The researchers screen-printed flexible infrared sensors from p- and n-type thermoelectric films and varied ink composition to tune film thickness. Measurements under a 10.6 µm CO₂ laser showed that thinner films generated larger voltages despite the intermediate films having higher absorptance; the thinnest paired-leg sensor reached 107.7 mV W⁻¹. Finite-element simulations reproduced the thickness trend, although the paired-leg thickness dependence was stronger in measurements than in the model. The demonstrated performance is laser-based; wearable and proximity-sensing uses are proposed appl

A flexible, printed sensor turns invisible light into electricity. Making its films thinner makes that signal stronger.

Highest measured responsivity in the thinnest paired-leg device

Bose et al. · Journal of Materials Chemistry A Read the paper ↗

Try taking material away.

Choose a film thickness. The same incoming light meets a different thermal mass—the amount of material that must warm up.

Two different thermoelectric materials make electrical contributions that add in the pair.

The thickest pair produces less voltage for the same incoming power.

Approximate reading from Figure 7f.

The key choice is thickness. Thinner films have less thermal mass and a denser microstructure. The measured electrical response is stronger; the authors link this to the larger temperature difference that develops in thinner films.

Invisible light. Visible consequences.

The sensor uses a temperature difference to generate its electrical signal.

A small spot of invisible light.

Infrared is light beyond the red end of our visible spectrum. In this study, a 10.6 μm laser illuminates the sensor’s central junction—the place where the two printed strips meet.

Absorbing more is only half the story.

The middle thickness absorbed the most infrared light. But the thinnest films delivered the strongest electrical response.

Signal per watt of light

The thinnest pair reaches the paper’s highest measured responsivity: 107.7 mV/W.

107.7 mV/W is reported in the text. ≈95 and ≈75 mV/W are approximate readings from Figure 7f. Nominal thicknesses: 12 ± 1, 23 ± 1 and 30 ± 1 μm.

Highest absorption. At the middle thickness.

Light absorption alone did not predict the strongest sensor. Film thickness also changes the thermal response and microstructure that shape the voltage.

Absorption reported for the intermediate films, Section 3.1.

The experiments and the thermal model.

Figure 7. Laser measurements compare the single-leg and paired sensors. Across the three architectures, the thinnest films give the highest responsivity.

Bose et al., Journal of Materials Chemistry A (2026), DOI: 10.1039/D6TA05248G. CC BY 4.0. Figures 7(a–f) and 9(a–d), cropped from the original; data unchanged.

The ink changes the architecture of the film.

These are actual scanning electron microscopy images: a beam of electrons reveals the films’ tiny grains and spaces. Compare the thicknesses to see the finer, more densely packed structure of the thin film.

The paper relates absorption to film morphology; the number of voids alone does not explain the absorption trend.

Bose et al. (2026), Figure 4 · CC BY 4.0 · Cropped and resized; microscope image and scale bar preserved.

Printed to bend. Built to sense.

The films were screen-printed onto flexible sheets. The work connects scalable fabrication, infrared detection and a simple material choice.

Flexible, with a measured history.

After 100 bending and unbending cycles around small rods, the researchers observed no cracks or breaks. Resistance still changed: the p-type films rose by up to 50%, while thinner films were more mechanically stable.

A faster thin-film response.

The thinnest sensor’s fitted rise and decay time constants were 3.87 s and 5.54 s. These measured times are distinct from the accelerated animation on this page.

Where the idea could go.

The authors point to wearable sensing, proximity detection and non-contact interfaces as possible applications. The measurements here establish the sensor response under laser illumination.

Tailoring thermal mass for enhanced infrared response in Sb–Bi–Te-based printed thermocouples

Amrutha Bose, Zirui Wang, Muhammad Irfan Khan, Leonard Franke, Apoorv Kulkarni, Stefan Heißler, Heidemarie Schmidt, Mohamed Hussein, Uli Lemmer and Md Mofasser Mallick

Journal of Materials Chemistry A · 2026 · 10.1039/D6TA05248G ↗

The virtual bench calculates V = R × P, using the highest reported responsivity or approximate graphical readings from Figure 7. It illustrates the direction and scale of the signal; it does not reproduce the individual experimental points or the finite-element simulation. Animation timing, film thickness in the drawing and the thermal colors are illustrative.

Responsivities used, in nominal thickness order 12, 23 and 30 μm: paired device 107.7, ≈95 and ≈75 mV/W; p-only ≈58, ≈45 and ≈35 mV/W; n-only ≈52, ≈51 and ≈40 mV/W. The selected power range ends at 13.37 mW, a power used in Figure 9. Room-temperature laser measurements and the 10.6 μm wavelength are described in Section 3.4.

The two materials are p-type Bi₀.₅Sb₁.₅Te₃ and n-type Bi₂Te₂.₇Se₀.₃. Their room-temperature Seebeck coefficients are about +120 and −143 μV/K (Section 3.3). In a pair, their opposite signs increase the voltage between the cooler ends. Film structure and absorption are discussed in Section 3.1; bending in Section 3.2; response times in Figure 8; finite-element modeling in Section 4. The paper’s conclusion calls the thinnest film 11 μm, within the reported 12 ± 1 μm interval used here.

The source paper and reproduced figure crops are licensed under Creative Commons Attribution 4.0. Source figure crops have not been recolored or changed. The graphical abstract is an AI-generated conceptual illustration, adapted from the device geometry and thermal-gradient idea in Figures 1 and 9; it is not an experimental result.

Bose et al. (2026), 10.1039/D6TA05248G · CC BY 4.0 · Crop only.