Self-sensing metallic material based on PZT particles produced by friction stir processing envisaging structural health monitoring applications

BT particles in this study. The SSM based on PZT particles showed a higher sensitivity than the SSM based on BT particles and processed aluminium. The inclusion of PZT particles improved the mechanical strength and decreased the electrical conductivity of the aluminium parts. The SSM based on PZT particles had a sensibility of 18 . 0 × 10 − 4 μ V / MPa and could detect solicitations with different frequencies, with the best performance observed under low-frequency solicitations. Additionally, EBSD, EDS, XRD and XPS analysis confirmed the existence of the PZT particles in aluminium parts. These results are promising and guarantee an increase in sensorial properties and the ability to self-monitor metal parts.


Introduction
Advances in Structural Health Monitoring (SHM) applications have resulted in the search for innovations, such as sensors, actuators, and built-in systems, to enable continuous monitoring throughout structural parts' life cycle [1][2][3].SHM can be carried out using an integrated monitoring system that includes contact surface sensors as well as sensors embedded in the monitored parts.Surface sensors are susceptible to damage from collisions and/or environmental action and embedding sensors, on the other hand, can be challenging and may result in a weakened part [4].In some applications, these technologies can be restrictive; for example, the incorporation of actuators can introduce inherent defects in their connections to the part, resulting in deterioration after a long period of use [5].Functionally Graded Materials (FGM) with high sensorial properties are expected to be a good solution to these problems [6].Until recently, the development of FGM was based on the graded compositional variation from high piezoelectric-low dielectric material to low piezoelectric-high dielectric material [7].Currently, Self-Sensing Materials (SSM) represent a different approach to obtain metal parts with sensorial properties.
Smart, self-sensing materials have the ability to revolutionize various fields by incorporating sensing capabilities directly into the material itself.Unlike traditional point-based sensors, these materials act as sensors throughout their entire structure, enabling continuous and distributed sensing.This approach has an enormous potential in areas such as infrastructure, aerospace systems, robotics, and biomedical devices.By utilizing stimulus-responsive properties, such as electrical conductivity, self-sensing materials can detect and report on various stimuli, including mechanical deformation, pressure [7], moisture content/humidity [8], pH [9], temperature [10,11] and others [12][13][14][15][16][17].
This offers benefits such as reduced implementation burdens, improved coverage, and real-time monitoring.With the ability to provide quick and easy electrical measurements, self-sensing materials have the potential to enhance human safety, prosperity, and health while minimizing the reliance on dense grids of sensors [18].A recent study highlighted the limitations of traditional point-based sensors and their associated costs.In the case of inspecting the fuselage of an aircraft, over 10,000 sensors would be needed to assess the entire structure, resulting in significant weight and cost increases [19].In contrast, self-sensing materials offer a solution by integrating sensing capabilities directly into the material, mitigating the need for excessive sensors and reducing weight.
Piezoelectric-based self-sensing is particularly appealing due to its ease of measurement and compatibility with data acquisition systems.These materials can be intrinsically piezoelectric or engineered through modifications with piezoelectric fillers, providing a wide range of material options.By enabling spatially continuous sensing and reducing the reliance on point-based sensors, self-sensing materials hold great promise in improving structural health monitoring, robotics, aerospace systems, and biomedical devices [20,21].
In recent years, attempts have been conducted to obtain smart materials with higher sensing and mechanical properties.Piezoelectric ceramic materials are technologically important in the field of smart materials [22][23][24][25][26]. Piezoelectric materials produce an electric charge under mechanical stress, which is directly proportional to the external force applied.When piezoelectric materials are used in a structure, an external load applied to this structure induces an output electric charge that allows to quantify the load [27][28][29].Under the same load, the health condition of different structural parts is reflected by the output electric charges of different piezoelectric sensors, allowing the structure's health to be assessed.The mechanical properties of such materials, including hardness and fracture toughness, were reported to be greatly improved [30].
Some authors have been studying piezoelectric composites, obtained by sintering piezoelectric ceramic particles, such as Lead Zirconate Titanate (PZT) and Barium Titanate (BT), with reinforcement metal particles, to improve material properties.These authors developed composites, such as PZT/Silver [30][31][32][33], PZT/Platinum [34,35] and BT/ Silver [36], and have found that the internal stress is relieved and occurs a dielectric and mechanical enhancement.Others reinforcement materials can also be used, such as, polymeric material [37,38], and Al 2 O 3 which improves fatigue resistance [39].However, the reported studies mainly focused on the effect of a small percentage of metal particles incorporated in piezoelectric ceramic matrix, i.e., volume fraction down to 30% metal [33,35].These types of materials are essentially used in sensors or actuators to increase their sensorial properties, and to develop nanoscale ferroelectric structures [40], but when the attentions are pointed to structural components, the smart material matrix should be metallic instead of piezoelectric ceramics, and the challenge for these applications increases.The matrix of these materials being metallic raises a set of challenges, that are linked with its high electrical conductivity properties.It is believed that the presence of piezoelectric ceramic powder in a metal matrix, such as, Nickel Aluminium Bronze/ BT [41], Bismuth ferrite/BT [42], 304 Stainless Steel/PZT [43,44] or Copper/BT [45] also enhances the material damping capacity [41], hardness, strength and exhibits superior cavitation resistance [46].Thus, the mechanical properties were improved, conducting enhanced hardness, wear resistance, corrosion resistance and low residual stresses [43,44].
Recently, a study demonstrated that the incorporation of piezoelectric ceramic particles within a metallic part can not only improve its mechanical properties, but also generate an electrical response when subjected to solicitations [47,48].The current study aims to investigate and explore the potential of incorporating piezoelectric particles in aluminium parts through the use of Friction Stir Processing (FSP) technology.However, there are other variants of FSP that allow incorporating particles in metallic matrix, as is the case of the variant presented by Vidal et al. [49] and Moreira et al. [50].Thus, PZT particles were utilized to enhance the sensitivity and technical features of the resulting SSM.Metallographic characterizations were performed to assess the distribution and concentration of the particles within the aluminium matrix, as well as the location and size of the particles.Additionally, mechanical properties were evaluated to ensure that the incorporation of piezoelectric particles did not negatively impact the mechanical behaviour of the base material.Finally, the sensorial properties of the SSM were assessed and its sensitivity was thoroughly evaluated.By conducting this study, a more in-depth understanding of the potential applications of self-sensing materials and the optimization of their properties can be achieved, opening up new possibilities for the development of advanced materials in a variety of industries.The employed technological process involves the transformation of metallic components into self-sensing materials, achieved by the strategic incorporation of PZT piezoelectric particles utilizing the innovative technique of FSP.This transformative procedure bestows upon aluminium components a unique capability, wherein they exhibit an electrical response to dynamic mechanical loads.This inherent electrical response can be harnessed as a sophisticated monitoring mechanism, providing real-time insights into the stress and strain experienced by the component throughout its entire lifecycle.So, future endeavors will need to materialize and deploy this technology effectively for the comprehensive lifecycle monitoring of metallic components.

Materials and methods
A SSM was produced using aluminium (Al) plates AA5083-H111 with dimensions of 203 (rolling direction) × 103 × 10 mm as base material.The nominal chemical composition of the Al plates is presented in Table 1.PZT piezoelectric micro-particles with the chemical formula PbTiZrO 5 and 99,9% of purity produced by Nanoshel Ltd. [51], were used to add the self-sensing ability to the aluminium plate.The piezoelectric and mechanical properties of the PZT particles are listed in Table 2.The piezoelectric constant (d 33 ) and the electromechanical coupling coefficient (k 33 ) are properties that are reflected when mechanical stress is parallel to the dipole, resulting in increased spontaneous polarization.The d 31 and k 31 effect is manifested by a perpendicular stress applied to the dipole moment, resulting in a transverse electrical charge.The PZT ceramic particles were produced from a solid solution of 52% of lead zirconate and 48% of lead titanate with a perovskite crystal structure [52].Additionally, BT piezoelectric micro-particles were utilized as a comparative method.Fig. 1 a,b show Scanning Electron Microscopy (SEM) images of the BT and PZT particles, respectively.The images show that the size of the BT particles is quite homogeneous, at approximately 2 μm.In contrast, the size of the PZT particles ranges from sub-micron sizes to about 20 μm.The PZT and BT particles were characterized by Zeiss Merlin VP compact SEM and applying gold coating with a layer thickness of 6 nm, via sputtering.

Self-sensing material fabrication
The PZT particles were embedded in the aluminium matrix by FSP according to the procedure described in [47].The groove dimensions employed (Fig. 2 b) were those that resulted in the best response to dynamic solicitations reported in [47].A set of FSP tools was used to process the material.The first tool was a pinless FSP tool with a left-hand scrolled shoulder, and the second one was a featureless concave shoulder with a triflute left-handed conical pin (pinned FSP tool).After compacting the particles in the groove, a pass with the pinless FSP tool was performed to close the groove and confine the particles inside.Four overlapped passes were then performed with the pinned FSP tool, all in the same direction.Then, a polarization process was carried out in a controlled environment at 90 • C applying a strong electrical field (9 kV/ mm) to create an asymmetry in the previously unorganized electrical dipoles' arrangements.Fig. 2 c depicts the FSP process to incorporate the PZT particles described.The FSP parameters are presented in Fig. 2 a.Additionally, two extra plates were used, one with BT particles (Al/BT) and one without any particles (Al Processed), to compare with the results of SSM based on PZT particles.
Temperature measurements were carried out while manufacturing the SSM using two K-type thermocouples placed between the backing plate and the AA5083-H111 plate, one on the advancing side (thermocouple AS) and the other on the retreating side (thermocouple RS), as depicted in Fig. 2 c.In addition, a Fluke Ti400 infrared camera was used for surface temperature measurements.

Self-sensing material characterization
Following SSM processing, samples were prepared for several characterization techniques.The macro and micrography, and X-Ray microtomography (μCT) samples were polished according to standard metallographic procedures.Then macro and micrography samples were etched in Keller reagent (2 mL HF, 3 mL HCl, 20 mL HNO 3 , and 175 mL H 2 O).
Macro and micrography optical analyses were performed with a Leica DMI 5000 M inverted optical microscope to examine the microstructure and particles' distribution.
The 3D and 2D micro-architectural morphology of the SSM based on PZT and BT particles was characterized by X-Ray microtomography (μCT), using a Phoenix V|TOME|X, GE.For that purpose, the SSMs were scanned at 200 kV and 120 μA.The scanning angular increment was 0.15 • , and the spatial resolution of 24.6 μm.The acquired image data were interpreted qualitatively and quantitatively using 3D tomographic reconstruction and analysis software.(Volume Graphics 3.04 software, Volume Graphics).
The SSMs samples were also mounted in a bakelite thermosetting resin with carbon filler, appropriate for examination in a SEM.The mounted samples were prepared using a Struers Tegramin automatic polisher with a final grinding step equivalent to 1200 grit paper, fol-    lowed by polishing with 1 and then 0.25 μm diamond suspension.The polished samples were slightly etched using Keller's etchant (≈ 10 s) to reveal the processed zone and particles' distribution.The sample preparation for Electron backscatter diffraction (EBSD) was the same but replacing the etching with a fine polishing step using a vibratory polisher and a non-crystallizing colloidal silica suspension for about 6 h to ensure a deformation-free.Additionally, a 2 nm layer of carbon was added by sputtering, in order to cover the non-conductive particles.
Optical macrographs reveal the macroscopic features of the processed zones and guide the choice of location for the energy dispersive Xray spectroscopy (EDX) measurements.The macrographs were created from multiple micrographs taken at ×20 magnification (using a Zeiss Axio Vert V1 MAT microscope equipped with EC Epiplan lenses) that were subsequently combined using the automated photomerge tool in Adobe Photoshop 2021.
The SEM and EDX were done using a Zeiss Merlin VP compact SEM equipped with Bruker XFlash EDX detector.The accelerating voltage was 10 kV, and the working distance was ≈ 10 mm.EBSD was carried out to assess the processed zone.The EBSD measurements were performed on a Zeiss Merlin VP compact SEM equipped with Bruker e-Flash HR EBSD detector.The EBSD analyses were performed with a step size of 0.1 μm.The accelerating voltage was 20 kV, and the working distance was approximately 20 mm.
The SSMs samples also characterized by X-Ray Diffraction (XRD) on a PANalyticalX'Pert Pro MDP diffractometer, with a copper anode (Cu Kα radiation) and an 1D X'Celerator detector.Measurements were obtained by continuous scanning in the 20-90 • (2θ) range.XRD data was analysed on a High Score Plus software.For X-ray photoelectron spectroscopy (XPS), a Kratos Axis Supra spectrometer equipped with a monochromatic Al Kα radiation source was used.Detailed spectra were recorded with a pass energy of 10 eV.Since slight differential charging occurred during the measurement, all binding energies were corrected a posteriori to C 1 s at 284.8 eV.
A Mitutoyo HM-112 hardness testing machine was used to measure the Vickers microhardness profile along the X direction of the processed plate.The top surface of the sample was machined, grounded and polished to obtain a homogenous surface condition.The spacing between consecutive indentions was of 1 mm for the base material and 0.5 mm for processed and thermal/mechanical affected regions.The load applied was 0.5 kgf for 10 s.
The electrical conductivity was characterized using eddy currents and potential drop measurement techniques.This electrical property was measured along a straight line along the X direction.The procedure was developed according Sorger et al. [56] and Santos et al. [57].So, to obtain a homogeneous surface of the samples, 1 mm of the top surface was machined.A pencil probe operating at 2 MHz (corresponding to a penetration depth of 0.019 mm for this alloy) and a NORTEC 600D impedance analyser was used to implement the eddy currents technique.A standard Jandel™ linear four-point probe with four straight aligned tungsten needles with a tip radius of 40 μm and with 0.635 mm probe spacing was used to implement the potential drop measurement technique.A Keithley SourceMeter 2450 was used to impose 80 mA and a Keithley Nanovoltmeter 2182 A was used to measure the voltage allowing the calculation of the electrical conductivity.
To assess the sensorial and tensile properties of the processed plates, uniaxial tensile test specimens were machined, according to the ASTM E8/E8M-13a standard, using a HAAS Super Mini Mill 2 CNC Machining Center.The geometry of the uniaxial tensile test specimens was the same than that used in [47].All specimens were produced with a thickness of 2 mm, to ensure accessibility to both sides of the nugget.
To characterize the mechanical behaviour of the SSM, uniaxial tensile tests were performed at room temperature using a servo-hydraulic MTS 312.21 testing machine with a load capacity of 100 kN.Additionally, fracture surface was analysed by SEM Hitachi High-Tech SU3800.
The sensorial properties of the SSM were assessed by measuring the electrical response when the specimens were subjected to dynamic loads.For this propose, a universal testing equipment MTS 312.21 was used to impose dynamic solicitations, and a Keithley Nanovoltmeter 2182 A was used to measure the electrical response, which was connected to a National Instruments DAQ and assisted by a LabView program for signal data processing.The experimental setup was described and presented in [47].To characterize the electrical response behaviour and sensorial properties of the SSM based on PZT particles (Al/PZT) two analyses were performed.First, the sensorial properties of Al/PZT were evaluated by subjecting it to a series of dynamic loads and measuring its electrical response, as described by Ferreira et al. [47].Finally, the electrical response characteristics of the SSM were investigated by measuring the electrical response when the SSM was subjected to the same load intensity but at three different frequencies (0.125 Hz, 0.25 Hz and 0.5 Hz).

Temperature measurements
Figure .3 b shows the temperature profile on the advancing (AS) and retreating (RS) sides recorded by the thermocouples during the 1st pass of FSP (groove closing processing) and the thermogram at time instant 76 s.On both plates (Al/PZT and Al/BT), thermocouple measurements revealed higher temperatures in the advancing side than that of the retreating side (Fig. 3 b).The Fig. 3 b also shows that the temperature measured by the thermocouples is higher for the Al/PZT plate compared to the Al/BT plate.Further analysis of the EDX data reveals that there are indications that the PZT particles are fragmented, while the BT particles remain approximately the same size.This may be one of the reasons why there was a greater increase in temperature during processing in the Al/ PZT plate, meaning that more heat is generated to fragment the PZT particles.It should be noted that the data presented in Fig. 3 b are only representative of the evolution and not the values reached at the nugget.Consequently, IR camera Fluke Ti400 was used to identify a value on the    stir surface.IR camera Fluke Ti400 measurements revealed peak temperatures about 280 • C during the 1st processing pass in Al/PZT Plate (Fig. 3 c).Because the thermocouples measurements were obtained under the plate, the temperature results were slightly lower than the IR camera measurements, as shown in Fig. 3 c.
The temperature was lower during the groove closing stage and was almost constant during processing stage, as it is observed in Fig. 4 a.During the groove closing process, the plastic deformation only occurs on the plate surface due to the use of pinless FSP tool to close the groove.Thus, the heat generation is lower than that of the following processing passes.During the processing state, with the pinned FSP tool, the plastic deformation increases and, consequently, the heat generation also increases.From Fig. 4, one can observe that on the processing stage, the maximum temperatures at each pass remain almost constant, around 340 • C for Al/PZT plate and 320 • C for Al/BT plate.So, the maximum temperature experienced by the material is not affected by the number of passes when using the same processing parameters and the same tool.
FSP involves the simultaneous input of heat and strain to the base material, resulting in a combination of static and dynamic recrystallization.During FSP, the simultaneous input of strain and heat initiates β phase transformation, as indicated by the phase diagram of Al -Mg alloy system shown in Fig. 5 a.The extent of recrystallization is determined by the strain rate and the material's peak temperature during FSP.So, the strain rate and peak temperature attained by the material are influenced by process parameters [58].The effects of this process on the microstructure, microhardness, and electrical conductivity of SSMs are explored in the following sections.During FSP, PZT and BT particles was subjected to temperatures of 340 and 320 • C, respectively.So, for processing temperature lower than Curie temperature, the perovskite special cell structure of PZT and BT will have structural distortion and spontaneous polarization.This occurs with PZT particles because their Curie temperature is 370 • C, which is higher than the processing temperature (340 • C).When processing temperature is higher than Curie temperature, the perovskite cell structure of piezoelectric ceramics will acquire a cubic lattice.BT particles exhibit this behaviour because their Curie temperature is 115 • C [47], which is lower than the processing temperature (320 • C).However, after cooling, it is expect that the BT lattice will present structural The map reveals the difference in particle concentration between the top "half" (higher particle concentration) and the bottom "half" (lower particle concentration) in the SZ.

Metallographic characterization
Macro and microstructural characterization allowed the analysis of the particles' distribution in the plates processed zone, and identify the main microstructural regions, namely the nugget, the region with the particles.Fig. 6 a-d presents the macro and microstructures of the SSM samples.There is an evident particle distribution in the nugget of Al/PZT and Al/BT plates proving that the FSP was able to promote particles' incorporation and distribution in the aluminium matrix.Furthermore, the particles were preferentially distributed in the advancing side, mainly in Al/PZT material.Fig. 6 e,f shows the stir zone of the aluminium processed plate and the thermal/mechanical affected region of friction stir processed material.The plates' stirred zone was subjected to grain refinement due to the dynamic recrystallization induced by FSP.Fig. 6 g-n shows the segmented and 3D μCT images of the Al/PZT and Al/BT plates.The particle distribution can be seen inside the nugget (red colour) using the green colour.As can be observed from Fig. 6 j, n, the particles were spatially distributed along the processed zone.Furthermore, it is possible to observe that the processed zone did not present internal porosities.
A sample of friction stir processed AA5083-H111 without any particles (Al Processed) was analysed by EDX to serve as a reference.Fig. 7 shows the optical macrograph of the defect free cross-section.The stirred zone (SZ) shows evidence of grain refinement and there is no evidence of any relevant heterogeneities in the SZ.The regions around the points marked "•a" and "•b" were selected for EDX analysis.The EDX maps (Fig. 7 a,b) identified the main elements in the alloy, namely Al and Mg, and some Si particles.
Fig. 8 shows the optical macrograph of the sample with embedded PZT particles.Four zones can be identified in the SZ, based on the different observable features and colour shade of the SZ.These regions are marked "•a", "•b", "•c", and "•d" in Fig. 8.The distribution of the Al (matrix), Pb and Ti (particles) elements in the sample was mapped via EDX but the signal intensity for Zr was too low, therefore this element was left out of the analysis.Low magnification EDX maps, presented in Fig. 8, support the distinction between the four zones, although the difference between regions "•b", and "•c" is slight at this magnification.Fig. 8 shows that the region labelled "•a"(Fig.10 a) has a more irregular distribution of particles than the remaining regions and that region "•d" (Fig. 10 d) seems to have a lower concentration of PZT particles.Fig. 10 a-d show high magnification (x2000 and x10000) maps of regions "•a", "•b", "•c", and "•d", respectively.Fig. 10 a highlight the irregular particle distribution in region "•a" along with some particle agglomerates which follow the stirred material flow lines.Fig. 10 b-d show that particles in regions "•b", "•c" and "•d".are more finely distributed, although particle distribution is not completely uniform, and areas of lower and higher particle density can be observed.The particles in region "•b" and "•d" are larger and more sparsely distributed compared with region "•c" where the particles are smaller but more densely distributed.Fig. 9 shows the largest particle found, which was located in region "•a".This particle has a roughly circular shape, with a radius of ≈59 μm, which is at the higher end of the original PZT particle size before FSP.Overall, most of the particles seem to be smaller than 10 μm suggesting that FSP resulted in particle fragmentation.
Fig. 11 shows the optical macrograph of the sample with embedded BT particles.In the SZ two main zones can be identified: a darker region, marked "•a"; and a lighter region marked "•b" in Fig. 11.
The distribution of the Al (matrix), Ba and Ti (particles) elements in the sample was mapped via EDX.Low magnification EDX maps, presented in Fig. 11, reveal that the darker region ("•a") has a higher concentration of BT particles.Conversely, the lighter region ("•b") in has lower concentration of BT particles.In region "•b" the particles are more homogeneously distributed whereas in region "•a" the distribution is slightly more irregular, with some discontinuity on the advancing side and a denser particle concentration zone on the retreating side.Fig. 13 shows high magnification (x2000) maps of regions "•a" (Fig. 13 a) and "•b" (Fig. 13 b), respectively.These maps confirm the higher particle concentration in region "•a" and show that the particles in this region tend to be larger than those found in region "•b".However, despite this tendency, some large particles can be found in region "•b" and, in fact, the largest particle overall was found in this region.The particle is shown in Fig. 12, and has a roughly elliptical shape, with a length of the longest axis ≈ 50 μm and length of the shortest axis ≈ 34 μm.Since the BT particles' size before FSP is < 2 μm it is possible that the particle in Fig. 12 is actually a dense agglomerate of particles instead of an individual particle, although that is not discernible from the figure.
Fig. 14 compares the inverse pole figure (IPF) and phase distribution maps obtained by EBSD measurements in the SZ of the sample processed without particles (Fig. 14 a,b), and the sample with PZT particles (Fig. 14 c,d).The IPF maps show that for both samples there is no preferential grain orientation (texture).It is also clear that the presence of the particles limits the size of the Al grains considerably.The phase map of the sample with PZT particles shows that the Al grains are surrounded by a non-indexed phase which is assumed to be PZT particles and/or intermetallic compounds formed during the processing, however this assumption requires further investigation.This suggests that the particles are located mostly around the Al grains, although some smaller particles can also be found inside larger Al grains (Fig. 14 d).
XRD performed in the piezoelectric particles and processed region of Al/BT and Al/PZT samples emphasize the existence of the BT and PZT particles, respectively, in the stirred zone, as shown in Fig. 15.The PZT pattern (Fig. 15 a) confirms the existence of PZT particles indexed to reference pattern (ICDD card 33-0784).The main reflections (101) and (110) located at 30.9 • and 31.3 • , respectively, are detected as well as the ones highlighted in the Fig. 15.BT particles are clearly visible in the processed sample (Fig. 15 b).The reflections at 22 • corresponds to the (100) and (001) reflections of the tetragonal structure (ICDD card: 005-0626) and with the more intense peaks of this structure positioned near 31 • corresponding to the (101) and (011) reflections.Some of the peaks are superimposed on the Al plane reflections (ICDD card 001-1176).
The chemical components was identified by XPS on the Al/PZT and Al/BT samples, as illustrated in Figs 16 and 17.The presence of Al, Mg, O, Pb, Zr and Ti elements was confirmed in the SSM based on PZT particles by the wide-range XPS spectra (Fig. 16).In addition, the presence of Al, Mg, O, Ba and Ti was confirmed in the SSM based on BT particles as shown in Fig. 17.The presence of C 1 s peak in Al/PZT and Al/BT samples is mainly related to adventitious carbon at the surface.In Al/PZT and Al/BT samples, the peaks of Al 2p and Mg 2p are located at 72.8 eV and 51.7 eV, respectively, indicating the major component of SSMs is AA5083-H111.The equivalent homogenous atomic concentrations of aluminium and magnesium of the samples are presented in Table 3.
In SSMs based on PZT particles, the presence of these particles is confirmed with the high-resolution spectra of the dominant peaks of Pb 4f and Zr 3d.The Pb and Zr showed both a unique chemical state, characterized by the characteristic doublet pairs, respectively: the Pb 4f 7/2 = 138.8eV and Pb 4f 5/2 = 143.6 eV for Pb; and the Zr 3d5/2 = 181.5 eV and Zr 3d3/2 = 183.9eV for Zr.These binding energies are typical for PZT [62].
Regarding the SSMs based on BT particles, the high-resolution spectra of the dominant peaks of Ba 3d confirmed the presence of BT particles in Al/BT sample.The Ba 3d photoelectron peaks have two   components in Al/BT sample, designated as a low energy peak at around 779.8 eV and the other known as at around 781.3 eV.The XPS measurements of the Al/PZT surface did not reveal a significant variation in the Ba 3d spectrum that could be attributed to the presence of oxide [63].
The peaks of Ti 2p, that are present in both samples (Al/PZT and Al/ BT) correspond to the doublets Ti 2p 3/2 and Ti 2p 1/2 at 457.7 and 463.4 eV, respectively.The splitting between Ti 2p 1/2 and Ti 2p 3/2 is 5.7 eV, corresponding to a normal state of Ti4+ [64].

Microhardness measurements and electrical characterization
To characterize the processed and non-processed zones and evaluate particles' distribution, microhardness measurements, potential drop measurements, and eddy current testing were performed in the transversal section (X direction) of the plates, 1 mm below the top surface.Hardness can be an indicator of processing conditions in terms of mechanical strength since it is directly proportional to it.Potential drop measurements allowed to obtain the electrical resistivity and conductivity in processed and non-processed zones.Eddy current testing is an expedited technique for assessing microstructural changes in materials, that is complementary to hardness measurements.In fact, other works [56,57,65,66] have shown that electrical conductivity is inversely proportional to hardness, and thus to the mechanical strength, with the advantage of not damaging the specimen's surface.As such, stirred zones have lower electrical conductivity, since more grain boundaries reduce the electronic mobility, while the thermal affected zones have higher conductivity due to the grain growth [56,57,65,66].
Fig. 18 depicts the hardness and electrical conductivity profiles for aluminium processed plate with PZT (Al/PZT).These results were also compared with aluminium processed plate without any particles (Al Processed) and aluminium processed plate with BT (Al/BT).In Fig. 18 a, it is observed a slight (≈ 20 HV0.5) hardness increase in the stirred zone when compared to the base material for Al/PZT plate.However, in the Al/BT plate, the presence of BT particles increased the hardness in stirred zone for ≈ 170 HV0.5, much higher than base material (≈ 90 HV0.5).So, a significant increase and uniform plateau hardness can be observed in the stirred zone for plates with build-in particles when compared to processed plate without any particles.The hardness increases in the plates stirred zone was caused by, at least, two phenomena, the grain refinement due to the dynamic recrystallization and the presence of the piezoelectric particles in nugget (Fig. 6).This result shows that the hardness increase comes mainly from the particle's presence and not from the grain refinement.
Potential drop measurements and eddy current testing are in good agreement with the obtained hardness profiles, as it is possible to observe in Fig. 18 c, d.The results show that the electrical conductivity of the plates is inversely proportional to microhardness and consequently to mechanical strength.PZT and BT particles significantly affect the electrical conductivity, it is possible to observe a decrease in the electrical conductivity in the nugget region of the plates with build-in particles.So, for Al/PZT and Al/BT, the results shown a decrement in electrical conductivity for values of the ≈ 25 IACS and ≈ 22 IACS, respectively (Fig. 18 c).The size of the PZT particles (with ranges from sub-micron sizes to about 20 μm) was higher than the size of the BT particle (at approximately 2 μm), and the electrical conductivity results shown that Al/PZT had a higher conductivity than Al/BT.Thus, SSM's electrical conductivity increased due to incorporation of larger particles.
The hardness and electrical conductivity profiles presented show that the inclusion of PZT and BT particles inside Al plates can modify the material properties, increasing mechanical strength and decreasing electrical conductivity.Furthermore, when compared to the Al/PZT plate results, the inclusion of BT particles increases the mechanical properties more.The obtained results could be a consequence of particles' size and geometry, i.e., for many FSP passes, the smaller particle size (BT particles) improves the mechanical properties of the SSM over the higher particle size (PZT particles), according to Mehdi et al. [67].

Tensile properties
The mechanical behaviour of the SSM (Al/PZT) was characterized using uniaxial tensile tests.These results were also compared to the mechanical behaviour of an Al/BT and an Al processed plates.Fig. 19 a, b depicts the engineering stress/strain curves, which show that the Al/ PZT has a higher yield strength than that of the Al processed due to the presence of PZT particles in the aluminium plate.However, Al/PZT plate presented a lower yield strength than the Al/BT plate.Furthermore, the incorporation of PZT ceramic particles within the aluminium matrix increases the brittleness of the plates (Fig. 19 b).This behaviour was concordant with macro and microstructural characterization, hardness, and electrical conductivity measurements.In addition, the yield strength, obtained using the 0.2% offset method, increased when compared with Al/processed and it decreased when compared with Al/ BT plate, as shown in Fig. 19 c.
As it is clear from Fig. 19, the incorporation of particles led to an increase in both the yield strength and ultimate tensile strength of the samples.However, in the case of the Al/PZT plate, the fracture strain decreased more significantly.The ultimate tensile strength of the Al/BT plate was 20.2% higher (298.8MPa) compared to the Al/processed plate (248.6 MPa).Similarly, the Al/PZT plate exhibited a 1.1% higher ultimate tensile strength (251.4MPa) compared to the Al/processed plate.However, both the Al/BT and Al/PZT plates had smaller elongations, with reductions of about 1.3% (14.15%) and 67.8% (4.62%) respectively, in comparison to the Al/processed plate (14.34%).
Well-developed deep dimples can be observed in the fracture surface of the SSM based on PZT particles (Fig. 19 d, e, f) and BT particles (Fig. 19 g, h, i) indicating a ductile failure mode.For Al/PZT, a bimodal distribution of smaller and larger dimples, a typical characteristic of metal matrix materials with well bonded particles, can also be observed.Larger dimples are generally associated with the particles, whereas  smaller dimples are associated with the ductile failure of the aluminium matrix.Fig. 19 f shows that some particles were fractured, there was decohesion or bonding with the matrix, again suggesting a weak interfacial bonding.As a result, the failure occurred at the interface of the nugget with the base material (Fig. 19 d).
For Al/BT exists a unimodal distribution of dimples, a typical characteristic of ductile failure of the metal matrix, as can be observed in Fig. 19 h, i. Fig. 19 i does not show particles, indicating that the failure occurred on the base material, as shown in Fig. 19 g, suggesting a strong interfacial bonding of the BT particles and the aluminium matrix.
The fracture surface revealed a ductile failure, and it also showed the superior bonding between the BT particles and the Al matrix, possibly because BT particles' size is smaller than PZT particles' size.
The findings depicted in Fig. 19 demonstrate that the incorporation of piezoelectric particles has a noteworthy influence on the plate's strength, particularly in the case of the Al/BT plate.The observed enhancement in the mechanical properties of the SSM can be attributed to factors such as: (i) the effective transfer of loads from the matrix to the fine reinforcements, (ii) significant grain refinement resulting from improved matrix dynamic recrystallization, and (iii) the mismatch in elastic and thermal expansion coefficients between the matrix and reinforcements.These conditions contribute to the failure of the Al/BT plate near the grips (Fig. 19 g).

Sensorial properties assessment
The sensorial properties of the SSM based on PZT particles (Al/PZT) were assessed by applying dynamic loads and measuring the electrical response.This assessment was also performed and compared with the Al/BT plate, as it is depicted in Fig. 20 a.This result shows that the Al/ PZT plate has a higher sensibility than the Al/BT plate, i.e., the Al/PZT had a sensibility of 18.0 × 10 −4 μV/MPa, which was 50% higher than the Al/BT plate (12.0 × 10 −4 μV/MPa).Thus, the SSM based on PZT particles provided the most evident and highest response to dynamic solicitations.This highest response could be attributed to the higher piezoelectric constant (d 33 / d 31 ) of PZT particles, which is 270/ -120 pC/ N for PZT materials and 190/ -78 pC/N for BT materials.
The effect of different loading frequencies on the electric response of a SSM based on PZT particles was also investigated.The influence of loading frequency on the electrical response of the SSM is evident in  to frequencies of 0.50 Hz, 0.25 Hz, and 0.125 Hz, respectively.Thus, it can be observed that the decreasing loading frequency results in an increase in the sensitivity of the Al/PZT plate.According to some authors [68,69], increasing the frequency causes a decrease in the dielectric constant, which reduces the permittivity of the material and its electron mobility.Thus, the electrical response of the SSM was expected to decrease slightly as loading frequency increased.So, this SSM may have the best performance when embedded inside metallic parts and subjected to low-frequency solicitations.The frequency of electrical response was an important characteristic studied by the authors.So, Fig. 20 c shows the electrical response of Al/ PZT plate when subjected to different loading frequencies (0.125 Hz, 0.25 Hz and 0.50 Hz with ≈ 90 MPa intensity).The results show that the frequency of electrical response remains the same when compared to the loading frequency.So, this SSM can detect solicitations with different frequencies.Additionally, as depicted in Fig. 20 c, it is evident that the aforementioned conclusion holds true.Specifically, the loading frequency of 0.50 Hz resulted in a lower electrical voltage response compared to the response generated by the load at 0.125 Hz and 0.25 Hz.

Conclusion
A new self-sensing material based on PZT particles was developed and demonstrated to be suitable for producing a metallic part with sensorial properties.The effect of PZT particles inside aluminium parts was investigated experimentally, and there was a grant of sensorial properties demonstrated by an electrical voltage response obtained through dynamic solicitations.The SSM based on PZT show a higher sensitivity compared with the SSM based on BT particles and processed aluminium.
The development of the self-sensing material based on PZT particles resulted in the following conclusions: • PZT and BT particles were successfully embedded in aluminium parts by FSP.• The Al/BT plate presents an heterogenous particle distribution, where two regions can be distinguished: a region consisting roughly of the top half of the SZ, and the lower half of the SZ.• In the Al/PZT plate, the SZ is more homogenous than that of the Al/ BT plate.• The BT particles seem to be more evenly distributed within each region of the SZ, while the PZT particles show a higher tendency to form agglomerates. • The particles limit the grain size in the SZ.
• The incorporation of particles led to an increase in both the yield strength and ultimate tensile strength of the Al/PZT and Al/BT samples.• The experimental results showed that SSM based on PZT particles has a sensibility of 18.0 × 10 −4 μV/MPa.
• This SSM can detect solicitations with different frequencies (0.125 Hz, 0.25 Hz and 0.50 Hz) and had the best behaviour when subjected to low-frequency solicitations (0.125 Hz).
These results are very promising because they guarantee an enhancement in sensorial properties and the ability for self-monitoring of metal parts.This research will improve structural health monitoring capability, using advanced materials, smart materials, and self-sensing

Fig. 3 .
Fig. 3. Temperature recorded during 1st pass processing (groove closing process): (a) experimental setup used to measure the temperature with the thermocouples; (b) time-evolution of temperature on the advancing (AS) and retreating (RS) sides; (c) experimental thermogram obtained by IR camera Fluke Ti400.
P.M.Ferreira et al.

Fig. 6 .
Fig. 6.Macro and microstructural characterization of self-sensorial material Al/PZT (a, b) and Al/BT (c, d), and processed aluminium (e, f); segmented μ-CT images of Al/PZT and Al/BT samples (g, h, i, k, l, m); the 3D representation of the μCT of Al/PZT and Al/BT samples (j, m).

Fig. 7 .
Fig. 7. Optical macrograph of the polished and etched FSP AA5083-H11 sample without particles and EDX maps of the (a) SZ and (b) base material of the FSP AA5083-H111.

Fig. 8 .
Fig. 8. Optical macrograph of the polished and etched SSMs sample based on PZT particles and low magnification EDX map of the Al/PZT sample.The map reveals the difference in particle concentration and distribution in the SZ.

Fig. 9 .
Fig. 9. EDX map showing the largest particle found in the SZ of the Al/PZT sample.×3000 magnification.
P.M.Ferreira et al.

Fig. 5 b
depicts the schematic diagram of the phase diagram of PZT piezoelectric ceramics as well as the lattice distortion in each temperature range.Furthermore, due to their identical perovskite structure, the phase diagram of BT piezoelectric ceramics is like that of PZT ceramics.

Fig. 11 .
Fig. 11.Optical macrograph of the polished and etched SSMs sample based on BT particles and low magnification EDX map of the Al/BT sample.The map reveals the difference in particle concentration between the top "half" (higher particle concentration) and the bottom "half" (lower particle concentration) in the SZ.

Fig. 12 .
Fig. 12. EDX map showing the largest particle found in the SZ of the Al/BT sample.x3000 magnification.

Fig. 14 .
Fig. 14.EBSD maps of the processed sample without particles (a, b), and the processed sample with PZT particles (c, d).The IPF figures (a, c) reveal the texture of the processed zones and the phase maps (b, d) show the phase distribution.The black colour in each map corresponds to non-indexed points.
P.M.Ferreira et al.

Fig. 18 .
Fig. 18.Samples characterization along the X direction: (a) macrography; (b) microhardness profile; (c) electrical conductivity obtained from Potential Drop with a linear four-point probe; (d) impedance change obtained from the Eddy currents technique.

Fig. 19 .
Fig. 19.Uniaxial tensile tests of self-sensorial materials: (a) engineering stress-strain curve; (b) elastic region; (c) yield strength.Fracture surface of the SSM based on the PZT particles (d, e, f) and BT particles (g, h, i).

Fig. 20 .
Fig. 20.Response to dynamic loads: (a) ΔV / Stress response of Al/PZT and Al/BT at 0.25 Hz; (b) the effect of different load frequencies on the electric response of an SSM based on PZT particles; (c) electrical response frequency of an SSM based on PZT particles when subjected to different load frequencies.

Table 3
Concentrations of major elements at surface of Al/PZT and Al/BT samples as determined by XPS; the brackets show the atomic fractions if oxygen and carbon is neglected.Other minor elements in the alloy were not detected.
P.M.Ferreira et al.