Skip to main content
NCBI home page
RSC Advances logoLink to RSC Advances
. 2026 Jul 2;16(34):33120–33134. doi: 10.1039/d6ra02511k

Pulse sensor for cardiovascular health monitoring based on self-poled piezo–tribo hybrid composites

Soham Kumar a, Jyotika Nanda b, Sam K Jacob a, Sreelakshmi S a, Prathipkumar S a, Gopal K Pradhan b, Pranab Mandal a,✉
PMCID: PMC13326701  PMID: 42395779

Abstract

Wearable devices, such as a pulse sensor, can be useful for real-time monitoring of cardiovascular health. Continuous and systematic monitoring of cardiovascular parameters can predict conditions such as coronary artery disease, heart failure, cardiac arrest, etc. However, most commercial wearable sensors are typically limited to macroscopic metrics such as heart rate and SpO2. This work introduces a sensitive piezoelectric–triboelectric hybrid composite-based pulse sensor that provides insight into an individual's cardiovascular health by measuring heart rate variability (HRV) metrics, including heart rate, P–P intervals (the time between successive heartbeats), and RMSSD (the root mean square of successive differences between intervals). The high sensitivity of the sensor is rooted in a dual strategy: first, the piezoceramic BaTi0.94Sn0.06O3 (BST6) is engineered at a morphotropic phase boundary, yielding an excellent piezoelectric coefficient of Inline graphic. Second, the device employs a layer-by-layer configuration incorporating a self-poling PDMS–MWCNT interfacial layer sandwiched between the PDMS–BST6 hybrid composite layers. This layered architecture facilitates self-poling by enhancing charge collection at the PDMS–MWCNT layer while suppressing recombination at the PDMS–BST6 hybrid composite layer. Further, Short-Time Fourier Transform (STFT) and Dynamic Time Warping (DTW) analyses of the recorded pulse wave suggest device stability over 100 days and evaluate heart condition. This study opens the scope for an IoT-enabled wearable system for long-term, real-time cardiovascular health monitoring and assessment.

Comprehensive view of material design for hybrid nanogenerator fabrication to advance wearable pulse sensor fabrication for cardiovascular health assessment.graphic file with name d6ra02511k-ga.jpg

1. Introduction

There is an increased risk of cardiovascular diseases (CVDs) in the older population worldwide.1 However, critical indicators can help in early prediction, identification, and diagnosis of CVDs and increase the chance of survival in patients.2,3 Currently, medical tools such as electrocardiography (ECG) and photoplethysmography (PPG) are widely used in healthcare systems to monitor cardiovascular health parameters, which require a specialized environment and trained personnel for operation.4 In contrast, commercial pulse monitoring systems are cost-effective and portable; however, they offer very limited parameters, such as heart rate (bpm) and peripheral capillary oxygen saturation (SpO2) levels. Detection of the arterial pulse waveform, including detailed cardiac features, enables measurement of critical HRV metrics such as R–R/P–P intervals (the time between successive heartbeats), RMSSD (the root-mean-square of successive differences between R–R/P–P intervals), ΔT (the time difference between the systolic and diastolic waves), etc. RMSSD is a primary marker of the parasympathetic nervous system and is highly sensitive to psychological stress, emotional regulation, and physical fatigue. On the other hand, continuous monitoring of the P–P interval can reveal the heart rhythm, the possibility of sinus arrhythmia, atrial fibrillation, or AV blocks. In this context, a wearable, portable pulse sensor that can measure various heart rate variability (HRV) metrics would offer significant benefits for real-time cardiovascular health monitoring and the risk of CVDs.

Multifunctional sensors are of great research interest nowadays due to their potential applications in the biomedical field and their versatility.4,5 The application of such sensors is not only limited to the biomedical field but also extended to soft robotics, wearable displays, and many other fields.6,7 The wearable multifunctional sensors are often based on three mechanisms: (i) piezoelectric, (ii) piezoresistive, and (iii) piezo-capacitive mechanisms.8–10 While piezoelectric sensors generate voltage under stress or vibration, piezoresistive sensors change resistance or current under stress.11 Piezo-capacitive sensors exhibit a change in capacitance when subjected to mechanical stress.12,13 Piezoelectric-based sensors are often used in biomedical devices due to their robustness and accuracy. Barium titanate (BTO), lead zirconate titanate (PZT), aluminium nitride (AIN), polyvinyl fluoride (PVDF), etc., are often used in these sensors as piezoelectric components. PZT-based devices are often used to determine pulse waves and detect systolic and diastolic peaks.14 However, Pb-containing devices/sensors are considered to be toxic and hazardous to health. At the same time, bulk piezoelectric materials such as BTO are excellent candidates for human pulse sensors due to their electromechanical coupling, but they often lack flexibility. To counter the flexibility issue, piezo–polymer-based composite sensors are also introduced.15 PVDF–ZnO-based sensors exhibit a clear pulse spectrum, but the signal amplitude is in the picoampere range.16 On the other hand, triboelectricity-based sensors are flexible but often suffer from noise due to their robust contact electrification mechanism.17,18 Interestingly, a combination of piezoelectric and triboelectric mechanisms is compatible for sensor fabrication due to their compatibility and similar operational features: mechanical-to-electrical energy conversion.19–22 The most common way to combine these two effects is to form a composite by mixing piezoelectric ceramics with a triboelectric polymer. The piezoelectric ceramic within the polymer matrix generates a piezoelectric potential under stress, and the polymer produces triboelectric charges upon contact with a different material.23,24

In this work, we demonstrate the fabrication and operation of a piezoelectric–triboelectric composite-based nanogenerator and pulse sensor for cardiovascular monitoring. The piezoelectric component is identified based on the study of Sn-doped BaTiO3 ceramics.25 The BaTi(1−x)SnxO3 solid solution offers several morphotropic phase boundaries (MPBs) in the range 0.03 ≤ x ≤ 0.12, where two or more polar phases coexist, resulting in a high piezoelectric coefficient.25 The high piezoelectric coefficient near the MPB region has been attributed to the joint effect of the anisotropy energy and the order parameter (polarization) at the phase boundary of two different polar phases.25 BaTi0.94Sn0.06O3 (BST6) was identified as the most suitable ceramic for sensor applications at room temperature (RT) due to its optimal electromechanical properties, i.e., high piezoelectric coefficient of Inline graphic and a moderate Curie temperature (TC) of 82 °C. The HNG based on these composites delivers an output voltage of 328 V and a power density of 0.9 mW cm−2. The high performance in the HNG is assisted by the excellent electromechanical response Inline graphic of BaTi0.94Sn0.06O3 at the morphotropic phase boundary. Secondly, integrating a PDMS–MWCNT self-poling layer into the device structure (Fig. 1) enhances charge collection and significantly improves the HNG device's voltage response. The composite-based pulse sensor offers a signal-to-noise ratio (SNR) of 12 dB, and a response time of 38 ms. That enables us to detect heart rate, P–P interval, and RMSSD of 74 bpm, 0.85 s, and 60 ms, respectively. Additionally, the comparative study of the fabricated pulse and ECG sensors reveals a strong linear correlation, with a Pearson correlation coefficient of 0.98. Further, dynamic time warping (DTW) analysis indicates high similarity between pulse waves over a 100 day interval, indicating the sensor's cyclability and reproducibility. The stability and durability of the current prototype device open an opportunity for integrated signal conditioning to enable future wireless transmission for real-time, portable cardiovascular monitoring.

Fig. 1. Schematic representation of composite fabrication, sensor fabrication, and cardiovascular parameter detection.

Fig. 1

Open in a new tab

2. Experimental section

2.1. Materials

BaCO3 (99%, LOBA Chemie Pvt. Ltd), TiO2 (99%, LOBA Chemie Pvt. Ltd), and SnO2 (99.99%, Sigma Aldrich) are used for the synthesis of Sn-doped BaTiO3 ceramics through the solid-state reaction route. Silver Conductive Lacquer (RS PRO) was used for electronic connections in subsequent measurements. SYLGARD™ 184 Silicone Elastomer Kit (DOW), tetrahydrofuran (99%, Avra), MWCNTs (TCI Chemicals), aluminium tape, copper wire, and double-sided tape are used for fabricating PNG and HNG devices.

2.2. Sn-doped BaTiO3 (BST) synthesis

Sn-doped BaTiO3 with chemical formula BaTi(1−x)SnxO3 is synthesised using conventional solid-state reaction. In brief, BaCO3, TiO2, and SnO2 are preheated at 180 °C for 12 hours, then weighed in a stoichiometric ratio. The measured chemicals are grinded in an isopropyl alcohol (IPA) medium. Later, the grinded chemicals are kept in an alumina crucible covered with a lead and heated at 750 °C for 12 hours at a heating and cooling rate of 5 °C min−1. After that, the powders are hand-grinded again and heated at 900 °C for 9 h at a heating and cooling rate of 5 °C min−1. Later, the powders are ball-milled in IPA for 1 hour and then heated at 1100 °C for 6 hours at a heating and cooling rate of 5 °C min−1. The collected powders are again ball-milled in IPA for 2 hours, an organic binder (polyvinyl butyral, 2 wt%) is added, and the mixture is then pressed into pellets (8 mm diameter, 0.5 mm thick). Finally, pellets are sintered at 1300 °C for 3 hours, with a heating and cooling rate of 3 °C min−1, to obtain dense pellets. These pellets are silver-painted on both sides for further electrical characterization. For HNG, PNG fabrication, the pellets are crushed and ball-milled for 6 hours to reduce particle size, then used for further processing. In total, 6 ceramic samples are prepared; a detailed description of the samples is tabulated in the SI (Table S1).

2.3. Preparation of composite active layer

First, the ball-milled BST ceramic is added to tetrahydrofuran (THF) and stirred. Next, Sylgard 184 (PDMS) base with a 10 : 1 weight ratio (10 wt PDMS : 1 wt BST) is added to the THF-BST solution and mixed using a Probe Sonicator at 260 W power for 1 hour in an ice bath. After that, the solution is stirred on a hot plate at 70 °C (1–2 hours) to remove THF by evaporation. Once the THF is evaporated, the solution is kept at room temperature (RT) for 1–2 hours to cool, and the curing agent (1 wt% of the PDMS base) is added once it reaches RT. This composite mixture, named P–B, is coated onto Al foil using a thin-film coater. The coated P–B layer is kept at 60 °C for 3 hours to dry and cure. The detailed description of a composite layer containing PDMS–BST6–MWCNT (P–B–M) is provided in the SI (Section S1). For the fabrication of different HNG device structures, P–B and PDMS–MWCNT (P–M) composite mixtures are also required. The synthesis process of the P–M composite mixture is presented in Fig. S1 and S2.

2.4. Pulse sensor fabrication

Two pulse sensor devices, PS-1 and PS-2, are fabricated using the composite layers P–B and P–B/P–M/P–B, respectively, as outlined in Table S1. Kapton tape of 150 µm thickness (with adhesive) is used as the top and bottom substrate, and Al tape is affixed on Kapton as the electrode. Cu wires are connected to the Al for electrical connections. The composite layer is sandwiched between two Al-Kapton substrates to form the sensor device, with no noticeable air gap; however, a small air gap will be present between the electrode and the composite layers if the electrode is not deposited directly on the composite layer. The sensor is fabricated as a wristband, and a detailed analysis of cardiac cycles and cardiovascular parameters is performed.

2.5. Device fabrication

Five different device structures, namely, PNG, PNGC, HNG, HNGC, and HNGL, are fabricated utilizing the composite layers PDMS–BST6 (P–B), PDMS–BST6–MWCNT (P–B–M), and PDMS–MWCNT (P–M) under various configurations as summarized in Table S2 and outlined below. For the device fabrication, the P–B layers (coated on Al foil, size 2 × 2 cm2) are affixed on the PET substrates, and Cu wires are connected to the Al foil. On the other hand, another Al tape is affixed on the PET substrate and placed on the top of the previously made assembly (P–B layer + Al-PET) without any air gap. Similarly, a Cu wire is connected to the top Al electrode, and the entire device assembly is sealed with anti-static tape and named as PNG. Similarly, PNGC devices are made by replacing the composite active layer with the P–B–M layer and are named PNGC. Similarly, hybrid nanogenerator (HNG) devices are fabricated. A small air gap is provided between the (active layer + Al-PET) assembly and the top electrode using double-sided tape. For the fabrication of layer-by-layer HNG (HNGL), the P–B layer is first coated onto the Al foil using a thin-film coater, then dried in an oven at 90 °C for 2–3 minutes. Again, the P–M layer is coated onto the dried P–B layer using a thin-film coater, then dried in an oven at 90 °C for 2–3 minutes. Additionally, the P–B layer is coated on top of the previously made layer and dried in a similar manner. Then, Al-PET is affixed to the previously assembled components (P–B/P–M/P–B) on Al-PET, and a small air gap is created with double-sided tape. For the electrical connection, Cu wires are connected to the top and bottom Al electrodes.

2.6. Characterization

For the structural characterization of BST ceramics and composite active layers, a powder X-ray diffractometer (PXRD, Malvern PANalytical Empyrean, Cu-Kα radiation, λ = 1.5406 Å) is used. Raman spectroscopy with a micro-Raman spectrometer (Renishaw inVia) is used to detect ferroelectric modes in BST ceramics and to characterize and map the presence and dispersion of MWCNTs in active layers. For the dielectric spectroscopy study, a custom-made setup coupled with an LCR meter (Keysight E4980A) is used.26 Ferroelectric and piezoelectric properties are measured with the help of the Precision LC II Ferroelectric Tester (Radiant Technologies, Inc.) coupled with a high-voltage amplifier (ELECTRICAL ENERGY LIMITED). Field Emission Scanning Electron Microscopy (FE-SEM, Zeiss Gemini SEM 450) is employed to study the morphology of the composite active layers, and Energy Dispersive X-ray spectroscopy (EDX) is used for elemental analysis of the BST ceramic. The sensor output is measured as voltage under pulse vibrations, and a DAQ6510 Data Acquisition and Logging Multimeter System (Keithley, Tektronix) is used for all sensor measurements. Signal-to-noise ratio (SNR) and response time (τ) for all sensors are calculated from the collected pulse wave signals. 9 trials, each of 120 s, have been conducted on five different volunteers of different ages (aged 24 to 31 years) to emphasize the sensor characteristics and study different cardiovascular parameters over time. A detailed analysis of cardiac cycles and cardiovascular parameters in a 28 year-old male volunteer is presented in the results and discussion sections. Short-Time Fourier Transform (STFT) and Dynamic Time Warping (DTW) analysis are used to study the changes in cardiovascular parameters over a 100 day interval. Informed consent letters were obtained from the volunteers for their active participation in the experiments involving the pulse sensor.

Open circuit voltage (VOC) and short circuit current (ISC) are measured from the PNG and HNG devices under a periodic force of ∼20 N at 2 Hz using a Mixed Domain Oscilloscope (MDO 34, Tektronix) and a DAQ6510 Data Acquisition and Logging Multimeter System (Keithley, Tektronix). The power density of the HNG devices is calculated using the formula,

2.6. 1

where U is the power density, VPP is the peak-to-peak voltage, RL is 10 MΩ load, and A is the area of the active layer.22 The error in VPP and IPP is calculated from 3 trials on a particular device, and 2 different devices of the same active layer are used throughout the study (a total of 6 trials per device). The error in the power density is calculated using the propagation-of-error method, and the same devices were used to report VPP and U in this study. The output performance of the fabricated devices is evaluated over six independent trials (n = 6). To assess the significance of the architectural improvements, a mean separation method (Tukey's Honestly Significant Difference (HSD)) was employed at a level of significance of α = 0.05, and the results are presented as mean ± standard deviation (SD) throughout the study.

3. Results and discussion

3.1. Sn-doped BaTiO3 ceramics characterization

3.1.1. Structural study on BST ceramics

Room temperature powder XRD patterns for all 6 samples (BST-3 to BST-12) show pure phase perovskite structure without any noticeable secondary phase, as presented in Fig. 2(a). The evolution of the peak in the vicinity of 2θ ∼ 45° is shown in the second panel of Fig. 2(a). The two distinct planes (002) and (200) are visible in the pure BT sample, indicating the presence of a tetragonal phase. As the Sn concentration increased in the BT matrix, the (002) and (200) planes merged towards lower 2θ values, indicating a transition to the cubic phase (Pm3̄m).27 However, several coexisting phases can be approximated in the presence of the (022) plane (orthorhombic phase) near 6 mol% Sn doping, which requires further analysis of the PXRD data.

Fig. 2. (a) PXRD graphs of samples BST-3 to BST-12 collected at 23 °C, (b) Curie temperature and piezoelectric coefficient (at 27 °C) of samples BST-3 to BST-12.

Fig. 2

Open in a new tab

The prominent dielectric transition from the ferroelectric to the paraelectric phase (Curie Temperature, TC) is visible for BST-3 near 105 °C. As we approach BST-12, the dielectric transition temperature shifts to lower temperatures and falls below the accessible temperature range (Fig. S3 and 2(b)). This indicates that the TC for BST-12 is very near RT. The TC for BST-3, BST6, and BST-9 is approximated to be 105 °C, 82 °C, and 58 °C, respectively, and is consistent with previous reports.28,29 The optimum TC (82 °C) of BST6 ceramic makes it a suitable candidate for sensing applications at RT. Furthermore, detailed structural studies of BST6 are carried out using PXRD and Raman spectroscopy. A significant shift of (200) plane is observed in the case of BST6 as compared to BaTiO3 (P4mm) PXRD patterns at RT (Fig. 2(a)). Moreover, the (002) planes are almost absent, and the presence of (022) planes is observed, indicating the orthorhombic (Amm2) characteristic of BaTiO3 at low temperature. The profile fit to the PXRD data is performed using three different space groups: P4mm, Amm2, and R3m. The goodness-of-fit parameters indicate that the sample lies on the phase boundary between the orthorhombic (Amm2) and tetragonal (P4mm) phases, as shown in Fig. 3(a). The fitting of the peak near 2θ ∼ 45° is presented in Fig. S4. The intensity patterns differ significantly from those of the tetragonal phase; neither shows a good fit to the orthorhombic phase. However, the mixed-phase model (Amm2 + P4mm) yields the best fit to the PXRD data, and the corresponding cell parameters are presented in Table S2.

Fig. 3. (a) Pawley fit to the room temperature PXRD data of BST6 using mixed phase model of (P4mm + Amm2), (b) room temperature Raman spectra of BST6, (c) room temperature polarization (P)–field (E) hysteresis loop and strain (S)–field (E) loop of BST6, (d) unipolar strain (S)–field (E) loop of BST6 at room temperature.

Fig. 3

Open in a new tab

While PXRD provides the average long-range crystal structure, Raman spectroscopy probes local structure and different bindings. The overall characteristic peaks of BST6 are similar to BaTiO3 (BTO). The presence of a sharp E + B1 mode at ∼305 cm−1 and the longitudinal optical (LO) mode at 715 cm−1 are definitive Raman signatures of the long-range ferroelectric order characteristic of the tetragonal phase of BaTiO3.28,29 Their presence in BST6 confirms that the material retains a strong ferroelectric character despite Sn doping, which is essential for a large piezoelectric effect. The appearance of a new band near 115 cm−1 and a dip near 125 cm−1 is highly significant. These features are attributed to disorder-activated scattering and interference of A(1)1 (TO) phonons.28,30 The substitution of Ti4+ with the chemically different Sn4+ ion disrupts the perfect periodicity of the crystal lattice. This local disorder is a direct consequence of the doping process. It is a common feature in systems near an MPB, where the local environment can fluctuate between different phase symmetries. The distinct feature near 192 cm−1 is assigned to a strain-driven orthorhombic phase (Fig. 3(b)).30,31 This finding provides direct spectroscopic validation of the mixed Amm2 (orthorhombic) + P4mm (tetragonal) phase model deduced from PXRD fitting. It confirms that phase coexistence is not merely a macroscopic averaging effect but is manifested in the material's local vibrational properties. The Raman data reveal a complex local structure characterized by the preserved long-range ferroelectric order, significant local lattice disorder (typical of MPB compositions), and coexistence of the orthorhombic phase.

3.1.2. Ferroelectric and piezoelectric studies

Polarization (P)–electric field (E) studies are performed for all the samples at RT. All the studied compositions show P–E hysteresis behaviour at RT with noticeable remanent polarization (Pr) and saturated polarization (Ps) values, as shown in Fig. S3(c). The hysteresis behaviour in the P–E loop confirms the ferroelectric nature of the samples at RT (Fig. 3(c)). Further, to verify the piezoelectric nature of the samples, the converse piezoelectric coefficient Inline graphic is calculated from the unipolar strain (S)–field (E) curve at RT under the 6 kV cm−1 applied field and presented in Fig. 2(b). In the studied range, BST-9 offers the highest Inline graphic value of 1740 pm V−1 at RT; however, the TC is 58 °C. On the other hand, BST6 offers Inline graphic value of 1241 pm V−1 (Fig. 3(d)), and the TC is 82 °C. As the BST6 ceramic offers moderate Inline graphic and a reasonably high Curie temperature, it is used in further composite fabrication for sensors and NGs.

3.2. Morphological studies and Raman spectroscopic studies on the active layer

The BST6 ceramics are used to fabricate the PDMS–BST6 (P–B), PDMS–BST6–MWCNT (P–B–M) composite layers. The P–B–M and P–B/P–M/P–B composite layers are used to fabricate the PS-1 and PS-2 sensors as described in Sections 2.3 and 2.4. To verify the presence of all components inside the composite active layers, PXRD and Raman spectroscopy are employed. The evident broad peak at 2θ ∼ 12.8° in PXRD patterns of composite active layers is an indication of the presence of PDMS inside the active layer. Moreover, for the case of P–B and P–B–M layers, all the perovskite sharp peaks are present along with the broad peak at 2θ ∼ 12.8°, confirming the presence of BST6 in the PDMS matrix of the composite active layers (Fig. 4(a)). However, the presence of MWCNT is not evident from the PXRD patterns of the composite active layers. On the other hand, all the characteristic Raman modes of PDMS are evident in the cases P–M, P–B, and P–B–M samples. Moreover, the vibrational modes of MWCNT (1345 cm−1 D-band, 1584 cm−1 G-band, 1614 cm−1 D′-band) are also detected in P–M, P–B, and P–B–M composite layers, confirming the presence of MWCNT within the composite active layer.30 The dotted lines in Fig. 4(b) indicate the vibrational modes of MWCNT in the Raman spectroscopic data. All the detected Raman modes for the polymer and perovskite BST6 are listed in Table 1.

Fig. 4. (a) Room temperature PXRD patterns of different composite layers. The dashed line at 2θ ∼ 12.8° indicates the PDMS characteristic peak, and the dotted lines near 2θ ∼ 22° and 44° are indications of (022) and (100) planes of MWCNT. (b) Room temperature Raman Spectra of different composite layers: PDMS, P–M, P–B–M, P–B, and MWCNT. (c) FE-SEM micrographs of (I) PDMS–BST6 (P–B) composite, (II) PDMS–BST6–MWCNT (P–B–M) composite, and (III) layered composite, i.e., PDMS–MWCNT (P–M) sandwiched between two PDMS–BST6 (P–B) layers. (d) Raman map of PDMS–BST6–MWCNT (P–B–M) composite layer using G-band intensity.

Fig. 4

Open in a new tab

Table 1. Raman active modes of PDMS, BST6, and MWCNT.

Raman modes
Samples PDMS BST6 MWCNTs
489 cm−1 615 cm−1 707 cm−1 305 cm−1 512 cm−1 1345 cm−1 1584 cm−1 1614 cm−1
Stretch Si–O–Si bond Stretch asymmetric Si–C Stretch symmetric Si–C E + B1 mode Vibration of oxygen atoms in TiO6 Structural disorder Tangential shear mode Defects
P–M ✓ ✓ ✓ ✗ ✗ ✓ ✓ ✓
P–B ✓ ✓ ✓ ✓ ✓ ✗ ✗ ✗
P–B–M ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓
Open in a new tab

The morphology of the composite active layers is studied using FE-SEM. The uniform distribution of the constituent components in the composite active layers is observed in the micrographs (Fig. 4(c(I) and (II))); no noticeable agglomeration of MWCNTs is found for the composite layers P–B and P–B–M. For the layer-by-layer composition, the average layer thickness is measured from a cross-sectional FE-SEM image. The average layer thickness is 120 ± 15 µm, and the total layer thickness is 350 ± 15 µm, as calculated from the micrographs.

The uniform distribution of MWCNT throughout the polymer matrix is essential for uniform device performance and uniform thickness of composite layers, as observed in our previous work and addressed in the present work.32 Furthermore, the uniformity of the composite layer (P–B–M) is probed using Raman mapping. Here, an area map is created from a 1000-point spectrum collected at 1 µm spacing within the dashed rectangle of the optical image shown in Fig. 4(d). The collected spectrum is analyzed using the G-band intensity of MWCNTs. The bright-blue regions on the area map represent the presence of MWCNT, whereas the dark-blue regions indicate relatively low abundance. However, the presence of MWCNT throughout the matrix is evident from Raman spectroscopy. PXRD and Raman spectroscopic analyses indicate the presence of all constituent components within the composite layers. Additionally, FE-SEM and Raman mapping indicate a uniform distribution of constituent components throughout the layers, which is essential for optimal sensor and NG performance.

3.3. HNG device performance and working principle

To understand the working principles of the device and compare their output performance, the composite layers P–B, P–B–M, and P–M are used to fabricate three HNG and two PNG devices as described in Table 2 and Fig. 6. In the PNG configuration, the absence of any air gap between the active layer and the electrodes reduces the chance of contact electrification (CE) or triboelectric effect.33,34 On the other hand, in the hybrid nanogenerator (HNG) configuration, the presence of an air gap between the active layer and the top electrode enables contact electrification or triboelectric effect in addition to the piezoelectric effect (PE). Under an external force, the output voltage in the PNG configuration is mainly due to stress-induced polarization whereas in the HNG configuration, it is due to the combined effect of contact electrification and stress-generated polarization.35,36 The working mechanisms of PNG and HNG are schematically presented in Fig. 5(a) and (b). In the PNG configuration with P–B composite, the device generates a peak-to-peak output voltage (VPP) of 5.0 ± 1.5 V under open-circuit conditions and a peak-to-peak current (IPP) of 82.0 ± 20.0 nA under short-circuit conditions, both under a 20 N force. In the HNG configuration with P–B composite, VPP increases to 72.0 ± 10.0 V and IPP increases to 1.5 ± 0.4 µA, as presented in Fig. 5(b) and 6(b). Moreover, the inclusion of MWCNT in the P–B–M composite results in an increase in VPP to 22.0 ± 5.0 V in the PNG configuration (referred as PNGC) and 110.0 ± 10.0 V in the HNG configuration (referred as HNGC). The increase in VPP and IPP (Fig. 6) is attributed to improved conductivity in the active layer, resulting from the presence of MWCNTs in the polymer matrix.32,37 Furthermore, to isolate the triboelectric contribution, only PDMS-based devices are subjected to the same experimental conditions and produce a VPP of ∼50.0 ± 6.0 V. Additionally, to reinforce our understanding, PDMS–SrTiO3-based devices are fabricated and studied under the same conditions. Interestingly, no noticeable change is observed in the devices' output voltage (VPP) (Fig. S5). This is due to the centrosymmetric nature of SrTiO3 (Pm3̄m), which is non-piezoelectric.

Table 2. Summary of the different devices' output performancea.

Device Top electrode Spacer Composite active layer Bottom electrode Mechanism V PP (V) Tukey grouping
TNG Al ✓ PDMS (P) Al CE 50.0 ± 6.0 c
PNG Al — PDMS–BST6 (P–B) Al PE 5.0 ± 1.5 e
PNGC Al — PDMS–BST6–MWCNT (P–B–M) Al PE + ECT 22.0 ± 5.0 d
HNG Al ✓ PDMS–BST6 (P–B) Al CE + PE 72.0 ± 10.0 c
HNGC Al ✓ PDMS–BST6–MWCNT (P–B–M) Al CE + ECT 110.0 ± 10.0 b
HNGL Al ✓ P–B P–M P–B Al CE + PE + SP 135.0 ± 10.0 a
HNGL-B1 Al ✓ P–M P–B Al CE + PE + RE 122 ± 10.0 a
HNGL-B2 Al ✓ P–M P–B P–M Al CE + PE + RE 22.0 ± 2.0 d
HNGL-B3 Al ✓ P–M P P–M Al CE + PE 75.0 ± 10.0 c
Open in a new tab
a

CE = contact electrification, PE = piezoelectric effect, ECT = enhanced charge transfer, SP = self poling, RE = recombination.

Fig. 6. Comparison of different output characteristics of PNG, PNGC, HNG, HNGC, and HNGL devices (a) peak to peak voltage (VPP) in open circuit condition, (b) peak to peak current (IPP) in short circuit condition, (c) peak to peak power density (U) under 10 MΩ load connected in parallel condition. The inset shows the variation of U of HNGL under different RL, (d) comparison of VPP under different applied forces for different devices.

Fig. 6

Open in a new tab

Fig. 5. Working principle and open circuit voltage (VOC) of (a) PNG device (PDMS–BST6 in PNG configuration), (b) HNG device (PDMS–BST6 in HNG configuration), (c) HNGL device (PDMS–MWCNT layer sandwiched between PDMS–BST6 layers).

Fig. 5

Open in a new tab

The layered composite structure P–B/P–M/P–B in the HNG device (referred to as HNGL) configuration shows an improved output performance compared to the other HNG and PNG devices discussed above. The HNGL device produced a VPP of 135.0 ± 10.0 V (maximum VPP ∼ 150 V, Fig. S6) in open-circuit conditions and an IPP of 2.0 ± 1.0 µA in short-circuit conditions under a 20 N force. Additionally, the HNGL device exhibits a high output power density of 440 µW cm−2, attributed to enhanced charge transfer (ECT) via the formation of nano-electrical bridges of MWCNTs within the polymer matrix. To assess the usability of the HNG devices as an alternative energy-harvesting mechanism in day-to-day life, the devices are tested during common human activities, such as hand pressing, bending, and toe pressing. Under hand-pressing (∼100 N), the HNG devices produce a VPP of ∼204 V, and the HNGC devices produce a VPP of ∼268 V. The maximum output voltage was recorded for the HNGL devices, with a VPP of ∼328 V and U of ∼947 µW cm−2 against 1 MΩ of RL, as presented in Fig. 6(c), (d) and S7.

The large output response in the HNGL device results from contact electrification, the piezoelectric effect, and self-poling, which arise from the conducting P–M layer and recombination-blocking P–B layers.38,39 The non-recombined charges from the triboelectric effect in the P–B layer generates an internal electric field which aligns the dipoles within the P–B composite layer. The detailed mechanism of self-poling and the contributions of various effects, e.g., contact electrification, piezoelectric effect, enhanced charge transfer, and recombination, are explained and summarized in Table 2 and Fig. S8 of the SI. Notably, the generated voltage from HNGL powers 20 blue LEDs in series, thereby demonstrating its power-generation capability (Fig. S9). For comparison, the output responses of similar PNG/HNG devices are presented in Table 3. Based on the output performance of HNGC and HNGL, the composites were further used for sensor fabrication and pulse-sensing applications. The sensor based on the P–B–M composite is named PS-1, and the P–B/P–M/P–B-based sensor is named PS-2; both sensors are studied in the following sections.

Table 3. Comparison of different BaTiO3-based HNG devices.

Composite system Maximum voltage (V) Maximum current (µA) Power density (µW cm−2) Remarks Ref.
BaTiO3/PDMS 372 81 1210 Microstructured and poled 40
PI (polyimide)/CNT@(BaTiO3) 305 104 530 Electrospun 41
PDMS/BCHT (Ba0.85Ca0.15)(Ti0.90Hf0.10)O3 + MWCNT 325 — 500 Microstructured 42
PDMS–BaTiO3 280 5.6 8.5 Microstructured 43
PDMS–BaTiO3 212 53 200 — 44
PVDF–BaTiO3 150 1.5 — — 45
PDMS–BST6 328 3 947.3 Self-poled This work
Open in a new tab

3.4. Sensing performance testing

To demonstrate the sensing capabilities, the PS-1 and PS-2 sensors are fabricated as a wristband (wrapped around the radial artery) to minimise noise from hand movement (as shown in Fig. 7(a) and (b)). The output voltage against the pulse pressure is compared for the PS-1 and PS-2 sensors under the same conditions. The PS-1 and PS-2 pulse sensors produce responses of 4.0 ± 2.0 mV and 6.0 ± 2.0 mV, respectively, at the same systolic pressure, with a background noise level of ∼280 µV, as shown in Fig. 7(c) and (d). Further, the signal-to-noise ratio (SNR) is calculated for the sensors using the relation,

3.4. 2

where, Vsignal is the amplitude pulse peak and Vnoise is the rms noise as presented in Fig. 7(e).46 A maximum SNR of 12 dB is found for the PS-2 device, among all the studied sensors. Moreover, the response time (τ) for the PS-2 device is found to be 38 ms (Fig. 7(f)). The acceptable SNR value and the reasonable response time of PS-2 suggest that it can measure a signal in the presence of background noise and all the significant peaks: systolic peak (P1), the reflected wave peak (P2), and the diastolic peak (P3) of a human cardiac cycle. Further measurements and analysis are carried out using the PS-2 pulse sensor due to its superior SNR and output-voltage performance.

Fig. 7. (a) Animated image of the PS-2 sensor, (b) real image of working and data collection setup for PS-2 sensor, (c) collected pulse wave from PS-2 sensor, (d) sensitivity comparison of different pulse sensors, (e) Signal to Noise Ratio (SNR) measurement on PS-2 sensor, (f) response time (τ) measurement for PS-2 sensor.

Fig. 7

Open in a new tab

3.5. Evaluation of different cardiovascular parameters

Important cardiovascular parameters, such as heart rate, P–P interval, and root-mean-square successive differences (RMSSD), are derived from pulse waves collected by the PS-2 sensor, as shown in Fig. 8. The average bpm (beats per minute/heart rate) is 74 ± 2, calculated from data collected with the PS-2 device.

Fig. 8. (a) Presentation of the complete cardiac cycle with representation of P–P interval, (b) Fast Fourier Transformation (FFT) of the raw pulse wave collected for 120 s interval, (c) distribution of P–P intervals for different data sets collected on a 28 year-old male volunteer, (d) distribution of RMSSD values for the collected P–P intervals.

Fig. 8

Open in a new tab

Further, the pulse waves (120 s set) are analyzed using Fast Fourier Transform (FFT) to extract significant information about the cardiac cycle. The FFT analysis shows a significant peak in the normalized amplitude near 1.23 Hz, assigned as the systolic peak (P1), and two relatively lower-amplitude peaks near 2.4 Hz and 3.5 Hz, assigned as P2 and P3 (Fig. 8(b)). The bpm calculated from the FFT analysis (60 × 1.23 = 73.8–74) closely matches the previously calculated average bpm. The pulse wave sets are further studied to extract the time intervals ΔT = P1–P3 and the P–P interval, which are key indicators of human cardiac health (Fig. 8(b)).47,48 The P–P intervals are further processed to calculate the RMSSD values using the formula,

3.5. 3

where PP indicates the P–P interval. RMSSD is a key indicator of human heart rate variability (HRV).49 Further, the HRV is an indicator of human mental state (fatigue, stress, anxiety, etc.).50 For the calculation of the P–P interval, the collected pulse waves are divided into small 20 s data sets, and 10 such data sets are considered to observe the distribution of P–P intervals. Fig. 8(c) presents one of those 10 data sets, which suggests that the P–P intervals range from 0.69 s to 1.01 s, with a mean of 0.85 s, all well within the normal range.48,51 The RMSSD values calculated from the P–P intervals are 0.060 ± 0.005 s (Fig. 8(d)), which indicates a normal adult male with stable, good heart health.52

To assess the correlation between the fabricated sensor and the medical-standard ECG, we extracted the P–P interval from the piezoelectric–triboelectric hybrid sensor and the R–R interval from the ECG. The P–P intervals and R–R intervals form a linear correlation and a Pearson's correlation coefficient of 0.98, as presented in Fig. S10. Further, the sensitivity of the pulse sensors was tested in the relative humidity (RH) range of 45–60%. Under these conditions, the device exhibited a negligible effect in the sensitivity to moisture changes (Fig. S11). This high degree of correlation and the negligible response to changes in moisture underscore the sensor's ability to precisely track cardiac features and confirm its potential as a non-invasive alternative to traditional biopotential electrodes for long-term physiological monitoring.

To understand the frequency-domain response of the pulse waves, Short-Time Fourier Transforms (STFT) are used, and a spectrogram is computed over 75 s cycles. Two such spectrograms are presented and compared in Fig. 9(a) and (b). The pulse spectrogram of day 1 (Fig. 9(a)) shows a consistent energy band (yellow) in the range of 1.2 to 2.2 Hz, indicating that the subject's resting heart rate ranges between 72 and 120 bpm. However, a dark line appears near 47 s, likely due to the subject's physical movement. Fig. 9(b) presents the pulse spectrogram for day 100, and similar characteristics are observed in the spectrogram as in the pulse spectrogram of day 1. Sharp vertical lines corresponding to systolic rise and dicrotic notch are observed in both spectrograms. On the other hand, no significant indication of missed bets, rhythmic gaps (faded vertical lines), or ectopic beats is found in any of the cases, emphasizing a consistent heart profile over 100 days. While FFT yields harmonic values (indicating systolic and diastolic peaks), STFT and spectrograms can reveal possible rhythmic gaps and ectopic beats. All the cardiovascular parameters collected in this study, along with the methods used to derive them, are summarised in Table 4.

Fig. 9. Pulse spectrogram images for the collected pulse waves from the PS-2 sensor (a) on day 1, (b) on day 100, (c) comparison of similarity of raw pulse waves collected from PS-2 sensor on day 1 and day 100, (d) cumulative mismatch matrix obtained from DTW for day 1 to day 100 pulse waves.

Fig. 9

Open in a new tab

Table 4. Summary of cardiovascular parameter deduction.

Parameter name Value Analysis method Remarks
Heart rate (bpm) 74 FFT Within normal range
P–P interval 0.85 s Peak detection (from the FFT filter) Within normal range
ΔT = P1–P3 0.28 s FFT filter Within normal range
RMSSD 0.060 s FFT filter Good heart health
Missing beats and ectopic beats Not found STFT No arrhythmia
Similarity score 0.1093 DTW Excellent heart health
Open in a new tab

3.6. Cyclability performance of the sensor

To assess the data reproducibility and cyclability of the PS-2 sensor, we evaluated its performance over a longer period (more than 100 days). The Dynamic Time Warping (DTW) algorithm is applied to the pulse wave sets from day 1 and day 100 to assess the similarity between them. Fig. 9(c) shows the similarity between raw pulse wave sets collected on day 1 and day 100 under normal conditions. The cumulative mismatch matrix (Fig. 9(d)) shows the morphological similarity between the pulse waves of day 1 and day 100. The dark pink valley along the diagonal suggests similarities among signatures, such as the systolic peak and dicrotic notch. The similarity score of 0.1093 indicates a high degree of similarity between the two pulse waves collected on day 1 and day 100, suggesting the sensor's reproducibility and cyclability, as well as the test subject's consistently healthy heart. The smooth transition of colours (from dark purple to yellow) indicates a steady, low-cost accumulation of mismatch. This visual supports a similarity score of 0.193, indicating high physiological stability. The spectrogram analysis complements the similarity score and provides significant information about the collected signal, helping differentiate unwanted interference within pulse-wave signals. The DTW analysis confirms a high degree of morphological similarity between the initial (day 1) and day 100 waveforms, demonstrating the sensor's repeatable response. Concurrently, STFT analyses revealed no evidence of signal damping or frequency-domain degradation, indicating that the hybrid composite maintains its mechanical and electrical integrity without significant fatigue over extended cycling.

4. Conclusions

In conclusion, we have developed a pulse sensor based on a piezoelectric–triboelectric composite for monitoring cardiovascular health indices. The high performance of the composite originates from the excellent electromechanical response of piezoelectric ceramic BaTi0.94Sn0.06O3, with Inline graphic. The HNG device, with a self-poling PDMS–MWCNT layer incorporated into the PDMS–BST6/PDMS–MWCNT/PDMS–BST6 structure, generates a large output voltage and power density of 328 V and 947.3 µW cm−2, respectively. The pulse sensor fabricated using the layered structure exhibits an output voltage of 6 mV, a response time of 38 ms, and an SNR of 12 dB in response to a radial arterial pulse. The large output response of the HNG device and the optimum sensitivity of the pulse sensor result from the self-poling PDMS–MWCNT interfacial layer, which enhances charge collection and suppresses recombination. The pulse sensor, despite a low output voltage of 6 mV, records an arterial pulse waveform and enables accurate measurement of heart rate, P–P intervals, and RMSSD. The Short-Time Fourier Transform (STFT) and Dynamic Time Warping (DTW) analyses suggest sensor cyclability and also the test subject's heart condition over a 100 day period. The present study highlights the potential of the composite layered structure for both energy harvesting and wearable bio-sensing technologies. However, the mV-level output signal is insufficient for direct wireless transmission to a portable device without significant amplification. Further studies will focus on developing a wearable pulse sensor with enhanced output voltage and integrating it with a low-power signal-conditioning circuit and wireless transmission modules to enable real-time, continuous monitoring of cardiovascular health for effective risk assessment.

Author contributions

S. K. conceptualized the project, performed the synthesis, structural, and morphological analysis of BST6 piezoelectric ceramic, designed the device characterization setup, fabricated and characterized devices, plotted figures, and drafted the manuscript. SKJ and SS assisted with synthesis and fabrication. PS assisted in formatting the draft. J. N. and G. K. P. performed the Raman measurements and analysis and worked on the draft. P. M. conceptualized the project and assisted in the manuscript. All authors contributed to the project as it developed.

Conflicts of interest

The authors declare no conflict of interest.

Supplementary Material

RA-016-D6RA02511K-s001
Download video file (27.5MB, mp4)
RA-016-D6RA02511K-s002

Acknowledgments

SK and PM acknowledge SRM University-AP, Andhra Pradesh (Grant No. SRMAP/URG/CF/2023-24/045, SRMAP/URG/GENERAL/2024-25/015 and SRMAP/URG/GENERAL-D/2025-26/019) and DST-FIST (Grant No. SR/FST/PS-I/2024/306), for the financial support and laboratory facilities. SK and PM thank Dr Pradyut K. Sanki for helpful discussion and research facilities. JN and GKP acknowledge the use of the Micro-Raman facility at the central research facility (CRF) of KIIT Deemed to be University, Bhubaneswar, and thank the Science and Engineering Board for financial support (CRG/2020/006190).

Data availability

The data used or generated in the article will be available on request.

Supplementary information (SI) is available. See DOI: https://doi.org/10.1039/d6ra02511k.

References

  1. Di Cesare M. Perel P. Taylor S. Kabudula C. Bixby H. Gaziano T. A. McGhie D. V. Mwangi J. Pervan B. Narula J. The heart of the world. Global Heart. 2024;19:11. doi: 10.5334/gh.1288. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Lin Q. Huang J. Yang J. Huang Y. Zhang Y. Wang Y. Zhang J. Wang Y. Yuan L. Cai M. Highly sensitive flexible iontronic pressure sensor for fingertip pulse monitoring. Adv. Healthcare Mater. 2020;9:2001023. doi: 10.1002/adhm.202001023. [DOI] [PubMed] [Google Scholar]
  3. Hou S. Chen X. Assi D. S. Feng Y. Wong C.-K. Tian J. Li J. Yu X. Guan B. Zhu X. Tactile-transparent wearable sensor for clinician-friendly pulse wave velocity monitoring and cardiovascular risk profiling. ACS Nano. 2025;19:32822–32835. doi: 10.1021/acsnano.5c11375. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Sun Y. Mao J. Cao L. Zheng X. Meng Q. Zhao S. Wang Z. Zhang Y. Cui D. Zheng H. Intelligent cardiovascular disease diagnosis system combined piezoelectric nanogenerator based on 2D Bi2O2Se with deep learning technique. Nano Energy. 2024;128:109878. doi: 10.1016/j.nanoen.2024.109878. [DOI] [Google Scholar]
  5. Yao C. Sun T. Huang S. He M. Liang B. Shen Z. Huang X. Liu Z. Wang H. Liu F. Personalized machine learning-coupled nanopillar triboelectric pulse sensor for cuffless blood pressure continuous monitoring. ACS Nano. 2023;17:24242–24258. doi: 10.1021/acsnano.3c09766. [DOI] [PubMed] [Google Scholar]
  6. Deng W. Zhou Y. Libanori A. Chen G. Yang W. Chen J. Piezoelectric nanogenerators for personalized healthcare. Chem. Soc. Rev. 2022;51:3380–3435. doi: 10.1039/D1CS00858G. [DOI] [PubMed] [Google Scholar]
  7. Yu Y. Nassar J. Xu C. Min J. Yang Y. Dai A. Doshi R. Huang A. Song Y. Gehlhar R. Biofuel-powered soft electronic skin with multiplexed and wireless sensing for human-machine interfaces. Sci. Robot. 2020;5:eaaz7946. doi: 10.1126/scirobotics.aaz7946. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Meng K. Xiao X. Wei W. Chen G. Nashalian A. Shen S. Xiao X. Chen J. Wearable pressure sensors for pulse wave monitoring. Adv. Mater. 2022;34:2109357. doi: 10.1002/adma.202109357. [DOI] [PubMed] [Google Scholar]
  9. Yang J. C. Kim J.-O. Oh J. Kwon S. Y. Sim J. Y. Kim D. W. Choi H. B. Park S. Microstructured porous pyramid-based ultrahigh sensitive pressure sensor insensitive to strain and temperature. ACS Appl. Mater. Interfaces. 2019;11:19472–19480. doi: 10.1021/acsami.9b03261. [DOI] [PubMed] [Google Scholar]
  10. De Fazio R. Mastronardi V. M. De Vittorio M. Visconti P. Wearable sensors and smart devices to monitor rehabilitation parameters and sports performance: an overview. Sensors. 2023;23:1856. doi: 10.3390/s23041856. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Zhu B. Ling Y. Yap L. W. Yang M. Lin F. Gong S. Wang Y. An T. Zhao Y. Cheng W. Hierarchically structured vertical gold nanowire array-based wearable pressure sensors for wireless health monitoring. ACS Appl. Mater. Interfaces. 2019;11:29014–29021. doi: 10.1021/acsami.9b06260. [DOI] [PubMed] [Google Scholar]
  12. Wang Z. L. Song J. Piezoelectric nanogenerators based on zinc oxide nanowire arrays. Science. 2006;312:242–246. doi: 10.1126/science.1124005. [DOI] [PubMed] [Google Scholar]
  13. Xu Q. Wen J. Qin Y. Development and outlook of high output piezoelectric nanogenerators. Nano Energy. 2021;86:106080. doi: 10.1016/j.nanoen.2021.106080. [DOI] [Google Scholar]
  14. Park D. Y. Joe D. J. Kim D. H. Park H. Han J. H. Jeong C. K. Park H. Park J. G. Joung B. Lee K. J. Self-powered real-time arterial pulse monitoring using ultrathin epidermal piezoelectric sensors. Adv. Mater. 2017;29:1702308. doi: 10.1002/adma.201702308. [DOI] [PubMed] [Google Scholar]
  15. He Q. Wang H. Zhang J. Ghahremani Arekhloo N. Karagiorgis X. Yalagala B. P. Skabara P. J. Heidari H. Zeze D. A. Hosseini E. S. A Wearable and Highly Sensitive PVDF–TrFE–BaTiO3 Piezoelectric Sensor for Wireless Monitoring of Arterial Signal. ACS Appl. Electron. Mater. 2025;7:7562–7571. doi: 10.1021/acsaelm.5c00548. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Yang T. Pan H. Tian G. Zhang B. Xiong D. Gao Y. Yan C. Chu X. Chen N. Zhong S. Hierarchically structured PVDF/ZnO core-shell nanofibers for self-powered physiological monitoring electronics. Nano Energy. 2020;72:104706. doi: 10.1016/j.nanoen.2020.104706. [DOI] [Google Scholar]
  17. Yi F. Zhang Z. Kang Z. Liao Q. Zhang Y. Recent advances in triboelectric nanogenerator-based health monitoring. Adv. Funct. Mater. 2019;29:1808849. doi: 10.1002/adfm.201808849. [DOI] [Google Scholar]
  18. Venugopal K. Panchatcharam P. Chandrasekhar A. Shanmugasundaram V. Comprehensive review on triboelectric nanogenerator based wrist pulse measurement: Sensor fabrication and diagnosis of arterial pressure. ACS Sens. 2021;6:1681–1694. doi: 10.1021/acssensors.0c02324. [DOI] [PubMed] [Google Scholar]
  19. Zhang J. He Y. Boyer C. Kalantar-Zadeh K. Peng S. Chu D. Wang C. H. Recent developments of hybrid piezo–triboelectric nanogenerators for flexible sensors and energy harvesters. Nanoscale Adv. 2021;3:5465–5486. doi: 10.1039/D1NA00501D. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Hu Y. Wang F. Ma Y. Ma S. Wang L. Recent Advances in Polyvinylidene Fluoride with Multifunctional Properties in Nanogenerators. Small. 2025;21:2412476. doi: 10.1002/smll.202412476. [DOI] [PubMed] [Google Scholar]
  21. Li X. Lin Z.-H. Cheng G. Wen X. Liu Y. Niu S. Wang Z. L. 3D Fiber-Based Hybrid Nanogenerator for Energy Harvesting and as a Self-Powered Pressure Sensor. ACS Nano. 2014;8:10674–10681. doi: 10.1021/nn504243j. [DOI] [PubMed] [Google Scholar]
  22. Hussain S. Z. Singh V. P. Rahman M. Onuralp D. B. Ordu M. Hybridized piezoelectric–triboelectric effects for self-powered sensing in thermally drawn nanocomposite fibers. J. Mater. Chem. A. 2026 doi: 10.1039/D6TA01231K. [DOI] [Google Scholar]
  23. Kim D. H. Dudem B. Yu J. S. High-performance flexible piezoelectric-assisted triboelectric hybrid nanogenerator via polydimethylsiloxane-encapsulated nanoflower-like ZnO composite films for scavenging energy from daily human activities. ACS Sustainable Chem. Eng. 2018;6:8525–8535. doi: 10.1021/acssuschemeng.8b00834. [DOI] [Google Scholar]
  24. Maity S. Sasmal A. Kar E. Sen S. Morphotropic phase boundary-assisted lead-free BaTiO3/PDMS composite-based hybrid energy harvester: a portable power source for wireless power transmission. ACS Appl. Energy Mater. 2023;6:7052–7064. doi: 10.1021/acsaem.3c00635. [DOI] [Google Scholar]
  25. Liu W. Wang J. Ke X. Li S. Large piezoelectric performance of Sn doped BaTiO3 ceramics deviating from quadruple point. J. Alloys Compd. 2017;712:1–6. doi: 10.1016/j.jallcom.2017.04.013. [DOI] [Google Scholar]
  26. Nagamalleswari K. Tulasirao P. Mandal P. A measurement setup for characterization of temperature dependence of impedance, dielectric permittivity, and pyroelectric current under a controlled environment. J. Electron. Mater. 2023;52:1625–1632. doi: 10.1007/s11664-022-09875-2. [DOI] [Google Scholar]
  27. Jin L. Qiao J. Hou L. Tian Y. Hu Q. Wang L. Lu X. Zhang L. Du H. Wei X. A strategy for obtaining high electrostrictive properties and its application in barium stannate titanate lead-free ferroelectrics. Ceram. Int. 2018;44:21816–21824. doi: 10.1016/j.ceramint.2018.08.286. [DOI] [Google Scholar]
  28. Ansari M. A. Sreenivas K. Effects of disorder activated scattering and defect-induced phase on the ferroelectric properties of BaSnxTi1-xO3 (0≤ x≤ 0.28) ceramics. Ceram. Int. 2019;45:20738–20749. doi: 10.1016/j.ceramint.2019.07.058. [DOI] [Google Scholar]
  29. Wang D. Fan Z. Rao G. Wang G. Liu Y. Yuan C. Ma T. Li D. Tan X. Lu Z. Ultrahigh piezoelectricity in lead-free piezoceramics by synergistic design. Nano Energy. 2020;76:104944. doi: 10.1016/j.nanoen.2020.104944. [DOI] [Google Scholar]
  30. Deluca M. Stoleriu L. Curecheriu L. P. Horchidan N. Ianculescu A. C. Galassi C. Mitoseriu L. High-field dielectric properties and Raman spectroscopic investigation of the ferroelectric-to-relaxor crossover in BaSnxTi1− xO3 ceramics. J. Appl. Phys. 2012;111:084102. doi: 10.1063/1.3703672. [DOI] [Google Scholar]
  31. Mayamae J. Vittayakorn W. Muanghlua R. Woramongkolchai S. Vittayakorn N. Effect of Sn content on the dielectric and piezoelectric properties of the ternary system (0.975-y) BaTiO3–0.025 SrTiO3–y BaSnO3. J. Mater. Sci. 2017;52:6928–6936. doi: 10.1007/s10853-017-0906-z. [DOI] [Google Scholar]
  32. Katragadda N. Kumar S. Jacob S. K. Sanki P. K. Nanda J. Pradhan G. K. Mandal P. Carbon Nanotube-Assisted Device Performance Improvement in Flexible Piezoceramic–Polymer Hybrid Nanogenerators. ACS Appl. Electron. Mater. 2023;5:6938–6946. doi: 10.1021/acsaelm.3c01341. [DOI] [Google Scholar]
  33. Cheng T. Shao J. Wang Z. L. Triboelectric nanogenerators. Nat. Rev. Methods Primers. 2023;3:39. doi: 10.1038/s43586-023-00220-3. [DOI] [Google Scholar]
  34. Wang Z. L. Wang A. C. On the origin of contact-electrification. Mater. Today. 2019;30:34–51. doi: 10.1016/j.mattod.2019.05.016. [DOI] [Google Scholar]
  35. Sriphan S. Charoonsuk T. Maluangnont T. Vittayakorn N. High-performance hybridized composited-based piezoelectric and triboelectric nanogenerators based on BaTiO3/PDMS composite film modified with Ti0. 8O2 nanosheets and silver nanopowders cofillers. ACS Appl. Energy Mater. 2019;2:3840–3850. doi: 10.1021/acsaem.9b00513. [DOI] [Google Scholar]
  36. Li H. Su L. Kuang S. Fan Y. Wu Y. Wang Z. L. Zhu G. Multilayered flexible nanocomposite for hybrid nanogenerator enabled by conjunction of piezoelectricity and triboelectricity. Nano Res. 2017;10:785–793. doi: 10.1007/s12274-016-1331-3. [DOI] [Google Scholar]
  37. Yan J. Jeong Y. G. Roles of carbon nanotube and BaTiO3 nanofiber in the electrical, dielectric and piezoelectric properties of flexible nanocomposite generators. Compos. Sci. Technol. 2017;144:1–10. doi: 10.1016/j.compscitech.2017.03.015. [DOI] [Google Scholar]
  38. Yaqoob U. Habibur R. M. Sheeraz M. Kim H. C. Realization of self-poled, high performance, flexible piezoelectric energy harvester by employing PDMS-rGO as sandwich layer between P (VDF-TrFE)-PMN-PT composite sheets. Composites, Part B. 2019;159:259–268. doi: 10.1016/j.compositesb.2018.09.102. [DOI] [Google Scholar]
  39. Yaqoob U. Uddin A. I. Chung G.-S. A novel tri-layer flexible piezoelectric nanogenerator based on surface-modified graphene and PVDF-BaTiO3 nanocomposites. Appl. Surf. Sci. 2017;405:420–426. doi: 10.1016/j.apsusc.2017.01.314. [DOI] [Google Scholar]
  40. Seong J. Bak B.-U. Lee D. Jin J. Kim J. Tribo-piezoelectric synergistic BaTiO3/PDMS micropyramidal structure for high-performance energy harvester and high-sensitivity tactile sensing. Nano Energy. 2024;122:109264. doi: 10.1016/j.nanoen.2024.109264. [DOI] [Google Scholar]
  41. Yan J. Wang K. Wang Y. Lv M. Yang G. Polyimide nanofiber-based triboelectric nanogenerators using piezoelectric carbon nanotube@ barium titanate nanoparticles. ACS Appl. Nano Mater. 2024;7:13156–13165. doi: 10.1021/acsanm.4c01685. [DOI] [Google Scholar]
  42. Sasmal A. Patra A. Maiti P. Sahu B. Krahne R. Arockiarajan A. Microstructuring of conductivity tuned piezoelectric polydimethylsiloxane/(Ba0. 85Ca0. 15)(Ti0. 90Hf0. 10)O3 composite for hybrid mechanical energy harvesting. Polym. Compos. 2025;46:11387. doi: 10.1002/pc.29692. [DOI] [Google Scholar]
  43. Sahu M. Vivekananthan V. Hajra S. Khatua D. K. Kim S.-J. Porosity modulated piezo-triboelectric hybridized nanogenerator for sensing small energy impacts. Appl. Mater. Today. 2021;22:100900. doi: 10.1016/j.apmt.2020.100900. [DOI] [Google Scholar]
  44. Li M. Li Z. Yaseen S. Cui N. Li P. Sun J. Liu J. Lü X. Gu L. High performance piezoelectric nanogenerator based on aligned single crystal BaTiO3 nanowires for self-powered applications. Chem. Eng. J. 2025:163834. doi: 10.1016/j.cej.2025.163834. [DOI] [Google Scholar]
  45. Zhao Y. Liao Q. Zhang G. Zhang Z. Liang Q. Liao X. Zhang Y. High output piezoelectric nanocomposite generators composed of oriented BaTiO3 NPs@ PVDF. Nano Energy. 2015;11:719–727. doi: 10.1016/j.nanoen.2014.11.061. [DOI] [Google Scholar]
  46. Charlton P. H. Marozas V. Mejía-Mejía E. Kyriacou P. A. Mant J. Determinants of photoplethysmography signal quality at the wrist. PLOS Digit. Health. 2025;4:e0000585. doi: 10.1371/journal.pdig.0000585. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Wu Y. Li Y. Wang X. Guo P. Luo H. Li X. Zhu X. Yang J. Intelligent ring for dynamic pulse wave monitoring using fine tubular triboelectric nanogenerators. Nano Energy. 2025;134:110581. doi: 10.1016/j.nanoen.2024.110581. [DOI] [Google Scholar]
  48. Tian G. Deng W. Yang T. Zhang J. Xu T. Xiong D. Lan B. Wang S. Sun Y. Ao Y. Hierarchical piezoelectric composites for noninvasive continuous cardiovascular monitoring. Adv. Mater. 2024;36:2313612. doi: 10.1002/adma.202313612. [DOI] [PubMed] [Google Scholar]
  49. Beydoun N. Quraishi S. A. Tolman E. Battarjee W. Weintraub A. Cobey F. Hong E. Association of preprocedural ultrashort-term heart rate variability with clinical outcomes after transcatheter aortic valve replacement: A nested, case-control, pilot study. Ann. Card Anaesth. 2022;25:318–322. doi: 10.4103/aca.aca_11_22. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Costa M. D. Davis R. B. Goldberger A. L. Heart rate fragmentation: a new approach to the analysis of cardiac interbeat interval dynamics. Front. Physiol. 2017;8:255. doi: 10.3389/fphys.2017.00255. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Yi Z. Huang J. Liu Z. Liu J. Yang B. Portable, wireless wearable piezoelectric arterial pulse monitoring system based on near-field communication approach. IEEE Electron Device Lett. 2019;41:183–186. [Google Scholar]
  52. Berntson G. G. Lozano D. L. Chen Y. J. Filter properties of root mean square successive difference (RMSSD) for heart rate. Psychophysiol. 2005;42:246–252. doi: 10.1111/j.1469-8986.2005.00277.x. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

RA-016-D6RA02511K-s001
Download video file (27.5MB, mp4)
RA-016-D6RA02511K-s002
RA-016-D6RA02511K-s002.pdf (1.8MB, pdf)

Data Availability Statement

The data used or generated in the article will be available on request.

Supplementary information (SI) is available. See DOI: https://doi.org/10.1039/d6ra02511k.

Articles from RSC Advances are provided here courtesy of Royal Society of Chemistry

RESOURCES