The strategy hinges on an electrolyte-gated transistor, a device in which the conventional insulating gate dielectric is replaced by an ionogel, here a network of poly(vinylidene fluoride-co-hexafluoropropylene) swollen with the ionic liquid 1-ethyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide. When a voltage is applied, ions crowd against the semiconductor surface and form an electric double layer just nanometers thick. That extreme thinness translates into enormous capacitance, greater than 1 microfarad per square centimeter at the carbon-nanotube gate electrode, and interfacial electric fields on the order of megavolts per centimeter. Unlike ordinary electrostatic gating, which merely accumulates carriers at the interface, fields this intense can cross the electrochemical threshold for electrolysis of the trace water dissolved in the ionic liquid, generating reactive oxygen species at the SnO surface.

Those reactive species are the agents of transformation. Applying a doping gate voltage of −1.5 volts drove oxygen into the near-interface region of the film, oxidizing tin from the 2+ to the 4+ state and converting p-type SnO into n-type SnO2. X-ray photoelectron spectroscopy tracked the chemistry: the ratio of the Sn4+ to Sn2+ peak areas rose from an average of 0.47 in the pristine film to 1.87 after doping. Cross-sectional high-resolution transmission electron microscopy then revealed a crystalline phase at the electrolyte interface with lattice spacings of 3.7 and 2.4 angstroms, matching the (110) and (200) planes of orthorhombic SnO2. A coherent crystallographic alignment between the SnO (002) plane near the substrate and the SnO2 (200) plane suggests the conversion proceeds by oxygen incorporation rather than by wholesale diffusion of the heavier tin atoms.

Depth-resolved electron energy loss spectroscopy at the oxygen K-edge sealed the case for localized conversion. Spectra acquired near the ionogel interface showed a peak roughly 6 electron volts above the first oxygen feature, the fingerprint of SnO2, while spectra from the film interior showed a narrower 3.5-electron-volt separation characteristic of SnO. The conversion is therefore confined to the region the electric field can reach, which matters for device design: in the converted transistors, electron transport is expected to run through the electrochemically formed SnO2 surface layer, even though the underlying SnO may still contribute to the total channel current.

The electrical results show that both polarities perform well at sub-1-volt operation. The p-type SnO transistor, measured at a drain voltage of −0.4 volts, delivered a maximum drain current of 1.1 microamperes at a gate voltage of −0.2 volts, a saturation mobility of 0.17 square centimeters per volt-second, an ON/OFF current ratio of 1.38 × 10^4 and a subthreshold swing of 0.163 volts per decade. The converted n-type SnO2 device, measured at +0.4 volts, carried a substantially larger saturation current of 10.6 microamperes at 0.6 volts, with a mobility of 1.17 square centimeters per volt-second, an ON/OFF ratio of 1.03 × 10^3 and a subthreshold swing of 0.277 volts per decade. Transient measurements gave turn-on and turn-off time constants of 28.6 and 7.7 milliseconds for the p-type device and 88.3 and 28.5 milliseconds for the n-type one.

Stability is a recurring worry for electrolyte-gated devices, but both polarities held up. Repeated transfer measurements over 15,000 seconds showed no noticeable current degradation or threshold-voltage drift for either device. The team also used gate-current analysis, analogous to cyclic voltammetry, to separate capacitive charging from genuine electrochemistry. During normal operation, both p-type and n-type devices drew small quasi-rectangular capacitive currents of roughly 0.2 and 0.5 nanoamperes, consistent with non-faradaic double-layer formation. At the −1.5-volt doping voltage, however, the gate current spiked to about 40 nanoamperes, nearly 200 times the capacitive baseline, and the associated faradaic charge density reached 9.01 × 10^15 per square centimeter, well above the roughly 10^14 ceiling achievable by purely electrostatic gating. Once the conversion was complete, the gate current returned to its capacitive shape.

The electrolyte itself proved to be a decisive variable. Polarity conversion succeeded with three different ionic liquids, including [EMI][BF4] and [BTMA][TFSI], indicating that the mechanism is not tied to one particular ion pair. With a 0.1-molar aqueous sodium chloride electrolyte, however, the conversion failed: linear sweep voltammetry showed water electrolysis beginning near 1.2 volts, before SnO could be oxidized, whereas [EMI][TFSI] remained stable up to about 3 volts. The ionic liquids occupy a useful middle ground, offering a wide electrochemical window while still carrying the trace water needed to supply reactive oxygen species. Conventional aqueous electrolytes, the authors conclude, simply cannot sustain the interfacial field required to drive the phase change.

With selective conversion in hand, the researchers moved from single transistors to circuits. A complementary inverter was first assembled from two p-type SnO transistors in series; the pull-down transistor was then converted in place to n-type SnO2, eliminating any separate n-type material processing. At a supply voltage of 0.9 volts, the inverter showed a sharp voltage transition with full rail-to-rail output swing within a narrow 0.1-volt input window and a voltage gain of 19.3. More complex gates followed from the same recipe: converting the series-connected pair among four p-type transistors produced a NAND gate, while converting the parallel-connected pair produced a NOR gate. Tested at a supply voltage of 0.7 volts with logic levels of −0.2 and 0.4 volts, both gates produced the correct output for every input combination, with the NAND outputting logic 1 for all inputs except (1,1) and the NOR outputting logic 1 only for (0,0).

The broader significance lies in what the approach avoids. Polarity control has mostly been demonstrated in atomically thin two-dimensional materials such as black phosphorus and tungsten diselenide, where gate-tuned Schottky barriers or selective van der Waals contacts can flip carrier type, but those films are extremely surface-sensitive and difficult to integrate at scale. Oxide semiconductors, by contrast, are chemically robust, thermally stable and compatible with low-temperature atomic layer deposition, which offers sub-nanometer thickness control and large-area uniformity, precisely the properties needed for deterministic, device-to-device polarity conversion. Stable p-type oxide semiconductors remain scarce, and SnO, with its high hole mobility rooted in the hybridized O 2p and Sn 5s valence band, is one of the few credible candidates; it is also less prone to swelling and structural disorder during electrolyte gating than organic semiconductors.

Limitations remain, and the authors are measured about them. State-of-the-art n-type oxide electrolyte-gated transistors based on zinc oxide or indium gallium zinc oxide still achieve higher mobilities and sharper switching, and the electron channel in the converted devices depends on a surface phase whose thickness is set by the electrochemical process. Even so, the demonstration that one ALD-grown oxide film can be patterned into both polarities after fabrication, using nothing more exotic than an ionogel, a carbon-nanotube gate and a carefully chosen voltage, offers a simplified and potentially scalable route to monolithic complementary oxide logic, a long-sought goal for thin-film electronics on glass, plastic and other large-area substrates.

Subject of Research: Electrochemical polarity conversion of tin oxide semiconductors for complementary logic circuits

Article Title: Selective Polarity Control of Metal Oxide Semiconductors for Complementary Logic Gates

Article References: Park, D. H., Ryu, S. H., Kim, M. S., Chung, T.-M., Kim, S. K., Baek, I.-H., & Lee, K. H. (2026). Selective Polarity Control of Metal Oxide Semiconductors for Complementary Logic Gates. Advanced Science, 13(59), Article e76809. https://doi.org/10.1002/advs.76809

Image Credits: AI Generated

DOI: 10.1002/advs.76809

Keywords: tin oxide, SnO2, electrolyte-gated transistor, electrochemical doping, polarity control, complementary logic, oxide semiconductors, atomic layer deposition, ionogel, thin-film electronics, inverter, NAND gate