Microfluidic-based redesign of a humidity-driven energy harvester

Abstract

Integrating microfluidic elements onto a single chip offers many advantages, including miniaturization, portability, and multifunctionality, making such systems highly useful for biomedical, healthcare, and sensing applications. However, these chips need redesigning for compatibility with microfluidic fabrication methods such as photolithography. To address this, we integrated microfluidics technology into our previously developed humidity-driven energy harvester to create a self-powered system and redesigned it so that it could be fabricated using photolithography and printing. The device comprises stacked electrodes, cation-exchange membranes, and microchannels. The multi-element version of the device generated ten times more voltage than the single-element version. Both versions produced stable patterns of voltage output with respect to the fluctuations in humidity in both controlled and real-world environments. Their potential as humidity sensors is supported by the correlations exhibited between humidity and voltage output. The capacity of the device to respond to changes in perspiration-induced changes in humidity suggests its usefulness as a power source for wearable sensors. This novel device element, which can be easily integrated into other microfluidic devices, is expected to provide a new approach to powering microfluidic-based wearable sensors.

---

Introduction

Integrating microfluidic elements onto a single chip is an important approach that can promote miniaturization, portability, reaction efficiency, cost efficiency, and multifunctionality.^1–3 Integrated microfluidics has multiple applications and is developing rapidly in the fields of biomedicine and healthcare, with applications in single-cell analysis,⁴ point-of-care testing,⁵ and drug delivery, among others,⁶ as well as in wearable sensors.⁷ In particular, wearable sweat sensors provide a promising application for microfluidic sensors.⁸ Therefore, integrating microfluidic power supplies into such sensors would be ideal to create self-powered microfluidic systems. Moreover, to integrate elements such as power supplies into microfluidic devices, it is important to redesign the devices to suit microfluidic technology. Currently, many device elements are designed using a bottom-up approach, in which the desired structure is created by assembling small components. In the integrated microfluidics approach, a top-down design and fabrication process are required, in which the desired structure is created by shaving off excess material from a large block of material. Such devices must therefore be redesigned so that they can be fabricated using microfluidic fabrication methods such as photolithography.

Recently, microfluidic devices that precisely control ion transfer using cation exchange membranes have been developed.^9–13 Their primary application is in analyzing biological samples such as cells¹⁴ and nucleic acids.¹⁵ In this application, ions locally concentrated in microchannels can act on biological samples to cause the concentration and separation of samples. Further, applications in power generation have also been reported,¹⁶ wherein power can be generated via reverse electrodialysis by filling two microchannels (chambers) connected via an ion-exchange membrane with electrolytes at different concentrations.

We recently presented a humidity-driven hygroelectric generator (i.e., cell) comprising a chamber structure (connected via a cation-exchange membrane), a hygroscopic liquid (such as LiCl_aq.), and AgCl electrodes.¹⁷ This device exhibits excellent performance in onsite power supply in wireless-sensor network systems because it can generate power in environments where solar power is unavailable (such as indoors or at night). Although the device has since been optimized both experimentally¹⁸ and theoretically,¹⁹ it achieves a low output voltage, and improvements in this output require device designs that can be integrated and serialized. Moreover, design improvements are required to facilitate microfluidic integration, considering that the device is relatively large (several centimeters per element) and rigid.

Therefore, to address these requirements, we have focused on integrating microfluidics into energy harvesters for sensing devices. In particular, we have redesigned our energy-harvester device¹⁷ as a self-powered microfluidic system that can be fabricated via photolithography and printing.

The use of a microfluidics-based design improves device integration. We designed and fabricated a multi-element device comprising ten devices connected in series, as well as a single-element device, to investigate the effects of microfluidics-based integration on power generation. We found that, by placing the device near the skin, it can generate a voltage from perspiration-induced changes in humidity. This novel device, which facilitates the integration of microfluidics technology, is expected to provide a new approach to powering various microfluidic-based wearable sensors.

Experimental

Device design

The device comprises a glass plate with patterned Ag/AgCl electrodes and ion-exchange membranes; two microchannels (open and closed to the atmosphere, respectively), separated using a polydimethylsiloxane (PDMS) wall, are placed on this plate (Fig. 1). The microchannels, connected by a Nafion cation-exchange membrane, are filled with a hygroscopic solution (LiCl_aq.). Along with the single-element device, a multi-element device, comprising ten elements in series, was tested.

Fig. 1 Design of the energy harvester. (a) Principle of power generation in dry and humid external environments (cross-sectional view). Device configuration of a (b) single-element device and (c) multi-element device.

The device operates as follows. In a dry environment, water evaporates from the solution in the open channel, creating a concentration gradient between the open and closed channels that generates a voltage between the two electrodes. Placing a load between the electrodes in this state causes the flow of Li⁺ ions. In the two channels, Cl⁻ ions are ionized from or bound to the AgCl electrode to achieve an electrical balance of ions in the solution. This transfer of Cl⁻ ions results in the transfer of electrons (Ag + Cl⁻ ⇌ AgCl + e⁻), thus generating electricity. In humid environments, the opposite phenomenon occurs, resulting in electron transfer in the opposite direction.

Device fabrication

For both the single- and multi-element devices, the following stepwise process was followed: (1) electrode formation via screen printing, (2) ion-exchange membrane formation via stencil printing, (3) microchannel fabrication via soft lithography, (4) microchannel bonding, and (5) AgCl formation.

The Ag electrodes were formed on a glass plate via screen printing. The screen-printed plates were fabricated using a plate-making machine (GOCCOPRO QS2536, Riso Kagaku Corporation, Japan) and computer-aided design software (AutoCAD, Autodesk, Inc., CA). Ag ink (RA FS 059 S, Toyo Ink Co., Ltd., Japan) was applied to a glass plate (slide glass, Matsunami Glass Ind., Ltd., Japan) using a screen-printing machine (NT-15TVA, Neotechno Japan Corporation, Tokyo, Japan) in which the constructed glass plate was installed. The Ag ink was then sintered by heating in an oven at 170 °C for 120 min.

The cation-exchange membrane was then constructed on the glass plate supporting the Ag electrode using stencil printing. First, to prepare a mold for the stencil, apertures were made in a silicone rubber film (thickness: 50 μm) using a laser cutter (CO₂ laser) (Speedy 360 flexx, Trotec Laser GmbH, Marchtrenk, Austria). The stencil mold was then placed on the glass plate supporting the Ag electrode. A 20% Nafion dispersion (663492, Sigma-Aldrich, St. Louis, MA) was dropped onto the plate, and the dispersion was spread on the plate using a plastic squeegee. After the stencil mold was removed, the Nafion was sintered by heating to 100 °C for 10 min on a hot plate. The dimensions of the prepared membranes were measured using a confocal laser microscope (LEXT OLS5000, Olympus Corporation, Tokyo, Japan).

Microchannels made of PDMS were fabricated using conventional photolithography and soft lithography.^20,21 In summary, a photoresist (SU-8 2100, Nippon Kayaku Co., Ltd., Tokyo, Japan) was applied to a silicon wafer and exposed to UV light through a photomask on which the channel geometry was printed. Subsequently, the photoresist was developed to create a convex mold. Liquid PDMS (10 : 1, mass ratio of the main agent-to-hardener) was cured by placing it on the convex mold and then heating it in an oven at 80 °C for 120 min. After curing, the PDMS was peeled from the mold to form a microchannel. Finally, a punch (BPP-10F, Kai Industries Co., Ltd., Gifu, Japan) was used to create an inlet (a hole with a diameter of 1 mm) to introduce liquid into the closed channel, and a surgical scalpel (no. 11, Akiyama-seisakusyo. Co. Ltd., Tokyo, Japan) was used to create a square opening (thus producing the open-channel reservoir).

The microchannel was bonded to the glass plate supporting the Ag electrode and cation-exchange membrane via oxygen-plasma bonding.²² To achieve this, the bonding surfaces of the glass plate and microchannel were irradiated with oxygen plasma using a plasma cleaner (PR510, Yamato Scientific Co., Ltd., Tokyo, Japan), which were then brought into contact with each other to undergo spontaneous covalent bonding. Only the microchannel bonding area was irradiated by the oxygen plasma to minimize the oxidation of the Ag electrode. The microchannel and glass plate (including the electrode and Nafion-based membrane) were lubricated with a drop of deionized water before the bonding surfaces were brought into contact. After contact, the residual deionized water was evaporated and removed by heating in an oven at 150 °C for 180 min.

To form AgCl on the Ag electrode area in contact with the solution, the microchannel was filled for 5 min with NaClO solution (4× dilution; Fujifilm Wako, Osaka, Japan), then rinsed with deionized water (the details are described in ESI† (Fig. S1 and Table S1)).

A hygroscopic liquid (20 wt% LiCl_aq.) was prepared as follows. LiCl (Fujifilm Wako) was added to the fabricated microchannels. The holes in the closed channels were then closed with polyimide tape to prevent the transfer of water vapor.

Device measurement

To characterize the device in a constant environment, the change in voltage was measured by subjecting it to periodic changes in humidity at a constant temperature. The devices were placed in a constant temperature and humidity chamber at 25 °C, with relative humidity changing from 30% to 90% every 4 h. The device was connected to a data logger for voltage measurement (LR5042, Hioki, Nagano, Japan) and an operation amplifier (non-inverting amplifier circuit, NJU7002D, Nisshinbo Micro Devices Inc., Tokyo, Japan). To measure the output, the charged device (i.e., at peak voltage) was connected to a load resistor (Fig. S2†).

To evaluate its performance in a real-world environment, the device was installed in an office and connected to the same data logger and amplifier. During measurement, a thermo-hygrometer (LR5001, Hioki, Japan) was installed to measure the humidity and temperature of the environment as well as the change in voltage of the device.

To demonstrate the use of perspiration, a sealed space was prepared on the surface of a person's arm, and a device was installed in that space. The change in voltage was measured using a data logger and amplifier similar to those used in the real-world (office) measurements.

This study was reviewed and approved by the AIST ergonomic experiment committee (approval no. HF2024-1463).

Results and discussion

Fabricated devices

The devices were fabricated using a stepwise process (Fig. 2a and b). By introducing sodium hypochlorite solution into the microchannel, the Ag electrode exposed to the liquid was partially chemically converted to AgCl. Although the two channels were separated by a PDMS wall and cation-exchange membrane, no liquid leakage occurred between them. The Ag electrodes formed were partially oxidized by the oxidizing plasma, turning black. The cation-exchange membrane was deposited via a simple process using stencil printing, without defects such as holes. A schematic of the membrane (separated by a PDMS wall) is presented in Fig. S3† (membrane thickness, 5.7 ± 0.75 μm; cross-sectional area, 0.014 ± 0.0015 mm²; n = 10). Although Nafion, which was used as a membrane material, was subject to swelling (as observed during our experiment), it did not break when exposed to water and could withstand long-term measurements. For the multi-element devices, we connected ten elements in series using the same fabrication process used to construct one device (Fig. 2b). As with the single-element device, the multi-element device was fabricated without liquid leakage from the channels and without membrane defects.

Fig. 2 The energy harvester, comprising open and closed microchannels. (a) A single-element device and (b) a multi-element device. Scale bars: 1 cm. Change in voltage in a (c) single-element device and (d) multi-element device during cyclic changes in humidity.

During the design stage, we predicted that PDMS would be unable to form strong bonds with the electrode and the cation-exchange membrane via oxygen-plasma bonding. This is because the electrodes and cation-exchange membranes formed by screen printing have different heights. To avoid leakage, we designed the device such that the contact surface between the PDMS and the electrode or cation-exchange membrane was as small as possible while making the bonding surface between the PDMS and glass as large as possible. In the fabricated device, although we did not verify the bonding between the PDMS and the cation-exchange membrane separating the channels, no leakage of liquid between the two channels was observed. This may be because the relatively large bonding surface between the PDMS and glass held the electrodes and cation exchange membrane firmly in place.

Changes in voltage with ambient humidity

For the single-element device, as for previously developed devices,^17–19 the voltage increased and then decreased in response to the change in humidity (Fig. 2c). At peak voltage, the voltage remained constant and did not change with humidity. Although similar results were observed for the previous device,^17–19 the change in peak voltage was smaller for the current device. This difference may be due to the fact that the cross-sectional area of the plane perpendicular to the direction of ion transfer was smaller in the present device than in the previous device, making it more difficult for water to move through the cation-exchange membrane. Ideally, cation-exchange membranes should allow only cations to pass through; however, in practice, they may also allow water to pass through. The simulations in our prior study¹⁹ revealed that the highest voltages were obtained when water migration was entirely absent. Therefore, this suggests that, in the present device, the higher voltage is due to the suppression of water migration.

In the multi-element device, as in the single-element device, the voltage increased and declined in response to the changes in humidity, achieving almost the same maximum and minimum voltage in each cycle (Fig. 2d). In the multi-element device, the maximum and minimum voltages were higher than those in the single-element device (five times the peak-to-peak voltage). These results indicate that, as with electrical devices in general, this device can be operated by amplifying the voltage when connected in series.

This device can potentially operate in a high-humidity environment; however, if left in that environment, it will not generate power because of the absence of change in humidity. Moreover, even if there is a high concentration of nitrogen or carbon dioxide in the environment, as long as there is a change in humidity, power generation will occur accordingly.

With regard to energy harvesting using micro/nanofluidic technology, there have been reports mainly on technologies that use moisture in the air²³ and salinity differences in the sea or rivers²⁴ as energy sources. The differences between these existing technologies and our proposed technology are as follows. In technologies that use prevalent moisture in the air, the higher the moisture content, the higher the potential for electricity generation. However, with our technology, the greater the fluctuation of the humidity content in the air, the greater the electricity generation capacity. Technologies that use the difference in salinity in seas or rivers are used in aquatic environments. However, our technology can be used in any environment with fluctuating humidity.

Power measurement

The power generated was calculated from the current and voltage measured during load-resistance testing (Fig. 3). For a single-element device, the maximum power was 2.4 nW (power density: 17 μW cm⁻², obtained by dividing the maximum output by the cross-sectional area of the cation exchange membrane, 0.014 mm²; Fig. S2†). The maximum power generated by the multi-element device was 16 nW, which was seven times higher than that generated by the single-element device. The load resistance at which the maximum power was obtained was ca. 500 kΩ for the single-element device and ca. 900 kΩ for the multi-element device. Typically, the optimum load resistance for achieving maximum power coincides with the internal resistance of the device.

Fig. 3 Power generation characteristics. (a) Current and voltage and (b) power in a single-element device. (c) Current and voltage and (d) power in a multi-element device.

The maximum output power of the previously developed device was 36.7 μW (with power density of 6.4 μW cm⁻²).¹⁸ In contrast, the new (single-element) device exhibited a maximum output that was several orders of magnitude lower, whereas its power density was three times higher. It is, therefore, possible to achieve a high output by increasing the size of the device or via the integration of multiple microfluidic elements.

The amount of power that can be obtained is small, but there is potential for application in devices that operate with minute amounts of power (e.g., nanowatt sensing^25,26).

Real-world measurements (in an office)

For both the single- and multi-element devices, as with the previous devices, the voltage changed in response to changes in humidity (Fig. 4a and b). The average absolute total changes in voltages achieved by the single- and multi-element devices were 14 mV and 67 mV, respectively. The relative humidity measured by the sensor and the voltage obtained using this device were correlated (Fig. 4c and d), suggesting the possibility of using this device as a humidity sensor. However, these data were not corrected for temperature, and the correlation may be even greater if temperature correction is performed. Typically, in humidity-driven power-generating devices, the ion concentration in the open channel remains relatively unchanged at constant humidity, resulting in a small voltage output. Therefore, for application as an energy harvester, it is important to install the device at a location with fluctuating humidity, such as in locations where the humidity varies from day to night. However, using such devices to generate electricity is difficult in air conditioned venues where the humidity is actively maintained at a constant value.

Fig. 4 Power generation characteristics of the devices under real-world conditions. Changes in voltage of (a) a single-element device and (b) a multi-element device in an indoor environment over a 10 day period. Voltage vs. humidity for (c) a single device and (d) a multi-element device.

Demonstration using sweat

To demonstrate the application of this device as a wearable energy harvester, the device was installed in a sealed space close to the skin, and the change in voltage was measured (Fig. 5). A single-element device was used for these measurements (Fig. 5a). During measurement, the subject sat on a chair and performed deskwork at a computer. After the seal was applied (“set”, Fig. 5b), the voltage gradually decreased. After one hour, the voltage became almost constant, reaching a minimum of ca. −30 mV, i.e., of the same order as the value obtained in our environmental testing. After the sealed space was opened to the atmosphere and the device was simultaneously removed from the arm (“release”, Fig. 5b), the voltage gradually increased, peaking after 4 h, then gradually decreasing to its initial value over the next 7 h. Therefore, the change in voltage was caused by an increase in humidity from sweat in the enclosed space on the arm where the device was confined. This suggests that long-term changes in voltage could be generated by periodically confining the device on the skin and later releasing it. With other types of wearable energy harvesters (e.g., heat, pressure, and electromagnetic harvesters), it is difficult to generate power if the user does not move (e.g., for elderly, disabled, or ill users) because there is no change in physical quantities.²⁷ On the other hand, even if a person is not moving, environmental conditions such as temperature and humidity,^28,29 emotional states such as stress and anxiety,³⁰ changes in hormones,³¹ and the effects of drugs,³² can all affect the amount of perspiration. Therefore, the proposed device has the potential to generate power even when the user is stationary.

Fig. 5 Measurement of the change in voltage using a single-element device installed in a sealed space near the skin. Setup for measurement: (a) cross-sectional conceptual diagram and (b) photograph. The inset shows a single-element device placed under the arm. (c) Changes in voltage. The space was enclosed at the “set” point and opened at the “released” point.

This device can potentially also be used to monitor perspiration from the skin by placing it on the skin's surface without sealing it to ensure airflow. This microfluidic device has various potential microfluidic-based wearable sensor applications, such as in devices like wristwatches that are regularly attached and removed.

Conclusions

We constructed a humidity-driven energy harvester using a microfluidics-based design and fabrication process involving printing and lithography. The device comprises electrodes, cation-exchange membranes, and microchannels placed on a flat surface. To examine the effects of combining multiple devices in series on power generation, we designed and fabricated a single-element device as well as a multi-element device comprising ten elements connected in series.

The findings revealed the power-generating capacity of the devices in real environments. Compared with other existing devices, this device generates less power per element but achieves higher power density. The output can be increased by enlarging the device or by integrating multiple microfluidic elements. The multi-element device generated a voltage that was higher than that generated by a single-element device. Based on real-world measurements, we observed a relatively strong correlation between environmental humidity and voltage; therefore, the device might be used as a humidity sensor with the potential to generate power from sweat on the skin.

In this study, we introduce a fabrication method and structure specific to microfluidics to form a connection between open and closed channels via a thin (small cross-sectional area) long cation-exchange membrane (acting as a nanochannel). This enabled the migration of ions while suppressing the migration of water between these two channels (which reduces the power generation characteristics¹⁹).

These findings suggest that this device can be used as a power source for microfluidic-based wearable sensors, novel microfluidics-based devices that can easily be integrated via microfluidics. In the future, the shape flexibility of this device can be enhanced by replacing the glass plate with a thin plate (e.g., less than 0.1 mm thick), which will allow the glass plate to bend easily.^33–37 This suggests the possibility of designing devices for installation in locations with curved or varying surface geometries. The novel design that we present here is likely to provide a new approach for self-powered microfluidic systems, such as power supplies for microfluidic-based wearable sensors and monitoring tools for environments sensitive to humidity (e.g., clean rooms, culture rooms, and sealed lens storage boxes).

Data availability

The data supporting this article have been included as part of the ESI.†

Author contributions

HH: conceptualization, methodology, investigation, formal analysis, data curation, writing – original draft, and visualization. YK: conceptualization, methodology, investigation, writing – review & editing, and project administration.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was commissioned by the New Energy and Industrial Technology Development Organization (NEDO) under Grant No. JPNP14004. We acknowledge Satoru Suzuki for his help with device fabrication and Takahiro Miura for his help with experimental support.

References

1. R. Ehrnström, Lab Chip, 2002, 2, 26N–30N.
2. L. M. Schmidt-Speicher and K. Länge, Curr. Opin. Electrochem., 2021, 29, 100755.
3. X. F. Ma, G. Guo, X. Y. Wu, Q. Wu, F. F. Liu, H. Zhang, N. Shi and Y. M. Guan, Micromachines, 2023, 14, 972.
4. H. Kutluk, M. Viefhues and I. Constantinou, Small Sci., 2024, 4, 2300206.
5. D. Liu, Y. Wang, X. Li, M. Li, Q. Wu, Y. Song, Z. Zhu and C. Yang, Aggregate, 2022, 3, e184.
6. S. E. Alavi, S. Alharthi, S. F. Alavi, S. Z. Alavi, G. E. Zahra, A. Raza and H. E. Shahmabadi, Drug Discovery Today, 2024, 29, 103936.
7. T. He, M. Ma, H. Li, F. Zhang, F. Liu, Z. Liu and X. Li, Sens. Actuators, A, 2022, 346, 113840.
8. F. P. Gao, C. X. Liu, L. C. Zhang, T. Z. Liu, Z. Wang, Z. X. Song, H. Y. Cai, Z. Fang, J. M. Chen, J. B. Wang, M. D. Han, J. Wang, K. Lin, R. Y. Wang, M. X. Li, Q. Mei, X. B. Ma, S. L. Liang, G. Y. Gou and N. Xue, Microsyst. Nanoeng., 2023, 9, 1.
9. S. J. Kim, Y. A. Song and J. Han, Chem. Soc. Rev., 2010, 39, 912–922.
10. S. J. Kim, S. H. Ko, K. H. Kang and J. Han, Nat. Nanotechnol., 2010, 5, 297–301.
11. M. Li and R. K. Anand, Analyst, 2016, 141, 3496–3510.
12. B. Berzina and R. K. Anand, Anal. Chim. Acta, 2020, 1128, 149–173.
13. H. Hirama, R. Otahara, K. Mogi, M. Hayase, T. Torii and H. Mekaru, Polymers, 2021, 13, 1069.
14. V. Yadav, N. Chong, B. Ellis, X. Ren, S. Senapati, H. C. Chang and P. Zorlutuna, Lab Chip, 2020, 20, 4273–4284.
15. K. P. McCarthy, D. B. Go, S. Senapati and H. C. Chang, Lab Chip, 2023, 23, 285–294.
16. T. C. Tsai, C. W. Liu and R. J. Yang, Micromachines, 2016, 7, 205.
17. Y. Komazaki, K. Kanazawa, T. Nobeshima, H. Hirama, Y. Watanabe, K. Suemori and S. Uemura, Sustainable Energy Fuels, 2021, 5, 3570–3577.
18. Y. Komazaki, T. Nobeshima, H. Hirama, Y. Watanabe, K. Suemori and S. Uemura, Jpn. J. Appl. Phys., 2023, 62, SC1012.
19. Y. Komazaki, K. Kanazawa, T. Nobeshima, H. Hirama, Y. Watanabe, K. Suemori and S. Uemura, Chem. Lett., 2022, 51, 561–565.
20. Y. Xia and G. M. Whitesides, Angew. Chem., Int. Ed., 1998, 37, 550–575.
21. J. C. McDonald, D. C. Duffy, J. R. Anderson, D. T. Chiu, H. K. Wu, O. J. A. Schueller and G. M. Whitesides, Electrophoresis, 2000, 21, 27–40.
22. A. Plecis and Y. Chen, Microelectron. Eng., 2007, 84, 1265–1269.
23. Y. Zhang, T. T. Yang, K. D. Shang, F. M. Guo, Y. Y. Shang, S. L. Chang, L. C. Cui, X. L. Lu, Z. B. Jiang, J. Zhou, C. Q. Fu and Q. C. He, Nat. Commun., 2022, 13, 3484.
24. M. Y. Xiong, K. W. Song and J. P. Leburton, Nano Energy, 2023, 117, 108860.
25. A. Sahafi, J. Sobhi and Z. D. Koozehkanani, Analog Integr. Circuits Signal Process., 2013, 75, 343–348.
26. B. Tar and U. Çilingiroğlu, IEEE J. Emerg. Sel. Top. Circuits Syst., 2014, 4, 284–291.
27. Z. Tabaie and A. Omidvar, Heliyon, 2023, 9, e14707.
28. V. Candas, J. P. Libert and J. J. Vogt, Eur. J. Appl. Physiol., 1983, 50, 223–234.
29. Q. Q. Wu, J. H. Liu, L. Zhang, J. W. Zhang and L. L. Jiang, Build. Environ., 2020, 183, 107117.
30. M. Harker, Skin Pharmacol. Physiol., 2013, 26, 92–100.
31. H. Lee, J. Petrofsky, N. Shah, A. Awali, K. Shah, M. Alotaibi and J. Yim, Tohoku J. Exp. Med., 2014, 234, 117–122.
32. W. P. Cheshire and R. D. Fealey, Drug Saf., 2008, 31, 109–126.
33. Y. P. Yuan, Y. Yalikun, S. Amaya, Y. Aishan, Y. G. Shen and Y. Tanaka, Sens. Actuators, A, 2021, 321, 112604.
34. Y. S. Hao, Y. Tanaka, Y. Hosokawa, M. Li and Y. Yalikun, Appl. Phys. Express, 2022, 15, 036502.
35. B. K. Dadhich, B. Panda, M. S. Sidhu and K. P. Singh, Sci. Rep., 2023, 13, 6286.
36. Y. Hao, C. Fang, Y. Yuan, K. Okano, R. Yasukuni, S. Cheng, Y. Tanaka, Y. Hosokawa, Y. Yang, M. Li and Y. Yalikun, IEEE Sens. J., 2023, 23, 1072–1080.
37. Y. P. Yuan, D. D. Ma, X. Liu, T. Tang, M. Li, Y. Yang, Y. Yalikun and Y. Tanaka, Lab Chip, 2023, 23, 3651–3661.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4lc00958d

---