Semiconductor-to-Metal-like Transition Behavior under Temperature Variation for Inkjet Printed PEDOT:PSS Tracks Embedded in Polymer

Herein, it is intended to show the effect of embedding an inkjet printed poly(3,4-ethylenedioxythiophene) polystyrene sulfonate (PEDOT:PSS) track in an insulator polymer, impacting its electronic transition behavior, as a consequence of temperature variation. A transition from semiconductor-to-metal-like behavior is observed, when the temperature is seen to exceed a certain value, which is of a nonchemical origin. Both the presented experimental and simulation results show how this transition really occurs. The proposed physical mechanism for explaining such a behavior is veri ﬁ ed with good repeatability. The main conclusion indicates consideration of special precautions, while enclosing inkjet-printed PEDOT:PSS-based tracks or sensors operating under ambient conditions, along with ﬂ uctuations. This conclusion can potentially be applied to any other inkjet printed conductive organic polymer ﬁ lm embedded in an insulator that ful ﬁ lls the conditions encountered in the experiments. The impact of this effect may be reduced and mitigated by using inkjet printing, in combination with other additive manufacturing technique. The results presented here are considered very important, as they lay the foundation for the correct compensation of the thermal drift of organic electronics-based circuits.

mechanical pressure, temperature, and infrared radiation detection. Any parasitic signal fed by the environment will cause sensor malfunction and jeopardizes the sensing operation. If we consider an example of a temperature sensor that gets utilized in an industrial or medical environment, the improper measurement of temperature reading will pose a threat to a correct diagnosis or could compromise the production standards. In the case of PEDOT:PSS-based infrared sensors, it would work on the bolometer principle, i.e., the resistance changes as the absorbed incident infrared radiation changes the temperature upon heating. [14] It is thus very essential to retain the resistance variation protected from any other influences (humidity, chemicals under vapor form) except that of temperature.
One of the straight forward ways to protect the PEDOT:PSS layer is to enclose/cover it, by inkjet printing, with an electrically insulating layer that allows an appropriate passage of the signal to be detected (pressure/temperature/infrared radiation). Thus, a low-cost yet digitally manipulative additive manufacturing technique like inkjet printing technology can be exploited to produce low-cost sensors, on demand. The device architecture considered in the article consists of a 3D sandwich-type structure composed of a PEDOT:PSS track embedded between two dielectric polymer layers (substrate and insulator). The 3D sandwich structure is realized using a laboratory scale inkjet printing technique. This type of structure was chosen because of several reasons: 1) PEDOT:PSS being well known to be a promising material for temperature and infrared sensors; [14] 2) implementation of encapsulation technique for PEDOT:PSS tracks to shield them from the environmental influences; 3) exploration of the operational range for implementation in temperature or infrared sensors as practical use. All these reasons contribute strongly toward the envisaged detailed investigation, of how the device behavior changes with respect to temperature change. We show experimentally, and explain how such a protective encapsulation can dramatically influence the temperature/heat flux measurement process, being capable to alter the functionality of the sensors based on PEDOT:PSS and, more generally, within organic semiconductors. From the experiments, it is known that an increase in humidity increases the track resistance of the air-exposed track, which is in accordance with the literature, [17] and reduces its thermoelectric voltage. Thus, one more condition is found that can be crucial to be fulfilled by the protective layer for a correct operation of the thermal sensors, except of those mentioned in the preceding paragraph. Additionally, considerations are also presented on how another low-cost additive manufacturing technique can be used to resolve this supplementary condition and ensure a correct operation of the thermal sensors. The use of a laminated object manufacturing (LOM) technique as another alternative offers the capability to encapsulate the PEDOT:PSS track in a void cell, whose ceiling (protective cover) can be kept independent of the direct contact. Besides, the inkjet printing alone cannot offer the creation of such void cells. This way, the electric response of the heating-based sensors remains unaffected by its encapsulation. It is for the first time when such an issue of the electric behavior fluctuations under thermal stimulation in 3D sandwich-type structures based on PEDOT:PSS and ways of its improvement are investigated and in-depth addressed. It is especially important now when research in the exploitation of PEDOT:PSS-based sensors in various application areas gains traction. [14,18] Knowing the real thermal drift behavior of such devices allows us for the correct compensation of the circuits containing organic electronics that are supposed to operate on a certain temperature range. Mitigation of this effect is vital in inkjet printed thermal sensors, among other sensors that are supposed to work in a wide temperature range and for organic-based energy harvesting devices. The article is organized such that Section 2 describes the methodology used for the experiments and the experimental results. The explanation of the physical phenomenon and simulation results is also presented here, whereas Section 3 is devoted to discussions. Among other aspects, the use of LOM technique is described here, and finally, the article summarizes with a conclusion section.

Methodology and Experimental Results
In order to determine, if the PEDOT:PSS material needs indeed any protection, a PEDOT:PSS track was deposited by inkjet printing on a polycarbonate substrate with the help of a laboratory scale DMP-2800 inkjet printer from Fujifilm Dimatix, 10 pL based DMC cartridges (drop space 40 μm -center to center distance between two drops). The bulk polymer PEDOT:PSS (high conductivity grade) was procured from Sigma-Aldrich and was dispersed in water with about 4 wt% concentration. The process of deposition via inkjet printing was repeated for 10 times in order to obtain a layer thickness of at least 1 μm for the PEDOT:PSS track. The track dimension was set to 20 mm length and 1 mm width. After the deposition process was accomplished, the printed track was thermally posttreated at 100°C (oven) for 30 min in air for removing the excess water from the printed layer. According to the optical profilometric measurements made by using a Photomap 3D system (FOGALE nanotech, France), the thickness of the tracks was found to be about 1.4 μm (see Figure 1).
The reconstruction of the lightwave phase is shown in the top left of Figure 1a; the green part indicates the regions of the field of view where phase retrieval is uncertain. This usually happens to occur during the measurement of rough surfaces. The same explanation applies for the image center below, where some spikes and sharp valleys are observed. These are because of the same uncertainty in the phase retrieval, as in the top left image where some narrow green regions can be noticed, along the length (left-right direction) of the track. As results from Figure 1b (bottom center) are analyzed, it could be said that the variation in the layer thickness profile was of the order of 10%. The electrical resistance of the inkjet printed track was measured as a function of temperature. The resistance was measured with a Keithley 2700 Multimeter/Data Acquisition System (Integra series). All the contacts to the PEDOT:PSS tracks were made with copper wires connected to the tracks with Carbon wire glue (Anders Products, USA). The present test was repeated over three different days. While the ambient temperature variation from day-to-day at the time of performing the measurement was found to vary less than 2°C, the humidity varied by 20%. The resistance versus temperature measurement for the first day within several samples is presented in Figure 2, while the day-to-day variation of sensitivity is shown in Table 1.
For evaluating the response of the PEDOT:PSS tracks, the temperature was increased slowly in steps of 2°C, swapped from 25°C to 40°C. Heating cycles were pursued, followed with a natural cooling down of the samples at ambient temperature. During this stage, the corresponding resistance values were noted, and the results were plotted in the figure, showing the change in resistance during heating and cooling cycles, inferring to a small hysteresis. Initially, the value of the resistance was found to be 3.48 kΩ, with an absolute error of %0.1 Ω. The implemented temperature was calibrated with a Pt sensor with an absolute error of 0.1 AE 0.05°C. From the depicted figure, it could be observed that as the temperature was increased/decreased, the electrical resistance proportionally decreased/increased, which pointed out an inverse correlation (tolerance AE 14 Ω). The same electrical characteristic was also observed for the same PEDOT:PSS track for all the three individual days. For the devices, a very negligible variation of the device behavior was observed, as the measurements were performed iteratively after a designated time span (1 week), from the date of manufacturing for the devices, indicating that the devices were already stabilized regarding its electrical characteristics. Therefore, the electrical resistivity of the PEDOT:PSS tracks was derived, corresponding to the variation in electrical resistance and layer geometry, with respect to the unit temperature step. This electrical resistivity could then after be standardized in form Ω/°C with respect to the thermal coefficient. The temperature coefficient resistivity values are summarized in Table 1, over 3 d. As mentioned earlier, the electrical resistance was seen to decrease with the increase in temperature over the considered temperature range. This is a behavior that is typical for a semiconductor/conjugated polymer as PEDOT:PSS is. We have observed from Table 1 a response variation of more than 20% has taken place between   www.advancedsciencenews.com www.pss-a.com the consecutive days, which lets the use of such a sensor material cumbersome for practical applications. We may thus conclude that the PEDOT:PSS track must be protected/encapsulated somehow from the influence of the environment, in order to acquire repeatable and reliable sensor response, as it is shown in Figure 3 for a PEDOT:PSS-based inductor coil that is embedded in an epoxy resin. The 3D sandwich-type inductor structure was manufactured on an acrylic precoated polyethylene naphthalate (PEN) substrate (125 μm thickness, Teonex Q65 HA from Dupont Teijin) with the help of a DMP-2831 inkjet printer from Fujifilm Dimatix, 10 pL based DMC cartridges (printing resolution of 1693 dpi, 2 layers). The 3D PEDOT:PSS inductor was deposited, using a commercial ink CLEVIOS F HC Solar from Heraeus Deutschland GmbH & Co. KG, [19] in two steps (inductor and bridge, respectively) on an acrylic-coated PEN substrate. After the deposition of the first layer, i.e., PEDOT:PSS inductor along with the initial contacts, it was thermally posttreated (cured) at 120°C for 10 min in a vacuum oven. Then, the deposition of epoxy resin was accomplished for the motivation to bypass the semiconductor track (bridge) from the interior of the inductor to the external electrode (shown as the dark blue horizontal segment in Figure 3a). For this reason, two specific locations were kept open, where the second layer of PEDOT:PSS could be extended from the top, just after this step. The deposited epoxy resin was again thermally cured at 130°C for 30 min also in the vacuum oven. Finally, the third PEDOT:PSS layer was deposited forming the contact stripe, which now establishes the serial electrical resistance connection between the inside and outside ends. As the epoxy layer was deposited all over the active inductor coil area, i.e., the conductive track, the inductor structure could then be defined as totally embedded. Copper wires were attached to the inductor coil by using carbon paste deposited exactly to the inkjet printed silver contacts that were made to the PEDOT:PSS inductor electrodes. The contact resistance was found to be less than 1 Ω. The DC resistance of the inductor was measured with the same Keithley multimeter and found equal to 33.06 kΩ at 22°C. The variation of the electrical resistance (DC) at the inductor was measured with respect to the change in temperature. The temperature was slowly varied in steps of 2°C, starting from 22°C to 40°C. A first heating cycle was pursued, followed with a natural cooling down of the sample at ambient temperature. The obtained results in form of change in resistance are plotted in Figure 4.
As it can be noticed from Figure 4a, the PEDOT:PSS inductor behaves like a semiconductor up to 30°C, i.e., its electrical resistance decreases with an increase in temperature. Above this temperature, the behavior changes unexpectedly and becomes that of a metal, i.e., resistance increases with an increase in temperature. On the other hand, once a cooling cycle is implemented, the inductor shows a semiconductor-like behavior starting from the beginning. During the cooling cycle, the tendency of change in resistance for the inductor coil was seen to follow the similar path. The temperature coefficient of resistance for the case of the www.advancedsciencenews.com www.pss-a.com inductor is shown in Table 2. We have noticed a very small hysteresis of the inductor resistance, immediately after cooling down to room temperature, of the order of few tens of Ohm; after that, the value recovered slowly to the initial ones. We consider that this is due to the mechanical relaxation of the conjugated polymer along with the embedded epoxy, soon after the heating-cooling cycle of the sandwich-type structure.
Compared to the case of not encapsulated PEDOT:PSS linear tracks that behaved like a semiconductor over the entire temperature range, the PEDOT:PSS material within the inductor seems to change its properties. Comparing the two cases, we can affirm undoubtedly that PEDOT:PSS cannot change its material properties just because of embedding. Moreover, the change of behavior is continuous, not abrupt, suggesting that there cannot be any phase transition-like phenomenon that explains the behavior. A chemical reaction at the interface between the PEDOT:PSS and the encapsulation layer is also excluded because of two reasons: 1) the chemical reaction would alter the conductivity at any temperature, not only after a certain temperature is exceeded; 2) the chemical reaction would take place only at the interface and not in the bulk volume of the semiconducting track because the solvents involved with the PEDOT:PSS ink and for the encapsulating epoxy layer are of different origins and are immiscible. Such a change at the interface would not alter the volume conductivity behavior. One mechanism that can possibly explain the experimental data is the contrast of the two thermal expansion coefficients of the two materials, i.e., PEDOT:PSS and epoxy resin, respectively. According to the literature, [20] PEDOT:PSS has a thermal expansion coefficient of 53 Â 10 À6 K À1 , while epoxy has a value of 92 Â 10 À6 K À1 . [21] Therefore, the mechanism that can be interpreted is, while the temperature is increased slowly, the resistivity of the printed PEDOT:PSS layer decreases, as in any semiconductor or conjugated polymer. However, at the same time, the PEDOT:PSS track experiences an mechanical elongation because of the higher thermal expansion of epoxy resin as compared to the inductor. It is shown by Latessa et al. [13] that the resistance of PEDOT:PSS increases with increase in the elongation strain. This way, two mechanisms compete for the variation in resistance (in fact in resistivity at the physical level) of the material with respect to temperature. Above a certain temperature, which is 30°C in the case of the inductor, the stretching stress takes the lead, and the overall resulting behavior is similar to that of the metal, i.e., an increase in resistivity with the increase in temperature. During the cooling cycle, the PEDOT:PSS is subjected to a mechanical compressive stress, which according to our experimental data does not contradict to the variation of resistivity with respect to the change in temperature. Here, we must underline the fact that under the mechanical stretching due to thermal expansion, the electrical conductivity of PEDOT:PSS occurring through free charge carrier percolation is decreased due to the mechanical elongation. On the other hand, during the cooling-down phase, i.e., below the ambient temperature, there is a compressive strain that at least preserves the percolation paths of the free charge carriers. At this present stage, we may not be able to clarify whether the percolation is further improved due to compression or not.
In order to check our hypothesized mechanism for the inductor case, we performed finite element simulations for determining the values of the strain, in the inductor at the temperatures used in the measurements. The simulations were performed by using ANSYS Multiphysics 2022 version software package. We have used a simplified model of the inductor and considered only those parts that are contributing majorly to the electrical resistance of the device. Thus, the contact pads and underpassing elements from interior end of the coil to the exterior were neglected. The material data were taken from refs. [20][21][22] and are presented in Table 3.
The spatial distribution of the strain in the inductor is depicted in Figure 5, while the extreme and average strain values are presented in Table 4 for each of the temperatures at which measurements were made. As it can be seen from Figure 5, the strain is relatively constant throughout the inductor, with the extreme values found especially at the corners.
We define the gauge factor G as the proportionality factor establishing the relative variation of the electrical resistance with the strain ε because of the piezoresistive effect (Latessa et al. [13] ) in the organic conjugated polymer/semiconductor as follows where R 0 is the initial (measured) resistance, while ΔR m is the difference between calculated and measured resistances,   respectively. The net variation ΔR net of the electrical resistance is given by the expression where ΔR m > 0 is due to the piezoresistive effect produced by the strain resulting from the different thermal expansion between PEDOT:PSS and its embedding material; ΔR T < 0 is due to the change of semiconductor resistivity because of temperature variation.
For calculating ΔR T , we have used the electrical resistance variation between ambient temperature (22°C) and the first step of temperature increase (24°C). The reason for that resides in the fact that the variation is small enough and near room temperature; in this way, the effect of thermal expansion will be negligible at this step. By dividing ΔR T by 2°C, we are able to obtain a temperature coefficient of resistivity β = À9.451 Ω°C. As we see in Figure 2 for the not encapsulated PEDOT:PSS track, the resistance decreases almost linearly, with the temperature increasing up to the maximum temperature used in the measurements (the regression factor R 2 = 0.998 for the linear fit). We may thus write where ΔT represents the temperature variation. At 30°C, the behavior of the inductor changes from semiconducting to metallic. For computing G, we equal (1) and (3) for ΔT = 8°C. We obtain the gauge factor G = 6.189. This value compares well with data described in the literature, being between the value of 0.48 mentioned in ref. [23] and 17.8 mentioned in ref. [13]. The respective two research teams have used different mechanical excitation mechanisms. In order to check the value of G, we calculated the values of the electrical resistance at all the measurement temperatures and compared them to the measured ones. The relative errors indicates us if our model for the resistance variation with respect to temperature is correct or not. From Table 4, we take the average strain value corresponding to the respective temperatures. By performing direct mathematical calculation, we obtain the data presented in Table 5. From Table 5, it could be noticed that the relative error does not exceed 0.2% (absolute value) at any of the measurement temperatures. This is an excellent agreement between estimated and measured values, even if we used a simplified (linear) model for ΔR T variation with respect to temperature. The graph showing the ratio of calculated to measured resistance at the considered temperatures is shown in Figure 6.   As it can be observed from Figure 6, the ratio of the measured to the calculated resistance is very close to 1 at each temperature in the measured range, the variation on the other hand staying within the limits between þ0.15% and À0.20%. This is a good agreement between theory and experiment. The reason for which there is not a perfect match is due to the simplified assumptions we have made in our model. For example, we have considered that both the temperature coefficient of resistivity and the thermal expansion coefficients do not depend on the temperature. We may thus affirm that the mechanism proposed by us is correct, and we have an apparent semiconductor-to-metal transition behavior of the embedded PEDOT:PSS track instead of a true transition. This conclusion can also explain why there is no such effect noticed on the not encapsulated PEDOT:PSS track deposited onto polycarbonate foil. The reason can be justified; polycarbonate has a thermal expansion coefficient close to that of PEDOT:PSS (68 Â 10 À6 K À1 versus 53 Â 10 À6 K À1 ). Thus, we believe that the effect of mechanical strain on the considered temperature range will be negligible in this case. A brief discussion of the prevailing phenomena is presented here. The variation of the resistivity with the increase in temperature can be explained by four mechanisms. The first two mechanisms are responsible for the decrease of resistivity, while the next two results in an increase in resistivity with respect to the temperature. The first one is based on the increased probability of hoping of free charge carriers between PEDOT cores as temperature increases. The second one is explained by the desorption of water molecules from the PEDOT:PSS track (when exposed to air) as temperature is increased. As it is known from the literature, PEDOT:PSS is declared to be a hygroscopic material, [17] making it useful for sensing humidity. The material's resistivity has a dependence to the humidity, which is primarily dependent on the water adsorption-desorption by the PSS component. When humidity is high, the insulating and hydrophilic PSS shell absorbs water and swells, resulting in an increase in the distance between the adjacent conductive and hydrophobic PEDOT cores. This leads to an increase in the resistivity of the PEDOT:PSS because charge carrier hoping in a conjugated polymer is the main conduction mechanism. When humidity decreases, water molecules desorb out from PSS, which causes the volume reduction of the PSS and hence of the distance between adjacent PEDOT cores. Therefore, the hoping phenomenon is enhanced, and the resistivity is reduced. The third mechanism consists of an increase in resistivity with temperature, as a consequence of increased scattering of free charge carriers on the thermal (vibrational) excitations of the molecules, while the fourth one is justified with the thermal expansion, which increases the hoping distance between PEDOT cores. The latter one is particularly important in our case because the charge transportation is triggered by the pullout produced by the more thermally expandable layers embedding the PEDOT:PSS track. Overall it is expected that these four mechanisms combinedly imply, which results for a nonembedded/encapsulated PEDOT:PSS track into decrease of resistivity with the increase in temperature, that is considered herein. Based on the data from Figure 2, we may consider that the third mechanism, that is, scattering-based mechanism, does not manifest significantly in the temperature range considered by us because the decrease in resistance here is uniform and shows no tendency to flatten. Also, in the case of the inductor, the second explanation, humidity driven evaporationbased mechanism is missing because the device is embedded in nonpermeable polymers. Thus, in the case of embedding strategy, things change as observed experimentally, and the fourth mechanism, forced by the thermal expansion of the dielectric polymer encapsulation, takes the lead at temperatures above 30°C. The effect of temperature on carbon paste contacts and copper wiring is also excluded because the linear, air-exposed PEDOT:PSS track containing the same materials does not show a similar temperature behavior (see Figure 2).

Discussion
There are two ways to diminish the effect of the thermal expansion on the behavior of embedded PEDOT:PSS-based electronic components. One of the most straightforward one would be to use an alternative embedding materials other than epoxy resin, for example, polymers that have a thermal expansion coefficient closer to that of PEDOT:PSS. The second route, according to ref. [20], would be to use PSS deprotonation. As it is known, [20] the PEDOT and PSS units are bonded by a dipole-dipole interaction between S in PEDOT and H in PSS. On the other hand, the C─C bond in a benzene ring (PSS) or thiophene (PEDOT) is up to 13 times stronger than the S─H bond. From this perspective, it follows that by adjusting the population of the S─H bonds by deprotonating PSS, the linear coefficient of thermal expansion can be changed. An increase of the thermal expansion coefficient by 57% is achievable, which is described in the publication from Music et al. [20] Another possibility to overcome the change in behavior is to use another additive manufacturing technique such as LOM. In this case, the PEDOT:PSS track is deposited via inkjet technology onto a substrate. After the inkjet deposition, the first layer is laminated to the substrate, having an opening where the track and its contacts are located. Then, a second layer is laminated to the first one, leaving the PEDOT:PSS enveloped within the closed cell, where the atmosphere is stable and encapsulated. This second layer lets a lateral opening for the external connections of the contacts for the PEDOT:PSS track, which could be either the inductor or any passive electronic device, that, after connecting the necessary wiring, are sealed. The overall www.advancedsciencenews.com www.pss-a.com schematic of the structure and technology steps is presented in Figure 7.
From the schematic, it can be noticed that this type of LOM-based embedding process is basically not applicable for developing pushing-type pressure sensors (as example) because this type of sensors needs intimate mechanical contact with the top layer in order to allow transmission of the exerting pressure directly to the PEDOT:PSS layer. The embedding process can still potentially be used here, but with the exception of the effect mentioned above, where the tuning of sensor response will remain mandatory. Another use of a cell-contained device architecture could be in the manufacturing of temperature sensors, where the electronic mechanism can be utilized by means of compensation for the thermal expansion, along with suitable software tools. While we have shown here the behavior changing effect only for the case of PEDOT:PSS, we may confirm that this can be projected to any other implemented inkjet printed set of polymeric materials, i.e., the organic conjugated polymer/ semiconductor embedded between two insulating layers, that have different values between their thermal expansion coefficients. In the particular case of PEDOT:PSS, the maximum operational temperature of 55°C is suggested, for suitable usage in electronics. Above this temperature, structural and chemical changes may occur. [24] One another particular field of application for which our results are of high interest is PEDOT:PSS-based thermoelectric devices. [25,26] A quite similar effect, but not of the same behavior change, can potentially appear also for the inkjet printed metallic tracks (gold, silver, copper, nickel nanoparticles based layers). In this case, the temperature coefficient of resistivity will remain metallic, but might show a higher value than that of an unembedded metallic track. This could be due to the fact that together with the natural variation of resistivity with temperature, there will be also a superposition of the thermal expansion effects that would stretch the metallic track, and thus increasing its resistivity. The electrical characteristic of piezoresistivity for metals is low, compared to semiconductors. Moreover, the metal nanoparticles are sintered during the postdeposition step, [27][28][29] in order to form a continuous conducting track. Thus, we may expect that the effect of piezoresistivity can also come into the play, even if enhancement of the net response to temperature variation is small. However, if the nanoparticles are not fully sintered, the change in resistivity caused by changes in the contact surface area between neighboring particles may add a substantial contribution to the overall variation in resistivity, with regard to the temperature change. It remains open to be determined, whether a hysteresis-like behavior of the track resistivity can take place or not, in the case of metal nanoparticles after several heating and cooling cycles. The two mechanisms might sum up, so that the overall measured temperature coefficient of resistivity is higher. This is good for applications where the thermal signature (either temperature or infrared radiation) is intended to be detected. The effect might be considered having originated from a different source, apart from that is described by us, for example, it could be due to a problem with the contacts under thermal stress. Here, we can underline that: 1) the contacts were the same for both the deposited PEDOT:PSS track on polycarbonate (unembedded), and for the embedded one on PEN substrate and 2) the temperature was always uniformly distributed over the whole surface area of the two devices, when performing resistance versus www.advancedsciencenews.com www.pss-a.com temperature measurements. As we have not observed such an effect for the case of unembedded devices (Figure 2), we may infer that the contacts are not contributing to this effect and the mechanism provided by us is in order. Besides, there is also no suspicion of a Seebeck effect because the temperature implemented was kept the same at both the contacts of the measured devices, and both the contacts for all the devices were similar.

Conclusion
We have shown experimentally and explained theoretically the mechanism by which an inkjet printed track of organic or conjugated polymer/semiconductor such as PEDOT:PSS embedded within an encapsulated polymer insulator reversibly changes its temperature coefficient of resistivity, from semiconductor-to-metal-like regime, as the temperature exceeds a certain value. This transition temperature value depends on the thermomechanical properties of the materials, forming the sandwich-type structure embedding the organic conjugated polymer/semiconductor. The material itself remains physically (electrically) a semiconductor; however, the piezoresistive effect takes the lead with respect to the genuine temperature coefficient of resistivity of the material, as the temperature is increased and produces the respective semiconductor-to-metal "transition". The effect is very important in practical applications, where such environment-sensitive organic polymers/semiconductors are used, from sensors mounted in automation systems to robotics-driven industrial manufacturing, and finally to energy harvesting systems, e.g., photovoltaic modules, that contain contacts/electrodes made with this type of material and protected by transparent polymers. A good knowledge of the thermal drift of a device ensures a predictable behavior, when the temperature varies and, by that, allows a robust compensation scheme in the circuit. A drift compensation based only on the semiconducting assumption for the material will be erroneous, as demonstrated by our work. While this phenomenon may be cumbersome for certain applications, we indicate ways of reducing or even mitigating this effect, either by using a careful choice of the semiconductor and insulating materials forming the 3D sandwich structure or by protecting materials in a void cell isolated from the ambient condition. The latter innovative approach was accomplished by using an additive manufacturing technique such as LOM for embedding the track in the respective void cell.