P.L.M.J. van Neer

Comparison of fundamental, second harmonic, and superharmonic imaging: a simulation study

In medical ultrasound, fundamental imaging (FI) uses the reflected echoes from the same spectral band as that of the emitted pulse. The transmission frequency determines the trade-off between penetration depth and spatial resolution. Tissue harmonic imaging (THI) employs the second harmonic of the emitted frequency band to construct images. Recently, superharmonic imaging (SHI) has been introduced, which uses the third to the fifth (super) harmonics. The harmonic level is determined by two competing phenomena: nonlinear propagation and frequency dependent attenuation. Thus, the transmission frequency yielding the optimal trade-off between the spatial resolution and the penetration depth differs for THI and SHI. This paper quantitatively compares the concepts of fundamental, second harmonic, and superharmonic echocardiography at their optimal transmission frequencies. Forward propagation is modeled using a 3D-KZK implementation and the iterative nonlinear contrast source (INCS) method. Backpropagation is assumed to be linear. Results show that the fundamental lateral beamwidth is the narrowest at focus, while the superharmonic one is narrower outside the focus. The lateral superharmonic roll-off exceeds the fundamental and second harmonic roll-off. Also, the axial resolution of SHI exceeds that of FI and THI. The far-field pulse-echo superharmonic pressure is lower than that of the fundamental and second harmonic. SHI appears suited for echocardiography and is expected to improve its image quality at the cost of a slight reduction in depth-of-field.

Dual-pulse frequency compounded superharmonic imaging

Tissue second-harmonic imaging is currently the default mode in commercial diagnostic ultrasound systems. A new modality, superharmonic imaging (SHI), combines the third through fifth harmonics originating from nonlinear wave propagation through tissue. SHI could further improve the resolution and quality of echographic images. The superharmonics have gaps between the harmonics because the transducer has a limited bandwidth of about 70% to 80%. This causes ghost reflection artifacts in the superharmonic echo image. In this work, a new dual-pulse frequency compounding (DPFC) method to eliminate these artifacts is introduced. In the DPFC SHI method, each trace is constructed by summing two firings with slightly different center frequencies. The feasibility of the method was established using a single-element transducer. Its acoustic field was modeled in KZK simulations and compared with the corresponding measurements obtained with a hydrophone apparatus. Subsequently, the method was implemented on and optimized for a setup consisting of an interleaved phased-array transducer (44 elements at 1 MHz and 44 elements at 3.7 MHz, optimized for echocardiography) and a programmable ultrasound system. DPFC SHI effectively suppresses the ghost reflection artifacts associated with imaging using multiple harmonics. Moreover, compared with the single-pulse third harmonic, DPFC SHI improved the axial resolution by 3.1 and 1.6 times at the -6-dB and -20-dB levels, respectively. Hence, DPFC offers the possibility of generating harmonic images of a higher quality at a cost of a moderate frame rate reduction.

Imaging beyond aliasing

A proper spatial sampling is critical for high quality imaging. If the sampling criterion is not met, artifacts appear in the image, generally referred to as grating lobes. For inspection efficiency the width of the field of view is becoming larger leading to an increase in the number of elements and therefore transducer complexity and cost. The development of volume scanning methods in the medical field poses its own problems. Here a matrix of piezoelements is used to scan a volume using electronic beam steering. The challenge is to connect 2500+ elements using <;256 channels. Most solutions use prebeamforming to reduce the data at a cost of image quality. Another option may be to reconstruct the non-aliased data from spatially aliased data. In this work a novel method to reconstruct nonaliased radio-frequency (RF) data from strongly spatially aliased RF data is investigated using simulations and experiments. The reconstruction method involves an iterative scheme using wave field extrapolation. No medium assumptions are made. It has the following steps: 1) A matrix containing zeros at the locations where signals need to be interpolated is created such that no aliasing occurs. 2) The dataset is inversely extrapolated to focus the wave energy. 3) A threshold is applied to the extrapolated data selecting such that aliasing artifacts are excluded. 4) The dataset is forward extrapolated such that the input data is obtained. Now the empty traces contain signal. 6) The original RF dataset is copied into the reconstructed dataset. 7) Steps 2 - 6 are performed iteratively using a progressively lower threshold. Aliased and non-aliased datasets were modeled based on point diffractors and reflectors of increasing width. The datasets were imaged using a wavenumber-frequency domain mapping. The error after reconstruction was 0.77%, 4.2% and 43%, for undersampling of a factor 2, 4 and 8, respectively. For point diffractors the reconstruction error was 0.32%, 3.3% and 7.2%, respectively. These results show the methods potential. It may also be used to reconstruct signals for dead array elements.

Membrane design of an all-optical ultrasound receiver

Ultrasound sensors such as piezoelectric transducers and CMUTs are successfully used for medical imaging. However, especially wiring of individual elements is difficult in the fabrication of small piezoelectric arrays, used in, e.g. the field of intravascular imaging. As an alternative, we designed a novel type of ultrasound receiver based on silicon-on-insulator technology. This receiver contains an optical microring resonator positioned on the acoustical membrane. The deformation of the membrane induces strain in the optical resonator resulting in an optical resonance shift that can be recorded.To determine whether this receiver is suitable as ultrasound sensor we designed three prototype elements and simulated their response. This paper presents the design and working principle of our ultrasound receiver and shows the modeling results of these elements. We found an optimum in the dimension of the element by varying the thickness with corresponding radius for a response at 1 MHz frequency using a finite element analyses. Furthermore we obtained a sensitivity of 3.4 microstrain/kPa when the response of a 80 μm element was modeled resulting in a minimum detection level of 590 Pa. The first acoustical simulations of a single element of this receiver array shows that it may be a suitable candidate for miniaturized non-electrical ultrasound receivers.

Stolt migration based iterative trace reconstruction

Proper spatial sampling is critical for high quality imaging. If the sampling criterion is not met, grating lobe artifacts appear in the image. In non-destructive testing applications efficiency is being increased by the steady enlargement of the field of view, resulting in more elements and thus a higher transducer complexity and cost. In the volume scanning methods used in medical applications, the challenge lies in connecting the 2500+ elements of a matrix array to <;256 channels. Usually pre-beamforming is used to reduce the data at the cost of image quality. An alternative is to reconstruct the non-aliased data from spatially aliased data. Last year we reported a reconstruction method based on wave field extrapolation, which performs best for a limited depth range. In this work the method is extended to cover the entire imaging range at once. Its performance is investigated using phased array data of a tissue phantom. To reconstruct the traces the technique uses an iterative scheme based on a fast wavenumber-frequency domain mapping (Stolt migration) and its inverse in combination with thresholding to exclude the aliasing artifacts in the imaging domain. A properly sampled dataset was recorded of a tissue mimicking phantom using a linear phased array transducer connected to a research scanner for full channel data capture. From this dataset, undersampled datasets were created by selecting a limited number of channels. The datasets were imaged using wavenumber-frequency domain mapping (Stolt migration). The reconstruction method significantly improved the image quality of the aliased datasets.

Feasibility study of photo thermal acoustic imaging of buried nanowires in Gate-All-Around (GAA) nanowire FETs

With the ongoing drive to integrate more functionality and processing power on the same semiconductor area, the device structures have become 3D. Such structures, like FinFETs and their successors GAA nanowire FETs, bring on new challenges to measure their geometry and material properties non-destructively at the nanometer scale. Photo Thermal Acoustic Imaging (PTAI) is a potential inspection method. Here, the sample is heated locally by a pulsed optical pump beam. This generates a propagating elastic wave in the sample, which is scattered by inclusions in the sample. Part of the scattered energy travels to the sample surface, where the surface deformations are sensed by an optical probe beam using interferometric detection. Thus, no physical contact for actuation or sensing is required. This work investigates the feasibility of PTAI for the detection and imaging of nanowires.

Iterative trace reconstruction of aliased radio-frequency data obtained using harmonic imaging: A feasibility study

It is critical to use a proper spatial sampling, otherwise images suffer from grating lobes. However, the cost of a medical ultrasound scanner is strongly related to the channel count of the receive electronics. This has led to channel reduction using multiplexing or in-probe pre-beamforming methods at the cost of image quality or frame rate. An alternative is to reduce the receive channel count and reconstruct the non-aliased data from spatially aliased data. Last year we reported on a wavenumber frequency domain mapping based iterative trace reconstruction method developed for fundamental imaging. However, harmonic imaging is often used in medical imaging to further improve the image quality. As the reconstruction method assumes linearity, it is not a-priori clear whether the reconstruction will work satisfactory in combination with harmonic imaging. Here, the feasibility of using the method for harmonic imaging is investigated using in-vivo linear array data. The reconstruction algorithm operates by iteratively focusing and defocusing of the data using an imaging algorithm and uses intermittent thresholding to suppress the aliasing artifacts in the imaging domain. Properly sampled plane wave transmission datasets were recorded of the right common carotid artery of a healthy volunteer using a linear array transducer attached to a research system. The reconstruction technique significantly improved the image quality of all aliased datasets for both the fundamental and second harmonic imaging modalities. In fact, the reconstruction quality was slightly better for the second harmonic imaging case.

Travelling standing waves: A feasibility study

Lately, there has been significant interest in the noninvasive manipulation of particles and liquids. The reported acoustic methods rely on either the acoustic radiation force or acoustic streaming. The latter can be used in developed flows to induce fluid velocities angled to the liquid flow direction. These methods often use standing wave fields induced locally through continuous compressional waves. This has a drawback for industrial applications, i.e. continuous flow reactors, the standing wave field is excited locally, whereas it should be excited along the entire pipe length. Solving this by using numerous actuators along the pipe length is impractical and cost prohibitive. In this work the feasibility is investigated of using guided waves to induce a standing wave field over a pipes radius, but traveling along a pipes length. Also, it is investigated whether this standing wave field significantly affects the flow velocity field and fluid mixing. The method relies on exciting Lamb waves in the wall of a liquid filled pipe, which partially refracted into the liquid. The frequency was chosen such that: 1) a radial resonance was excited in the liquid, and 2) the reflected waves in the liquid interfere constructively with the refracted wave energy. The experimental geometry consisted of a copper pipe with an attached transmitting piezo of such dimensions that L(0,1) waves were excited. A 2nd piezo was mounted to detect and optimize the guided wave intensity. The pressure in the liquid was measured using a hydrophone. The effect of the pressure waves on the mixing of salt injections was measured using conduction probes at the inlet and outlet of the pipe. The 7th and 8th order standing wave fields were measured in the liquid at 650 and 770 kHz with transmit efficiencies of 20 and 22 kPa/V, respectively. This demonstrates the feasibility of inducing radial standing wave fields traveling along a pipe. Also, the salt concentration curves were altered in shape and surface area, if the ultrasound was switched on. This suggests that the convective mixing was increased by the induced radial fluid velocity components.

First measurements on a novel type of optical micro-machined ultrasound transducer (OMUT)

Several types of ultrasound sensors have been developed and are used in the field of medical imaging. Conventional transducers are made of piezo-electric material and show good practical performance. However, when the piezo-electric elements need to be small (below 100 μm × 100 μm), these transducers face challenges in fabrication as well as the electrical impedance matching of the elements. As an alternative, we fabricated an optical micro-machined ultrasound transducer (OMUT). This sensor contains an optical micro-ring resonator, which is coupled to a photonic waveguide, and integrated onto an acoustical membrane. The OMUT is build with standard silicon-on-insulator (SOI) technology, allowing for easy fabrication. In this paper, we present the first measurement results of the sensor. Our prototype has a -6 dB bandwidth of 19% and a noise equivalent pressure (NEP) of 0.5 Pa. These first acoustical measurements show that this prototype may form the basis of future ultrasound transducers.

Ultrasound transmission spectroscopy: In-line sizing of nanoparticles

Nanoparticles are increasingly used in a number of applications, e.g. coatings or paints. To optimize nanoparticle production in-line quantitative measurements of their size distribution and concentration are needed. Ultrasound-based methods are especially suited for in-line particle sizing. These methods can be used for opaque dispersions and at high concentrations. However, using ultrasound to measure nanoparticles is challenging: despite the use of high frequencies the scattering is close to the Rayleigh regime (ka ≪ 1) and the information contained in the measurements is limited. In this work the performance of an ultrasonic particle sizing method is evaluated using SiO2 nanoparticles. The measurement method is based on ultrasound transmission spectroscopy. The presence of nanoparticles affects the propagation of ultrasound in the medium, which is measured over a frequency band of 50 - 250 MHz. The wave propagation effects are then interpreted using the inversion of a physics model. The investigated dispersions consisted of SiO2 nanoparticles (1.4 and 2.0 vol%) dispersed in water. Four batches, provided by Nano-H S.A.S., had monomodal size distributions with mean sizes 150, 300, 420 and 440 nm. Two bimodal size distributions were investigated: 1) a mix of 50% 302 nm and 50% 422 nm particles, and 2) a mix of 50% 150 nm and 50% 422 nm particles. As a reference the size distributions were measured using an optics based Malvern Zetasizer. The mean particle sizes and concentrations were similar to the reference, with differences between 4.5 and 19% and between 3 and 15%, respectively. The shape of the particle size distributions obtained by the ultrasonic instrument were similar to that of the reference. Also, the ultrasound instrument was able to produce correct results for both mono- and bimodal size distributions. The temperature of the mixture did not have a significant influence on the results.