IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control ,
vol. 59, no. 12,
December
2012
2631
Broadband High-Frequency Measurement of Ultrasonic Attenuation of Tissues and Liquids Johannes Bauer-Marschallinger, Thomas Berer, Hubert Grün, Heinz Roitner, Bernhard Reitinger, and Peter Burgholzer Abstract—The ongoing expansion of the frequency range used for ultrasonic imaging requires increasing attention to the acoustic attenuation of biomaterials. This work presents a novel method for measuring the attenuation of tissue and liquids in vitro on the basis of single transmission measurements. Ultrasound was generated by short laser pulses directed onto a silicon wafer. In addition, unfocused piezoelectric transducers with a center frequency of 50 MHz were used to detect and emit ultrasound. The laser ultrasound method produces signals with a peak frequency of 30 MHz. In comparison to piezoelectric generation, pulse laser excitation provides approximately 4 times higher amplitudes and 20% larger bandwidth. By using two excitation methods in succession, the attenuation parameters of porcine fat samples with thicknesses in the range of 1.5 to 20 mm could be determined quantitatively within a total frequency range of 5 to 45 MHz. The setup for liquid measurements was tested on samples of human blood and olive oil. Our results are in good agreement with reports in literature.
I. Introduction
M
ethods for measuring ultrasonic attenuation in biological tissues and liquids have been a subject of scientific research since the development of medical ultrasonic imaging. These measurements are normally based on piezoelectric transducers with a narrowband frequency spectrum of up to 10 MHz. Attenuation data for higher frequencies are rare; however, they are of interest for recent developments in the field of medical ultrasonic imaging. This work presents a method for measuring acoustic attenuation up to frequencies of 75 MHz on the basis of ultrasonic generation by short laser pulses and piezoelectric transducers. The information loss resulting from acoustic attenuation determines how far inside the body images can be acquired and what resolution those images may achieve. Images of tissue from deep within the body may have less resolution than images from beneath the skin, because ultrasonic attenuation in the subcutaneous fat layer in-
Manuscript received June 26, 2012; accepted September 19, 2012. This work has been supported by the Austrian Science Fund (FWF), projects TRP102-N20 and S10503-N20; the European Regional Development Fund (EFRE) in the framework of the EU program Regio 13; the Christian Doppler Research Association; the Federal Ministry of Economy, Family, and Youth; and the federal state of Upper Austria. J. Bauer-Marschallinger, T. Berer, B. Reitinger, and P. Burgholzer are with Christian Doppler Laboratory of Photoacoustic Imaging and Laser Ultrasonics, Linz, Austria, and with Research Center for NonDestructive Testing GmbH (RECENDT), Linz, Austria (e-mail:
[email protected]). H. Grün and H. Roitner are with Research Center for Non-Destructive Testing GmbH (RECENDT), Linz, Austria. DOI http://dx.doi.org/10.1109/TUFFC.2012.2504 0885–3010/$25.00
creases with frequency [1]. Because high-frequency applications in ultrasonic imaging are currently under development, strength and frequency dependency of attenuation beyond 10 MHz are increasingly of interest. For example, traditional pulse–echo ultrasound has been extended to frequencies up to 100 MHz and used to investigate the human dermis [2]–[4]. A promising technology called photoacoustic tomography (PAT) [5], [6] also operates with broadband and high-frequency ultrasonic signals. PAT combines the advantages of diffuse optical imaging (high contrast) and ultrasonic imaging (high spatial resolution). In addition to many other factors, the accuracy of the acquired images depends on acoustic attenuation [7]. It is not very surprising that reconstructed images from attenuated signals have less contrast and lack sharpness. For example, in the case of early detection of breast cancer with PAT, small tumors in the mammary gland might not be visible without compensation of attenuation. Models of acoustic attenuation in photoacoustic imaging and possible compensation methods on the basis of Szabo’s equation [8] are proposed by, e.g., Roitner and Burgholzer [9] and La Rivière et al. [10]. Another approach developed by a group at the University College of London [11], [12] utilizes a modified wave equation with fractional Laplacian [13]. Of course, the performance of compensation depends on the noise level in the measured signals and an accurate knowledge of the attenuation parameters. Raju and Srinivasan [3] and Bamber [14] give an overview of studies concerned with ultrasonic attenuation of fat tissue. The attenuation parameters are well known for frequencies up to 10 MHz (see, e.g., [15], [16]). However, for higher frequencies, not quite as many investigations (e.g., [2], [17], [18]) have been conducted. This work contributes to the development of measuring methods of attenuation of tissues and liquids within the frequency range utilized by PAT and other high-frequency techniques. Additionally, it extends the frequency range for which in vitro attenuation of porcine tissue has been investigated to 45 MHz. The presented method uses laser-generated ultrasound [19], [20] and unfocused, broadband, high-frequency piezoelectric transducers. Laser ultrasound has long been used as a contactless investigation method for material testing and characterization up to frequencies of several hundred megahertz. Oraevsky et al. used laser-generated ultrasound to measure optical properties of the tissue [21], however, there is no study known to the authors which uses laser ultrasound for measuring acoustic attenuation of tissue.
© 2012 IEEE
2632
IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control ,
vol. 59, no. 12,
December
2012
Section II provides an overview of the measurement system and a description of the piezoelectric transducers used and the application of laser ultrasound. The section ends with a detailed explanation of the setup for tissue and liquid measurements. The samples investigated and used for testing the system are described in Section III. Section IV specifies the signal processing, the signal loss model, and the determination of the parameters describing the ultrasound attenuation of the samples. Section V presents results for subcutaneous fat of pig, human blood, and olive oil. The paper concludes by discussing the method and by comparing the results of the measurements with previously published studies. II. Measurement System A. Overview Ultrasonic attenuation was determined by comparing transmission measurements (see Fig. 1) of investigated samples and distilled water. The unfocused ultrasound beam was generated either by piezoelectric transducers or by laser ultrasonic excitation. Both generation methods are described in detail and compared in the three following subsections. After traveling through the sample/water path, the ultrasound waves were detected by a piezoelectric transducer. For each sample and configuration of ultrasound generation, a reference measurement was performed in water with the same emitter–detector geometry as that used for the sample measurements. All experiments were done at room temperature (22°C to 24°C) in an air-conditioned room. B. Generation With Piezoelectric Transducer Unfocused piezoelectric immersion transducers (V358SU, Panametrics, Waltham, MA) with a beam diameter of 6.35 mm, center frequency of 50 MHz, and 40 MHz fullwidth at half-maximum (FWHM) were used to emit and detect ultrasound. To maximize the received signals, the transducers must be aligned very accurately. Fig. 2 shows the sensitivity of the signal strength to deviations of coaxial alignment for the range of emitter–detector distances used in this study. Deviation of parallel alignment leads also to a decreased bandwidth, as Fig. 3 shows. The behavior shown also occurs for emitter–detector distances of 27 and 40.25 mm.
Fig. 1. Measurement principle: an unfocused ultrasound beam propagates through the sample. The detected signal is compared with a reference measurement in distilled water.
Fig. 2. The root-mean-square ultrasound signal generated and received by piezoelectric transducers is very sensitive to (top) out-of-axis displacement and (bottom) deviations of parallel alignment of the transducers.
The transducers were controlled by a pulser/receiver (5073PR-40-E, Olympus NDT Inc., Waltham, MA; max. gain 39 dB, min. gain −49 dB). The output of the amplifier is linear up to 1 V peak when terminated by 50 Ω. The receiver gain is set to yield a voltage peak well below that limit. Two settings of the energy parameter of the pulser were used for the measurements. They are denoted E1 and E2 throughout the paper. The energy of the signals generated with the E2 setting is 2.25 times higher than for the E1 setting. C. Generation With Laser Ultrasound The second ultrasound excitation method we used is based on a laser ultrasonic technique [19], [20]. In this case, a short laser pulse illuminates an opaque target. Depending on the pulse duration and energy, this leads to thermoelastic or ablative generation of ultrasound waves. In the first case, the absorbed energy leads to an abrupt temperature rise at the surface of the target and, hence, to thermoelastic expansion of the illuminated volume and the emission of ultrasonic waves. If the absorbed light power is high enough, the surface is vaporized and the
Fig. 3. Spectra of ultrasound pulses generated and detected by the piezoelectric transducers for small deviations from parallel alignment and an emitter–detector distance of 9.65 mm.
bauer-marschallinger et al.: broadband high-frequency measurement of ultrasonic attenuation
resulting rebound of the expanding plasma causes the emission of ultrasound waves within the target. This is called ablative generation. The laser pulses were generated by a frequency-doubled Nd:YAG laser at wavelength of 532 nm, with a pulse duration of 6 ns, a repetition rate of 20 Hz, and pulse energies between 12 and 65 mJ (5% standard deviation). The diameter of the laser beam was about 6 mm. Because ablative excitation may hamper the comparability of sample and references measurements because of changes of the optical absorption rate and geometry of the surface, the pulse energy was chosen to be below the limit for ablative laser-ultrasound generation. Two effects were checked to determine this energy limit of approximately 65 mJ. An immediate effect is a visible light flash emitted by the plasma. Moreover, the surface of the target is blackened if ablation occurs. The thermoelastically generated ultrasound signals show negligible variation (
