Mechanoluminescence (ML), an interesting light emission phenomenon that directly convert widely distributed mechanical stimuli to photon^1,2, has attracted huge attention in widespread applications including visualized stress distribution^3–5, electronic signatures^1,6, temperature sensing^7–9, health diagnosis¹⁰, mechanically driven light emission device¹¹, and flexible electronic skin^4,6,12–14. Numerous inorganic ML materials such as ZnS:Mn/Cu^6,15,16, SrAl₂O₄:Eu^17,18, CaZnOS:Cu¹⁹, and LiNbO₃:Pr^20,21, have been investigated. Nerveless, ZnS is still considered as one of the most representative ML materials, which exhibits advantages of self-recoverability²², stable ML emission²³, free of pre-excitation²⁴, feasibility in color-tuning²⁵, high brightness, and easy fabrication.

Normally, these ML materials are prepared into powder like samples and mixed with highly elastic organic materials such as polydimethylsiloxane (PDMS)^26–28, epoxy resin^29,30 and so on, to demonstrate the ML properties by applying different forms of stress. However, these powder like samples are mainly composed of polycrystalline materials. Under macroscopic stress, the forces between microscopic crystals are very complex, involving not only the influence of crystal interfaces, but also the analysis of stress conditions³¹, which ultimately makes the analysis of ML mechanisms particularly difficult, and greatly limits the accuracy of stress sensing, causing difficulties in the practical applications. Thus, there is an imperative need to develop non-polycrystalline integrated ML materials to simultaneously study the real ML mechanism and device based stress sensing applications.

Stress-threshold is an important indicator reflecting the ML accuracy and performance. Recently, researchers have conducted research on the stress-threshold of ZnS system. Jha et al. systematically studied the stress-threshold of the macrocrystal, microcrystal, and nanocrystal ZnS:Mn, 30 and 0.8 MPa for the macrocrystal and microcrystal, respectively³². It is revealed that the threshold is highly related to the piezoelectric field inside the ZnS:Mn crystals. Particularly, Kleckner et al. used the nanoscopy tools to precisely study the ML thresholds of single ZnS:Mn particles³³. Great heterogeneity of the stress-thresholds was observed, ranging from 0.233 to 47.15 MPa, with the lowest threshold obtained from the particle that is abundant with stacking disorder. In 2024, Yuan et al. reported a state-of-the-art low threshold force of 0.1 N for the ZGGO:Cr by precise control of defect content and trap distribution¹⁴. While the defect-control strategy has significantly reduced the threshold, it remains worthwhile to explore new approaches to feasibly lower the threshold force for ML emission. Furthermore, in the previously researches, the optical properties of the ML materials have been well investigated while the electrical changes such as variations in carrier concentration and conductivity under mechanical stimuli is lacking, mainly due to the poor conductivity and complexity of the inorganic-organic composite. The deficiency of electrical properties study hinders the deep-insight understanding of the real ML mechanism. Therefore, a combined study of mechano-electro-photon behaviors is of great significance.

Under the current perspective, the generation of non-balanced free carriers is a key to the ML performance. Based on the origin of non-balanced carriers, two mainstream mechanism models, piezoelectric-induced carrier detrapping model^20,34,35 and the triboelectric-induced electroluminescence model^5,36,37, have been proposed and developed. Regarding the carrier detrapping model, it is believed that the charge carriers trapped at various defects are released to the conduction or the valance band under the mechanical stimulus, generating non-balanced carriers for ML emission. For the triboelectric-induced electroluminescence model, non-balanced carriers are generated through the carrier transfer from the counterpart triboelectric materials, and has been well confirmed by Xie et al.¹⁶ As mentioned above, the non-balanced carriers for ML emission rely on either the detrapping of the electrons through piezoelectric effect or the electron injection by triboelectric effect, both of which result in limited non-balanced carrier generation. Conversely, the valence band (VB) of ML materials contains abundant electrons, but they are hardly activated by mechanical stimulation due to the substantial energy required for activation. For the photodetector design, researchers have integrated counterpart materials with the optoelectronic materials, forming a heterostructure in which type Ⅱ electron transition occurs^38–41. Herein, type Ⅱ electron transition means the electron in the VB of one material is directly activated to the conduction band (CB) of the counterpart material, so that lower energy could be efficiently utilized. This strategy may dramatically increase the non-balanced carriers engaged in ML emission and subsequently significantly improve the ML performances, however, no relevant study has been conducted.

In this work, we propose constructing a ZnS:Mn/ZnO heterostructure with a type Ⅱ energy band alignment, where the CB of ZnO (CB_ZnO)is moderately higher than the VB of the ZnS:Mn film (VB_ZnS), aiming to directly activate the electrons in the VB of the ZnS for ML emission. The heterostructure design may also lower the stimulation threshold by introducing additional piezoelectric field via lattice mismatch and a built-in electric field through heterojunction formation at the ZnS:Mn/ZnO interface. Using standard photolithography, we patterned the film into a 50×50 µm² square array, forming a regular heterostructure with discrete emission centers. Each square functions as a pixel for mapping force distribution and recording dynamic forces. Remarkably, the ZnS:Mn/ZnO heterostructure exhibited a record-low threshold of 0.05 N (4 kPa). Additionally, two gold (Au) electrodes were formed on the ZnS:Mn/ZnO heterostructure to enable combined mechano-electro-photon studies. Pulsed electric current and 4-fold increasement in ML intensity were observed for the ZnS:Mn/ZnO heterostructure, suggesting the successful activation of type Ⅱ electron transition upon newton force stimulus. Compared with the limited trapped carriers by defects or the carrier transfer by triboelectric effect, the VB of ML materials contain large number of electrons that are available for activation. Therefore, the type II electron transition proposed in this work may dramatically increase the number of non-balanced free carriers involved in the ML process, thereby have a revolutionary impact on the efficiency and fundamental understanding of ML emission.

Deposition of ZnO:Mn films: Quartz glass with the size of 25×25×1 mm³ was selected as the substrate for ZnO film growth. Prior to deposition, the substrates were ultrasonically cleaned in acetone, methanol, and deionized water in sequence for 10 mins, respectively. ZnO films doped with Mn were deposited on quartz glass substrates by co-sputtering technique (PVD-75, Kurt J. Lesker, USA) with a thickness of 200 nm. Various powers of Mn target, 2.5, 3, 3.5, 4 and 5 W were studied, and their photoluminescence are presented in Figure S12. When the substrate was heated to 200^oC, the chamber was filled with argon at a working pressure of 4 mTorr. Two fixed radio frequency (RF) power of 100 W and 3.5 W were applied to the ZnO and Mn targets, respectively. EDS was utilized to measure the element mappings (Figure S14), and the Mn content (3.5 W) was measured to be 0.6 at%.

ZnS:Mn/ZnO heterostructure construction: To fabricate the ZnS:Mn/ZnO heterostructure, an AZ5214 photoresist mask was used to pattern the ZnO:Mn film deposited above, and a AlN film (80 nm) was grown as a barrier layer to prevent the ZnO layer beneath from sulfidation in the following annealing process. On the other hand, porous anodic aluminum oxide template (Pore diameter 200 nm and period of 450 nm) was covered on ZnO:Mn film before sulfidation annealing, to fabricate the ZnS:Mn/ZnO-AAO heterostructure. As for the sulfidation annealing process, 3g sulfur powder was added to the quartz bottle placed in the tube furnace (OTF-1200X-III, Hefei Kejing Materials Technology Co., China), films were heated to 700^oC with a rate of 10^oC min^− 1 and annealed for 60 min in the argon gas for sulfidation, followed by slowly cooling down (5^oC min^− 1) to room temperature. The anneal process in detail is presented in Figure S13. The sulfur powder was simultaneously heated and kept at 200^oC for the whole annealing process.

Characterizations: Surface morphology, crystal structure, and chemical composition of all films were characterized using SEM (APREO S, Thermoscientific, Holland), XRD (Ultima IV diffractometer, Rigaku), and XPS (Nexsa, TFS). Cross-sectional images and ZnS:Mn/ZnO interface were obtained using transmission electron microscopy (TEM, Titan Cubed Themis G201, FEI). The film thickness was measured by a surface-profile measurement system (Dektakxt, Bruker, USA). PL spectra were recorded using a FLS-1000 photo-luminescence spectrometer (Edinburgh Instruments Ltd., UK). The equipment used to obtain ML spectroscopy was a home-built measuring system consisting of a stepper motor, a digital push-pull force gauge, a linked displacement platform and the QE65pro fiber optic spectrometer (OceanInsight, UK). The UV-vis absorption spectra were carried out by using a UV-vis spectrophotometer (UV 3600 plus, Shimadzu, Japan). Electrical measurements were conducted at room temperature using an AES-4TH testing system equipped with a probe station.

We systematically investigated the ML behaviors of four ZnS:Mn thin films including traditional film, porous ZnS:Mn film prepared through sulfidation, the ZnS:Mn/ZnO-LT, and the ZnS:Mn/ZnO-AAO. These films were directly grown on the quartz glasses and no pre-irradiation was conducted before ML measurement. A spectrometer in Fig. 3f was utilized to record the ML properties. As shown in Fig. 3a, the traditional ZnS:Mn film exhibits very weak ML emission, even when subjected to a strong force of 13 N. In sharp contrast, the sulfidated ZnS:Mn film (Fig. 3b) and the ZnS:Mn/ZnO-LT heterostructure (Fig. 3c) exhibit 12- and 15-fold stronger ML emission, respectively. Moreover, the ML behaviors of the ZnS:Mn/ZnO-AAO film with thicknesses of 200 and 400 nm exhibit similar improvements (Figure S6 and S7), further confirming the key role of the heterostructure.

The ML intensities at various stimuli of the ZnS and ZnS:Mn/ZnO-LT heterostructure are summarized in Fig. 3d. It is seen that the films with heterostructure demonstrate pronouncedly higher ML than the pristine ZnS:Mn film, especially at the cases of small stimulus (< 7 N). Within the force range of 1–7 N, the emission intensity exhibits a linear relationship with the applied force. For a direct comparison, their ML spectra stimulated at 1 N are depicted in Fig. 3g. The ML intensities of the ZnS:Mn/ZnO-LT film is about 4-fold stronger than that of the porous ZnS:Mn film. As shown in Fig. 3e, the ML of the ZnS:Mn/ZnO-LT heterostructure films are strong that they can be directly seen by naked eyes under a small stimulus of 1 N. To study its performances at small stimulus, the ML curves of the ZnS:Mn/ZnO-LT film range from 0.05 to 0.8 N were recorded and shown in Fig. 3h. Interestingly, the ZnS:Mn/ZnO heterostructure exhibits clear ML behavior under these small stimuli. To our knowledge, 0.05 N is the record-low stress threshold of the reported ML materials. Low stress-threshold is of great importance for the development of next-generation stress sensors, the electronic skin, and the humanoid robots since it can expand the detection to rather small stress together with much higher resolution. The ZnS:Mn/ZnO heterostructure almost has no stress-threshold, together with the linear relationship between the ML intensity and the stimulus force in Fig. 3i, making it considerable attractive for widespread practical applications. Furthermore, the ML performances under stimulus with frequency of 500 Hz were also studied. As depicted in Figure S8, the ML peaks observed in the near-infrared spectrum are centered at 754, 775, 852, and 970 nm, which can be attributed to the formation of (Mn)_n clusters⁴⁷. The application of high-frequency stimulation enhances the likelihood of (Mn) _n cluster formation and promotes the energy transfer from the excited Mn²⁺ ions to the (Mn)_n clusters, consequently resulting in the emergence and intensification of the near-infrared emission band.

As depicted in Figure S4a, the transmittance of the ZnS:Mn/ZnO-LT film is smaller than both the ZnO and the ZnS film, suggesting enhanced light absorption because of the ZnS:Mn/ZnO heterostructure. Figure S4b demonstrates that the optical bandgap of the ZnO is approximately 3.28 eV, and the bandgap of the ZnS is around 3.64 eV. Figure S4c and 4d depicts the PL spectra of the ZnS:Mn/ZnO-LT films excited by 330 nm light and the PLE spectra at 585 nm. Figure S4e shows that the time-resolved decay curve of the traditional ZnS:Mn film is consisting of two exponential decay components. Differently, the porous ZnS:Mn film and the ZnS:Mn/ZnO-LT in Figure S4f exhibits three exponential decay components. The values of the decay curve in detail are summarized in Table S1. Compared to porous ZnS film in Figure S4f, the ZnS:Mn/ZnO-LT exhibits a higher proportion of 𝜏₂ occupancy, A₂. This suggests that the heterostructure introduces more defects⁵. Moreover, thermoluminescence test was also conducted, as shown in Figure S5, but no thermoluminescence was observed both for the porous ZnS:Mn film and the ZnS:Mn/ZnO-LT with various thicknesses.

To characterize the coupled mechano-electro-photon behaviors of the ZnS:Mn/ZnO-LT heterostructure, two Au electrodes were deposited on the film surface to simultaneously collect electrical responses, while ML emission was recorded under the same mechanical stimulus. Figure S9 shows the I-V curves of the porous ZnS:Mn film and the ZnS:Mn/ZnO-LT film. As ZnS is a wide-bandgap semiconductor and highly resistive, the current is rather small. In contrast, the I-V curve of the ZnS:Mn/ZnO-LT heterostructure device exhibits rectifying characteristic of diode, suggesting formation of heterojunction, and the conductivity of the ZnS:Mn/ZnO-LT is much larger, ~ 40 nA, due to the higher conductivity of the ZnO.

Figure 4d shows the current responses of the ZnS:Mn/ZnO-LT heterostructure to force stimuli varied from 0.2 to 7 N. Interestingly, pulsed currents are clearly observed once the ZnS:Mn/ZnO-LT is applied to a force through Al needle, and recovers to the original value when the friction is stop. The larger the force is, the higher the pulsed current △I is. As shown in Fig. 4e, the current increases rapidly upon stimulation, with a current increase of △I = 9.5 µA at 3 N. In sharp contrast, there is no pulsed current both for the ZnO and the ZnS:Mn film in Fig. 4a and 4b, even at a larger stimulus of 13 N. For a given semiconductor, the carrier mobility is unlikely to change; thus, it can be deduced that free carriers are generated in the ZnS:Mn/ZnO heterostructure upon force stimulus. In contrast, neither the pristine ZnO nor the ZnS:Mn film generates free carriers, or the number of free carriers generated is smaller. The difference in the free carrier generation between the heterostructure and the pristine film is of great interest, since it implies new mechanism of free carrier generation in the ZnS:Mn/ZnO-LT heterostructure. Moreover, the ML behaviors of the ZnS:Mn/ZnO-LT heterostructure were simultaneously recorded upon the same stimuli in electrical signal collection and shown in Fig. 4c and 4f. Figure 4g shows the plots of ML intensity as a function of newton force, suggesting excellent linearity between them. The magnitude of the pulsed current △I upon newton force are depicted in Fig. 4h. It is seen that the △I gradually increase with the force within 5 N and sharply increases to 20 µA at 7 N. In Fig. 4i, it is observed that the higher the pulsed current generated upon the force stimulus, the stronger the ML intensity is.

Besides the process (i), it is deduced from Fig. 4d that the electrons in the VB_ZnS are directly activated to the CB_ZnO upon force stimulus. This process is denoted as process (ii) in Fig. 5b, resulting in the additional non-balanced holes in the ZnS. For the photodetector research, the phenomenon of electrons transition from the VB of one material to the CB of the adjacent material has been well recognized, and named it as type Ⅱ electron transition^40,41. Herein, it is the first time to propose and confirm the electron transition from the VB of one material to the CB of adjacent material in the ML study. To distinguish the electron transition within one material, we also named it as type Ⅱ electron transition. Since the conductivity of ZnO is better than the ZnS, the conductive channel of the ZnS:Mn/ZnO heterostructure is mainly ZnO. The type Ⅱ electron transition from VB_ZnS to CB_ZnO results in the increase of free electrons in ZnO and thus the improved conductivity of the heterostructure, well explaining the pulsed current observed in Fig. 4d. Moreover, the non-balanced holes in ZnS generated by type Ⅱ electron transition also pronouncedly enhance the electron-hole pairs’ non-radiation combination, resulting in a significantly improvement in ML emission as shown in Fig. 3d. In Figure S11, the type Ⅱ electron transition is further confirmed by 520 nm photon irradiation. Compared with the limited trapped carriers by defects or the carrier transfer by triboelectric effect, the VB of ML materials contain large number of electrons that are available for activation. Thus, the type II electron transition in the ML behavior observed herein may significantly increase the number of non-balanced free carriers, thereby have a revolutionary impact on the efficiency and intensity of ML emission. This discovery provides a novel and general strategy for improving ML performance and advances the understanding of the ML mechanism.

Considering the integration advantages of ZnS:Mn/ZnO heterostructure and the innovative carrier generation mode, this study further designed a pixel-array ML integrated application mode. Under the dynamic stress application, the ZnS:Mn/ZnO heterostructure array realizes the pixelated trajectory recording of scratching and handwriting by converting mechanical stress into luminescence through the ML process. As shown in Fig. 6a, the scratching trajectory is visualized in real-time via image acquisition. A complementary metal oxide semiconductor (CMOS) image sensor is employed to capture the luminescence of the ZnS:Mn/ZnO heterostructure array under the stress. Using a 20x20 pixel array, the handwriting of “SZU” was successfully recorded in Fig. 6d. The ML intensity is derived by extracting the gray values of the captured image, and further calibrated to the value of force using the relationship shown in Fig. 4g. In this way, the approximate magnitude of the force applied during the “SZU” writing process was capable to record and present in in Fig. 6d. Notably, the ZnS:Mn/ZnO heterostructure array was fabricated via standard semiconductor processes on a wafer scale, enabling the integration of a photodetector array with the heterostructure array chip as depicted in Fig. 6c, directly converting the stimulus to the force/stress values, may providing dual-modes sensing of force or stress in the future. Owing to its extremely low stress-threshold, the heterostructure array-based system holds potential for applications in ethology research in Fig. 6c and the development of high-precision force-recognition mechanical hands in Fig. 6b. This design is the first time to transform the application mode of ML materials from polycrystalline powder to integrated array, result in the widespread promotion to develop the ML chips by utilizing the mechano-optical conversion. Due to its high efficiency and passive characteristics, it will have a significant impact on the current electrical stress sensing mode.

In summary, we have successfully constructed a ZnS:Mn/ZnO heterostructure with type Ⅱ energy band alignment. Standard semiconductor technique was utilized to fabricate the ZnS:Mn/ZnO heterostructure array chips, forming regular heterostructures and emission centers in square array. Each emission square functions as a pixel for mapping force distribution and recording dynamic forces, realizing the chip-based stress sensing applications. Pulsed electric current of 20 µA was simultaneously observed with bright ML emission, confirming the successful activation of type Ⅱ electron transition upon force stimulus. Notably, a record-low threshold of 0.05 N is achieved by introducing piezoelectric field at the heterostructure interface through lattice mismatch and a build-in electric field, providing a novel approach for lowering the ML stress-threshold. Compared with the limited trapped carriers by defects or the carrier transfer by triboelectric effect, the VB of ML materials contain large number of electrons that are available for activation. Thus, the type II electron transition in the ML behavior observed herein may significantly increase the number of non-balanced free carriers involved in the ML process, thereby have a revolutionary impact on the efficiency and intensity of ML emission.