Touchless user interfaces are gaining traction due to growing hygiene concerns in public spaces like ATMs and kiosks. Existing solutions, such as voice recognition, eye tracking, near-field communication, radio-frequency-based gesture recognition, and motion detection, have limitations. Commercial systems often utilize NIR cameras, such as the Microsoft Kinect (time-of-flight, requiring calibration) and Leap Motion Controller (hand gesture, limited interaction space). ShadowSense technology, relying on shadow detection, also presents certain constraints. This research aims to address these limitations by developing a visually transparent NIR-sensitive organic photodetector (OPD) array as a touchless interface, which can be seamlessly integrated onto existing displays, improving accuracy and eliminating field-of-view issues. The use of reflected NIR light from fingers or a penlight provides the input signal for screen control.

The existing literature extensively covers various touchless interface technologies. Voice recognition systems, while convenient, can be unreliable in noisy environments. Eye-tracking technology requires precise calibration and may be susceptible to user variability. Near-field communication (NFC) has limited range and requires specialized tags. Radio-frequency-based gesture recognition, like Google's Project Soli, has challenges regarding accuracy and power consumption. Existing gesture or motion-detection systems often require large spaces and sophisticated processing. Commercial systems such as Microsoft Kinect and Leap Motion Controller, while providing touchless interaction, suffer from drawbacks in calibration requirements and limited working ranges. This research leverages the advantages of NIR light reflection from human skin while addressing the shortcomings of previously reported systems.

The researchers developed a large-area, solution-processed NIR-sensitive OPD array. The device features an inverted OPD stack with a 300nm active film—a blend of PCE-10 (donor polymer) and IEICO-4F (non-fullerene acceptor)—coated on a printed copper grid transparent conductive electrode (TCE). An amorphous indium gallium zinc oxide (a-IGZO) layer served as the electron transport layer. The OPD layer was photolithographically patterned for transparency, followed by a thermally evaporated MoO3 hole transport layer and an indium tin oxide (ITO) electrode, finally topped with a laminated barrier film. Electro-optical modeling, using numerical electro-optical and two-dimensional finite element modeling (FEM), optimized the design of the NIR-sensitive OPDs, specifically the copper grid's line width and pitch. Simulations were validated against experimental results, identifying an optimal pitch of 10–20 µm with a 1 µm line width. The patterned OPD subpixel array was optimized through an optical transmittance prediction model, considering photoactive and non-photoactive stack components. This model aimed to maximize visible light transmittance (VLT) and EQE while maintaining adequate photocurrent. The performance of discrete OPDs was then characterized through J-V measurements, linearity analysis, EQE and spectral responsivity assessments, and detectivity calculations. The fabrication and characterization of a 16 × 16 OPD array, incorporating a 14 × 14 parallel OPD subpixel array for each main pixel, are also described, including its linearity, uniformity, and transient response analysis. Two demonstration videos showcase the touchless interface using a NIR-emitting penlight and gesture recognition.

The study successfully fabricated a visually transparent NIR-sensitive OPD array with a VLT of approximately 70%. Electro-optical modeling played a crucial role in optimizing the design, achieving an external quantum efficiency (EQE) of 36% at 850nm. The OPDs exhibited a detectivity of approximately 10¹² Jones at 850 nm, with a low dark current density of around 10⁻¹⁰ mA cm⁻². The 16 × 16 OPD array demonstrated a linear photoresponse and high uniformity. Transient response measurements showed rise and fall times of 2.4 ms and 2.8 ms, respectively. The touchless user interface demonstrations, utilizing both a NIR-emitting penlight and gesture recognition with integrated NIR LEDs, successfully controlled a laptop display, showcasing the practical applications of the developed technology. Specifically, the penlight demo used 800, 960, and 1200 Hz signals for right click, left click, and idle states, respectively. The gesture recognition demo used the variation in light intensity due to changing distance between the finger and imager to detect click gestures. In both demos, bandpass filtering was used to remove background noise, ensuring both a high pointer position update rate and high signal-to-noise ratio.

The results demonstrate the successful development and implementation of a novel touchless user interface. The high VLT, excellent photodetectivity, and low dark current of the OPD array contribute significantly to the system's functionality. The use of electro-optical modeling and optimization of the device design highlights the importance of theoretical modeling in achieving desired performance parameters. The successful implementation of both penlight and gesture-controlled interfaces proves the versatility of the proposed technology. The use of readily available fabrication techniques makes the approach scalable and cost-effective. The achieved position accuracy, particularly for the gesture demo, is sufficient for many applications. The integration of the imager onto standard displays without substantial impact on visibility adds to the practical feasibility of this technology.

This research successfully developed a solution-processed, large-area, visually transparent NIR-sensitive OPD array for touchless user interfaces. The optimized design, achieved through electro-optical modeling, resulted in high VLT and photodetectivity. The demonstrated penlight and gesture-controlled interfaces highlight the technology's potential for diverse applications, such as ATMs, interactive whiteboards, and electronic signage. Future work could explore higher-resolution arrays, more sophisticated gesture recognition algorithms, and the integration of this technology into a broader range of display types.

While the current technology demonstrates promising results, certain limitations should be considered. The resolution of the 16x16 array (4.2 pixels per inch) might not be sufficient for all applications. Further research is needed to increase the pixel density and improve the overall resolution. The current implementation of gesture recognition is limited to simple click gestures; more complex hand movements might require more sophisticated algorithms and potentially additional hardware components. The reliance on a bandpass filter to eliminate background noise necessitates further study to evaluate performance under varied lighting conditions and enhance robustness against ambient light interference. The performance under extremely harsh lighting conditions needs to be explored.