1. Introduction

Ultrasound sensors with tens of MHz frequency range and a high signal-to-noise ratio (SNR) are sought-after for high-resolution biomedical imaging [1–3] and non-destructive industrial monitoring [4–6]. Conventional piezoelectric ultrasound sensors have a narrow frequency range with a low SNR, and are incapable of detecting high acoustic frequencies. Optical fiber-based ultrasound sensors with high sensitivity are promising alternatives to piezoelectric ultrasound sensors. Ultrasound sensors in polymer optical fiber (POF)-based fiber Bragg gratings (FBG) [7,8] have been proposed and demonstrated, providing ultrasound detection at frequencies of around 25 MHz contributed by the low Young’s modulus. However, they have limitations in higher frequency ultrasound detection due to the large grating bandwidth from sub-nanometer to several nanometers. The relatively small slope at the edge of the reflection spectrum and large diameter of the POF-based FBG leads to low ultrasound sensitivity. Strain variations on the order of nano-strain induced by high ultrasound frequencies can not be detected in such grating-based sensors. Detection of weak ultrasound signals requires a fiber device with a steep optical spectral response capable of producing measurable intensity changes from nano-strain acoustic perturbation.

Tapered fiber-based ultrasound sensors have been proposed and demonstrated for the detection of ultrasound frequencies up to 150 kHz [9]. With the core size of sub-micrometers, the nano-strain variation induced intensity change becomes detectable, especially in dual-core microfibers (DCM) with a small core-to-core distance, because the fractional change of core size induced by ultrasound pressure results in a measurable change in the transmitted power. A DCM is analogous to a Mach-Zehnder interferometer (MZI) where each of the even and odd modes represents an arm of the MZI. The phase difference between the even and odd modes is modulated by ultrasound signals leading to intensity variation at the outputs of the DCM, which allows for direct detection using a photo-detector and eliminates the need for phase recovery as is the case in single-core taper-based acoustic sensing [10,11]. DCM-based ultrasound sensing with a frequency range of several hundred kHz has been demonstrated by directly measuring the output intensity change [12]. A multi-mode interferometer acoustic sensor based on single-mode fiber (SMF) taper was proposed and demonstrated for the detection of acoustic signals with frequencies up to tens of kHz [13]. The interference spectrum of multi-mode interferometers has low contrast over a large free spectral range (FSR) leading to a relatively small maximum-spectral-slope, which reduces the ultrasound detection sensitivity, and as a result, ultrasound signals with a frequency beyond tens of kHz can not be detected. A frequency response around 25 MHz is detected by introducing elliptical bubbles into the SMF taper to couple power into a larger number of modes and increase the maximum-spectral-slope of the interference spectrum [14]. Ultrasound detection at high frequencies beyond 25 MHz could not be detected as the maximum-spectral-slope is limited by the large FSR of a few nanometers that results from the small refractive-index difference between the modes. In addition, SMF tapers have a large Young’s modulus (73 GPa for silica [15]), leading to low sensitivity as ultrasound waves induce a small variation to taper length and refractive index. As SMF taper-based ultrasound sensors have a diameter of a few micrometers, the SMF taper sensors are fragile making them difficult to operate in a harsh environment.

Recently, mechanically robust As$_2$Se$_3$-PMMA tapers have been designed and demonstrated for ultrasound sensing at frequencies up to 34 MHz [16]. As the lower Young’s modulus of the As$_2$Se$_3$ and PMMA (17.8 GPa for As$_2$Se$_3$ and 3.5 GPa for PMMA [17]) compared to that of silica, the strain sensitivity of the As$_2$Se$_3$-PMMA taper is six times higher than that of silica-based tapers [18,19], which allows the highly sensitive broadband ultrasound sensing. The maximum-spectral-slope and FSR of the interference spectrum are tailorable by varying core size and distance between two cores of the dual-core As$_2$Se$_3$-PMMA microfiber, which enables the design and implementation of ultrasound sensors with a targeted sensing performance. In addition, the design of the DCM of different parameters enables tailorable refractive index difference between even and odd modes, which can be performed to have optical phase comparable to acoustic phase modulation at high frequencies of tens of MHz, leading to detectable high-frequency ultrasound waves.

In this paper, we demonstrate broadband ultrasound sensing based on fused dual-core chalcogenide-PMMA microfibers. DCMs are designed and fabricated with submicron core diameters for increased ultrasound sensitivity. Ultrasound frequency range and SNR of the fabricated DCMs are measured using a piezoelectric transducer (PZT) as an ultrasound source using two arrangements. In the first arrangement, the DCM sensor and the PZT are placed on a 2 mm-thick aluminum plate, and experimental results show a higher SNR and a larger frequency range when the DCM has a smaller core diameter and a closer core separation. In the second arrangement, the aluminum plate is removed and the DCM is placed in direct contact with the PZT to eliminate acoustic attenuation due to propagation through the plate medium allowing for the measurement of higher acoustic frequencies. Ultrasound sensing with an average SNR of 31 dB and a broadband frequency range from 20 kHz to 80 MHz is achieved in a compact dual-core As$_2$Se$_3$-PMMA with a core diameter of 0.5 $\mathrm{\mu}$m, core separation of 0.445 $\mathrm{\mu}$m and waist length of 1 cm. The obtained ultrasound response is compared with the PZT response that is characterized by measuring electrical reflection coefficient S$_{11}$, validating the high sensitivity of the fused dual-core As$_2$Se$_3$-PMMA microfiber sensor.

2. Fabrication, simulation and working principle

The fabrication of dual-core As$_2$Se$_3$-PMMA fibers has been reported in [20,21]. To make an As$_2$Se$_3$-PMMA preform with a fused dual-core shape, an assembly with two As$_2$Se$_3$ fibers and a PMMA tube is placed on a spinning lathe and heated at 220 $^{\circ }$C for 48 hours. Figure 1(a) presents the cross-section image of the polished end of a fused dual-core As$_2$Se$_3$-PMMA fiber. The ratio of the core separation (${\rm D}_{\textrm{core}}$) and core diameter d${\rm }_{\textrm{core}}$ is ${\rm D}_{\textrm{core}}$=0.89d${\rm }_{\textrm{core}}$. The fused dual-core fiber is then tapered using the heat-brush method [22]. A sample profile of a fused DCM with a waist length (${\rm L_w}$) of 1 cm and an As$_2$Se$_3$ core diameter of 2 $\mathrm{\mu}$m is shown in Fig. 1(b). The section between 2.5 cm and 3.5 cm along the microfiber corresponds to the waist region. One As$_2$Se$_3$ core of the dual-core As$_2$Se$_3$-PMMA fiber is butt-coupled with single-mode fibers and the butt-coupling interfaces are fixed using UV-cured epoxy, as illustrated in Fig. 1(c). Four fused dual-core microfibers with the same waist length of 1 cm and different diameters are fabricated and utilized in this investigation. Images and dimensions of the four DCMs are shown in Fig. 1(d). The four microfibers have an As$_2$Se$_3$ core diameter of 4 $\mathrm{\mu}$m, 2 $\mathrm{\mu}$m, 0.75 $\mathrm{\mu}$m, 0.5 $\mathrm{\mu}$m and a PMMA cladding diameter of 225.9 $\mathrm{\mu}$m, 113.5 $\mathrm{\mu}$m, 42.4 $\mathrm{\mu}$m and 28.0 $\mathrm{\mu}$m, respectively.

 

Fig. 1. (a) Optical microscope image of a fused dual-core As$_2$Se$_3$-PMMA fiber. (b) The relationship between a tapered core diameter and the fiber length. (c) Schematic of a coupled DCM. (d) The side views of four DCMs.

Download Full Size | PDF

Both the even and odd modes are equally excited when a laser is launched into one of the input cores of the DCM. The electric field distributions of even and odd modes of a fused dual-core microfiber are shown in Fig. 2(a) and Fig. 2(b), respectively. The calculated field distributions are obtained using refractive-index values of 2.674 and 1.481 for As$_2$Se$_3$ cores and PMMA cladding, respectively. Light propagation along a DCM is simulated using the beam propagation method (BPM), as shown in Fig. 2(c). The light power transfers back and forth between the two cores with a constant period ($\Lambda$) when propagating along the DCM. The coupling length is defined as the length required for the power to transfer from one core to the other given by $L_c=\Lambda /2= \pi /\left (\beta _{e}-\beta _{o}\right )$, where $\beta _{e}$ and $\beta _{o}$ are the propagation constant of the even and odd modes, respectively. The coupling lengths are 4.0 ${\rm \mu}$m, 9.2 ${\rm \mu}$m, 98.9 ${\rm \mu}$m and 458.6 ${\rm \mu}$m for the four samples with an As$_2$Se$_3$ core diameter of 0.5 $\mu$m, 0.75 $\mu$m, 2 $\mu$m, 4 $\mu$m, respectively. The short spatial period of 4.0 $\mu$m gives rise to high-frequency ultrasound sensing. The difference between the phases of even and odd modes ($\phi _{d}$) in the DCM is given by $\phi _{d}=2\pi z \Delta n/\lambda$, where $\lambda$ is the wavelength of light, $z$ is the fiber length and $\Delta n$, given by $\Delta n=n_{e}-n_{o}$, is the refractive-index difference with $n_{e}$ and $n_{o}$ being the effective refractive-indices of the even and odd modes. The output power of the DCM is given by $P=P_{0}\cos ^{2}[\pi z \Delta n/\lambda ]$ in As$_2$Se$_3$ core 1 when the input light is injected in the same core, where $P_0$ is the input power. Due to the larger phase difference change over wavelength in the slow axis of the dual-core taper in comparison to the fast axis, the spectral slope at the quadrature point of the transmission spectrum is larger in the slow axis, which leads to a higher sensitivity. The value of $\Delta n$ in the slow axis of the DCM as a function of $\lambda$ is calculated for different core diameters from 0.5 $\mathrm{\mu}$m to 21.25 $\mathrm{\mu}$m when the ${\rm D}_{\textrm{core}}$/d${\rm }_{\textrm{core}}$=0.89 using the finite element method (FEM). The phase difference between even and odd modes at each core diameter of the microfiber, including the core diameters in the waist region and transition region, is calculated in a step of 0.1 $\mathrm{\mu}$m according to the tapered fiber profiles. The $\phi _{d}$ is estimated by adding the calculated phase difference at each core diameter of the DCM. Figure 2(d) shows the calculated transmission spectra of the four fused DCMs with the same waist length of 1 cm over core diameters of 4 $\mathrm{\mu}$m, 2 $\mathrm{\mu}$m, 0.75 $\mathrm{\mu}$m and 0.5 $\mathrm{\mu}$m. The FSR of transmission spectra shows a sharp decrease with reduced core diameter, as plotted in the red curve in Fig. 2(e). The FSR of the transmission spectrum for the DCM with a core diameter of 0.5 $\mathrm{\mu}$m is around 70 times smaller than that of the DCM with 4 $\mathrm{\mu}$m. The spectral slope is calculated from the first derivative of the function $P(\lambda )/P_0$. The maximum-spectral-slope (${\rm S_m}$) at the quadrature point of the normalized transmission spectrum is presented by the black curve in Fig. 2(e). The value of ${\rm S_m}$ for the DCM with a core diameter of 0.5 $\mathrm{\mu}$m (${\rm S_m}$=14.1) is around 74 times larger than that of the DCM with a core diameter of 4 $\mathrm{\mu}$m (${\rm S_m}$=0.19). The large value of ${\rm S_m}$ for the fused DCM with a small core diameter is contributed by the small FSR of the transmission spectrum.

 

Fig. 2. Numerical simulations of fused dual-core As$_2$Se$_3$-PMMA microfibers. (a) Even mode profile. (b) Odd mode profile. (c) Field distribution of transmitting light. (d) Normalized transmission spectra for four DCMs. (e) Calculated ${\rm S}_{\rm m}$ and FSR of the transmission spectra as a function of core diameter.

Download Full Size | PDF

Real-time ultrasound measurement in a dual-core As$_2$Se$_3$-PMMA microfiber is based on the detection of power change instead of the spectral shift of the transmission spectrum at one of the output cores. Using a photo-detector to measure the ultrasound induced power variation at the output of the DCM, the measured AC voltage ($V_{S}$) is given by [23]

3. Experimental results and discussion

Figure 3 presents a schematic setup for ultrasound sensing using a dual-core As$_2$Se$_3$-PMMA microfiber. Light from an Erbium-doped fiber amplifier (EDFA) is linearly polarized using a linear polarizer (LP) and is launched into one of the input cores of the DCM. The polarized light is aligned to the slow axis of the DCM using a polarization controller (PC). The transmission spectrum of the DCM is recorded by an optical spectrum analyzer (OSA). The transmission spectrum represents the interference between the fields propagating in the even and odd modes of the DCM and varies periodically with wavelength due to the different wavelength-dependence of $\Delta n$. Wavelength dips of the transmission spectrum in the DCM are obtained when the phase difference of two modes satisfies the condition $\phi _{d}(\lambda )=(2m+1) \pi$. The normalized transmission spectra of the four fused DCMs are plotted in Fig. 4(a). The value of FSR decreases with reducing core diameter due to the large phase difference of even and odd modes. The FSR of the transmission spectra for different core diameters is plotted by the red curve in Fig. 4(b). The maximum-spectral-slope ${\rm S_m}$ of the normalized transmission spectra at the quadrature points for the four DCMs is measured as a function of core diameter, as shown in the black curve in Fig. 4(b). The value of ${\rm S_m}$ for the DCM with a core diameter of 0.5 $\mathrm{\mu}$m (${\rm S_m}$=20.3 nm$^{-1}$) is 72.5 times larger than one with a core diameter of 4 $\mathrm{\mu}$m (${\rm S_m}$=0.28 nm$^{-1}$), in close agreement with the numerical simulations in Fig. 2(e). The ${\rm S_m}$ for the submicron core microfiber is tens of times larger than that of gratings [23,24] and multi-mode tapers [25].

 

Fig. 3. The schematic experimental setup of ultrasound sensing. EDFA: Erbium-doped fiber amplifier; LP: linear polarizer; PC: polarization controller; AFG: arbitrary function generator; PZT: piezoelectric transducer; PD: photo-detector; OSA: optical spectrum analyzer; ESA: electrical spectrum analyzer; OSC: oscilloscope.

Download Full Size | PDF

 

Fig. 4. (a) Measured transmission spectra of four DCMs. (b) Maximum slopes (S${\rm _m}$) and FSR of the transmission spectra as a function of core diameter.

Download Full Size | PDF

For ultrasound detection, the EDFA is replaced by a continuous wave (CW) light from a tunable laser, and the polarization of the laser is adjusted by the PC to align with the slow axis of the dual-core microfiber. The pump power of incident light is 2 dBm. To get the highest SNR and largest ultrasound frequency range, the wavelength of the CW light is tuned to the wavelength of maximum-spectral-slope at a quadrature point of the transmission spectrum. A PZT with a central frequency of $\sim$3.8 MHz is glued onto a 2 mm-thick aluminum plate, and is driven by a sinusoidal electrical signal with a peak-to-peak voltage (Vpp) of 10 V from an arbitrary function generator (AFG). The fused dual-core is placed on the surface of the aluminum plate, 5 mm away from the PZT. The acoustic wave propagation from the PZT results in a variation of the refractive indices of the even and odd modes and the waist length, inducing a wavelength shift in the transmission spectrum. The power of the CW light is changed by the acoustic wave-induced spectral shift, which is detected by a photo-detector (PD) and recorded by an electrical spectrum analyzer (ESA) and an oscilloscope (OSC). To reduced the acoustic attenuation when the acoustic wave propagates through the aluminum plate and increase ultrasound detection range, the fused dual-core As$_2$Se$_3$-PMMA microfiber is directly placed on the PZT without the aluminum plate.

Figure 5(a) presents the measured signals from the photo-detector as a function of time for the four fused dual-core As$_2$Se$_3$-PMMA microfibers when the PZT is driven by the function generator with a frequency of 100 kHz and a voltage of Vpp=10 V. The output signals show a sinusoidal waveform with a frequency of 100 kHz as the ultrasound wave induces a periodic perturbation on the dual-core fiber when it propagates along the surface of the aluminum plate. The peak-to-peak voltage of the measured sinusoidal signals is plotted in Fig. 5(b), showing a sharp increase with decreasing core diameter. The peak-to-peak voltage for a fused DCM with a core diameter of 0.5 $\mathrm{\mu}$m is 65 times greater than that of 4 $\mathrm{\mu}$m, which is close to the 72.5 fold increase in the maximum slopes ${\rm S_m}$. The small difference between the increase of Vpp and the increase of ${\rm S_m}$ is due to the reduced ${\rm \Delta} \lambda _S$ for a dual-core fiber with a large core diameter [18,20]. The large output voltage for the small-core microfiber ensures high-sensitivity ultrasound sensing.

 

Fig. 5. (a) The ultrasound response of four DCMs with different core diameters in the time-domain at the ultrasound frequency of 100 kHz. (b) Output voltage as a function of core diameter.

Download Full Size | PDF

To characterize the ultrasound sensing frequency response of the dual-core As$_2$Se$_3$-PMMA microfibers, the frequency is varied from 20 kHz to 50 MHz in step of 100 kHz. SNR is measured at each ultrasound frequency using ESA. The PD and ESA are placed in another room of 20 meters away to avoid the detection of radiated electromagnetic waves from an acoustic source setup, in which the PZT is connected to the function generator by a SubMiniature version A (SMA) connector. The SNR of the detected signal for each of the dual-core As$_2$Se$_3$-PMMA microfibers is plotted in Fig. 6(a). The SNR shows several peaks at around 4.0 MHz, 12.5 MHz, 21.1 MHz, 29.2 MHz and 37.8 MHz, corresponding to the central frequency and high odd order harmonics of the PZT. The SNR increases as the core diameter of the DCM decreases, especially at the low-frequency range below 5 MHz. The SNR of the DCM with a core diameter of 0.5 $\mathrm{\mu}$m at hundreds of kHz is 3 orders of magnitude higher than the SNR of a DCM with a core diameter of 4 $\mathrm{\mu}$m. A maximum SNR of >80 dB for the DCM with a core diameter of 0.5 $\mathrm{\mu}$m and a cladding diameter of 28.0 $\mathrm{\mu}$m is detected at 200 kHz as plotted in Fig. 6(b). The maximum detected frequency as a function of core diameter is plotted in Fig. 6(c) showing an increase of maximum detected frequency by 15 MHz as the core diameter decreases from 4 $\mathrm{\mu}$m to 0.5 $\mathrm{\mu}$m. This is because high spectral slope at the quadrature point of the transmission spectrum for the dual-core taper with small core diameter. The frequency range for the dual-core As$_2$Se$_3$-PMMA microfiber ultrasound sensor is 30 MHz larger and SNR is up to 30 dB higher than those of silica-based fiber microfiber sensor using the same PZT as an ultrasound source [26]. The larger detection frequency range and higher SNR for the dual-core As$_2$Se$_3$-PMMA microfiber sensor are due to the smaller Young’s modulus, the larger spectral slope at the quadrature point of the transmission spectrum than that of the silica microfiber. Moreover, compared with the silica microfiber-based ultrasound sensing, the dual-core microfiber with a core diameter of 0.5 $\mathrm{\mu}$m is coated by PMMA cladding with a diameter of 28 $\mathrm{\mu}$m, which provides more mechanical robustness and reduces the chances of breaking during operation.

 

Fig. 6. (a) SNR as a function of frequency for four DCMs with different core diameters. (b) Maximum SNR for the DCM with a core diameter of 0.5 $\mathrm{\mu}$m at 200 kHz sine function. (c) Maximum frequency response as a function of core diameter.

Download Full Size | PDF

Ultrasound sensing SNR as a function of acoustic frequency is measured for 1 cm-long dual-core As$_2$Se$_3$-PMMA microfibers with the same core diameter of d${\rm }_{\textrm{core}}$=0.75 $\mathrm{\mu}$m and different core separations of ${\rm D}_{\textrm{core}}$=0.99d${\rm }_{\textrm{core}}$ and ${\rm D}_{\textrm{core}}$=0.89d${\rm }_{\textrm{core}}$, and the results are plotted in Fig. 7. The maximum detection frequency for the dual-core As$_2$Se$_3$-PMMA microfiber with a core separation of ${\rm D}_{\textrm{core}}$=0.89d${\rm }_{\textrm{core}}$ increase 7.8 MHz, compared with that of the DCM with a core separation of ${\rm D}_{\textrm{core}}$=0.99d${\rm }_{\textrm{core}}$. The transmission spectrum of the DCM with ${\rm D}_{\textrm{core}}$=0.89d${\rm }_{\textrm{core}}$ has a smaller FSR and a sharper spectral slope at the quadrature point than those of the DCM with ${\rm D}_{\textrm{core}}$=0.99d${\rm }_{\textrm{core}}$, as plotted in the inset of Fig. 7, due to the larger rate-of-change of $\phi _d$ with $\lambda$ for the DCM with a smaller ${\rm D}_{\textrm{core}}$. The value of maximum-spectral-slope for the DCM with ${\rm D}_{\textrm{core}}$=0.89d${\rm }_{\textrm{core}}$ (${\rm S_m}=5.2$ nm$^{-1}$) is more than twice larger than that of the DCM with ${\rm D}_{\textrm{core}}$=0.99d${\rm }_{\textrm{core}}$ (${\rm S_m}=2.5$ nm$^{-1}$), which leads to a larger sensitivity resulting in a larger ultrasound frequency range for the ultrasound measurement. To increase the ultrasound sensing maximum detection frequency, a dual-core As$_2$Se$_3$-PMMA microfiber with a submicron core diameter and a close core separation is favorable.

 

Fig. 7. SNR of two DCMs with different core separation ${\rm D}_{\textrm{core}}$ as a function of frequency. Inset: Transmission spectra for the two microfibers.

Download Full Size | PDF

Ultrasound acoustic waves from the PZT are attenuated when traveling through the aluminum plate, and the attenuation increases as the ultrasound frequency increases resulting in a lower SNR at higher frequencies, leading to the understated measurement of the actual frequency detection range of the fiber ultrasound sensor. To eliminate the effect of acoustic attenuation on the maximum ultrasound frequency detection of the sensor, the compact dual-core chalcogenide-PMMA microfiber with waist length of 1 cm, core diameter of 0.5 $\mathrm{\mu}$m and core separation of 0.445 $\mathrm{\mu}$m is directly fixed on the PZT without the aluminum plate avoiding the propagation attenuation of acoustic waves. The SNR as a function of ultrasound frequency in the DCM is measured, as presented by the red curve in Fig. 8(a), showing a broadband ultrasound range from 20 kHz to 80 MHz with an average SNR of 31 dB. The frequency range for the SNR above 10 dB is 68 MHz, and SNR is 5-10 dB for the frequency range of 68-80 MHz. Compared with the ultrasound sensing with the 2 mm thick aluminum plate as plotted by the red curve in Fig. 6(a), the maximum detection frequency increases by 40 MHz for ultrasound sensing without the aluminum plate due to the elimination of the attenuation of acoustic waves. The ultrasound detection with the maximum frequency of 80 MHz offers great potential for high-resolution biomedical imaging.

 

Fig. 8. (a) SNR of the fused DCM with a core diameter of 0.5 $\mathrm{\mu}$m without the aluminum plate, and ${\rm S}_{11}$ of the PZT for frequencies from 300 kHz to 80 MHz. Inset: Close-up of high frequency range. Frequency response at 79.95 MHz (b) with connecting fiber sensor; (c) without connecting fiber sensor. (d) Detected radiated electromagnetic waves and microfiber sensor ultrasound response.

Download Full Size | PDF

Furthermore, the amplitude of the generated ultrasound wave from the PZT reduces as the frequency increases for the same driving voltage amplitude. Therefore, the frequency detection range of the sensor must take into account the reduced amplitude of the generated ultrasound signal from the PZT at higher frequencies. Acoustic waves with different frequencies from the PZT are characterized by measuring electrical reflection coefficient S$_{11}$ in a S-parameter network analyzer [27]. S$_{11}$ is equal to the ratio of the power of a reflected electromagnetic wave and the power of an incident electromagnetic wave on the PZT. S$_{11}$=0 dB implies that all the power of the incident electromagnetic wave is reflected from the PZT and nothing is radiated. The S$_{11}$ value for a frequency range from 300 kHz to 80 MHz is measured, as presented in the black curve in Fig. 8(a). We observe the first dip at 4 MHz and several dips with a period of $\sim$8 MHz corresponding to the central frequency and high odd order harmonics of the PZT. At higher frequencies, PZT has little absorption to driving power as measured in S$_{11}$ parameter, as presented in the inset of Fig. 8(a), and acoustic periods decrease, resulting in a reduction in the amplitude of the generated acoustic waves and the fiber sensor displacement. Indeed, the SNR decreases at high frequencies due to the reduced phase change of the dual-core taper. Figure 8(a) shows that the peaks of SNR coincide with the dips in the S$_{11}$ frequency response of the PZT confirming that the amplitude of generated acoustic waves from the PZT correlates with the amplitude of the electromagnetic reflection from the PZT. Due to the non-zero radiated acoustic wave of the PZT response at the non-resonant frequencies and the large acoustic wave induced-strain at low ultrasound frequencies, the ultrasound signals with high SNR are obtained. The weak amplitude of ultrasound waves from the PZT at higher frequencies indicates that a wider ultrasound detection frequency range for the As$_2$Se$_3$-PMMA microfiber sensor can be measured in the presence of an ultrasound source that can produce waves at such high frequencies.

Finally, it is essential to ensure that the measured signal at the ESA does not include the radiated signal from the acoustic source setup. The ultrasound responses at 79.95 MHz in the cases with and without connecting the DCM sensor are measured as shown in Fig. 8(b) and Fig. 8(c), respectively. The ultrasound response at 79.95 MHz is selected rather than 80 MHz to eliminate the artifact harmonics that appear at integer multiples of the reference frequency of 10 MHz of the function generator. The measured response in Fig. 8(c) shows that radiated signal from the acoustic source setup is not captured by the ESA. The measurement is repeated for ultrasound frequencies from 20 kHz to 80 MHz without connecting the DCM sensor, and no signal is detected above the noise floor, which indicates that the radiated electromagnetic wave from the acoustic source setup does not affect the ultrasound measurement. To further ensure that the radiated electromagnetic waves from the acoustic source setup do not affect the ultrasound measurement, the power of the radiated electromagnetic waves as a function of frequency is measured by the ESA when PD is turned off, as presented by the blue curve in Fig. 8(d). The power of the radiated electromagnetic waves from the acoustic source setup is at least 15 dB below the noise floor measured when the PD is turned on, confirming that the radiated electromagnetic waves do not affect the high-frequency ultrasound measurement using the dual-core As$_2$Se$_3$-PMMA microfiber sensor.

4. Conclusion

In conclusion, we have demonstrated broadband ultrasound sensing based on fused dual-core As$_2$Se$_3$-PMMA microfibers. The ultrasound detection frequency range and SNR have been measured in dual-core microfibers with different core diameters and core separations. The dual-core taper sensor with a large spectral slope at the quadrature point of the transmission spectrum contributes to high-frequency ultrasound sensing. Ultrasound sensing with a broadband acoustic frequency range from 20 kHz to 80 MHz and an average SNR of 31 dB is achieved in a compact, mechanically robust, dual-core microfiber with a waist length of 1 cm, an As$_2$Se$_3$ core diameter of 0.5 $\mathrm{\mu}$m and a core separation of 0.445 $\mathrm{\mu}$m. The fused dual-core As$_2$Se$_3$-PMMA microfiber sensors with high-performance ultrasound detection can potentially be used for a variety of practical applications including biomedical and industrial high-resolution imaging.