Shahrzad
Zahertar
a,
Hamdi
Torun
*a,
Ran
Tao
ba,
Pep
Canyelles-Pericas
ca,
Jingting
Luo
ba,
Qiang
Wu
a and
Yong-Qing
Fu
*a
aFaculty of Engineering & Environment, University of Northumbria, Newcastle upon Tyne, NE1 8ST, UK. E-mail: hamdi.torun@northumbria.ac.uk; richard.fu@northumbria.ac.uk
bShenzhen Key Laboratory of Advanced Thin Films and Applications, College of Physics and Optoelectronic Engineering, Shenzhen University, 518060, China
cDepartment of Integrated Devices and Systems, MESA+ Institute, University of Twente, Enschede, 7522 NB, The Netherlands
First published on 31st January 2022
A key challenge in developing lab-on-a-chip devices is integrating various functionalities such as liquid manipulation and sensing on a single platform, which conventionally requires different technologies to be separately optimized. In this paper, we design circular or ring-shape interdigitated structures on a piezoelectric substrate that can be used for both sensing and liquid manipulation for point-of-care applications. We present acoustofluidic functionalities including streaming, jetting and nebulisation using the circular surface acoustic wave device to demonstrate the acoustofluidic manipulation capabilities of the actuator. We then optimise the design inspired by the double split-ring resonator geometry so that the device geometry can be used as a high performance wireless and electromagnetic sensor. We present the wireless-sensing capabilities of the device with case studies in monitoring glucose concentration and the fermentation process of yeast cells.
Wireless sensing is desirable for LoC devices especially when they are used in wearable configurations where the active sensors and sensor arrays can be located away from a readout unit.15,16 Among different technologies for wireless sensing, electromagnetic metamaterials with their exceptional high-quality factors have been demonstrated effectively.17,18 Metamaterials are artificially engineered structures that possess properties that are not readily available in nature, for example, those designed materials having negative values of permittivity and permeability at the same time. Split-ring resonators (SRRs) are among the basic building blocks of metamaterials.19–21 When exposed to electromagnetic radiation, these structures can resonate due to their interaction with radiation. Their fundamental resonant frequency depends on their geometry and the composition of the materials around them.22 Altering any of these parameters will result in a change in the resonant frequency, and this mechanism has been utilized for various sensing applications including biomedical,23–26 chemical,27–30 and environmental sensors,31,32 as well as in food industry.33,34 Metamaterials have been demonstrated for different bands in the electromagnetic spectrum from microwaves to terahertz frequencies.35,36 Metamaterial-based resonators can be integrated into LoC devices due to their sensing capabilities. However, they are incapable of liquid manipulation; therefore, they cannot form an LoC device on their own.
On the other hand, surface acoustic wave (SAW) actuators have been extensively studied for the manipulation of liquids particularly with volumes of microlitres or even smaller.37–39 SAW devices are commonly composed of interdigitated electrodes that are patterned on a piezoelectric substrate. They have been widely used in acoustofluidics to manipulate droplets for sample preparation, sorting, patterning, mixing, pumping, jetting, and nebulisation.40,41 In addition, SAW devices have been successfully demonstrated as biomolecular and environmental sensors.42–47 However, numerous environmental factors and variations in material characteristics such as conductivity, pressure, mass loading, temperature, humidity and viscosity can influence the SAW sensing mechanism41 making the specificity of these devices complicated. Also, achieving high sensitivity levels in molecular sensing requires scaling down the size of the device, which is challenging due to the complexity attributed to the miniaturization and fabrication process.48
Recently, we proposed a concept of combining electromagnetic metamaterial-inspired sensors and surface acoustic wave (SAW) actuators on a single platform for LoC applications,49,50 and we demonstrated that the electromagnetic sensing capabilities of metamaterials and the liquid droplet manipulation capabilities of SAW devices can be combined in a single structure. Combining liquid manipulation and sensing capabilities in a single structure can make it possible to realise integrated LoC platforms with advance functionalities. For example, the devices can be used to analyse liquid samples dynamically while they are mixed with a different liquid on the spot. In addition, these devices can be operated in an array formation where each device will be a single analyser unit. Such an architecture can be used as a multiple target analyser if each unit is functionalised with different agents.
In this paper, we propose a new device based on circular interdigitated transducers (IDTs) that are configured in a split-ring resonator (SRR) geometry. The circular design of IDTs focuses traveling SAWs in a focal point and acoustic energy from various directions are accumulated in this point, which can be beneficial for effective mixing/streaming of particles/cells and efficient jetting and nebulisation of droplets.49,51,52 We verify this by utilising a standard circular IDT design and by experimenting various acoustofluidic functions such as streaming, jetting and nebulisation. In the next step, we have optimised our design for performing metamaterial-based sensing. Our optimised design includes an additional set of IDTs compared to our standard design and is inspired by a double split ring resonator (DSRR) geometry. DSRRs in the concept of metamaterials include two individual SRRs with gaps positioned on the opposite direction from each other and are implemented to increase the effective capacitance of the resonator. The new design incorporating circular DSSR-inspired IDTs can be advantageous in several domains including (1) lowering the resonant frequency in a specific frequency band, to be employed as multi-band sensors; (2) controlling the frequency by altering the position of the gap on one of the resonators, and (3) designing a more compact device at a target frequency.53,54 In this paper, we demonstrate the metamaterial-based sensing capabilities of the DSRR-inspired device based on different experiments where we measured glucose concentration of droplets, characterised various dielectric materials, and monitored the yeast fermentation process over time.
In the first set of acoustofluidic experiments, a thin glass slide of 100 μm thick was placed on top of the device with a thin layer of water in between, to form a superstrate design. A 5 μl deionised (DI) water droplet consisting of polystyrene microbeads (with a diameter of 10 μm) was placed on top of the glass slide to monitor the streaming inside the droplet at a power within the range of 1.1–8.4 W and a frequency of 9.8 MHz. For the nebulisation experiment, a 2 μl water droplet was placed on the substrate and the power and frequency were set to 15 W and 9.8 MHz, respectively. In the second set of acoustofluidic tests, a 1 μl water droplet was placed in the centre of the IDTs, and a power of 4.6 W at 9.8 MHz was applied to achieve atomisation. In the third set of acoustofluidic experiments, the central region on the surface of the standard circular IDT device was treated with 1% CYTOP solution (L-809A) by drop-coating and the surface was heated to 150 °C for ∼5 minutes. A 2 μl water droplet was placed in the centre of the IDTs and a power in the range of 0.5–5 W at 9.8 MHz was applied to achieve streaming and jetting. All these acoustofluidic experiments were captured by utilising a standard video camera.
For electromagnetic characterisation, a copper loop antenna with a perimeter of 8.8 cm was connected to one port of a vector network analyser (Agilent Technologies, N5230A) and was utilised to excite the device-under-test electromagnetically. The S11 response of the antenna was captured to characterise the device. Different sensing experiments were performed using the DSRR-inspired circular IDTs. In the first sensing experiment, different glucose concentrations (0–250 mg dl−1) were prepared and a 30 μl droplet of each concentration was placed in the central region of the IDTs. Next, the S11 spectra were obtained for each concentration sequentially and the experiment for each concentration was repeated at least five times to ensure the repeatability and reliability of the device. In the second sensing experiment, small glass containers with a capacity of 2 ml were utilised and filled with different materials including glucose powder, glucose solution, and dried yeast powder. The containers were placed on the gap region of the inner circles, and the S11 spectra were recorded for each bottle sequentially and each experiment was repeated at least five times. In the third sensing experiment, dried yeast powders were added to a solution of glucose and the fermentation process was monitored by observing the frequency response of the device for an hour. For the control experiment, 1 ml of DI water was added to the dried yeast inside the container and the same procedure was applied.
By applying RF power to the electrodes of the fabricated device at its resonant frequencies, propagating SAWs along the surface can be generated. When these waves encounter a droplet along their path, some of the energy of the wave is leaked inside the droplet, and depending on the power applied, different phenomena can happen. In the experiments, we used a superstrate glass; therefore, the Lamb waves are generated and then they propagate through the glass slide through superstrate conditioning, thus leading to streaming inside of the droplet and nebulisation of the droplet at different powers. Fig. 1(b) to (d) show the streaming performance of the device by increasing the applied power from 1.1 W to 8.4 W, respectively, when a microdroplet consisting of polystyrene microbeads (5 μm in diameter) was placed on the surface of the glass. The achieved streaming was more pronounced when the applied power was higher as it can be seen by comparing Fig. 1(b) to (d).
When the applied power was further increased, the nebulisation of the water droplet was achieved when a power of 15 W was applied to the IDT electrode pads. The nebulisation process over time is shown in Fig. 1(e) to (g), respectively. Fig. 1(h) shows the higher magnification nebulisation behaviour of a 1 μl water droplet driven by an RF power of ∼7.2 W. We can clearly see that when the SAW energy is dissipated into the droplet, acoustic wave pressure drives the liquid droplet, thus resulting in the droplet vibration and strong capillary waves on the parent droplet surface. When the power is high, the amplitudes of the capillary waves increase significantly, and the capillary wave quickly becomes unstable, resulting in whipping, pinch-off and jetting of satellite droplets of about tens to hundreds of microns and formation of mist and satellite droplets as shown in Fig. 1(h).
Then, we treated the central region of the IDTs by drop-coating a hydrophobic layer of CYTOP to demonstrate the jetting effect. A 2 μl DI water droplet was placed at the centre of the IDTs and the applied power was set to 0.5 W and then to 5 W at 9.8 MHz. Streaming (Fig. 1(i)) and jetting (Fig. 1(j)) of the droplet were achieved at these powers, respectively, as the Rayleigh waves were leaked into the droplet.
![]() | (1) |
Ceff = Csurface + Cgap | (2) |
![]() | (3) |
The CST simulation results of the standard circular IDTs compared to DSRR-based circular IDTs are depicted in Fig. 2. The boundary conditions that were applied in the modelling are explained in the Methods section and the schematic of the device relative to the directions of the electric and magnetic fields is illustrated in the inset of Fig. 2. The results show that although only one pair of IDTs would be enough for generating acoustic waves on the surface of the device, an additional set of IDTs lowers the fundamental magnetic resonance of the device for sensing applications as expected without the need of increasing the diameter of the rings. The fundamental resonant frequency of the standard circular IDT design is at ∼8.1 GHz, whereas the fundamental resonant frequency of the DSRR-based design is ∼3.8 GHz.
The simulation results presented in Fig. 2 confirm that the DSRR-inspired geometry is used to set the magnetic resonant frequency of the device below 4 GHz so that it can be measured effectively using a simpler setup. A more conventional alternative would be increasing the diameter of the device at the expense of increased footprint that exposes the device more to the environmental factors. Also, when used for molecular sensing applications, increasing the size of the device usually decreases its sensitivity since the ratio of the molecular interaction area to the overall device area decreases. For these reasons, we have selected the DSRR-shape design to perform the electromagnetic characterisation and the sensing experiments. The results are reported in the next section.
Fig. 3 shows the electromagnetic properties of the DSRR-inspired design obtained through simulations and experiments under the same boundary conditions. The directions of the electric and magnetic fields used in the simulations are illustrated in the inset of Fig. 3(a) with respect to the device. In this case, the electric field polarizes the device in the y-direction and supports two circulating current paths on the right and left sides of the ring with respect to the contact pads depicted in the figure inset, resulting in a resonant frequency of 3.68 GHz. In the next step, we placed a cone-shape structure to model a droplet inside the inner circle of the device with varying electric permittivity to investigate the effects of changing the dielectric constant of the medium in the vicinity of the device. In one setting, the relative permittivity was set (εr = 1) to mimic a vacuum, and in the other setting, the permittivity was set to εr = 80 to model the water droplet. The reflection scattering parameters for both cases are presented in Fig. 3(a). For a higher dielectric constant (εr = 80) of the material of the droplet, the resonant frequency is shifted 11 MHz towards higher frequencies. This shift is due to the presence of additional dielectric load that alters the distribution of the circulating current along the surface of the device. We also extracted the current density patterns that are induced in the device in both cases for a better understanding of the behaviour of the device. The results are depicted in Fig. 3(c and d). The intensity of the circulating current is higher around the inner loop in the case of εr = 80, leading to an electrically smaller path. Thus, the electrical parameters are smaller, and the resonant frequency is larger. This phenomenon explains a lower magnetic resonant frequency of vacuum compared to the case where a water droplet is present in the centre of the structure.
We performed experiments to confirm our simulation results in the next phase. We used a copper loop antenna and connected it to one port of the VNA. The electromagnetic wave is fluctuated around the antenna and the direction and strength of the electric and magnetic fields vary spatially and temporally with respect to the antenna position. The direction of the electric and magnetic fields can excite the device differently, thus leading to a distinct resonant behaviour.56,58
We placed the fabricated resonator underneath the loop antenna to couple with each other as shown in the inset of Fig. 3(b). After measuring the reflection parameters of the coupled system, we added a water droplet at the centre of the inner circle and measured the S11 spectrum again. The experimental characterisation results are summarised in Fig. 3(b). The resonant frequency of the device shifts towards higher frequencies (∼21 MHz) with the presence of a DI water droplet as predicted by the simulations. The slight discrepancies between simulations and experiments are attributed to several factors. One of them being the use of the loop antenna in the experimental setup, whereas planar boundaries were used in simulations to excite the resonator. Also, in simulations, the ideal environment is considered for ease of use. In addition, there may be slight differences between the material properties loaded from the simulator library and the actual experimental setting.
The average Δf values versus different concentrations of glucose are plotted in Fig. 4(c), and the error bars represent the standard deviation of the shift in frequency based on ten different measurements for each concentration. The resonant frequency of the device increases linearly within the measurement range with a sensitivity of 17.5 kHz mg−1 dl−1, which is the slope of the fitted line to the data. Based on the presented experimental data and considering the standard deviation of the obtained results, the device can detect the changes in the glucose concentration with a resolution of 23 mg dl−1 or 1.28 mmol l−1. The device exhibits excellent linearity at least to a concentration of 300 mg dl−1.
We first obtained the S11 spectra of the device when the container was empty. For the next steps, we used identical containers that were filled with 0.1 g of dried yeast, 0.1 g of glucose powder, glucose solution with a concentration of 0.1 g ml−1, and 1 ml of water. We repeated the experiment by placing each specific container on the exact position on the gap of the inner circular IDTs and recorded the data. The obtained results are plotted in Fig. 5(a). The dielectric properties of the measured materials are summarised in Table 1.
Material | Relative permittivity | Average resonant frequency shift (MHz) | Standard deviation (MHz) |
---|---|---|---|
Water | 80 | −22.31 | 1.15 |
Glucose solution | 76 | −19.81 | 0.84 |
Glucose powder | 3–4 | −11 | 0.35 |
Dried yeast | 2 | −11.5 | 0.5 |
Air | 1 | −9 | 0.5 |
The frequency of the device (f0), when there is no container placed on top of it, is 3.5 GHz. By utilising the glass container together with different materials, the dielectric constant of the area around the resonator is altered and increased compared to the case when there is no container. This results in a change in the effective capacitance of the sensor. Increasing the dielectric constant results in the increased effective capacitance, and therefore, the decreased resonant frequency. Among the materials inside the container, air (empty glass container) has the lowest relative permittivity (εr = 1) and water has the highest value (εr = 80). The relative permittivity of the dried yeast powder and glucose powder are 2 and 3–4 (ref. 60), respectively, and the relative permittivity of the glucose solution is ∼76,59 which is slightly lower than that of the DI water. Hence, it is expected that among these materials, water would have the lowest frequency, whereas the empty bottle would exhibit the highest frequency. The shift in resonant frequency (Δf) with each container with respect to the resonance of the device (f0) was calculated and the results are summarised in Fig. 5(b). The points in Fig. 5(b) represent the average shift in the resonant frequency of the device in the presence of an external substance for all the repeated experiments for each container. The error bars represent the standard deviations of the obtained shift in the frequency for each set of experiments per container repeated at least five times. The average shifts of the resonance per container for the empty bottle, dried yeast, glucose powder, glucose solution and water with respect to the operating frequency of the device are 9, 11.5, 11, 19.81, and 22.31 MHz, respectively, towards lower frequencies.
C6H12O6 (glucose) → 2C2H5OH (ethanol) + 2CO2 (carbon dioxide) |
For our experiment, we prepared a solution of glucose with a concentration of 0.1 g ml−1. In the next step, we added 0.1 g of dried yeast on top of this solution. Afterwards, we placed the container to the specified location in the previous experiment, covering the gap of the inner circular IDTs of the resonator for a duration of 140 minutes, and the S11 spectra were recorded every 10 minutes. The S11 spectra are presented in Fig. 6(b). A narrower range in the x-axis of the S11 spectra is also selected to illustrate the shift more clearly and is presented as an inset in Fig. 6(b). After positioning the container, we noted the initial resonant frequency before the process as the reference frequency of the system. We observed that during this experiment, the resonant frequency shifted towards lower frequencies and stabilized after about an hour. The results of the shift in frequency over time are plotted in Fig. 6(a). During the monitoring duration, the glucose concentration in the solution decreases; therefore, the relative permittivity increases around the sensor. At the same time, ethanol is produced which has higher relative permittivity compared to air and the same principle applies.27,62 The fermentation reaction leads to an increase in the effective capacitance of the device and as a result, to a lower resonant frequency. Also, by the accumulation of ethanol over time, the fermentation will be ceased, thus leading to the stabilization of the resonant frequency change, Δf, following an exponential decay in time, t:
Δf = 4.618 × e−0.02965t − 3.533 | (4) |
This journal is © The Royal Society of Chemistry 2022 |