If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
High-intensity focused ultrasound transducers operating at 4 MHz based on lead-free piezoceramics from the sodium bismuth titanate (NBT) family are described. First, the piezoelectric material (Pz12X) is evaluated from the standpoint of transducer design and its important characteristics, including temperature dependance of several parameters such as dielectric and mechanical coefficients. Then, the performance of six transducers of the same design is evaluated in terms of electro-acoustic efficiency and its dependency on the operating acoustic power level up to 30 W. Overall, the initial electro-acoustic efficiency of three independent transducers is approximately 50% at low acoustic power levels and slightly drops down to 42% as the input electric power reaches 10 W. This process is stable and fully reversible. Moreover, the stability of electro-acoustic efficiency over extended power burst cycling is studied using another two transducers up to 95 × 103 power bursts of 250-ms duration and acoustic power of 10 W. This protocol is beyond the typical clinical use of similar devices in practice. No significant changes in electro-acoustic performance are noted. Additionally, the input electric power and the output acoustic power, together with the temperature of the piezoelectric component, are evaluated simultaneously over the period of one power burst. It is found that the maximum operating temperature over a high-input electric power burst of 600 J is below 60°C, which defines the operational limit for such devices, as the de-poling temperature of the lead-free material is around 85°C. It is found that the lead-free material from the NBT family is also a promising alternative to lead-based PZT-type materials in high-power therapeutic ultrasound.
) has become a more acceptable and widespread non-invasive modality for treatment of a variety of medical conditions. HIFU treatment of several clinical indications, such as uterine fibroids, essential tremor and brain tumors (
) has already gained approval in both the European Union and the United States, as well as other countries. In fact, the Focused Ultrasound Foundation has listed more than 140 indications that can be potentially treated with HIFU, with many being added to the list every year (
). The harsh operating conditions of HIFU devices constitute the main reason that HIFU manufacturers and the research community typically build their ultrasonic devices using the well-established system of piezoelectric materials composed of lead zirconate titanate (PZT) having a lead content >50 wt%.
Since the first European Restriction of Hazardous Substances (ROHS) directive became effective in 2006 (EU Regulation 2002/95/EC), there has been a general incentive to replace all lead-containing components with non-regulated alternatives. This has pushed through significant changes in the electronics industry, in particular for companies marketing their Conformité Européenne (CE)–marked products in the European Union and its associated countries. The PZT-based elements responsible for generation of ultrasound signals in both diagnostic and therapeutic ultrasound devices have, however, so far been exempted from the ROHS restrictions under various rules and limitations. These exemptions have been driven primarily by arguments that proven alternatives do not exist. With the subsequent updates of the ROHS rules in 2013 and 2019, more substances are becoming restricted, and exemptions are gradually removed or reduced. It is therefore very likely that restrictions with significant impact on both diagnostic and therapeutic ultrasound products could be implemented in the future as advances in material science gradually offer realistic lead-free options to PZT.
The state-of-the-art lead-free piezoelectrics represent different families of materials, including barium titanate (BaTiO3 or BT), potassium sodium niobate ([KxNa1 –x]NbO3 or KNN), lithium sodium niobate ([Li1 –xNax]NbO3 or LNN), sodium bismuth titanate ([Na0.5Bi0.5]TiO3 or NBT), and bismuth-based layered perovskite structures such as Bi4Ti3O12 (BiT) and SrBi4Ti4O15 (SBT). Most of these materials have been intensively investigated over the last 20 years, with promising results, but without gaining wider industrial acceptance (
). Some of the materials that are comparable in performance to PZT-based materials require very advanced manufacturing techniques and, therefore, are hardly suitable for large-scale production, are expensive and have low reproducibility. The most promising lead-free piezoelectric materials to date were considered to be a group of alkali niobates, in particular the potassium sodium niobate family materials (KNN family) materials, that exhibit relatively high piezoelectric performance (
). However, most alkali niobate materials require careful control of the composition and advanced powder preparation techniques. Reliability and reproducibility are also known issues for alkali niobate materials that need to be specifically addressed (
). All these factors prevent the wide commercial use of KNN-family materials. As an alternative to the alkali niobates, Bi-based materials such as sodium bismuth titanate Na0.5Bi0.5TiO3 (NBT) and bismuth potassium titanate Bi0.5K0.5TiO3 (BKT) have recently attracted wide interest as a possible alternative to lead-containing piezoceramics (
). Compared with KNN or PZT, the NBT-based materials have lower piezoelectric properties and relatively low depolarization temperature; however, they exhibit a relaxor-type behavior, meaning high strain/displacement at high fields, high coercive field (Ec) and relatively high mechanical quality factor Qm (compared with KNN). Such materials are therefore potentially attractive, for example, for actuator applications, where high displacement is needed, or in HIFU applications, where high driving voltages are used (
). Furthermore, the relative simplicity of manufacturing of such ceramics (in comparison with, e.g., the KNN family) makes them attractive for industrial applications, thus making them one of a few lead-free piezoelectric alternatives commercially available at the moment.
The main purpose of this work was to assess the viability of a new commercially available lead-free alternative based on the NBT family by evaluating its short- and long-term performance in HIFU transducers. This is accomplished regardless of how they compare with PZT-based devices, with the paramount question being: Can Pz12X devices be used in clinical conditions for extended periods?
The performance of the transducers is assessed on the basis of electro-acoustic efficiency studied at different power levels, including clinical levels. The tested transducers operating at 4 MHz are therefore not designed for a particular purpose or medical indication. However, very similar devices operating from 3 to 10 MHz can be applied for subcutaneous treatments in cosmetology or treatment of shallow organs such as lymph nodes (
). Use of the same overall design concept, but with a piezoelectric element designed for operating at frequencies of 10 to 25 MHz, can be directly applicable to a very wide range of dermatological indications including actinic keratosis and basal cell carcinoma (
The main section of the article starts with a detailed description of materials and methods used in regrading properties of Pz12X, transducer design and fabrication, as well as measurements of electro-acoustic efficiency under different power conditions. Results of various tests are provided afterward, followed by a thorough discussion emphasizing several aspects of using BNT-family piezoelectric material for HIFU application. The article ends with conclusions and the outlook for future work.
Characterization of Pz12X
Dielectric and piezoelectric characterizations of the Pz12X material samples were performed at room temperature, using an LCR meter (Model 4310, Wayne Kerr Electronics, Bognor Regis, UK) and a piezometer (PM300, Piezotest Pte. Ltd., Singapore). The coupling coefficients, frequency constants and mechanical quality factors were determined using an impedance analyzer (E4990A, Agilent, Palo Alto, CA, USA) according to EN 50324-2. The same impedance analyzer together with an environmental test chamber (Model 9028, Delta Design, San Diego, CA, USA) was used for impedance measurements of coupling coefficients and quality factors as a function of temperature up to 150°C.
Additionally, dielectric characterization of the samples in a wide temperature range (25°C–600°C) was performed using a high-temperature probe (ProboStat, NorECs, Oslo, Norway) equipped with a high-temperature sample holder and a conventional tube furnace together with the LCR meter Model 4310. To avoid any unnecessary charge accumulation in the samples caused by the pyroelectric effect, the samples were electrically shorted between the consecutive measurements during heating and cooling. Ferroelectric properties of the samples have been evaluated in a wide temperature range (20°C–200°C) using a ferroelectric test system (TF Analyzer 1000, aixACCT GmbH, Aachen, Germany). Statistics as well as a two-sample t-test of selected parameters of Pz12X components were calculated using Minitab 18 (Minitab LLC, State College, PA, USA).
A batch of six HIFU transducers operating at around 4 MHz was fabricated. The devices were housed in 3-D-printed structures as depicted in Figure 1A, with the front of the device radiating toward the focal point and with an air acoustic backing. The core of each device was a lead-free piezoelectric component made from Pz12X (CTS Ferroperm Piezoceramics, Kvistgaard, Denmark) in the shape of a section of a hollow sphere with a diameter of 20 mm and a radius of curvature (focusing radius) of 15 mm (Fig. 1B). The electrical matching circuit was attached to the top of the device, minimizing the length of the electrical connections. The transducers were electrically matched to 50 Ω at parallel fundamental resonance. The fabricated transducers were used in a few different configurations/experiments, and a full list of the devices is given in Table 1.
To examine in detail the performance of a Pz12X focusing component under realistic conditions, a special transducer was built with the addition of a contactless infrared (IR) temperature sensor “looking” directly on the surface of the piezoelectric component, as depicted in Figure 2A. A contactless sensing was selected to eliminate the impact of the radiofrequency (RF) signal on temperature sensing. For the same reason, fully shielded wiring was applied as illustrated in Figure 2B. An integrated IR temperature sensor (MLX90614, Melexis NV, Ieper, Belgium) programmed in pulse width modulation (PWM) mode was used to enable direct sampling of the output signal.
Measurements of output acoustic power
The transducers were evaluated using measurements of electro-acoustic efficiency, which is an excellent indicator of the overall performance of the devices, giving direct insight into conversion efficiency between the electric and acoustic power. The measurements of acoustic power were performed using a radiation force balance (RFB, Precision Acoustics, Dorchester, UK) in a suspended target configuration, depicted in Figure 3A, with de-gassed (oxygen level < 2 mg/L) and de-ionized water at 23°C.
The radiation force balance returns a direct reading of the vertical component of acoustic pressure ,
where m is the mass reading from the microbalance, g = 9.81 m/s2 is the gravitational acceleration, c = 1490 m/s is the speed of sound in water and ξ = 0.98 is the target calibration factor. To estimate the total acoustic power, one needs to account for propagation losses and focusing through inclusion of the propagation factor βp and focusing factor βf, respectively. The total output acoustic power Pa is given by
α = 2.3 × 10–4 MHz2/cm is the attenuation factor in water, f is the operating frequency (in MHz) and z (here in cm) is the propagation distance to compensate.
The focusing factor, bf, is given by
where is the transducer focus half-angle given by , a = 10 mm being the aperture of the transducer of focal radius F = 15 mm (radius of curvature). To protect the target from damage caused by high acoustic intensity, the measurements were performed in so-called post-focal fashion; that is, the target was located below the focal point. In the results presented, the distance from the focal point to the target was kept equal to 8 mm, also defining the propagation compensation distance as per eqn (3). Therefore, the measured Pa represented acoustic power at the focal plane.
Finally, the electro-acoustic efficiency, ηea, is defined as the ratio of total acoustic power Pa to input electric power Pe as
The devices were driven by a function generator (33600A series, Keysight Technologies, Santa Rosa, CA, USA) through an inhouse-built class AB broadband 90-W, 50-dB-gain RF amplifier at their fundamental frequency. The input electric power Pe was measured using a –30-dB directional coupler (DDS-1 Stenbock Enterprise LLC, Aurora, OR, USA). The output signal from the directional coupler was digitized and analyzed using an oscilloscope (DSOX3024T, Keysight Technologies), finally producing the values of forward power, Pef (total output power), as well as the reflected, Per, electric power (resulting from impedance mismatch). Finally, the input electric power, Pe, delivered to the transducers was calculated using the formula
To avoid overheating of transducers, the measurements of electro-acoustic efficiency were performed using a limited power duty cycle in a broad range of power levels (from 0.5 to ∼10 W). At each power level, five bursts of output acoustic power were acquired using the RFB software (RFB2, Precision Acoustics, UK) as well as values for forward and reflected electric power. The data were analyzed afterward using in-house-developed GNU Octave scripts, analyzing and averaging data through every power burst and combining results at each power level, producing effectively final electro-acoustic curves. An example of such analysis at a selected power level (representing one point on the electro-acoustic conversion curve) is illustrated in Figure 3B. In the presented results, the power-on time was equal to 3 s with a repetition interval of 15 s, resulting in a power duty cycle of 20%. This power cycling regime was found sufficient to preserve the integrity of transducers having the design as per Figure 1A, simultaneously giving enough time to stabilize the readings from the microbalance otherwise known for a significant inertia when used for dynamic measurements.
Power burst cycling
Power burst cycling was performed to examine the stability of the output of Pz12X components in an HIFU application under real clinical conditions, where the transducers normally operate at an extended number of cycles delivering energy to a tissue. Two scenarios of operation were examined: delivery of 30 × 103 bursts of 1.25 J of acoustic energy each (Pa = 5 W, burst time tb = 250 ms), which is considered a typical scenario in dermatology or cosmetology applications, and delivery of twice the dosage at 2.50 J in more than three times the number (95 × 103) of bursts. In both cases, a new device was used (TranD and TranE, respectively) to eliminate the impact of usage history on the results. The bursts were repeated with a 2-s period, which represented the worst-case scenario for devices used in continuous mode, where the applied part is allowed to be in contact with the patient for less than 10 min as per IEC 60601-1 (
International Electrotechnical Commission (IEC) 60601-1-11: Medical electrical equipment— Part 1-11: General requirements for basic safety and essential performance—Collateral Standard: Requirements for medical electrical equipment and medical electrical systems used in the home healthcare environment.
) (contact temperature should be <42°C). The power burst cycling was interrupted at a pre-defined number of accumulated cycles to perform the measurement of electro-acoustic efficiency. Because of the pre-defined cycle period and number of cycles, the measurements were performed over the course of several days for each transducer, with a repeated process of water degassing at the beginning of each day.
The electro-acoustic efficiency measurement as described in the previous section was a semimanual process taking a few hours per single electro-acoustic efficiency curve. Therefore, a fully automated setup was employed in the power burst cycling test, where the power cycling was combined with measurement of electro-acoustic efficiency at a pre-defined number of generated cycles. The setup comprised the same hardware as described in the preceding section. The control was, however, performed by a Python 3.6.6 script integrating electric power measurement, acoustic power measurement and control of the power cycling. This enabled an integrated measurement of electro-acoustic efficiency and power burst cycling in a single procedure. The acoustic power was adjusted according to eqns (2)–(4) in a similar fashion as in the measurements of individual electro-acoustic efficiency curves.
Single-burst performance with temperature measurements
Single-pulse performance was determined using a specially adapted transducer with a built-in IR sensor. The output PWM signal of the sensor was directly sampled using an oscilloscope (DSOX3024T, Keysight Technologies, Santa Rosa, CA, USA). This enabled simultaneous measurement of input electric power and output acoustic power (using RFB), as well as surface temperature of the device, during each single burst of power. Because of the complexity of the measurement as well as the use of general-purpose instrumentation, the overall sampling period of all input signals was limited to some 10 ms, limiting the use of the setup at shorter power burst times (on the order of hundreds of milliseconds), however making it suitable for analysis of bursts longer than 3 s.
The unknown surface emissivity necessary for correct IR measurement of temperature was estimated in a separate experiment using a sample of Pz12X component attached to a hot plate using thermal paste, with simultaneous measurement of the surface temperature using MLX90614 and a reference thermocouple-J (Keysight Technologies, Santa Rosa, CA, USA) connected to an acquisition module (34972A, Keysight Technologies). The compensation was performed in the temperature range 23°C to 110°C.
Even though the single-burst performance tests were carried out at very high energy levels per burst (up to ∼600 J) the intention was not to go beyond the permanent damage level. For this reason, the transducer was cooled to room temperature at the end of every burst.
Lesioning in tissue-mimicking gel
To illustrate the ultimate clinical capabilities of the fabricated transducers, TranA was used to create thermal lesions in egg white–containing tissue-mimicking phantoms at different energy levels per burst. The tissue-mimicking phantom gel was prepared according to the recipe described in
in the form of rectangular cuboids fitting the test setup holder. The transducer TranA was additionally equipped with a 3-D-printed clip-on adaptor forming a coupling liquid chamber (see Fig. 4A), effectively turning the open-end transducer into a handpiece that could be used outside of a water tank in the test setup, as depicted in Figure 4B. The coupling liquid (de-ionized water) was held inside the chamber using a 10-μm-thick low-density polyethylene (LDPE) film. The distance between the end of the clip-on chamber and the focal point is called the nominal penetration depth (
) as it defines the level at which the ultrasonic energy will be focused with reference to the surface. In the results presented, the nominal penetration distance was equal to 4 mm.
Similar to the previously described experiments, the device was driven by a function generator (33600A series, Keysight Technologies) through the in-house-built 90-W RF amplifier at the fundamental frequency. The duration of each burst was pre-determined by the exact number of cycles pre-programmed on the function generator; each burst was triggered manually.
Properties of Pz12X material and components
Selected nominal values of material parameters of Pz12X are given in Table 2 together with properties of industry standard lead-based materials Pz26 and Pz27 (CTS Ferroperm Piezoceramics, Kvistgaard, Denmark) representing typical values for so-called hard- and soft-doped materials, for comparison. The selected values are relevant for design and operation of transducers, especially those operating at high power levels. The nominal values are given at room temperature, measured at 1 kHz.
Table 2Nominal values of selected material parameters of Pz12X and industry standard lead-based materials Pz26 and Pz27 at room temperature
Figure 5A and 5B illustrate dielectric properties of Pz12X in a very broad temperature range. The most important result is the de-poling temperature of Pz12X, Td = 85.2°C, at which temperature the material fully loses its piezoelectric properties, and the transducer effectively becomes permanently non-functional. This also changes the value of the dielectric constant of the material at room temperature. The de-poling temperature has a profound impact on how Pz12X-based devices should be operated and what applications they can address. It appears that Pz12X does not exhibit a sharp maximum of dielectric permittivity because it is a relaxor-type piezoelectric material, exhibiting typical order-disorder phase transition (
). It can also be noted that in the operating temperature range (from 20°C to 85°C), the dielectric permittivity is changing significantly, which is a challenge for appropriate impedance matching of transducers.
The temperature-induced de-poling process of Pz12X is also well illustrated in Figure 6A, which depicts both coupling coefficients (planar and thickness modes) as a function of temperature. However, both coefficients are relatively stable up to around 80°C. Mechanical quality factors furthermore drop noticeably with rising temperature, indicating increasing mechanical losses caused by temperature.
Another important material parameter is the coercive field, Ec, defining the safety margin of operation for the driving electric voltages. As indicated in Figure 6B, the coercive field of Pz12X is also strongly dependent on temperature. However, up to the de-poling temperature, Td, the coercive field remains at a relatively high level above 1.5 kV/mm, noticeably higher than that of PZT materials at room temperature (
Even though the results for Pz12X are preliminary, one could attempt to assess the batch and batch-to-batch variability, which has been one of the major issues with the introduction of lead-free materials to the market (
). Two separate batches of Pz12X focusing components were manufactured. Basic statistics of the selected component properties that are important in view of transducer manufacturing are given in Table 3. The variability within the batches can be evaluated using standard deviation values, while batch-to-batch variability is better reflected by a two-sample t-test for means.
Table 3Basic statistics of selected parameters at component level of two production batches
p Value of two-sample t-test for means
Sample size, n
Capacitance, C, pF
4004.5 ± 84.9
4130.0 ± 121.1
Dielectric loss, tan δ
3.29 ± 0.19
3.35 ± 0.16
Piezo coefficient, d33, pC/N
–146.7 ± 10.9
–150.1 ± 5.8
Serial resonance frequency, fs, kHz
3879 ± 18
3899 ± 13
Parallel resonance frequency, fp, kHz
4270 ± 45
4267 ± 34
Thickness coupling factor, kt
45.42 ± 1.65
44.25 ± 1.55
p Values in boldface represent statistically significant differences at the confidence level of p < 0.05.
The performance tests started with assessment of the electro-acoustic efficiency of three different transducers (TranA to TranC) in a range of forward electric power from 0.5 W to approximately 10 W, conducted in three consecutive runs. The test started with a so-called virgin run; that is, the transducer was not exposed to any high-power drive beforehand. This type of test quickly indicates the short-term performance as it can reveal some irreversible changes in the transducers caused by exposure to high power during measurements of electro-acoustic efficiency. Figure 7A and 7B illustrate the power dependence of reflected power, as well as the power dependence of electro-acoustic efficiency of a selected transducer (TranB), over the course of three consecutive test runs. The data can be interpreted as input (Fig. 7A) as well as input/output (Fig. 7B) characteristics of the transducer. The TranB data illustrated in Figure 7A and 7B were very similar to those registered for TranA and TranC. Already at this initial testing stage, the measured maximal output acoustic power was in the range 4 to 5 W, which was at a level required in typical clinical applications. Summary curves for all tested transducers are given in Figure 8 together with error bars indicating the 1 × standard deviation calculated from three consecutive runs.
The power cycling can be used as an indicator of the long-term stability of Pz12X-based HIFU devices. Two transducers (TranD and TranE) were used in the tests. TranD tests represent a very typical clinical scenario of usage, in which around 30 × 103 bursts of acoustic energy of a maximum of 1.25 J is expected throughout the lifetime of a medical transducer. This number can vary for different manufacturers but is a good starting point when considering a new device design or/and piezoelectric material. TranE was exposed to twice the energy per burst, namely, 2.5 J, and approximately triple the number of bursts (>90 × 103). Before and after power cycling input curves of TranD and TranE are provided in Figure 9A. Results of electro-acoustic efficiency measurements at low and high power (∼1 and ∼15 W, respectively) are given in Figure 9B as a function of accumulated number of bursts.
The data in Figure 10A–F confirm the hypothesis of favorable operating conditions, even at far more severe driving power and energy levels than needed for relevant clinical use. This includes forward electric power up to 30 W and burst times up to 20 s, where the maximum registered temperature around the apex of the piezoelectric component was not higher than approximately 60°C. This temperature corresponds well to the recommended operating temperature for Pz12X given in Table 2, as well as other results given in Figures 5 and 6.
Figure 11 illustrates thermal lesions produced by the TranA device with the clip-on coupling medium chamber at four different energies. The lesions were produced in order from right to left, with approximately 3-s intervals between bursts. Figure 11 thereby indicates that the Pz12X transducer can produce repeatable thermal lesions in a tissue-mimicking phantom; moreover, the lesions are reproducible at each tested energy level to the extent that no loss of performance is noticeable within the tested number of lesions per energy level (up to 6).
Pz12X in HIFU application
Considering the ongoing regulatory activities in the EU to eliminate lead-containing products through gradual additions and tightening of the regulation, as well as the necessity of becoming more environmentally friendly, it becomes increasingly important to better understand how material performance and transducer behavior can be optimized when using the new generation of lead-free alterative piezoelectric materials. The distinction between “hard” and “soft” piezoelectric materials is, first, somehow less obvious in the case of lead-free materials as lead-free materials exhibit a mixture of properties of those two categories, thus producing a new situation for transducer designers/manufacturers. All old design rules might therefore no longer be applicable in the case of lead-free piezoelectrics.
This problem is illustrated for Pz12X, a lead-free material from the NBT family. Table 1 summarizes its nominal properties in comparison with well-known hard- and soft-doped lead-based (PZT) materials. It is clear that Pz12X exhibits characteristics that do not follow any of these two conventional classifications, which in turn makes its application for specific transducer types less obvious.
When it comes to the operating temperature of piezoelectric-based devices, it is usually (in the case of PZT) the Curie temperature, TC, that is of concern. The thermal de-poling of Pz12X occurs significantly before the dielectric permittivity reaches its maximum as depicted in Figure 5A. The low de-poling temperature of Pz12X, Td = 85°C, is one of the values having the biggest impact on how the devices should be operated, as it is significantly lower compared with that of PZT. Another important issue is the relatively high dielectric and elastic loss, which is more than 10-fold higher in the case of elastic loss and 4-fold higher in dielectric loss in comparison to hard-doped PZT. In general, three categories of losses in piezoelectric materials should be considered: elastic, dielectric and piezoelectric. As argued in
, elastic losses are by far more prevalent compared with dielectric losses in the case of devices operating at resonance, such as HIFU transducers. This already places Pz12X somehow between the “soft” and “hard” categories, enabling lead-free piezoceramics to enter HIFU applications. In this case, the figure-of-merit (
). Pz12X is therefore, in some respects, more an analog of lead-free “hard” piezoelectric and is therefore a reasonable choice for HIFU applications, even if some of its properties do not place it in the “hard” or “soft” category.
Table 3 summarizes basic statistics of two analyzed batches of ready-made components. The standard deviation helps in understanding the variability within each batch. Most of the values are within 10% of the relative standard deviation (SD/mean), which is a good indicator as the industry standard is sometimes up to ±20%. The critical p values help to assess the statistical difference between the means of two analyzed batches (batch-to-batch variability). Capacitance and serial resonance frequency, as well as thickness coupling, exhibit statistical differences with p values < 0.05; however, as the relative differences between the means are less than 10%, it can be concluded that in practice there is good batch-to-batch reproducibility with industry standards set at 20%. Differences in capacitance of the two batches might have at least two origins: differences in dielectric permittivity of the raw piezoelectric material and/or differences in geometry (thickness). Statistically significant differences in the serial resonance frequency suggest that it is more likely the second option, as lower capacitance of batch A would translate into thicker component, that is, lower serial resonance frequency. Lack of statistically significant differences in dielectric loss and the piezoelectric coefficient, d33, which are purely material-related properties (independent of geometry), suggests that production of the raw material is stable with acceptable repeatability.
The initial electro-acoustic efficiency curves of three analyzed devices exhibit behavior typical of such devices with a quasi-linear decrease in the efficiency with rising input power. This is attributed mainly to increased elastic losses in the piezoelectric component at elevated temperature caused by high-power operation and electric losses in the matching circuit. The curves are very similar for every test run, indicating that any change in parameters occurs mainly as a result of temperature increase, and the changes are reversible. This suggests that driving conditions during electro-acoustic efficiency measurement are such that the de-poling temperature of Pz12X is not reached. The analysis of the acoustic power curves (an example given in Fig. 3B) suggests that during the power bursts, the acoustic power remains relatively constant. Both Figures 7 and 8 indicate very repeatable behavior of the transducers. Figure 7A illustrates a linear increase in the ratio of reflected to forward power as the input power increases, reaching almost 10% at approximately 10 W of forward electric power. This suggests that power-induced changes occur most likely in the impedance matching circuit. These potential changes are, however, reversible and can be compensated by implementing a calibration curve dependent on input power. The drop in electro-acoustic efficiency as depicted in Figures 7B and 8 is due mostly to losses in the piezoelectric component, as the efficiency is plotted against the effective input electric power, eqn (6), adjusted for the electric mismatch.
Results of stability tests of electric-acoustic efficiency over an extended number of cycles suggest that Pz12X is a real alternative to hard PZT for certain classes of HIFU devices. Both tests at clinical dosage, as well as extended clinical dosage, are promising, despite the relatively low de-poling temperature of Pz12X and suboptimal design of the transducer itself (e.g., no special care for thermal management was given). This can also be attributed to favorable operating conditions, such as limited duty cycle, limited energy of single burst (<2.5 J) and, therefore, limited operating temperature of the piezoelectric component. Similar devices operate in clinical practice, where the energy limit is defined by the maximum clinical effect and maximum contact temperature of the applied part with the patient (<42°C) given by the safety standard IEC 60601-1 (
International Electrotechnical Commission (IEC) 60601-1-11: Medical electrical equipment— Part 1-11: General requirements for basic safety and essential performance—Collateral Standard: Requirements for medical electrical equipment and medical electrical systems used in the home healthcare environment.
Data illustrated in Figure 9A indicate the input characteristics have good stability, even for the transducer exposed to a high dose of energy and larger number of power cycles. The curve at the end of the test with N = 95 × 103 for TranE is slightly above the initial curve; however, this change is acceptable from the application point of view. The two curves for TranD overlap, indicating no change in reflected power caused by power cycling in the typical clinical usage scenario. Similarly, electro-acoustic efficiency at low (∼1 W) and high (∼15 W) power indicates no change (within the measurement error) in the range of power cycles tested. Even though the TranD exhibits overall better performance (higher ηea than TranE), the performance remains unchanged with the accumulated number of power cycles, suggesting that the operating conditions of Pz12X components, and in particular operational temperature, permit the use of Pz12X in the application described.
Combined temperature and acoustic power measurements provide deeper insight into the performance of the HIFU devices studied. Data illustrated in Figure 10A, and to some extent in Figure 10D, support the hypothesis that an increase in electric mismatch (increase in reflected power) at lower power levels (here 10 W) is due mainly to temperature increase of the electric components in the impedance matching circuit. This is supported by the fact that the temperature increase of the piezoelectric component is below 5°C–6°C and, therefore, has very limited impact on the piezoelectric coupling, elastic and dielectric losses, as well as the effective capacitance change of the component. At higher power levels (Fig. 10B, 10C, 10E, 10F), the acoustic output decreases with time and increased temperature of the piezoelectric component. Here the mechanism is somehow more complex, as several thermal and non-thermal processes take place. Of course, the losses in the electric matching circuit remain. The acoustic output power decreases at the same time because a significant thermally induced change in dielectric permittivity causes de-tuning, that is, change in the resonance frequency of the device. As the driving frequency remains unchanged, the overall efficiency drops. Additionally, as the temperature increases, the elastic losses in the piezoelectric material increase (see Fig. 6A), causing a further drop in electro-acoustic efficiency and producing extra heat in the component. This mechanism, if not controlled through the limited duty cycle, can ultimately lead to a runaway situation, where the temperature increases beyond the de-poling level and consequently leads to destruction of the device. Of course, this situation should be avoided in practice. Data illustrated in Figure 10F indicate an approximate boundary for operation of the studied device, during which the power burst attains the maximum recommended operation temperature of approximately 60°C.
Another noticeable effect in Figure 10C and 10F is the “dip” in temperature level just after the power has been switched off. This occurs because the streaming effect results in forced cooling of the piezoelectric component. Further analysis of this effect is, however, beyond the scope of this article.
The lesioning experiment gives the ultimate overview of the capabilities of the lead-free devices discussed, as it represents the expected clinical effect on the human body. Similar, tadpole-like and egg-like shaped lesions produced by PZT-based HIFU devices have been reported (
). Correspondence of the depth at which the lesions are formed to a 4-mm penetration depth of the clip-on coupling medium chamber is well controlled. This confirms that lead-free HIFU transducers can produce predictable and repeatable lesions and therefore can be used in clinical settings.
Conclusions and Outlook
This work represents the first characterization of a new lead-free material, Pz12X, used in a transducer relevant for medical HIFU therapy. Incorporating such lead-free piezoelectric materials into both diagnostic and therapeutic ultrasound transducers is becoming increasingly relevant as the restriction of lead-containing components is expected to be tightened further in coming iterations of the European ROHS legislation.
Pz12X is concluded to be difficult to classify in terms of established categories from the traditional PZT system, in which primarily soft materials are selected for diagnostic transducers while hard materials are used for therapeutic devices. Pz12X has a low de-poling temperature, high dielectric losses and large variation in permittivity with temperature, which match the characteristics of soft PZT materials, while, at the same time, a very high coercive field makes it comparable to hard PZT. A limited duty cycle, as well as limited energy levels, of certain classes of HIFU devices are thereby the potential enablers for use of Pz12X in such devices. The devices developed in this work were concluded to be able to operate at relevant clinical levels of acoustic power as well as duty cycles, while maintaining temperatures in the piezoelectric components well below their relatively low de-poling temperature.
Data from single-burst performance tests and extended power cycling furthermore indicate that the devices maintain their electro-acoustic efficiency without signs of deterioration far beyond clinically relevant levels in terms of both power and required stability after cycling. The recommendation published in
that a piezoelectric device should survive at least 109 power cycles was directly fulfilled by the test devices, as 95 × 103 power bursts used in the test of TranE (td = 250 ms, at f = 4.378 MHz) is equivalent to 95 × 103 × 1.0945 × 106 = 10.4 × 1011. To obtain a higher level of confidence on this critical issue, these data should be complemented by further extensive long-term cycling and stability testing under various controlled conditions.
The overall results are concluded to be very positive and to form a very promising basis for work on substitution of lead-containing piezoceramics, such as Pz12X, in HIFU devices and other devices operating with limited duty cycles and power levels.
It should be noted that the results and analyses presented here are semiquantitative because of the limited scope and initial nature of the work. The work can, however, be further deepened by inclusion of a fully quantitative analysis of devices based on, for example, a lumped circuit model for better understanding of loss mechanisms in the device, which are concluded to be of increased importance in the case of lead-free devices compared with those made with PZT. Further work can also be supplemented by detailed numerical analysis (e.g., using the finite-element method) of the thermal behavior of the device. Finally, the output acoustic field can be studied further experimentally and numerically in direct comparisons with comparative PZT-based devices.
Conflict of interest disclosure
The research described was funded by TOOsonix A/S and CTS Ferroperm Piezoceramics A/S, each supporting their respective teams. No external funding was received.
Sintering behavior of lead-free (K,Na)NbO3-basedpiezoelectric ceramics.
60601-1-11: Medical electrical equipment— Part 1-11: General requirements for basic safety and essential performance—Collateral Standard: Requirements for medical electrical equipment and medical electrical systems used in the home healthcare environment.