RESEARCH ARTICLE | APRIL 02 2025

Integrated strategy for icing/fogging mitigation with

electromagnetic metamaterials and thin film surface
acoustic waves 
Chi Zhang ; Huiling Ong ; Hamdi Torun ; Jikai Zhang ; Luke Haworth ; Nicholas L. Theodorou ;
Prashant Agrawal ; Weipeng Xuan ; Jinkai Chen ; Dengmu Cheng ; Jikui Luo  ;
Yong-Qing Fu 

Appl. Phys. Rev. 12, 021402 (2025)

https://doi.org/10.1063/5.0241048


View Export
Online Citation

Articles You May Be Interested In

Methodologies, technologies, and strategies for acoustic streaming-based acoustofluidics

Appl. Phys. Rev. (March 2023)

Thermal annealing effect on ZnO surface acoustic wave-based ultraviolet light sensors on glass substrates
Appl. Phys. Lett. (May 2014)

16 September 2025 14:30:42

A traveling surface acoustic wave-based micropiezoactuator: A tool for additive- and label-free cell lysis
Biomicrofluidics (September 2024)
 Applied Physics Reviews ARTICLE pubs.aip.org/aip/are

Integrated strategy for icing/fogging mitigation

with electromagnetic metamaterials and thin film
surface acoustic waves
Cite as: Appl. Phys. Rev. 12, 021402 (2025); doi: 10.1063/5.0241048
Submitted: 28 September 2024 . Accepted: 11 March 2025 .
Published Online: 2 April 2025

Chi Zhang,1,2,3 Huiling Ong,1 Hamdi Torun,1 Jikai Zhang,1 Luke Haworth,1 Nicholas L. Theodorou,1
1 4 4 1
Prashant Agrawal, Weipeng Xuan, Jinkai Chen, Dengmu Cheng, Jikui Luo,2,3,a)
and Yong-Qing Fu1,a)

AFFILIATIONS
1
Faculty of Engineering and Environment, Northumbria University, Newcastle upon Tyne NE1 8ST, UK
2
College of Information Science and Electronic Engineering, Zhejiang University, Hangzhou 310027, China
3
International Joint Innovation Center, Zhejiang University, Haining 314400, China
4
College of Electronics & Information, Hangzhou Dianzi University, Hangzhou, China

a)
Authors to whom correspondence should be addressed: [email protected] and [email protected]

16 September 2025 14:30:42

ABSTRACT
Icing, fogging, and frosting cause safety hazards, reduced energy efficiency, and operation difficulties in various sectors including aerospace
and renewable energy. Traditional methods for mitigating these hazards are often based on active transducers that are either inconvenient,
energy intensive, or utilizing chemicals that are detrimental to the environment and lacking long-term stability. To tackle the challenges of in
situ monitoring and mitigating fogging and icing hazards on structural surfaces, we explored an integrated platform by combining electro-
magnetic (EM) metamaterials and piezoelectric thin film-based surface acoustic wave (SAW) technologies. Icing monitoring was performed
using EM metamaterial based on SAW electrodes with advantages of wireless and non-contact detection, and effective de-icing functions
were achieved through harnessing mechanical vibrations, acousto-thermal, and acoustic streaming effects generated by the SAWs. This inte-
grated platform is modular and scalable up for practical applications requiring fogging/icing detection and prevention systems. Our results
have shown that the resonant frequency of the metamaterial device was decreased with accumulation of condensation on the surface of the
device, which showed the fulfillment of sensing and monitoring. Results also showed that as the applied SAW power was increased, the time
taken for de-fogging and de-icing were significantly decreased.
Published under an exclusive license by AIP Publishing. https://doi.org/10.1063/5.0241048

I. INTRODUCTION fogging and icing.11,12 For example, in cold weather, windshields of

Formation of fogging and icing on structural surfaces has caused vehicles, airplanes, and ships are prone to fogging or icing, which sig-
severe hazards, energy reduction, and operation issues in various fields, nificantly reduces the visibility and causes potential accidents.11,12
including aerospace,1,2 energy/renewable energy,3–5 and transporta- Condensation of fog, frost, or ice on the surface of solar panels pre-
tion.6,7 For example, icing on the plane’s surface disrupts aerodynamics vents them from effectively absorbing solar radiation and reduces their
of the wings, reducing lift-up capability, increasing flight resistance, power generation efficiency.11,12 Therefore, it is crucial to find solu-
and increasing the risk of fatal accidents, whereas accumulation of ice tions to effectively monitor, detect, or prevent the condensation, fog,
over power lines changes the shape and weight of cables, hindering frost, or ice, while simultaneously deploying effective active methods of
their normal operation.8–10 Icing of wind turbine blades significantly de-icing and de-fogging systems.
reduces their electrical energy conversion efficiency, and accelerates The common detection methods for fogging and icing include
surface damage, along with potential shedding hazards. One of the optical,13 ultrasonic,4 microwave,14–16 thermal or infrared,7 and
least frequently studied but widely faced issues is the performance deg- impedance modulation5 based sensing technologies. Among them,
radation of optical lenses, glass windows, or photonic devices due to electromagnetic (EM) or microwave sensing technology has attracted a

Appl. Phys. Rev. 12, 021402 (2025); doi: 10.1063/5.0241048 12, 021402-1
Published under an exclusive license by AIP Publishing
 Applied Physics Reviews ARTICLE pubs.aip.org/aip/are

lot of attention due to its advantages such as simple structure, low de-fogging, and sensing of ice and fog formation without mutual inter-
manufacturing cost, high sensitivity, wireless, and non-contact detec- ferences, avoiding the requirement of additional precision sensors for
tion. The mechanism of microwave sensing is to monitor the fre- fog and ice monitoring, simplifying the design of ice protection
quency responses of the microwave resonator. For example, Wiltshire system.
et al.16 reported an ice sensing system based on a split-ring resonator
II. DESIGN AND PRINCIPLE OF THE INTEGRATED
(SRR), which is a type of planar microwave resonator. Shah et al.17
PLATFORM
recently developed a smart ice detection and removal system utilizing
multi-frequency microwave SRR sensors for location-specific ice sens- A. Design of the integrated platform
ing. The SRR structures were one of the first examples of electromag- Figures 1(a) and 1(b) show the conceptual diagrams and experi-
netic metamaterials which are commonly circular or rectangular rings mental setup of the proposed integrated platform, which consists of a
with one or multiple splits. The geometry of these structures deter- loop antenna connected to a vector network analyzer (VNA) and a
mines their resonant frequencies, and they exhibit high-quality factors SAW device made on a ZnO/glass substrate for sensing. Details of the
exceeding several hundred in microwave frequencies.18 When they deposition process of the ZnO films and fabrication process of the
were used for ice sensing applications, the scattering parameters of the SAW device are provided in the supplementary material. The antenna
devices were measured for detecting the accumulation or melting of was used to generate the electromagnetic field which excites the IDT
ice.16 Similarly, Niksan et al. reported real-time wireless monitoring of configuration-based EMS, and was fabricated using the copper wire
ice by employing an SRR array with a sensing distance up to 33 cm.14 with a diameter of 0.1 mm wound into a circular loop with a diameter
Many active and passive methods have been proposed to prevent of 2.25 cm. To protect the antenna from external environmental inter-
or remove the hazards caused by the accumulation of fog and ice. ference, it was placed within a rectangular box with the top cover
Methods for removal of fogging mainly include resistance heating and thickness of 0.5 mm, made using a 3D printer. The SAW device was
accelerating air circulation or increasing evaporation rate of condensed placed on top of the box at the center position as shown in Fig. 1.
water on the surface of the structures,19,20 whereas most de-icing meth- A signal generator (Aim TTi, TG5011A) was used to produce RF
ods include mechanical deformation/vibration, electrothermal, and signals, which were amplified using a power amplifier (Amplifier
photothermal ones.21,22 However, these methods have shortages of Research, Model 75A250) to drive the SAW device. In the anti-fogging
either large power consumption and low efficiency, or complex proc- and de-fogging experiments, the fogging on the IDT electrode was cre-
essing procedures, defects, pollution, degradation and/or damage to

16 September 2025 14:30:42

ated by using a nebulizer (Omron Ultrasonic Nebulizer NE-U17)
the surface of the structure. Recently, an emerging technology for which generated water aerosols to be condensed onto the device with
fog/ice mitigation has been developed to incorporate piezoelectric thin the controlled flow rates. The pipe for the humid air had a diameter of
film surface acoustic wave (SAW) devices directly onto structural sur- 1.5 cm, and its outlet was located 2 cm directly above the center of the
faces such as silicon, glass, or metals, to produce acoustic wave propa- IDTs. For the de-icing experiments, we have prepared small ice beads
gation.11,23–26 Such wave propagation along the surfaces in various with different volumes, which were obtained from a cold chamber.
modes of Rayleigh, Lamb, or shear horizontal ones generates interfacial
vibration, localized heating and streaming to achieve anti-fogging,
anti-icing, and de-icing functions. De-icing and de-fogging using SAW
technologies were recently reported to consume much less energy and
are more energy efficient compared to the traditional methods such as
resistive heating.25
It is possible to combine the wireless sensing capabilities of SRR-
based metamaterials and the excitation capabilities of SAW devices by
designing the SAW interdigital transducer (IDT) electrodes into a
modified SRR geometry.27–32 Such an integrated system has previously
been demonstrated to perform multiple functions in microfluidic sys-
tems including particle/cell manipulations, tissue engineering, and
chemical synthesis.23,33 These devices have shown that a single device
unit can be used as both a SAW actuation source and an electromag-
netic metamaterial sensor (EMS). The activation functions can be pro-
vided by the conventional SAW devices, while the SAW devices can be
directly used as the EMS for sensitive ice monitoring, attributed to its
high-quality factors. However, there were no previous studies to inte-
grate these two technologies for icing or fogging protection.
In this study, we report an integrated metamaterial-based SAW
platform to monitor fogging/icing and achieve anti-fogging, de-fog-
ging, anti-icing, and de-icing functions. A single set of IDTs has been
applied as a conventional SAW device, generating SAWs for de-
fogging and de-icing, as well as an EMS when coupled with a loop
antenna to monitor the presence and contents of fog and ice in real- FIG. 1. Conceptual illustrations of the integrated platform: (a) application scenarios
time. Such an integrated system allows for simultaneous de-icing, and (b) operation and experimental setup.

Appl. Phys. Rev. 12, 021402 (2025); doi: 10.1063/5.0241048 12, 021402-2
Published under an exclusive license by AIP Publishing
 Applied Physics Reviews ARTICLE pubs.aip.org/aip/are

They were then placed on the IDT electrode of the SAW device.
Additionally, another type of de-icing experiment was conducted by
placing the SAW device on a refrigeration plate in a cold chamber. A
K-type thermocouple (RS PRO, Type K Exposed Junction) was placed
on the surface of the SAW device to monitor the surface temperature.

B. De-icing and de-fogging mechanisms

De-fogging and de-icing functions were achieved using the inte-
grated platform when the SAW device was powered using the power
amplifier, based on four effects induced by SAWs.25 These include
mechanical vibration and thermoacoustic effect (vibration or acoustic
wave agitation can accelerate internal molecular motions, raising inter-
nal energy of the object and the temperature), acoustic streaming effect
(internal circulation is generated in the liquid), and acoustic radiation
forces (which are applied to the internal particles in liquid). For de-
fogging, the removal rate of condensed water on the surface of the
device is accelerated by combined effects of thermoacoustic, acoustic
streaming, and mechanical vibration. For de-icing, the nanoscale
earthquake-like mechanical vibrations weaken the ice adhesion on the
substrate surface, while the thermoacoustic effect accelerates the melt-
ing of the ice. Moreover, during the SAW induced de-icing process,
tiny cracks are often generated at the solid/ice interfaces and these
cracks propagate in ice due to continuous interfacial vibrations, thus
weakening the adhesion of ice to the glass. Simultaneously, due to the
localized acousto-thermal heating effects, the ice at the interfaces starts

16 September 2025 14:30:42

to be partially melted before it is changed into liquid.
FIG. 2. Working principle of the integrated platform: (a) structural dimensions of
C. SAW operation SAW device, (b) S11 spectra of the SAW device obtained from both experiments
Conventional lithium niobate (LiNbO3) SAW devices are often and simulation, (c) induced circulating current and the equivalent circuit of the SAW
used for acoustofluidics studies. However, they are normally brittle, device operated as an EMS, and (d) measured and simulated EMS S11 results of
antenna. Surface charge distributions at the IDTs of the SAW device at (e)
and excessive mechanical and thermal stresses can easily cause fail- 2.86 GHz and (f) 3.31 GHz.
ure.34,35 In this study, we used piezoelectric film ZnO based SAW devi-
ces fabricated on a glass substrate which have been demonstrated
successfully for various sensing and optical applications.36,37 effectively coupled into the liquid on the device’s surface. The de-
Moreover, the ZnO/glass SAWs have a relatively low temperature coef- fogging and de-icing experiments were performed using the Rayleigh
ficient of frequency (TCF), i.e., the resonant frequency will not be wave mode. For this purpose, the output frequency of the signal genera-
changed significantly due to minor changes in environmental tempera- tor was fixed at 9.88 MHz. It is worth noting that such the ZnO/glass
tures. They also have high light transmission property (>80%) and SAW device with a low frequency has been proved to provide good de-
possess good optical transparency.38,39 icing and de-fogging performance.11 In this study, effectiveness of de-
Figure 2(a) shows the IDT pattern of the SAW device used in this fogging and de-icing performance was demonstrated within an effective
work, where the IDT wavelength is k ¼ 300 lm, and the dimensions surface area of 1.0  1.6 cm2 near the IDT region of the SAW device.
are L1 ¼ 14.7 mm, L2 ¼ 8.1 mm, L3 ¼ 5 mm, W1 ¼ 12.71 mm, W2
¼ 1 mm, and g ¼ 2.71 mm. The resonant frequency of the Rayleigh D. Metamaterial based sensing
wave of the SAW device was determined by the phase velocity v The IDT metallic layer can be coupled with the loop antenna for
(2639 m/s) of the sound wave and the IDT wavelength k, using wireless sensing applications. When the IDT layer is excited by the elec-
fSAW ¼ v/k. Finite element analysis was used to study the vibration tromagnetic (EM) waves from the antenna, circulating currents can be
modes and scattering parameters of the ZnO/glass SAW device, with induced on the structure at distinct resonant frequencies. For instance,
the obtained results shown in Fig. 2(b) (the red curve). The simulated when the magnetic field component of the EM wave is perpendicular
frequency of the Rayleigh wave is 9.88 MHz which closely matches the to the IDT layer, the induced current is circulated around the hollow
calculated result. The blue curve in Fig. 2(b) shows the scattering area of the IDT layer, a phenomenon referred to as magnetic coupling
parameter S11 obtained using a VNA at room temperature. The resonance. In this case, the IDT layer can be modeled as an LC resonant
Sezawa mode wave, corresponding to a frequency of 19.76 MHz, also circuit with a lumped element model as shown in Fig. 2(c).28 The reso-
appears and aligns well with the simulation result. The inset images in nant frequency of the IDT-based EMS can be expressed as30
Fig. 2(b) show the simulated results of different vibration patterns for
1
the SAW device. The Rayleigh wave is primarily concentrated on the fEMS ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffi ; (1)
surface of the device, allowing the wave energy to be more easily and 2p Leff Ceff

Appl. Phys. Rev. 12, 021402 (2025); doi: 10.1063/5.0241048 12, 021402-3
Published under an exclusive license by AIP Publishing
 Applied Physics Reviews ARTICLE pubs.aip.org/aip/are

where Leff and Ceff are the effective inductance and capacitance of the to the gap area or the permittivity of the object changes, a larger fre-
structure, respectively, and mainly determined by the shape and size of quency shift of fEMS2 will occur. On the other hand, the frequency sig-
the IDT electrode. Additionally, the permittivity of the material sur- nal of fEMS3 is more sensitive to attachment (fog or ice) in the IDT
rounding the IDT electrode significantly affects the value of Ceff. region. Since the IDT region has a higher vibration amplitude (tens to
Therefore, the frequency fEMS will be changed as fog or ice is formed hundreds of pm) as compared to areas outside of the IDTs when the
on the surface of the IDT layer. Circulating currents can also be SAW device is supplied with an RF power, it is considered more effec-
induced by electric field component applied parallel or perpendicular tive for removing fog or ice in this area.21 Therefore, fEMS3 was used as
to the IDT fingers, often called the electrically coupled resonance. a sensing parameter to monitor the fog and ice accumulation in this
Fundamental and higher order resonances can be obtained by moni- study, and the metamaterial was designed to have working frequencies
toring scattering parameters (in particular, the S11 reflection spec- ranging from 2 to 8 GHz.
trum) of the loop antenna using the VNA. From the S11 spectrum,
III. EXPERIMENTAL RESULTS AND DISCUSSION
individual dips corresponding to different resonant frequencies and
environmental changes can be monitored by detecting the shift in the A. Fog monitoring and prevention
resonant frequencies. Fog monitoring capability and de-fogging/anti-fogging perfor-
Figure 2(d) presents the simulation and experimental results of mance of the integrated platform were first investigated, which was
the S11 spectra of the metamaterial devices. The simulation was per- conducted at room temperature (20  C). To investigate the fog moni-
formed using the commercial electromagnetic simulation software, toring capability, the nebulizer was turned on, and the generated water
CST Studio Suite. As shown in Fig. 2(d), there are two distinct reflec- aerosols were deposited onto the surface of the SAW device to create a
tion dips corresponding to the resonant frequencies, at layer of condensate. Figure 3(a) shows images of the SAW device’s sur-
fEMS2 ¼ 2.86 GHz and fEMS3 ¼ 3.31 GHz, respectively. The charge dis- faces taken at different stages, where the air flow rate of the nebulizer
tributions on the device surface at these two resonant frequencies are was set to be 8 l/min. Here, the power and air velocity of the nebulizer
illustrated in Figs. 2(e) and 2(f). At 2.86 GHz, the induced current were set to be their maximum values. The volume of water droplets
forms a loop around the hollow area of the IDT layer (i.e., the area condensed on the surface of the SAW device gradually increased dur-
between the pad and IDT). In contrast, at 3.31 GHz, the induced cur- ing the fog generation process. The measured S11 values of the loop
rent path is symmetric along the symmetrical axis of the IDT electrode. antenna at different durations are shown in Fig. 3(b), which reveals
Similar to the findings reported in our previous work,32 the frequency

16 September 2025 14:30:42

that the reflection dips corresponding to fEMS3 are shifted to the lower
signal of fEMS2 was found to be more sensitive to changes in the hollow frequency region as the condensate on the device surface is increased.
area between the two pads. This means that when an object is attached This is because the relative permittivity of water (ewater ¼ 78.2) is much

FIG. 3. Fog monitoring and prevention:

(a) the images of the SAW device’s surfa-
ces taken at different stages during the
water vapor condensation process, (b) the
S11 spectra of the loop antenna at differ-
ent times, (c) DfEMS3 as a function of time
at different flow velocities of humid air,
and (d) the dependence of DfEMS3 on time
during the anti-fogging process with differ-
ent RF signal powers.

Appl. Phys. Rev. 12, 021402 (2025); doi: 10.1063/5.0241048 12, 021402-4
Published under an exclusive license by AIP Publishing
 Applied Physics Reviews ARTICLE pubs.aip.org/aip/are

larger than that of air (eair ¼ 1);38 therefore, the effective capacitance of study, the nebulizer was turned on for 35 s and then was turned off,
the IDT structure increases with the increase in condensate, resulting allowing a certain amount of condensation water (estimated to be
in a decrease in fEMS3. Figure 3(c) shows the relationship between the 10–25 ll) to be deposited on the surface of the SAW device. After 15 s,
frequency shift of fEMS3 (i.e., DfEMS3) and time at different flow the signal generator and power amplifier were turned on for de-
rates of humid air. As the flow rate of humid air was increased, fogging, with the output power of the power amplifier set to be 1.9 W.
which was controlled by the nebuliser, the rate of the frequency Figure 4(a) shows the photos of the SAW device taken at different
shift was increased. This is because a higher flow rate leads to a stages of the de-fogging process. When the SAW device was powered
faster accumulation rate of condensate. The increasing rates of on, the condensate in the left IDT region (i.e., on the device’s side close
fEMS3, with the humid air flow rates of 8 and 4 l/min, were calcu- to the pad) was evaporated first, which is probably because of the
lated to be 0.8 and 0.2 MHz/s, respectively. higher current flowing through the left IDT.
Anti-fogging performance of the integrated platform was then Figure 4(b) shows the S11 spectra of the loop antenna measured
investigated by activating the nebulizer to create fog, and simulta- at different stages of the de-fogging process, revealing that the reflec-
neously the signal generator and power amplifier were switched on to tion dips are shifted to the higher frequency region as the condensate
generate SAWs. Here, the air velocity of the nebulizer was set to be on the device surface was removed. Figure 4(c) shows the DfEMS3 val-
8 l/min. Figure 3(d) shows the dependence of DfEMS3 with time as the ues as a function of time as the RF signal power was increased from
RF signal power was increased from 0 to 1.9 W. The decreasing rate of 0.36 to 1.9 W. It is clearly seen that after the nebulizer was turned on
fEMS3 was reduced when the RF signal power was applied, and the for 35 s, the value of fEMS3 was decreased by about 35 MHz. After the
larger the RF signal power applied, the more significant the rate of SAW power was switched on, the frequency quickly returned to its ini-
decrease. As mentioned earlier, the combined effects of acousto- tial value. The time required for the de-fogging process was shown to
thermal, acoustic streaming, and mechanical vibration effects pre- decrease as the amplitude of the RF signal was increased. When the
vented the accumulation of fog generation, accelerating the removal or RF signal power was 1.9 W, the required de-fogging time was
evaporation of fog and moisture condensation on the surface of the 61 s. However, when the SAW power was decreased to 0.36 W, it
SAW device. took 152 s to complete the de-fogging process. It was also found
that during the de-fogging process, the frequency shift of fEMS3 was
initially increased rapidly but then increased slowly, which was
B. De-fogging more pronounced when the RF signal power was relatively high.

16 September 2025 14:30:42

De-fogging performance of the integrated platform was investi- This could be explained by the fact that the de-fogging’s efficiency
gated by first activating the nebulizer in the chamber to form a layer of of the SAW device in the IDT region was stronger than that out-
thick fog, before the signal generator and power amplifier were side the IDT region. When de-fogging began, the condensed water
switched on for starting the de-fogging. The evaporating power and air was evaporated first at the IDT region, followed by those outside
velocity of the nebulizer were set to be their maximum value. In this the IDT region.

FIG. 4. Simultaneously monitoring fogging

and de-fogging: (a) the photos of the SAW
device taken at different stages of de-
fogging processes, (b) the S11 spectra at
different de-fogging stages, (c) the depen-
dence of DfEMS3 on time at different RF
signal power, and (d) the variation of fEMS3
as a function of time at different spraying
times.

Appl. Phys. Rev. 12, 021402 (2025); doi: 10.1063/5.0241048 12, 021402-5
Published under an exclusive license by AIP Publishing
 Applied Physics Reviews ARTICLE pubs.aip.org/aip/are

Another de-fogging experiment was conducted, and the output fEMS3

Q¼ ; (2)
SAW power was fixed at 1.9 W. The nebulizer was turned on for 13, 25, fBW
and 35 s, respectively, before being turned off. After another 15 s, the
signal generator and power amplifier were turned on for de-fogging. where fBW is the bandwidth at half of the lowest reflection coefficient.
Figure 4(d) shows the variation of fEMS3 as a function of time. Results fEMS3 is the measured resonant frequencies and was obtained when dif-
showed that the longer the spraying time, the higher the condensation ferent amounts of condensation water were deposited on the surface of
water content, and therefore, longer time was required for de-fogging. the SAW device. In this experiment, the nebulizer was turned on for
When the spraying times were 13, 25, and 35 s, the required de-fogging 0–45 s and then turned off, and the nebulizer was set to its maximum
times were about 24, 40, and 61 s, respectively. evaporating power and air velocity. After the S11 became stable, the
In summary, the integrated platform is capable of effectively mon- value of fEMS3 was recorded ten times.
According to the S11 spectra shown in Figs. 3 and 4, the obtained
itoring, preventing, and removing fog or condensation on the surface of
Q factor of the IDT EMS at the resonant frequency of fEMS3 was larger
SAW devices, and its performance is largely dependent on the magni-
before the nebulizer was turned on and was also larger after de-fogging
tude of the RF signal power. The maximum power level that we
was completed. The quality factor decreased during the spraying pro-
reported here is relatively high at 1.9 W considering the area of the devi-
cess with the increase in condensed water on the surface of the SAW
ces, leading to a power density of 0.56 W/cm2. We should address that
device, mainly due to the loss induced by the presence of water. Figure
we did not design and apply the RF matching system. Therefore, there
5(a) shows the relationship between the Q factor and time during the
are about one fourth of RF powers estimated to be reflected. The real
spraying stage. The frequency responses of the sensor exhibit a good
power applied to the structure should be much smaller. It is important
linearity as a function of time within the measurement range. The
to note that this is not representative of a fundamental limit, and
quality factor has maintained a value within a range of 100–800, dem-
power-efficient SAW transducers have been demonstrated by optimiz-
onstrating that the resonant frequencies can be measured with a good
ing wave modes, applying matching circuit, and selecting the best fre-
accuracy. The nonlinearity of the quality factor vs time was attributed
quencies for different applications.40 A recent work reported the energy
to the scattering and coupling losses when the condensation water was
density requirement for a LiNbO3 SAW device at 0.29 W/cm2 which is
in contact with the SAW device.
on par to that of electrothermal heaters at 0.25–0.45 W/cm2. These val-
Figure 5(b) shows the recorded frequency shifts when different
ues are not significantly different than our current device’s ones.41
amounts of condensation water were accumulated on the surface of

16 September 2025 14:30:42

the SAW device (here the nebulizer’s working time gradually increased
C. Factors affecting sensing performance from 0 to 45 s with the other parameter fixed). It was noticed that the
The quality factor Q was utilized to evaluate the sensor perfor- measurement shift of fEMS3 increased with the increase in condensation
mance in terms of the losses and resonator bandwidth, which was cal- water. Before introducing the fogging, the maximum shift of the reso-
culated using the following formula: nant frequency was 0.41 MHz, whereas after 45 s of moisture

FIG. 5. Sensing performance of the inte-

grated platform: (a) the dependence of Q
factor and DfEMS3 on time, (b) the fre-
quency shift of DfEMS3 with different
amounts of condensation water, (c) the
dependence of DfEMS3 and the surface
temperature of the SAW device against
time when the RF signal power is 3 W (c)
and 1.5 W (d).

Appl. Phys. Rev. 12, 021402 (2025); doi: 10.1063/5.0241048 12, 021402-6
Published under an exclusive license by AIP Publishing
 Applied Physics Reviews ARTICLE pubs.aip.org/aip/are

spraying, the maximum change was increased to 1.42 MHz. When the of the RF signal and temperature on the sensing performance of
change was increased to a certain value, the frequency of fEMS3 could the integrated platform were minimal.
not be used to reliably detect the condensation water content on the
surface of the SAW device.
To study the effects of the RF signal and temperature on the D. Ice monitoring and de-icing functions
fog sensing performance of the integrated platform, the signal gen- Two sets of experiments were conducted to evaluate the capabil-
erator and power amplifier were turned on for 50 s while keeping ity of the integrated platform for ice monitoring and de-icing. In the
the nebulizer turned off. Figures 5(c) and 5(d) show the depen- first set of experiments, frozen ice droplets with different sizes were
dence of DfEMS3 and the surface temperature of the SAW device as first prepared in the cold chamber and then transferred onto the sur-
a function of time when the RF signal powers were set at 1.5 and face of the SAW device. Here, the experimental setup used was the
3 W, respectively. The results indicated that the temperature of the same as that shown in Fig. 1(b), and the room temperature was 20  C.
SAW device was quickly increased (from 20 to 45  C for 3 W RF A micropipette was used to place droplets with volumes of 1.3, 2.0, 3.3,
signal power and from 20 to 31  C for 1.5 W RF signal power) after and 5.0 ll on a cooling plate, the temperature of which was set to be
the RF signal was applied. The higher the power of the RF signal, 16  C. After the droplet was frozen, we then moved the ice crystals
the faster the temperature increased. However, except for slight onto the surface of the SAW device.
fluctuations of DfEMS3 when the power amplifier was turned on or The relative permittivity of ice is 3.2,15 whereas that of water is
off, there was no significant dependence of DfEMS3 on temperature. 42
80. When ice melts into water, the effective capacitance of the IDT
The fluctuations were mainly caused by the ambient mechanical structure and substrate increases drastically due to the rise in permit-
vibrations when the switch was turned on or by pulsed currents tivity, resulting in a decrease in resonant frequency of DfEMS3. Figure
due to the switching state being switched off. Therefore, the effects 6(a) illustrates the relationship between DfEMS3 and time as the ice

16 September 2025 14:30:42

FIG. 6. Ice monitoring and de-icing: (a) the relationship between DfEMS3 and time when ice droplet of different volumes melts naturally, (b) the frequency shift of fEMS3 as a func-
tion of time with different RF signal power, (c) the DfEMS3–t curves when the droplet melted naturally and when the power amplifier was turned on during the droplet melting pro-
cess. (d) and (e) The photos of the ice droplet at different times during the natural melting process of droplet and the de-icing progress, respectively.

Appl. Phys. Rev. 12, 021402 (2025); doi: 10.1063/5.0241048 12, 021402-7
Published under an exclusive license by AIP Publishing
 Applied Physics Reviews ARTICLE pubs.aip.org/aip/are

crystals of different volumes melted naturally. The results indicated rate of the resonant frequency during the natural melting of ice droplet
that the resonant frequency shift gradually decreased during the ice was calculated to be 0.43 MHz/s (from t ¼ 20 s to t ¼ 60 s), whereas
melting process and eventually stabilized. The total change in DfEMS3 with an RF signal of 1.9 W, the decrease rate was increased to
was increased with the volume of the ice droplet, demonstrating that 1.84 MHz/s (from t ¼ 35 s to t ¼ 40 s).
the sensing system can distinguish between ice and water. In the second set of experiments, the SAW device was fixed on a
For de-icing, the RF signal with a frequency of f ¼ 9.88 MHz was cooling plate, and the loop antenna was placed directly above the SAW
applied to the SAW device for generating the de-icing effect. device, as shown in Fig. 7(a). Both the antenna and SAW device were
Simultaneously, we tracked the changes of resonant frequency fEMS3 placed inside a square box with a size of 40  40  40 cm3. To mini-
induced by the IDT configuration to monitor ice melting status. Figure mize the influences of ice melting on the antenna, a smaller SAW
6(b) shows the recorded DfEMS3–t plot when the RF signal to SAW res- device with an acoustic resonance frequency of 40 MHz was used,
onator was set at 0, 0.36, 0.95, and 1.9 W, respectively. It is evident that where the IDT wavelength was k ¼ 64 lm, and the dimensions were
the rate of decrease in fEMS3 increased with the RF signal power. L1 ¼ 3 mm, L2 ¼ 1.5 mm, L3 ¼ 3.8 mm, W1 ¼ 5.6 mm, W2 ¼ 0.5 mm,
Moreover, in the case where the SAW device was powered by an and g ¼ 0.8 mm.
amplified RF signal, the total decrease in fEMS3 was greater than that A water droplet of 5 ll was put onto the IDT area of the SAW
observed during the natural melting of ice droplet. We observed that device using a pipette, and then the cooling plate was turned on. After
when the RF signal power exceeds 0.95 W, the DfEMS3–t curves bend the water was frozen, the signal generator and amplifier were then
upwards after reaching their lowest point. This was because the melted switched on. The effect of ice temperature on the de-icing effect of sur-
water began to evaporate after spreading over the surface of the SAW face acoustic waves was investigated. The RF signal power was fixed at
device. Figures 6(d) and 6(e) show photographs of the ice droplet, 0.95 W, and the temperatures of the cooling plate were set to be 3.8,
taken at different durations during the natural melting process of 9.8, 12, and 16  C, respectively. Figure 7(b) shows the DfEMS3–t
droplet [corresponding to the blue curve in Fig. 6(b)] and the melting curves obtained at different temperatures. The results indicate that
process of droplet when the RF signal was 1.9 W [corresponding to the lower ice temperatures resulted in longer de-icing times, which is
yellow curve in Fig. 6(b)]. expected since ice at lower temperatures requires more energy to melt.
Figure 6(c) presents a comparison of DfEMS3–t curves between Figure 7(c) shows the dependence of DfEMS3 on time at different RF
the natural melting of droplet (blue curve) and the melting process of signal power levels, with the cooling plate temperature fixed at
droplet when the power amplifier was turned on (red curve). In the 16  C. As clearly shown in this figure, the lower RF signal power

16 September 2025 14:30:42

experiment corresponding to the red curve, the power amplifier was leads to a lower de-icing speed, similar to the results shown in
activated at t ¼ 35 s, with the RF signal power set to be 1.9 W. The Fig. 6(b). The blue and red curves in Fig. 7(d) depict the results of peri-
results indicated that the decrease rate of DfEMS3 was significantly odic de-icing tests with RF signal powers of 3.21 and 0.95 W, respec-
increased when the RF signal was applied. Specifically, the decrease tively. In these tests, the power amplifier and signal generator were

FIG. 7. Ice monitoring and de-icing: (a) experimental setup for de-icing process, (b) the dependence of DfEMS3 on time at different ice temperatures with the RF signal power
set at 0.95 W, (c) the DfEMS3–t curves at different RF signal power with the temperature of the cooling plate set at 16  C, (d) the results of periodic de-icing tests when the RF
signal power is 3.21 and 0.95 W, (e) the dependence of the surface temperature of the SAW device and fEMS3 on time under different RF signal power, and (f) dependence of
the surface temperature of the SAW device and fEMS3 vs time for the SAW device applied with a power of 5.78 W.

Appl. Phys. Rev. 12, 021402 (2025); doi: 10.1063/5.0241048 12, 021402-8
Published under an exclusive license by AIP Publishing
 Applied Physics Reviews ARTICLE pubs.aip.org/aip/are

TABLE I. Comparative analysis of our integrated device with existing de-icing devices.

Specific energy/
De-icing method power requirements Energy consumption/savings Additional notes

Electrothermal cyclic de-icing 4.7 to 8.7 kW/m 2

 Requires a continuous heated
parting strip.
Thermal–mechanical expulsion 3.1 kW/m2  Eliminates the continuous heated
de-icing43 parting strip and reduces run-
back ice.
Pulsed defrosting method44 8.3 kW/m2 Energy consumed is <1% of the Offers faster snow removal.
total generated power. 20
higher energy savings in Alaska
compared to Chicago.
This work 5.1–10.2 kW/m2 Opportunity to investigate the Integrated capability for active
fundamental limit. 2.9 kW/m2 ice monitoring and prevention.
has been achieved in LiNbO3
SAW devices.40

turned on for 20 s and then turned off for 40 s, repeated cyclically for surface of the SAW device. When the RF signal was applied, the melt-
4 min. The cooling plate’s temperature was maintained at 16  C. The ing of this rime ice led to a decrease in DfEMS3.
results obtained showed that DfEMS3 was decreased as the ice was The periodic variations of temperatures and resonant frequen-
melted and returned to the initial value when the melted water was cies of DfEMS3 as a function of time are shown in Fig. 7(f). Although
refrozen, demonstrating the effectiveness of the EMS in distinguishing the DfEMS3 values were varied periodically with temperature changes
between ice and water. Additionally, for the same de-icing duration, a when there was no droplet on the surface of the SAW device, the

16 September 2025 14:30:42

higher RF signal power resulted in a greater decrease in DfEMS3, further total frequency variations were much smaller than those with the
confirming that RF power significantly impacts de-icing efficiency. droplet on the device surface. Therefore, we can confirm that this
A simple analysis was performed to help understand if the integrated platform can effectively monitor the melting of ice in
energy efficiency is due to effects of vibrational modes or due to the real-time.
joule heating. An ice drop was measured with around 5.6 mg in In terms of power densities, Krammer and Scholz43 reported that
mass, and thus 1.87 J were required to overcome latent heat and a thermal–mechanical expulsion de-icing system only required
result in the phase change from ice to water. Approximately an addi- 3.1 kW/m2 and also lowered the chance of runback ice,43 which was
tional 0.19 J was determined to be required to raise the temperature comparable to 4.7–8.7 kW/m2 for an electrothermal cyclic de-icing sys-
from 16 to 0  C. The natural melting (joule heating) took 77 s, tem. Li et al.44 also reported that utilizing a pulsed defrosting method
whereas 1.9 W of SAW device power melted the ice in around 15 s, required much less energy with a much faster rate, and the maximum
and 1.4 W in around 30 s. Since the 1.9 W outperformed the 1.4 W energy density was demonstrated to be 8.3 kW h m2. Based on our
in efficiency (absorbed energy divided by energy supplied), we have estimation, the operating power of our SAW device over the effective
concluded that this is partially due to vibration. In future work, this surface area was from 5.1 to 10.2 kW/m2. Therefore, our future
will be further investigated. study will be focused on optimizing configurations and designs of both
To investigate the effect of substrate temperature on the resonant the SAW device and metamaterials to further reduce power consump-
frequency shift DfEMS3 of the IDT EMS, comparative experiments were tion. Table I summarizes a comparative analysis of the energy con-
conducted without putting the droplet on the surface of the SAW sumption of our integrated device with existing de-icing devices.
device. The temperature of the cooling plate was fixed at 16  C, and
nitrogen flowed into the chamber at a rate of 1 l/min to remove mois- IV. CONCLUSIONS
ture from the air [as shown in Fig. 7(a)]. Figure 7(e) shows the depen- In this study, we developed an integrated platform combining
dence of the surface temperature of the SAW device (the dashed SAW devices with microwave electromagnetic metamaterials on
scatter lines) and the resonant frequency shift DfEMS3 (the solid scatter ZnO/glass substrates for simultaneous sensing and removal of fog
line) as a function of time under different RF signal powers. The results and ice on structural surfaces. Such an integrated platform consists
indicate that when the RF signal was applied, the surface temperature primarily of a pair of IDTs, a SAW device, and a readout loop
of the SAW device initially increased rapidly and then gradually stabi- antenna. When the IDT layer was used as a planar microwave reso-
lized. Each stabilized temperature was higher with a higher RF signal nator, the platform allowed real-time monitoring of the presence of
power. As the temperature increased, the value of DfEMS3 decreased by fog and ice on the surface of the device, as well as their quantities.
about 2 MHz, and this variation in DfEMS3 due to temperature changes The results indicated that the electromagnetic resonant frequency of
was independent of the RF signal power. This phenomenon could be the device was decreased with the accumulation of condensation on
attributed to the condensation of minor residual moisture in the air the surface of the device and returned to its initial value when this
within the chamber, which formed rime ice on the low-temperature condensation was removed. The freezing of the liquid droplets on the

Appl. Phys. Rev. 12, 021402 (2025); doi: 10.1063/5.0241048 12, 021402-9
Published under an exclusive license by AIP Publishing
 Applied Physics Reviews ARTICLE pubs.aip.org/aip/are

surface of the device affected the change of resonant frequency, which DATA AVAILABILITY
was increased when the droplets were frozen, and vice versa. On the The data that support the findings of this study are available from
other hand, when the IDTs were used in conjunction with the piezo- the corresponding authors upon reasonable request.
electric thin film as an actuator (SAW device), the platform could be
effectively used for de-icing. Effects of different volumes of the ice
REFERENCES
droplets and surface temperature on the de-icing performance were 1
K. L. Sørensen, A. S. Helland, and T. A. Johansen, “Carbon nanomaterial-based
systematically investigated. We have shown that accumulation of con-
wing temperature control system for in-flight anti-icing and de-icing of
densation such as icing and fogging was significantly delayed or pre- unmanned aerial vehicles,” in IEEE Aerospace Conference Proceedings (IEEE,
vented when the IDTs were excited by electrical signals via the 2015).
2
combined effects of acoustic-thermal, acoustic streaming, and A. Shah, O. Niksan, M. C. Jain, K. Colegrave, M. Wagih, and M. H. Zarifi,
mechanical vibrations. The novelty of this study is the demonstration “Microwaves see thin ice: A review of ice and snow sensing using microwave
of a single IDT-based transducer that can be used both as a SAW techniques,” IEEE Microwave Mag. 24(10), 24–39 (2023).
3
D. Szab
o et al., “Icing analysis of Klece-Logatec transmission line with two-
device for actuation and as an electromagnetic resonator-based wire-
level icing model,” in Flexitranstore: Special Session in the 21st International
less sensor for de-icing and de-fogging applications. This can lead to Symposium on High Voltage Engineering (ISH 2019) (Springer International
decoupling of actuation and sensing functionalities using a single Publishing, 2020), pp. 107–115.
device. In the future, we plan to test the durability of the devices, 4
E. Madi, K. Pope, W. Huang, and T. Iqbal, “A review of integrating ice detec-
especially when they are exposed to harsh environmental conditions tion and mitigation for wind turbine blades,” Renewable Sustainable Energy
expected in de-icing applications. Rev. 103, 269–281 (2019).
5
M. C. Homola, P. J. Nicklasson, and P. A. Sundsbø, “Ice sensors for wind tur-
SUPPLEMENTARY MATERIAL bines,” Cold Reg. Sci. Technol. 46(2), 125–131 (2006).
6
M. San-Juan, O.
Martín, B. J. Mirones, and P. De Tiedra, “Assessment of effi-
See the supplementary material for details of deposition process ciency of windscreen demisting systems in electrical vehicles by using IR ther-
of the ZnO films on glass substrates by magnetron sputtering and fab- mography,” Appl. Therm. Eng. 104, 479–485 (2016).
rication process of the IDT electrodes by photolithograph and left-off 7
T. Rashid, H. A. Khawaja, and K. Edvardsen, “Review of marine icing and
process for the SAW devices. anti-/de-icing systems,” J. Mar. Eng. Technol. 15(2), 79–87 (2016).
8
N. Lawrentschuk, N. E. Fleshner, and D. M. Bolton, “Laparoscopic lens fog-
ging: A review of etiology and methods to maintain a clear visual field,”
ACKNOWLEDGMENTS

16 September 2025 14:30:42

J. Endourol. 24(6), 905–913 (2010).
9
The authors acknowledge the Engineering and Physical Sciences T. Karkantonis, A. Gaddam, H. Sharma, G. Cummins, T. L. See, and S. Dimov,
“Laser-enabled surface treatment of disposable endoscope lens with superior
Research Council of UK (EPSRC EP/P018998/1), the UK Fluids antifouling and optical properties,” Langmuir 38(37), 11392–11405 (2022).
Network Special Interest Group of Acoustofluidics (EP/N032861/1), 10
D. Kreeft, E. A. Arkenbout, P. W. J. Henselmans, W. R. Van Furth, and P.
the EPSRC Centre for Doctoral Training in Renewable Energy Breedveld, “Review of techniques to achieve optical surface cleanliness and their
Northeast Universities (ReNU) for funding through grant EP/S023836/ potential application to surgical endoscopes,” Surg. Innovation 24(5), 509
1, and the International Exchange Grant (IEC/NSFC/201078) through (2017).
Royal Society and the Natural Science Foundation of China (NSFC).
11
H. L. Ong et al., “Integrated transparent surface acoustic wave technology for
active de-fogging and icing protection on glass,” Mater. Chem. Phys. 304,
The authors also gratefully acknowledge the support of the China 127842 (2023).
Scholarship Council (CSC) for providing funding for this research. 12
H. L. Ong et al., “Advances in hybrid icing and frosting protection strategies for
optics, lens, and photonics in cold environments using thin-film acoustic
AUTHOR DECLARATIONS waves,” Adv. Eng. Mater. 2402139 (published online) (2025).
13
K. J. Anderson and M. D. Ray, “SLD and ice crystal discrimination with the
Conflict of Interest optical ice detector,” SAE Technical Paper No. 2019-01-1934, 2019.
The authors have no conflicts to disclose.
14
O. Niksan, K. Colegrave, and M. H. Zarifi, “Battery-free, artificial neural
network-assisted microwave resonator array for ice detection,” IEEE Trans.
Microwave Theory Tech. 71(2), 698–709 (2023).
Author Contributions 15
R. Kozak, B. D. Wiltshire, M. A. R. Khandoker, K. Golovin, and M. H. Zarifi,
Chi Zhang and Huiling Ong contributed equally to this work. “Modified microwave sensor with a patterned ground heater for detection and
prevention of ice accumulation,” ACS Appl. Mater. Interfaces 12(49), 55483–
Chi Zhang: Conceptualization (equal); Data curation (equal); 55492 (2020).
Investigation (equal); Methodology (equal); Software (equal); Validation
16
B. Wiltshire, K. Mirshahidi, K. Golovin, and M. H. Zarifi, “Robust and sensitive
(equal); Visualization (equal); Writing – original draft (equal); Writing – frost and ice detection via planar microwave resonator sensor,” Sens. Actuators,
B 301(May), 126881–126886 (2019).
review & editing (equal). Huiling Ong: Data curation (equal); Formal 17
A. Shah et al., “Multi-location microwave ice sensor for integrated, zone-based
analysis (equal); Writing – review & editing (equal). Hamdi Torun: de-icing,” IEEE Trans. Aerosp. Electron. Syst. (published online) (2025).
Supervision (equal). Jikai Zhang: Formal analysis (equal). Luke 18
H. Torun, F. Cagri Top, G. Dundar, and A. D. Yalcinkaya, “An antenna-
Haworth: Data curation (equal). Nicholas L. Theodorou: Investigation coupled split-ring resonator for biosensing,” J. Appl. Phys. 116(12), 124701
(equal). Prashant Agrawal: Writing – review & editing (equal). Weipeng (2014).
Xuan: Methodology (equal). Jinkai Chen: Data curation (equal).
19
J. Chu, G. Tian, and X. Feng, “Recent advances in prevailing antifogging surfa-
ces: Structures, materials, durability, and beyond,” Nanoscale 15(27), 11366–
Dengmu Cheng: Formal analysis (equal). Jikui Luo: Methodology 11402 (2023).
(equal); Supervision (equal); Writing – review & editing (equal). Yong- 20
F. Deng, Z. Chen, C. Wang, C. Xiang, P. Poredos, and R. Wang, “Hygroscopic
Qing Fu: Conceptualization (equal); Methodology (equal); Supervision porous polymer for sorption-based atmospheric water harvesting,” Adv. Sci.
(equal); Writing – review & editing (equal). 9(33), 2204724 (2022).

Appl. Phys. Rev. 12, 021402 (2025); doi: 10.1063/5.0241048 12, 021402-10
Published under an exclusive license by AIP Publishing
 Applied Physics Reviews ARTICLE pubs.aip.org/aip/are

21
A. Kenzhebayeva et al., “A mini-review on recent developments in anti-icing 33
X. Liu, X. Chen, Z. Yang, H. Xia, C. Zhang, and X. Wei, “Surface acoustic wave
methods,” Polymers 13(23), 4149 (2021). based microfluidic devices for biological applications,” Sens. Diagn. 2(3), 507–
22
Q. He, K. Li, Z. Xu, J. Wang, X. Wang, and A. Li, “Research progress on con- 528 (2023).
struction strategy and technical evaluation of aircraft icing accretion protection 34
P. Yu, S. Huang, Z. Ren, J. Liu, and Y. Liu, “Investigation on effects of thermo-
system,” Chin. J. Aeronaut. 36(10), 1–23 (2023). mechanical coupling on residual stress in lithium niobate in surface acoustic
23
Y. Q. Fu et al., “Advances in piezoelectric thin films for acoustic biosensors, wave device,” Mater. Sci. Semicond. Process 172, 108083 (2024).
35
acoustofluidics and lab-on-chip applications,” Prog. Mater. Sci. 89, 31–91 M. Gruber, I. Kraleva, P. Supancic, J. Bielen, D. Kiener, and R. Bermejo,
(2017). “Strength distribution and fracture analyses of LiNbO3 and LiTaO3 single crys-
24
H. Fallahi, J. Zhang, H. P. Phan, and N. T. Nguyen, “Flexible microfluidics: tals under biaxial loading,” J. Eur. Ceram. Soc. 37(14), 4397–4406 (2017).
Fundamentals, recent developments, and applications,” Micromachines 10(12), 36
G. Abbas, F. Adamu-Lema, and F. Adriyanto, “2013 Index IEEE Electron
830 (2019). Device Letters Vol. 34,” IEEE Electron Device Lett. 34(12), 1593–1670 (2013).
25
D. Yang et al., “Nanoscale ‘earthquake’ effect induced by thin film surface 37
J. H. Lim, C. K. Kong, K. K. Kim, I. K. Park, D. K. Hwang, and S. J. Park, “UV
acoustic waves as a new strategy for ice protection,” Adv. Mater. Interfaces electroluminescence emission from ZnO light-emitting diodes grown by high-
8(2), 2001776 (2021). temperature radiofrequency sputtering,” Adv. Mater. 18(20), 2720–2724
26
L. Haworth et al., “Reduction of ice adhesion on nanostructured and nanoscale (2006).
slippery surfaces,” Nanotechnol. Precis. Eng. 6(1), 013007 (2023). 38
J. Zhou et al., “Deposition of aluminum doped ZnO as electrode for transparent
27
S. Zahertar, Y. Wang, R. Tao, J. Xie, Y. Q. Fu, and H. Torun, “A fully integrated ZnO/glass surface acoustic wave devices,” Surf. Coat. Technol. 320, 39–46
biosensing platform combining acoustofluidics and electromagnetic metamate- (2017).
rials,” J. Phys. D 52(48), 485004 (2019). 39
W. B. Wang et al., “Transparent ZnO/glass surface acoustic wave based high
28
S. Zahertar et al., “A flexible PVDF-based platform combining acoustofluidics performance ultraviolet light sensors,” Chin. Phys. B 24(5), 057701 (2015).
and electromagnetic metamaterials,” in IEEE International Conference on 40
A. Winkler, R. Br€ unig, C. Faust, R. Weser, and H. Schmidt, “Towards efficient
Flexible and Printable Sensors and Systems (FLEPS) (IEEE, 2020), p. 9239454. surface acoustic wave (SAW)-based microfluidic actuators,” Sens. Actuators, A
29
R. Tao et al., “Flexible and integrated sensing platform of acoustic waves and 247, 259–268 (2016).
metamaterials based on polyimide-coated woven carbon fibers,” ACS Sens. 41
S. Jacob et al., “Surface acoustic waves equip materials with active de-icing func-
5(8), 2563–2569 (2020). tionality: Unraveled glaze ice de-icing mechanisms and application to centimeter-
30
S. Zahertar et al., “An integrated flexible platform of electromagnetic metama- scale transparent surfaces,” Adv. Mater. Technol. 8(16), 2300263 (2023).
terials and acoustofluidics on Kapton,” in IEEE International Conference on 42
C. M€atzler, “Microwave permittivity of dry snow,” IEEE Trans. Geosci. Remote
Flexible and Printable Sensors and Systems (FLEPS) (IEEE, 2021). Sens. 34(2), 573–581 (1996).
31
S. Zahertar et al., “An integrated platform for metamaterial-based sensing and 43
P. Krammer and D. Scholz, Estimation of Electrical Power Required for Deicing
surface acoustic wave-based acoustofluidics utilising circular interdigital trans- Systems (Hamburg University of Applied Sciences (HAW), Germany, 2009).

16 September 2025 14:30:42

ducers,” Sens. Diagn. 1(2), 270–279 (2022). 44
L. Li et al., “Enabling renewable energy technologies in harsh climates with
32
S. Zahertar et al., “Flexible platform of acoustofluidics and metamaterials with ultra-efficient electro-thermal desnowing, defrosting, and deicing,” Adv. Funct.
decoupled resonant frequencies,” Sensors 22(12), 4344 (2022). Mater. 32(31), 2201521 (2022).

Appl. Phys. Rev. 12, 021402 (2025); doi: 10.1063/5.0241048 12, 021402-11
Published under an exclusive license by AIP Publishing