Highly Efficient Capture and Quantification of the Airborne Fungal Pathogen Sclerotinia sclerotiorum Employing a Nanoelectrode-Activated Microwell Array.

In this study, we present a microdevice for the capture and quantification of Sclerotinia sclerotiorum spores, pathogenic agents of one of the most harmful infectious diseases of crops, Sclerotinia stem rot. The early prognosis of an outbreak is critical to avoid severe economic losses and can be achieved by the detection of a small number of airborne spores. However, the current lack of simple and effective methods to quantify fungal airborne pathogens has hindered the development of an accurate early warning system. We developed a device that remedies these limitations based on a microfluidic design that contains a nanothick aluminum electrode structure integrated with a picoliter well array for dielectrophoresis-driven capture of spores and on-chip quantitative detection employing impedimetric sensing. Based on experimental results, we demonstrated a highly efficient spore trapping rate of more than 90% with an effective impedimetric sensing method that allowed the spore quantification of each column in the array and achieved a sensitivity of 2%/spore at 5 kHz and 1.6%/spore at 20 kHz, enabling single spore detection. We envision that our device will contribute to the development of a low-cost microfluidic platform that could be integrated into an infectious plant disease forecasting tool for crop protection.

Introduction

Infectious plant diseases caused by microorganisms such as fungi and bacteria are one of the main factors affecting crop production, resulting in huge economic losses to farmers and growers.[1−3] Among the list of numerous diseases, Sclerotinia stem rot (SSR) is of particular importance due to its wide host range and harmful effects.[4,5] SSR is caused by the necrotrophic fungal pathogen Sclerotinia sclerotiorum, affecting more than 400 plant species worldwide, including several economically important crops such as canola, soybean, sunflower, and carrot.[4,6,7] SSR, also commonly known as white mold, is particularly devastating to the canola industry, the world’s second-largest oilseed crop.[8] Yield losses due to SSR can be as high as 50%, causing severe financial losses[9] and making it the greatest threat to canola production. Microscopic spores produced by the fungus disseminate throughout the fields in wind currents, representing the primary source of inoculum initiating SSR epidemics.[6,10] Currently, chemical control employing fungicides is the main strategy for the management of SSR.[11] Although this approach can be highly effective, fungicides are economically inefficient when applied routinely and with no indication of disease risk. Ideally, farmers must apply fungicides during specific time frames and only when necessary, that is, when spores are present in the field but before symptoms are visible.[11,12] However, as SSR outbreaks are hard to predict, farmers typically apply fungicides routinely and without any objective information on the risk of SSR development, a decision that costs time, drastically reduces the profits, and affects the environment.

Current methods for predicting SSR development are imprecise. Risk assessment checklists[13,14] and weather-based forecasting models[15−17] were the first systems developed for this purpose. Although simple and field-specific, the checklists are time-consuming, labor-intensive, and do not include any measurement of airborne inoculum. On the other hand, weather-based systems lack field specificity and are based exclusively on weather parameters. The measurement of airborne inoculum levels is a critical factor influencing the accuracy of SSR forecasting systems, as it indicates the amount of pathogen available for its development. Another popular strategy to assess the risk of SSR is through petal infestation measurements. The identification of spores in plant petals using the agar test is the most common approach.[18−20] The percentage of petals infested is then correlated to the risk of disease development.[18] A major limitation of this method is the incubation time delay, which can be significant and especially harmful when weather conditions are favorable for SSR development.[12] Molecular diagnostic methods such as polymerase chain reaction (PCR) can also be used to measure airborne inoculum,[21−23] by analyzing either plant petals collected from the field or samples obtained through air sampling using commercially available spore-traps that are installed in the field. Although measurements obtained with PCR are relatively accurate and can reduce the time delay encountered with the agar test, farmers still need to ship the sample to a testing laboratory. In addition, they are expensive and bulky, limiting their usefulness for on-site applications.

A platform that can efficiently measure levels of S. sclerotiorum airborne inoculum with features such as portability, simplicity, and low cost is deemed an essential component in the development of a reliable on-site SSR forecasting system and is yet to be developed. Nano- and micromachined devices emerge as an alternative to provide such a platform.[20,24] In this study, we designed a microfluidic device based on a nanoelectrode-activated microwell array for the capture and quantification of S. sclerotiorum spores. Microwell arrays have been commonly used for high-throughput cell sequencing[25−29] and cell pairing,[30,31] as well as for cancer cell identification and characterization.[32−35] These devices are normally designed with thousands of microwells, and cells are generally detected and analyzed by imaging and microscopy. One common approach is to capture particles or cells into the microwells using gravity.[28,36−38] To increase the capture efficiency, as well as to obtain selective capture, active trapping mechanisms such as dielectrophoresis (DEP) have been used.[30,39−42] DEP is an electrokinetic phenomenon widely employed for the selective manipulation of polarizable particles such as bacterial and mammalian cells within a spatially nonuniform electric field.[43−45] Recent reports of devices using microwell array and DEP include a microfluidic device composed of 3600 microwells for double-sub-Poisson single-cell RNA sequencing,[25] an electroactive device with 300 000 microwells for the molecular analysis of tumor cells,[34] and a microfluidic chip with more than 3000 microwells for the capture and subsequent analysis of cancer cells, including the characterization of cell apoptosis via immunostaining and fluorescence in situ hybridization (FISH).[32]

Here, we developed an inexpensive and portable microfluidic device with a total of 190 picoliter wells, which were fabricated on top of coplanar nanothick aluminum electrodes. These electrodes were employed for the dielectrophoresis-driven capture of spores and subsequent on-chip detection using non-faradaic electrochemical impedance spectroscopy (nF-EIS). We extensively characterized our device and unambiguously demonstrated a spore trapping rate of more than 90%. The impedimetric quantification of single spores was also demonstrated, with a platform design that allows us to address each column in the microwell array individually. The device presented here provides a unique approach for the capture and quantification of S. sclerotiorum airborne inoculum and a solution of value to the agricultural sector. We believe that the novel application of our device will leverage the development of a platform technology for infectious plant disease prevention.

Results and Discussion

Design and Operating Principle of the Device

Our microfluidic device (Figure a) was designed and fabricated with a total of 20 aluminum nanoelectrodes (100 nm thick, 20 μm wide, and 6 μm gap) and upon which an array of 190 microwells made of SU-8 resist was fabricated (Figure b,c). Each electrode can be addressed individually, and between each pair, 10 microwells were placed. To focus the flow of spores toward the center of the device, in which the electrodes were placed, the microfluidic channel was designed with a constriction and each alternate column in the microwell array was shifted in the y-axis direction to ensure that a microwell is always under the path of a flowing spore (Figure b).

Figure 1

Microfluidic device based on a nanoelectrode-activated microwell array. (a) Assembled microfluidic device. (b) Microscopic image of nanoelectrodes and microwells. The array has 190 microwells in total, 10 microwells per electrode pair (column) with 20 aluminum nanoelectrodes in total. (c) Helium ion microscopy (HIM) image of microwells made of SU-8. (d) Schematic representation of the cross-sectional view of the assembled device. (e) Nanoelectrodes can be combined into an interdigitated structure for spore capture in all of the microwells of the device using DEP. (f) After DEP capture, nanoelectrodes are operated individually for electrochemical impedance spectroscopy (EIS) measurements column by column.

According to previous reports,[20,24,46] the detection of ∼10 spores/m3 of air allows an 8-day advanced forecast of SSR outbreaks. This threshold expressed in, for example, spores per milliliters of solution will, of course, depend on the collection volume of the air-sampling system. In our device, we chose a low number of microwells based on this detection range. On the other hand, the diameter and depth of microwells (Figure d) were designed based on the average size of S. sclerotiorum spores, which have ellipsoidal shapes, in the range of around 2–5 μm × 7–15 μm (Figure S1). The size of the spores is in agreement with what was previously reported.[47,48] The operating principle to capture S. sclerotiorum spores into our microwells is based on the DEP-induced force. The DEP force acting on a polarizable particle such as a cell or spore arises from the interaction between a nonuniform electric field and the particle’s induced dipole. Depending on the dielectric properties of the particle and the surrounding medium, the DEP force can be defined as positive or negative, depending on whether the particle is attracted toward the region of maximum electric field gradient or repelled from it.[49] The time-averaged DEP force FDEP acting on a particle of radius r is given bywhere E is the amplitude of the electric field and εp* and εm* are the relative complex permittivities of the particle and medium, respectively, each given by ε* = ε + σ/jω, where ω is the frequency of the applied electric field, ε the permittivity, σ is the conductivity of the particle or medium, and j = √(−1). The nonuniform electric field for DEP capture is generated by applying a sinusoidal voltage to the electrodes, which can be configured as an interdigitated electrode (IDE) structure during the process of trapping spores into the microwells (Figure e). The electrode configuration is externally controlled by switches in a custom-made chip holder (Figure S2). Once the spores are captured, non-faradaic electrochemical impedance spectroscopy (nF-EIS) measurements can be performed to quantify the spores column by column. nF-EIS is a label-free detection technique that measures the electrical current of an electrode–electrolyte system in response to an applied AC potential with no redox species in solution.[50−52] During nF-EIS measurements (Figure f), nanoelectrodes can be operated individually, allowing the measurement of each column in the microwell array and providing the ability to determine spore occupancy per column.

DEP-Assisted Capture of Spores

To evaluate the performance of our device to capture S. sclerotiorum spores, two sets of experiments were designed. In the first set, the occupancy distribution of the captured spores was examined as a function of the applied flow rate. Devices with microwell diameters of 20 and 15 μm were tested. Prior to loading spores, ethanol was slowly injected into the device at a flow rate of 2 μL/min to remove air bubbles within the microwells. Afterward, our DEP buffer was introduced into the channel at the same flow rate for 10 min. Subsequently, 10 μL of stained S. sclerotiorum spores at a low concentration of ∼4.4 × 104 spores/mL was pumped into the device at a fixed flow rate of 0.2 μL/min.

While the spore suspension was flowing, a sinusoidal signal of 20 Vpp and 300 kHz was applied to the electrodes to enable a positive DEP capture. Once the entire spore volume (10 μL) was pumped, a washing step was implemented to remove the remaining spores on the SU-8 surface by increasing the flow rate to 15 μL/min for 2 min while keeping the DEP signal on. Flow rates of 0.4 and 0.8 μL/min were also examined, and three independent experiments were performed for each flow rate. A high spore occupancy of 91.23% was achieved in devices with microwell diameters of 20 μm, subjected to a flow rate of 0.2 μL/min (Figure a), and with more than 70% of the microwells occupied by at least two spores. As expected, spore occupancy decreases as the flow rate increases, with an average occupancy of 83.11 and 71.86% for flow rates of 0.4 and 0.8 μL/min, respectively. It was also verified that the percentage of single spores increased from 16.49 to 24.38% when the flow rate increased from 0.2 to 0.8 μL/min. The average velocity of spores increases as the flow rate increases and thus the drag force acting on them, decreasing the number of spores captured with the same DEP signal amplitude. The same trend was observed in devices with microwells of a 15 μm diameter (Figure b) and almost no triplets were present after the washing step.

Figure 2

Spore occupancy employing DEP-assisted loading. Each bar shows the mean value of three independent experiments and represents the number of microwells occupied by zero, one, two, and three or more spores. The total number of microwells is 190. (a) Spore occupancy as a function of the applied flow rate for devices in which microwells have a 20 μm diameter and employ a sinusoidal DEP signal of 20 Vpp at 300 kHz. (b) Same as (a) but with microwells of a 15 μm diameter. (c) Spore occupancy as a function of microwell depth: the diameter of the microwells is 20 μm, and the applied DEP signal is 20 Vpp at 300 kHz. (d) Same as (c) but with microwells of a 15 μm diameter. Error bars represent the standard deviation.

The difference in occupancy levels between the two devices can be attributed to two main reasons: a reduction in the cross-sectional area of the channel due to a smaller constriction and a larger attenuation of the electric field due to smaller microwells. The microfluidic channel constriction (Figure b) is 420 and 320 μm for devices with microwells of 20 and 15 μm diameter, respectively. By the continuity equation,[53] we know that as the cross-sectional area of our microchannel is reduced, the mean velocity of spores increases for a constant flow rate. Therefore, for the same flow rate, spores are flowing around 32% faster in devices with microwells of a 15 μm diameter. Furthermore, SU-8 is an insulator, and as the diameter of the microwells is reduced, a larger attenuation of the electric field and thus the DEP effect is obtained. Video S1 shows the process of spore capture, and the fluorescence microscopy images of the spores captured at different flow rates are shown in Figure S3. In the second set of experiments, the influence of microwell depth with regard to the occupancy distribution of the captured spores was evaluated at a constant flow rate of 0.2 μL/min. In addition to the previously described chips showcasing a microwell depth of 10 μm, devices with microwell depths of 5 and 20 μm were also fabricated. The DEP signal amplitude and frequency, the concentration of spores, and the injected volume were kept the same as in the previous experiment. As expected, a high spore occupancy (92.35%) was observed in devices with microwells of a 20 μm diameter and 5 μm depth (Figure c). On the other hand, the occupancy in devices with microwells of a 15 μm diameter and 5 μm depth was on average 79% with more than 45% of the microwells occupied by doublets (Figure d).

During the washing step of devices with 5 μm depth, a few spores were observed escaping the microwells, particularly those in which there were two or more spores already. This issue was solved by increasing the voltage to 22 Vpp during the washing step. Notably, the spore capture efficiency decreased drastically for both microwell diameters when the depth of microwells was 20 μm (Figure c,d). This can be attributed to the decrease in the intensity of the electric field gradient with the increasing distance from the electrode surface. We carried out numerical simulations using commercially available software (COMSOL Multiphysics) to investigate this effect. The simulation domain and the electric field generated by the electrodes are shown in Figure S4a, where the electric field intensity |E| is represented with a color map. As per eq , the gradient of the squared electric field ∇|E|2 is directly proportional to the DEP force acting on the spores; therefore, we defined the effective gradient as ∇|E|eff2 = ∫∇|E|2dA, which represents the magnitude of the electric field gradient (∇|E|2) integrated over the white dashed rectangle of area A above the microwells (Figure S4a), with a fixed height of 10 μm and width given by the microwell diameter. The simulation of the effective gradient as a function of SU-8 thickness (Figure S4b) shows that as the thickness of the SU-8 layer increases, the effective gradient decreases exponentially, and therefore, the DEP force also decreases.

This means that as the microwell depth increases, the DEP capture efficiency will decrease exponentially, which was clearly observed in our experiments. It is also important to point out that we also tested spore loading in all our devices using nothing but gravity. However, in all cases, all of the microwells were empty, which indicates that the DEP force was the dominant force in capturing the spores into microwells. Figure shows the spores captured inside microwells using DEP. Based on these results, it is evident that our devices can effectively capture spores into the microwells using DEP. Devices with a 20 μm diameter and 5 μm depth provided higher capture efficiency and flexibility to account for the spore size variability. Devices with a 15 μm diameter and 5 μm depth provided a lower number of spores per microwell when compared to devices with microwells of a 20 μm diameter, and the overall capture efficiency could be increased by, for instance, reducing the flow rate, at the expense of an increase in capture time.

Figure 3

HIM images of S. sclerotiorum spores inside microwells captured by the dielectrophoresis force. (a) Single spore in a microwell of a 20 μm diameter; (b) single spore in a single microwell of a 15 μm diameter. The applied DEP signal was 20 Vpp at 300 kHz.

Impedimetric Quantification of S. sclerotiorum Spores

After capturing the spores, we employed nF-EIS to quantify them. During this process, nanoelectrodes are operated individually (Figure f), allowing for impedance measurements of each column in the microwell array by applying an AC potential from the impedance analyzer to the respective pair of electrodes located beneath each microwell (Figure b). To test the performance of our quantification method, we first pumped a solution with a low concentration of spores (∼2 × 104 spores/mL) at a flow rate of 0.2 μL/min into the device. Subsequently, we used DEP to capture single spores into a fixed column of the microwell array. Every time a single spore was captured in a microwell of the respective fixed column, the DEP signal was turned off and the impedance spectrum was recorded from 5 kHz to 1 MHz. This process of capturing spores in a fixed column is shown in Figure S5. Flowing spores were monitored using the microscope camera to ensure that no spore was captured in a microwell that was already occupied. Thus, the DEP signal was turned back on only when a spore was flowing in the direction of an empty microwell of the same column. During the impedance spectrum measurement, which takes ∼20 s, none of the captured spores escaped from the microwells, even when the DEP signal remained off for a longer time. The typical magnitude and phase response of a single column in the microwell array as a function of the number of captured spores and for devices with microwells of a 20 μm diameter is shown in Figure a,b. As expected in a capacitive-based sensor, magnitude curves decrease (Figure a) and phase curves tend to −90° (Figure b) as the frequency increases. Figure c depicts the experimental Nyquist plot and the equivalent circuit model with such responses. Rm models the solution resistance, which is in parallel with the solution capacitance Cs. The constant phase element (CPE) models the electrical double-layer at the electrodes, all in parallel with a parasitic capacitance Cp, which accounts for parasitic effects introduced by the connection cables, chip holder, and substrate. The simplified equivalent circuit is well-studied and commonly used to describe the electrode–electrolyte interface in interdigitated electrode sensors.[54] By fitting the experimental data to the equivalent circuit (Figure S6), we verified that the captured spores (Figure ) will mainly induce changes in Rm, Cs, and CPE, contributing to the total impedance change of the system. These curves clearly show that changes in the number of captured spores can effectively modulate the impedance response of our device. Seven spores per column were captured since, as the microwells of a single column were filled, it became more difficult to prevent two spores from occupying the same microwell. To account for variations in the impedance response of each column in the microwell array, we defined the normalized impedance aswhere Zspore is the impedance response due to the captured spores in microwells and Zbuffer is the impedance response given by the buffer. Figure d shows the calibration curve for the normalized impedance magnitude at frequencies of 5 and 20 kHz, in which each point represents the average value of three independent experiments (N = 3) performed in different columns of the microwell array. By fitting the experimental data with a linear regression, an R2 value of 0.9390 with a slope of 0.02 was obtained for 5 kHz. However, for 20 kHz, the R2 value was 0.940 with a slope of 0.016. We can express the sensitivity as the percentage of impedance change per captured spore by simply multiplying the slopes by 100, yielding a sensitivity of 2%/spore at 5 kHz and 1.6%/spore at 20 kHz. These curves clearly indicate that it is possible to quantify spores with our device and that the sensitivity decreases with frequency, which was expected by the spectra obtained in Figure a. The linearity, on the other hand, was very similar for both frequencies. During our DEP experiments, we observed that more than one spore could be captured in a single microwell, and thus, we carried out an experiment to determine the impedance change due to spores in a single microwell. The calibration curve (N = 3) for the normalized impedance magnitude is shown in Figure e. No more than four spores could be captured in the same microwell; the R2 and slope values obtained for 5 kHz were 0.8061 and 0.017, respectively, while the values for 20 kHz were 0.7869 and 0.007, respectively. The microwell dimensions limit the number of spores that can be captured, and as the number of spores increases, the area of exposed electrodes is reduced. Furthermore, spores tend to stack on top of each other partially or completely as the number of spores increases, thus reducing the electric field perturbation caused by these spores, which leads to a reduction in the total impedance change. As such, these specific measurements deviate from a simple linear correlation and a more complicated model would have to be used to improve the coefficient of determination.

Figure 4

Impedimetric quantification of the captured spores in devices with microwells of a 20 μm diameter. (a) Magnitude of impedance versus frequency as a function of the number of single spores captured in a column of the microwell array. (b) Phase of impedance versus frequency. (c) Nyquist plot with an equivalent circuit model. (d) Calibration curve of the normalized impedance magnitude at 5 and 20 kHz (N = 3). (e) Calibration curve of the normalized impedance magnitude at 5 and 20 kHz when spores are captured in a single microwell (N = 3). (f) Calibration curve of the normalized impedance phase at 50 and 20 kHz (N = 3). Arrows in (a–c) indicate the shift given by the increase in the number of spores. Error bars in (d–f) represent the standard deviation.

Lastly, we also noticed that the impedance phase could also be employed to quantify spores, especially in the mid-frequency range. The calibration curve (N = 3) for the normalized impedance phase is shown in Figure f. At 50 kHz, an R2 value of 0.9368 was obtained with a slope of 0.007, while for 100 kHz, the R2 was 0.9284 with a slope of 0.005. Although the impedance phase is less sensitive than magnitude, it could still be employed as an alternative parameter to quantify spores. We repeated the same experiments for chips with microwells of a 15 μm diameter (Figure ). With a smaller diameter, a smaller portion of the electrode surface area is in contact with the solution, yielding considerably larger magnitude values (Figure a,c) when compared to microwells of a 20 μm diameter (Figure a,c). The calibration curve for the normalized impedance magnitude (N = 3) at frequencies of 5 and 20 kHz is shown in Figure d. A sensitivity of 0.7%/spore at 5 kHz and 0.4%/spore at 20 kHz was obtained. When capturing spores in a single microwell, no more than three spores could be captured, obtaining a calibration curve with a slope of 0.003 at 5 kHz and 0.001 at 20 kHz (Figure e). Lastly, the calibration curve (N = 3) for the normalized impedance phase is shown in Figure f. At 50 kHz, an R2 value of 0.9736 was obtained with a slope of 0.002, while for 100 kHz, the R2 was 0.9750 with a slope of 0.001. The sensitivities achieved are larger than the basic accuracy of our measurement instrument (0.05%) for both microwell diameters, validating the reliability of our results. The lower sensitivity presented by devices with microwells of a 15 μm diameter can be mainly attributed to the reduction of the exposed surface area of the measuring nanoelectrodes, which naturally increases the double-layer impedance or equivalently reduces the double-layer capacitance of the measuring system. This is a well-known effect for reducing the sensitivity in impedance measurements.[55] According to these results, devices with 20 μm diameter microwells are considerably more sensitive for spore quantification.

Figure 5

Impedimetric quantification of the captured spores in devices with microwells of a 15 μm diameter. (a) Magnitude of impedance versus frequency as a function of the number of spores captured in a column of the microwell array. (b) Phase of impedance versus frequency. (c) Nyquist plot with an equivalent circuit model. (d) Calibration curve of the normalized impedance magnitude at 5 and 20 kHz (N = 3). (e) Calibration curve of the normalized impedance magnitude at 5 and 20 kHz when the spores are captured in a single microwell (N = 3). (f) Calibration curve of the normalized impedance phase at 50 and 20 kHz (N = 3). Error bars in (d–f) represent the standard deviation.

Lastly, and to test the selectivity of our method, a mixed solution of S. sclerotiorum spores and Fusarium graminearum spores was prepared (Figure S8). The mixed sample was introduced into the device under the same settings as previous experiments. Only the target spores were selectively captured in the microwells by DEP, while F. graminearum spores flowed to the outlet drain of our device (Video S2). This result agrees with the dielectrophoretic separation of spores previously reported.[20]

Our primary goal is to develop a device that can effectively and reliably quantify S. sclerotiorum spores in solution and potentially be integrated with commercially available spore-trap samplers into an SSR forecasting system. Spore-trap samplers are instruments employed for sampling airborne particles in a liquid collection medium, which facilitates their integration with microfluidic devices. The spore-trap sampler (Cyclone, Burkard Manufacturing) to which we intend to connect the presented device for future testing is shown in Figure S7.

Through extensive experimentation and characterization, we have demonstrated the feasibility of our device for the capture and accurate quantification of S. sclerotiorum spores, well within the sensitivity requirements needed for SSR forecasting applications.[46] The selectivity of our microwell traps is based on the dielectric properties of spores. Different types of spores have different dielectric properties, which in turn allows for their differentiation.[20] Thus, the frequency of the applied DEP signal could be easily tuned to allow for the selective capture of a specific type of spore within a mixed sample flowing through the device. Besides the selectivity based on the different dielectric properties, the microwells in our current device also provide a physical restriction to the size of particles that can be captured.

When compared to other approaches,[20] some key issues have been improved. Our current quantification method is based on static rather than dynamic impedance measurements, which increases the sensitivity and reduces the need for instruments with an ultrafast response time. Moreover, the DEP signal is applied with the same electrodes that are employed for impedimetric sensing, avoiding the need for a dedicated structure for DEP and another one for sensing. Lastly, we eradicated clogging issues by increasing our microchannel size but without losing the sensitivity to single spores. When compared to more traditional methods, such as the petal test or PCR, our approach has several advantages, such as not requiring an incubation time, expensive reagents, or the highly labor-intensive process of petal collection. Although PCR can achieve the detection of a few nanograms of S. sclerotiorum DNA,[22] its complex operation and instrumentation hinder their potential application in the field.[24] Our proposed method is simple, portable, and sensitive to single spores, which can allow us to provide an accurate and early warning system to avoid SSR outbreaks.

Furthermore, and to the best of our knowledge, only a few microfluidic platforms employing microwell arrays and impedimetric sensing have been reported,[42,56] which are primarily restricted to medical applications. These devices are fabricated with two electrode planes, one under the microwells and the other one on top, reducing the impedance measurement to the whole array, limiting the device sensitivity, and making it immune to changes that occur in single microwells. In contrast, with our approach, each column in the array can be measured individually, which allows the detection of single particles in a single microwell.

Our next steps will focus on developing a portable impedance analyzer and automating the switches that are responsible for controlling the mode of operation of the electrodes. This, for instance, will allow having a sector of the array configured in DEP mode while the other sectors are simultaneously measuring the impedance.

Conclusions

The primary goal of any SSR forecasting system is to reduce the unnecessary application of fungicides. For this, the effective and rapid quantification of S. sclerotiorum airborne inoculum is essential. However, the current lack of simple, cost-effective, and portable platforms that can capture and quantify S. sclerotiorum airborne inoculum has hindered the development of an efficient early warning system.

The device presented here remedies these limitations through the unique integration of a microfluidic platform and a label-free quantification method that uses dielectrophoresis to reliably capture S. sclerotiorum spores in a picoliter well array. The spores in the microwells are subsequently quantified using non-faradaic electrochemical impedance spectroscopy employing coplanar nanothick aluminum electrodes.

Microwell arrays with different diameters and depths were fabricated and extensively tested to determine the optimal conditions for spore capture and quantification. We demonstrated a highly efficient spore trapping rate of more than 90% and the detection of single spores, satisfying the sensitivity requirements to provide an early warning of SSR outbreaks. Our device and methodology could also be easily extended to other fungal pathogens affecting agricultural food crops. Due to characteristics like simplicity, cost-effectiveness, and portability, we believe that the future integration of our device with high-throughput spore-trap samplers has great potential for crop protection applications, such as the on-site forecasting of SSR.

Methods

Device Fabrication

Microfluidic devices were fabricated using standard photolithography processes on 500 μm thick glass substrates with 4 in. diameter. The substrates were first cleaned with piranha solution (3:1, H2SO4/H2O2) for 15 min. Immediately after this, 100 nm of aluminum was sputtered on top of the substrates. Electrodes were patterned using a positive photoresist AZ1512 (EMD Performance Materials Corp.), which was spread at 500 rpm for 10 s, then increased to 5000 rpm for 40 s, and finally baked at 100 °C for 60 s. Afterward, the photoresist was exposed under UV light at 100 mJ/cm2 using a mask aligner (ABM-USA, Inc.) and developed using an AZ 400k 1:4 developer (EMD Performance Materials Corp.). The metal layer was subsequently etched using aluminum etchant type A (Transene Company Inc.). The electrodes were fabricated with a width of 20 μm and a gap of 6 μm between them. Using a second photomask, microwells were fabricated on top of the electrodes using the negative photoresist SU-8 (Kayaku Advanced Materials Inc.) with a thickness of 10 μm. Thicknesses of 5 and 20 μm were also fabricated. SU-8 2010 was spread on top of the substrates containing patterned electrodes at 500 rpm for 15 s and then increased to 3500 rpm for 30 s to form 10 μm thick layers. These substrates were soft baked at 65 °C for 2 min and then for 4 min at 95 °C. UV light exposure was done at 100 mJ/cm2, and subsequently, post exposure, the substrates were baked at 65 °C for 2 min and then for 5 min at 95 °C. Finally, the substrates were developed for 1 min using the SU-8 developer (Kayaku Advanced Materials Inc.). Additional substrates with SU-8 layers of 5 and 20 μm thickness were fabricated using SU-8 2005 and SU-8 2015, respectively. Each substrate provides six devices, three with microwells of 20 μm in diameter and three with 15 μm in diameter.

To obtain the microfluidic channels, a master mold for poly(dimethylsiloxane) (PDMS) molding was fabricated on a prime silicon wafer of 4 in. diameter using SU-8 2015 with a thickness of 18 μm. A 10:1 mass ratio of PDMS base and curing agent (SYLGARD 184 silicone elastomer kit) was poured onto the master mold and cured in an oven at 100 °C for 1 h. Afterward, the polymerized PDMS was peeled off and inlet/outlet holes were created on the channels using a disposable biopsy punch (Robbins Instruments Inc.) and subsequently cleaned with IPA and Milli-Q water.

Device Bonding and Assembly

Microfluidic channels on PDMS structures were irreversibly bonded to the fabricated glass devices containing the SU-8 microwells by silanization. First, the channel side on the PDMS was exposed to oxygen plasma using a reactive-ion etching machine (Trion Technology, Inc.). After the surface activation, the channel side was immersed in a liquid solution containing 99% (3-aminopropyl)triethoxysilane (APTES) for 45 s. Afterward, the PDMS was washed with Milli-Q water and dried using nitrogen gas. Immediately after this, the PDMS and the glass devices were carefully aligned and brought into contact. The structure was baked on a hot plate at 150 °C for 1 h while a standard calibration weight of 200 g was applied on top. Finally, 21G stainless steel connectors were inserted into the input and output holes in the PDMS and connected to PTFE tubing (Elveflow Microfluidics).

Spore Production and Reagents

The process to obtain S. sclerotiorum spores has been described in our previous reports.[20,24] In summary, compact masses of hyphae, called sclerotia, were buried in wet sand and incubated at 10 °C until they germinated carpogenically to form apothecia. The spores that were released by apothecia were captured on filter paper disks using a vacuum pump. To prepare spores in solution, the filter paper disks were cut into small pieces (∼2 mm × 10 mm) and inserted in a 2 mL centrifuge tube containing 1.5 mL of ultrapure Milli-Q water with a resistivity of 18.2 MΩ/cm (Sigma-Aldrich). Subsequently, the tube was shaken for 45 s at 1500 rpm using a digital vortex mixer (Fisher Scientific). The piece of paper was then removed from the tube, and the solution was filtered using a cell strainer (PluriSelect) with a 20 μm mesh. During DEP experiments, the spores were resuspended in our DEP loading buffer, consisting of 1% w/v bovine serum albumin (BSA) in Milli-Q water to avoid nonspecific binding of spores. Furthermore, our DEP buffer was chosen due to its nontoxicity to spores and its low conductivity, which reduces Joule heating and facilitates DEP trapping. The spores were stained to facilitate identification and imaging using acridine orange (Thermo Fisher Scientific).

F. graminearum culture was maintained on PDA plates, and further macroconidia spores were obtained by culturing the fungus in a synthetic nutrient-poor broth (SNB) medium with 1% sucrose to induce the formation of spores (KH2PO4 1 g, KNO3 1 g, MgSO4·7H2O 0.5 g, KCl 0.5 g, glucose 0.2 g, sucrose 0.2 g) on a shaker (150 rpm) at room temperature for 7 days. The spores were separated by filtration of the liquid culture through 20 μm filter and further centrifugation.

All other chemicals used were of analytical grade and obtained from Sigma-Aldrich.

Spore Fixation

The spores were fixed for HIM imaging using 4% paraformaldehyde in phosphate-buffered saline (PBS) with 0.1% Triton X-100. First, the spores were introduced into the device and captured using DEP. After this, 1 mL of the paraformaldehyde solution was pumped into the device and left to rest for 15 min. Second, PBS 1× was introduced for 10 min to rinse the microwells. Lastly, the captured spores were dehydrated by introducing ethanol of graded concentrations, which were 20% for 5 min, 40% for 5 min, 60% for 5 min, and 80% for 5 min.

Instrumentation and Experimental Setup

A custom-made chip holder based on pogo-pins (Mill-Max Corp.) was employed to electrically connect our microfluidic device to all measurement equipment. A set of switches in the holder allowed us to control the signal applied to each electrode. The flow of spores in the solution within the microfluidic channel was generated and controlled using a syringe pump (New Era Pump Systems Inc. NE-4000). During DEP experiments, sinusoidal signals were applied to the electrodes via the chip holder using a function generator (Rigol DG822) through a bipolar 10× amplifier (Tabor Electronics 9250). An oscilloscope (Tektronix TDS 2012B) was also used to monitor the applied signal. During the process of DEP capture, our device was placed on the viewing stage of an upright fluorescence microscope (Amscope FM820TMF143) integrated with a CCD camera (Sony ICX825ALA) for imaging and video recording. nF-EIS measurements were performed using a high-precision impedance analyzer (Zurich Instruments MFIA) controlled by the software LabOne.