To evaluate the resolution of the CMUT, small objects such as screws (8 mm in length, 2.9 mm in diameter) and polycarbonate cubes (4 mm sides) were used as test samples. For the screw test, the object was mounted on a 3D-printed square-shaped holder attached to an automated XYZ stage. The screw, along with its holder, was systematically moved along the z axis (bottom-to-top direction) in controlled increments of 2 mm while keeping the x and y coordinates constant. The initial position of the XYZ stage was (28, 70, 155) mm, and the movements continued until (28, 70, 202) mm. Throughout this process, the CMUT remained stationary, secured in a 3D-printed holder facing the test object. CMUT signals were collected at each position as the object moved incrementally. The test setup is shown in Fig. , with an arrow indicating the direction of motion and an enlarged image of the screw in front of the CMUT. In the second setup, polycarbonate cubes were used as test objects and kept stationary in a water-filled glass beaker. The CMUT, secured in a holder facing the objects, was mounted on an XYZ stage to allow precise and controlled movements during the experiment. The data were collected at each position above the objects, as well as once away from the test objects, where only the beaker was in the CMUT's focus, for comparative reference purposes. Figure  shows the test setup and a magnified image of the three transparent polycarbonate cubes.

To investigate how the resolution of the CMUT is affected by distance, this experiment was conducted using a hex nut with a thickness of 2.5 mm and a length of 5.45 mm. The CMUT was positioned facing downward toward the object and was incrementally moved upward along the z axis using the XYZ stage. The experiment setup, with enlarged images of the test object and CMUT, is shown in Fig. . Starting from the initial position (0, 0, 0), the height was increased in 5 mm steps until (0, 0, 45), with data being recorded at each step.

To examine the effect of object size on resolution, the experiment was repeated using a smaller hex nut measuring 4 mm in length and 1.5 mm in thickness. Pulse echo responses were recorded at three measurement points: 0 mm (minimum distance), 25 mm, and 50 mm. For comparison, the larger hex nut with a 5 mm length (used previously) was also tested under the same conditions, allowing for direct evaluation of size-dependent resolution changes. Figure  shows the two hex nuts.

To test the ability of the CMUT for topography mapping, the CMUT was moved in controlled steps using an automated XYZ stage over the 3D-printed test objects, and data were captured at each step. Two 3D structures as shown in Fig.  were designed and printed to be used as test objects. The first test object was an irregular structure containing protrusions of varying heights (10, 15, 20 mm) on a 5 mm base. The CMUT was moved linearly over the test object, and a total of 30 data points were recorded. A Python script was designed to analyze the pulse echo data obtained from the CMUT as it moves across the object. It plots the time-of-flight (ToF) values against the positions along the scan, facilitating the interpretation and analysis of the object's dimensions. To further explore this ability, the second test structure was designed, containing 3D representation of the letters “E”, “N”, “A”, and “S”. First, the letter E was mapped, and the acquired data were then processed through a Python script to visualize the responses in a 3D contour plot. The image on the left side of Fig.  shows the grid pattern (with x, y coordinates) in which the CMUT was moved and from which the data were captured for the letter E. In order to improve the measurement accuracy, the number of grid points was increased to map the letter N, as shown in the image on the right side of Fig. . The acquired data were then processed similarly to render a 3D contour plot.

To evaluate the minimum detectable object size, an experiment was conducted using a screw measuring 8 mm in length and 2.9 mm in diameter. The screw, mounted on a holder, was incrementally moved along the z axis in 2 mm steps, from (28, 70, 155) mm to (28, 70, 202) mm, while recording pulse echo responses. Figure  shows the stacked pulse echo signals collected at each position. The first detectable echo from the screw appeared at a z position of 160 mm and persisted until 166 mm. Beyond this point, echoes from the larger holder structure became dominant. In the pulse echo plots, the screw’s reflection was observed at around 81 µs, while the holder's stronger echo appeared earlier at approximately 75 µs. These results demonstrate that the CMUT is capable of detecting small objects (as small as 8 mm) and provides pulse echo responses that span the physical length of the object, thereby confirming its high resolution and sensitivity.

To conduct a more comprehensive analysis of the resolution aspect of the CMUT (in terms of the smallest detectable object size), three polycarbonate cubes measuring approximately 4×4 mm each were selected as test objects. The three pulse echo responses of the CMUT when it was directly above the cubes and one away from them and facing only the beaker are shown in Fig. . The results clearly demonstrate distinguishable echo signals from the polycarbonate cubes. The echo signals from the cubes are marked in rectangular boxes which are followed by the signal from the beaker. It can be seen that the echo pulses from the cubes appear in the time range of 64–66 µs, which is significantly earlier than the echo from the beaker, which appears at 69 µs, hence proving the CMUT's ability to detect and distinguish objects as small as 4×4 mm (length × breadth), marking the resolution of the CMUT at 4 mm. It is important to note that this resolution value could potentially be even better and is yet to be tested.

To investigate the resolution as a function of distance, a hex nut was used as test object, and the CMUT data were collected at different distances above it. The pulse echo responses obtained from the CMUT at each distance are shown in Fig. . The signals marked in the circles are received from the hex nut, and the following peak is from the beaker. It can be seen from the plots that CMUT resolution remains unchanged in the distance range of 40 mm. While there is no change in the resolution with distance, the signal amplitude does decrease to some extent with increasing distance owing to the attenuation in the medium. The signal amplitude versus distance is also depicted in Fig.  (right).

To examine the impact of object size on resolution, two hex nuts measuring 5.45 and 4 mm were used as test objects. Measurements were taken at three positions, starting from a minimum distance of approximately 1 cm from the CMUT (taken as 0 mm start position) and extending up to 50 mm. The results presented in Fig.  indicate that resolution remains consistent within the tested range (50 mm), demonstrating good accuracy in terms of resolution.

To analyze the received CMUT signals obtained from the linear scan of the irregular structure with varying heights, a Python script was written that takes the receive signal data and identifies the ToF value of the echo pulse using a rising-edge detection function. Based on this value, the distance is computed using Eq. (), where the speed of sound in water is taken to be 1500 m s⁻¹ assuming a constant water temperature (experiment conducted at room temperature). 1Distance=(ToF×speedofsoundinwater)2 The result of the Python script yielded a plot of depth (computed distance from the CMUT to the object) versus the distance traversed along the x axis (Fig. ), which closely reproduced the test structure and its dimensions, proving the CMUT's ability for topography mapping. Minor discrepancies were observed, likely due to the mounting of the test object on an uneven container surface or due to limited measurement points. Increasing the measurement points could enhance the reconstruction accuracy. To further analyze this ability, the test structures with a 3D representation of letters were used. The letters were situated on a 10 mm thick base structure and were individually mapped by the CMUT. These signals were then processed through a Python script that sequentially reads measurement data at each point, identifies the peak-to-peak voltage of the signal, and associates it with the corresponding coordinate on the grid. Subsequently, these data are visualized through both 3D contour plots for enhanced interpretation. The initial mapping focused on the letter E, which measured 16×10.7 mm², with a distance of 2.81 mm between its arms. A total of 35 measurements were taken to cover the test structure. Pulse echo responses, along with their corresponding ToF readings obtained at the coordinates (0, 0), (0, 9) and from the beaker for reference, are depicted in Fig. . Signal (a) represents the top surface of the 3D alphabet, signal (b) corresponds to the top surface of the base plane, and signal (c) depicts the beaker (with the CMUT moved away from the test structure). For a preliminary analysis of the results, ToF values were identified from the plots manually using a data analysis and plotting software. These values were then used to compute the distance values using Eq. () as described previously and are shown in Table . Preliminary analysis of the obtained data revealed that the difference between the computed distance values of signal (a) and (b) is approximately 9 mm closer to the thickness of the 3D letter E (10 mm), and the difference between signal (c) and signal (b) is approximately 10 mm, corresponding to the thickness of the base structure (10 mm). This proves the ability of the CMUT for dimension mapping.

To visualize these data, a Python script was used to compute and plot the peak-to-peak amplitude of these response signals using the greatest amplitude method and to map them to their respective coordinates. The resulting 3D and 2D plots are shown in Fig. . The plot reveals a distinct structure resembling the letter E, yet some discrepancies are evident, likely due to the limited number of measurement points in the data collection process. Given the narrow 2.8 mm spacing between the arms in the structure E and the 3 mm measurement step, the arms are not clearly distinguishable. To enhance the results, the second letter, N, was mapped with increased measurement points. Subsequently, the data were processed using the Python script to generate a 3D plot of amplitudes, as illustrated in Fig. . The 3D plot exhibited an improved replication of the letter N due to both an increased number of measurement points and the inherent simplicity of the letter's structure itself.