Perovskite solar cells have a defect problem with two faces, and until now, fixing both faces at once has been chemically self-defeating. A team spanning Northwestern University and the Daegu Gyeongbuk Institute of Science and Technology (DGIST) reports a way around the contradiction: a pair of passivating molecules engineered to ignore each other while each does its own job on the perovskite surface. Their champion inverted single-junction cell reaches 27.4% power conversion efficiency, with a certified quasi-steady-state efficiency of 25.8% and more than 1,000 hours of operation at 85 °C before falling to 90% of its starting efficiency.
The result, described in a manuscript from the group of Bin Chen and Edward H. Sargent at Northwestern with Jongmin Choi’s group at DGIST, is a chemistry result first and an efficiency result second. The efficiency matters — 27.4% is near the top of what single-junction perovskites have shown — but the mechanism is what other laboratories can actually reuse: a modular way to treat chemically diverse defects instead of choosing which half of the defect population to leave alone.
The problem: two defect polarities, one self-defeating mixture
A perovskite film is an ionic crystal, and its surface is where the crystal terminates imperfectly. Missing ions, undercoordinated metal centers, and atoms sitting on the wrong lattice sites all create charged point defects that act as trap states: they capture electrons or holes and let them recombine without producing current. These non-radiative recombination centers eat voltage and efficiency, and they come in both polarities. Undercoordinated lead ions (Pb²⁺) are cationic, electron-deficient sites; iodine interstitials and lead-iodide antisite defects are electron-rich, anionic ones.
Chemists have long had a tool for each polarity. Lewis bases — molecules that donate an electron pair — coordinate to the electron-deficient cationic sites. Lewis acids — molecules that accept electron density — address the electron-rich anionic sites. Used individually, each works. Used together, they fail, for the most elementary reason in chemistry: an acid and a base react with each other. Mixed in solution or on a surface, they form an adduct that quenches both passivators before either reaches its defect. The standard workaround has been to pick one polarity and accept the other defect population, or to design a single multifunctional molecule carrying both characters, which the authors note tends to suffer from rigid internal charge structure, uneven surface coverage, and inconsistent performance.
The new paper’s move is to stop trying to make acids and bases coexist by weakening their chemistry, and instead make them coexist by gating their geometry. The authors call it chemical orthogonality, and they define it precisely: the mutual interaction energy between the acid and the base must be substantially weaker than the binding energy of each passivator to its own target defect, so that solution-phase association between the two is weak and reversible, and each remains independently reactive toward its defect site when the film is deposited.
The gating design: a directional acid meets a bulky base
The design has two complementary halves, and neither works without the other.
The acid half is a family of fluorinated aromatic iodides: iodobenzene, 4-iodobenzotrifluoride, 1-iodo-3,5-bis(trifluoromethyl)benzene (IBTFMB), and iodopentafluorobenzene (IPFB). Fluorination strengthens the σ-hole — a region of positive electrostatic potential on the far side of the iodine atom, opposite the carbon-iodine bond — which makes the molecule a halogen-bond donor toward electron-rich defects. But the σ-hole is highly directional: it demands a precise, near-linear alignment between the acid’s iodine and the base’s lone pair before orbital overlap can occur. That directionality is a constraint the designers exploit.
The base half is a family of sp³-hybridized phosphines: tris(diethylamino)phosphine (TDEP), tris(1-pyrrolidinyl)phosphine (TPP), tris(dimethylamino)phosphine (TDMP), and trimethylphosphine (TMP). A phosphine’s lone pair sits in an umbrella-like geometry, and the substituents around the phosphorus form a steric pocket that blocks the one approach geometry the σ-hole requires, suppressing the acid-base orbital overlap that would form an adduct. The asymmetry is the point: a defect site on the perovskite surface is sterically accessible from many approach angles and imposes no such strict directional requirement, so the acid and the base, which cannot easily find each other, can each easily find their defects.
Screening across the four acids and four bases, with forty devices per combination, showed that the details matter and that bigger is not always better. The bulkiest phosphines, TDEP and TPP, had the strongest intrinsic donor character by calculation but performed worse in devices, because their bulk also impeded access to defect sites. The most compact base, TMP, paired with the strongest acid, IPFB, over-associated — individually strong passivators that lacked sufficient steric separation and quenched each other after all. The winning region was the balanced middle: IBTFMB as the acid and TDMP as the base, at a 3.5:1 molar ratio (17.5 mM IBTFMB, 5 mM TDMP).
The spectroscopic evidence for non-interaction is direct rather than inferred from device performance. In phosphorus-31 NMR, the IBTFMB–TDMP mixture showed a chemical shift within 0.01 ppm of the free base — essentially no electronic perturbation, no persistent adduct. As a positive control, the authors forced contact at high concentration in an equimolar mixture and watched the phosphorus resonance move from 122 ppm to a cluster between 50 and 70 ppm, the signature of the tight adduct that the working pair avoids.
The evidence: complementary trap suppression, measured twice over
A passivation claim that rests only on a device curve is a claim about one cell. This paper’s case rests on defect profiling, and the profiling shows the two passivators doing different jobs — which is exactly what orthogonality predicts.
Drive-level capacitance profiling found trap densities concentrated near the charge-transport interfaces, as expected for surface-dominated losses. After the combined treatment, the trap density at the electron-transport-layer/perovskite interface fell by roughly two orders of magnitude, from 10¹⁸ to 10¹⁶ cm⁻³ at 5 kHz. Thermal admittance spectroscopy, which resolves traps by energy depth, showed distinct suppression patterns: the acid broadly reduced the trap spectrum, particularly halide-related shallow states and antisite-derived deeper states, while the base more effectively suppressed the deepest trap band, consistent with passivation of undercoordinated Pb²⁺ and cation-interstitial defects. The perovskite surface in these films is iodine-poor, so the base does the heavier lifting — the acid’s targets are less abundant — but the acid’s contribution is what makes the suppression broad rather than partial.
There is a third step in the recipe, mechanistically distinct from the pair. After the acid-base treatment, the team applies propane-1,3-diammonium iodide (PDAI₂), a field-effect passivator previously reported by the same group. PDAI₂ does not coordinate to defects; it introduces a surface dipole field that repels minority carriers from the interface, and X-ray photoelectron spectroscopy confirmed it does not perturb the acid-base coordination equilibrium.
The device numbers, handled in order
The devices are inverted (p-i-n) single-junction cells with the structure FTO/NiOₓ/2PACz–Me-4PACz/perovskite/C₆₀/BCP/Ag, measured through a 0.049 cm² aperture mask. That area is the first honesty constraint on the headline: this is a small laboratory cell, not a module, and every number below belongs to that scale.
The champion cell from the combined treatment reached 27.4% on a current-voltage sweep, with a short-circuit current density of 26.4 mA/cm², an open-circuit voltage of 1.187 V, and a fill factor of 87.43%. Three different efficiency figures circulate for this device, and they are not contradictory — they are different protocols. The 27.4% is the champion sweep. The same device held 26.7% under in-house maximum-power-point steady-state tracking. And under an asymptotic quasi-steady-state protocol, the devices achieved a certified 25.8%. The gap between a champion sweep and a certified quasi-steady-state value is normal: the certified protocol is deliberately the hardest to flatter, holding the device at maximum power until the output stabilizes rather than sweeping quickly past it. Quote 25.8% for the conservative number and 27.4% for the champion; neither is a module specification.
The stability result is the more consequential number for anyone thinking beyond the laboratory. Under the ISOS-L-3 protocol — continuous one-sun illumination at 85 °C in ambient air, tracked at maximum power point — the treated devices reached T90, the time to fall to 90% of initial efficiency, after 1,030 hours, against 423 hours for the control. Two caveats travel with that figure. First, the stability-test devices started at 23.6% initial efficiency, not the 27.4% champion value; the 1,030-hour clock ran on representative devices, not the record cell. Second, the test is a hot, illuminated, ambient-air stress test, not a damp-heat test: no 85 °C/85% relative-humidity data appears in the manuscript text consulted, so nothing here speaks to the moisture-driven degradation that has historically been perovskites’ hardest failure mode.
Why a modular framework matters more than one record
The paper’s discussion makes a claim larger than its champion cell, and it is worth taking seriously because it is testable. Earlier bimolecular approaches, the authors argue, always ended up addressing only one defect polarity, because introducing both an acid and a base meant complexation. Multifunctional single molecules addressed both polarities but with internal competition and uneven coverage. Steric-electronic gating decouples the two problems: choose an acid family and a base family that are orthogonal to each other, then tune each independently against its own defect population. In principle, the same framework extends beyond this acid-base pair to other defect chemistries and compositions — the design rule is about relative interaction energies and approach geometry, not about IBTFMB and TDMP specifically.
That generality is also where the result connects to the stability thread this desk has followed through October at the single-junction level: recent coverage has chased perovskite durability through interface chemistry, capturing mobile ions before they reach reactive layers and engineering crystallization to leave fewer defects behind. This paper attacks the same problem from the opposite side — accept that both defect polarities will exist on a real surface, and treat both simultaneously without letting the treatments cancel each other.
It is also worth placing the result against the field’s commercial backdrop, covered last week in the Caelux–INA Solar agreement: gigawatt-scale perovskite-on-glass partnerships aimed at 2028 production. That story is about manufacturing bets and targets; this one is about laboratory interface chemistry. The laboratory result does not shorten the manufacturing road: a 0.049 cm² cell with a three-step surface treatment still has to survive translation to large-area coating, where uniformity, throughput, and yield decide whether a treatment is a process or a curiosity.
What this is not: one stack, one scale, and an open certification question
The honest limits are specific. First, the demonstration is one device stack — an inverted cell on NiOₓ with a carbazole-phosphonic-acid hole-selective bilayer — and one perovskite composition family. Whether the same acid-base pair is orthogonal on other surfaces, with different defect energetics, is untested here. Second, the scale is laboratory: a 0.049 cm² aperture, spin-coated treatments, thirty devices per condition for the control comparison. Nothing here addresses module-area uniformity. Third, the stability evidence covers one stress axis — heat plus illumination in air — and not humidity, thermal cycling, or outdoor exposure. Fourth, the certification deserves a footnote of its own: the manuscript text renders the certifying body as “National Laboratory of the Rockies (NLR),” a name that does not correspond to any established certification laboratory this desk recognizes, so the certified figure should be quoted as the manuscript’s reported value pending clarification. The in-house steady-state value of 26.7% is the more directly checkable number in the meantime.
What to watch
Generalization comes first. The first follow-up that applies a sterically gated acid-base pair to a different perovskite composition — a wide-bandgap top cell for tandems, for instance — and reports comparable trap suppression will show whether the framework is a design rule or a one-pair trick.
A damp-heat result would convert the stability question from a single-axis claim into a commercial one. The ISOS-L-3 result at 85 °C is genuinely demanding, but perovskite commercialization has always been decided substantially by moisture, and a treatment that holds up under heat and light still has to prove itself under humidity.
Scale-up data — slot-die or blade-coated versions of the treatment, or any large-area demonstration — would answer the process question, since spin-coating a two-component solution plus a sequential third step is straightforward in a laboratory and nontrivial at meter scale. And the certification record deserves a follow-up of its own: a named certifier and a published certificate for the quasi-steady-state value would remove the one oddity in an otherwise carefully documented paper.
The bottom line
A Northwestern–DGIST team has shown that two defect passivators with opposite chemical characters can work on the same perovskite surface without destroying each other — not by weakening their chemistry, but by gating their geometry, so that a directional fluorinated aromatic acid and a bulky phosphine base each reach their own defects while being unable to reach each other. The champion inverted cell reaches 27.4%, holds 26.7% under steady-state tracking, carries a reported certified quasi-steady-state 25.8%, and survives 1,030 hours at 85 °C under illumination before falling to 90% of its starting efficiency, against 423 hours for the control.
It is a laboratory result on a 0.049 cm² cell, on one stack, with humidity stability untested and one certification detail unclear. It is also a reusable design rule: make the acid-base interaction weaker than each passivator’s defect binding, and enforce that ordering with sterics. The number to watch is not 27.4%. It is the second pair — the first acid-base combination chosen by this logic, on a different composition, that passivates both defect polarities at once.
No investment recommendation is made or implied anywhere in this article.
Sources and further reading
- Shin, D., Kitade, S., Duan, L., Liu, C., Li, C., et al. “Sterically gated Lewis acid and base pairs enable orthogonal defect passivation in perovskite solar cells.” Manuscript hosted by the U.S. Department of Energy’s Office of Scientific and Technical Information (OSTI) — the primary source for the mechanism (steric-electronic gating, σ-hole directionality, IBTFMB–TDMP pair at 3.5:1), the device figures (27.4% champion, 26.4 mA/cm², 1.187 V, 87.43% fill factor, 26.7% in-house steady-state, certified quasi-steady-state 25.8%), the trap-density reduction (10¹⁸ to 10¹⁶ cm⁻³ at the electron-transport interface), the device stack (FTO/NiOₓ/2PACz–Me-4PACz/perovskite/C₆₀/BCP/Ag, 0.049 cm² aperture), and the stability result (ISOS-L-3, 85 °C, 1,030 hours to T90 vs. 423 hours for the control, from 23.6% initial efficiency). Corresponding authors Bin Chen and Edward H. Sargent (Northwestern); Jongmin Choi (DGIST). No journal, volume, or DOI is stated in the manuscript text consulted, and none is claimed here.
- Perovskite-Info (October 27, 2026), “Perovskite solar cells reach 27.4% by repairing two types of defects at the same time” — trade coverage of the same result, the source for the Northwestern–DGIST framing; it carries no journal citation and no stability data, which is why this article works from the manuscript directly.
- Excitonic.com, “INA Solar’s 3 GW Caelux Deal and the Perovskite-on-Glass Bet” — the recent perovskite manufacturing coverage this article distinguishes itself from: a commercial announcement, not a laboratory interface result.
A note on sourcing: the manuscript text consulted renders the certifying laboratory as “National Laboratory of the Rockies (NLR).” No such certification laboratory is known to this desk, and the certified 25.8% figure is therefore reported here as the manuscript’s value, not as an independently confirmed certification. No damp-heat (85 °C/85% relative humidity) data, control-device efficiency, or module-area result is stated, because none appears in the manuscript text consulted. Background on perovskite defect chemistry, Lewis acid-base passivation, and ISOS stability protocols is drawn from the standard perovskite literature.