atoms into methanol, or an accounting different from that stoichiometry.
Three later reports move one coordinate and not both.
Chauhan, Rana, Gupta, Kalita, Thadhani, and Manna oxidized methane with O₂ over an iron site in a metal–organic framework, in a batch reactor at 125 °C for 40 hours, and reported 94 percent selectivity in the liquid phase and about 90 percent overall selectivity to methanol (doi:10.1038/s41467-024-54101-8 (opens in a new tab)). The charge in that experiment is given as 34 mmol of methane, and the carbon balance accounts for 0.58 mmol of methane consumed (0.572 mmol of carbon-containing products, 98.6 percent balance). Dividing 0.58 by 34 gives 1.7 percent conversion. That division is done here from those two reported quantities; the paper does not print “1.7 percent” as the conversion. The same section describes 0.5 mmol of methanol with 5.8 μmol of iron, which is about 86 turnovers over the 40 hours, again calculated here from those two quantities. A conversion of about 13 percent in the same paper belongs to a two-temperature sequence aimed at acetic acid, not to the methanol condition. The reactor is a batch autoclave, not a steady continuous window.
Li, Shi, and co-workers ran ceria under continuous flow at atmospheric pressure (doi:10.1021/acscatal.5c05829 (opens in a new tab)). At 350 °C and the shorter contact time they examined, methane conversion was 3.8 × 10⁻⁴ percent. Methanol accounted for 80 percent of the carbon-containing products at 350 °C, and a stability test held about 85 percent selectivity and a methanol rate of about 3.3 μmol per gram of catalyst per hour over 55 hours. Lengthening the contact time raised conversion to 3.4 × 10⁻² percent and cut the methanol rate from 3.3 to 0.9 μmol g⁻¹ h⁻¹ by over-oxidation. Selectivity near 80 percent and a conversion of 5 percent are not the same experiment.
Nallapareddy and Underwood, using a non-thermal plasma and active product removal, report methanol selectivity above 30 percent at methane conversion above 50 percent, and a single-step methanol yield of 21.4 percent, near atmospheric pressure (doi:10.1021/acssuschemeng.4c01595 (opens in a new tab)). That is the high-conversion, moderate-selectivity corner. It is not 80 percent selectivity.
No combination opened for this problem is 80 percent carbon selectivity to methanol at 5 percent once-through conversion with O₂, in a steady window, above an empty reactor, at a turnover that rules out a stoichiometric oxidant.
Specification
The oxidant is O₂, diluted as the procedure specifies. No hydrogen peroxide, nitrous oxide, or other terminal oxidant is used, and no sacrificial reductant is added. The reactor is continuous. The reported point is a steady window of at least 2 hours, not the integral of a pulsed stoichiometric reaction and not a sealed batch held overnight.
Methane conversion is the moles of methane consumed divided by the moles of methane fed in that window. Carbon selectivity to methanol is the moles of carbon in methanol divided by the moles of carbon in all detected carbon-containing products, both measured by calibrated gas chromatography. Methanol dissolved in a liquid phase, or trapped by an adsorbent and then released, is quantified and included. An empty reactor of the same material, volume, temperature, pressure, and flow is run the same day.
The process is catalytic. If the catalyst contains a metal, the moles of methanol in the window are at least 100 times the moles of metal in the bed. If it contains no metal, the moles of methanol are at least 100 times the moles of catalyst, using the formula weight the procedure states for the molecule or the repeat unit. A bed that states no formula weight cannot use the metal-free rule, because the turnover would not be a number. A ¹³CH₄ experiment accounts for the carbon in the methanol.
Pass threshold
Carbon selectivity to methanol is greater than 80 percent at a methane conversion of at least 5 percent, the methanol yield exceeds the empty-reactor yield measured side by side, the turnover rule is met, and the ¹³C experiment agrees.
Dummer and co-workers state that when only one oxygen atom of O₂ is incorporated and methane is the only reductant, the maximum theoretical methanol selectivity is 80 percent. Sitting the pass mark on that ceiling would grade the stoichiometric limit of the one-oxygen path. The pass mark is therefore selectivity greater than 80 percent, which is a pass mark defined here, together with the 5 percent once-through conversion those authors treat as the meaningful floor. The parity pair of about 80 percent at 5 percent remains the comparison in the sources. None of the 2024–2025 results cited above occupies both greater-than-80-percent selectivity and 5 percent conversion under the reactor rules in the specification. The turnover of 100 and the 2-hour window are pass marks defined here, so that a stoichiometric metal–oxo disguised as a catalyst, a metal-free bed with no stated formula weight, or a 40-hour batch at about 1.7 percent conversion, does not pass.
Controls
- The empty reactor is mandatory. Homogeneous gas-phase oxidation is real; Dummer and co-workers describe reproducible methanol selectivities around 40 to 60 percent at about 5 percent conversion in premixed systems, with once-through yields near 2.5 mol percent. A filled reactor that does not beat the empty tube fails.
- ¹³CH₄ labeling: the ¹³C fraction in the methanol matches the feed within the precision of the measurement.
- A methane-free oxygen run produces no methanol.
- Carbon balance over the window closes to within 10 percent relative. A missing 30 percent of the carbon fails the run even if the detected products look clean. The 10 percent closure is a pass mark for analytical honesty.
Allowed
Heterogeneous or soluble catalysts, aqueous or gas-phase operation, non-thermal plasma, and any temperature or pressure the apparatus can hold, provided the oxidant is O₂, no sacrificial reductant is added, and the empty-reactor control is run at those same conditions. A methanol-adsorption strategy is allowed if the methanol is then recovered and quantified in the same window.
Not allowed
Hydrogen peroxide, periodate, nitrous oxide, or other stoichiometric oxidants. A reductant such as NADH, hydrogen, or carbon monoxide whose job is to spare one oxygen atom. A stoichiometric metal–oxo reagent, even if it is called a catalyst. Looping unreacted methane in the conversion arithmetic so that a low once-through conversion looks like a high one. Quoting a yield from a single liquid extraction with no gas-phase analysis. Reporting a batch integral in place of the 2-hour continuous window. Pairing a selectivity measured at one contact time with a conversion measured at another.
Partial success
Selectivity to methanol is at least 80 percent at a methane conversion of at least 1 percent, in the same 2-hour continuous window, the empty-reactor and isotope controls pass, and the turnover rule is met. One percent conversion at that selectivity is a lesser pass mark. It is not a solution. The review treats conversion near 5 percent as the meaningful floor. A 40-hour batch near 1.7 percent conversion does not meet the window, so it is not partial success on these terms either.
Sources
Dummer, Willock, He, Howard, Lewis, Qi, Taylor, Xu, Bethell, Kiely, and Hutchings, Chemical Reviews, 2023, 123, 6359–6411, doi:10.1021/acs.chemrev.2c00439 (opens in a new tab). The parity pair of about 80 percent selectivity at 5 percent conversion, the competitive reading of about 75 percent at no less than 5 percent conversion, the gas-phase selectivities of about 40 to 60 percent at about 5 percent conversion, the once-through yield near 2.5 mol percent, and the 80 percent ceiling of the monooxygenase stoichiometry are taken from that review. The underlying 1995 Foulds and Gray paper was not re-read here; the parity pair is cited as Dummer and co-workers report it.
Chauhan, Rana, Gupta, Kalita, Thadhani, and Manna, Nature Communications, 2024, 15, doi:10.1038/s41467-024-54101-8 (opens in a new tab). About 90 percent overall methanol selectivity and 94 percent liquid-phase selectivity, 40-hour batch, 34 mmol of methane charged, 0.58 mmol consumed. The 1.7 percent conversion and the roughly 86 turnovers are calculated here from those reported quantities, as stated above.
Li, Shi, and co-workers, ACS Catalysis, 2025, 15, 20496–20511, doi:10.1021/acscatal.5c05829 (opens in a new tab). Continuous-flow ceria. The conversions 3.8 × 10⁻⁴ percent and 3.4 × 10⁻² percent, the 80 percent share of carbon-containing products, the roughly 85 percent selectivity, and the 3.3 and 0.9 μmol g⁻¹ h⁻¹ rates are taken from that paper.
Nallapareddy and Underwood, ACS Sustainable Chemistry & Engineering, 2024, 12, 9144–9155, doi:10.1021/acssuschemeng.4c01595 (opens in a new tab). Selectivity above 30 percent at conversion above 50 percent, and a methanol yield of 21.4 percent.
The turnover floor, the 2-hour window, the carbon-balance tolerance, and the partial-success conversion are pass marks defined here.
A rechargeable cell without lithium
A cell containing no lithium, cobalt, or nickel matches a lithium iron phosphate cell built beside it, on energy and on both efficiencies.Claim
A rechargeable cell that uses none of lithium, cobalt, or nickel stores and returns as much energy, per gram of active material, as a lithium iron phosphate cell tested in the same hardware, and it does so at least as efficiently, both in charge recovered and in energy recovered.
Why it matters
Lithium iron phosphate cells are the cost and safety benchmark that a lithium-free cell has to meet if it is going to replace them in stationary storage. Cobalt and nickel are the other constrained metals in common cathodes. A half-cell against a lithium-metal counter electrode does not answer the question, and a cathode-only watt-hour figure leaves out the anode. The grade is a full cell, weighed, next to a lithium iron phosphate cell that had to work.
Current best (with sources)
Voß and co-workers modeled large-format cells at a 90 μm cathode coating. In their hard-carbon-basic case the lithium iron phosphate pouch cell is 214 Wh kg⁻¹ and 507 Wh L⁻¹. Layered-oxide sodium-ion cells in that model span 147–206 Wh kg⁻¹ and 333–419 Wh L⁻¹. Polyanion sodium-ion cells span 158–192 Wh kg⁻¹ and 310–366 Wh L⁻¹. Prussian-blue-analogue cells span 161–185 Wh kg⁻¹ and 260–302 Wh L⁻¹. They write that several sodium cathodes, especially nickel-rich layered oxides, match or surpass lithium iron phosphate at the material level, and that the cell-level gap is largely the hard-carbon anode against graphite. Nickel-containing layered oxides are excluded by this problem even when the model puts them closest to the lithium iron phosphate pouch cell. The nickel-free classes in that model, polyanions and Prussian blue analogues, sit below it.
Those are modeled pouch and prismatic cells, not coin cells. This problem does not convert 214 Wh kg⁻¹ into a coin-cell pass mark. The grade is the cell built beside the candidate, because coin-cell watt-hours depend on loading, electrolyte fill, and formation. No absolute watt-hour target is cited.
Specification
Both cells are coin cells in identical cases, built in the same week and cycled on the same tester at 25 ± 2 °C. The control is a lithium iron phosphate cathode against a graphite anode. Its electrolyte and voltage window are those on the certificate of the lithium iron phosphate powder used. The candidate cell contains no lithium, cobalt, or nickel in the electrodes, the electrolyte, the binder, or the separator. Current collectors are aluminium, copper, or titanium. Every powder and salt is identified, and the absence of those three elements is taken from the supplier assay or from an elemental analysis of the electrodes.
Both cells are cycled at the same specific current: half the certificate specific capacity of that lithium iron phosphate, expressed as milliamperes per gram of cathode active mass, and applied to both cathodes. The candidate does not choose a gentler current by declaring a larger theoretical capacity. Each cell uses its own preregistered voltage window. The control uses the supplier’s window.
The anode mass is the mass of everything on the anode current collector that is not the current collector, weighed before assembly. A metal anode is weighed in full. Unreacted excess counts. The cathode active mass is the mass of the cathode storage material, excluding binder and conductive carbon, weighed before assembly. Those two masses are the denominator.
Cycling runs for 30 cycles. Cycles 1 through 4 are formation and are reported but not graded. Graded quantities are the averages, over cycles 5 through 30, of:
- specific energy, the discharge energy in watt-hours divided by the sum of the anode mass and the cathode active mass, as defined above
- Coulombic efficiency, discharge capacity divided by charge capacity
- round-trip energy efficiency, discharge energy divided by charge energy
Thirty cycles at this rate is about five days of cycling after assembly, so the grade fits in two weeks. The cycle count is a pass mark defined here, chosen for that reason. It is not a claim about calendar life or about a pouch cell.
Pass threshold
Over cycles 5 through 30, the candidate’s average specific energy, average Coulombic efficiency, and average round-trip energy efficiency are each at least those of the control. Over the same cycles, the average discharge capacity per geometric cathode area is also at least the control’s. There is no absolute watt-hour target. The control measured beside the candidate is the target. Thirty cycles is the whole of the graded life. Cycle life beyond that window is outside this grade, because a longer test does not fit the two-week clock.
Controls
The control cell must itself behave like a working lithium iron phosphate cell: its discharge capacity on cycle 10 is at least 80% of the certificate specific capacity, and its average Coulombic efficiency over cycles 5 through 30 is at least 99%. Eighty percent and 99% are pass marks for the control, so that a weak control cannot hand the candidate an easy comparison. If the control fails those marks, the batch is rebuilt. It is not graded.
A cell that vents, leaks, or is removed from the tester fails. Capacity and energy are taken from the tester’s calibrated current, voltage, and time, not from a manufacturer’s advertisement. The same channel calibration is used for both cells.
Allowed
Sodium, potassium, magnesium, calcium, zinc, aluminium, or organic electrode materials, including metal anodes of those elements other than lithium. Aqueous or non-aqueous electrolytes that obey the element rule. Any anode and cathode pair inside the element rule. The entire anode mass still counts.
Not allowed
Lithium salts, lithiated cathodes, cobalt, or nickel, including nickel foam, nickel tabs, and stainless-steel parts in contact with the electrolyte. Comparing a candidate half-cell that has a lithium-metal, sodium-metal, or other metal counter electrode against a full lithium iron phosphate cell, unless that metal electrode is the anode of a full cell and its whole mass is in the denominator. Quoting specific energy per gram of cathode only. Discarding cycles from the average. A primary cell. Treating formation cycles 1 through 4 as graded cycles. A cathode so thin that its areal discharge capacity falls below the control’s. Quoting cycle life beyond 30 cycles as part of this grade.
Partial success
The candidate’s average specific energy over cycles 5 through 30 is at least half the control’s, and its average Coulombic efficiency and average round-trip energy efficiency are each at least the control’s. Half the control energy is a pass mark for this lesser bar. Matching the efficiencies without matching the energy is not a solution.
Sources
Pouch-cell figures of 214 Wh kg⁻¹ and 507 Wh L⁻¹ for lithium iron phosphate, and the sodium-ion ranges quoted above: Voß, Gruber, Mitterfellner, Plöpst, Degen, Schmuch, and Lux, Energy & Environmental Science, 2025, 18, 8104–8129, doi:10.1039/D5EE00415B (opens in a new tab). No coin-cell watt-hour number is cited, because the grade is the side-by-side cell. The formation exclusion, the 30-cycle length, the shared specific current, the mass definitions, the 80% and 99% control checks, and the partial-success ratio are pass marks defined here.
Open point: the Voß model is not a 30-cycle coin cell. A thin sodium-metal anode and a nickel-free cathode could still clear the active-mass comparison over this short window even though the modeled hard-carbon pouch cells do not. If an attempt clears the three averages on the rules written here, the honest revision is a longer graded window, not a denial of the data. This file does not invent a cycle life the model did not measure.
The structure is in the spectra
A preregistered program, with nobody editing the answer, names the constitution and relative configuration of most of a blinded set from a standard NMR and mass-spectrometry packet.Claim
A packet of ordinary spectra is enough for a frozen procedure to name the constitution and the relative configuration of a small molecule that has never been published. A person may start the procedure. A person may not choose or repair the structure after seeing the packet.
Why it matters
Structure assignment sits between a routine measurement and every decision that follows it: which isomer the reaction made, which impurity is in the batch, which metabolite is in the extract. A laboratory can already record the packet in an afternoon. What it cannot yet do is trust a single, unedited output the way it trusts the mass measurement. Solving this would turn structure elucidation into a check a second laboratory can grade, including for a model.
Current best (with sources)
Published top-1 rates on structure-from-spectrum benchmarks do not clear the grade below.
Xue and co-workers report a top-1 accuracy of 92.07 percent across the input conditions of their own benchmark, and 85.10 percent when the inputs are one- and two-dimensional NMR without a molecular formula or fragments (Analytical Chemistry, 2025, doi:10.1021/acs.analchem.5c03783 (opens in a new tab)). Their abstract also says the model was used on six previously uncharacterized natural products from Magnolia officinalis and on six unexpected synthetic products. Fang and co-workers, re-evaluating named models on an experimental, scaffold-split benchmark, report a top-1 sequence accuracy for that model of 15.49 percent with the molecular formula supplied and 11.75 percent without it (arXiv:2601.15763, 2026, HTML (opens in a new tab)). Their Table 1 marks the large Mind-Gaussian set associated with the same work as non-experimental. Neither number is a relative-configuration score on unpublished molecules.
Hu, Chen, Rotskoff, Kanan, and Markland report that a multitask model finds the exact molecule 69.6 percent of the time within the first 15 predictions, on simulated proton and carbon spectra, for molecules of up to 19 heavy atoms, with no molecular formula given (ACS Central Science, 2024, doi:10.1021/acscentsci.4c01132 (opens in a new tab)). That figure is a top-15 rate on simulated spectra.
Chen and co-workers report that a diffusion model places 38 of 45 structures first when experimental chemical shifts are supplied and the two-dimensional spectra are reconstructed from those shifts together with the molecular structures (arXiv:2510.26231, PDF (opens in a new tab)). Reconstructing the two-dimensional spectra from the answer is a different task from reading a raw packet. On eleven natural products with raw experimental data, the same preprint places three structures first.
Tandem mass spectrometry is a different packet. Nowatzky and co-workers state that in silico methods in the 2016 CASMI challenge reached a recall of up to 34 percent on previously unknown compounds, and that identification rates in CASMI 2022 were below 30 percent (Nature Communications, 2025, doi:10.1038/s41467-025-57422-4 (opens in a new tab)). CASMI scores compound identification from mass spectra. It does not score relative configuration from a nuclear-magnetic-resonance packet.
An open doubt: the 92 percent author-reported figure and the 15 percent independent figure cannot both describe the same experiment. Until the spectra, the split, and the stereochemical rule are the ones in the specification below, neither figure is a solution of this problem. The Fiehn Lab results page did not load when this file was checked (a DNS error); the “below 30 percent” sentence is taken from Nowatzky and co-workers, who cite that page.
Specification
An independent laboratory prepares 20 compounds. Each has a molecular weight between 150 and 500 and at most three stereocenters. At least 10 of the 20 have at least one stereocenter, so a method that only names planar constitutions cannot fill the set. None has a published spectrum or a published structure on the day of the challenge; the laboratory records the search. Before the packet is released, a second chemist, who does not know the synthetic route, solves the structure from the packet alone and writes it down. A packet that this second chemist cannot solve is replaced. That chemist’s solution is a screen for an ambiguous packet. It is not an entry, and it is not shown to the attempt.
Each packet contains: the molecular formula from high-resolution mass spectrometry, an infrared spectrum, a proton spectrum, a proton-decoupled carbon spectrum, a COSY, an HSQC, and an HMBC. If the molecule has stereocenters, a NOESY or ROESY is included. Solvent and spectrometer frequency are stated. The packet does not contain a name, a starting material, a yield, or a drawing. The two-dimensional spectra are the spectra that were recorded, not spectra rebuilt from a candidate structure.
The attempt registers the procedure before the packets are opened: the program, the weights, and every setting that can change the output. It then returns one structure per packet within 24 hours, as the molfile or SMILES string the program writes, with relative stereo. Absolute configuration is not required. Nobody edits that file after the program stops. No further experiment is allowed. Searching a spectral database with the packet is not allowed; the compounds are unpublished so that a match would be a failure of the set, and the set would be replaced.
Grading compares the returned structure with the structure the laboratory made. Constitution is compared by the non-stereo layers of a standard InChI. Relative configuration is compared separately, at every stereocenter the packet’s NOESY or ROESY can support. A correct constitution with a wrong relative configuration does not get full credit.
Pass threshold
At least 16 of the 20 structures are correct in constitution and in relative configuration at every stereocenter the packet’s NOESY or ROESY can support. At least 8 of those 16 are among the compounds that have a stereocenter.
Sixteen of twenty, and 8 of those 16, are pass marks defined here. It is high enough that a few misses do not pass, and it asks for a single committed structure rather than a rank inside a long list. The second-chemist screen keeps the set solvable from the packet. The locked program is what keeps a person from supplying the answer the screen already showed a person can find.
Controls
The second chemist’s solution is archived before the attempt is scored and is the key, together with the synthetic characterization the laboratory already holds. The grader does not consult the attempt’s reasoning, only the structure in the answer file. The grader also checks that the answer file matches the program’s output and was not edited.
A packet whose InChI matches more than one constitution consistent with the formula is not used. Two graders score disagreements. A disagreement that is not resolved from the packet counts the structure wrong.
Allowed
Any program, model, or ranking rule that was registered before the packets were opened and that writes the answer file itself. Training on published spectra is allowed. The compounds in the challenge are unpublished.
Not allowed
A person assigning the structure, or choosing among the program’s candidates, after the packet is opened. Editing the answer file. The synthetic procedure, the reagents, or a photograph of the sample. A database lookup of the spectrum or of the molecular formula against a structure collection. Returning a list and asking to be graded on the best entry. Rebuilding a two-dimensional spectrum from a proposed structure and treating that rebuilt spectrum as data. Absolute-configuration claims are ignored, unless they contradict a relative configuration that was required.
Partial success
At least 16 of the 20 structures are constitutionally correct, with relative configuration ignored. That partial bar is the same count, on an easier question. It is not a solution. A method that gets the skeleton and misses every stereocenter has not finished the job the packet was built to ask.
Sources
The set size, the 16-of-20 marks, the molecular-weight window, and the 24-hour clock are pass marks defined here.
Cited measurements, each opened for this draft:
- Xue et al., Analytical Chemistry, 2025, 97, 22603–22614, doi:10.1021/acs.analchem.5c03783 (opens in a new tab). Author-reported top-1 rates of 92.07 percent and 85.10 percent.
- Fang et al., arXiv:2601.15763, 2026, https://arxiv.org/html/2601.15763 (opens in a new tab). Independent experimental top-1 sequence rates of 15.49 percent and 11.75 percent. The version opened has a placeholder DOI.
- Hu, Chen, Rotskoff, Kanan, and Markland, ACS Central Science, 2024, 10, 2162–2170, doi:10.1021/acscentsci.4c01132 (opens in a new tab). 69.6 percent within the first 15 predictions, on simulated spectra.
- Chen et al., arXiv:2510.26231, https://arxiv.org/pdf/2510.26231 (opens in a new tab). 38 of 45 with two-dimensional spectra reconstructed using the structures; 3 of 11 natural products from raw experimental data placed first.
- Nowatzky et al., Nature Communications, 2025, doi:10.1038/s41467-025-57422-4 (opens in a new tab). CASMI 2016 recall up to 34 percent; CASMI 2022 identification rates below 30 percent.
A molecule drawn before the assay
For most of a blinded set of proteins, one newly designed drug-like molecule binds nearly as tightly as that protein’s tool compound.Claim
For a protein disclosed with its assay, and not with the structure of its tool compound, one small molecule, designed before any new experiment, binds that protein nearly as tightly as the tool compound and is selective against a named counter-target.
Why it matters
Hit-finding today is a search: many compounds are bought or made, and the assay keeps a few. A method that can spend its one guess on a molecule the laboratory has never tested, and have that molecule land next to a real tool compound, would change the first week of a drug-discovery project. The later problems of solubility, permeability, and safety would still remain. They are asked together in a separate problem. This one asks only for binding and selectivity, on purpose, so a failure can be read.
Current best (with sources)
Prospective, experimental hit-finding has produced ligands. It has not produced one new molecule per target, made in a short registered route, matching a tool compound across a blinded set.
In CACHE Challenge #1, 23 teams each nominated up to 100 commercial compounds for the WDR domain of LRRK2, a site with no known ligand. Seventy-three of 1,955 procured compounds bound with a dissociation constant below 150 µM. After a second, analog round, seven series bound in two orthogonal assays with affinities from 18 to 140 µM (Li, Ackloo, Schapira, and co-workers, bioRxiv, 2024, doi:10.1101/2024.07.18.603797 (opens in a new tab)). Those affinities sit far weaker than the 10 nM to 1 µM tool compounds this problem uses as the reference.
The published report of CACHE Challenge #2 states that 1,957 predicted compounds were tested against the RNA site of the SARS-CoV-2 helicase Nsp13, that 0.7 percent bound in a surface-plasmon-resonance assay, and that two scaffolds reached a dissociation constant below 10 µM (Journal of Chemical Information and Modeling, 2025, doi:10.1021/acs.jcim.5c00535 (opens in a new tab)). CACHE Challenge #3, as announced by the organizers, tested 1,739 predicted molecules and called four of them chemically novel starting points (Conscience, 10 December 2024 (opens in a new tab)). CACHE Challenge #4 tested 1,688 compounds against the TKB domain of CBLB at 50 µM. One team produced a clearly novel hit with experimental confirmation. A second confirmed binder was chemically related to ligands already in the patent literature (CACHE Challenge #4 results (opens in a new tab)). Each of these challenges allows on the order of a hundred commercial compounds per team, and then a follow-up round. This problem allows one new structure and one backup, both submitted before any assay, and grades them against a tool compound run on the same plate.
Thomas and co-workers synthesized nine generated adenosine A2A ligands for one receptor. Eight of the nine bound, three of them in the nanomolar range, and the two strongest were co-crystallized (Nature Communications, 2025, doi:10.1038/s41467-025-60629-0 (opens in a new tab)). The authors state that people still filtered the designs before synthesis. The campaign is one target, nine compounds, and a receptor whose ligand chemistry is already extensively explored. It is the strongest recent single-target design result found for this draft, and it does not meet a six-of-eight blinded comparison with a tool compound.
Atz and co-workers synthesized two generated PPARγ ligands. One had an EC50 of 1.5 µM at PPARγ and 0.24 µM at PPARδ, with a calorimetric dissociation constant of 0.8 µM. The other had a selective EC50 of 2.3 µM at PPARγ. The longer routes were six and five linear steps (Nature Communications, 2024, doi:10.1038/s41467-024-47613-w (opens in a new tab)). Two molecules, one protein, and no side-by-side tool-compound window across a set.
An open doubt: CACHE #1 is still a preprint, and the public CACHE #2, #3, and #4 write-ups do not all report a uniform dissociation-constant cutoff. The pattern they share is what this threshold uses: many compounds tested, few confirmed, affinities weaker than a developed tool compound, and a second round of analogs. A later paper that clears six of eight targets on the terms below should retire or raise this problem.
Specification
The laboratory preregisters 8 protein targets. Each has a biochemical assay the grading laboratory already runs on a plate reader, by absorbance, fluorescence, or luminescence, a published tool compound, and a counter-target assay named in advance. The tool compound’s structure is not disclosed until the procedures are locked. The laboratory does disclose the target, the assay, the counter-target, and that the tool compound’s published half-maximal inhibitory concentration, or dissociation constant, in an assay of the same type lies between 10 nM and 1 µM. That window is a selection rule for the targets, so that the tool compound is a real ligand and not a fragment that anything might beat, and not a picomolar drug against which a tenfold miss would still be extraordinary for the wrong reason.
Within 24 hours of disclosure, the attempt submits one structure per target and a route of at most four linear steps from commercial materials, with a written procedure. A second structure and route may be submitted at the same time and is used only if the first route does not produce the compound. Nothing is redesigned after assay data.
The laboratory runs the procedure as written. The isolated compound has purity of at least 95 percent by quantitative proton NMR. It passes a drug-likeness filter taken from Lipinski and co-workers: no more than one of the following is violated: more than 5 hydrogen-bond donors, more than 10 hydrogen-bond acceptors, molecular weight above 500, calculated octanol–water partition coefficient above 5. The laboratory, not the attempt, checks that the compound is not a previously reported ligand of that target in ChEMBL or PubChem on the preregistration date.
The assay is a dose–response of at least five concentrations, run side by side with the tool compound and with vehicle. The counter-target assay is run at the same time.
Pass threshold
At least 6 of the 8 compounds are isolated as specified and, in the side-by-side assay, show a half-maximal concentration within tenfold of the tool compound measured in that same run, with a dose–response, an inactive vehicle, and at least tenfold weaker activity on the counter-target.
Six of eight, the tenfold windows, the four-step limit, and the 95 percent purity are pass marks defined here. The tenfold comparison is against the tool compound in the same plate, not against the published number, so that assay drift does not decide the grade.
Controls
Vehicle wells are inactive. A known inactive analog, named at preregistration for each target, is run once and is inactive. The tool compound’s half-maximal concentration in the run lies within tenfold of its published value; if it does not, the assay is repeated rather than scored. Tenfold of the published value is a pass mark for assay validity.
The independent route execution is the synthesis control. A procedure that does not produce the drawn compound fails that target.
Allowed
Any design method. Any reaction in the written route. One backup compound submitted before any assay.
Not allowed
A compound that is already a reported ligand of the target. A peptide, an oligonucleotide, or a structure that violates the Lipinski filter as defined above. Screening a library, or making a third compound after seeing the assay. Submitting a list and asking to be graded on the best member. Changing the route after a failed step. Quoting a published potency for the tool compound instead of the side-by-side measurement. Using a tool compound outside the 10 nM to 1 µM selection window. A structure that differs from the tool compound, or from any reported ligand of that target, only by adding, removing, or replacing one methyl, fluoro, or hydroxy group. The tool structure is compared after the procedures are locked. A match of that kind fails the target.
Partial success
At least 4 of the 8 compounds show a dose–response, beat the vehicle, and land within a hundredfold of the side-by-side tool compound, with purity and the drug-likeness filter satisfied. A hundredfold and 4 of 8 are pass marks for this lesser bar. Partial success is not a solution.
Sources
Lipinski, Lombardo, Dominy, and Feeney, Advanced Drug Delivery Reviews, 1997, 23, 3–25, doi:10.1016/S0169-409X(96)00423-1 (opens in a new tab). The abstract states that poor absorption or permeation is more likely when there are more than 5 hydrogen-bond donors, more than 10 hydrogen-bond acceptors, the molecular weight is greater than 500, and the calculated log P is greater than 5. This problem uses the common reading that at most one of those four is violated. That reading is a pass rule layered on the paper, and it is open to revision. Purity uses the quantitative proton NMR assay of Pauli and co-workers, Journal of Medicinal Chemistry, 2014, doi:10.1021/jm500734a (opens in a new tab).
The 10 nM to 1 µM window, the tenfold and hundredfold bars, and the 6-of-8 count are pass marks defined here.
Cited prospective tests, each opened for this draft:
- Li et al., bioRxiv, 2024, doi:10.1101/2024.07.18.603797 (opens in a new tab). CACHE #1: 73 of 1,955 below 150 µM; seven series from 18 to 140 µM.
- CACHE Challenge #2, Journal of Chemical Information and Modeling, 2025, doi:10.1021/acs.jcim.5c00535 (opens in a new tab). 0.7 percent of 1,957 bound; two scaffolds below 10 µM.
- CACHE Challenge #3 announcement (opens in a new tab), 10 December 2024. Four chemically novel hits from 1,739 molecules.
- CACHE Challenge #4 results (opens in a new tab). 1,688 compounds; one clearly novel confirmed hit.
- Thomas et al., Nature Communications, 2025, doi:10.1038/s41467-025-60629-0 (opens in a new tab). Eight of nine A2A compounds bound, on one target.
- Atz et al., Nature Communications, 2024, doi:10.1038/s41467-024-47613-w (opens in a new tab). Two PPARγ designs, EC50 values 1.5 µM and 2.3 µM.
As strong as the PET beside it, and it unzips
A polymer matches a PET control in a standard tensile test and returns its monomer under a catalytic cap, at milder heat than a cited PET glycolysis.Claim
One polymer is strong in the way a bottle-grade polyester is strong, and it comes apart to its monomers under heat milder than a documented glycolysis of poly(ethylene terephthalate), with a base and catalyst loading that stays catalytic, then goes back together, three times, inside two weeks.
Why it matters
Poly(ethylene terephthalate) is strong enough for bottles and fiber, and it can be glycolyzed back to monomer, but the documented conditions are hot. A weaker polyester that unzips in warm methanol does not replace it. The joint grade is the tensile test against a polyester molded the same day, and a closed chemical loop that does not spend a stoichiometric base to get there.
Current best (with sources)
Cao and co-workers ran catalytic glycolysis of poly(ethylene terephthalate) at 180 °C for 1 hour and reported a bis(2-hydroxyethyl) terephthalate yield approaching 95.5%. That is a real depolymerization of the incumbent polymer. It is above the temperature cap of this problem, so a process that needs 180 °C does not pass, whatever its yield.
Han, Zhang, Zhou, Zhao, Odelius, and Hakkarainen report a triarylmethane polyester, P3Ad, with a maximum tensile stress of 51.66 ± 0.91 MPa and an elongation at break of 167.91 ± 16.39% on injection-molded bars, which they describe as similar to poly(ethylene terephthalate) by comparison with published values. They depolymerized 20 g in refluxing methanol, about 66 °C, in 15 minutes, with 10 wt% potassium hydroxide. Isolated dimethyl adipate was 90%. The triarylmethane monomer was 98% as a crude yield and 88% after recrystallization. One repolymerization with titanium isopropoxide gave a polymer they describe as mechanically indistinguishable from the original. Six mechanical reprocessing cycles retained tensile properties. That is the nearest chemical loop found for this draft.
It does not pass this problem. Ten weight percent potassium hydroxide is 1.78 mmol per gram of polymer. That loading equals 5 mol% only for a repeat unit of about 28 g mol⁻¹. Polyester repeat units are heavier, so the loading is a larger mole fraction than 5 mol%. The tensile numbers were not measured against a poly(ethylene terephthalate) bar molded and pulled the same day on the same machine. The chemical loop was shown once, not three times. The comparison plot against commercial plastics is not a substitute for the side-by-side test.
Specification
Tensile specimens of the candidate and of an unfilled bottle-grade poly(ethylene terephthalate) control are molded to the same geometry and tested on the same machine, the same day, by ASTM D638. At least five specimens of each are tested. The graded values are the median ultimate tensile strength and the median elongation at break.
The candidate is then depolymerized at a temperature of at most 120 °C, within 12 hours, to at most two defined monomers. Hydroxide, alkoxide, and any other catalyst, taken together, are at or below 5 mol% relative to repeat units charged. Each monomer is isolated. The isolated mass of each is at least 90% of the mass required by the repeat-unit stoichiometry of the polymer charged. Each isolated monomer is at least 95% pure by quantitative proton NMR, using an internal standard as described by Pauli and co-workers. The monomers are repolymerized by a written procedure to the same polymer, molded, and tested again. This loop is performed three times. After the third repolymerization, the medians are compared with the candidate’s own first medians, not with the polyester control.
The clock starts when the first tensile specimens are molded and ends when the third recycled set is tested. It is at most 14 days. A procedure that cannot finish three chemical loops in that window fails even if one loop looked perfect.
The candidate may be poly(ethylene terephthalate) itself, or another polymer. The joint requirement is the point. A material that is strong and does not unzip, or that unzips and is weak, fails.
Pass threshold
The candidate’s first median tensile strength and first median elongation at break are each at least 80% of the poly(ethylene terephthalate) control’s medians. Each cycle returns at least 90% of each monomer at the purity above, at no more than 120 °C, within 12 hours, at a combined base and catalyst loading of at most 5 mol%. After three cycles, both medians remain at least 80% of the candidate’s first medians. The 14-day clock holds.
Eighty percent, 90 percent, 95 percent purity, 5 mol%, 120 °C, 12 hours, three cycles, and 14 days are pass marks defined here. The temperature cap is set below the 180 °C glycolysis cited above, not offered as a record of the mildest depolymerization anyone has run.
Controls
The control resin is identified by its grade and by a proton spectrum consistent with poly(ethylene terephthalate). Specimens of both materials are the same thickness. ASTM D638 warns that specimen preparation, speed, and environment change the comparison, which is why the control is molded and tested beside the candidate. Five specimens and the median, rather than the best specimen, are the pass rule defined here.
A depolymerization with no catalyst and no base, if the procedure claims one, is run once on a small charge. It must not itself reach the 90% yield in 12 hours at the same temperature. If it does, the reaction is not the one the procedure describes, and the run is invalid.
Carbon recovered in the isolated monomers, plus identified side products, accounts for at least 90% of the carbon charged. That closure is a pass mark for the mass balance.
Allowed
Any backbone. Up to two constitutional monomers, as in a diacid and a diol. A catalyst and a base that together stay at or below 5 mol% relative to repeat units. Enzymes, if they meet the 12-hour, yield, and loading rules. Methanol or another solvent, provided the solvent is not counted as a monomer and the 5 mol% cap is still met.
Not allowed
Fillers or fibers that the polyester control does not also contain. Reporting tensile strength of a composite against an unfilled control. Calling an oligomer mixture a monomer. A depolymerization above 120 °C. A stoichiometric hydroxide, including the 10 wt% potassium hydroxide procedure cited above. A cleavage reagent above 5 mol% relative to repeat units, including neat acetic acid or another carboxylic acid used as the solvent. Quoting a literature tensile strength for poly(ethylene terephthalate) instead of the side-by-side test. One successful unzipping with no repolymerization. Mechanical reprocessing counted as a chemical cycle.
Partial success
The first tensile medians meet the 80% comparison, and one cycle meets the monomer yield, purity, temperature, time, and 5 mol% loading. The later cycles may fall short. One cycle is not a solution. A polymer that unzips and is not remade has not closed the loop. P3Ad, as reported, does not meet this lesser bar either, because of the base loading and the missing side-by-side polyester control.
Sources
ASTM D638-22, the standard test method for tensile properties of plastics, astm.org/d0638-22.html (opens in a new tab). The method says tensile properties vary with specimen preparation, speed, and environment, and that comparative tests have to control those factors. Cao, Liang, Yang, Zhu, Deng, Li, Elimelech, and Lu, Nature Communications, 2024, 15, 6266, doi:10.1038/s41467-024-50702-5 (opens in a new tab), for glycolysis at 180 °C and the yield approaching 95.5%. Han, Zhang, Zhou, Zhao, Odelius, and Hakkarainen, Angewandte Chemie, 2025, doi:10.1002/ange.202421431 (opens in a new tab), for the P3Ad tensile figures, the 10 wt% potassium hydroxide methanolysis at about 66 °C, the monomer yields, and the single chemical repolymerization. Monomer purity uses the quantitative proton NMR assay of Pauli and co-workers, Journal of Medicinal Chemistry, 2014, doi:10.1021/jm500734a (opens in a new tab). The conversion of 10 wt% KOH into 1.78 mmol g⁻¹, and the statement that this equals 5 mol% only near 28 g mol⁻¹, is arithmetic on that loading. The 80%, 90%, 5 mol%, 120 °C, 12-hour, three-cycle, and 14-day requirements are pass marks defined here.
Amide bonds without a coupling reagent
One catalyst, at or below 5 mol percent, couples a blinded set of acids and amines in isolated yields that match side-by-side HATU reactions, and the stoichiometric byproduct is water.Claim
A carboxylic acid and an amine become an amide in the hands of one catalyst, with water as the only stoichiometric byproduct, and the isolated yield matches a coupling-reagent reaction run beside it. The same written procedure is used for an unhindered primary amine, a secondary amine, an electron-poor aniline, and a hindered acid.
Why it matters
Amide formation is one of the reactions a medicinal-chemistry project runs most often, and it is still usually done with a stoichiometric coupling reagent. That reagent is the waste, the cost, and a common cause of epimerization. A catalyst that matches it on the pairs people actually submit, rather than on the pairs a scope table was built around, is the result that would retire the reagent.
Current best
Borate and boronic-acid catalysts already make many amides with water as the byproduct. They do not yet do it as one locked procedure across the four classes below.
Procter, Alamillo-Ferrer, Shabbir, Britton, Bučar, Dumon, Rzepa, Burés, Whiting, and Sheppard reported a fluorinated borate, B(OArF)₃, for difficult amides (doi:10.1039/D4SC07744J (opens in a new tab)). On 3-fluorophenylacetic acid with 2-aminopyridine, that borate gave 43 percent yield in an 18-hour screen. Extending the same reaction to 66 hours gave amide 4 in 57 percent isolated yield. Changing the solvent to tert-butyl acetate raised the isolated yields of amides 3 and 4 above 80 percent in 20 hours. The scope that follows is not one recipe: the footnotes use 2 mol percent or 5 mol percent of the isolated borate, or 10 to 15 mol percent of borane–dimethyl sulfide to generate it, in tert-butyl acetate or in toluene, for 20 hours, 24 hours, 40 to 45 hours, or 72 hours. Electron-poor anilines and hindered secondary anilines are reported to react, with toluene chosen for some of them because of side reactions in tert-butyl acetate. 1-Adamantylamine showed low reactivity. Aromatic carboxylic acids, including heterocyclic ones, “proved less reactive, giving low yields even after prolonged reaction times.” A multigram run of one amide gave 59 percent yield at 5 mol percent catalyst, and another gave 71 percent. In the same screen, 2-chlorophenylboronic acid gave no amide from the aminopyridine pair, and a DATB catalyst gave 20 percent. On the easier pair of benzoic acid and benzylamine, 10 mol percent of 2-chlorophenylboronic acid gave 79 percent.
That is a real advance on coordinating and electron-poor amines, bought with a change of solvent, loading, and time, and with aromatic acids left at low yield. It is not one procedure, frozen before the pairs are known, matched against HATU on a blinded set.
Specification
The grading laboratory already runs amide couplings. A laboratory that would be learning the workup during the test does not grade this problem.
The laboratory preregisters 12 acid–amine pairs. At least three pairs use a primary aliphatic amine, three a secondary amine, three an aniline or other electron-poor amine, and three a hindered acid, either α,α-disubstituted or an ortho-substituted benzoic acid. The pairs are disclosed together. No pair is one of the example numbers in the 2025 Sheppard scope.
Within 24 hours, and with no experiment on the test pairs, the attempt submits one catalyst system and one procedure: identity, loading at or below 5 mol percent, solvent, temperature, time, and workup. The procedure may name a drying agent. It may not name a stoichiometric activator. One backup procedure may be submitted at the same time. The backup, if it is used, is applied to every pair or to none. A footnote that changes the solvent for the anilines and the loading for the hindered acids is a second procedure, and it is not allowed unless it was the single backup and is then used for the whole set.
The laboratory runs that procedure, and runs a HATU comparison on the same pair, same scale, same day. The HATU conditions are fixed so they cannot be weakened: 1.1 equivalents of HATU, 2 equivalents of N,N-diisopropylethylamine, N,N-dimethylformamide, 0 °C to room temperature, 12 hours, then the same chromatography conditions used for the catalytic product. Both products are isolated. The catalytic product is at least 95 percent pure by quantitative proton NMR. Yields are isolated masses of amide divided by theoretical mass.
Pass threshold
On at least 10 of the 12 pairs, the isolated yield of the catalytic amide is at least the isolated yield of the matched HATU reaction and at least 40 percent, at the purity above, and no stoichiometric byproduct other than water is formed. The 10 include at least 2 pairs from each of the four classes.
Ten of twelve, the class floors, 5 mol percent, 40 percent, and 95 percent purity are pass marks defined here. The comparison with HATU is the side-by-side experiment, not a universal percentage copied from a paper. Forty percent is a floor so that matching a failed HATU reaction does not count. It sits below the yields above 80 percent, and below the 59 and 71 percent multigram yields, that Procter and co-workers report for individual amides after per-substrate adjustment. A HATU reaction that was run badly does not help: if the HATU yield on an unhindered primary amine is below 70 percent, that pair is repeated. Seventy percent is a pass mark for the control on that subclass only.
Controls
Each pair has a no-catalyst reaction, same solvent, temperature, and time. It must not reach the catalytic yield. If the uncatalyzed reaction matches the catalyzed one, that pair does not count as catalysis.
The HATU product and the catalytic product must be the same amide by NMR and high-resolution mass spectrometry. A rearranged product fails even if the mass is high. Where the acid or the amine has a stereocenter that can epimerize, the enantiomer or diastereomer ratio of the catalytic amide must be at least that of the HATU product run beside it. A more epimerized catalytic product fails that pair.
Water is the stoichiometric byproduct required by the procedure’s equation. A hidden activator that is consumed one-to-one with the acid fails the pair. The laboratory checks the spent reaction for the coupling reagents named in “Not allowed” if the procedure’s mass balance suggests one was present.
Allowed
Any catalyst, including an enzyme, within the loading rule. A drying agent that is not a coupling reagent. Molecular sieves. One backup procedure submitted before any experiment on the test pairs, used for every pair or for none.
Not allowed
Carbodiimides, uronium salts, phosphonium salts, and other stoichiometric coupling reagents. Preforming an acid chloride or a mixed anhydride with a stoichiometric activator and then calling the acylation step catalytic. More than 5 mol percent of catalyst, including a borane precursor counted as the boron it delivers. A different solvent, temperature, or loading for one class unless that change is the single backup procedure applied to every pair. Editing conditions after the first yield is known. Choosing the twelve pairs from a published scope table of the same catalyst.
Partial success
At least 8 of the 12 pairs meet the yield, purity, and HATU comparison, every failure is one of the hindered acids, and each of the other three classes has at least two passes. Eight of twelve, confined to that class, is the lesser bar. It shows a general unhindered amidation and an unsolved hindered one. It is not a solution.
Sources
Yields are judged against the side-by-side HATU experiment rather than against a published yield. The 5 mol percent ceiling, the 10-of-12 count, the class floors, the 40 percent absolute floor, the 70 percent floor on unhindered HATU controls, and the partial-success rule are pass marks defined here.
The 2025 scope is Procter and co-workers, Chemical Science, 2025, 16, 4718–4724, doi:10.1039/D4SC07744J (opens in a new tab). The 43 percent screen yield, the 57 percent isolated yield at 66 hours, the yields above 80 percent after the solvent change, the low yields on aromatic acids, the low reactivity of 1-adamantylamine, the 59 and 71 percent multigram yields, the 0 percent and 20 percent comparison catalysts, and the 79 percent yield with 10 mol percent 2-chlorophenylboronic acid are taken from that paper. The variation in loading, solvent, and time is taken from the footnotes of its scope scheme.
Purity uses the quantitative proton NMR assay of Pauli and co-workers, Journal of Medicinal Chemistry, 2014, 57, 9220–9231, doi:10.1021/jm500734a (opens in a new tab).
The named anomer of a glycoside
A procedure written in advance returns the specified anomer for most of a blinded set of sugar couplings, including 1,2-cis cases with no participating ester.Claim
Given a glycosyl donor supplied as a hemiacetal and an alcohol, and a drawing that names one anomer of the glycoside, a procedure written before any experiment on that pair returns that anomer in quantity. The set includes 1,2-cis glycosides that have no participating group at C2.
Why it matters
The biological activity of an oligosaccharide often sits in which anomer was formed. Neighboring-group participation makes many 1,2-trans linkages reliable and does not make the 1,2-cis linkage. A method that returns whichever anomer the donor and the acceptor happen to prefer still leaves the cis bond as a separate research project. A procedure that hits the named anomer, on pairs it did not choose, is the result that would make oligosaccharide synthesis a planned route rather than a screen.
Current best
Park, Harper, Kuhl, Kwan, Liu, and Jacobsen reported a macrocyclic bis-thiourea that catalyzes stereospecific substitution of glycosyl chlorides (doi:10.1126/science.aal1875 (opens in a new tab)). On the model pair, α-mannosyl chloride and benzyl alcohol, catalyst 6 gave 88 percent yield and an α:β ratio of 8:92. The product configuration followed the configuration of the glycosyl chloride: inversion, for both α- and β-chlorides. The paper describes β(1,6) and β(1,3) linkages across several sugars as good to excellent, and states that the more demanding β(1,4) linkage could be produced only when the steric demand of the nucleophile was reduced. The donors in that study were glycosyl chlorides of a chosen anomer, not hemiacetals, and the anomer of the product was the one inversion provides. An 8:92 ratio is about 12 to 1. The main text does not attach a higher numerical ratio to a general pair, and none is invented here.
Lourenço and Ventura, reviewing 1,2-cis glycosylation in 2024, write that construction of the 1,2-cis bond is still a major challenge, that the reactions remain highly dependent on the donor–acceptor pair, and that the donor and the conditions are tuned for the case in hand (doi:10.1039/bk9781837672844-00023 (opens in a new tab)). That is the opposite of one procedure, written before the pairs are known, that returns either named anomer from a hemiacetal.
docs/SELECTION.mddescribes a Nature Synthesis paper (s44160-025-00846-z (opens in a new tab)) as reporting electrophilic-bromine activation with inversion across a wide set of monosaccharides, and as calling a general method for every linkage unrealized. This problem does not treat that description as a pass. The donor here is the hemiacetal as drawn. The anomer is the one named in advance, including the 1,2-cis pairs with no participating group. The procedure is locked before the pairs are known. Copying a published inversion of a preformed halide for that exact pair fails. No yield or ratio is taken from that paper here, because this draft did not re-open its tables.Specification
The laboratory preregisters 10 donor–acceptor pairs. Each donor is supplied as a hemiacetal. An anomeric mixture is acceptable. The protecting groups are already in the drawing and are not redesigned after disclosure. Each acceptor is a specified alcohol. The product drawing names one anomer, α or β.
At least four pairs are 1,2-cis and have a non-participating substituent at C2: an ether, a benzyl ether, an azide, or a hydrogen, and not an ester, a carbonate, or an amide. At least three pairs are 1,2-trans. The donors include at least three of the following: glucose, galactose, mannose, a glucosamine, and a 2-deoxy sugar. At least three acceptors are secondary alcohols. The pairs are disclosed together.
Within 24 hours, and with no experiment on the test pairs, the attempt submits one procedure per pair, or one general procedure: reagents, catalyst and loading, solvent, temperature, time, and workup. One backup procedure per pair may be submitted at the same time. The laboratory runs the procedure once, on 0.1 to 0.2 mmol of donor.
The anomeric ratio is the proton-NMR integral of the named anomer divided by the integral of the other anomer, measured on the crude mixture before chromatography, against a weighed internal calibrant. The named anomer is then isolated. Its purity is at least 95 percent by absolute quantitative proton NMR. Identity includes a carbon spectrum and a correlation that places the new glycosidic bond. A chemist who did not write the procedure confirms that the isolated material is the named anomer.
Pass threshold
At least 8 of the 10 pairs, including at least 3 of the 4 designated 1,2-cis pairs, give the named anomer in an isolated yield of at least 40 percent, with a crude anomeric ratio of at least 19 to 1, at the purity above.
Eight of ten, 3 of 4, 40 percent, 19 to 1, and 95 percent purity are pass marks defined here. Nineteen to one is set above the 8:92 ratio, about 12 to 1, reported for the Jacobsen model pair, and it is not a ratio copied from a best example whose number was not in the text opened here. Forty percent is a floor for an isolable sample. The cis quota, the hemiacetal starting point, and the ban on reprotecting are what keep a stereospecific substitution of a preformed chloride from being scored as a general solution.
Controls
For two pairs chosen before the procedures are opened, one of them a 1,2-cis pair, the laboratory runs the reaction with no catalyst and no promoter, otherwise as written. That blank must not itself reach a 19-to-1 ratio in favor of the named anomer. If it does, the donor or the conditions already determine the anomer and the pair is replaced before grading.
The crude spectrum must show both anomeric regions. A single peak is credited only if an independently prepared sample of the other anomer, or a deliberately anomerized sample, resolves from it on the same method and a spike shows which peak is which. Chromatography that enriches the named anomer does not change the crude ratio, which is the number that counts.
Allowed
Any catalyst, promoter, or leaving-group activation that the written procedure names, including conversion of the hemiacetal into a glycosyl halide or imidate inside that procedure. Enzymes. Computation and literature lookup before the procedures are locked. The one backup procedure submitted with the first procedure.
Not allowed
Redesigning the protecting groups after the pairs are disclosed. Supplying, or buying, a single anomer of a glycosyl chloride or imidate that was not in the preregistered drawing and calling the subsequent inversion a prediction. Separating anomers and discarding the wrong one without the crude ratio meeting the bar. A participating C2 ester added to a pair the laboratory marked as non-participating. Screening promoters on the test pair, or on a close analog made for the purpose, after disclosure. Counting a 1,2-trans outcome on a participating donor toward the cis quota.
Partial success
At least 6 of the 10 pairs reach a crude anomeric ratio of at least 10 to 1 and an isolated yield of at least 30 percent, including at least 2 of the 4 cis pairs, and the blanks pass. These figures are pass marks for the lesser bar. Partial success is not a solution.
Sources
The counts, the 40 percent and 30 percent yields, the 19-to-1 and 10-to-1 ratios, and the cis quota are pass marks defined here.
Cited measurements:
- Park, Harper, Kuhl, Kwan, Liu, and Jacobsen, Science, 2017, 355, 162–166, doi:10.1126/science.aal1875 (opens in a new tab). Catalyst 6, α-mannosyl chloride, benzyl alcohol: 88 percent yield, α:β = 8:92. Stereospecific inversion. The β(1,4) linkage required a less hindered nucleophile. Broader examples are described as good to excellent without a second numerical ratio being taken from the main text.
- Lourenço and Ventura, Carbohydrate Chemistry, 2024, pp. 23–43, doi:10.1039/bk9781837672844-00023 (opens in a new tab). The 1,2-cis linkage is described as still dependent on the donor–acceptor pair and on conditions tuned for that pair. No volume number is cited because the record used here does not give one.
Purity uses the absolute quantitative proton NMR assay of Pauli and co-workers, Journal of Medicinal Chemistry, 2014, 57, 9220–9231, doi:10.1021/jm500734a (opens in a new tab).
Candidates
These specifications are public and are not on the v1 list. A candidate moves onto the twelve only when docs/SELECTION.md is updated in the same change.
A superconductor you can leave on the bench
A material kept at ambient pressure shows both zero resistance and a Meissner effect at or above standard room temperature.Claim
A substance prepared in air, and measured in air at ambient pressure, is a superconductor at or above standard room temperature: its electrical resistance collapses and it expels magnetic field, on two separately made samples.
This file is not on the v1 list. Grading the Meissner effect needs a magnetometer, which the criteria treat as outside a standard laboratory. The claim stays, as a candidate, until that rule changes.
Why it matters
A superconductor that did not need liquid nitrogen or a diamond-anvil cell would change magnets, transmission, and sensors. Resistance that falls on cooling is not enough, and neither is a pellet that sits over a permanent magnet. The two signatures have to appear on the same specimen, at ambient pressure, after the specimen has been left in air, and a second batch has to do it again.
Current best (with sources)
Wen and co-workers describe HgBa₂Ca₂Cu₃O₈₊δ as the cuprate with the highest critical temperature at ambient pressure, and they report crystals with T_c ≈ 134 K. The same paper states that the critical temperature of that compound reaches 164 K under pressure. Pressure is outside this problem. Deng, Chu, and co-workers report a pressure-quench protocol that leaves HgBa₂Ca₂Cu₃O₈₊δ superconducting at 151 K once the pressure is released. One hundred and fifty-one kelvin is still far below 298 K, and this problem additionally requires 24 hours in air and both signatures on two batches. Neither report is used as a pass mark.
Wu and co-workers observed both a resistive and a magnetic transition in the yttrium–barium–copper oxide system between 80 and 93 K at ambient pressure. That material is the instrument control, not a record to beat on temperature. It cannot itself pass this problem. This problem does not cite a room-temperature critical temperature, because it is not using one.
Specification
Standard room temperature here means 298 K, the conventional standard temperature. Ambient pressure means 1.00 ± 0.05 bar of air. The sample sits in laboratory air at that pressure for 24 hours before the measurement, and the measurement itself is at that pressure. A diamond-anvil cell, or any measurement above 1.05 bar, does not count.
Two signatures are required on the same specimen, then repeated on a second specimen made in a separate batch.
- Four-point resistance, with the current reversed and the two voltage readings averaged, so that a thermal voltage is not mistaken for a resistance. The resistance at 298 K, divided by the resistance of the same specimen at 30 K above the claimed onset or at 328 K, whichever is higher, is at least as small as the same ratio measured the same day on a YBa₂Cu₃O₇₋δ pellet below its own transition. The denominator is therefore always taken at least 30 K above 298 K.
- Direct-current magnetometry in the same applied field used for the YBa₂Cu₃O₇₋δ pellet that day. The field is one at which that pellet shows a clear diamagnetic transition, and it is written down. The graded moment is the zero-field-cooled diamagnetic moment at 298 K, per geometric volume of the specimen. It is at least the zero-field-cooled diamagnetic moment per geometric volume of the pellet below its transition.
The geometric volume is the measured outer volume of the specimen. Composition is reported by powder diffraction and by elemental analysis. The laboratory states what the sample is, not only that a pellet was pressed.
Pass threshold
Both signatures pass on both batches at 298 K and at a pressure of 1 bar, after 24 hours in air. The copper control below shows neither signature.
The temperature, the pressure, the 24-hour wait, the two-batch rule, the current reversal, and the 30 K offset used to form the resistance ratio are pass marks defined here. YBa₂Cu₃O₇₋δ is an instrument control. The 134 K and 151 K reports are the ambient-pressure context. They are not targets to match.
Controls
- The YBa₂Cu₃O₇₋δ pellet, measured the same day on the same instruments and in the same applied field, shows both a resistance drop and a zero-field-cooled diamagnetic transition. If it does not, the instruments are not trusted and the candidate is not graded.
- A copper specimen of similar shape shows no resistance drop and no diamagnetic transition on the same sweep.
- The candidate’s diamagnetic signal must be present at 298 K on cooling from above the claimed onset and absent above that onset. A signal that is present at all temperatures is a magnetic impurity, not a Meissner effect. Field-cooled and zero-field-cooled curves are both reported. The graded curve is the zero-field-cooled curve.
- Both batches pass. One pearl of a pellet is not enough.
Allowed
Any composition the elemental analysis can describe, including organics and non-copper oxides. Ambient-pressure synthesis of any kind. A commercial magnetometer and a four-point probe.
Not allowed
Reporting resistance without magnetometry, or magnetometry without resistance. A transition that exists only under pressure, or only after a pressure quench if the released specimen does not then meet the 298 K, 24-hour air, and two-batch rules. A sample that must be kept in an inert atmosphere or that decomposes in the 24-hour wait. A film so thin that the volume used in the magnetization ratio is chosen after the fact. Levitation of a pellet over a magnet, by itself. A resistance taken at one current direction only.
Partial success
Both signatures are present on both batches at or above 273 K, the resistance ratio is at least 1,000 relative to the same specimen above the onset, and the zero-field-cooled diamagnetic moment per volume is at least ten times the copper blank and is temperature-dependent. Matching the yttrium–barium–copper oxide volume susceptibility is not required for the partial bar. A resistance drop without a diamagnetic transition is not a partial success. The 273 K floor, the factor of 1,000, and the factor of ten are pass marks for this lesser bar. Two hundred and seventy-three kelvin is still above the 134 K and 151 K reports cited here.
Sources
Wu, Ashburn, Torng, Hor, Meng, Gao, Huang, Wang, and Chu, Physical Review Letters, 1987, 58, 908, doi:10.1103/PhysRevLett.58.908 (opens in a new tab), for the resistive and magnetic transition between 80 and 93 K at ambient pressure. Wen and co-workers, npj Quantum Materials, 2025, doi:10.1038/s41535-025-00735-w (opens in a new tab), for HgBa₂Ca₂Cu₃O₈₊δ at T_c ≈ 134 K as the highest ambient-pressure cuprate in that paper, and for the statement that the critical temperature reaches 164 K under pressure. Deng, Chu, and co-workers, Proceedings of the National Academy of Sciences, 2026, doi:10.1073/pnas.2536178123 (opens in a new tab), for the reported 151 K ambient-pressure transition after a pressure quench. The 298 K requirement, the pressure window, the 24-hour wait, and the partial-success factors are pass marks defined here.
The crystal form named in advance
For most of a blinded set of flexible organic molecules, one submitted crystal structure matches the withheld single-crystal X-ray structure.Claim
Given the diagram of a new organic molecule and the conditions under which a crystal of it was grown, one predicted crystal structure matches the single-crystal X-ray structure that has been withheld.
Why it matters
The crystal form decides solubility, stability, and whether a tablet or a formulation can be made twice. Companies still find late polymorphs after years of screening. A method that names the form actually grown, before it sees the diffraction, would let a solid-state group decide which experiment to run and would give a sharp test of energy models that today return long lists.
Current best (with sources)
The standing blinded comparison is the seventh crystal-structure-prediction blind test, run by the Cambridge Crystallographic Data Centre from 2020 to 2022 and published in 2024.
In the structure-generation half, a match was a 30-molecule overlay, hydrogens omitted, distance and angle tolerances of 35 percent, and a root-mean-square deviation below 1.0 Å. Lists ran to about 1,500 structures. For the agrochemical XXXI, which has three rotatable bonds, 10 of 17 groups generated form A and 9 of 17 generated form B somewhere in the list. For the drug candidate XXXII, which has eleven rotatable bonds, 3 of 13 groups generated form A and 2 of 13 generated form B (Hunnisett and co-workers, Acta Crystallographica B, 2024, 80, 517–547, doi:10.1107/S2052520624007492 (opens in a new tab)).
In the ranking half, periodic dispersion-corrected density-functional methods were the most consistent. For XXXII they agreed with each other and placed the experimental forms more than 4 kJ/mol above the predicted global minimum, at low temperature and at ambient temperature (Hunnisett and co-workers, Acta Crystallographica B, 2024, 80, 548–574, doi:10.1107/S2052520624008679 (opens in a new tab)). Finding a packing inside a long list, and naming that packing as the form to expect, are different outcomes. The second paper’s result on XXXII is that the experimental forms were not ranked first.
An open doubt: target XXVIII in the same test was published by an outside group while the test was running, so that target is not evidence about a blind prediction. This problem uses the generation paper’s overlay rule and asks for one structure, on molecules flexible enough that the seventh test did not rank the experimental form first.
Specification
The laboratory holds six unpublished single-crystal structures. Each molecule contains only C, H, N, O, S, F, and Cl, has a molecular weight between 250 and 550, has at least eight rotatable bonds, and has one molecule in the asymmetric unit. The structure has no unresolved disorder. On the morning the diagrams are released, none of the six is in the Cambridge Structural Database; the laboratory records the search. A molecule that fails one of these checks is replaced before release.
The laboratory releases, for each molecule, the constitutional diagram with relative stereochemistry, the crystallization solvent, the temperature, and the method (evaporation, cooling, or vapor diffusion). It does not release cell parameters, a powder pattern, a photograph, or a space group. The attempt has seven days to submit one crystallographic information file per molecule: the single structure it expects under those conditions. A list is not a submission for the full grade.
Grading uses the withheld single-crystal structure. The structures match when a 30-molecule overlay, hydrogens omitted, with distance and angle tolerances of 35 percent, fits all 30 molecules with a root-mean-square deviation below 1.0 Å. That match rule is the one used in the seventh blind test, adopted here so two laboratories will call the same pair a match. The laboratory archives the overlay.
Pass threshold
At least 4 of the 6 submitted structures match the withheld experimental structures under the overlay rule above.
Four of six, the eight-rotatable-bond minimum, the seven-day clock, and the decision to submit one file are pass marks defined here. The 1.0 Å and 30-molecule criteria are taken from the generation paper cited above. Eight rotatable bonds is a selection rule aimed at the unsolved end of that test: three rotatable bonds were often generated somewhere in a long list, and eleven were not placed first by the ranking methods.
Controls
Before release, a second crystallographer overlays the experimental structure on itself with the same rule and confirms a match. A structure that fails this self-match is replaced.
The attempt does not receive the diffraction frames, the solved structure, or a powder pattern until after the six files are archived. The grader compares files, not narratives. If the overlay program returns a match and a visual check shows a different packing, the visual check wins and the target fails. The disagreement is archived.
Allowed
Any energy model, search, or machine-learning potential. Any amount of computation during the seven days. Use of the Cambridge Structural Database for other molecules. Prediction of disorder is allowed only if the major component still meets the overlay rule; the targets themselves are ordered.
Not allowed
A powder pattern, a cell, or a space group of the target. A list of structures from which the closest is chosen after grading. A submission filed after any diffraction of that batch has been shown to the attempt. A molecule already in the Cambridge Structural Database. Changing the submitted file after the deadline. Counting a match to a different polymorph from the one grown under the stated conditions.
Partial success
At least 4 of the 6 experimental structures appear in a list of at most 50 structures submitted, under the same deadline and the same overlay rule. Fifty and 4 of 6 are pass marks for this lesser bar. Partial success shows that the packing was generated. It is not a solution, because the form was not named.
Sources
The overlay rule (30 molecules, 35 percent tolerances, root-mean-square deviation below 1.0 Å) is taken from Hunnisett et al., Acta Crystallographica Section B, 2024, 80, 517–547, doi:10.1107/S2052520624007492 (opens in a new tab). Table 3 of that paper is the source for the generation counts: XXXI, 10 of 17 and 9 of 17; XXXII, 3 of 13 and 2 of 13. The ranking result, experimental forms of XXXII more than 4 kJ/mol above the global minimum, is from Hunnisett et al., Acta Crystallographica Section B, 2024, 80, 548–574, doi:10.1107/S2052520624008679 (opens in a new tab).
Four of six, eight rotatable bonds, one submitted structure, the list of 50, and the seven-day clock are pass marks defined here.
Four properties, one molecule
On most of a blinded set, one molecule drawn before any measurement meets potency, selectivity, solubility, and permeability in the same round.Claim
For a protein disclosed with its assay, its tool compound, and its counter-target, one new small molecule is drawn before any measurement on it and, in a single round of experiments, is potent, selective, soluble, and permeable.
Why it matters
A tight binder that cannot dissolve, cannot cross a membrane, or cannot ignore the protein next door does not become a drug. Projects learn those facts in sequence, and each fact starts a new design round. A method that hits all four on the first made molecule, for more than one target, would collapse that sequence into one experiment a standard laboratory can finish.
Current best (with sources)
Pieces of the four properties have been predicted, or measured, separately. A blinded one-molecule round that hits all four has not.
The 2025 ASAP–Polaris–OpenADMET challenges asked participants to predict properties of compounds that had already been made in an antiviral program: potency at two coronavirus main proteases, and, separately, microsomal stability, kinetic solubility, log D, and MDR1–MDCK permeability (potency challenge (opens in a new tab), ADMET challenge (opens in a new tab)). The split was chronological and the test values were hidden. Predicting a number for an existing structure is a different task from drawing a new structure that then meets the numbers.
Atz and co-workers made two generated PPARγ ligands and measured several properties on them. Compound 1 had an EC50 of 1.5 ± 0.2 µM at PPARγ, 0.24 ± 0.05 µM at PPARδ, a calorimetric dissociation constant of 0.8 ± 0.1 µM, a log D of 1.5, and a PAMPA effective permeability of 3.9 × 10^-6 cm/s. Compound 2 had a selective EC50 of 2.3 ± 0.7 µM at PPARγ, a log D of 1.7, and a PAMPA permeability of 14 × 10^-6 cm/s. In a 50-target panel, both stayed under 50 percent effect at 10 µM except at PPARγ. The routes were six and five linear steps (Nature Communications, 2024, doi:10.1038/s41467-024-47613-w (opens in a new tab)). The paper reports log D and PAMPA. It does not report a kinetic-solubility pass against a threshold, and the campaign is two molecules for one receptor, chosen as the top of a generated list, with scoring models trained on known ligands.
Munson and co-workers synthesized 32 generated compounds aimed at MEK1 and mTOR together. Most had cell-growth IC50 values in the 1–10 µM range, and four of six retested compounds reduced both phosphorylation readouts by more than half at 1 µM. The authors state that the method, as published, does not provide molecules optimized for absorption, distribution, metabolism, excretion, and toxicity (Nature Communications, 2024, doi:10.1038/s41467-024-47120-y (opens in a new tab)).
Thomas and co-workers report an 8-of-9 binding hit rate, including three nanomolar ligands, for generated A2A compounds on a single receptor, with human filtering before synthesis (Nature Communications, 2025, doi:10.1038/s41467-025-60629-0 (opens in a new tab)). Selectivity against A2B was examined. Solubility and permeability were not the grade.
CACHE challenges #1 through #4 test many commercial compounds per team and confirm sparse hits, generally at micromolar concentrations. They are summarized in the companion problem on single-property design. They do not grade solubility and permeability on the same molecule that matched a tool compound.
An open doubt: PAMPA and a biochemical counter-target are not a full picture of permeability or selectivity in a person. They are the measurements a chemistry laboratory can finish in the same fortnight as the potency assay. A kinetic-solubility cutoff of 100 µM is a pass mark defined here, chosen so the result is an isolated number rather than a comparison with an insoluble tool compound. It is open to revision if a side-by-side standard becomes the field’s usual control.
Specification
The laboratory preregisters 4 protein targets. Each has a biochemical assay the grading laboratory already runs on a plate reader, by absorbance, fluorescence, or luminescence, a published tool compound, and a counter-target assay named in advance. The tool compound’s structure is not disclosed until the procedures are locked. The laboratory does disclose that its published half-maximal concentration in an assay of the same type lies between 10 nM and 1 µM. The four targets are disclosed together.
Within 48 hours, and before any new measurement on the designed compounds, the attempt submits one structure per target, a route of at most four linear steps from commercial materials, and a written procedure. One backup structure and route may be submitted at the same time and is opened only if the first procedure does not yield the compound. Nothing is redesigned after the first spectrum or the first assay plate.
The laboratory runs each procedure as written and isolates the product. Purity is at least 95 percent by quantitative proton NMR with a weighed internal calibrant. The compound passes the same Lipinski filter used in the companion design problem: at most one violation among more than 5 hydrogen-bond donors, more than 10 hydrogen-bond acceptors, molecular weight above 500, and calculated log P above 5. The laboratory checks that the structure is not a reported ligand of that target in ChEMBL or PubChem on the preregistration date.
Four measurements are then run on that isolated solid, in one round:
- A dose–response of at least five concentrations, side by side with the tool compound and with vehicle.
- The same format on the named counter-target.
- Kinetic solubility at pH 7.4, by shake-flask and HPLC, or by a plate assay the laboratory has already validated, reported in micromolar.
- Parallel artificial membrane permeability (PAMPA) at pH 7.4, reported as an effective permeability in cm/s.
The four targets may be synthesized and assayed in parallel. The whole round finishes within ten working days of the procedures being opened.
Pass threshold
At least 3 of the 4 compounds pass all four of the following, on the isolated solid:
- half-maximal concentration within tenfold of the tool compound measured on the same plate
- at least tenfold weaker activity on the counter-target, with an inactive vehicle
- kinetic solubility of at least 100 µM at pH 7.4
- PAMPA effective permeability of at least 2 × 10^-6 cm/s
Three of four, the tenfold windows, 100 µM, 2 × 10^-6 cm/s, four linear steps, 95 percent purity, and the ten-day clock are pass marks defined here. The tenfold potency comparison uses the same plate, so assay drift does not decide the grade. The solubility and permeability numbers are pass marks, not records copied from a table.
Controls
Vehicle wells are inactive. For each target, a known inactive analog named at preregistration is inactive in the potency assay. The tool compound’s half-maximal concentration in the run lies within tenfold of its published value; otherwise the assay is repeated and not scored. Tenfold of the published value is a pass mark for assay validity.
On every solubility plate the laboratory runs a named high-solubility reference and a named low-solubility reference. The high reference must read at least tenfold higher than the low reference, or the plate is repeated. On every PAMPA plate the laboratory runs a named high-permeability reference and a named low-permeability reference. The high reference must exceed the low reference by at least fivefold, or the plate is repeated. The candidate is graded against the pass marks above, and the references only decide whether the plate is valid.
A procedure that does not produce the drawn compound fails that target. Purity below 95 percent fails that target even if the assays look favorable.
Allowed
Any design method. Any reaction in the written route. One backup compound, submitted before any assay or spectrum of the test compounds. PAMPA lipids and solubility media that the laboratory names in the preregistration.
Not allowed
A compound that is already a reported ligand of the target. A second round of analogs after any result. A list graded on its best member. A peptide or an oligonucleotide. A Lipinski filter violation beyond the one allowed above. Quoting a published potency, solubility, or permeability in place of the side-by-side or same-week measurement. Substituting a cellular IC50 for the biochemical assay, or a calculated log P for the kinetic-solubility measurement. Changing the route after a failed step. A structure that differs from the tool compound, or from any reported ligand of that target, only by adding, removing, or replacing one methyl, fluoro, or hydroxy group. The tool structure is compared after the procedures are locked.
Partial success
At least 3 of the 4 compounds meet the potency and selectivity bars, with purity and the drug-likeness filter satisfied, and miss solubility, permeability, or both. Three of four on those two endpoints is a pass mark for this lesser bar. Partial success is the companion single-property problem on a smaller set. It is not a solution of this one.
Sources
Lipinski, Lombardo, Dominy, and Feeney, Advanced Drug Delivery Reviews, 1997, 23, 3–25, doi:10.1016/S0169-409X(96)00423-1 (opens in a new tab). The “at most one violation” reading is a pass rule layered on that paper. Quantitative proton NMR purity follows Pauli and co-workers, Journal of Medicinal Chemistry, 2014, doi:10.1021/jm500734a (opens in a new tab).
The 100 µM solubility mark, the 2 × 10^-6 cm/s permeability mark, the tenfold windows, 3 of 4, and the four-step limit are pass marks defined here.
Cited experiments, each opened for this draft:
- Atz et al., Nature Communications, 2024, doi:10.1038/s41467-024-47613-w (opens in a new tab). Two PPARγ designs with the EC50, log D, and PAMPA values quoted above.
- Munson et al., Nature Communications, 2024, doi:10.1038/s41467-024-47120-y (opens in a new tab). Thirty-two dual MEK1/mTOR compounds; ADMET left for later.
- Thomas et al., Nature Communications, 2025, doi:10.1038/s41467-025-60629-0 (opens in a new tab). Eight of nine A2A binders on one target.
- Polaris challenges, 2025: antiviral potency (opens in a new tab) and antiviral ADMET (opens in a new tab). Blinded prediction for compounds already made.
The yield named before the flask
On most of a blinded set, the isolated yield of one execution of a locked procedure falls close to the yield named in advance.Claim
For a fully written reaction the attempt has never run, the isolated yield named in advance agrees with the isolated yield of a single execution of that procedure.
Why it matters
A route planner that can draw a product and a route planner that can say how much of it will be in the flask are different tools. The second one decides whether a campaign is worth starting, whether a robot should spend the substrate, and whether a predicted condition is actually better than the one already written. Isolated yield, measured once, is a number a teaching laboratory can produce. Predicting it for reactions that have not been published is still open.
Current best (with sources)
The closest 2025 prospective test predicts whether a product ion appears, inside one reaction family, after a very large screen of that same family.
Götz and co-workers ran about 50,000 three-component reactions on a 3-microliter scale, from 193 substrates, and kept a public set of about 40,000 combinations analyzed by liquid chromatography–high-resolution mass spectrometry (Science Advances, 2025, doi:10.1126/sciadv.adw6047 (opens in a new tab)). Models predicted which of several products formed. When every building block had already appeared in training, accuracy on products A–C was 93 ± 0.2 percent. When none of the three building blocks had appeared, accuracy was 78 ± 3.3 percent. In a prospective plate, 147 of 149 reactions predicted to give product A did so. For 12 new initiators, the model called the formation of product A correctly in 533 of 593 valid reactions. The authors note that product A formed in more than four of five reactions, so for that product a classifier that always says “yes” is already strong, and only the easier splits clearly beat it. The endpoint is detection of a product ion in a mixture. It is not an isolated yield, and it is not a set of different reaction classes under locked preparative procedures.
An open doubt: a chemist who knows a reaction class can often guess a yield band. The ten-point window below is a pass mark defined here, tight enough that a flat guess fails when the set contains both high-yielding and failed reactions. If a later public set shows that experienced chemists already land inside ten points on nine of twelve locked procedures of this kind, the window should be tightened or the problem retired. No such graded, multi-class, isolated-yield challenge was found for 2023–2026.
Specification
An independent laboratory writes 12 procedures before the attempt sees them. Each procedure states the substrates, reagents, equivalents, solvent, concentration, temperature, time, and workup, at a scale between 0.2 and 0.5 mmol. The substrates are commercial or already in the laboratory’s inventory. The exact transformation, meaning those substrates under a procedure of the same reaction class, is absent from the literature and from the Open Reaction Database on the day of preregistration; the laboratory records the search.
The 12 procedures cover at least three reaction classes, with at least three procedures in each class. The classes are named in the preregistration and are ordinary preparative classes a synthetic laboratory already runs, such as a cross-coupling, an amide coupling, and a reductive amination. The laboratory may choose other named classes. It may not choose twelve variants of one class.
The procedures are released together, without yields. Within 48 hours, and without running any of the 12, the attempt submits one predicted isolated yield for each procedure, as a percentage to the nearest whole number. The laboratory, or a robot following the same written steps, then executes each procedure once. A deviation from the written procedure fails that entry. Isolated yield is the mass of product obtained after the written workup and the written purification, corrected by quantitative proton NMR against a weighed internal calibrant, divided by the theoretical mass.
After the yields are measured, the laboratory checks that the highest and lowest isolated yields in the set differ by at least 40 percentage points. If they do not, the set is invalid and is replaced with a new set of 12. The attempt’s predictions are not used to decide which procedures are replaced. Forty percentage points is a pass mark for the set, so that a constant prediction cannot hide inside a flat set.
Pass threshold
At least 9 of the 12 predicted yields lie within 10 percentage points of the isolated yield measured for that procedure.
Nine of twelve and the 10-point window are pass marks defined here. They are not a reported error from a leaderboard. Ten points is about the width of a careful quantitative-NMR yield, doubled, so the grade is a real prediction and not a quarrel about integration.
Controls
For every isolated product, a blank NMR sample of the internal calibrant in the same solvent is recorded the same day. The product signals used for the assay are free of the calibrant and of residual solvent. Two of the twelve reactions, chosen by the laboratory before the predictions are opened, are assayed twice by two people. The two yields must agree within 5 percentage points, or the batch is invalid and those two procedures are repeated. Five points is a pass mark for assay agreement.
For two further procedures, chosen the same way, a blank is run with the catalyst or the coupling reagent omitted and every other step unchanged. The blank must give below 5 percent isolated yield of the product by the same assay. If it does not, the batch is invalid. Five percent is a pass mark for the blank.
Identity is assigned by a chemist who did not write the prediction, from a proton spectrum, a carbon spectrum, and a high-resolution mass measurement. An ambiguous assignment fails that entry.
Allowed
Any model, lookup of other reactions, or calculation. A robot may execute the locked procedure. A person may execute it. The executor follows the procedure as written.
Not allowed
Running any of the 12 reactions, or a close analog made for the purpose, before the yields are submitted. Editing a predicted number after the flask is started. Changing equivalents, temperature, time, or workup. Reporting a range and asking to be scored on the nearer end. Using a published yield for the same substrates. Substituting a chromatographic area percent for the isolated, NMR-corrected yield. Dropping a reaction from the set after seeing that the prediction missed.
Partial success
At least 9 of the 12 predictions lie within 20 percentage points of the measured isolated yield, and the blanks and the duplicate assays pass. Twenty points and 9 of 12 are pass marks for this lesser bar. Partial success is not a solution.
Sources
Nine of twelve, the 10-point and 20-point windows, the 40-point spread required of a valid set, the 5-point duplicate limit, and the 5 percent blank are pass marks defined here. Quantitative proton NMR follows Pauli and co-workers, Journal of Medicinal Chemistry, 2014, doi:10.1021/jm500734a (opens in a new tab).
The formation-prediction results quoted above are from Götz, Richards, Stepek, Takahashi, Huang, Bertschi, Rubi, and Bode, Science Advances, 2025, 11, eadw6047, doi:10.1126/sciadv.adw6047 (opens in a new tab).
A new phase, then the powder pattern
Most of a blinded set of predicted inorganic crystals are synthesized once and match an independent powder-diffraction refinement.Claim
A crystal structure that is predicted to be a new, stable inorganic phase is made by one written procedure, and an independent powder pattern shows that phase as the majority of the solid.
Why it matters
Computational catalogues now hold far more “stable” crystals than synthesis has ever confirmed. Energy above a convex hull is a useful filter and a weak promise: a file can be stable at zero kelvin and still be a known disordered solid solution, or a mixture, or a phase that never appears in the crucible. A grade that requires one firing, a majority phase, and a refinement done by someone who did not write the prediction would turn those catalogues into materials a laboratory can hold.
Current best (with sources)
Large catalogues are real. Prospective, independently refined new phases are not established at the rate the catalogues suggest.
Merchant and co-workers report that graph networks found more than 2.2 million structures stable relative to earlier computational catalogues, of which 381,000 lie on the updated convex hull. The models predict energies to 11 meV per atom, and the precision of a “stable” call is above 80 percent when the structure is supplied and 33 percent per 100 trials from composition alone. Of structures already in the Inorganic Crystal Structure Database, 736 match structures the search also produced (Nature, 2023, doi:10.1038/s41586-023-06735-9 (opens in a new tab)). Those 736 matches are overlap with experiments done independently of a preregistered synthesis campaign. They are not a campaign.
Szymanski and co-workers built the A-Lab, a robotic solid-state laboratory, and reported a run against targets drawn from the Materials Project and from that search. The article page, as opened for this draft, states that the laboratory synthesized 36 of 57 targets over 17 days, that manual refinement of each of those 36 patterns required the target phase for a satisfactory fit, that many samples still contained prominent byproducts, and that four further calls were inconclusive by powder diffraction alone. An author correction was published on 19 January 2026 (Nature, 2023, doi:10.1038/s41586-023-06734-w (opens in a new tab)). Leeman, Liu, Stiles, Lee, Bhatt, Schoop, and Palgrave re-examined the original claim of new materials and concluded that no new materials had been discovered in that work. They report that automated Rietveld analysis was not reliable, and that about two thirds of the claimed successes are likely known, compositionally disordered versions of the ordered predictions (PRX Energy, 2024, 3, 011002, doi:10.1103/PRXEnergy.3.011002 (opens in a new tab)).
Prein and co-workers selected candidates with a synthesizability model, fired them once in a benchtop muffle furnace, and characterized 16 samples by powder diffraction. Seven of the 16 matched the target and two more matched a polymorph. Of the successes, three were already in the Inorganic Crystal Structure Database and marked as synthesized. Three compositions had no structure in that database: Y2MnFeO6 was present at about 3.3 percent in a mixture of known polymorphs; Nd3BTeO9 was obtained nearly pure, and the authors note a 2023 synthesis paper that lacked a deposited structure; Eu2WO6 was a non-target polymorph at about 88 percent. The authors write that most successes were isostructural with known compounds and that many entries marked theoretical were already known (NeurIPS 2025 AI for Science workshop, arXiv:2511.01790 (opens in a new tab)).
An open doubt: the A-Lab correction and the Leeman reanalysis do not yet share one public table in which every pattern has been refined both as the ordered target and as the best known disordered model. This problem makes that comparison the grade. Prein’s workshop paper is the attempt’s own refinement, not an independent one. How many of their seven would survive the rule below is unanswered here.
Specification
The attempt deposits eight predicted crystal structures, as crystallographic information files, before any synthesis. Each composition is absent, on the deposit date, from the Inorganic Crystal Structure Database and from the experimental entries of the Materials Project. The attempt includes a short search record. A structure that fails the search is replaced before synthesis begins.
For each target the attempt also deposits one procedure: precursor powders that a named supplier sells, the molar ratios, the mill or grinding step, the atmosphere (air or one named cylinder gas), a maximum temperature at or below 1100 °C, and a hold of at most 24 hours. The scale is 200 to 500 mg. No flux, no high-pressure cell, and no second firing.
The laboratory runs each procedure once, as written. It then hands the powder and the target file to a crystallographer who did not help generate the prediction. That person collects a powder pattern on a laboratory diffractometer and performs a Rietveld refinement. The refinement tries, at minimum, three models: the predicted ordered structure; the best matching known phases of those elements from the two databases above; and a compositionally disordered model in which those elements share sites, when such a model is crystallographically possible. Scale factors, lattice parameters, and a simple background are refined. Atomic coordinates may be refined only inside the restraints of the starting model.
A target counts only if the predicted phase is the majority phase, at least 70 percent by refined weight, and the weighted-profile residual of the model that includes the predicted phase is lower by at least 2 percentage points than the best model that excludes it. Seventy percent and 2 percentage points are pass marks defined here. They exist so a trace peak, or a fit that a known solid solution matches equally well, cannot pass.
The eight procedures may run in parallel. Synthesis and refinement finish within ten working days.
Pass threshold
At least 4 of the 8 targets meet the majority-phase rule and the residual rule above, and each of those four was absent from the two databases on the deposit date.
Four of eight is a pass mark defined here. It is set above the fraction of genuinely new, majority-phase structures that the opened 2023–2025 reports can support, once known phases, disordered models, and trace constituents are treated as misses.
Controls
A silicon or corundum standard is run on the diffractometer the same week, and the refined lattice of the standard agrees with its accepted lattice within 0.2 percent, or the patterns are recollected. Two tenths of a percent is a pass mark for instrument validity.
For one target chosen by the laboratory before the procedures are opened, an empty-crucible blank is heated through the same program. Its pattern must not show the predicted phase. If it does, the furnace run is invalid.
The crystallographer’s refinement, the alternative models, and the residuals are archived. A narrative that the phase “looks present” does not replace the weight percent and the residual comparison. If two qualified refiners differ by enough to change a pass into a fail, the target fails.
Allowed
Any stability model, substitution rule, or generative model, used before the files are deposited. Any precursor the procedure names, inside the temperature and atmosphere limits. Computation after the deposit is irrelevant, because the files are frozen.
Not allowed
A second procedure after the pattern is seen. A human edit to the target structure after the pattern is seen. Counting a known phase, or a disordered model that fits as well, as the predicted compound. High pressure, a hydrothermal bomb, or a temperature above 1100 °C. A thin film in place of a powder. An automated refinement that no person has checked against the alternative models. A composition that was already in the databases on the deposit date.
Partial success
At least 4 of the 8 patterns show the predicted phase at 30 percent or more by refined weight, with the same residual rule, even if the phase is not the majority. Thirty percent and 4 of 8 are pass marks for this lesser bar. Partial success is a detected phase in a mixture. It is not a made material.
Sources
Seventy percent, 30 percent, the 2-point residual gap, 4 of 8, 1100 °C, and the 0.2 percent standard check are pass marks defined here.
Cited reports, each opened for this draft:
- Merchant et al., Nature, 2023, doi:10.1038/s41586-023-06735-9 (opens in a new tab). 2.2 million structures below earlier hulls; 381,000 on the updated hull; 736 matches to existing experimental structures.
- Szymanski et al., Nature, 2023, doi:10.1038/s41586-023-06734-w (opens in a new tab). Article page: 36 of 57 after manual review; correction dated 19 January 2026.
- Leeman et al., PRX Energy, 2024, 3, 011002, doi:10.1103/PRXEnergy.3.011002 (opens in a new tab). Reanalysis: no new materials established; disordered known phases; automated Rietveld unreliable.
- Prein et al., arXiv:2511.01790, https://arxiv.org/html/2511.01790 (opens in a new tab). Seven of 16 characterized samples matched a target, with the novelty limits quoted above.
Fluorine at a chosen carbon
Replace one preregistered hydrogen with fluorine, and not the others, on most of a blinded set of molecules, including unactivated sites.Claim
Given an ordinary organic molecule and one carbon–hydrogen bond marked in advance, a written procedure replaces that hydrogen with one fluorine and does not do the same at the other carbon–hydrogen bonds. The marked carbon may be an unactivated methyl or methylene sitting next to a C–H bond the usual fluorination catalysts would prefer.
Why it matters
Fluorine is how a candidate’s metabolism, acidity, and conformation are tuned, and it is usually installed by building the molecule from a fluorinated starting material. A late fluorination counts only if the site was chosen. A method that fluorinates the weakest C–H bond on a steroid is useful, and it is not the same result.
Current best
Liu, Huang, Cheng, Nielsen, Goddard, and Groves reported a manganese porphyrin that fluorinates aliphatic C–H bonds with fluoride ion and a stoichiometric iodine(III) oxidant (doi:10.1126/science.1222327 (opens in a new tab)). The abstract states that simple alkanes, terpenoids, and steroids were fluorinated at otherwise inaccessible sites in 50 to 60 percent yield, and that the site of C–H cleavage is set by an oxomanganese(V) intermediate. Decalin was fluorinated predominantly at the C2 and C3 methylene positions, bornyl acetate gave the exo-5-fluoro product, and 5α-androstan-17-one was fluorinated in the A ring. The site is the one that intermediate selects. A later protocol from the same laboratory, covering the porphyrin and a manganese salen catalyst, states that isolated methylene groups and unactivated benzylic sites are accessible and that the reactions typically afford 50 to 70 percent yield of the monofluoride. Celestolide and ibuprofen gave about 50 percent isolated yield, on a few hundred milligrams, in one to eight hours (doi:10.1038/nprot.2013.144 (opens in a new tab)). Those are still the sites the method addresses, not a carbon marked in advance against a more reactive neighbor.
Wang, Ruskin, Marques, Garrison, and Lectka surveyed single-atom late-stage fluorination of natural products and drugs in 2025 and present it as a collection of methods, nucleophilic and electrophilic, each with its own substrate class (doi:10.1021/acs.chemrev.5c00363 (opens in a new tab)). Nothing opened for this problem accepts a blinded set, a carbon marked by someone else, and a procedure frozen before any experiment on that set.
Specification
An independent laboratory preregisters 12 molecules. Each has a molecular weight between 180 and 500, at least four constitutionally distinct carbon–hydrogen sites, and only the elements C, H, N, O, S, F, Cl, and Br. Fluorine already in the drawing is allowed and is not the marked site. On each structure the laboratory marks exactly one carbon that bears at least one hydrogen. The required product is the monofluoride at that carbon. A gem-difluoride, an alcohol, or a fluoride at any other carbon does not count.
At least four of the twelve marked carbons are unactivated, by a rule fixed before the structures are released. A marked carbon is unactivated only when all three of the following are true: it is a methyl or a methylene; it is not benzylic, not allylic, and not adjacent to oxygen, nitrogen, or sulfur; and the same molecule contains at least one benzylic C–H or tertiary aliphatic C–H. The attempt sees the marked structures.
The structures are revealed together. Within 24 hours, and with no experiment on any of the 12, the attempt submits one written procedure. Only the substrate changes from molecule to molecule. The laboratory runs that procedure once, as written, on 0.1 to 0.3 mmol of substrate. One backup procedure may be submitted in the same 24 hours. If it is used, it applies to every substrate or to none.
The site ratio is measured on the crude mixture, before chromatography, by fluorine NMR against a weighed internal calibrant: the integral of the specified monofluoride divided by the sum of the integrals of every other monofluoride. The specified product is then isolated. Identity is a proton spectrum, a fluorine spectrum, a carbon spectrum, and an HSQC or a fluorine–proton correlation sufficient to place the fluorine. A chemist who did not write the procedure does the placement. Purity of the isolated monofluoride is at least 95 percent by absolute quantitative proton NMR.
Pass threshold
At least 9 of the 12 substrates give the specified monofluoride in an isolated yield of at least 30 percent, with a crude site ratio of at least 10 to 1, and at least 3 of those 9 are among the 4 unactivated sites.
The count, the 3-of-4 unactivated requirement, the 30 percent yield, the 10-to-1 ratio, and the 95 percent purity are pass marks defined here. Thirty percent is below the 50 to 60 percent yields Groves and co-workers report for sites their catalyst selects, and it is a floor for isolation rather than a record. The unactivated-site count is what a benzylic-only method cannot clear. Ten to one means the marked carbon accounts for at least about 91 percent of the monofluorinated material.
Controls
A no-reagent blank, identical except for the absence of the catalyst or fluorinating reagent named in the procedure, is run for two substrates chosen before the procedures are opened. One of the two is unactivated. The blank must not produce the specified monofluoride above 2 percent yield by the same fluorine-NMR assay. If it does, the batch is invalid.
The crude fluorine spectrum must be integrated against the calibrant, not against a chromatogram that has already discarded the other isomers. If the assigning chemist cannot place the fluorine, the substrate fails.
Allowed
Any catalyst, fluoride salt, electrophilic fluorinating reagent, light source, or electrochemical cell that the written procedure names. Computation before the 24-hour clock starts is allowed. A stoichiometric oxidant is allowed if it is named in the procedure. The single backup procedure submitted with the first procedure is allowed.
Not allowed
Installing a halogen, a boron substituent, a diazonium salt, or a directing group on the way to the fluoride. Deoxyfluorination of an alcohol that is not already the marked carbon in the preregistered drawing. Halogen exchange at a carbon that was not the marked C–H. Counting a gem-difluoride, or a mixture of monofluorides, as the product. Screening conditions on the test substrate, or on a close analog made for the purpose, after the structures are revealed. Reporting only the eight activated sites and treating the unactivated four as optional.
Partial success
At least 6 of the 12 substrates meet a 15 percent isolated yield and a 5-to-1 site ratio, at least 1 of the 6 is an unactivated site, and the blanks pass. These figures are pass marks for the lesser bar. Partial success is not a solution.
Sources
The success counts, the unactivated-site rule, the 30 percent and 15 percent yields, the site ratios, the 2 percent blank ceiling, and the 95 percent purity are pass marks defined here.
Cited measurements:
- Liu, Huang, Cheng, Nielsen, Goddard, and Groves, Science, 2012, 337, 1322–1325, doi:10.1126/science.1222327 (opens in a new tab). Aliphatic C–H fluorination in 50 to 60 percent yield at sites selected by the oxomanganese intermediate. The decalin, bornyl acetate, and androstanone descriptions are from that abstract.
- Liu, Huang, and Groves, Nature Protocols, 2013, 8, 2348–2354, doi:10.1038/nprot.2013.144 (opens in a new tab). Isolated methylenes and unactivated benzylic sites; typically 50 to 70 percent yield of the monofluoride. Celestolide and ibuprofen, about 50 percent isolated, a few hundred milligrams. Individual yields in the substrate tables were not re-tabulated here.
- Wang, Ruskin, Marques, Garrison, and Lectka, Chemical Reviews, 2025, 125, 9382–9428, doi:10.1021/acs.chemrev.5c00363 (opens in a new tab). A survey of late-stage fluorination by substrate class, not a blinded site-selection test.
Purity uses the absolute quantitative proton NMR assay of Pauli and co-workers, Journal of Medicinal Chemistry, 2014, 57, 9220–9231, doi:10.1021/jm500734a (opens in a new tab). The 50 to 70 percent sentence and the celestolide and ibuprofen figures are taken from the protocol abstract, not from a line-by-line check of every example in its tables.
An acyclic all-carbon quaternary center
A procedure written before any experiment returns the named enantiomer of an acyclic all-carbon quaternary center for most of a blinded set.Claim
Given an achiral or racemic starting material and a drawing of one enantiomer of a product, a procedure written with no experiment on that substrate forms an acyclic all-carbon quaternary stereocenter as that enantiomer. The set is not ten examples of one privileged reaction.
Why it matters
A carbon that bears four different carbon substituents is a shape that medicinal chemistry and natural-product synthesis both use and both struggle to build, especially when that carbon is not in a ring. Cyclic quaternary centers have a long list of reliable catalytic methods. Acyclic ones are where a route still changes to avoid the stereocenter, or resolves it at the end. A first procedure that returns the named enantiomer, including on constructions that are not the two best-known ones, is the result that would make those centers ordinary intermediates.
Current best
The reviews separate cyclic from acyclic, and the best new method occupies one acyclic subclass.
Feng, Holmes, and Krische reviewed enantioselective construction of acyclic quaternary carbon stereocenters and described it as still formidable beside the much larger body of cyclic methods (doi:10.1021/acs.chemrev.7b00385 (opens in a new tab)). Griffiths and Franckevičius, on catalytic allylic alkylation of acyclic enolates, write that the acyclic case is harder because the enolate geometry is not fixed the way it is in a ring (doi:10.1002/chem.202304289 (opens in a new tab)). Xu, Zhou, Zhu, and Zhou reviewed catalytic desymmetrizations that produce all-carbon quaternary centers and still describe the broader construction as a long-term challenge (doi:10.1038/s44160-023-00406-3 (opens in a new tab)).
Chen, Jiang, Yao, Feng, Zhong, Huang, and Shi then reported a nickel-catalyzed enantioselective C(sp³)–H amination that desymmetrizes a gem-dimethyl group and gives β-amino acid derivatives with an acyclic all-carbon quaternary center in high yields and with enantioselectivity up to greater than 99 percent ee (doi:10.1021/jacs.5c16651 (opens in a new tab)). The abstract does not state a single isolated-yield figure for the whole scope, and none is invented here. The transformation is one bond construction, on one substrate pattern. It shows that greater than 99 percent ee is reachable inside that pattern. It does not show a procedure, written before a blinded multi-class set is revealed, that returns the named enantiomer for constructions outside it.
Specification
The laboratory preregisters 10 transformations. Each is a single step from a commercial or previously prepared achiral or racemic starting material to one specified enantiomer. The new stereocenter in the product is acyclic and all-carbon: it bears four different carbon substituents, it is not a ring atom, and none of the four bonds is to hydrogen, nitrogen, oxygen, or a halogen.
Before the attempt sees the set, the laboratory marks at least four transformations as unprivileged. A transformation is unprivileged only if the product is not made by either of the following: allylic alkylation, or another enolate alkylation, of a β-ketoester, a β-ketoamide, a malonate, or an α-cyano carbonyl; or C–H functionalization of a gem-dimethyl group. The laboratory records, for each of the four, which bond is formed. The four bonds are not the same bond construction. The other six transformations may be privileged, so that a method cannot pass by declining every case except the one it already knows.
Within 24 hours of disclosure, and with no experiment on a test substrate or on a stand-in made to resemble it, the attempt submits one procedure per transformation: catalyst or reagent, loading, solvent, temperature, time, and workup. Loading of any chiral catalyst or ligand is at most 5 mol percent. One backup procedure per substrate may be submitted at the same time.
The laboratory runs the procedure once. The isolated product is identified by NMR and high-resolution mass spectrometry. The enantiomer ratio is measured by chiral chromatography against a racemic sample of the same constitution, injected on the same method the same day. The specified enantiomer is assigned by a chemical correlation to a compound of known configuration, a vibrational circular-dichroism comparison, or an X-ray structure of a heavy-atom derivative. Optical rotation alone is not sufficient. Purity of the isolated enantiomer is at least 95 percent by absolute quantitative proton NMR.
Pass threshold
At least 8 of the 10 transformations, including at least 3 of the 4 unprivileged ones, give the specified enantiomer in an isolated yield of at least 40 percent and an enantiomer ratio of at least 95 to 5, at the purity above.
The count, the unprivileged quota, the 40 percent yield, the 95-to-5 ratio, the 5 mol percent ceiling, and the 95 percent purity are pass marks defined here. Ninety-five to five is 90 percent enantiomeric excess. The Shi result of greater than 99 percent ee shows that a higher ratio is already available inside gem-dimethyl C–H amination, which is why that reaction cannot fill the unprivileged quota. Forty percent is a floor for an isolable sample, not a yield taken from that paper.
Controls
The racemic reference must show both enantiomers. A single peak is credited only if the racemate resolves on that method and a spike identifies the named enantiomer. A reaction run without the chiral catalyst, otherwise identical, is performed for two substrates chosen in advance, one of them unprivileged. That control must not itself reach 95 to 5. If it does, the substrate or the assay is contaminated and the batch is invalid.
The laboratory’s assignment of the four unprivileged transformations is checked against the definitions above before the procedures are opened. A transformation that is in fact an enolate alkylation of a β-dicarbonyl, or a gem-dimethyl C–H functionalization, does not count toward the quota.
Allowed
Any chiral catalyst, ligand, or enzyme named in the procedure, within the loading rule. A different catalyst for each substrate. Computation and literature lookup before the procedures are locked. The one backup procedure submitted with the first procedure.
Not allowed
A stoichiometric chiral auxiliary that is attached and removed. Classical resolution, or separation of enantiomers on a chiral stationary phase presented as a synthesis. A product whose quaternary center is in a ring. Filling the unprivileged quota with four examples of one bond construction, or with the two excluded reaction classes. Changing solvent, catalyst, or temperature after the first experiment. More than 5 mol percent of the chiral catalyst. Quoting enantioselectivity without the absolute-configuration assignment named above.
Partial success
At least 6 of the 10 transformations reach an enantiomer ratio of at least 90 to 10 and an isolated yield of at least 40 percent, including at least 2 of the 4 unprivileged ones, and the blanks pass. These figures are pass marks for the lesser bar. Partial success is not a solution. Six privileged successes and no unprivileged success are not partial success.
Sources
The counts, the yield, the ratios, the loading, and the purity are pass marks defined here.
Cited sources:
- Feng, Holmes, and Krische, Chemical Reviews, 2017, 117, 12564–12580, doi:10.1021/acs.chemrev.7b00385 (opens in a new tab). Acyclic quaternary carbon stereocenters are described as a remaining challenge beside cyclic methods.
- Griffiths and Franckevičius, Chemistry – A European Journal, 2024, 30, e202304289, doi:10.1002/chem.202304289 (opens in a new tab). Acyclic allylic alkylation is harder because the enolate geometry is not fixed.
- Xu, Zhou, Zhu, and Zhou, Nature Synthesis, 2023, 2, 1020–1036, doi:10.1038/s44160-023-00406-3 (opens in a new tab). Desymmetrization routes to all-carbon quaternary centers, with the broader problem still described as open.
- Chen, Jiang, Yao, Feng, Zhong, Huang, and Shi, Journal of the American Chemical Society, 2025, 147, 39060–39066, doi:10.1021/jacs.5c16651 (opens in a new tab). Nickel-catalyzed gem-dimethyl C–H amination, acyclic all-carbon quaternary centers, enantioselectivity up to greater than 99 percent ee. No scope-wide yield is cited because the abstract does not give one.
Purity uses the absolute quantitative proton NMR assay of Pauli and co-workers, Journal of Medicinal Chemistry, 2014, 57, 9220–9231, doi:10.1021/jm500734a (opens in a new tab).
Direct air capture, measured heat in
A sorbent takes carbon dioxide from 400 ppm air and gives it back, over ten cycles, at a capacity and a measured heater energy past the resin cited here.Claim
One solid sorbent, used for ten cycles without being replaced, captures carbon dioxide from air at about 400 ppm and releases it as a concentrated stream, with a cyclic working capacity and a measured heater energy that are both past the vapor-promoted resin process cited here.
Why it matters
Taking carbon dioxide out of air is only useful if the heat to get it back off the sorbent is small and the sorbent keeps its capacity. Equilibrium uptakes and tabulated heats of adsorption leave out the sensible heat of the bed and the water that real air carries. The grade is carbon dioxide collected from a humid 400 ppm feed, divided into the electrical energy that actually went into the heater, on one sample, ten times.
Current best (with sources)
Wang, Qu, Ding, Webley, and Li report a primary-amine resin whose carbon dioxide uptake is 1.11 mmol g⁻¹ at 20 °C and 400 ppm. In their vapor-promoted desorption cycles the working capacity stayed at 0.9–1.0 mmol g⁻¹ over nine cycles. They calculate a thermal energy of 10.4 MJ per kilogram of CO₂ for the solar-heated prototype, and they give 12–15 MJ kg⁻¹ for temperature-vacuum swing on the same resin, including 0.5–1 MJ kg⁻¹ of electrical energy for evacuation. The abstract reports carbon dioxide of 97.7% purity together with water, and recovery of 98% of the adsorbed CO₂ in the sunlight-driven prototype. Desorption in that process is driven near 100 °C by the vapor, without a vacuum pump.
A heat of adsorption in kilojoules per mole is not this measurement. Their 10.4 MJ kg⁻¹ already includes sensible heat and the latent heat of water, which is why it sits well above a bare reaction enthalpy.
Specification
At least 1.00 g of dry sorbent is packed in a tube. The dry mass is recorded before cycle 1. The feed is air, or synthetic air, at 1.00 ± 0.05 bar and 25 ± 2 °C, with the inlet carbon dioxide measured at 400 ± 40 ppm and the relative humidity at 50 ± 10%. Inlet concentration, humidity, and flow are measured, not assumed.
Each cycle is adsorption until the outlet has returned to 95% of the inlet concentration, or for a shorter preregistered time, then desorption. Desorption heats the bed from 25 ± 2 °C to a final temperature of at most 120 °C. The product gas is collected until the outlet carbon dioxide returns to within 5% of the inlet value. Water is condensed. The carbon dioxide collected is the integral of measured outlet flow and measured dry-gas concentration. Purity is the mole fraction of carbon dioxide in that collected gas on a dry basis.
Heater energy is the electrical energy delivered to the heater from the start of the ramp until collection stops, measured by a watt-hour meter or by integrated voltage and current. No blank heat is subtracted. A tabulated enthalpy does not count.
Cycle 1 is reported and not graded. Cycles 2 through 10 are graded. The same sorbent charge is used for all ten cycles. It is not reloaded, washed, or regenerated outside the measured heater step. The ten cycles finish within 14 days.
Working capacity is the moles of carbon dioxide collected in desorption, divided by the initial dry mass. Energy is the heater energy divided by the mass of carbon dioxide collected. Both are averages over cycles 2 through 10.
Pass threshold
Over cycles 2 through 10, the average working capacity is at least 1.50 mmol of CO₂ per gram of dry sorbent, the average heater energy is at most 6.0 MJ per kilogram of CO₂ collected, and the average dry-gas purity is at least 95 mol% CO₂. The bed returns to 25 ± 2 °C between cycles, and the desorption temperature never exceeds 120 °C.
1.50 mmol g⁻¹, 6.0 MJ kg⁻¹, 95 mol%, 120 °C, ten cycles, 1.00 g, and 14 days are pass marks defined here. The capacity mark is above the 0.9–1.0 mmol g⁻¹ cyclic capacity and above the 1.11 mmol g⁻¹ uptake of the resin cited above. The energy mark is below the 10.4 MJ kg⁻¹ calculated for that vapor-promoted prototype. Neither number is a claim that no other paper has printed a larger equilibrium uptake.
Controls
- A tube packed with glass beads of the same bed volume, taken through one adsorption and desorption under the same flows and the same temperature program, collects less than 5% of the carbon dioxide collected in the candidate’s cycle 2. If it does not, the analyzer or a leak is producing the signal, and the sorbent is not graded.
- The inlet concentration is measured on every cycle. A feed that drifts outside 400 ± 40 ppm voids that cycle.
- Flow is measured at the outlet during desorption. An assumed flow fails the attempt.
- The watt-hour measurement is checked the same week against a resistive load of known power. The check is reported.
Allowed
Any solid sorbent, including amines, salts, and frameworks. A co-adsorbed water layer, if it is picked up from the humid feed and not added as liquid water or steam from outside. Electrical heat. A tube, a heater, a flow meter, and a carbon dioxide analyzer.
Not allowed
A feed above 1,000 ppm carbon dioxide, including flue gas. Liquid amine scrubbing. Adding steam, purge gas, or vacuum and omitting their energy. Subtracting a sensible-heat blank so that only a reaction enthalpy remains. Reporting an equilibrium uptake in place of the carbon dioxide collected on cycles 2 through 10. Replacing or reloading the sorbent between graded cycles. A desorption above 120 °C. A sample below 1.00 g dry mass.
Partial success
Over cycles 2 through 10, the average working capacity is at least 1.20 mmol g⁻¹ and the average heater energy is at most 8.0 MJ kg⁻¹, with dry-gas purity at least 95 mol% and the same temperature cap. 1.20 mmol g⁻¹ and 8.0 MJ kg⁻¹ are pass marks for this lesser bar. They are past the cited cyclic capacity and past the cited 10.4 MJ kg⁻¹, and they are not the full mark. A high capacity at an unmeasured enthalpy is not partial success.
Sources
Resin uptake of 1.11 mmol g⁻¹ at 20 °C and 400 ppm, cyclic working capacity of 0.9–1.0 mmol g⁻¹, calculated thermal energy of 10.4 MJ kg⁻¹, the 12–15 MJ kg⁻¹ comparison, 97.7% purity, and 98% recovery: Wang, Qu, Ding, Webley, and Li, Nature Communications, 2024, 15, 9745, doi:10.1038/s41467-024-53961-4 (opens in a new tab). The 1.50 mmol g⁻¹, 6.0 MJ kg⁻¹, 1.20 mmol g⁻¹, 8.0 MJ kg⁻¹, 95 mol%, 120 °C, 1.00 g, and 14-day requirements are pass marks defined here. Open point: other sorbents have been reported with equilibrium uptakes, at 400 ppm, above 1.11 mmol g⁻¹. Those uptakes are not cyclic heater-energy measurements on a humid feed, so this problem does not treat them as having cleared the joint bar. A later paper that does report both numbers under a comparable humid cycle should move the marks.
Solar hydrogen from pure water
A photocatalyst splits pure water at ambient pressure under one sun, and the six-hour solar-to-hydrogen efficiency is past the sheet results cited here.Claim
A particulate photocatalyst, as a sheet or a suspension in one vessel, splits pure water into hydrogen and oxygen at ambient pressure under one-sun illumination, with no applied bias and no sacrificial reagent, at a solar-to-hydrogen efficiency past the pure-water sheet results cited here, for six continuous hours.
Why it matters
Hydrogen from sunlight and pure water avoids an electrolyser and a purified feed. Photovoltaic electrolysis already reaches high efficiencies and is a different device. The photocatalyst version is still near one percent in the sheet experiments that define the field, and several of those efficiencies were measured below atmospheric pressure or above room temperature. The grade is hydrogen and oxygen from pure water, at 1 bar, under a measured one-sun flux, for an afternoon.
Current best (with sources)
Wang, Hisatomi, Domen, and co-workers report a SrTiO₃:La,Rh / Au / BiVO₄:Mo sheet that splits pure water at pH 6.8 with a solar-to-hydrogen efficiency of 1.1% and an apparent quantum yield above 30% at 419 nm. The body of that paper places the 1.1% result at 331 K and 10 kPa, and a related point of 1.0% at 318 K and 5 kPa. A carbon-conductor sheet, SrTiO₃:La,Rh / C / BiVO₄:Mo, is reported at 1.2% at 331 K and 10 kPa, retaining 80% of that efficiency at 91 kPa, and the paper states that 1.0% was then the highest value among Z-scheme pure-water systems operating at ambient pressure. A 2024 review from the same line of work still treats photocatalyst-sheet efficiencies above 1%, and practical solar-to-hydrogen efficiencies of about 1% at most, as the state of that approach. A 100 m² array of SrTiO₃:Al panels in that review is a scale demonstration, not a higher one-sun efficiency.
Two higher numbers are a different experiment and do not clear this problem. Fu, Wu, and co-workers report 2.47 ± 0.03% for a two-cell system bridged by an iodide redox couple, with the hydrogen cell in aqueous hydrohalic solution. Hansora and co-workers report a perovskite photoelectrode module with a module-level solar-to-hydrogen efficiency of 11.2% under one sun, unbiased, on a 16 cm² artificial leaf. That is a wired photoelectrode module, not a particulate sheet in pure water. This problem does not grade it.
Specification
The reactor holds pure water with a resistivity of at least 1 MΩ·cm at the start. No buffer, salt, redox mediator, or sacrificial donor is added. The photocatalyst is a powder, a printed sheet, or a film of particles, in that one vessel. No electrical bias is applied. The headspace is 1.00 ± 0.05 bar. The water is not externally heated above 40 °C. The water temperature is recorded.
Illumination is continuous for 6 hours. The spectrum is AM1.5G. The irradiance at the sample plane is 100 ± 5 mW cm⁻², measured the same day with a calibrated thermopile or reference cell. The illuminated geometric area is at least 1.00 cm² and is fixed before the run.
Hydrogen and oxygen are quantified by gas chromatography against a calibration mixture. The amount of each gas uses the measured headspace volume or the measured outlet flow, not a bubble count. Solar-to-hydrogen efficiency is
(moles of H₂) × 237 kJ mol⁻¹ / (incident energy in kilojoules),
expressed as a percent. The 237 kJ mol⁻¹ is the Gibbs energy of H₂O(l) → H₂ + ½ O₂ at 298 K, fixed here so that every attempt uses the same numerator. A paper that used a different numerator is not recomputed.
The graded efficiency is the value over the full 6 hours. The efficiency over the first hour and over hours 2 through 6 are both reported. Hours 2 through 6 must be at least 80% of the first-hour value.
Pass threshold
The 6-hour solar-to-hydrogen efficiency is at least 5.0%. The molar ratio of hydrogen to oxygen over those 6 hours is between 1.90 and 2.10. The dark control and the no-catalyst control pass. The efficiency over hours 2 through 6 is at least 80% of the first hour. The water stays at or below 40 °C, and the headspace stays at 1 bar.
Five percent, the ratio window, the 80% hold, 40 °C, 1 bar, 1.00 cm², and 6 hours are pass marks defined here. Five percent is above the 1.0%, 1.1%, and 1.2% sheet results cited above, and those results were not all taken at 1 bar and at or below 40 °C.
Controls
- A dark run of the same reactor, same catalyst, and same duration, with the lamp off, produces less than 5% of the hydrogen measured in the illuminated run.
- An illuminated run with no photocatalyst, same water and same lamp, produces less than 5% of that hydrogen.
- The hydrogen-to-oxygen ratio is inside the window above. Oxygen missing, or hydrogen from a sacrificial donor, fails the attempt.
- The irradiance measurement is reported. A nominal lamp setting with no measurement at the sample plane fails the attempt.
Allowed
Any particulate semiconductor, cocatalyst, or sheet architecture, including a conductive underlayer that is not connected to a power supply. A suspension stirred in one vessel. A solar simulator.
Not allowed
An applied voltage or a wired photoanode and photocathode. A photovoltaic cell plus an electrolyser. A sacrificial donor, including methanol, triethanolamine, or silver ion. A redox mediator that carries charge between two vessels. A hydrohalic acid as the hydrogen source. A pressure below 0.95 bar. External heating of the water above 40 °C. Quoting an apparent quantum yield in place of the solar-to-hydrogen efficiency. An illuminated area chosen after the run, or an area below 1.00 cm². Using 286 kJ mol⁻¹, or any numerator other than 237 kJ mol⁻¹, in the graded efficiency.
Partial success
The 6-hour solar-to-hydrogen efficiency is at least 2.0% under the same water, pressure, temperature, area, ratio, and control rules, and hours 2 through 6 hold at least 80% of the first hour. Two percent is a pass mark for this lesser bar. It is above the 1.0–1.2% sheet results and below the 5.0% mark. The 2.47% two-cell iodide system does not meet it, because the electrolyte and the two vessels are excluded. Partial success is not a solution.
Sources
Sheet efficiency of 1.1% for pure water, with the 331 K and 10 kPa condition in the body: Wang, Hisatomi, Domen, and co-workers, Nature Materials, 2016, doi:10.1038/nmat4589 (opens in a new tab). Carbon-conductor sheet at 1.2% at 331 K and 10 kPa, 80% of that efficiency retained at 91 kPa, and the statement that 1.0% was the highest ambient-pressure Z-scheme pure-water result then: Journal of the American Chemical Society, 2017, doi:10.1021/jacs.6b12164 (opens in a new tab). The 2024 review treating sheet efficiencies as above 1% and practical values as about 1% at most: Frontiers in Science, 2024, doi:10.3389/fsci.2024.1411644 (opens in a new tab). Two-cell 2.47 ± 0.03% result with an iodide couple: Fu, Wu, and co-workers, Nature Communications, 2025, doi:10.1038/s41467-025-56314-x (opens in a new tab). Perovskite module at 11.2%: Hansora and co-workers, Nature Communications, 2025, doi:10.1038/s41467-025-59597-2 (opens in a new tab). The 5.0%, 2.0%, 237 kJ mol⁻¹, 40 °C, 1 bar, and 6-hour requirements are pass marks defined here. Open point: a pure-water, one-vessel, one-sun result above 2% at 1 bar may exist and was not found in the 2024 review or the papers opened for this draft. If it does, the partial bar moves.
Polyethylene back to polyethylene
Unfilled polyethylene is converted to isolated ethylene and polymerized back to polyethylene that matches the starting resin in a side-by-side tensile test.Claim
A weighed sample of unfilled polyethylene is converted to isolated ethylene, that ethylene is polymerized back to polyethylene, and the new polymer matches the starting resin in a tensile test run the same day.
Why it matters
Polyethylene is the plastic made in the largest amount, and it has no ester or amide to unzip. Cracking it makes a mixture of olefins. Returning a mixture of ethylene and propylene to a cracker is not the same as handing back the monomer and making the same polymer again. The grade is the isolated ethylene, then a tensile bar of the polymer made from it.
Current best (with sources)
Bi, Chen, Lin, Han, and co-workers report carbon yields of ethylene and propylene together of up to 79% from polyethylene, over two zeolite stages, the first at 260 °C and the second at 540 °C, with no added hydrogen and no added olefin. They state that conventional catalytic cracking keeps the combined ethylene and propylene yield below 25%, and that pulsed-heating routes have reached 36–45% combined. The catalyst remained active over five cycles in their test. The main text reports the sum. It does not give the ethylene share of that 79%, and it does not polymerize the ethylene back to polyethylene or report a tensile test of a remade resin.
A route that co-feeds ethylene to make propylene is a different transformation. This problem does not count propylene, and it does not allow an olefin co-feed.
Specification
The starting material is an unfilled high-density polyethylene, identified by grade. A portion is molded into ASTM D638 bars. At least five bars are tested. The median ultimate tensile strength and the median elongation at break are the reference.
A separate weighed charge of the same resin, at least 1.00 g, is converted in a reactor whose carrier gas is nitrogen or argon. No hydrogen and no olefin is co-fed. The hottest zone is recorded. Ethylene is isolated as a gas or a condensed liquid, not left as a peak in a chromatogram of a mixed stream. The isolated ethylene is at least 95 mol% pure by gas chromatography against a calibration mixture. Carbon in that isolated ethylene is the amount of isolated gas, from its mass or from its measured volume, pressure, and temperature, multiplied by the ethylene mole fraction from that chromatogram. That carbon is divided by the carbon charged in the polyethylene. The graded yield is this carbon yield, not the gross mass of an impure cut.
Identified products, including propylene, alkanes, and residue, are quantified. Together with the isolated ethylene they account for at least 90% of the carbon charged, or the run is invalid.
The isolated ethylene, and no other monomer, is polymerized by a written procedure to polyethylene. The product is molded into bars of the same geometry as the starting resin and tested on the same machine, the same day, at least five bars. The clock from charging the polyethylene to the end of that tensile test is at most 14 days.
Pass threshold
Carbon in the isolated ethylene is at least 80% of the carbon charged. The remade polymer’s median ultimate tensile strength and median elongation at break are each at least 80% of the starting resin’s medians. Purity, carbon closure, and the 14-day clock hold.
Eighty percent carbon yield, 80% of each tensile median, 95 mol% purity, 90% carbon closure, 1.00 g, and 14 days are pass marks defined here. Eighty percent as ethylene alone is above the reported 79% combined ethylene-plus-propylene yield, so it is above either share of that sum. The tensile comparison is not a number that paper reports.
Controls
- A reactor with no polyethylene, taken through the same temperature program and the same carrier flow, produces ethylene amounting to less than 5% of the ethylene isolated from the real charge. If it does not, the ethylene did not come from the polymer.
- The starting resin and the remade polymer are molded to the same thickness and pulled at the same speed. ASTM D638 warns that preparation, speed, and environment change the comparison. Five specimens and the median, rather than the best bar, are the pass rule defined here.
- Gas chromatography of the isolated ethylene is calibrated the same day with a mixture of known ethylene content.
Allowed
Any catalyst, including zeolites, at any loading. Any temperature the apparatus can hold, including a stage near the 540 °C cited above. A laboratory polymerization of the isolated ethylene, including a Ziegler–Natta or metallocene procedure in a rated pressure vessel. Nitrogen or argon as the carrier.
Not allowed
An olefin co-feed, a hydrogen co-feed, or any carbon-containing feed other than the polyethylene and the catalyst. Counting propylene, butenes, or a mixed C₂–C₃ stream as ethylene. A chromatogram yield of ethylene that is not isolated and then polymerized. Polymerizing purchased ethylene. Quoting a literature tensile strength for polyethylene instead of the side-by-side test. Fillers or fibers in the remade bars that the starting bars do not contain. A charge below 1.00 g.
Partial success
Carbon in the isolated ethylene is at least 40% of the carbon charged, the isolated ethylene is at least 95 mol% pure, and the remade polymer’s median ultimate tensile strength is at least 80% of the starting median. Elongation need not meet the 80% comparison. Forty percent and the strength comparison are pass marks for this lesser bar. They are not a solution. The main text of the 79% paper does not report this remake. If the ethylene share of that 79% is later read off and is already above 40%, the yield half of this lesser bar is not beyond that paper, and the tensile half is the part that remains.
Sources
Combined ethylene and propylene yields up to 79%, the statement that conventional cracking is below 25% combined, the 36–45% pulsed-heating range, and the stage temperatures 260 °C and 540 °C: Bi, Chen, Lin, Han, and co-workers, Nature Chemical Engineering, 2025, doi:10.1038/s44286-025-00290-y (opens in a new tab). ASTM D638-22, astm.org/d0638-22.html (opens in a new tab), for the tensile method and the warning that comparative tests have to control preparation, speed, and environment. The 80% isolated-ethylene yield, the 40% partial yield, the tensile ratios, the 95 mol% purity, and the 14-day clock are pass marks defined here. Open point: the ethylene-versus-propylene split of the 79% is not stated in the main text opened for this draft, so this problem does not assign an ethylene yield to that paper.