Ultrasound Signal Detection with Multi-bounce Laser Microphone

Abstract

The multi-bounce laser microphone utilizes optical methods to detect the displacement of a gold-covered thin film diaphragm caused by ultrasound signal pressure waves. This sensitive all-optical sensing technique provides new opportunities for advanced ultrasound imaging as it is expected to achieve a higher detection signal-to-noise ratio (SNR) in a broader spectrum, as compared to conventional ultrasonic transducers. The technique does not involve signal time-averaging and the real-time enhancement in detection SNR stems from the amplification of signal strength due to multiple bouncing off the diaphragm. The system was previously developed for detecting acoustic signatures generated by explosives and were limited to lower than 10 kHz in frequency. To demonstrate its feasibility for biomedical imaging applications, preliminary experiments were conducted to show high fidelity detection of ultrasound waves with frequencies ranging from 100 kHz to in excess of 1 MHz. Experimental results are also presented in this work demonstrating the improved detection sensitivity of the multi-bounce laser microphone in detecting ultrasound signals when compared with a commercial Fabry-Perot type optical hydrophone. Furthermore, we also applied the multi-bounce laser microphone to detect photoacoustic signatures emitted by India ink when a LED bar is used as the excitation source without signal averaging.

I. Introduction

Diagnostic ultrasound is one of the most common imaging modalities in the medical imaging field. One critical component of this technology is designing sensitive receivers for detecting megahertz ultrasound signals. Currently, piezoelectric transducers have been the most widely used detectors for ultrasound signals. However, newer imaging techniques, such as photoacoustic imaging, require broader detection bandwidth and more sensitive detectors in order to work with significantly weaker clinical signals. As mentioned in [1], conventional piezoelectric transducers may be limited in their ability to detect the full frequency bandwidth with high sensitivity, due to restricted transducer size and material composition. Sensitive and broadband ultrasound signal detecting techniques are needed for the development of advanced ultrasound imaging.

For clinical diagnosis and surgical monitoring and guidance, many imaging modalities are able to non-invasively show regional anatomy and position details, but few of them are able to provide functional information. Monitoring the concentration of metabolism-related species or bio-markers could also be helpful during surgery and in interventional imaging. In retinal imaging, for example, Ophthalmoscope and Optical Coherence Tomography (OCT) allows ophthalmologists to identify micron-scale anatomical features for diagnosis and surgical guidance. Some general physiological indices that correlate with health and diseases can also be accessed through biochemical analysis. However, to determine whether the tissue in a specific anatomic location is injured, dead or healthy, more detailed physiological information is required. This may include but not be limited to metrics of blood circulation such as blood flow, oxygen saturation (sO2), carbon dioxide (pCO2) and other in vivo indices [2]. Photoacoustic imaging represents a potential solution for this specific demand and has garnered research interests in the areas of structural and functional imaging. Currently, challenges to application of photoacoustic imaging fall broadly into two area: 1) the sensitivity of detection provided by the transducers and 2) safety concerns and clinical flow interruption due to laser hazards [3].

The multi-bounce laser microphone is an all-optical sensing technique previously developed for detecting acoustic signatures generated by explosives below 10 kHz. It is also easier to build up a compact assembly of the sensing components. This system may act as a useful detecting tool to address the challenges in application of photoacoustic imaging as it is expected to achieve a higher signal-to-noise ratio (SNR), in a broader spectrum, as compared to conventional transducers. To demonstrate its feasibility in biomedical imaging field, the experimental design in this work strategically applied this technique to the detection of ultrasound signals ranging from 100 kHz to more than 1 MHz. We also used the multi-bounce laser microphone to detect photoacoustic signatures emitted by India ink with a LED bar working as the excitation source.

II. Methods

A. Experiment Configuration and Devices

The experimental configuration shown in Fig. 1 is developed from the design in [4]-[5]. A pulsed laser source operating at 100 kHz pulse repetition rate (PRR) is used to build up the multi-bounce laser microphone. The laser beam splits into reference and signal beams. The reference beam goes directly to the photo -electromotive force (photo-EMF) detector (Brimrose Corp., Baltimore, MD). Pressure waves from single element piezoelectric transducer (V318-SU, Olympus Scientific Solutions Americas Inc., Waltham, MA) propagate through the phantom to a gold-coated thin film diaphragm. The diaphragm that attached to the phantom surface functions as a mirror where the signal beam interrogates on the propagating pressure waves before reaching the detector. Another gold mirror is placed in front of the diaphragm for multiple bounces between gold surfaces. The photo-EMF detector decodes the phase difference between reference and signal beams by producing electrical currents. When detecting the pressure waves, the phase difference caused by the displacement of the diaphragm surface would be accumulated through multiple bounces of the laser signal beam, leading to amplified response expressed in output currents, therefore enhancing the sensitivity of detection.

Fig. 1.

Design of the multi-bounce laser microphone experiment. (A)Multiple bounces of the laser beam indicated by bright spots on the gold-covered diaphragm viewed through an infrared viewer; (B)Multiple bounces between gold mirror and gold-coated diaphragm; (C) Assembled phantom: ① diaphragm, ② silicone phantom and ③ 3D-printed phantom holder.

Detection of photoacoustic signatures from India ink is also presented under the illumination of 850 nm LED pulses with 135 ns pulse width, all without involving any sample averaging commonly seen in prior publications like [6], in which different number of averages are used to minimize the noise.

B. Phantom and Holder Design

Different materials, like agar, plastic and silicone, were used to construct the phantoms for the experiments reported here. The two types of silicone materials used in the experiments are high-performance platinum cure liquid silicone compounds Dragon Skin 10 Medium and high-impact urethane rubber Simpact 85A (Smooth-on Inc., Macungie, PA).

The silicone material and curing agent were mixed by a prescribed volume ratio. The mix was stirred well and poured into the 3D-printed holder (made of Polylactic Acid, PLA). The silicone compounds would then cure to the desired form in the holders. The 3D-printed phantom holders were designed to work as a mold and improve the stability of the resultant phantom. Fig. 1 (C) shows the assembly of 3D-printed holder, silicone phantom and gold-coated diaphragm.

C. Optical Hydrophone Test

A commercial Fabry-Perot type optical hydrophone was used as a reference to monitor the acoustic pressure level on the diaphragm surface when the single-element ultrasound transducer was used as the source of acoustic agitation. The fiber end of the optical hydrophone (Fibre-optic Hydrophone System, Precision Acoustics Ltd., Dorset, UK) was directed perpendicularly to the phantom’s front surface, detecting ultrasound signal emitted by the single element ultrasound transducer. The very same phantom was then interrogated by the laser microphone and the outputs from these two different modalities were then compared under similar acoustic agitation conditions.

D. Data collection and Signal Analysis

An oscilloscope (Digital Oscilloscope TDS 3034C, Tektronix Inc., Beaverton, OR) was used for output data display and acquisition. Due to the very limited laser PRR of the laser microphone, a sampling oscilloscope-like approach was adopted to provide sufficient samples of the ultrasound waves whose frequencies were significantly higher than the PRR of the laser microphone. With the envelope-detection function of the oscilloscope, multiple frames of data were collected and overlaid prior to data acquisition.

The oscilloscope exported each dataset with 10000 data points. Data of continuous wave detection would be further processed with envelop detection function and frequency analysis in MATLAB.

III. Experiment Results

A. Continuous ultrasound signal detection

The single element transducer V318-SU is driven by sine wave signal generated from an arbitrary function generator and amplified by the radio frequency amplifier (RFA). For tests with RFA, the peak-to-peak driving voltage for the single element transducer is about 220V, while for tests carried out at Brimrose Lab without RFA, it is set at 10 to 26V. Fig. 2 shows temporal traces (lasts for 10 μs) collected using digital oscilloscope and frequency analysis output of 500 kHz ultrasound signal detection using different systems.

Fig. 2.

Continuous wave signal detected by different systems with silicone phantoms

Sine wave signals can be observed sitting on top of the pulses and achieving about 1.5V peak-to-peak in amplitude when detecting with 13-bounce laser microphone. When using optical hydrophone, the detection presents about 0.0025V peak-to-peak in output amplitude. In both cases, the peak value of the detected signal is located around 500 kHz in spectrum.

The signal-to-noise ratio (SNR) for the ultrasound signal detection test is determined in Eq. (1). $A_{signal}$ is the peak absolute FFT value, while $A_{noise}$ is the absolute FFT value at around 10 times higher of the signal peak frequency.

$$SNR=20\ast {\log }_{10}\left(\frac{A_{signal}}{A_{noise}}\right)dB.$$ (1)

Table I. shows all calculated SNR results of tests with different experiment design.

TABLE I.

Test Results of Ultrasound Signal Detection with Multi-bounce Laser Microphone

Phantom	Test Results
	System a	SNR (dB)	Experiment Description b
Agar	M	40	220V, one-bounce
	O	51	220V
Plastic	M	38	220V, one-bounce
	O	37	220V
Silicone	M	32	220V, one-bounce
	O	40	220V
	M	38	220V, 3-bounce
	M	42	10V, 13-bouncec
	M	46	15V, 13-bouncec

^a.Indicates which system is used. ‘M’ for laser microphone and ‘O’ for fibre-optic hydrophone.

^b.Description includes driving voltage tor the single element transducer and indicates number of bounces if using multi-bounce laser microphone system.

^c.Experiments are carried out at Brimrose Lab

B. Pulsed mode signal detection

The tests are carried out with a similar configuration for continuous wave testing. To realize the pulsed mode ultrasound signal excitation, another arbitrary function generator is applied to control on-and-off for the driving signal. The single element transducer is driven by a pulsed mode ultrasound signal which has 3 cycles of sine wave periodically appear every 10 to 40 μs. Fig. 3 presents detected temporal traces of both continuous wave and pulsed mode waves. The sine wave cycles enclosed by the yellow rectangular boxes represent the detected ultrasonic pulses depicted in different time scales. Looking at the trace of pulsed mode ultrasound signal that has 3 cycles every 40 μs and the continuous wave signal trace, we can find that the pulse cycles were detected at approximately same location where the cycles appeared in the continuous wave trace.

Fig. 3.

Pulsed wave signals detected by different systems with agar phantoms

C. Photoacoustic signal detection

The photoacoustic signal traces detected by multi-bounce laser microphone are collected through the oscilloscope when one LED bar of 850 nm wavelength and the 135 ns pulse width was used to illuminate samples of India ink contained in a medical tubing with the inner diameter of 0.9 mm. After preliminary signal processing involving moving averages, results of 3 traces are presented in Fig. 4.

Fig. 4.

Photoacoustic signal detected by multi-bounce laser microphone

The traces are quite noisy but a similar pattern can be found among all the traces in the time window between 10 to12 us. The trend (indicated by the red curve in Fig. 4) is climbing to local maximum peaks at around 10.9us and 11.1us, then going down to local minimum peaks at around 11us and 11.2us. Note that the results shown in Figure 4 does not involve averaging over multiple of traces which is typically found in prior research involving the use of LED for photoacoustic signature generation.

IV. Conclusion

The experiments conducted here prove that the multi-bounce laser microphone is capable of detecting megahertz range continuous wave ultrasound signals as well as some pulsed wave signals. Some of SNR results of the tests are listed in Table I. It clearly demonstrated that the SNR of the multi-bounce laser microphone can be improved as the number of bounces increases. To date, we have succeeded in introducing up to 21 bounces in this system. Various aspects of the laser microphone approaches are being pursued and further improvement in detection SNR in excess of a factor of 10 has been achieved and will be presented in upcoming publications.

Fig. 5 shows a comparison between signals detected with the multi-bounce laser microphone system and the traditional optical hydrophone system. Even with much lower driving voltage to the single element ultrasound transducer, which means a less powerful emitted ultrasound signal to be detected, the multi-bounce laser microphone system still outperforms the conventional optical hydrophone in the SNR comparison for detected ultrasound signals.

Fig. 5.

Comparison between ultrasound signal detected with different systems

For the photoacoustic signal detection test using LED bar as the light source, all the results are obtained without using any multi-trace averaging that are commonly found in prior photoacoustic detection work in order to minimize the noise and boost up detection SNR. Such improved detection sensitivity by the laser microphone technology is likely to help achieve higher frame rate for photoacoustic imaging using low-power excitation light source including the LEDs.

With various improvements in device design, the number of bounces deployed by the laser microphone can be further increased, especially in relatively slow regime of MHz in ultrasonic frequencies, leading to further enhancement in detection sensitivity. Work is also underway to explore the detection of photoacoustic signatures with broader frequency bandwidth than what we have presented in this paper. We have demonstrated that the multi-bounce laser microphone technique has the potential to detect ultrasound and photoacoustic signals generated by much lower illumination power sources, such as LED bars, which could not be consistently achieved with conventional piezoelectric ultrasound transducers. In many clinical scenarios, photoacoustic imaging with high power laser is not yet allowed due to safety issues related to laser exposure. Switching to an LED illumination source and maintaining the illumination power at a safe level may be a logical next step moving the photoacoustic imaging into more clinical applications. The higher SNR enhanced by this multi-bounce laser microphone system is definitely a key innovation for the detection of the relatively weaker acoustic signals generated by illumination with an LED system.