Optical detection of ultrasound is impractical in many imaging scenarios because it often requires stable environmental conditions. We demonstrate an optical technique for ultrasound sensing in volatile environments with miniaturization and sensitivity levels appropriate for optoacoustic imaging in restrictive scenarios, e.g. intravascular applications.
Optical sensors of ultrasound are a promising alternative to piezoelectric techniques, as has been recently demonstrated in the field of optoacoustic imaging. In medical applications, one of the major limitations of optical sensing technology is its susceptibility to environmental conditions, e.g. changes in pressure and temperature, which may saturate the detection. Additionally, the clinical environment often imposes stringent limits on the size and robustness of the sensor. In this work, the combination of pulse interferometry and fiber-based optical sensing is demonstrated for ultrasound detection. Pulse interferometry enables robust performance of the readout system in the presence of rapid variations in the environmental conditions, whereas the use of all-fiber technology leads to a mechanically flexible sensing element compatible with highly demanding medical applications such as intravascular imaging. In order to achieve a short sensor length, a pi-phase-shifted fiber Bragg grating is used, which acts as a resonator trapping light over an effective length of 350 µm. To enable high bandwidth, the sensor is used for sideway detection of ultrasound, which is highly beneficial in circumferential imaging geometries such as intravascular imaging. An optoacoustic imaging setup is used to determine the response of the sensor for acoustic point sources at different positions.
Ultrasound detectors play a key role in many imaging applications. Conventionally, ultrasound is detected by piezoelectric transducers, which transform pressure waves into voltage signals1. In optoacoustic imaging, ultrasound is generated via a process of thermal expansion by illuminating the object with high-power modulated light2-6. Although piezoelectric transducers are the method of choice in optoacoustic applications, their use often hinders miniaturization mainly because miniaturized piezoelectric transducers are often characterized by low sensitivity. Additionally, since piezoelectric transducers are optically opaque, they may severely interfere with light delivery to the imaged object, limiting possibilities for usable imaging configurations. Light that is back-scattered from the object to the transducer may also limit the proper detection of ultrasound and complicate the design of the imaging system due to optically induced parasitic signals in the transducer7.
Optical detectors of ultrasound have been recognized as a possible alternative to piezoelectric transducers that offers many benefits in optoacoustic imaging scenarios8-12: They are often transparent and can be usually miniaturized without loss of sensitivity. The working principle of optical detectors is interferometric detection of the minute deformation created in the optical medium due to the presence of ultrasound. Often, optical resonators are used to enhance detection sensitivity by trapping light in the perturbed medium for extended durations, thus increasing the effect of the deformation on the phase of the optical signal. In those cases, optical detection schemes are based on monitoring variations in the resonance wavelength, which directly relate to structure deformations in the resonator. Most commonly, narrow-linewidth continuous wave (CW) techniques are used in which a CW laser is tuned to the resonance wavelength. Small shifts in the resonance wavelength change the relative position of the laser’s wavelength within the resonance, thus causing variations in the intensity of the transmitted/reflected laser light, which can be readily monitored. However, if the resonance shifts are too strong, e.g. owing to large variations in pressure, temperature, or vibrations, the resonance may shift completely away from the laser’s wavelength, effectively saturating the detector13.
Pulse interferometry14 offers a solution to the limitation of signal saturation and enables ultrasound detection under volatile environmental conditions. In contrast to narrow-linewidth CW schemes, pulse interferometry employs a wideband pulse source to illuminate the resonator. In this case, the resonator acts as a bandpass filter, transmitting only those wavelengths that correspond to its resonance frequency, while the resonance shifts are detected by measuring the wavelength variations in the optical signal at the resonator’s output, e.g. by using a Mach-Zehnder interferometer locked to quadrature14,15. An automatic reset circuit is used to immediately restore the interferometer’s working point in the case it is lost due to extreme variation in environmental conditions. Because of the relatively broad bandwidth of the source, the resonance wavelength stays within the illuminated band even under strong perturbations, enabling stable detector operation even under harsh ambient conditions. The use of a coherent source for interrogation, i.e. optical pulses, facilitates low-noise detection.
The corresponding pulse interferometry system used in our experiments is shown in Figure 1. The pulse laser used for interrogation produced 90 fsec pulses at a repetition rate of 100 MHz with output power of 60 mW and spectral width of over 100 nm. The optical filter had a FWHM spectral width of approximately 0.4 nm and was tuned to the frequency of the resonance. Following the filter, an optical amplifier was used to compensate for the significant loss in the filtering. Additional filtering was applied after the amplification stage to reduce amplified spontaneous emission from the amplifier. The resonator used in our experiments is a pi-phase-shifted fiber Bragg grating (π-FBG)8, manufactured by Teraxion Inc. Particularly for the medical application of ultrasound sensing, π-FBGs have the benefit of being all-fiber components, and thus robust and small. Figure 2 shows a comparison between the dimensions of the optical fiber used in this work and a 15 MHz miniaturized intravascular ultrasound (IVUS) piezoelectric transducer. Some alternative resonance-based detection approaches, such as micro-ring resonators fabricated in planar waveguides, require coupling fibers at the component’s input and output, either leading to more fragile devices or hindering miniaturization. In contrast, π-FBGs are in-fiber components, and do not require additional fiber coupling. The resonance in π-FBGs is created by the pi phase shift in their center; light is trapped around the pi phase shift over portion of the fiber which is considerably shorter than the length of the grating itself. In our experiments, the π-FBG had a length of 4 mm and coupling coefficient of κ = 2 mm-1 and its sensitivity was distributed non-uniformly along its length, with the sensitivity exponentially decreasing from the grating’s center with a rate of κ. The full-width-half-maximum (FWHM) of the sensitivity distribution (SD) was approximately 350 µm. The resonance width of the grating is determined by both its length and its coupling coefficient according to the following equation:
(1)
where λ is the resonance wavelength and neff is the effective refractive index of the mode guided in the fiber8.
To assess whether the π-FBG detector is appropriate for imaging applications, its spatially dependent response needs to be measured over a wide frequency band. However, this task is extremely challenging when conventional acoustic techniques are used. We therefore employ an optoacoustic method for ultrasound detector characterization16 in which a dark microscopic sphere embedded in transparent agar serves as an optoacoustic point source. In our experiment, the microscopic sphere has a diameter of approximately 100 µm and is illuminated with high power nanosecond optical pulses with a repetition rate of 10 Hz, pulse duration of approximately 8 nsec, and average power of 200 mW. The optical energy deposited in the microscopic spheres generates broadband ultrasound signals owing to the optoacoustic effect. The π-FBG detector is translated relatively to the microscopic sphere to obtain its spatially dependent acoustic response. Figure 3 shows an illustration of the optoacoustic experiment. Generally, this technique can be employed to characterize different kinds of ultrasound detectors.
1. Optoacoustic Characterization of the π-FBG Detector
2. Estimation of Robustness and Sensitivity of the π-FBG Detector's Performance
Figures 4a and 4b respectively show the signals and their corresponding spectra from the microscopic sphere at a distance of 1 mm from the fiber for three offsets from the center of the π-FBG. The offsets are given in the z direction, as depicted in Figure 3. Clearly, The optical detector’s sensitivity to high frequency ultrasound (f > 6 MHz) is anisotropic and is highest when the center of the π-FBG is directly above the microscopic sphere. Despite the high acoustic impedance mismatch between the silica fiber and water, no distinct resonances are observed at frequencies above 6 MHz, leading to a well-defined sharp optoacoustic signal, required for imaging applications. Generally, although the resonance frequencies at f < 6 MHz may be used for sensing, their use for imaging would require their incorporation in a model-based reconstruction model, significantly complicating the image formation algorithm17.
Figure 5 shows a comparison between ultrasound signals measured using a pulse source and a low-coherence source. Because of the low sensitivity obtained by the low-coherence source, an optoacoustic source with a higher magnitude was used in comparison to the one used in the experiment of Figures 3-4. Namely, the optoacoustic source was a graphite rod with a diameter of 0.7 mm, positioned at an approximate distance of 1.5 mm from the fiber and illuminated with the same laser used in the experiment of Figures 3-4. A significant reduction in sensitivity of a factor of 18 is observed for the signals detected with the low-coherence source. The lower sensitivity that is obtained in the case of the incoherent wideband source is inherent as the wideband spectrum of the source is generated by a random process. In contrast, in the coherent pulse source, the wideband response is a result of a deterministic process.
Figure 1. The optical setup used for ultrasound detection. The sensing element is a pi-phase-shifted fiber Bragg grating, and the read-out system is based on pulse interferometry. Click here to view larger image.
Figure 2. A size comparison between a commercial intravascular ultrasound probe with a central frequency of 15 MHz and the optical-fiber based sensing element used in this work. Click here to view larger image.
Figure 3. An illustration of the optoacoustic setup used for measuring the acoustic response of the optical detector. A dark microscopic sphere illuminated with high-power nanosecond pulses constitutes an acoustic point source, which is translated in three dimensions to obtain a spatially dependent acoustic response of the detector. Click here to view larger image.
Figure 4. The signals (a) and their corresponding spectra (b) detected from the microscopic sphere (as depicted in Figure 3) at a distance of 1 mm from the fiber for three offsets from the center of the π-FBG. The spectra are compared to the spectrum of an ideal spherical source with a diameter of 100 µm. Click here to view larger image.
Figure 5. A comparison between ultrasound signals obtained using a pulse source and a low-coherence source. A significant reduction in sensitivity is observed for the signals detected with the low-coherence source. Click here to view larger image.
In conclusion, a new optical method for ultrasound detection is introduced, which is based on a combination of a π-FBG and pulse interferometry. The technique is especially suited for optoacoustic imaging applications owing to the transparency of the sensing element, which enables almost arbitrary object illumination patterns. In contrast, standard piezoelectric based ultrasound detectors are opaque and thus block some of the optical paths to the imaged object, leading to bulky imaging setups. The developed optical detector can thus facilitate the miniaturization of optoacoustic technology and its clinical translation.
The physical and mechanical properties of the sensing element depend on the fiber used. Commercially available single mode fibers are relatively durable and small. For example, in silica fibers, such as the one used in this paper, diameters of 250 µm or smaller and breakage bend radii of less than 1 cm are standard. Plastic fibers may also be used and may have better mechanical properties; however, the fabrication of high quality FBGs is currently commercially available only in silica fibers.
The design of the pi-phase-shifted FBG determines the sensitivity and spatially dependent acoustic response of the optical detector. Generally, it is desired that the resonance be as narrow as possible to achieve maximum sensitivity. However, the width of the resonance measured in Hertz, must be higher than the desired acoustic bandwidth for the detector to allow its proper operation. Additionally, a high-quality π-FBG is currently a custom-made product whose fabrication requires high-precision manufacturing capabilities offered by only few companies.
Pulse interferometry is used for reading the signal from the optical sensing element and enables robust performance under volatile environmental conditions. The bandwidth of the source determines the tradeoff between robustness and performance: If the bandwidth is chosen to be too small, it will cover the resonance only for weak perturbations. If the bandwidth is too large, only a fraction of the energy at the input of the FBG will be transmitted. The bandwidth is controlled by optical bandpass filters, which also provide an additional benefit of reducing noise in the system due to amplified spontaneous emission.
The sensitivity field of the ultrasound detector plays an important role in optoacoustic imaging applications. It is therefore recommended that the response of the detector be characterized before its incorporation in an optoacoustic setup. In our experiments, the π-FBG provides good sensitivity at high frequencies (f > 6 MHz) only when the point source is positioned close to the grating’s center (Figure 4). This suggests that the detector has a relatively non-diverging sensitivity field. Therefore, when used in optoacoustic imaging experiments, it is highly beneficial for illumination to be delivered to regions in which high ultrasound detection sensitivity is obtained.
The authors have nothing to disclose.
D. R. acknowledges support from the German Research Foundation (DFG) Research Grant (RA 1848/1) and the European Research Council Starting Grant. V. N. acknowledges financial support from the European Research Council Advanced Investigator Award, and the BMBF's Innovation in Medicine Award.
π-FBG | Teraxion Inc. | Custom made device | |
microscopic sphere | Cospheric LLC | BKPMS 90-106um- 10g | 100 µm polyethylene microsphere |
Femto-second pulse laser used for interrogation | Menlo Systems GmbH | T-Light Femtosecond Laser | |
Optical filter | Optoplex Corporation | 2-Port Optical Tunable Filter (50 GHz) | |
Optical amplifier | Amonics | AEDFA-PM-PA-35-B-FC | Benchtop 35dB Gain Pre Amp Polarization Maintaining EDFA |
50/50 coupler | OZ-Optics | FUSED-22-1550-8/125-50/50-3S3S3S3S-3-0.5-PM | Fused 2×2 fiber splitter with 0.5 meter long, 3mm OD PVC jacketed 1550nm 8/125μ PM fiber pigtails, 50/50 split ratio in the slow axes of PM fibers and with super FC/PC connectors on all ports. |
Fiber holder | Thorlabs | T711/M-250 | Metric, Post-Mountable Fiber Clamp, 250 µm |
Agar for microbiology | Sigma Aldrich | 05039-500G | |
Nano-second pulse laser used for generating the optoacoustic signals | Opotek | VIBRANT Arrow 532 type I | |
Graphite rod | Faber-Castell | 120700 | Faber-Castell Pencil Leads – 0.7 mm |