The research and development efforts on flexible hybrid electronics (FHEs) have been rising rapidly due to its wide range of applications in healthcare, consumer products, automotive, aerospace, and energy industries. [1][2][3][4][5][6][7] The conventional lithography-based processes for manufacturing electronics are often complicated, time and energy-consuming, relatively expensive, and produce a large amount of waste materials (washing-cleaning-disposal requirements). [8] Thus, much effort has been put toward inventing new techniques for direct printing and additive manufacturing of electronics on desired flexible and rigid substrates, [9][10][11][12][13][14][15][16][17][18] such as polyimide and polyethylene terephthalate (PET). [19][20][21] Different fabrication strategies have been proposed for additive nanomanufacturing of electronics. [22][23][24][25][26][27] The most common direct-write technologies in the field of printed electronics are inkjet printing (IJP) and aerosol jet printing (AJP). [6,[28][29][30][31][32] These technologies allow the precise deposition of liquids containing functional materials. Recently, Patil et al. printed silver (Ag) nanowires (resistance < 50 Ω sq -1 ) on flexible substrates via inkjet printing. [33] Gilshtein et al. demonstrated inkjet-printed indium tin oxide (ITO) patterns with a resistivity of 3.1 × 10 -3 Ω cm on soda-lime glass. [34] Chen et al. showed aerosol jet printing of Ag with a sheet resistance of 1.13 × 10 -2 Ω/m 2 on cellulose fiber paper substrate. [31] However, the current ink formulations used in IJP and AJP processes make device fabrication more complicated because they require toxic solvents and additives that limit the substrate's compatibility and hinder the device's performance. [35,36] Also, nozzle clogging in IJP is a common and extremely complex phenomenon. [37] Furthermore, to guarantee a high conductivity of the printed structures, aerosol printing usually requires a high sintering temperature (e.g., ~280 °C) and a long sintering time (e.g., 12h on a glass substrate). This limits the AJP to only a few types of substrates. [31] As a result, a new direct printing and patterning method is needed to overcome the challenges. Moreover, the new printing and patterning methods also need to be capable of printing structures with good reliability under repeated mechanical bending and unbending processes.
Motivated by these challenges, here, we demonstrated a novel additive nanomanufacturing technique recently developed in our lab for manufacturing FHEs and sensors by printing conductive Ag and ITO on different flexible platforms such as polyimide and PET substrates.
Conductive Ag is typically used to print electronic tracks, electrodes and antennas. [38,39] Transparent conductors such as ITO are utilized in the printing of components such as flexible strain sensors, solar cells, and organic light-emitting diodes (OLEDs). [40][41][42][43] The main principle of our ANM printer is the in-situ formation of pure and dry nanoparticles produced by condensation of laser-plasma plume in atmospheric Ar pressure, which is then guided through a nozzle onto a substrate that is sintered in real-time, forming the desired patterns. The mechanical reliability of ANM-printed samples has been shown by investigating the resistance change of the samples via bending and cycling tests. FHEs and sensors such as near-field communication (NFC) tag antenna, ITO-based strain sensor and lighting circuits were fabricated on polyimide and PET substrates to demonstrate their performances in this printing approach. This novel ANM and device manufacturing method is a promising technology for future manufacturing of the FHEs and sensors directly from dry and pure nanoparticles generated and sintered in on-demand and in real-time. Figure 1 shows the schematic illustration of the ANM printer consisting of a microchamber, a rotating target, a nozzle, and a gas flow feeder system. A Coherent COMPex excimer laser (KrF, pulse duration 25 ns) beam was divided into two paths; one for laser ablation (LA) and the other for laser sintering (LS) processes in the ANM system. These beams were focused onto the surface of targets and substrates via UV convex lenses. Depending on the type of the materials and substrates, LA and LS beams' energy were adjusted to achieve the formation and sintering of desired nanoparticles. Focused LA beam hits the rotating targets, which will result in the formation of plasma plume that interacts with background Ar gas resulting in condensations and formation of nanoparticles on-demand. These nanoparticles are then directed toward the substrate through a nozzle with different diameters ranging from 100 µm to 300 µm using a carrier gas. The LS beam with suitable energy sinters and crystallizes the arrived nanoparticles onto the substrate's surface in real-time. As more nanoparticles arrive, the deposition increases and forms a continuous film.
The substrates were placed on an XY positioning stage and moved according to the preprogrammed path. The distance between the apex of the nozzle and the substrate was kept constant at ~ 0.5 mm. Commercially available Ag and ITO targets with 99.99% purity were used for the ablation and formation of nanoparticles. The diameter of the nozzles used in ANM printer determined the feature size of the printed lines. The thickness of the printed lines was controlled by printing speed as well as the number of layer-by-layer printed paths.
It should be noted that since we used only an excimer laser for the ablation and sintering process, the repetition rate was always the same for ablation and sintering processes. Typically, lasers with shorter wavelengths (higher photon energy) and shorter pulse widths (high peak power) are suitable for the efficient ablation of a wide range of solid targets. Also, depending on the type of materials and substrates involved in the process, various sintering energies should be employed.
Moreover, the shallow absorption depth of UV laser used here allowed us to slightly sinter the generated nanoparticles without damaging the underlying substrates.
shows an example of a 2 × 2 cm 2 printed Ag NFC tag antenna (four loops with a total length of ~32 cm and ~1.38 Ω cm resistance) printed on a polyimide substrate. The mounted M24LR04E IC was programmed to launch our group's website. As shown in Figure 1c, the group website pops up on display by holding the cellphone in the proximity (~ 1 cm) of the printed NFC tag. Figure 1d demonstrates the printed FHEs circuit on the polyimide substrate via ANM process.
To confirm the nanoparticle formation and analyze their size distribution, we performed scanning electron microscopy (SEM) and scanning transmission electron microscopy (STEM) measurements on the as-generated nanoparticles. While the LS beam was blocked, we collected laser-ablated Ag and ITO nanoparticles (ablation energy (AE) = 2 J cm -2 , gas flow rate (GFR) = 2.8 SLPM, sintering/crystallization energy (S/CE) = blocked, repetition rate (RR) = 10 Hz). The SEM images of the unsintered and laser-sintered nanoparticles are shown in Figure 2a and d.
and e shows the bright-field (BF) STEM images of the captured Ag and ITO nanoparticle, which were directly deposited onto the TEM grids for measurements. According to the STEM images, the size of the generated nanoparticles was in the range of 3-10 nm diameters.
The unsintered nanoparticles form a fluffy structure that is loosely attached to the substrate. When these nanoparticles were laser sintered during the process, they fused together to form a solid line.
and f shows the sintered Ag and ITO nanoparticles. The sintering energy was 0.08 J cm - 2 and 0.13 J cm -2 for Ag and ITO, respectively. Figure 3 shows a cross-sectional SEM image of the printed Ag line confirming the sintering of the ablated nanoparticles on SiO2 substrate. SEM images in Figure 4 show the morphological and electrical resistance evolution of the sintered Ag nanoparticles at different sintering laser energies. By increasing the crystallization energy while other parameters (AE = 2 J cm -2 , RR = 40 Hz, GFR = 2.8 SLPM, laser beam overlap= 93%, deposition time in each point=20 ms, number if printed path= 6, and nozzle diameter size= 300 µm) are kept constant, the nanoparticles start to fuse together. The highest porosity was seen at the lower sintering energy (~0.03 J cm -2 ), and the lowest porosity was achieved at the higher laser sintering energy (~0.11 J cm -2 ), as shown in Figures 4a and i, respectively. It should be noted that the laser energies below the abovementioned lowest threshold did not impact the particles, and energies above the higher threshold resulted in reablation of the deposited particles. Moreover, the electrical resistivity of the printed lines decreased as the sintering laser energy increased (shown in the upper right corner of the figures). As it has shown in Figure 4a-i, 40 Ω cm and 7 Ω cm resistance were measured for lowest and highest sintering laser energies, respectively.
For the given process parameters (AE = 2 J cm -2 , S/CE= 0.09 (Ag) and 0.39 (ITO) J cm -2 , GFR = 2.8 SLPM, laser beam overlap= 93%, deposition time in each point= 20 ms and nozzle diameter size= 300 µm), the thickness variation of the printed Ag and ITO lines on the polyimide substrates was analyzed at four different laser repetition rates (i.e., 10 Hz, 20 Hz, 30 Hz, and 40 Hz), as shown in Figure 5. In the case of Ag (Figure 5a), at 10 Hz repetition rate, the thickness increased from ~1 µm to ~60 µm when the number of printed paths increased from 1 to 20. Also, while the repetition rate increased (e.g. 40 Hz), the thickness of printed Ag layers increased in each path compared to a lower repetition rate (e.g. 10 Hz). This could be attributed to the fact that the ablation rate and hence the number of in-situ generated and sintered nanoparticles increase at higher repetition rates. The influence of gas flow rate (GFR), repetition rate (RR), and laser sintering/crystallization energy (S/CE) on thickness and resistance of printed TiO2 and ITO is comprehensively discussed in our previous work, [2] where we showed that the thickness of the ANM-printed TiO2 is directly proportional to the number of printed paths and repetition rate.
Here, we have also performed a systematic study to investigate the influence of repetition rates and the number of printed paths on the printing speed and electrical conductivity of printed Ag lines on polyimide substrates. Similar information for ITO lines are reported in our previous study. [2] Hz repetition rate, the number of printed paths should be around 30 times. The Ag resistance could be tuned from 0.02 Ω cm to ~200 KΩ cm by adjusting the abovementioned process parameters.
As a result, the ANM-printed lines and materials could be potentially used to realize fully printed FHEs. The print resolution of the ANM was tested down to 100 µm using common commercial printing nozzles. It should be noted that printing thinner lines (e.g., < 100 µm) may be possible using smaller diameter nozzles since our ANM technique uses nanoparticles with an average size of 3-10 nm. Also, the sport size of the sintering laser can be reduced down to a few microns with suitable optics. Therefore, a combination of a nozzle with a smaller hole and the sintering laser with a smaller sport size can facilitate smaller line widths.
To test the reliability of the printed lines, we printed a set of Ag lines with the experimental parameters shown in Table 1. The reliability of these ANM-printed lines was investigated under the static and cyclic bending studies. Consequently, mechanical strain such as bending, cycling, and stretching may change the electrical performance of the printed lines. Thus, we performed insitu measurements to measure the changes in the resistance of the printed lines under different bending radii and compressive stresses. Figure 7a shows the schematic of the bending/cycling test setup. For instance, we tested different bending radii ranging from 77 mm to 16 mm equivalent to 0.08% to 0.5% strain. Figure 7b-d shows the variations in the electrical resistance of printed Ag lines as a function of different strains at three different numbers of cycles. The resistance change (ΔR/R0) and the strain on the device were calculated using the equations:
where R0, Rs, tsubstrate, Rbend are initial resistance, under-stress resistance, substrate thickness, and bending radius, respectively. [44,45] In the case of Ag, the resistance increased as it was bent for just one cycle, but the values of the resistance did not change drastically from the initial resistance values (e.g., R0 = 10.4 Ω cm to Rs = 11.5 Ω cm at the maximum strain of 0.5%). As it is shown in Figure 7b-d, the resistance increased more at higher bending strains (lower bending radius). As shown in Figure 7b, the largest change in resistance was for conditions where the initial resistance was 22.4 Ω cm, and the final resistance was 25.5 Ω cm at 0.5% strain. Also, the smallest change in the resistance seen in Figure 7b was for resistance changing from an initial resistance value of 10.8 Ω cm to a final resistance of 10.9 Ω cm at 0.08% strain.
The results from the 50 cycling tests are demonstrated in Figure 7c. According to the results, the resistance increased only slightly. The mean average of the overall change in resistance was 1.97 Ω cm. Thus, these Ag samples could tolerate a large strain with a slight increase in their electrical resistance. The results for 1000 cycles for different strains are presented in Figure 7d. Also, we observed that during the first 50,000 cycles, the resistance increased for a 10 Ω cm Ag line up to ~20 Ω cm (i.e., 100%), then the resistance started to increase continuously to more than ~38 Ω cm at 1,000,000 cycles. Relaxation time for this sample was around ~1 hour, and the final measured resistance was ~27 Ω cm. Low resistance deviation in bending radius beyond 40 mm (0.19% strain) and after numerous folding/unfolding cycles demonstrates the reliability of ANM printing process for manufacturing flexible electronics and sensors. Similar to the IJP and ASJ printing, the increase in electrical resistance could be attributed to the formation of cracks in deposited materials. When the substrate is stretched, the high flexibility of the substrate and the less stretchable Ag causes crack formation, which ultimately causes the change of the resistance.
Higher bending and stress accelerated the dislocation and crack formation in the sintered materials.
ITO is a transparent conducting layer that has been used in numerous applications. Although ITO has a low sheet resistance and excellent transparency, printed ITO lines on flexible substrates are usually micro/nanocrystalline structures with grain boundaries and pores, which might form micro/nano cracks under tensile strains. These changes in properties can be exploited for strain sensors, touchscreens, and motion detectors applications. [41] We used a servohydraulic fatigue testing machine for displacement-controlled stretching tests, as shown in Figure 8a. ITO lines were printed onto the PET substrates (55 mm×5 mm×0.175 mm) using process parameters reported previously. [2] Stretching tests for the printed ITO lines on PET were performed for different stretch levels. The amount of stretch was controlled with a precision of 0.18% to 0.9% displacement. All of the stretched samples were subjected to 100 cycles. For instance, Figure 8b shows the change in the resistance when subjected to 0.18% to 0.9% displacement for a printed ITO sample with initial resistance of ~1.18 KΩ cm. For the case of 0.18% displacement, the resistance change was less than 5%, while for 0.36% displacement, it is about 16%. For the case of 0.54% to 0.72% displacement, the resistance change was around ~50%.
Applying 0.9% displacement resulted in a gradual resistance change of up to ~400%. Above 0.9% displacement, permanent substrate plastic deformation was observed, resulting in resistance overload. The cycling frequency was kept at 1 Hz for all the samples. The bending experiment was also used to demonstrate how the ITO resistance changes at different bending radii, as shown in Figure 8c. The initial resistance of printed samples was between 420 Ω cm to 1600 Ω cm. Figure 8d shows the normalized resistance variance of strain sensor during 20 seconds at 0.19% strain and 9 cycles. The inset image in Figure 8d shows the printed ITO strain sensor on PET substrate. Furthermore, Figure 8e, f shows the SEM images of an ITO line before and after stretching. As it is shown in Figure 8f, under 0.9% displacement applied strain, small cracks appeared in the line that dramatically increased its resistance. Increased strain levels increase the crack density and widen the spacing between them, and hence the resistance increases.
Adversely, when the cracks close, the resistance decreases.
Ag meander-shaped ribbon (300 μm width and 1.5 cm length) was ANM-printed as a temperature sensor, as shown in Figure 9a. The sample was placed onto a hotplate to measure the sensor response, and the temperature measurement range was performed from 30 °C to 80 °C with 10 °C temperature increment and 5 minutes wait time at each temperature.
We designed and constructed several device prototypes to demonstrate the flexibility and applicability of our ANM process in printing FHEs on a polyimide substrate (Figure 10). As it is
In summary, we demonstrated the dry printing and additive nanomanufacturing of FHEs and sensors on flexible polyimide and PET substrates. In-situ laser-generated nanoparticles inside a microchamber were used as the building blocks for printing conductive and functional lines and patterns. A real-time laser sintering process was used to sinter the nanoparticles onto the substrate as they exit the nozzle at room temperature and atmospheric pressure. Thin Ag and ITO patterns with resistances down to 0.02 Ω cm and 100 Ω cm, respectively, were achieved by controlling the number of printed paths and repetition rates while using optimum sintering laser energy.
Mechanical reliability tests including cycling (0.08% to 0.5% strain) and stretching (0.18% to 0.9% displacement) were performed on the additively manufactured Ag and ITO lines on 127 µm thick polyimide and 200 µm thick PET substrates. The results showed that all the Ag lines are still conductive at the highest testes strain, even after 1,000,000 cycles. Also, ~5% to ~400% change in resistance of ITO-printed lines under 0.18% to 0.9% displacement was observed. Conductive NFC antenna tag, strain sensor, temperate sensor, and FHEs were printed on polyimide and PET substrates, and their reliable performance was confirmed. The dry nature, printing flexibility, purity, mechanical reliability, and device performance characteristics demonstrate the potential use of the developed system as an alternative and transformative approach for future printed flexible electronics and sensors.
The ~12.5 Ω cm temperature sensor was printed using ANM printer and the following parameters.
The ablation energy, sintering/crystallization energy, gas flow rate, and repetition rate were set at 2 J cm -2 , 0.07 J cm -2 , 2.8 SLPM, 20 Hz, respectively. 15 times printed path was employed with 93% laser beam overlap, 20 ms delay on each spot, and 300 µm nozzle diameter size.
FHE circuit printed while ablation energy, sintering/crystallization energy, gas flow rate, and repetition rate were 2 J cm -2 , 0.09 J cm -2 , 2.8 SLPM, 20 Hz, respectively. Laser beam overlap, delay on each spot, nozzle diameter size, and the number of the printed paths set to 93%, 20 ms, 300 µm, and 20 times, respectively.
SEM images were obtained using Zeiss Crossbeam 550 microscope. The voltage and current were 5 KV and 2 nA, respectively.
STEM imaging was performed using a Cs corrected JEOL NeoARM operating at 80kV. Nanoparticles were directly deposited onto the TEM grid by holding the TEM grids under the printer nozzle for a few seconds. Samples were then studied under TEM without further sample preparations. This collection method ensured the purity and cleanliness of the collected nanoparticles.
KEYENCE VHX-6000 series microscope was used to measure the thickness of ANM-printed samples.
Tabletop MTS Landmark servohydraulic test system with a load capacity of 25kN was used to conduct displacement-controlled stretching tests.