Flexible Self‐Powered Organic Photodetector with High Detectivity for Continuous On‐Plant Sensing

Organic photodetectors (OPDs) exhibit performance on par with inorganic detectors (e.g., Si) but can be ultrathin, ultra‐lightweight, flexible, and mechanically resilient, opening up opportunities for novel applications including optical sensors for continuous human and plant health monitoring. Here, a high‐performance flexible self‐powered OPD designed for on‐plant optical sensing is developed. The OPD employs an electrode consisting of Ag nanowires (NWs) embedded in a UV‐curable resin to achieve a flexible and thin form factor. In addition, the OPD active layer consisting of D18‐Cl and Y6 is sequentially cast to reduce dark current. The flexible OPD is sensitive to 400–950 nm wavelengths and exhibits photodetector characteristics comparable to state‐of‐the‐art rigid OPDs. The responsivity reaches values of 0.47 A W−1 and specific detectivity exceeds 1012 Jones. Owing to the embedded Ag NW electrodes in a thin substrate (t = 20 µm) and sequentially cast active layers, the detector demonstrates excellent bending stability. The photocurrent remains steady across 4000 cycles with a bending radius of 2 mm. The flexible OPD is demonstrated to effectively detect plant uptake of the rare‐earth metal terbium and sense time‐dependent chlorophyll fluorescence. Thus, this work highlights the potential for OPDs as on‐plant sensors to advance precision agriculture.


Introduction
17][18] Chlorophyll fluorescence, in particular, can be used to monitor plant photosynthetic activity, providing valuable insights into the response of plants to biotic and abiotic stresses and identifying spatial regions of infection or disease when combined with other imaging techniques. [18]Fluorescence measurements have also been utilized at the field or canopy level to assess disease at a larger scale. [19,20]Current commercial chlorophyll fluorometers typically employ relatively bulky detector configurations not designed for extended on-plant use, limiting their effectiveness.Plant biological processes change dramatically over a day and under various environmental conditions due to their endogenous biological clocks, so multi-dimensional imaging at one time per day, as is often employed in many drone-based sensing strategies, is insufficient to capture extended health characteristics. [21]In contrast, on-plant sensors attach to the plant and collect data on various environmental factors, such as light, temperature, humidity, and soil moisture levels, with demonstrations reviewed by Lee et al. [22,23] The information gathered by these sensors can be used to optimize growing conditions, track the health of the plant, and improve agricultural practices.On-plant sensors can also provide early warning signals of potential diseases or stress, allowing farmers to take timely action to mitigate these issues. [23]A previous on-plant optical sensor was demonstrated that employed a bendable fluorescent fiber to monitor chlorophyll and plant growth. [24]However, the sensor required an external spectrometer and was limited in the operating bending radius.In addition to plant health monitoring, on-plant sensors have the potential to detect the presence of specific contaminants, such as rare-earth elements (REE), to assist soil remediation. [25]REEs are used in various expanding technologies, including automotive, batteries, and renewable energy, and thus, there is increased soil contamination occurring throughout the REE life cycle. [25,26]s these technology sectors expand, maintaining soil quality through soil remediation will become increasingly critical.Onplant optical sensors can monitor the uptake of specific contaminants in plants through fluorescence detection.
Organic photodetectors (OPDs) are well-suited for continuous on-plant monitoring.[29][30][31] Organic photovoltaics have been demonstrated in ultrathin form factors with a powerto-weight density far greater than Si-based devices. [32]This opens up opportunities for the detectors to conform to non-planar surfaces, [28,33,34] and surfaces that cannot bear the weight of heavy objects (e.g., a leaf).Flexible OPDs have been fabricated utilizing flexible conductive electrodes with polyethylene terephthalate (PET), [35] polyimide, [33,36] and polyethylene naphthalate substrates. [37]However, they typically exhibit performance below that of rigid detectors, [35] remain limited in the degree of flexibility, [33,37] or require a voltage bias for effective operation. [36]ere, we demonstrated a flexible self-powered OPD capable of stable, robust, long-term detection.To achieve these target characteristics, a combination of design and fabrication techniques was implemented.A highly flexible and mechanically robust transparent conductive electrode (TCE) was fabricated by embedding Ag nanowires (NWs) in a thin (≈20 μm) UV-curable resin. [38]To ensure highly sensitive detectors (i.e., high specific detectivity), a low dark current is necessary.[44] We utilized a similar sequential cast-ing technique using the polymer D18-Cl and SMA Y6 for effective dark current suppression and increased detectivity.D18-Cl and Y6 have been demonstrated in high-performance organic photovoltaic applications, [45][46][47] with molecular structures and energy band structures illustrated in Figure 1a. [47]These organic semiconductors also have complimentary spectral absorption, as shown in Figure 1b, providing sensitivity to chlorophyll fluorescence (650 to 800 nm), as well as REE fluorescence (e.g., terbium (Tb) fluorescence ≈ 550 nm).The fabricated OPDs displayed responsivity of nearly 0.5 A W −1 , specific detectivity of over 10 12 Jones, as well as high-frequency response (>8 kHz 3 dB cutoff frequency).The Ag NW electrode and completely fabricated OPDs both showed excellent bending stability.The photocurrent and detectivity of the OPDs are shown to remain nearly constant after 4000 bending cycles at a bending radius of 2 mm.A finite element analysis provides insight into the improved bending stability when using the Ag NW electrode over commercially available indium tin oxide (ITO) coated PET.The flexible OPDs are the first demonstration of sequentially cast photoactive layers on a highly flexible TCE for effective dark current suppression, increased detectivity, and stability with bending to a small radius of curvature.The OPDs are then demonstrated to be effective onplant optical sensors, measuring time-resolved chlorophyll fluorescence in Tradescantia zebrina (inchplant) and detecting the uptake of Tb through fluorescence response in Phytolacca americana (pokeweed).These demonstrations highlight the effectiveness of the OPDs for extended optical sensing on plants.

Results and Discussion
The fabricated OPD device architecture is illustrated in Figure 1c.Control rigid OPDs were fabricated with ITO glass for the front electrode, while the flexible OPDs used a flexible TCE consisting of Ag NWs embedded in a UV resin.The design was based on a resin-embedded zinc oxide (ZnO)-encapsulated Ag NW electrode by Booth et al., which utilized a bimodal blend of Ag NWs coated with a ZnO sol-gel solution embedded in a UVcurable resin. [38]However, the OPDs demonstrated here employed a conventional architecture where the ZnO is replaced with PEDOT:PSS. [38]The result was a highly flexible electrode with a sheet resistance of ≈18 Ω □ −1 and an average visible transmittance of 80% (Figure S1, Supporting Information).The HTL PEDOT:PSS was applied to both rigid and flexible electrodes, followed by sequentially casting the photoactive semiconductors D18-Cl and Y6.Finally, the electron transport layer (ETL) of PFN-Br was spin-coated on the active layer, followed by the vacuum deposition of Ag as the back electrode.

Photodetector Performance
To evaluate the performance of the rigid and flexible OPDs, we first considered the light and dark current density across a range of bias voltage from −1 to 1 V.The results of the voltage sweep can be seen in Figure 2a, with the rigid and flexible OPDs showing nearly identical photoresponse under 100 mW cm −2 white light, and comparable dark currents.Flexible detectors typically exhibit dark currents that are an order of magnitude higher than rigid detectors. [35]The dark current was effectively suppressed in the OPDs through sequential casting of the active layer, as evidenced by comparing the dark current from rigid OPDs with a sequentially cast to a blended solvent cast active layer, shown in Figure S2, Supporting Information.The sequential casting pro-cess reduced the dark currents by over one order of magnitude compared to the blend solution cast OPD without a change in responsivity under light.This is attributed to the film morphology and local order of the blend and sequentially cast active layers, which were analyzed using atomic force microscopy and grazing incidence wide-angle X-ray scattering, seen in Figure S3, Supporting Information.The active layer surface exhibited fibrillar morphology for both the sequential and blend cast film and displayed similar local ordering seen through X-ray scattering.The external quantum efficiency (EQE) and responsivity of the rigid and flexible detectors were measured at 0 V bias with results given in Figure 2b.Both the rigid and flexible OPDs demonstrate EQEs of over 75% for much of the spectrum between 450 and 825 nm, with the rigid OPD EQE being slightly higher in the lower wavelength range ( < 700 nm).This is likely due to the slightly higher transmittance of the transparent conducting electrode and substrate in this region of the spectrum (Figure S1, Supporting Information).The responsivity of the OPDs, calculated as R = EQE ×  1240 , steadily increased between wavelengths of 400 and 800 nm.The rigid and flexible OPD reached a maximum value of 0.473 and 0.467 A W −1 , respectively, at ≈800 nm, among the highest responsivity seen in organic photodiode detectors and comparable to Si diodes. [30]All responsivities were measured at 0 V bias, demonstrating the self-powered operation of the OPDs.Specific detectivity (D*) is a figure of merit to characterize the photodetector's sensitivity compared to the noise floor normalized by area and bandwidth.This metric is particularly important for on-plant fluorescence measurements, as fluorescence intensity can be weak (e.g., <1 μW cm −2 ).D* is defined as where R is responsivity, A is photodetector active area, and S n is the noise spectral density in units of A Hz −1/2 .The noise spectral density of the detectors is reported in Figure S4, Supporting Information.The rigid OPD maintained slightly lower values due to the greater dark current suppression.The rigid and flexible OPDs exhibited D* of 5.46 × 10 13 and 3.63 × 10 12 Jones, respectively, displayed in Figure 2c.These values are among the highest seen in self-powered OPDs and are sufficient to capture the low fluorescence intensities for on-plant monitoring.D* is often reported assuming shot noise from dark current dominates the total OPD noise.With this assumption, the values are even higher at 3.65 × 10 14 and 2.43 × 10 13 Jones for the rigid and flexible OPDs, respectively.
For extended on-plant sensing, the stability of the OPDs photocurrent is crucial.The 0 V bias photocurrent of both rigid and flexible OPDs was measured regularly over 600+ h, with results shown in Figure 2d.Both OPDs demonstrated excellent stability, maintaining over 95% of their initial photocurrent during the measurements.The dark current was monitored over the same period to analyze time-dependent D* shown in Figure S5, Supporting Information.The dark current was measured at −0.5 V bias to accurately capture changes in performance, as the 0 V bias dark current can be inconsistent due to slight ambient light variation.To monitor the change in detectivity with time, D* was calculated assuming shot noise from the dark current dominated total OPD noise and that the detector's responsivity dropped at the same rate as 0 V bias photocurrent under white light.Both rigid and flexible OPDs showed a consistent drop in dark current, which resulted in a slight increase in D* after 600 h.The rigid OPD demonstrated a 5.4% increase, while the flexible OPD showed an 8.5% increase in D*.
The high response speed of the photodetectors is essential for capturing high-frequency signals and changes in the fluorescent signature with time.For example, the fluorescent lifetime of the REE Tb is ≈1 ms, [48] so a rise and fall time of the detector far below this value is necessary to accurately sense fluorescence.The photocurrent responses of the rigid and flexible OPDs were tested under a green LED modulated at increasing frequency, with the flexible OPD response shown in Figure 3a-d (rigid OPD measurements can be seen in Figure S6, Supporting Information).Rise/fall times were measured for the OPDs, displayed for a flexible OPD in Figure 3e and for a rigid OPD in Figure S6, Supporting Information.The rise/fall times for the rigid and flexible photodetector are 39/40 and 81/77 μs/μs, respectively, sufficient to detect the fast time-dependent measurements needed for fluorescence measurements.Specifically, the OPD fall time is over an order of magnitude smaller than the Tb fluorescence lifetime, so the levels of Tb in the leaves can be precisely measured.The relative response of the OPDs was measured as the frequency was systematically increased, with results shown in Figure 3f.The 3 dB cutoff frequencies for the rigid and flexible OPDs were 12.8 and 8.2 kHz, respectively.A summary of the performance characteristics of the rigid and flexible OPDs, including responsivity, D*, stability, and time-dependent response, can be seen in Table 1.

Bending Stability
The mechanical behavior of the OPD with Ag NW resin electrode was first considered using finite element analysis.Ag NWs were modeled assuming a 30 nm thick continuum composite layer with 20% loading of Ag in the resin.We compared the mechanical behavior to an ITO-based TCE on 127 μm PET substrate to determine the benefits of the embedded Ag NW approach.The bending radius of the detector was varied by placing it between two plates and reducing the distance between the plates as illustrated in Figures 4a and S7, Supporting Information.We expect the flexible OPDs to exhibit both concave and convex bending when on a plant, and FEA results suggested similar stress magnitude in either bending direction.Thus, in our analysis, we focused on concave bending only.Stresses induced by bending were captured using von Mises stress, [49] which is commonly used to determine failure stress in ductile materials.The von Mises stress in the electrode and active layer, with a decreasing bending radius, is given in Figure 4b.The stress in the active layer and electrodes was reduced by more than an order of magnitude when transitioning from the ITO-PET substrate to the Ag NW Resin electrode.Such a reduction was achieved due to the thinner substrate and replacing the thick and high Young's modulus ITO.A summary of the full FEA stress analysis can be seen in Figure S8, Supporting Information, with the full analysis parameters described in the supporting information.Bending stability for the Ag NW resin electrodes was then experimentally measured, bending up to 4000 cycles with a bending radius of 2 mm.Sheet resistance was measured regularly during the bending cycles, with results displayed in Figure 4c.The sheet resistance of the Ag NW electrode increased very slightly over the 4000 cycles, from 18 to 19.4 Ω □ −1 .This was only a 7.7% increase in the sheet resistance of the electrode, well within the range for effective OPD performance.The ITO PET electrode was also cyclically bent at a bending radius of 2 mm, with results shown in Figure 4b.After just 100 cycles, the sheet resistance increased by over 500%.The bending stability of the flexible OPDs was tested with the same experimental conditions and was found to maintain 96% of the original photocurrent after 4000 bending cycles with a 2 mm bending radius.In addition, there was a 21% reduction in dark current and a subsequent increase in specific detectivity over the bending cycles, displayed in Figure S9, Supporting Information.The excellent bending stability is attributed to the thin form factor of the flexible OPD and the use of flexible Ag NW electrodes, leading to minimal stress in each layer.Additionally, the embedded Ag NW electrodes maintain a smooth surface morphology with low roughness even after repeated bending cycles, minimizing the possibility of an increase in dark current (Figure S10, Supporting Information).

Plant Fluorescence Sensing
The performance of the flexible OPD was tested through two fluorescence demonstrations.First, fluorescence from the rare earth metal Tb, uptaken into pokeweed leaves, was measured.The presence of REEs in leaves provides valuable information about the condition of the soil and can inform the efficacy of soil remediation.Further, plants have the potential to be implemented as REE extraction tools.Detecting Tb in leaves has several factors to consider.First, any spectrally broadband autofluorescence signal will be dominated by the chlorophyll fluorescence (650-800 nm), which needs to be distinguished from the Tb fluorescence (425-625 nm).Second is the lifetime of chlorophyll and Tb fluorescence, which are ≈1 ms and 1 μs, respectively.To detect the presence of Tb in leaves, we used a modulated LED and measured the signal at 0.25 ms after the LED was switched off.Since chlorophyll's fluorescence lifetime is considerably shorter than that of Tb, we could sense the decay of Tb's fluorescence without interference from chlorophyll.The experimental setup for Tb fluorescence detection can be seen in Figure 5a.Kapton tape was used as a high-pass absorption-based excitation rejection filter to minimize the signal from the LED source, while also providing low autofluorescence. [50,51]For statistically significant detection, we modulated the LED for 100 on/off cycles and captured the normalized response.The contaminated pokeweed was grown in soil that included 50 mm TbCl 3 in water and was compared to plants grown in soil without TbCl 3 .The mean signal, along with the associated 95% confidence interval can be seen in Figure 5b, showing a significantly higher signal in the leaf exposed to the 50 mm Tb solution.To ensure the light we detected was from Tb fluorescence, we also measured the mean signal and 95% confidence interval 2 ms after the LED switched off.Theoretically, these values should be equivalent between the two leaves, as these measurements were taken after the lifetime of Tb fluorescence.The results can be seen in Figure 5b, showing an insignificant difference between the 0 and 50 mm signals.This confirms the presence of Tb in the pokeweed plant grown with 50 mm TbCl 3 solution, demonstrating the efficacy of the flexible OPD as a detector for rare earth metal plant uptake.
The flexible OPD was also used to capture the time-dependent chlorophyll fluorescence of an inchplant leaf while laminated directly on the leaf surface of a living plant, with the experimental setup and results seen in Figures 5c,d and S11, Supporting Information.The response of chlorophyll fluorescence can be seen for a dark-adapted leaf in two locations with clear differences in chlorophyll density (pictured in Figure 5c).The time-dependent response followed the expected behavior for a leaf, exhibiting an initial peak in fluorescence intensity followed by a gradual drop to a steady-state response. [52]Additionally, there was stronger chlorophyll fluorescence in the leaf region with higher chlorophyll density.Chlorophyll fluorescence is widely used in plant health monitoring at the cellular, leaf, and whole plant scales.Interactions with pests and pathogens can alter the metabolic function of plants, and detecting the changes in chlorophyll fluorescence can inform the condition of plants at multiple scales.Employing varying LED configurations would make gathering the various chlorophyll fluorescence characteristics possible. [53]

Conclusion
Highly flexible, robust OPDs were fabricated displaying high responsivity, D*, and stable operation.The highly flexible and lightweight OPDs showed detector characteristics on par with state-of-the-art rigid OPDs while retaining high performance over repeated bending cycles, highlighting their compatibility with extended testing on plants.High detectivity and selfpowered operation were realized using a sequential casting of the high-performing active layer materials, suppressing the dark current while maintaining high responsivity for high D* at 0 V bias.The thin form factor of the Ag NW resin electrode reduced stress in the film under bending, leading to stable performance after 4000 bending cycles.The conformal nature of the detector enabled placing the detector on a leaf without significantly changing the leaf's natural position.The effective detection of low-intensity fluorescence signals was then highlighted in two demonstrations: measuring fluorescence from the REE Tb using time-gating and measuring spatial differences in chlorophyll fluorescence across a leaf.
As the population on earth continuously increases, the need for efficient and sustainable agriculture is at an all-time high.Precision farming and plant phenotyping can help improve agriculture output with the assistance of sensors and detection methods to monitor plant health changes due to environmental stress, disease, or pathogens.Our conformable on-plant OPDs can be employed on leaves for extended time with minimal impact on plant growth for continuous monitoring of plant conditions.Organic electronics are persistently advancing with ultralight and flexible form factors, exhibiting power-to-weight densities that surpass the capabilities of inorganic materials.We envision the combination of conformal OPDs with conformal OLEDS or micro-LEDS to achieve in-field health monitoring using fluorescence detec-tion schemes with a minimal operating burden to advance agricultural productivity and sustainability.

Experimental Section
Materials: D18-Cl, Y6, and PFN-Br were purchased from 1-Material and used without modification.PEDOT:PSS was purchased from Heraeus and used without modification.The D18-Cl solution was prepared with a concentration of 7 mg mL −1 in 0.95:0.05chlorobenzene:tetrahydrofuran and mixed overnight at 80 °C.Y6 was prepared with a concentration of 9 mg mL −1 in chloroform and mixed overnight at room temperature.PFN-Br was prepared with a concentration of 0.5 mg mL −1 in methanol and mixed for 2 h at room temperature before casting.High aspect ratio Ag NWs (D = 30 nm, L = 100-200 μm) dissolved in IPA at a concentration of 20 mg mL −1 were purchased from ACS Material, LLC.Low aspect ratio Ag NWs (D = 60 nm, L = 20-30 μm) dissolved in IPA at a concentration of 10 mg mL −1 were synthesized as described previously. [38]The UV Resin used was Norland Optical Adhesive (NOA) 63, purchased from Edward Optics.ITO-coated PET (5 mil) was purchased from Sigma Aldrich.
Device Fabrication: Sequentially cast OPDs were fabricated with the device structure consisting of TCE/PEDOT:PSS/D18-Cl/Y6/PFN-Br/Ag.Rigid devices utilized ITO-coated glass as the TCE.ITO glass substrates were cleaned with DI water, acetone, and isopropanol for 15 min, followed by UV-Ozone treatment for 20 min.Flexible TCE fabrication is described in detail elsewhere, [38] with the exception of the PEDOT:PSS replacing the protecting ZnO layer on the Ag NWs.The bimodal Ag NW solution was 1:1 high:low aspect ratio Ag NWs with a total concentration of 6.67 mg mL −1 .The Ag NWs were spin-coated on an n-Octadecyltrichlorosilane (OTS)coated glass substrate at 2500 rpm for 30 s and annealed at 150 °C for 10 min.The Ag NWs were patterned to the desired electrode shape using Scotch tape.PEDOT:PSS was spin-coated at 4000 rpm for 30 s and annealed at 150 °C for 15 min.NOA 63 was spin-coated at 3000 rpm using slow acceleration and cured under a 365 nm UV lamp for 30 min, leading to a film thickness of ≈20 μm.Once cured, the electrode was removed from the fabrication glass using a blade and placed upon PDMScoated glass for subsequent fabrication steps.For both rigid and flexible devices, PEDOT:PSS was spin-coated at 4000 rpm and annealed at 150 °C for 15 min, followed by D18-Cl (2500 rpm), Y6 (4000 rpm), and PFN-Br (3000 rpm).100 nm Ag was vacuum deposited under ≈1 × 10 −6 mbar through a shadow mask (device area of 0.076 cm 2 ).UV resin was spincoated onto the fabricated device for encapsulation.
Film and Device Characterization: Absorbance spectra of the films and transmittance spectra of the flexible electrodes were obtained using UVvis spectroscopy with an Ocean Optics Jazz spectrometer.The calculation of average visible transmittance is AVT = ∫ T()P()b s ()d ∫ P()b s ()d , where T() is transmittance, P() is the photopic response of the human eye and b s () is the AM1.5 solar spectrum.The responsivity/EQE was measured using a Newport 67 011 light source, Oriel Cornerstone 130 monochromator, Keithley 2700 Data Acquisition System, SR570 Current Preamplifier, and SR810 Lock-in Amplifier.The noise spectral density was measured using an SR570 Current Preamplifier and Picoscope 4424 PC oscilloscope.The noise density frequency was selected near the 3 dB cutoff frequency for the rigid and flexible OPDs.Photocurrent stability was measured in a N 2 controlled environment, with samples kept in darkness between measurements.Frequency response was measured using a 532 nm LED modulated using a BK Precision 4017A function generator and measured using an SR570 Current Preamplifier and PicoScope 4424 PC Oscilloscope.AFM images were obtained using an Asylum MFP-3D-BIO.Physical bending tests were conducted on a custom bending stage with a bending frequency of ≈1 cycle/s.The bending tests were conducted in ambient atmosphere.Finite Element Analysis: To simulate the physical bending of the flexible OPD, an adaptive curvature model was created in ANSYS 2022 R2.The geometry of two plates and the flexible OPD was made in SolidWorks.The flexible OPD was subjected to an initial bending radius of 1000 mm to ensure the composite began bending in the correct direction upon plate displacement.The two plates were 5 mm wide, 100 mm long, and 1 mm thick and modeled as rigid.In between two plates, the composite flexible OPD was bent displacing a plate downward while keeping the bottom plate fixed.The top plate was displaced downward to a maximum of 99 mm to create a 0.5 mm radius of curvature for the OPD.The composite was kept in place by clamping the two edges of the OPD to the smaller ends of the plates.Full FEA details, including material properties, composite layer thicknesses, and analysis parameters, are described in the Supporting Information.
Fluorescence Measurements: For Tb fluorescence experiments, the exciting LED was 365 nm wavelength.The flexible OPD was covered with Kapton tape to act as a high-pass filter.The LED was pulsed with a square wave at 10 Hz, with the output response continuously being recorded using a PicoScope 4424 PC Oscilloscope.The low fluorescence signals were amplified using a ThorLabs AMP100 transimpedance amplifier at 1 MV A −1 .The normalized signal was obtained at 0.25 and 2 ms after the LED was switched off.The data from 100 cycles was recorded, and a mean and 90% confidence interval was calculated.The relative response between the 0 and 50 mm leaves was recorded for comparison.At 0.25 ms, the 50 mm leaf was expected to show a higher signal than the 0 mm leaf due to Tb fluorescence.At 2 ms, the 0 and 50 mm leaves were expected to show the same signal, as 2 ms is longer than the fluorescence lifetime of Tb.For the chlorophyll fluorescence experiments, a 450 nm LED was used to excite the chlorophyll.Kapton tape was used to isolate the chlorophyll fluorescence from the exciting LED.The response was recorded using a PicoScope 4424 PC Oscilloscope.

Figure 1 .
Figure 1.a) Chemical structure and energy band structure of donor (D18-Cl) and acceptor (Y6) materials.b) Absorbance spectra of D18-Cl, Y6, and sequential cast blend.c) Flexible OPD structure.d) Photograph of a flexible OPD laminated on a leaf.

Figure 2 .
Figure 2. Rigid and flexible OPD characteristics.a) Absolute current density versus voltage in the dark and under white light (100 mA cm −2 ), b) responsivity and EQE at 0 V bias, c) specific detectivity at 0 V bias, and d) normalized photocurrent stability.

Figure 3 .
Figure 3. Photocurrent response of flexible OPD with green LED at a) 16 Hz, b) 128 Hz, c) 1 kHz, and d) 8.2 kHz.e) Rise and fall time of flexible OPD with green LED (0-90%).f) Frequency response of rigid and flexible OPD with green LED.The dashed line shows the 3 dB cutoff frequency.

Figure 4 .
Figure 4. a) Image of flexible OPD FEA stress analysis with a bending radius of 10 mm.b) Finite element stress analysis of flexible OPD with varying bending radius.Dashed lines refer to the OPD device structure with ITO PET as the TCE, while solid lines refer to the OPD device structure with Ag NW UV resin TCE.c) Sheet resistance stability of the Ag NW UV resin TCE and ITO PET TCE under 4000 bending cycles.d) Photocurrent stability under white light at 100 mA cm −2 for flexible OPD under 4000 bending cycles.The bending radius is 2 mm.The inset and right pictures show the OPD in the experimental bending stage.

Figure 5 .
Figure 5. a) Experimental setup for Tb fluorescence measurements.Inset describes the ideal experimental behavior.b) Mean and 95% confidence interval of the relative fluorescence response between pokeweed grown with Tb solutions of 0 and 50 mm at two different times after the 360 nm LED is switched off.The sample size is 100 LED cycles.c) Experimental setup for chlorophyll fluorescence measurements.Inset, picture indicating the two locations on the inchplant leaf tested with different coloration indicative of different chlorophyll densities.d) Time-dependent chlorophyll fluorescence of inchplant for two leaf locations with a 450 nm incident LED.