Plant Tattoo Sensor Array for Leaf Relative Water Content, Surface Temperature, and Bioelectric Potential Monitoring

This study presents a wearable plant tattoo sensor array designed for continuous monitoring of leaf temperature, relative water content, and biopotential. Current plant wearable sensor technologies often require relatively bulky substrates for sensor support and adhesives for leaf attachment, which potentially can hinder plant growth and affect long‐term measurements. The multifunctional tattoo sensor array overcomes these issues by adhering directly to the leaf surface without the need for additional supporting structures or glues. This array includes a biopotential electrode, a resistive temperature sensor, and an impedimetric water content sensor, all constructed using laminated gold‐on‐polymer thin‐film patterns. Due to their mechanical flexibility, stretchability, and conformability, the sensors can seamlessly attach to leaves via van der Waals force. Performances of these sensors are evaluated to explore plant responses under diverse growth environments. This sensor array is capable of both short‐term and long‐term monitoring, offering continuous data and detailed insights into plant physiological responses to various stress conditions.


Introduction
Given pressing climate change challenges, the adoption of precision agriculture has become increasingly vital. [1]Precision agriculture utilizes a data-centric approach to enhance decisionmaking for agricultural management at a subfield resolution cross space. [2]This method promotes the efficient use of water, fertilizers, and other chemicals, such as pesticides and herbicides; therefore, it improves resource-use efficiencies and minimizes negative environmental impacts, such as greenhouse gas DOI: 10.1002/admt.202302073emissions into the air and fertilizer runoff into water systems. [3,4][27][28][29] Recent advances in the biomedical field have focused on improving wearability, comfort, durability, and measurement modalities. [30][33][34][35][36][37][38][39][40][41][42][43][44][45][46][47][48] These sensors offer invaluable insights into the responses of plants to various biotic and abiotic stress factors.Despite this progress, most wearable plant sensors are constructed on flexible substrates, such as polydimethylsiloxane [34] poly(ethylene terephthalate), [44] and polyimide, [40,46,47] facilitating their conformal attachment to the leaf surface.These sensor substrates, however, impose physical constraints on the attached leaves; this limitation has the potential to impede gas exchange, photosynthetic activity, and transpiration on the leaf surface, potentially hindering prolonged monitoring.Therefore, there is an urgent need to design wearable plant sensors that can be integrated more seamlessly with the leaves, allowing for continuous monitoring with minimal interference with plant growth.
Utilizing wearable plant sensors to monitor leaf relative water content, bioelectric potential (biopotential), and surface temperature is crucial for understanding plant responses to various stresses and improving agricultural management.First, relative water content indicates the hydration status of plants in Figure 1.a-e) Schematic representation of the processes for fabricating and applying a plant tattoo sensor array onto a leaf.Initially, an Au layer is deposited on a PET film, which is then laminated onto a water-activated adhesive (WAA) sheet to form an Au/PET/WAA composite a).The Au/PET film is then cut out from the composite to form the patterns of three sensors using a precision cutting machine b).The Au/PET film in the unwanted areas is peered off from the WAA, leaving a biopotential electrode, a temperature sensor, and a water content sensor, on the WAA layer c).To transfer these sensors onto a leaf, water is sprayed onto the WAA layer, weakening the adhesive d).Once the support layer is removed, the sensors are attached directly to the leaf surface e).f,g) Optical images of the plant tattoo sensor arrays attached to the leaf surfaces of a maize plant f) and a Pothos plant g).
relation to their maximal water-holding capacity at full turgidity.Methods like optical reflectance, [49] near-infrared spectroscopy, [50] terahertz waves, [51] and electrical impedance [52] have been explored for measuring the leaf relative water content.However, these techniques often lack the capability for real-time, long-term monitoring or require substantial equipment costs.For instance, impedance analysis employing electrodes is a simple and popular method to measure leaf water content, but it poses challenges due to its physical impact on plant growth. [53]Moreover, plants produce biopotential signals from the movement of ions and charged particles across cell walls, which change rapidly in response to external stimuli.These variations are observable in various plant tissues and are useful for studying plant responses, especially regarding hydration and irrigation. [54]Biopotential measurement across leaf regions using conductive nanosheets has been significant in studying plant photosensitivity. [55]Additionally, leaf temperature plays a critical role in evaluating plant status, indicating key processes, such as photosynthesis and transpiration.Recently, the development of several multisensory wearable platforms has enabled continuous microclimate and physiological monitoring for plants. [10,11,34,47,48,53,56]These technological advancements aid in making informed decisions for agricultural management, such as irrigation and pest control, ultimately promoting plant growth and sustainability.
This paper reports a versatile tattoo-like sensor array that can be applied directly to a leaf surface to monitor its relative water content, surface temperature, and biopotential signals (Figure 1).Distinct from other plant wearables that often require relatively bulky substrates for support and adhesive tape for attachment, our plant tattoo sensors adhere seamlessly to the leaf without the need for additional support structures or glue.This design minimizes the physical impact of the sensor array on essential leaf functions.The array is composed of a biopotential electrode, a resistive temperature sensor, and an impedimetric water content sensor, all constructed from gold and polyethylene terephthalate (Au/PET) thin-film patterns.Owing to their high flexibility, stretchability, and conformability, these sensors can attach to the leaf surface via the van der Waals force between the Au and the leaf.The functionality and effectiveness of this sensor array are demonstrated through its application on the leaves of variously treated plants to examine the plant responses under varying illumination and water conditions.

Design and Fabrication
Figure 1a-e illustrates the overall design and manufacturing of the presented Au/PET-based tattoo sensor array and its application to a leaf surface.The biopotential electrode, resistive temperature sensor, and impedimetric water content sensor are formed by patterning Au/PET into stretchable shapes to enable conformable leaf attachment.Specifically, the biopotential electrode comprises four interconnected pads, each measuring 4 mm × 4 mm.This configuration provides sufficient surface coverage to capture bioelectronic signals from the leaf surface.These pads are connected to each other using 800 μm-wide filamentary and serpentine wires.The temperature sensor operates as a thermistor that changes its resistance with temperature; it is designed with a serpentine structure, measuring 800 μm in width and having a 1.5 mm radius.The relative water content sensor incorporates a circular pad electrode with a diameter of 1 mm and a thin ring-like electrode to facilitate electrical impedance measurements.Variations in leaf water content cause changes in the impedance of the leaf.It is noteworthy that all three sensors in the array are placed with the Au side in contact with the leaf, while the ultrathin PET layer is exposed to air.This arrangement allows attaching the sensors to the leaf through van der Waals force between the Au and leaf.
The fabrication process started with the deposition of a 10 nmthick Ti layer and a 100 nm-thick Au layer onto a 1.4 μm-thick PET film (PolyK Technologies, USA) via e-beam evaporation.The Auon-PET composite was utilized over an Au layer because of the enhanced stretchability and ease of handling.The ultrathin PET film was chosen due to its elasticity and robustness.The Au/PET film was laminated onto a water-activated adhesive (WAA; Silhouette America, USA) sheet using a laminator (Orthland, USA).After heating the combined Au/PET/WAA film at 75 °C for 2 min to reactivate the adhesive, the composite sheet was attached to a cutting mat and placed in a programmable cutting machine (Silhouette America, USA) with the Au side facing the cutting blade.The machine was set to carve a pre-defined open-mesh design using AutoCAD (Figure 1a).After cutting, the Au/PET/WAA sheet was carefully separated from the mat and then sprayed with water to deactivate the adhesive on the WAA layer.This step allowed for the removal of unnecessary Au/PET portions using tweezers (Figure 1b).Consequently, only the desired patterns for the surface temperature, biopotential, and water content sensors remained on the support layer of the WAA sheet (Figure 1c).
To transfer the formed Au/PET patterns from the WAA sheet onto a leaf, the Au/PET/WAA layer was positioned with the Au electrodes against the leaf surface and the WAA layer facing outward.The adhesive bond between the Au/PET and WAA sheet was neutralized by spraying water onto the WAA layer (Figure 1d).Subsequently, the WAA sheet was manually removed, leaving the Au/PET sensor patterns attached to the leaf surface via van der Waals force (Figure 1e).Once connected to external signal readout circuits, these tattoo sensors can provide real-time readings of leaf temperature, biopotential, and water content levels.
Each sensor was connected to its respective instrument through small lightweight alligators clipped to the Au/PET pads of the sensor.The wearable biopotential electrode was interfaced with an Arduino microcontroller for voltage measurement and was accompanied by a metal reference electrode clipped to the root of the plant.The wearable temperature sensor was connected to a multimeter (34410A, Agilent, USA) to measure changes in electrical resistance.The relative water content sensor was connected to an impedance analyzer (4294A, Agilent, USA) to perform impedance measurements at various frequencies.Additionally, to monitor relative humidity (RH) in the air near the leaf where the tattoo sensor array was attached, a commercial RH sensor (Sensirion SHT30, Mouser Electronics, USA) was placed 2 cm away from the leaf.Furthermore, to provide ground-truth temperatures at the leaf, a handheld non-contact infrared surface thermometer (IRT2DP, Tramex, USA) was utilized.

Results and Discussion
To examine the attachment of the wearable sensors on plant leaves, images were taken using environmental scanning electron microscopy (E-SEM) (Quanta-FEG 250, FEI, USA) (Figure 2).These images show that the Au/PET electrodes of the sensors adhere and conform to the leaves of maize plants.The seamless attachment was facilitated through the van der Waals force, negating the need for adhesives.
The gravimetric method was used to obtain ground-truth data on the relative water content of leaves. [57]Fresh leaf samples were stored in a tube containing distilled water.The tube was refrigerated at 4 °C for 24 h, allowing the leaves to reach full turgor.Subsequently, the leaves were removed from the tube and dried using fiber wipes.The Au/PET water content sensor was transferred to the leaf to record electrical impedance values.Leaf weight was determined by an analytical balance.Both the weight and impedance of the samples were tracked at 30 min intervals at room temperature (23 °C).Subsequently, the relative water content or RWC was calculated using RWC (%) = W−DW TW−DW × 100, where W represents the sample fresh weight, TW is the sample turgid weight, and DW is the sample dry weight.As shown in Figure 3a, the relative water content of a maize plant leaf gradually decreased over time owing to evaporative losses to the atmosphere.Concurrently, the real component of the impedance, which was measured using the attached water content sensor, was found to decrease as the frequency increased from 100 to 1000 Hz (Figure 3b).An equivalent circuit 53] was utilized as the basis for interpreting the impedance (Z) responses captured by the water content sensor (see the inset of Figure 3b).The coupling between the sensor electrode and leaf was characterized by two capacitors in parallel with two resistors.The electrical conduction pathway was modeled using constant-phase elements (CPEs) and resistances (Rs), which correspond to different physiological components: the electrode/epidermis interface (CPE E and R E ), the epidermis/extracellular fluid interface (CPE I and R I ), and the extracellular fluid present within the leaf (R F ).The impedance function, Z(), is described as follows: At lower measuring frequencies where w approaches 0, the impedance simplifies to the summer of the resistance: This equation suggests that at low frequencies, the impedance primarily reflects the cumulative resistance of the electrode/epidermis interface, the epidermis/extracellular fluid interface, and the extracellular fluid.The dominance of extracellular fluid resistance in the Z-response is attributed to the low capacitance of the cell membranes, exhibiting a poor dielectric constant at low frequencies.Consequently, the relative water content was determined by analyzing the real component of the impedance measured using the sensor.Notably, after ≈240 min (Figure 3a), the leaves began to show signs of wrinkling, indicative of drought conditions, particularly when the relative water content decreased to ≈50%.This also led to the detachment of the sensor from the leaf.The impedance values ranged from ≈450 kΩ at 1000 Hz to 10 MΩ at 100 Hz (Figure 3b).150 Hz was chosen as the optimal frequency because of the greater sensitivity and measurement reliability at lower frequencies.Figure 3c shows the correlation between the impedance measured at 120, 150, and 400 Hz and the relative water content.Owing to the leaf wrinkling and deformation resulting from the significant water loss, the effective dynamic range of the water content sensor was determined to be above 60% relative water content.Furthermore, impedance responses were observed to fluctuate more at lower frequencies such as 120 Hz, whereas higher frequencies such as 400 Hz yielded lower sensitivity (44.4 kΩ % −1 ).At 150 Hz, the sensor demonstrated an almost linear correlation between the impedance and relative water content, yielding a sensitivity of 129.4 kΩ % −1 .Therefore, a calibration curve correlating the impedance and relative water content was established at 150 Hz (Figure 3c).
The Au/PET temperature sensor features a temperature coefficient of its resistance denoted as S.This coefficient indicates the extent of resistance change corresponding to each degree of temperature change.The resistance of the sensor was determined using the formula R = R ref [ Figure 3d shows a linear relationship between the measured resistance and temperature.Therefore, the temperature sensitivity of the sensor was determined to be 0.082 Ω °C −1 .
The mechanical flexibility and stretchability of the tattoo sensors were verified through a series of deformations applied to the leaf surface, including bending, stretching, and twisting in various directions and degrees of curvature, as shown in Figure 4a.These sensors maintained a conformal and firm attachment to the leaf surface across all types of deformation.Impedance and resistance measurements were conducted for the water content sensor and the temperature sensor, respectively (Figure 4b,c).The recorded impedance and resistance values exhibited minimal change under these deformation conditions.However, as shown in Figure 4d, repeated bending of the leaf to an angle of ≈90 degrees, as displayed in image 10 of Figure 4a, over fifty-five times led to the detachment of parts of the Au/PET electrodes from the leaf surfaces.This vigorous bending likely caused significant stress to the leaf and the electrodes, weakening their adhesive bond and resulting in the electrode peeling.Despite this, the sensors demonstrated consistent performance across various deformations.
The effectiveness of the tattoo sensor array was evaluated by transferring these sensors to the leaf of a maize plant.This aimed to examine how irrigation and light conditions could impact the leaf surface temperature, biopotential, and relative water content.In addition, the commercial RH sensor was placed 2 cm off the leaf to monitor RH changes due to water exchanges occurring in the stomata of the leaf.This study was conducted in a growth chamber where illumination, air temperature, and irrigation were controlled.To ensure consistent measurements, all three sensors in the array were attached to the same leaf, reducing discrepancies caused by different measurement sites.
Figure 5a reveals two distinct peaks in biopotential immediately after the plant was watered twice at different times.The first peak ≈15:00 presented a 55 mV amplitude, while the second peak ≈15:33 had a 170 mV amplitude.The presence of the peaks might be ascribed to the uptake of water and dissolved nutrient ions, resulting in an ion concentration gradient and a subsequent increase in the biopotential.Although the biopotential reverted after water absorption, it went to a new level that was still below the level observed before the irrigation (Figure 5b,c).It was also found that several minutes after the plant was watered, the leaf surface temperature dropped by 0.1-0.2°C due to the increased transpiration during which water moved from the soil to the plant, and part of the water evaporated through the stomata, carrying the heat away from the leaf surface, and thus, leading to the evaporative cooling.As a result, the RH near the leaf was found to increase due to the increase of water evaporation into the air.Furthermore, the top panel of Figure 5a demonstrates a good agreement between the leaf surface temperatures measured with the wearable temperature sensor and those obtained from a commercial noncontact infrared thermometer.A notable advantage of the tattoo sensor over its commercial counterpart is its ability to perform continuous direct measurements on the leaf surface, in contrast to the commercial sensor which only provides discrete data at specific time intervals.
Notably, when the growth light was turned off at 16:30, a decrease in temperature and a fluctuation in biopotential were observed (Figure 5a).The temperature decline was likely attributed to reduced photosynthesis by the plant.Without illumination, the stomata of the leaf stayed closed, which might normally lead to a rise in leaf temperature; however, this potential increase was offset by the cooling effect of the surrounding air in the absence of environmental light, resulting in a noticeable reduction in leaf surface temperature.
The relative water content sensor was set to record impedance changes every 10 min at the selected frequency of 150 Hz (Figure 5d).The impedance of the sensor when unattached from the leaf and exposed to air, exhibited a value of 8.3 MΩ.Upon its attachment to the leaf surface, the impedance decreased.The bottom panel of Figure 5a shows the leaf's relative water contents were derived from the impedance values obtained in Figure 5c.There was a noticeable decrease in the relative water content recorded ≈15:00.After the irrigation with 500 mL of water, the relative water content increased to 82%, prompting the plants to transpire and release water from the leaf through the stomata.An additional 500 mL of water was applied ≈15:30.Possibly because of the overwatering condition, there was a minimal impact on the relative water content (Figure 5a; the bottom panel).As a result, only a small change in the RH of the surrounding air was observed (Figure 5a; top panel).Furthermore, switching off the light at 16:30 resulted in a decline in the relative water content, which was attributed to reduced photosynthesis and transpiration, as indicated by the lowered RH in the air surrounding the plant.Detailed analysis of the impedance changes when the light was alternatively turned on and off revealed meaningful dynamics (Figure 5a; bottom panel).Specifically, it was noticed that there was a lag of ≈5 min following the switch-off of the light, attributed to the gradual closing process of the stomata.On the contrary, when the light was turned back on, both the leaf's relative water content and the RH near the leaf increased, as the stomata reopened to facilitate transpiration.This reaction to light features the capacity of the plant to manage photosynthesis and gas exchange in response to changing light conditions.
In an extended study spanning ≈10 days, the effectiveness of the sensor array was further examined on a maize plant leaf in a controlled environment.Figure 6a displays the variations in the leaf surface temperature and the nearby RH.Notably, when the growth light was turned on, the leaf temperature rose because the heat from direct illumination outweighed the cooling effect of increased transpiration.Simultaneously, the nearby RH was found to increase due to the opening of the stomata and the release of water vapor from the leaf.Figure 6b presents the biopotential signal reading from the attached biopotential electrode.Upon irrigation, there was a ≈100 mV increase in biopotential, indicative of the plant absorbing water and nutrients from the soil.This absorption produced an ion concentration gradient, resulting in a biopotential shift.The biopotential also varied with the switching of light between the on and off states, causing light-induced and dark-triggered membrane depolarization, known as action potentials.Specifically, Figure 6b shows a transient depolarization of ≈220 mV after 18 h of darkness, followed by an extended hyperpolarization phase and another depolarization phase.A similar pattern of depolarization and hyperpolarization was noted upon re-exposure to darkness, albeit over a short duration.Figure 6c illustrates the changes in the leaf's relative water content over time, correlating with RH fluctuations near the leaf.Three periods exemplify the dynamic changes in leaf water content.In period 1 (Figure 6d), the leaf's relative water content increased when the light was turned on, indicating water absorption from the soil.This peaked at 76%, after which the rate of absorption slowed significantly.Period 2 (Figure 6e) showed a slight rise in water content following irrigation, followed by a decrease due to transpiration, reflecting the plant's initial uptake of soil water and subsequent use in photosynthesis and transpiration.These fluctuations were in line with the observed RH increase near the leaf.Finally, period 3 (Figure 6f) demonstrated a simi-lar pattern of rising relative water content and surrounding RH with the activation of the growth light.Therefore, this extended study utilizing the tattoo sensor array on the leaves provided high temporal resolution data in leaf surface temperature, biopotential, and relative water content of the plant responding to varying environmental conditions.The sensor array will contribute to a deeper understanding of plant responses to environmental stimuli, enabling potential applications in agriculture and plant biology.

Conclusion and Discussion
Integrating wearable electronics and sensors with plants presents a new opportunity to enhance our understanding of plant physiology and their interaction with the environment.We have developed a wearable tattoo-like sensor array designed for application on plant leaves, which allows for the real-time monitoring leaf's temperature, biopotential, and relative water content.The sensor array utilizes a composite layer of Au and PET, creating stretchable patterns formed on water-activated adhesive or watersoluble tape through an efficient cut-and-transfer technique.A key advantage of our sensor array is its ability to adhere to leaves via van der Waals forces between the Au and the leaf surface, eliminating the need for additional support structures or adhesive and thereby ensuring a seamless integration.Thanks to its remarkable flexibility and conformability, this sensor array operates effectively over extended periods, providing detailed insights into plant responses under various stress conditions.
Previous work on transferring sensing materials and structures primarily assisted by water-soluble tapes focused on human skin applications, [25][26][27][28][29][30] with some efforts extending to using substrates such as PDMS or polyimide to facilitate adhesion, which, however, could affect plant growth.Our work has demonstrated the successful transferring of sensors onto leaf surfaces, marking a significant expansion of tattoo-like wearable sensing electronics into the agricultural domain, given the unique texture and roughness of leaf surfaces compared to human skin.
The potential applications of tattoo-like wearable plant sensors extend beyond their current capabilities of measuring relative water content, temperature, and biopotential levels.By incorporating appropriate sensing elements designed to monitor physical (e.g., strain and light exposure) [58,59] and biochemical parameters (e.g., nutrient levels, virus presence, and volatile organic compound emissions), [21,[60][61][62] these devices could provide a holistic view of plant health, growth dynamics, and responses to environmental stresses.Such comprehensive monitoring could significantly improve agricultural practices by enhancing fertilization, irrigation, and pest control strategies.For instance, early detection of plant diseases through specific biomarkers or metabolic changes could facilitate prompt interventions, thus reducing chemical usage.Additionally, integrating wireless communication [63,64] and energy harvesting [65] technologies into these plant wearables could enable remote and continuous field monitoring.However, developing non-invasive, durable wearable sensors for plants presents significant challenges.These include ensuring sensor longevity, robustness, accuracy, efficient power use, and cost-effectiveness.An important approach to advancing tattoo-like plant wearable technology involves exploring a wider array of sensing materials and structures capable of adhering to leaf surfaces via van der Waals forces, ensuring stable attachment.[68] Furthermore, processing the vast data collected by these sensors demands sophisticated computational resources and models.Overcoming these challenges requires innovative approaches and collaborative efforts across various disciplines, yet the rewards could be transformative for agriculture, enhancing crop yields and contributing to environmental conservation.

Figure 2 .
Figure 2. E-SEM images of Au/PET electrodes on a maize plant leaf.a) Top view; b) Angled view; and c) Detailed view of the Au/PET electrode adhering to the maize leaf surface.

Figure 3 .
Figure 3. a) Gravimetric measurement for the relative water content of a maize plant leaf over time at room temperature.b) Real component of the impedance of the relative water content sensor as a function of frequency from 100 to 1000 Hz at different time instances during the gravimetric measurement in (a).The inset shows the leaf anatomy and the equivalent RC circuit representing the electrode/leaf coupling modeled as a series of two circuits of capacitances in parallel with resistances.c) Real component of the impedance of the relative water content sensor at 120, 150, and 400 Hz against the leaf relative water content.Error bars represent the standard deviation of the impedance measurements (n = 5).d) Resistance of the wearable temperature sensor as a function of temperature.
1 + S(T -T ref )], where R represents the resistance at temperature T, and R ref is the resistance at a reference temperature T ref .

Figure 4 .
Figure 4. Characterization of the plant sensors on a Pothos leaf surface under various deformations.a) Optical images of the sensors in different states: ① Unstressed state; ② Stretched perpendicular to the midrib; ③ Stretched along the midrib; ④ Twisted around the midrib; ⑤-12 Bent at varying angles and directions.b,c) Electrical impedance response of the relative water content sensor b) and resistance response of the temperature sensor to the various manipulations depicted in (a).Error bars indicate the standard deviation from the mean (n = 55).d) Partial detachment of Au/PET electrodes from the leaf surface after the leaf was bent at an angle of ≈90 degrees (see panel 10 in (a)) and then returned to its flat state for more than fifty-five times.

Figure 5 .
Figure 5. a) Top: Leaf surface temperature (left Y-axis) measured using the wearable temperature sensor in comparison to a non-contact infrared thermometer.The right Y-axis represents the RH measured by a commercial RH sensor positioned 2 cm from the leaf surface.Middle: Biopotential measured using the wearable electrode at the leaf surface.Bottom: Leaf relative water content measured using the wearable impedimetric sensor.b,c) Detailed views of the biopotential signals tracked by the wearable sensor during two scheduled irrigations.d) Impedance measured at specific time instants with 10 min intervals using the impedimetric relative water content sensor.

Figure 6 .
Figure 6.a) Long-term measurements of leaf surface temperature under varying light conditions using the wearable sensor (left Y-axis).The RH of nearby air was also monitored using a commercial RH sensor (right Y-axis).The white and grey boxes indicate the on and off states of the growth light, respectively.Also, two irrigations were scheduled.b) Leaf biopotential measured from Day 7 to Day 9. c) Leaf relative water content and RH of nearby air with the light alternately turned on and off.d-f) Detailed view of the leaf water content and RH during three distinct periods.