An interface-engineered HMS/PANI thermoelectric device for 3D printer waste heat harvesting

Abstract

The development of flexible, high-performance thermoelectric materials is necessary due to the growing demand for sustainable energy conversion from waste heat sources. In this work, we present higher manganese silicide/polyaniline (HMS/PANI) composite thin films that exhibit a hierarchical interface design, which partially decouples the typically interrelated thermoelectric parameters for waste heat recovery from working 3D printers. Structural characterization studies – XRD and SEM – confirm the well-dispersed HMS network with good interfacial quality. Through systematic composition optimization, we achieved enhanced electrical conductivity, 438 S m⁻¹, and a simultaneous increase in the Seebeck coefficient of 72 μV K⁻¹ at 80% HMS content. This enhancement results from hierarchical composite engineering, achieved by percolation-driven charge transfer at HMS/PANI interfaces and energy filtering effects, which also increase carrier concentration. The optimized composition exhibits an enhanced power factor of 2.3 μW (mK²)⁻¹. Phonon engineering suppresses the lattice thermal conductivity relative to pure PANI, thereby maximizing the thermoelectric performance of composites. I–V characterization confirms the ohmic transport behaviour across all compositions. Device output characterization demonstrates a maximum power generation of 0.33 nW at ΔT = 70 K. Real-time application demonstration on an operating 3D printer reveals stable performance by generating 6–8 mV output, validating its functionality for autonomous IoT energy harvesting systems. This work establishes HMS/PANI composites as a flexible thermoelectric material where sophisticated composite design transcends individual phase limitations, paving the way for practical thermoelectric energy harvesting from diverse waste heat sources.

---

1. Introduction

Thermal energy represents one of the most abundant yet underutilized energy resources in the world. Industries, electronic gadgets, computational facilities, and manufacturing systems continuously generate enormous quantities of waste heat as an inevitable byproduct of their operations. Manufacturing processes alone account for approximately one-third of global energy consumption, with over 50% of industrial energy being dissipated as waste heat into the environment. This substantial energy source presents a critical opportunity for sustainable energy conversion.^1–3 The current recovery technologies have limitations in efficiency, flexibility, and stability.^4–6 Among the emerging manufacturing technologies, three-dimensional (3D) printing has rapidly proliferated across diverse applications, with the global 3D printing market expanding at a notable rate. During the 3D printing process, particularly in fused deposition modeling (FDM) and stereolithography techniques, the heated print beds maintain temperatures of 60–80 °C continuously during operation.^7,8 This substantial waste heat is generated and dissipates into the surrounding environment without being put to productive use. This represents an ideal, yet overlooked, opportunity for real-world demonstration of waste heat recovery technology, offering accessible and practical thermal gradients, characteristic of low-grade heat sources that would benefit numerous IoT energy harvesting systems and autonomous devices.^7,9,10

The most prominent way to convert waste heat into energy is through thermoelectric generators (TEGs). Thermoelectric generators convert thermal energy directly into electrical energy through the Seebeck effect. This offers distinct advantages such as solid-state operation without any moving parts, high reliability, silent operation, low maintenance, and environmental compatibility.^7,11,12 These inherent advantages make thermoelectric technology an ideal waste heat harvesting technology in diverse operational environments ranging from wearable devices to industrial facilities.^2,13 However, the widespread implementation of TEGs remains constrained due to the limited thermoelectric performance of currently available materials. The commercial thermoelectric material based on bismuth telluride (Bi₂Te₃) has limitations concerning its toxicity, which restricts environmental acceptability, high material cost, and brittleness, preventing its integration into flexible architectures.^11,14–16 These factors have motivated intensive research toward alternative thermoelectric materials capable of overcoming these limitations.

Among the thermoelectric materials, higher manganese silicide (HMS) is a prominent thermoelectric material for mid-temperature applications. It has the advantages of Earth-abundant materials as raw materials, low cost, and environmental acceptability.¹⁷ Incorporation of inorganic thermoelectric fillers into polymer matrices represents a promising strategy to overcome individual material limitations through synergistic property engineering. Also, conducting polymers offer reliable sustainability, tunable electrical properties, and ease of processing; they are becoming increasingly critical for next-generation, scalable energy harvesting devices and autonomous sensor networks.¹⁸ Conducting polymers, particularly polyaniline (PANI), have emerged as promising candidates for next-generation thermoelectric applications. They offer a unique combination of solution processability, enabling diverse fabrication methods, and mechanical flexibility suitable for wearable and textile integration.¹⁹ They also have the advantages of low cost, environmental stability and limited toxicity. Numerous approaches for enhancing polyaniline's thermoelectric limits through structural and compositional engineering have been thoroughly investigated in recent years. For instance, incorporating inorganic fillers like Bi₂Te₃ into electrospun PANI, along with polyvinylpyrrolidone (PVP) networks, improves the figure of merit by decoupling thermal and electrical transport.²⁰ Other successful approaches include developing more ternary nanocomposites such as PANI/WS₂/CNTs²¹ and embedding metal sulfides or multinary nanofillers (e.g., NiS or Mg_0.99Cu_0.01Ag_0.97Sb_0.99) to boost the Seebeck coefficient via interfacial energy filtering.^22,23 However, many of these systems rely on complex architectures or face thermal stability limits at elevated temperatures. To address this, we introduce higher manganese silicide as an attractive filler material for PANI composites. This combination is favorable for imparting a semi-metallic character with high intrinsic carrier density, enabling substantial electrical conductivity, environmental stability, non-toxicity, and thermal stability.¹⁴ Recent developments in flexible thermoelectrics have shown that interrelated thermoelectric parameters can be significantly decoupled by using interfacial regulation techniques and logically designed organic–inorganic boundaries.²⁴ When HMS fillers are strategically incorporated into PANI matrices, multiple property enhancements can be achieved.

Despite extensive thermoelectric material research, systematic validation in practical real-world operating environments remains limited in the literature. Most thermoelectric studies employ laboratory–controlled temperature differentials under stable thermal conditions. The thermal environment of operating 3D printers presents an ideal real-world demonstration, as continuous heat generation from the operating bed maintains a stable thermal gradient with room temperature. Successful device demonstration under actual 3D printer operation beyond laboratory setups validates the reproducibility, durability, and practical applicability of the thermoelectric system for IoT energy harvesting systems, wearable health monitoring systems, and autonomous low-power electronics.

This work presents a comprehensive investigation of an optimized HMS/PANI composite thin film for high-performance thermoelectric waste heat harvesting. The study systematically progressed from the characterization of fundamental materials through device fabrication to real-time application demonstration on 3D printer equipment. We have synthesized and characterized HMS/PANI composites across a composition range from 0% to 80% HMS content. We have conducted structural and morphological analysis through XRD and SEM techniques. Comprehensive transport measurements were made using a Hall effect measurement system. The interplay between carrier concentration, mobility, electrical conductivity, and thermopower was explored to understand how percolation and interface engineering jointly optimize the electrical conductivity and Seebeck coefficient. A flexible thermoelectric device was fabricated from the optimized composite composition and characterized for power output under controlled thermal gradients. The practical device function was demonstrated in an operating 3D printer environment, to validate the real-time applicability for waste heat energy conversion. This demonstration paves the way for the practical utility of TEGs for battery-free IoT energy harvesting systems powered by ambient waste heat without requiring laboratory infrastructure or artificial thermal control.

2. Experimental

2.1 Materials required

Aniline, extra pure, 99.5% was purchased from Sisco Research Laboratories Pvt. Ltd. Potassium dichromate, 99% purity, from Sigma Aldrich, was used as an oxidant. Conc. hydrochloric acid was purchased from Sisco Research Laboratories Pvt. Ltd. Ethyl cellulose powder from Sigma Aldrich was used as a binder. Ethyl acetate was used as the solvent for paste making. Manganese powder (metal basis) and silicon lumps (0.1–2.5 cm) from Alfa Aesar were used to synthesize higher manganese silicide (HMS) powder through a solid state reaction. Cotton fabric was used as the flexible substrate. Silver paste and copper tape were used for electrodes and interconnects.

2.2 Synthesis of higher manganese silicide

The higher manganese silicide (HMS) used in this study was synthesized via a solid-state reaction method. Stoichiometric amounts of elemental Mn and Si powders with the nominal composition of MnSi_1.7 were weighed and mechanically ground. The elemental powder mixed homogeneously was pelletized using a hydraulic press with a 10 mm diameter under 4 tons of pressure. The compacted pellet was sealed in an evacuated quartz tube. The vacuum sealed quartz tube with the pellet was then heat-treated in a furnace at 900 °C for 24 hours. After cooling to room temperature, the heat-treated pellet was subsequently ground into a fine powder and sieved using 400 mesh sieves. The obtained powder was characterized by means of XRD and used for incorporation into the PANI matrix.

2.3 Synthesis of PANI and HMS-PANI composites

PANI and HMS-PANI composites were prepared via in situ oxidative polymerization of aniline in acidic aqueous solution using potassium dichromate (K₂Cr₂O₇) as the oxidant. Briefly, the required amounts of aniline (9 ml), conc. HCl (10 ml) and deionized water (10 ml) were taken in a round-bottom flask. HMS powder was dispersed in the monomer solution at the desired weight fractions (10, 30, 50, and 80 wt% HMS relative to PANI's mass). The solution was stirring for 30 min at a temperature of 0 °C–10 °C under a N₂ atmosphere as shown in Fig. 1. 25 ml of 0.33 M potassium dichromate (K₂Cr₂O₇) was added dropwise to the solution. The polymerization was allowed to proceed for 4 h. The resulting composite was filtered and washed (deionized water and ethanol) to remove unreacted monomers and oxidative impurities and dried at room temperature. Pure PANI was synthesized under similar conditions without HMS addition (Fig. 1a). The use of K₂Cr₂O₇ as an oxidant for PANI synthesis has been reported²⁵ and provides a straightforward route to conductive emeraldine salts under acidic conditions.

Fig. 1 (a) Synthesis of HMS/PANI composites by the in situ polymerization method, (b) composite paste making technique, and (c) thin film coating of HMS/PANI composites on fabric using the blade coating technique.

2.4 Film fabrication

A composite paste was prepared by re-dispersing the dried composite powder in an appropriate binder system to achieve blade-coatable viscosity. In this work, ethyl cellulose and ethyl acetate were used as a binder system. The required amounts of ethyl cellulose, ethyl acetate (1 : 1), and composite powder were taken in a paste maker jar and sonicated for 15 min. The jar was placed in the paste maker and allowed to spin for 1 h. The resulting composite paste consistency was achieved as shown in Fig. 1b. Cotton fabric substrates were cleaned and fixed to a flat support. Films were deposited using the blade-coating technique (both sides) and dried at room temperature for 15 minutes. To ensure reproducibility and standardize the thermoelectric performance metrics, the material loading of the composite film was carefully controlled. The geometric dimensions of the active area were measured, and the flexible substrates were weighed using a high-precision microbalance before and after deposition. The optimized HMS/PANI film exhibited an average uniform thickness of 400 µm and an active material mass loading of 30.71 mg cm⁻². The resulting composite-coated fabrics were visually uniform, as shown in Fig. 1c.

2.5 Characterization

Powder X-ray diffraction (XRD) patterns were collected on a diffractometer from Bruker, D8 Advance, with Cu Kα radiation (λ = 1.5406 Å). The thermal stability of the composites was analysed by thermogravimetric (TGA) analysis using a TA Instruments system. The morphology of the thin film was studied to confirm the coating coverage of threads using a scanning electron microscope (SEM) from EVO 18 Research. X-ray photoelectron spectroscopy (XPS) analysis was carried out using a ULVAC-PHI VersaProbe 4. Hall effect measurements were performed to observe transport trends using an Ecopia measurement system in a standard Van der Pauw configuration. The measurements were conducted at room temperature under an applied static magnetic field of 0.51 T. Due to known limitations and contact-resistance artifacts in evaluating the apparent Hall mobility of semi-disordered polymers, the bulk electrical conductivity of the composite films was rigorously evaluated and corroborated using a standard four-probe measurement setup. The Seebeck coefficient of the thin films was measured using a home-built setup shown in Fig. S4.²⁶ Briefly, the sample was suspended between two temperature-controlled blocks. A steady-state temperature gradient (ΔT) was established across the sample using a voltage supply, and the exact temperatures at both ends were monitored using fine-gauge K-type thermocouples. The corresponding thermoelectric voltage (ΔV) across the film was simultaneously recorded using a high-precision digital logger (HIOKI-LR8431-20). The thermal conductivity of the thin films was measured via the MTPS approach using an MTPS H1142 from C-Therm Technologies. I–V characterization of the device was conducted using a Keithley 2450. To ensure the reliability and reproducibility of the reported thermoelectric properties, measurements were repeated on samples for each composition. The plotted data points represent the average values, with error bars indicating the standard deviation.

3. Results and discussion

3.1 XRD analysis

Fig. 2a shows the XRD patterns of pure HMS, pure PANI, and HMS/PANI composites containing 10%, 30%, 50% and 80% HMS by weight. The characteristic diffraction peaks of HMS clearly indicate high crystallinity and phase purity. This HMS peak persists across all composite samples, confirming the retention of the HMS crystal structure even at high polymer content. Pure PANI shows a broad diffraction peak, confirming its amorphous polyaniline. The progressive appearance of HMS peaks in composites ensures the successful physical blending and the presence of both phases. As the HMS content increases from 10% to 80%, the intensity of HMS-related peaks at the specific planes such as (1015), (200), (2111), (2115), (220), (1130), and (3015) becomes more prominent. This correlation between filler content and peak intensity confirms the homogeneous dispersion and the dominant role of HMS in the overall crystallinity of composites at higher loadings. The absence of new or shifted peaks indicates that there is no chemical reaction between HMS and PANI during composite formation. Instead, the XRD pattern represents a physical mixture of intact phases, making it suitable for flexible thermoelectric device fabrication, where phase preservation is essential.

Fig. 2 (a) XRD pattern of the HMS/PANI composites along with pure PANI & HMS. (b) SEM images of the fabricated thin films: (i) the pure PANI thin film, (ii) the 50% HMS content thin film, and (iii) the 80% HMS content thin film.

3.2 SEM analysis

The SEM images in Fig. 2b are presented at the uniform scale of 20 μm, which offers insights into the surface morphology of different composites. The pure PANI thin film (Fig. 2b(i)) shows an amorphous, granular texture with some degree of interconnected porosity. This is typical of coated PANI and serves as the baseline for the polymer matrix. The film with 50% HMS content demonstrated that the HMS is embedded and well-dispersed throughout the PANI matrix. At 80% HMS, the image represents a surface that is densely packed with HMS, with PANI serving primarily as a binder. The increased surface coverage of HMS indicates successful integration without agglomeration. Across all composites, HMS elements were well integrated into the PANI matrix, with a negligible amount of aggregation. This uniformity and matrix filler adhesion are significant for stable mechanical properties and effective electron transport in flexible thermoelectric fabrics.

3.3 TGA analysis

To ensure the operational reliability of the flexible thermoelectric generator, the thermal stability of the HMS/PANI composites was evaluated via TGA (Fig. S1). The thermograms demonstrate that the initial weight loss occurring below 100 °C is associated with moisture evaporation, while the major degradation of 80% HMS/PANI does not begin until 220 °C. The TGA results confirm that the composite possesses excellent thermal headroom. Consequently, the active material remains structurally intact and chemically stable, preventing any thermally induced degradation of the electrical conductivity or power factor during continuous device operation.

3.4 XPS analysis

The XPS analysis and schematic band alignment are shown in Fig. 3. The optical band gaps (E_g) and work functions (Φ) of the constituent materials were used to develop the energy band structure based on the literature. Pure PANI exhibits a wider band gap (E_g = 3.6 eV) and a lower work function (Φ_PANI ≈ 4.42 eV) compared to the narrow band gap (E_g ≈ 0.7 eV) and higher work function (Φ_HMS ≈ 4.78 eV) of the HMS filler.^27–30 To equilibrate the Fermi levels, electrons are eventually driven to go from the PANI matrix to the HMS in composites. This phenomenon is confirmed by the high-resolution XPS N 1s spectra shown in Fig. 3b and c. The other possible constituent's peaks are shown in SI, Fig. S2. In the composite, the primary N 1s peak of pure PANI, which is centered at 397.30 eV, gradually moves to a higher binding energy of 397.6 eV. The electron transfer to HMS is physically confirmed by this +0.3 eV shift, which directly indicates a decrease in electron density surrounding the PANI nitrogen backbone. According to the electrostatic charge potential model, this reduction in valence electron density decreases the electrostatic shielding of the nucleus, resulting in a stronger effective nuclear charge and the correspondingly higher binding energy for the core electrons. This charge redistribution causes a potential energy barrier (ΔE) at the organic–inorganic interface by bending the energy bands, as shown in the band alignment schematic diagram (Fig. 3a). This barrier effectively scatters low-energy carriers while permitting the transmission of high-energy carriers. This confirms the transport behavior, validating the interfacial energy filtering model as the primary mechanism for the simultaneous decoupling of carrier mobility and the Seebeck coefficient.

Fig. 3 (a) Schematic band alignment diagram of the HMS and PANI interface, (b) N 1s XPS spectrum of pure PANI, and (c) N 1s XPS spectrum of the HMS/PANI composite (50%).

3.5 I–V characterization

To understand the voltage and current flow in the films, I–V measurements have been performed for all the different composition thin films. Fig. 4a shows the I vs. V plot recorded within the range of −1 V to +1 V. From the graph, it is evident that HMS/PANI thin films demonstrate progressive enhancement in electrical conduction with HMS loading. While pure PANI exhibits a minimal current response, 10% HMS shows a suppressed current due to isolated particles. From 30% HMS, the current increases to 0.032 mA as HMS networks begin percolating. At 50% and 80% HMS, the current reaches 0.045 and 0.086 mA, respectively, with a 90-fold enhancement over pure PANI. However, all the composites show a linear ohmic behaviour, indicating HMS/PANI interface formation.³¹ The maintained ohmic response also confirms that the interfacial band provides beneficial energy filtering for thermoelectric performance enhancement.

Fig. 4 (a) I–V characterization of all the composition composite thin films, (b) carrier concentration of HMS/PANI composite thin films, and (c) mobility of the HMS/PANI composite thin films.

3.6 Electron transport properties

The electron transport behaviour of the composite thin films was studied using a Hall measurement system. Fig. 4b shows the carrier concentration values measured for all the different compositions from pure PANI to the 80% HMS sample. Furthermore, the Hall coefficient (R_H) was extracted for all compositions. As detailed in the SI (Fig. S3), the R_H values for all composite films are consistently positive. From the values, it is evident that p-type conductivity occurs in the thin films, confirming that holes are the predominant carriers. Also, the data show a dramatic increase in carrier concentration from 1.31 × 10¹³ cm⁻³ in pure PANI to 4.21 × 10¹⁹ cm⁻³ at 80% HMS – which is a 3-million-fold enhancement. This sharp rise in carrier density represents a key factor governing the improved thermoelectric behaviour of the composites. It is noted that the pure PANI film exhibits a relatively low baseline carrier concentration (Fig. 4b). This is attributed to partial de-doping that occurs during the rigorous water and ethanol washing steps used to ensure absolute polymer purity. This phenomenon is consistent with literature reports of lightly-doped pure PANI films exhibiting carrier concentrations in the range of ∼10¹⁴ cm⁻³.³² However, this moderate baseline is advantageous; upon the incorporation of HMS, interfacial charge transfer successfully elevates the composite's carrier concentration into the optimal regime. HMS, on the other hand, acts as a degenerate semiconductor with a high carrier concentration (∼10²¹ cm⁻³).³³ The two phases go through Fermi-level alignment when integrated into PANI. This causes spontaneous electron transfer from HMS into the PANI matrix, hence increasing the carrier population within the polymer.

At 10% HMS, carrier concentration remains limited due to the small interfacial area. As the HMS fraction increases (30% and 50%), the formation of interconnected HMS networks significantly enlarges the available interface for charge transfer. This leads to a non-linear increase in carrier concentration. At a high HMS content (80%), the carrier concentration begins to saturate as the composite's Fermi level approaches that of bulk HMS.

The carrier mobility of HMS–PANI composites shows a non-monotonic dependence on HMS content (Fig. 4c). Pure PANI exhibits high mobility (2.7 × 10³ cm² V⁻¹ s⁻¹), while adding 10% HMS causes a dramatic collapse to 2.31 cm² V⁻¹ s⁻¹. This is because the isolated HMS particles act as strong scattering centers that disrupt PANI's ordered transport pathways. However, as the HMS loading increases to 30% and 50%, the mobility partially recovers to 9.3 × 10² cm² V⁻¹ s⁻¹ and 2.84 × 10² cm² V⁻¹ s⁻¹, respectively. This recovery indicates that the composite has reached its percolation threshold. The HMS particles begin to physically connect, forming continuous, percolating low-resistance pathways that facilitate easier charge transport. This percolation-driven recovery and the sensitivity of transport to filler connectivity are reported for polymer–filler composites.³⁴

At higher HMS loadings (50–80%), mobility falls again (1.03 cm² V⁻¹ s⁻¹ at 80% HMS), reflecting a transition to an interface-dominated, metal-like transport regime where abundant interfaces and grain-boundary scattering reduce carrier mobility. Such interface and grain-boundary scattering effects are described by classical models (e.g., Mayadas–Shatzkes) and by recent studies of interface-rich composites.^35,36

This behaviour is consistent with reports of competing scattering and percolation effects in thermoelectric and conductive polymer composites, where energy-selective and interface-engineering strategies have been used to balance conductivity and Seebeck performance.^37,38

3.7 Thermoelectric properties

The thermoelectric properties of the thin film are presented in Fig. 5. The measured properties demonstrate a systematic and favourable evolution of the electrical conductivity (σ), Seebeck coefficient (S), power factor (PF), and thermal conductivity components (κ_total, κ_e, and κ_l) as a function of HMS filler loading in PANI composites. The detailed analysis has been discussed below.

Fig. 5 (a) Measured electrical conductivity, Seebeck coefficient, and calculated power factor of HMS/PANI composite thin films and (b) measured thermal conductivity and calculated electronic and lattice thermal conductivity.

3.7.1 Electrical conductivity. Electrical conductivity increases substantially with HMS loading (Fig. 5a). Pure PANI shows an electrical conductivity of approximately 1 S m⁻¹, which rises to 437 S m⁻¹ at 80% HMS content. This enhancement is attributed to the percolation network formation and synergistic carrier transport evolution. Due to the high carrier density of HMS, it acts as a charge reservoir. At low HMS loadings, particles remain sparsely distributed, but as the percolation threshold (∼30–50%) is approached, interconnected HMS networks develop, following power-law scaling σ(p)∝(p − p_c)^t characteristics. Carrier mobility exhibits non-monotonic behaviour (peaking at 30% HMS then declining at higher loadings), reflecting the transition from mobility-limited to carrier-concentration-limited transport. The 6 orders of magnitude increase in carrier density dominates the conductivity term (σ = neμ), yielding a net enhancement, despite 3 orders of magnitude suppression in mobility.^39–41 Ohmic I–V response across all compositions confirms high-quality HMS/PANI interfaces enabling efficient charge transport. This percolation-driven mechanism, combined with optimized carrier dynamics, validates the composite's enhanced electronic performance for thermoelectric applications.

3.7.2 Seebeck coefficient. The Seebeck coefficient increases progressively from 16 μV K⁻¹ in pure PANI to 72 μV K⁻¹ at 80% HMS (Fig. 5a). Notably, the maximum Seebeck coefficient is achieved precisely in the regime where carrier mobility is heavily suppressed (as seen in Fig. 4c). This inverse relationship is a direct result of the interfacial energy filtering effect. The abundant HMS/PANI grain boundaries create potential barriers that preferentially scatter low-energy carriers, thereby reducing the overall mobility while allowing high-energy carriers to participate in transport. This filtering process increases the mean energy of the carrier population, leading to the observed enhancement in the Seebeck coefficient. This energy-dependent scattering mechanism effectively increases the Seebeck coefficient according to the equation derived from Boltzmann transport theory:where the selective carrier filtering enhances the logarithmic term, boosting thermopower despite increased conductivity. Also, the introduction of HMS modifies the composite's total density of states at the Fermi level. Specifically, HMS can introduce sharp features in the density of states near the Fermi level, which, according to the Mott relation, directly enhance the Seebeck coefficient.⁴² The systematic increase from 10% to 80% HMS suggests that the composite's electronic structure continuously evolves toward a configuration with an enhanced Seebeck response. This is not simply additive mixing but rather a synergistic engineering of the band structure, where PANI's character combines with HMS's metallic features to create intermediate states with superior thermoelectric properties.

3.7.3 Power factor. The power factor is defined as

PF = S²σ (2)

encapsulating the fundamental trade-off between electrical conductivity and the Seebeck coefficient. The calculated power factor is shown in Fig. 5a. Since both the parameters increase concurrently across the composition range, the power factor also shows the same trend. This behaviour validates the theoretical prediction of composite thermoelectrics, where high performance is achieved at high HMS content, as band structure engineering and percolation effects converge. The maximum PF of ∼2.3 μW (mK²)⁻¹ represents an improvement over pure PANI or pure HMS, demonstrating the substantial benefit of composite design for thermoelectric device fabrication.

3.7.4 Thermal conductivity. Total thermal conductivity (κ_total) remains relatively modest across the composition range (Fig. 5b). The graph reveals that the electronic thermal conductivity (κ_e) increases at 80% HMS, while the lattice thermal conductivity (κ_l) shows an inverse trend, decreasing from 0.05079 W mK⁻¹ to 0.04163 W mK⁻¹. The increase in electronic thermal conductivity with HMS loading is directly linked to the rise in electrical conductivity through the Wiedemann–Franz law:

κ_e = LTσ (3)

Where L is the Lorenz number and T is the absolute temperature. As HMS-induced conductive pathways develop, more charge carriers participate in heat transport, proportionally increasing κ_e. Conversely, the dramatic suppression of lattice thermal conductivity is a significant finding reflecting enhanced phonon scattering at the HMS/PANI interfaces. Overall, this behaviour is desirable for thermoelectric applications, as reduced total thermal conductivity directly enhances the figure of merit:

The low thermal conductivity enables significant ZT enhancement by decreasing the denominator. The dimensionless figure of merit (ZT) has been calculated to thoroughly assess the thermoelectric performance (Fig. S5). At room temperature, the optimized 80% HMS/PANI composite is estimated to have a maximum ZT of around 0.015. However, this reflection of the flexible polymer system's intrinsically moderate ZT.

3.8 Device output

The hybrid composite film was cut to a size of 5 mm width and 20 mm height to act as the p-type leg with a uniform thickness of 400 µm. These legs were placed on a Kapton substrate with silver paste and connected with copper tape to form a thermoelectric device (Fig. 6a and b). The final, fully assembled flexible thermoelectric device had overall macroscopic dimensions of 27 mm × 50 mm × 1 mm (inclusive of the Kapton tape substrate, composite fabric, silver paste, and copper tape connections). The 80% HMS/PANI thin film device demonstrates a reliable thermoelectric output performance across varying thermal gradients. Current–voltage (I–V) characteristics reveal progressive enhancement with increasing temperature difference (ΔT), with the maximum current output reaching approximately 0.2 µA at ΔT = 70 K (Fig. 6c). Power output exhibits a characteristic parabolic response, with a maximum power output reaching 0.33 nW at ΔT = 70 K and 3.5 mV applied voltage (Fig. 6d), demonstrating temperature-dependent enhancement of the Seebeck-driven power generation. To evaluate the device's power generation capabilities, the output power was plotted as a function of the external load resistance (Fig. 6e). As is characteristic of thermoelectric generators, the power output initially increases with load resistance, reaching a distinct peak before declining. This maximum power output of 0.33 nW at ΔT = 70 K is achieved when the external load resistance equals the internal resistance (∼32 kΩ) of the HMS/PANI flexible device, confirming optimal impedance matching for maximum power transfer. The power density increases substantially with ΔT, indicating the efficient thermal-to-electrical energy conversion. The linear I–V response confirms ohmic contact quality, while the consistent power generation across all temperature differences validates the device's reproducibility and stability. These results confirm that the 80% HMS composition, optimized through comprehensive thermoelectric characterization, delivers substantial electrical power output suitable for a practical flexible energy harvesting application.

Fig. 6 (a) Graphical representation of the fabricated flexible thermoelectric device, (b) photo of the fabricated thermoelectric device, (c) measured output current of the fabricated device with varying ΔT, (d) calculated power output of the fabricated device with varying ΔT and (e) power output vs. load resistance (P vs. R curve).

3.9 Real-time application

To study the practical applicability of the optimized 80% HMS/PANI thermoelectric device, its real-time performance was demonstrated by integrating the device into an operating 3D printer (Fig. 7c). The device was adhered to a pre-heated printing bed at 55 °C while the ambient atmosphere was at room temperature (Fig. S6). This created and maintained a natural thermal gradient of approximately 30–35 °C. Under these in situ operating conditions, the thermoelectric device generated a voltage ranging from 6 to 8 mV (Fig. 7d). Through this, the device's capability to convert waste heat from common electronic equipment into electrical energy is demonstrated (Fig. S7). This output voltage is consistent and reproducible in practical uncontrolled environments without requiring controlled temperature differentials or laboratory setups. The stable voltage generation under dynamic conditions despite thermal fluctuations inherent to the 3D printing process confirms the device's durability and reliability for real-world applications. Similarly, the device was pasted on various hot objects, and the output was analysed (Fig. 7b). This demonstration exemplifies the device's suitability for flexible, wearable, and ambient energy harvesting applications where temperature gradients are typically small and dynamic.

Fig. 7 (a) Photo of the TE device pasted on a water bottle at RT, (b) photo of the TE device pasted on a water bottle with hot water, (c) schematic representation of TE device output measurement from the operating 3D printer bed, and (d) recorded output voltage of the TE device from the operating 3D printer bed.

While the flexible HMS/PANI generator thrivingly harvests low-grade waste heat, its power output necessitates further system-level integration for practical IoT or wearable applications. Because direct, continuous operation of standard electronics is not feasible, these modules must be coupled with power management circuits for voltage conditioning and micro-capacitors for energy storage. By utilizing a duty cycling approach, energy is continuously accumulated and released in intermittent bursts. These flexible generators can serve as viable components for ultra-low-power autonomous sensor nodes.

As shown in Table 1, the current HMS/PANI prototype's absolute maximum power output (0.33 nW) is low when compared to flexible generators that use conventional inorganic heavy-metal alloys (e.g., Bi₂Te₃ and Sb₂Te₃) or highly conductive carbon nanotubes (CNTs). The variance is primarily caused by the polymer matrix's intrinsically low electrical conductivity when compared to metallic or entirely inorganic systems, as well as the current module's minimal number of legs (n = 4). However, the primary objective of this work is to develop a composite that is entirely devoid of heavy metals, inexpensive, and simple to process, rather than to meet the absolute power of toxic telluride alloys. The HMS/PANI flexible generator successfully demonstrates competitive voltage generation (6–8 mV) under practical conditions, validating its potential as an environmentally benign alternative for low-grade waste heat harvesting.

Table 1 Comparison of the TE device output with those in the literature

Materials		Substrate	Dimension of leg (cm)	No. of leg	Heat source	ΔT (K)	Output voltage (mV)	Maximum power (nW)	Ref.
p-Type	n-Type
Bi2Te3	Co:ZnO	Fabric	2 × 0.5	4	Human body (wrist)	25	0.7	85	43
Sb2Te3	Ag2Se	Polyimide	2 × 0.4	2	Human body (while running)	20	10	65	44
Cu–Bi0.5Sb1.5Te3/PEDOT: PSS	—	Polyimide	2.5 × 0.8	8	Human wrist	—	7.7	—	45
CNT/PANI	PEI/CNT	Yarn	—	80	Human wrist	10	4.7	377	46
—	CNT	PET	9.6 × 1	—	Hot water	27.5	4.5	2500	47
CuAl2O4	Cu0.94Co0.06Al2O4	Fabric	1 × 1	2	Human wrist	2.7	2.4	304	48
HMS/PANI	—	Fabric	2 × 0.5	4	Operating 3D printer bed	70	6–8	0.33	This work

---

4. Conclusion

This work demonstrated the successful development and characterization of HMS/PANI composite thin films as flexible thermoelectric materials for practical energy harvesting applications. At an optimal 80% HMS loading, the electrical conductivity reached 438 S m⁻¹, driven by percolation-induced interconnects and interfacial charge transfer. Concurrently, the interfacial energy filtering effect successfully decoupled interrelated transport parameters, yielding an enhanced power factor of 2.3 μW (mK²)⁻¹, while interfacial phonon scattering effectively suppressed the lattice thermal conductivity. The fully assembled flexible generator delivered a maximum power output of 0.33 nW at ΔT = 70 K. Crucially, real-time testing on an operating 3D printer bed generated a stable 6–8 mV output, successfully demonstrating the composite's practical viability as a sustainable power source for future ultra-low-power IoT systems. In addition, the existing HMS/PANI flexible generators show great stability and feasibility for low-grade heat harvesting. Future research will concentrate on improving the trade-off between conductivity and thermopower. Specifically, integrating higher-mobility inorganic fillers, engineering more optimized continuous conductive architectures, and employing precise interfacial doping strategies hold great promise for elevating the power factor of these flexible composites to be highly competitive with state-of-the-art materials.

Author contributions

Nishath Begum Jamal Mohammed: conceptualization, investigation, methodology, and writing – original draft. Arunkumar Chandrasekhar: conceptualization, supervision, funding acquisition, and writing – review & editing. Rajasekar Parasuraman: conceptualization, supervision, validation, and writing – review & editing. Pandiyarasan Veluswamy: project administration and formal analysis.

Conflicts of interest

The authors claim that there are no conflicts of interest to declare.

Data availability

The data are available upon reasonable request from the corresponding author.

Supplementary information (SI) is available. See DOI: https://doi.org/10.1039/d6nr00059b.

Acknowledgements

We gratefully acknowledge the financial support provided by the Vellore Institute of Science and Technology, Vellore, INDIA, through the VIT International fund (VIN) award 2024-2025 (Award No. [VIN2024-01]). We thank the NSNE laboratory (Nano Sensors and Nano Energy Laboratory), the Chemistry Research Laboratory located in SJT Annex 102. We further acknowledge the support of the SMILE Research Laboratory, IIITDM Kancheepuram with the Department of Science and Technology (DST), Government of India, under Grant No. CRG/2021/006188.

Schematic diagrams in Fig. 1a–c, 6a and 7c were generated using Perplexity Pro (DALL·E 3 model, December 2025), followed by arranging them using a Microsoft PowerPoint presentation, and reviewed by the authors for scientific accuracy.

References

1. M. Rahimi, et al., Emerging electrochemical and membrane-based systems to convert low-grade heat to electricity, Energy Environ. Sci., 2018, 11, 276–285.
2. S. Hur, et al., Low-grade waste heat recovery scenarios: Pyroelectric, thermomagnetic, and thermogalvanic thermal energy harvesting, Nano Energy, 2023, 114, 108596.
3. W. Wu, et al., Enhancing the waste heat utilization of industrial park: A heat pump-centric network integration approach for multiple heat sources and users, Energy Convers. Manag., 2024, 306, 118306.
4. H. Huo, Y. Xuan and T. Meng, Enhancing thermoelectric conversion efficiency of hydrogel-based supercapacitors by the three-dimensional ion channels hydration, J. Energy Storage, 2024, 80, 110437.
5. J. M. Cansino, M. Ordóñez and M. Prieto, Decomposition and measurement of the rebound effect: The case of energy efficiency improvements in Spain, Appl. Energy, 2022, 306, 117961.
6. J. Huang, et al., Rational Design of Electrode Materials for Advanced Supercapacitors: From Lab Research to Commercialization, Adv. Funct. Mater., 2023, 33, 2213095.
7. Z. Li, et al., Thermo-electrochemical cells enable efficient and flexible power supplies: From materials to applications, Energy Storage Mater., 2025, 74, 103902.
8. M. Zhang, L. Shi, P. Hu, G. Pei and G. Shu, Carnot battery system integrated with low-grade waste heat recovery: Toward high energy storage efficiency, J. Energy Storage, 2023, 57, 106234.
9. M. S. Iqbal, et al., Recent advances in triboelectric nanogenerators: Mechanism, rational designing and applications, Mater. Today Energy, 2024, 46, 101732.
10. W. Liu, C. Yuan and X. Wang, Hybrid Flexible Thermoelectric Temperature Sensors for Early Warning, Adv. Funct. Mater., 2025, 36(12), e19209.
11. J. Oliver, R. Malet, R. Aragonés, R. Voces and C. Ferrer, Waste heat recovery unit for energy intensive industries thermoelectricity harvesting. In 2020 IEEE 29th International Symposium on Industrial Electronics (ISIE), IEEE, 2020, pp. 1079–1084.
12. K. Manojkumar, et al., Flexible Thermoelectrics for Wearable Electronics: Trends and Benchmarks in Solid-State and Ionic Materials, Textile Architectures, Interface Engineering, and Device Performance, Adv. Electron. Mater., 2025, 12(2), e00396.
13. W. Wu, et al., Efficient transformation of industrial waste heat to chemical energy: a hybrid regenerator with preheating and internal heating, Appl. Therm. Eng., 2026, 284, 129208.
14. V. Rathi, R. Brajpuriya, R. Gupta, K. P. S. Parmar and A. Kumar, Graphene-derived composites: a new Frontier in thermoelectric energy conversion, Energy Adv., 2024, 3, 389–412.
15. M. M. R. Al-Fartoos, A. Roy, T. K. Mallick and A. A. Tahir, Advancing Thermoelectric Materials: A Comprehensive Review Exploring the Significance of One-Dimensional Nano Structuring, Nanomaterials, 2023, 13, 2011.
16. Z. Liang, et al., Substrate-Free Inorganic-Based Films for Thermoelectric Applications, Adv. Mater., 2025, 37, 2416394.
17. W. Liu, Z. G. Di Chen and J. Zou, Eco-Friendly Higher Manganese Silicide Thermoelectric Materials: Progress and Future Challenges, Adv. Energy Mater., 2018, 8, 1800056.
18. N. Jain and Y. O. Waidi, The Multifaceted Role of 3D Printed Conducting Polymers in Next-Generation Energy Devices: A Critical Perspective, JACS Au, 2025, 5, 411–425.
19. M. Chen, D. Xie, H. Zhou and P. Zong, PANI-Based Thermoelectric Materials, Organics, 2025, 6(3), 33.
20. J. S. Khan, et al., Thermoelectric properties of polyvinylpyrrolidone/polyaniline-mixed Bi2Te3 composites prepared by electrospinning, Polym. Compos., 2023, 44, 2745–2756.
21. M. Singh, A. K. Gautam, M. Faraz and N. Khare, Flexible, Freestanding Ternary Nanocomposite Thick Film of Polyaniline/WS2/CNTs for Efficient Thermoelectric Application, ACS Appl. Energy Mater., 2024, 7, 133–144.
22. G. Arslanturk, et al., Nickel sulfide-filled polyaniline hybrid composites for high-performance thermoelectric materials at room temperature, Synth. Met., 2026, 317, 118048.
23. A. Sharma, N. S. Chauhan and T. Mori, Energy filtering and anisotropic structural response in polyaniline:CSA hybrids for flexible thermoelectrics, J. Mater. Chem. A, 2026, 14, 8014–8030.
24. Y. Li, et al. Enhanced Thermoelectric Performance of PEDOT/PSS by Interface Regulation of Hydroxylated Fluorinated Graphene, ACS Appl. Polym. Mater., 2025, 7, 3309–3317.
25. M. M. Ayad and M. A. Shenashin, Polyaniline film deposition from the oxidative polymerization of aniline using K2Cr2O7, Eur. Polym. J., 2004, 40, 197–202.
26. L. P. Chandrasekar, P. Veluswamy, H. Ikeda and S. Mohandos, Exploring the synergistic effects of secondary phase in copper aluminate via stannous doping for thermoelectric applications, Surf. Interfaces, 2025, 62, 106245.
27. R. P. Mccall, et al., Spectroscopy and Defect States in Polyaniline, Phys. Rev. B, 1990, 41(8), 5202.
28. O. Abdulrazzaq, et al. Tuning the work function of polyaniline via camphorsulfonic acid: An X-ray photoelectron spectroscopy investigation, RSC Adv., 2015, 5, 33–40.
29. D. B. Migas, V. L. Shaposhnikov, A. B. Filonov, V. E. Borisenko and N. N. Dorozhkin, Ab initio study of the band structures of different phases of higher manganese silicides, Phys. Rev. B:Condens. Matter Mater. Phys., 2008, 77(7), 075205.
30. G. Kim, et al., Synchronized enhancement of thermoelectric properties of higher manganese silicide by introducing Fe and Co nanoparticles, Nano Energy, 2020, 72, 104698.
31. M. Shalini, et al., Facet dependent ultralow thermal conductivity of zinc oxide coated silver fabric for thermoelectric devices, Sci. Rep., 2024, 14(1), 27210.
32. S. M. Giripunje and J. Ghushe, 99-P010, Int. J. Appl. Phys. Math., 2012, 2, 234–236.
33. Z. Li, et al., Isoelectronic Re-Ge-codoped higher manganese silicides with enhanced thermoelectric properties via band optimization, charge transfer, and phonon scattering, J. Eur. Ceram. Soc., 2023, 43, 4799–4807.
34. W. Bauhofer and J. Z. Kovacs, A review and analysis of electrical percolation in carbon nanotube polymer composites, Compos. Sci. Technol., 2009, 69, 1486–1498.
35. A. F. Mayadas and M. Shatzkes, Electrical-resistivity model for polycrystalline films: the case of arbitrary reflection at external surfaces, Phys. Rev., 1970, 1(4), 1382.
36. F. Corbella, et al., An Application of the Mayadas-Shatzkes Model of Thin Film Resistivity to Ag-Based Thermal Control Coatings, ACS Appl. Mater. Interfaces, 2025, 17, 51431–51447.
37. X. Guan and J. Ouyang, Enhancement of the Seebeck Coefficient of Organic Thermoelectric Materials via Energy Filtering of Charge Carriers, CCS Chem., 2021, 3, 2415–2427.
38. T. A. Yemata, et al., Interface engineering in polymer thermoelectric composites: harnessing energy-filtering effects to overcome the Seebeck-conductivity trade-off, J. Mater. Chem. A Mater., 2025, 13(42), 35976–35998.
39. L. He and S. C. Tjong, Low Percolation Threshold of Graphene/Polymer Composites Prepared by Solvothermal Reduction of Graphene Oxide in the Polymer Solution, Nanoscale Res. Lett., 2013, 8(1), 132.
40. C. Kim, T. Kim and J. Cho, Selective Charge Carrier Transport and Bipolar Conduction in an Inorganic/Organic Bulk-Phase Composite: Optimization for Low-Temperature Thermoelectric Performance, ACS Appl. Mater. Interfaces, 2024, 16, 5036–5049.
41. S. Goswami, S. Nandy, E. Fortunato and R. Martins, Polyaniline and its composites engineering: A class of multifunctional smart energy materials, J. Solid State Chem., 2023, 317, 123679.
42. L. P. Chandrasekar, P. Veluswamy, H. Ikeda and S. Mohandos, Enhancing thermoelectric performance in flexible fabric-based Mo-doped CuAl₂O₄: Insights into carrier type modification and electrical conductivity optimization, Ceram. Int., 2024, 50, 48330–48342.
43. B. Lincoln, R. Annie Sujatha, P. Veluswamy and A. Majumdar, A hybrid ceramic-based flexible thermoelectric nanogenerator with enhanced thermopower for human energy harvesting, Energy Convers. Manag., 2023, 292, 117364.
44. D. Yang, et al., Flexible power generators by Ag2Se thin films with record-high thermoelectric performance, Nat. Commun., 2024, 15(1), 923.
45. Y. Wang, et al., Bi0.5Sb1.5Te3/PEDOT:PSS-based flexible thermoelectric film and device, Chem. Eng. J., 2020, 397, 125360.
46. H. Li, et al., High-Performance Polyaniline-Coated Carbon Nanotube Yarns for Wearable Thermoelectric Generators, ACS Appl. Mater. Interfaces, 2024, 16, 17598–17606.
47. W. Zhou, et al., High-performance and compact-designed flexible thermoelectric modules enabled by a reticulate carbon nanotube architecture, Nat. Commun., 2017, 8(1), 14886.
48. L. P. Chandrasekar, et al., The spinel-based pliable thermoelectric device for room temperature application, Mater. Chem. Phys., 2024, 322, 129520.

---