introduced to improve the films' performance has also increased the time and energy required to produce them. In particular, the high-temperature, long-duration (~500 • C for up to 18 h) sintering process negated the time and energy advantage of additive manufacturing over traditional manufacturing methods [16][17][18][19][20]. More information and comparison of the high energy consumption and long duration between different methods and literatures are shown in Table S1 in the supplemental file.

Existing literature demonstrated the film thickness dependence only for thin films (hundreds of nanometers or several micrometers) [22][23][24]. Han et al. used chemical vapor deposition (CVD) to fabricate p-type Sb 2 Te 3 thin films at a thickness range of 1-16 µm and achieved a maximum electrical conductivity of 666 S/cm at the lowest thickness (1 µm) [23]. Wang et al. used electron-beam evaporation to fabricate p-type Sb 2 Te 3 thin films with thicknesses between 10 to 100 nm. A maximum electrical conductivity of 700 S/cm was also achieved at the lowest thickness (10 nm) with decreases in defects and interfaces [22]. Physical techniques like CVD and physical vapor deposition (PVD) are costly and time-consuming. They grow films by repeatedly depositing thin layers, which could introduce extra defects, interparticle interfaces, and void formations. As an alternative, in this work, we have used a cost-effective and energy-efficient additive manufacturing technique to print thermoelectric films and TEGs with thicknesses in the range of hundreds of microns. However, the thickness effect in printed composite films (up to 300 µm) has not been systematically studied. This work aims to systematically investigate the effect of thickness together for next-generation f-TEGs with optimized performances fabricated via energy-efficient and scalable manufacturing processes.

Our previous works reported a low-energy-input printing method for fabricating n-BTS and p-BST composite films and single-leg devices using the synergistic effects of (1) a small amount of chitosan binder (0.05 wt%), ( 2) heterogeneous (mixed nanoscale and microscale) TE particles, and (3) applied mechanical pressure of 20 MPa combined with curing at a low temperature of 120 • C for 30 mins [25][26][27][28][29]. Table S2 and S3 (Supporting Information) compare this current research work with our previously reported printed TE composite work. While the results of these experiments were promising, the performance of films produced by our low-energy input process was lower than those produced by the energy-intensive sintering process. Therefore, we hypothesize that the reduction in film thickness, similar to what was observed in previously researched thinner films (1-16µm) produced through vapor deposition, will lead to a reduction in the occurrence of defects increasing electrical conductivity in our 170 µm films when compared to our 300 µm films. More generally, defect concentration and nature within TE films seem critical in optimizing film performance. With this in mind, we introduced a fourth synergistic factorthickness variationto our process, intending to produce films with micro and nanostructures that optimized electrical conductivity while simultaneously reducing thermal conductivity, effectively decoupling the two properties. Finding high-performance n and p-type films will allow high-energy density f-TEG devices to be manufactured at scale without expending much time or energy.

Each of the four synergetic factors influences the micro and nanostructure and interfaces of the films differently. We assume that tuning the weight ratio of the naturally occurring and low-cost Chitosan binder will produce composite microstructures that will not affect the electrical connection among active TE particles while facilitating interfacial thermal resistance among particles and the polymer micron-sized particles that provide a large mean free path for charge carriers and nanosized particles and interfacial defects facilitating phonon scattering. Additionally, tuning external uniaxial pressure will initiate lattice defects, resulting in high-weighted mobility and phonon vibrations. We hypothesize that a film with a lower thickness (~170 µm compared to 300 µm) will have fewer micron-scale defects (grains, grain boundaries, and interfaces), resulting in less charge carrier scattering and high electrical conductivity. Moreover, we assume different film thicknesses will experience the effect of other synergetic factors differently. An important aspect of the proposed low-thermal budget additivemanufacturing method is that it promises to eliminate the need for the current state-of-the-art long-duration and high-temperature curing process, decouple electrical and thermal conductivity, and achieve highperformance composite TE films. This results in a process whose only energy-consuming component is the application of mechanical pressure, which had previously been used together with the high-temperature curing process [23,24,30].

Thus, this work demonstrated the fabrication of n and p-type freestanding films using this low-thermal budget approach to overcome the issues of the high-temperature curing process to allow for widespread adoption of TEGs as illustrated in Fig. 1. The effect of thickness on thermoelectric parameters was explored. Air stability and flexibility tests showed the real-world durability of our TEG manufacturing technique. Finally, the f-TEG device's capability as a self-sufficient power source was demonstrated by integrating it with a step-up circuit to light a blue LED and charge a capacitor. Fig. 1 illustrates the importance of all synergistic factors to produce scalable f-TEGs with high energy density.

The free-standing films achieved high electrical conductivity for both chitosan p-BST-100 mesh (660 S/cm) and chitosan n-BTS-100 mesh (286 S/cm) films of 170 µm owing to the smaller number of micro-scale defects (pores, interfaces, grain boundaries) compared to 240 µm and 300 µm films. The nano-sized defects, particles, pores, and interfaces resulted in a low thermal conductivity of 0.77 W/m-K for p-BST and 0.65 W/m-K for n-BTS films. We decouple electrical and thermal conductivity using interfacial engineering and a low-thermal budget additive manufacturing approach. High electrical conductivity and low thermal conductivity resulted in a high ZT of 0.89 for p-BST and 0.50 for n-BTS composite films, which is 25% higher than our previously reported ZT for p and n composites [25][26][27]. Further, a 6-couple flexible device was stencil printed using p-BST and n-BTS composite inks. A maximum open circuit voltage of 91 mV and a maximum power output of 357 µW was achieved at a temperature difference (ΔT) of 40 K. The device exhibited a power density of 5.0 mW/cm 2 , a 40% improvement from our previously best-reported TEG devices [17][18][19][20][21][25][26][27]30]. A flexibility test was performed by bending the TEG device at a 3-cm radius for 1000 cycles. Electrical resistance was chosen to evaluate the flexibility of the TEG device due to its sensitivity to cracking with bending [18]. The TEG device experienced only 4.3% increase in internal resistance and no measurable change in voltage at a ΔT of 40 K, as seen in Fig. S16. An air stability test was conducted by exposing the TEG device to ambient air for more than six months, showing an increased resistance of 23% and no measurable change in voltage at a ΔT of 40 K. Finally, the TEG device capability as a self-sufficient power source in a real-world application was demonstrated by lighting a blue LED and charging a capacitor at a ΔT of 17 K when integrated with a step-up circuit. These demonstrations proved the potential application of printed f-TEGs as a self-sufficient power supply for wireless sensor devices to monitor physical parameters continuously.

The materials used in the research, 100 mesh Bi 0.5 Sb 1.5 Te 3 (BST-100) and Bi 2 Te 2.7 Se 0.3 (BTS-100) were purchased from Wuhan Dongxin Inc., China. The 100 mesh TE particles contain a mixed distribution of micron and nano-sized TE particles [25]. Chitosan powder (MP Biomedicals, LLC), Dimethyl sulfoxide (DMSO, ≥99.7%, extra dry, ACROS Organics), and Salicylic acid were purchased from Fisher Scientific. The 5 mm thick non-adhesive Kapton substrate was purchased from Caplinq, Canada, and used as masking tape.

Im a g e J s o f tw a r e w a s u s e d f o r t h e S EM fim a g e s ( F fi g .

, 3 9 , 4 6 , 5 6 , 5 7 , 6 0 , 6

F

( 5 ) . 0.89 for p-BST 170 µm and a ZT of 0.50 for n-BTS 170 µm composite films at room temperature. The higher ZT achieved for p-type and n-type composite films is among the best-reported values for printed films using a low-thermal budget method with a low curing temperature (120 • C for 10 min). The achieved power density of 5.0 mW/cm 2 at a temperature difference of 40 K for 6 couple BST-BTS devices stands as a 40% improvement from our previously best reported TEG. Further, the stability and flexibility tests showed that the fabricated TEG device is suitable for practical and real-world applications. Demonstrating LED lighting and capacitor charging at a temperature difference of 17 K confirms the device's capability as a self-sufficient power source for wireless sensor devices. We envisage that the high-performance, flexible, scalable, and air-stable TEG device, fabricated with energy-efficient methods, can be a self-sufficient and long-lasting power source for many wireless sensors to monitor physical parameters continuously.