A near-autonomous AI chemist improves a challenging reaction in medicinal chemistry
Two titles, both real. The heading above is how the lab announced this work. The document actually behind it is titled “TEMPO Improves Generality and Decreases Oxidative Deboronation in Chan–Lam Couplings of Primary Sulfonamides” — everything below is read from that document.
A 10,080-reaction robotic screen found that adding TEMPO to Chan–Lam couplings of primary sulfonamides raises yield and cuts a competing side reaction, nearly doubling the share of high-yielding reactions.
It gives medicinal chemists a cheap, validated fix for a coupling reaction common in drug synthesis that has long suffered from boronic-acid degradation.
Jan Rzymkowski · Shuyuan Zhang · Artur Chołuj · Aleksander Szkółka · Mateja Dud · Mateusz Bruno-Kamiński · Jan Busz · Michał Sadowski · Grzegorz Wojciechowski · Jan Kulczycki · Mariusz Gruza · Tadija Radusinović · Maria Wyrzykowska · Szymon Kapuściński · … — meet the researchers →
Two readings, equal authority
How to choose: The paper’s words is verbatim — use it when you need to quote, or to judge how they write. Plain language is a paraphrase written for comprehension — use it when you want the idea fast. Neither is a summary of the other; they are two doors into the same room.
“Primary sulfonamides are valuable motifs in medicinal chemistry but remain challenging substrates for Chan–Lam N-arylation because of low nucleophilicity and boronic acid degradation. Here, we report a high-throughput study of TEMPO-promoted Chan–Lam coupling between primary sulfonamides and arylboronic acids. Across two microscale screening campaigns comprising 10,080 reactions, we evaluated oxidant identity, oxidant loading, copper source and loading, base, solvent, temperature, and substrate structure. Stoichiometric TEMPO emerged as a surprising yet uniquely effective additive, improving desired C–N bond formation while suppressing oxidative deboronation relative to oxidant-free and most of the strongly oxidizing conditions. Under the optimized condition, using 2 equivalents of TEMPO and 20 mol% Cu(OAc)₂, the mean estimated product yield increased to 25.2% (from 16.6%), and the fraction of reactions exceeding 30% yield increased over twofold to 37.5% (from 15.6%). Bench-scale validation confirmed the beneficial effect of TEMPO in eleven of fourteen representative substrate pairs, with LC-PDA-MS yields improving over twofold in the majority of cases. Notably, we observe consistent gains for electron-poor boronic acids across the high-throughput campaign and the bench-scale validation. Evaluation of structurally related additives showed that 4-hydroxy-TEMPO maintained comparable performance, offering a potentially lower-cost and more readily removable alternative to TEMPO. These results identify aminoxyl additives as a surprising tool for improving primary sulfonamide Chan–Lam couplings, with potential application on the industrial scale.”
Primary sulfonamides are useful building blocks in drug design but are hard to attach to aryl groups because the boronic acid coupling partner tends to degrade during the reaction. The authors ran a 10,080-reaction robotic screening campaign and found that adding the mild, stable radical TEMPO — instead of a strong oxidant — both improved the desired coupling and reduced that degradation. A bench-scale follow-up confirmed the effect, and a cheaper TEMPO variant worked almost as well.
What this paper defines
Every definition below is the paper’s own sentence, with its locator. The plain gloss is a reading aid and is marked as one.
Chan–Lam coupling
“now widely known as the Chan–Lam (Chan–Lam, CL) coupling, enables the formation of C–N, C–O, and C–S bonds through the copper-catalyzed cross-coupling of organoboron reagents (such as boronic acids, pinacol esters, stannanes, or siloxanes) with various N–H, O–H, and S–H containing nucleophiles.”1. Introduction
In plain terms: A copper-catalyzed reaction that links a boron-containing reagent to a nitrogen-, oxygen-, or sulfur-containing partner to form a new bond.
oxidative deboronation
“oxidative deboronation is a key side reaction, in which oxidative cleavage of the carbon–boron bond generates phenolic byproducts.”1. Introduction
In plain terms: An unwanted side reaction where the boron-containing reagent breaks down into a phenol instead of forming the desired product.
robustness (screening criterion)
“we prioritized robustness across the substrate set, defined as the fraction of reactions giving >30% estimated yield.”3. First High-throughput Campaign: Oxidant Screen
In plain terms: How the authors judged a condition as 'good': not by average yield, but by what fraction of many different substrate pairs cleared a 30% yield bar.
TEMPO
“TEMPO (2,2,6,6-tetramethylpiperidinyloxyl) was selected among others as a promising candidate because it is a mild, shelf-stable radical oxidant known to modulate Cu redox chemistry, although its precise role in Chan–Lam C–N coupling remains unresolved.”1. Introduction
In plain terms: A stable, mild radical compound already known to help copper catalysts turn over in other reactions, tested here as an additive for this coupling.
What they actually did
Each step is a synthesis. Open any step to see the paper’s own sentence it was derived from, with its locator — so nothing here floats free of the source.
- The team used a previously published high-throughput dataset on the same substrate class to set the starting baseline reaction conditions.
Trace this step to the paper
“Gandhi, Brown, Doyle, and co-workers recently generated a 3,904-reaction high-throughput dataset to identify effective conditions for primary sulfonamide Chan–Lam N-arylation”1. Introduction
- They designed a substrate matrix of 12 primary sulfonamides and 8 boronic acids (96 pairs) and ran 10,080 reactions across two microscale HTE campaigns.
Trace this step to the paper
“Across two HTE campaigns, we performed 10,080 microscale Chan–Lam reactions covering 12 primary sulfonamides and 8 boronic acids, which gave 96 unique substrate-pair combinations in each HTE campaign (Figure 1).”2. Experimental Description
- Reactions were run at microliter scale in 96-well plates with automated liquid handling and quantified by LC-PDA-MS against an internal standard.
Trace this step to the paper
“All reactions were performed on a microliter scale in 96-well plates using automated liquid-handling workflows. The reactions were run at a 14 mM concentration and a total reaction volume of 28.6 µL, enabling broad exploration of chemical space while minimizing material consumption.”2. Experimental Description
- In the first HTE campaign, they screened 10 different oxidants across 53 condition sets and judged conditions by how many substrate pairs cleared a 30% yield threshold, rather than by average yield alone.
Trace this step to the paper
“we prioritized robustness across the substrate set, defined as the fraction of reactions giving >30% estimated yield.”3. First High-throughput Campaign: Oxidant Screen
- This first screen identified TEMPO with potassium carbonate in a DMA/diglyme solvent mix at 60 degrees C as the best-performing condition.
Trace this step to the paper
“the best-performing condition used TEMPO (1 eq.), K₂CO₃ (2 eq.), and DMA/diglyme (3:7, v/v) at 60 °C for 18 h.”3. First High-throughput Campaign: Oxidant Screen
- A second HTE campaign of 4,992 reactions then optimized TEMPO loading, copper source/loading, temperature, base, and solvent, and also tested structural analogues of TEMPO.
Trace this step to the paper
“the second high-throughput campaign was designed to optimize the TEMPO-promoted N-arylation. Conditions varied in TEMPO loading (from 0.25 to 2 equivalents), copper catalyst source and loading (from 2.5 mol% to 40 mol%), as well as temperature, base and solvent.”5. Second High-throughput Campaign: TEMPO Optimization
- That optimization selected a final condition of 2 equivalents of TEMPO with 20 mol% copper acetate.
Trace this step to the paper
“The selected combination used 2 equivalents of TEMPO and 20 mol% Cu(OAc)₂”5. Second High-throughput Campaign: TEMPO Optimization
- Four substrate pairs were advanced to bench (milligram) scale to test whether the HTE-observed TEMPO benefit reproduced under more practical conditions.
Trace this step to the paper
“Four substrate pairs were advanced to bench-scale validation to test the reproducibility of the TEMPO effect observed in the HTE campaign.”4. Bench-scale validation of TEMPO-promoted conditions compared to optimized copper loading without TEMPO
- A further bench-scale test paired ten different boronic acids with two sulfonamides to see how boronic-acid electronics affect the TEMPO benefit.
Trace this step to the paper
“An additional test intended to more systematically examine the influence of boronic acid identity was performed with 20 mol% copper loading for both variants (Table 2).”4. Bench-scale validation of TEMPO-promoted conditions compared to optimized copper loading without TEMPO
- They tested TEMPO-related aminoxyl additives and a non-radical piperidine control to check whether the benefit depended specifically on TEMPO's radical structure.
Trace this step to the paper
“We tested TEMPO, TEMPO-derived aminoxyl radicals: 4-hydroxy-TEMPO, 4-oxo-TEMPO, and 1,2,2,6,6-pentamethylpiperidine (PMP) as a sterically related non-radical piperidine control (Figure 7).”5.2. Evaluation of TEMPO-Related Additives
Exactly what was run, and how
What they reported — and what they left out
This is a wet-lab chemistry paper; despite OpenAI's co-authorship, the text describes no AI/ML model, so there are no model settings to report.
The numbers they report
In the first screen, TEMPO raised both the hit rate and the mean yield relative to no added oxidant, at the same copper loading.
no-oxidant: 9.9% of reactions >30% yield, mean 12.2%; with TEMPO: 26.6% hit rate, mean 19.6%
See it in the paper
“the reaction performed without an added oxidant gave only 9.9% of reactions above the 30% yield threshold, with a mean estimated yield of 12.2%. Adding TEMPO improved both metrics, increasing the hit rate to 26.6% and the mean estimated yield to 19.6% (Figure 2A).”3. First High-throughput Campaign: Oxidant Screen
Bench-scale tests confirmed the TEMPO benefit in most of the four validation pairs, with a large improvement in three of them.
3 of 4 pairs improved, up to a four-fold increase
See it in the paper
“in three out of four cases, addition of TEMPO significantly increased the product/IS ratio relative to the no-oxidant control, with up to a four-fold improvement.”4. Bench-scale validation of TEMPO-promoted conditions compared to optimized copper loading without TEMPO
In the one bench-scale pair where TEMPO did not clearly help, results were comparable to the no-oxidant control.
product/IS ratio 2.12 (TEMPO) vs. 2.60 (no oxidant)
See it in the paper
“TEMPO and no-oxidant conditions gave comparable results (product/IS ratio: 2.12 vs. 2.60, respectively).”4. Bench-scale validation of TEMPO-promoted conditions compared to optimized copper loading without TEMPO
Across ten tested boronic acids, TEMPO's benefit was clearest for electron-poor ones, roughly doubling product amount, but absent or reversed for two others.
benefit in 7 of 10 boronic acids tested, roughly twofold average increase
See it in the paper
“The benefit of the TEMPO addition was clearly demonstrated for all electron-poor boronic acids (Entries 1-3, 6-8, 10) and one of electron-rich boronic acids (Entry 9), averaging a roughly twofold increase in the measured product amount, while it was not observed in the remaining two or addition of TEMPO even led to lower relative amount of product (Entries 4 and 5).”4. Bench-scale validation of TEMPO-promoted conditions compared to optimized copper loading without TEMPO
The final optimized TEMPO condition more than doubled the fraction of reactions clearing the 30% yield bar, relative to no oxidant.
15.6% (no oxidant) to 37.5% (optimized TEMPO condition)
See it in the paper
“increasing the percentage of reactions achieving >30% yield from 15.6% under the no-oxidant control to 37.5%.”5. Second High-throughput Campaign: TEMPO Optimization
Raising copper loading alone also helped, but far less than the optimized TEMPO condition did.
See it in the paper
“Increasing copper loading alone also improved performance, but to a much smaller extent, indicating that the benefit of the optimized condition cannot be explained by copper loading alone”5. Second High-throughput Campaign: TEMPO Optimization
At 20 mol% copper, mean yield rose steadily as TEMPO loading increased to 2 equivalents.
16.7% (no TEMPO) to 25.3% (2 eq. TEMPO)
See it in the paper
“This trend is particularly evident at 20 mol% copper loading, where the mean yield increases from 16.7% without TEMPO to 25.3% when paired with 2 equivalents of TEMPO.”5.1. Increased TEMPO loading improves product yield and reduces oxidative deboronation
Increasing TEMPO loading also reduced the oxidative deboronation side reaction at 20 mol% copper.
37.3% (oxidant-free) to 30.2% (2 eq. TEMPO)
See it in the paper
“increasing the TEMPO concentration reduced oxidative deboronation from 37.3% under oxidant-free conditions down to 30.2% at 2 equivalents of TEMPO (Figure 6).”5.1. Increased TEMPO loading improves product yield and reduces oxidative deboronation
A separate side reaction, oxidative homocoupling, stayed minor across conditions.
between 2.2% and 5.9%
See it in the paper
“The mean estimated yield of oxidative homocoupling remains small (between 2.2% and 5.9%).”5.1. Increased TEMPO loading improves product yield and reduces oxidative deboronation
Among structurally related additives, only 4-hydroxy-TEMPO matched TEMPO's performance.
See it in the paper
“Among the tested additives, only 4-hydroxy-TEMPO maintained performance comparable to TEMPO.”5.2. Evaluation of TEMPO-Related Additives
The pyrazole-containing sulfonamide S05 performed notably well despite being highly substituted.
approximately 28% mean yield
See it in the paper
“the performance of the highly functionalized pyrazole derivative S05, giving approximately 28% mean yield, is particularly noteworthy”5. Second High-throughput Campaign: TEMPO Optimization
The ortho-dimethylamino boronic acid B02 performed better than its close analogue B01 under the optimized conditions.
overall mean yield of 21%
See it in the paper
“the ortho-dimethylamino analog B02; despite its electronic similarities to B01, showed improved yields under the optimized TEMPO-containing conditions, with overall mean yield of 21%”5. Second High-throughput Campaign: TEMPO Optimization
Overall, the optimized TEMPO condition raised mean yield and doubled the fraction of high-yielding reactions relative to no oxidant.
mean yield 16.6% to 25.2%; >30%-yield fraction roughly doubled
See it in the paper
“The optimized condition increases mean product yield from 16.6% to 25.2% and doubles the fraction of reactions exceeding 30% yield relative to the no-oxidant control.”Conclusions
Bench-scale validation held up for most of the fourteen representative substrate pairs tested overall.
11 of 14 pairs, LC-PDA-MS yields improving over twofold in the majority of cases
See it in the paper
“Bench-scale validation confirmed the beneficial effect of TEMPO in 11 of 14 representative substrate pairs, with LC-PDA-MS yields improving over twofold in the majority of cases.”Conclusions
The overall campaign is described as the largest Chan–Lam HTE screen reported to date.
10,080 microscale reactions
See it in the paper
“we report the largest high-throughput screening of Chan–Lam coupling of primary sulfonamides, comprising 10,080 microscale reactions across a diverse substrate matrix.”Conclusions
What they assert, beside what they showed
Left is the claim in the paper’s own words. Right is the data offered for it. Where the two do not fully meet, a gold band names the distance.
TEMPO's benefit doesn't come from being a strong oxidant but from creating a milder redox environment that limits boronic-acid decomposition.
“This suggests that the benefit of TEMPO does not come from stronger oxidation, but from providing a mild aminoxyl-mediated redox environment that supports N-arylation without accelerating organoboron decomposition.”
“In contrast, TEMPO decreased the mean oxidative deboronation estimated yield relative to the no-oxidant control while also improving the desired product yield.”
3. First High-throughput Campaign: Oxidant ScreenThe optimized TEMPO condition substantially outperforms the no-oxidant baseline on the key screening metrics.
“The optimized condition increases mean product yield from 16.6% to 25.2% and doubles the fraction of reactions exceeding 30% yield relative to the no-oxidant control.”
“increasing the percentage of reactions achieving >30% yield from 15.6% under the no-oxidant control to 37.5%.”
5. Second High-throughput Campaign: TEMPO OptimizationThe TEMPO improvement generalizes across substrate classes rather than being limited to one type of sulfonamide or boronic acid.
“The improvement is seen for a variety of sulfonamides and boronic acids, hence not limited to only one substrate class.”
“addition of TEMPO (blue) improved the mean yield for the majority of substrates (with the exception of B01 and S04 that appear slightly worse) compared to no-oxidant control (red).”
5. Second High-throughput Campaign: TEMPO Optimization4-hydroxy-TEMPO (TEMPOL) is a viable, cheaper substitute for TEMPO at industrial scale.
“TEMPOL stands as a substantially less expensive alternative for the synthesis of sulfonamide-containing APIs.”
“Among the tested additives, only 4-hydroxy-TEMPO maintained performance comparable to TEMPO.”
5.2. Evaluation of TEMPO-Related AdditivesSimply adding an oxidant is not enough to improve this coupling; the specific oxidant identity matters.
“In conclusion, adding an oxidant is not sufficient to improve primary sulfonamide Chan–Lam coupling.”
“most other oxidants either reduced the overall yield or gave no significant improvement.”
3. First High-throughput Campaign: Oxidant ScreenHow they frame it, and what they want next
Their framing
The authors frame TEMPO's effect as counterintuitive: it is a much milder oxidant than the alternatives typically used, yet it outperforms them. They repeatedly tie the yield gain to suppression of a specific competing pathway (oxidative deboronation) rather than to raw oxidizing strength, and close by pitching the result as industrially relevant through a cheaper TEMPO analogue that is easier to remove from the product.
Register: The paper alternates between confident causal statements ('TEMPO behaved differently: it improved C–N bond formation while reducing oxidative deboronation') and explicit hedges around mechanism and generality, most notably flagging its own mechanistic explanation as unconfirmed.
Where they hedge
“though further mechanistic studies are required to confirm this hypothesis.”Our contributions
“the study evaluated only a limited number of examples within any given nucleophile class”1. Introduction
“Although modest in absolute terms, this improvement is meaningful given the breadth of the screen, which included 96 distinct substrate pairs.”3. First High-throughput Campaign: Oxidant Screen
What they say it means
- Mild aminoxyl radicals like TEMPO may have a broader, underexplored role as redox mediators in copper-catalyzed cross-coupling beyond this specific reaction.
the paper’s words
“These data suggest a previously underexplored role for aminoxyl radicals in primary sulfonamide Chan–Lam coupling.”Conclusions
- Because the TEMPO benefit holds even at low copper loading, the protocol may suit process settings that need to minimize residual metal in the product.
the paper’s words
“The beneficial effects of TEMPO on reaction yield persist even at reduced copper loadings (2.5 mol% to 10 mol%), underscoring its potential utility in process chemistry settings where minimizing residual transition-metal content is often desirable.”Conclusions
What they call for next
- Further mechanistic work is needed to confirm that TEMPO's benefit really comes from suppressing boronic-acid decomposition.
the paper’s words
“though further mechanistic studies are required to confirm this hypothesis.”Our contributions
- A more exhaustive optimization of solvent and base choice is left for future work rather than completed here.
the paper’s words
“We decided to keep solvent and base intact in the optimized conditions (Figure 1c) and move a more exhaustive optimization of solvent and base to future work.”5. Second High-throughput Campaign: TEMPO Optimization
Limitations they state
“though further mechanistic studies are required to confirm this hypothesis.”Our contributions
“with the exception of B01 and S04 that appear slightly worse”5. Second High-throughput Campaign: TEMPO Optimization
“Two sulfonamides and one boronic acid were replaced in the second screen because the material was exhausted.”5. Second High-throughput Campaign: TEMPO Optimization
Moves worth stealing
States the paper's contributions as an explicit numbered list right after the introduction, each framed as a novelty claim.
“Our contributions are as follows:”
Opens the abstract with the practical problem (why this substrate class is hard) before naming the fix, so the reader feels the need before the solution.
“Primary sulfonamides are valuable motifs in medicinal chemistry but remain challenging substrates for Chan–Lam N-arylation because of low nucleophilicity and boronic acid degradation.”
Frames its own central result as counterintuitive relative to field assumptions, repeating the word 'surprising' in both the abstract and the conclusions.
“Stoichiometric TEMPO emerged as a surprising yet uniquely effective additive”
Where else this leads
Published alongside it
The nearest publications in time, across all three labs.
- Introducing LifeSciBench OpenAI
2026-06-17 - Diffuse AI Control on Fuzzy Tasks Anthropic
2026-06-15 - Artificial Minds, Human Disagreement: The Politics of AI Consciousness Google DeepMind
2026-06-15 - From AGI to ASI Google DeepMind
2026-06-12
What this page was built from
The corpus manifest listed this as an OpenAI blog post about an 'AI chemist,' but the supplied text is the full underlying wet-lab chemistry manuscript ('TEMPO Improves Generality and Decreases Oxidative Deboronation in Chan–Lam Couplings of Primary Sulfonamides'), whose own title is used here per instruction; the manuscript text is complete (abstract through references) but contains no explicit publication date or standalone URL/DOI for itself, so pub_date and source_url are carried over from the corpus manifest's announcement-page metadata; one author's surname is printed as 'WłodarczykPruszyński' with no separating hyphen (likely 'Włodarczyk-Pruszyński') and is preserved verbatim rather than corrected.