Enhanced thermoelectric properties of self-assembling ZnO nanowire networks encapsulated in nonconductive polymers

The near-room temperature thermoelectric properties of self-assembling ZnO nanowire networks before and after encapsulation in nonconductive polymers are studied. ZnO nanowire networks were synthesized via a two-step fabrication technique involving the deposition of metallic Zn networks by thermal evaporation, followed by thermal oxidation. Synthesized ZnO nanowire networks were encapsulated in polyvinyl alcohol (PVA) or commercially available epoxy adhesive. Comparison of electrical resistance and Seebeck coefficient of the ZnO nanowire networks before and after encapsulation showed a significant increase in the network's electrical conductivity accompanied by the increase of its Seebeck coefficient after the encapsulation. The thermoelectric power factor (PF) of the encapsulated ZnO nanowire networks exceeded the PF of bare ZnO networks by ~ 5 and ~ 185 times for PVA- and epoxy-encapsulated samples, respectively, reaching 0.85 μW m−1 K−2 and ZT ~ 3·10–6 at room temperature, which significantly exceeded the PF and ZT values for state-of-the-art non-conductive polymers based thermoelectric flexible films. Mechanisms underlying the improvement of the thermoelectrical properties of ZnO nanowire networks due to their encapsulation are discussed. In addition, encapsulated ZnO nanowire networks showed excellent stability during 100 repetitive bending cycles down to a 5 mm radius, which makes them perspective for the application in flexible thermoelectrics.

reducing k without affecting σ.To date, most of the research on the ZnO thermoelectric properties has focused on investigating ZnO thin films or ZnO powders-based ceramics [34][35][36][37] .Reported for these materials, Seebeck coefficients (S) were generally measured at elevated temperatures of 500-1000 °C and ranged from ~ − 100 to ~ − 380 µV K −135 , with some reports claiming S of ZnO powders-based densified ceramics up to ~ − 600 µV K −1 measured at room-temperature 34 .Knowledge of thermoelectric properties of ZnO nanowires or nanowire networks is limited to a few reports.Zappa et al. 12 reported room-temperature S values of obtained by thermal evaporation ZnO nanowires ranging from ~ 110 to -190 µV K −1 , which correlate with the theoretical S values (from − 120 to − 150 µV K −1 at temperatures 300-600 K) 38 .
Despite the importance of the development of flexible thermoelectric materials application in wearable devices and waste heat harvesting from curved surfaces 39 , there is a minimal number of reports on the application of ZnO for flexible thermoelectrics at near-room temperatures 22,40 , where ZnO is used in the form of powder filler for electrically conductive polymer 40 or deposited in the form of thin film on a flexible substrate 22 .Most state-ofthe-art flexible thermoelectric materials developed for low-grade heat-to-power conversion are polymer-based composites with different nanostructured fillers, such as carbon nanotubes, bismuth and antimony chalcogenides, or hybrid fillers [41][42][43][44][45] .Commonly, the matrix for the polymer-based composite is electrically conductive polymers.However, despite the satisfactory efficiency of the electrically conductive polymers-based thermoelectric composites, reaching ZT up to 0.4 43 , the use of electrically conductive polymers as a matrix has some significant drawbacks.Mostly, these polymers have p-type conductivity and, thus, are more suitable for developing p-type composites rather than n-type.In turn, n-type electrically conductive polymers have poor stability in air and moisture, resulting in a significant decrease in their conductivity 46 .In addition, the poor processability of many electrically conductive polymers significantly complicates the synthesis of the composites.
Recently, our group demonstrated a novel approach for fabricating flexible n-type thermoelectric thin films by encapsulating nanostructured Bi 2 Se 3 -MWCNT and Sb 2 Te 3 -MWCNT hybrid networks in non-conductive polymers and application of these networks in flexible thermoelectric generators [47][48][49] .However, to date, the research of non-conductive polymer-or epoxy-based composites containing nanostructured ZnO filler was focused on the investigation of their mechanical, thermal, electrical, and optical properties for applications not related to thermoelectrics [50][51][52][53] .
In this work, self-assembling ZnO nanowire networks were synthesized by thermal oxidation of metallic Zn nanowire network pre-deposited on a solid (glass) or flexible (mica) substrate using low-pressure thermal evaporation.After oxidation, ZnO nanowire networks were encapsulated in PVA or commercially available epoxy adhesive using a drop-casting technique.The electrical and thermoelectric properties of the ZnO nanowire networks before and after encapsulation were studied at room temperature and compared.In addition, temperaturedependent thermoelectric properties of encapsulated in PVA and epoxy adhesive ZnO networks were studied in the target temperature range 180-380 K and compared.The encapsulated ZnO nanowire networks' bending behavior and mechanical and electrical stability were evaluated.To our knowledge, the thermoelectrical properties of PVA-and epoxy-encapsulated ZnO nanowire networks were investigated for the first time.

Deposition of metallic Zn nanowire networks
Zinc foil (99.95%, thickness 50 µm, Goodfellow GmbH, Hamburg, Germany) was cut in pieces of approx.2.5 × 6.0 cm (weight 0.22 g ± 0.02 g) and used as Zn source material for the deposition of Zn nanowire network.Before the synthesis, the Zn foil and glass substrates (Avantor, Radnor, PA, USA) were cleaned in acetone and isopropanol and dried under nitrogen gas (N 2 ) flow.For the fabrication of flexible samples, freshly cleaved mica sheets (~ 60 µm thick, Goodfellow, Huntingdon, UK) Pre-cleaned Zn foil were placed into the center of the horizontal single-zone quartz furnace tube (OTF-1200X, MTI Corp., Richmond, CA, USA).The substrates were placed into the furnace tube in the area where the maximum temperature during the process will reach 300-350 °C.Before the synthesis, the quartz tube was flushed for a few minutes with N 2, pumped down to a pressure of 0.2 Torr, and sealed.For the evaporation of Zn source material, the furnace was heated up to 450 °C in the centre of the tube at a rate of 15°C min −1 and kept at this temperature for the next 60 min, followed by natural cooling down to room temperature.

Oxidation of Zn nanowire networks to ZnO
Substrates with the pre-deposited Zn nanowire networks were placed in the centre of open to the ambient air tube of the furnace tube (OTF-1200X, MTI Corp., Richmond, CA, USA).The furnace was heated to 500 °C at a rate of ~ 16°C min −1 and kept at this temperature for the next 60 min, followed by natural cooling to room temperature.

Encapsulation of ZnO nanowire network with PVA and epoxy adhesive
For the encapsulation in PVA, the 5 wt% solution of PVA (Celvol E 04/88S) and deionized water DIW) was ultrasonicated for 20 min and stirred for 30 min at a temperature of 70 °C.PVA-DIW solution was uniformly applied over the ZnO network and dried in air at 50 °C for 12 h.Epoxy adhesive (Henkel AG&Co.KGaA, Dusseldorf, Germany) was drop-casted on the surface of the ZnO nanowire network sample, followed by its spread throughout the network due to the capillary forces and dried under ambient conditions for 24 h.After encapsulation, the wt.% ratio of ZnO to encapsulating polymer was approximately 85:15.

Morphological and structural characterization
The structure and morphology of prepared ZnO nanowire networks were inspected using a field-emission scanning electron microscope (SEM) Hitachi S-4800 (Hitachi Ltd., Tokyo, Japan).Powder X-ray diffraction (XRD) measurements were performed on a Bruker D8 Discover diffractometer (Bruker Corp., Billerica, MA, USA) using copper radiation source (Cu Kα = 1.54180Å) with Bragg-Brentano geometry and a LynxEye (1D) detector.The divergence and anti-scattering slits were set at 0.6 mm and 8 mm, respectively.The patterns were recorded from 10° to 70° on the 2θ scale, using a scan speed of 0.5 s/0.02°.For the identification of the diffraction peaks, the ICDD database PDF-2/Release 2021 was used (Ref.cards PDF 01-078-9363 Zn and PDF 01-075-6445 ZnO).Fourier transform infrared (FTIR) spectra were measured with Bruker Vertex 70v spectrometer equipped with an attenuated total reflection module with diamond crystal.Measuring range 400-4000 cm −1 , resolution ± 2 cm −1 , 20 scans per spectrum.Measurements performed in vacuum, 2.95 hPa.

Electrical and thermoelectric characterization
For electrical and thermoelectric measurements, electrical contacts (Au-Pd alloy, thickness 25 nm) were sputtered on the ZnO nanowire networks by ion beam coater Gatan 681 (Gatan, Inc., Pleasanton, CA, USA).The distance between the electrodes was 8-12 mm, and the sample width was 4-7 mm.Current-voltage curves of the sample were measured by a Keithley 6487 picoammeter/voltage source.Room-temperature Seebeck coefficient measurements were conducted under ambient conditions using a lab-made device reported elsewhere 48 and calibrated by NIST Standard Reference material 3451 (NIST, Gaithersburg, MD, USA) for low-temperature Seebeck coefficient.Thermoelectric voltage, generated by the ZnO nanowire network, was registered by HP 34401A multimeter (Hewlett-Packard Company, Palo Alto, CA, USA) and custom software for automatic data recording.The temperature gradient between the sample sides did not exceed 5% from the absolute temperature at which the measurements were performed.The thermoelectric voltage generated by the samples before and after encapsulation was calculated as the mean of five consecutive measurements, and ± standard deviation [SD] was estimated from these five measurements.Measurements of temperature dependencies of the Seebeck coefficient, resistance, and thermal conductivity were performed under high vacuum conditions using the thermal transport option (TTO) of the physical property measurement system (PPMS DynaCool9T, Quantum Design, San Diego, CA, USA) in the temperature range from 180 to 380 K.The measurement errors were estimated by the PPMS software MultiVu by fitting different models to the experimental data.

Bending tests
Two-point bending tests down to a 5 mm radius were performed using a custom experimental setup reported elsewhere 54 , allowing simultaneous bending of the sample and measuring its electrical current at a constant voltage applied to the sample.The resistance of the samples was calculated from applied to the sample voltage (0.1 V ± 1 mV) and measured current ([SD] ± 100 pA).For the data representation, the average resistance for three consecutive measurements was calculated, and [SD] was calculated using Excel for the Microsoft 365 function STDEV.S.

Results and discussion
Deposited on a substrate, a metallic Zn nanowire network typically consists of randomly arranged interconnected nanowires of average radii 50-100 nm and length of a few μm (Fig. 1a,b).The thicknesses of the Zn nanowire network layers on glass substrates varied between 10 and 100 μm.XRD patterns obtained for the Zn nanowire network revealed diffraction peaks corresponding respectively to the (002), (100), (101), and (102) crystallographic planes of hexagonal phase Zn (Fig. 1e, black curve).The intensity of these diffraction peaks corresponds to that of the reference Zn metal, thus indicating no predominating growth direction of Zn nanowires.No zinc oxide phases were detected, thus proving that deposition conditions were optimal for forming metallic Zn nanowires.After oxidation, the oxidized Zn nanowire network changed its color from grey to white, characteristic of ZnO.XRD patterns obtained for the oxidized nanowire network revealed diffraction peaks corresponding to (100), (002), ( 101), (102), (110), and (103) crystallographic planes of wurtzite ZnO (Fig. 1e, green curve).No diffraction peaks corresponding to metallic Zn were detected, indicating a complete transformation of metallic Zn nanowires to ZnO.SEM inspection of the oxidized nanowire network revealed the change in the surface morphology of the nanowires from smooth to grainy (Fig. 1c,d) and the increase of radii of the nanowires from 50-100 nm to 100-200 nm.The average thickness of the ZnO nanowire network layer on a glass substrate also increased compared to the thickness of the initial Zn nanowire network by ~ 50%.
Presumably, the increase of the average radii of the ZnO nanowires and, consequently, of the total thickness of the ZnO nanowire network compared to the initial Zn nanowires may be related to the solid state up diffusion model of Zn oxidation 26 .After the initial ZnO layer is formed at the surface of the Zn nanowire, Zn 2+ diffuses from the Zn core to the surface, driven by force generated by a contact electrical field at Zn-ZnO interface, and then forms the ZnO with O 2-in the air.The diffusion process continues until the oxidation is complete, leaving the core of the nanowire hollow.This presumption is supported by the detailed SEM inspection of the crosssection of as-grown ZnO nanowire networks, showing that the nanowires have a tubular structure (Fig. 1f).The tubular structure of the ZnO nanowires was observed throughout the cross-section of the sample; no significant differences in the morphology of the ZnO nanowires in the top and bottom layers of the network were observed.This observation is consistent with the reports on the growth of ZnO nanowires by oxidation of metallic Zn nanowires pre-fabricated by nebulized spray pyrolysis, where the transformation of smooth, fully metallic Zn nanowires to grainy ZnO nanotubes was observed 29 .
The images and schematics of the encapsulation of ZnO nanowire networks are shown in Fig. 2a and b.The drop-casting of the polymer over the ZnO nanowire network results in penetration of the polymer throughout the network to the substrate, as shown in Fig. 2c, illustrating the bottom surface of the encapsulated ZnO nanowire network.The current-voltage curves of synthesized ZnO nanowire networks before and after encapsulation showed linear behavior, indicating good electrical contacts between the nanowires throughout the nanowire network both before and after encapsulation (Fig. 2d and e).
Resistances of different samples of not-encapsulated ZnO nanowire networks were 10 4 -10 6 Ω, depending on the linear dimensions of the samples.The estimated electrical conductivity σ of these samples was ~ 0.15 S m −1 .Encapsulation of the ZnO nanowire networks resulted in the increase of their electrical conductivity σ in comparison to the initial resistance of not-encapsulated network by a factor of ~ 3 for ZnO network encapsulated in PVA (further in text referred as ZnO-PVA), and by a factor of ~ 18 for ZnO network encapsulated in epoxy adhesive (further in text referred as ZnO-epoxy), reaching 0.42 and 2.75 S m −1 respectively (Table 1).
Presumably, the increase in the conductivity of the encapsulated networks may be explained by the interaction between the surfaces of ZnO nanowires and the encapsulating binder.It has been shown previously that PVA/ ZnO films exhibit higher electrical conductance in comparison with the bare ZnO films due to the contribution of the PVA-ZnO interface, which becomes the main conductance path for the electrons due to the chemisorption of water molecules from PVA, which can be described by the following equations 55 : (1) where O v is the ZnO lattice oxygen, and V o is the oxygen vacancy in the ZnO lattice.Water chemisorption (Eqs.1-3) results in the formation of hydrogen bonds with ZnO and oxygen ionization in the lattice sites of n-type ZnO followed by formation of hydroxyl (OH) groups and generation of free electrons, thus creating a conduction path along the ZnO-PVA interface 55 .Regarding the ZnO-epoxy system, it has been shown previously by the other research groups that the interaction of ZnO nanoparticles with epoxy involves the formation of hydrogen bonds between the epoxide groups and free hydroxyl (OH) groups attached to the ZnO surface or absorbed by ZnO water molecules, and results in the formation of dual nanolayer interface between the ZnO and epoxy 52,56 .Performed FTIR characterization of the ZnO and ZnO-epoxy samples revealed the presence of (2)  a low-intensity peak at approximately 3000-3500 cm −1 (Fig. 3), which can be attributed to the -OH absorption peak in epoxy adhesive 57 .The first nanolayer of the interface consists of strongly bonded to each other and the ZnO surface epoxy segments.The high number of hydrogen bonds in the first ZnO-epoxy interface nanolayer mitigates the interface polarization and enhances the transport of charge carriers, making this layer electrically conductive 58 .Presumably, the mitigation of the interface polarization is the reason for the significantly higher electrical conductance of the ZnO-epoxy compared to ZnO-PVA.
Increase in the electrical conductivity of the ZnO nanowire networks after their encapsulation in PVA or epoxy adhesive was accompanied by simultaneous increase of the absolute values of the Seebeck coefficient │S│ by ~ 22% for ZnO-PVA (from 175 to 225 µV K −1 , Fig. 2f, Table 1) and by ~ 225% for ZnO-epoxy (from 175 to 570 µV K −1 , Fig. 2f, Table 1).The increase of the │S│ of the ZnO nanowire networks after their encapsulation in PVA may be related to the presence of oxygen vacancies and other defects formed at the surfaces of ZnO nanowires and the ZnO-PVA interface, serving as charge trapping sites.In the case of the ZnO-epoxy system, the second nanolayer of the ZnO-epoxy interface formed in the ZnO-epoxy system consists of loose epoxy segments, facilitating the easy transfer of charge carriers, as well as having a high probability of defects, which may act as charge trapping sites 58,59 .The high surface-to-volume ratio of ZnO nanowire networks and the presence of charge trapping defects at ZnO-PVA and ZnO-epoxy interfaces may result in a decrease of charge carrier concentration in the ZnO-binder system, which, in accordance with Mott's relation between the Seebeck coefficient and charge carrier concentration would result in the increase of the Seebeck coefficient 60 : where S is the Seebeck coefficient, k B is Boltzmann's constant, e is the electron charge, h is Planck's constant, T is the absolute measurement temperature, m* is the effective mass of the carrier (m* = 0.23 m 0 (electron mass) for ZnO 61 ), and n is charge carrier mobility.The charge carrier concentrations estimated for ZnO, ZnO-PVA, and ZnO-epoxy systems using Eq. ( 4) were 2.5•10 18 cm −3 , 1.7•10 18 cm −3 , and 0.4•10 18 cm −3 respectively.The charge carrier mobility can be estimated by the following equation: The calculated using Eq. ( 5) values of charge carrier mobility were 0.00375 cm 2 V −1 s −1 (bare ZnO), 0.015 cm 2 V −1 s −1 (ZnO-PVA), and 0.4 cm 2 V −1 s −1 (ZnO-epoxy).Probably, the very low values μ estimated for bare ZnO nanowire network are related to the string scattering of the charge carriers due to the sponge-like architecture of the network, consisting of tubular nanowires with grainy structure.Encapsulation of the nanowire network in PVA and epoxy adhesive resulted in the increase of mobility of the charge carrier concentrations by ~ 4 and ~ 106 times, respectively.This result supports the hypothesis that the interface layer between the ZnO and the encapsulating polymer facilitates and enhances the transport of charge carriers.
The calculated PF of the bare ZnO nanowire network was ~ 4.6 nW m −1 K −2 at room temperature, which is insufficient for near-room temperature thermoelectric applications.However, the PFs of the PVA-and epoxyencapsulated ZnO-nanowire networks increased by a factor of ~ 5 and ~ 185, respectively, compared to the www.nature.com/scientificreports/bare ZnO networks, reaching ~ 0.85 µW m −1 K −2 for ZnO-epoxy samples (Table 1).This value was impressively higher than PF values reported by other groups for ZnO-based flexible thermoelectric films (0.06 and 3.5•10 -6 µW m −1 K −2 respectively for ALD-deposited ZnO thin films on polyimide substrates 22 and PEDOT:PSS/ ZnO microparticle composite 40 ), and twice as high compared with the state-of-the-art n-type PVA-encapsulated Bi 2 Se 3 -MWCNT flexible networks (0.4 µW m −1 K −2 ) 48 .The calculated ZT at room temperature of the bare ZnO, and ZnO-PVA were in the range of 10 -8 -10 -7 (Table 1).However, due to the high Seebeck coefficient and relatively low thermal conductivity, the ZT of the ZnO-epoxy sample reached ~ 3•10 -6 , which is by 3 orders of magnitude higher than obtained for the PVA-Bi 2 Se 3 -MWCNT flexible thermoelectric films (Table 1).These results illustrate good perspectives of ZnO-epoxy composites for near-room temperature low-grade heat conversion, as well as in nano-and micro-power applications, such as, for example, environmental health monitoring.The temperature dependencies of the thermoelectric properties of the ZnO-PVA and ZnO-epoxy samples in the target temperature range for low-grade heat conversion (180-380 K) showed that Seebeci coefficient of ZnO-PVA sample was relatively stable in this range, showing gradual decrease down to 80% of the room-temperature value (S RT ) with the decrease of temperature down to 180 K, and increase up to 115% of the S RT with the increase of temperature up to 380 K (Fig. 4a, blue dots).
The temperature increase of the Seebeck coefficient of ZnO-PVA samples was accompanied by a decrease in the resistance, which was expected for semiconductor behavior (Fig. 4b, blue dots).With the temperature increase from 180 to 380 K, the resistance of the ZnO-PVA sample decreased from ~ 230% of room-temperature value R RT to ~ 70% (Fig. 4b).In contrast, ZnO-epoxy showed similar to ZnO-PVA sample behavior of the Seebeck coefficient and resistance at temperatures below 300 K (Fig. 4b, red dots).However, with the temperature increase above 300 K, the Seebeck coefficient of the ZnO-epoxy sample showed a rapid increase, reaching ~ 200% from the S RT at 370 K (Fig. 4a, red dots).Such increase in the Seebeck coefficient was accompanied by the U-torn of the resistance of the ZnO-epoxy sample at temperatures above 300 K, reaching ~ 135% of the R RT at 370 K. Presumably, such behavior of the ZnO-epoxy sample at temperatures above 300 K was related to the changes in the ZnO-epoxy interface, for example, to the temperature-promoted activation of charge traps when the temperature approaches the operational limit of the epoxy adhesive and it starts transition to a viscous state, resulting in changes in charge carrier concentration or charge carrier filtering at the ZnO-epoxy interface, and further increase of the Seebeck coefficient.In turn, the thermal conductivity of ZnO-PVA and ZnO-epoxy did not show pronounced dependence on temperature in the temperature range 200-330 K (Fig. 4c).However, ZnO-PVA samples showed up to 50% lower compared to the room-temperature value k RT thermal conductivity in the temperature range 180-200 K, which may be related to the partial freezing of the PVA (Fig. 4c, blue dots).In contrast, ZnO-epoxy samples showed lower up to 50% thermal conductivity at temperatures around and above 350 K, which may be related to the changes in the epoxy structure at elevated temperatures (for example, partial transition to a viscous state).Temperature dependencies of the ZT of ZnO-PVA and ZnO-epoxy samples showed a gradual increase of the ZT of ZnO-PVA with the increase of temperature, reaching ~ 3.3 higher ZT at 380 K compared to its room-temperature value Z RT (Fig. 4d, blue dots).In contrast, ZT of ZnO-epoxy showed a rapid increase at temperatures above 330 K, reaching ~ 30 times higher ZT value at temperature 370 K compared to the Z RT (Fig. 4d, red dots).These results suggest epoxy-encapsulated ZnO nanowire networks for low-grade waste heat conversion applications in the temperature range 330-370 K, for example, from hot pipes or other similar heated surfaces.
In addition, bending tests performed for PVA-encapsulated ZnO nanowire samples (Fig. 5a) showed that the resistance of the sample tends to decrease reversibly with the decrease of the bending radius (Fig. 5b).Presumably, the decrease in the resistance is related to the increased number of direct mechanical and electrical contacts between the ZnO nanowires when bent, resulting in an increase in the effective cross-section of the sample.More rapid return to the initial resistance values during unbending of the sample may be related to the elastic properties of the PVA as well as to the changes in the electrical contacts within the ZnO nanowire network during the bending, i.e., destruction of several contacts and formation of a new ones.The total decrease in the resistance of the sample was within 6% from the initial resistance R 0 of the not-bent sample (Fig. 5b).
ZnO-PVA encapsulated samples also showed very good stability during 100 consequent bending cycles down to a 5 mm radius (Fig. 5c).The resistance of the sample showed an insignificant increase within 0.5% from the initial resistance value R 0 of the not-bent sample during the first ten bending cycles, followed by stabilization of the resistance within 1% from R 0 for the subsequent 90 bending cycles.

Conclusions
Self-assembling ZnO nanowire networks were fabricated by thermal oxidation in air pre-deposited by thermal evaporation metallic Zn nanowire network.XRD studies of the oxidized Zn nanowire networks proved a complete transition of metallic Zn to ZnO.At the same time, an in-depth SEM investigation revealed that the resulting ZnO nanowires have tubular structure, presumably resulting from the diffusion of Zn 2+ from the core of the nanowire to the surface during the oxidation process.As-grown ZnO nanowire networks showed electrical conductivity of approximately 0.15 S m −1 and Seebeck coefficients of ~ -175 µV K −1 , which is comparable with some theoretical and experimental Seebeck coefficient values reported for ZnO nanowires obtained by thermal evaporation of ZnO (~ − 110 to 190 µV K −1 ).After the encapsulation in polyvinyl alcohol (PVA) and commercially available epoxy adhesive, ZnO nanowire networks showed a significant increase in the electrical conductivity by a factor of ~ 3 and ~ 18 for ZnO-PVA and ZnO-epoxy systems, respectively, reaching the value of 2.75 S m −1 for the epoxy-encapsulated sample.The increase in electrical conductivity was attributed to the formation of electrically conductive channels at the ZnO-encapsulating polymer interface.The increase in the electrical conductivity after the encapsulation of the ZnO nanowire networks was accompanied by the increase in the Seebeck coefficient of ZnO-PVA and ZnO-epoxy systems in comparison with not encapsulated ZnO nanowire networks by ~ 22% and ~ 225% for ZnO-PVA and ZnO-epoxy systems respectively, which was attributed to the changes in charge carrier concentrations in these systems caused by the presence of oxygen vacancies on the surface of ZnO, as well as defect sites at the ZnO-polymer interface, which may serve as charge traps.The absolute value of the Seebeck coefficient of the ZnO-epoxy sample, measured at room temperature, reached ~ − 570 µV K −1 , which is comparable to the best values of the room-temperature Seebeck coefficient reported for ZnO powders-based densified ceramics (~ − 600 µV K −1 ) and is significantly higher compared to the Seebeck coefficients reported for the vast majority of the state-of-the-art ZnO-based thermoelectric materials and ceramic composites, measured at elevated (500-1000 °C) temperatures.The maximal power factor and ZT at room temperature were shown by the epoxy-encapsulated ZnO nanowire networks and reached 0.85 µW m −1 K −2 and ~ 3•10 -6 , respectively, which significantly exceeded the power factor values reported previously for similar (ZnO-and/or non-conductive polymers-based) flexible thermoelectric materials.A study of the temperature dependencies of the thermoelectric properties of the encapsulated ZnO nanowire networks in the target temperature range of 180-380 K showed that ZT of these materials can be improved by a factor of ~ 3 and ~ 30 for ZnO-PVA and ZnO-epoxy samples respectively with the increase of temperature up to 350-380 K.In addition, bending tests of the ZnO-PVA samples showed a reversible decrease in the resistance of the sample up to ~ 6% from the resistance value of the unbent sample when bent down to a 5 mm radius, which was attributed to the formation/destruction of new direct electrical connections between the ZnO nanowires during the bending.ZnO-PVA samples also showed very good mechanical and electrical stability during 100 consecutive bending cycles down to a 5 mm radius.The fluctuations of the resistance of the sample during the repetitive bending tests did not exceed 1% from the initial resistance of the sample.The simplicity of fabrication, flexibility, and significantly enhanced thermoelectric properties of PVA-and especially epoxy-encapsulated self-assembling ZnO nanowire networks open a new path for applying such systems in near-room temperature flexible thermoelectrics for low-grade heat conversion.

Figure 1 .
Figure 1.SEM images of deposited on a glass substrate Zn nanowire networks at low (a) and high (b) magnification; ZnO nanowire networks after oxidation of pre-deposited Zn nanowire networks at low (c) and high (d) magnification; XRD patterns of Zn nanowire networks (black curve) and ZnO nanowire networks (green curve); (f) left panel-an SEM image of the cross-section of as-grown ZnO nanowire network, revealing nanotube-like structure of the ZnO nanowires; right panel-magnified SEM images of the selected areas.Yellow arrows indicate nanowire cross-sections where tubular structure can be observed.

Figure 2 .
Figure 2. (a) photographs and (b) schematics of the encapsulation process of the ZnO nanowire networks; (c) scanning electron microscope image of the bottom surface of the encapsulated ZnO nanowire network; (c,d) Examples of current-voltage curves of ZnO nanowire networks: (c) before (black curve) and after (green curve) encapsulation in polyvinyl alcohol (PVA), and (d) before (black curve) and after (red curve) encapsulation in the epoxy adhesive; (f) absolute values of thermally generated by ZnO nanowire networks voltage U T vs. temperature difference between the sides of the network: before encapsulation (black circles), after encapsulation in PVA (green circles), and after encapsulation in epoxy adhesive (red circles).

Figure 4 .
Figure 4. Temperature dependencies relative to the room-temperature (RT) values of (a) Seebeck coefficient; (b) resistance; (c) thermal conductivity; (d) thermoelectric figure of merit of ZnO nanowire networks encapsulated in PVA (blue dots) and epoxy adhesive (red dots).

Figure 5 .
Figure 5. (a) A schematic indicating the bending direction of the PVA-encapsulated ZnO networks.(b,c) relative changes in the electrical resistance of the PVA-encapsulated ZnO networks vs (b) its bending radius and (c) consecutive bending cycles.R is the resistance of the bent sample, and R 0 is the initial resistance of the notbent sample.

Table 1 .
2lectrical conductivity σ, Seebeck coefficient S, power factor PF, thermal conductivity k, and calculated figure of merit ZT of the ZnO nanowire networks before and after encapsulation, measured at room temperature under ambient conditions and comparison with the results published for other similar flexible thin films.*Thermalconductivitydata for the bare ZnO are taken from the ref.2.