Copper Nanowires for Transparent Electrodes: Properties, Challenges and Applications

: Transparent electrodes are essential elements of devices bearing a screen or display, as well as solar cells, LEDs etc. To overcome the drawbacks presented by indium tin oxide, nanomaterials have been proposed for a long time as alternatives. Metal nanowires are particularly interesting for their high intrinsic electrical conductivity. Copper nanowires have attracted wide interest due to the low cost and high abundancy of the starting material. However, they are easily oxidized thus suitable strategies must be devised to prevent it. This review discusses the fundamental properties and challenges of copper nanowires, focusing on the efforts made to make them longer and thinner then the strategies to prevent oxidation and to join them in the network are presented. After that, mechanical properties are summarized and applications are presented, before conclusions and perspectives are ﬁnally given. Cu-graphene structures were obtained by Ahn et al. using a chemical vapor deposition process [51]. Here the NW network was tested for 30 days in ambient conditions with little Rs change, while for pure Cu NWs the Rs increased sharply after less than ﬁve days. Composites with graphene sheets which were exfoliated in the liquid phase were recently reported by Navik et al. [52]. Here the network stability was investigated for 46 days, showing a 27% increase in ambient condition and a 92% increase at 60 ◦ C, with a signiﬁcant improvement compared to bare Cu NW networks [52].


Introduction
Transparent conductors are key components of a wide range of devices that are common in everyday use. Such devices include displays, touch screens, light-emitting diodes (LEDs), solar cells, memory devices, to name a few [1]. While commercial devices employ indium tin oxide (ITO) as transparent conductor, this material is far from ideal due to the toxicity of indium and its availability issues that eventually will drive the cost up. Moreover, the expanding technological interest in flexible electronics rules out ITO due to its intrinsic brittleness, which would bring the material to failure over bending [2,3]. The last two decades have thus seen significant research and development effort from both the scientific community and the industry to find suitable alternatives.
In order to serve as transparent conductor, a material should have the lowest possible sheet resistance, Rs, associated with the highest possible transmittance, T, usually measured and reported at 550 nm. To match ITO performance a transparent conductor should have Rs < 100/Ω/sq and T > 90% [4]. Carbon nanotubes have been initially proposed as alternatives, as they ensure high conductivity and transparency [5][6][7][8]. However, their performance never reached that of ITO. After the discovery of graphene, this has also been proposed as ITO alternative, which only matched, and surpassed, ITO performance when grown as a single large area graphene sheet [9], while sheets produced by liquid phase exfoliation showed lesser performance than carbon nanotubes [10][11][12].
Silver nanowires (NWs) were proposed as an exciting ITO alternative given their high conductivity due to the metallic character. Indeed, Ag NWs were able to match ITO performance both in terms of Rs and T [13][14][15][16]. However, Ag is also expensive and not very abundant [17], thus NWs made of other metals have been pursued. For example copper (Cu) has been proposed as an alternative being almost as conductive as Ag and at the same time much cheaper and more abundant [17]. Indeed, the pioneering work by Rathmell et al. paved the way for the utilization of Cu NWs as ITO alternatives [18][19][20]. Cu NWs however suffer from the issues common to Cu in general, i.e., the presence of a native oxide as well  [18]. Copyright  Other synthetic methods to improve NW morphology were also reported. Zhang et al. reported a synthesis in a liquid crystalline medium using hexadecylamine and cetyl trimethylammonium bromide (CTAB) and Cu acetylacetonate as precursor [28]. Such synthesis allowed the formation of NWs with 78 nm average diameter and length over 100 μm, with Rs = 90 Ω/sq at 90% transmittance. Compared to Ref. [19], for the same T the halving of Rs is most likely due to the much longer length, despite the wider diameter. Guo et al. obtained very thin NW (16 nm) with 40 μm length carrying out a reaction in oleylamine using a Ni-based catalyst [29]. A high transmittance of 93% was reported, associated with a low Rs of 51 Ω/sq. Compared to previous reports, the improvement here is mainly due to the very small NW diameter.
Another synthetic method to produce very long NWs was reported by Mayousse et al., who used octadecylamine and CuCl2 reacting at 165 °C for 140 h, to obtain films with Rs = 55Ω/sq at T = 94% [30]. Unlike many other reports, the authors here also study haze, intended as the amount of light that is scattered rather than transmitted. While haze is detrimental for applications like displays or LED as they decrease transmittance, this can be seen as beneficial for solar cell applications as it allows the scattered light to be reabsorbed and thus utilized for energy production. Haze evaluation is thus related to the specific application. It was found that haze is somehow related to the transmittance, and is minimum for maximum transmittance [30]. Interestingly, the authors report diameters of a few hundred nm associated with lengths lower than 50 nm. The good performance is likely justified as the length distribution also includes structures as long as several hundred μm, which may be responsible for the low Rs at such high T. A similar finding was reported by Cui et al., who obtained NWs with 17 μm length and 17 nm diameter for a performance of 35 Ω/sq at T = 90% and haze down to 1% at T = 94% ( Figure 2) [31]. The authors used a silane as reducing agent for CuCl2 in oleylamine and oleic acid. It has been shown that haze is related to NW diameter [32] so this is a further reason to try reducing NW diameter, where haze is to be minimized, beside improving T. Another method for improving T was found to be purification [33,34]. Here, NWs were purified by three cycles of centrifugation after adding n-hexane to the aqueous solution. NWs settled on the bottom of the hexane phase separating from other particles. T up to 97% were reported, associated to Rs = 103 Ω/sq. [33]. Alternatively, a method to obtain high purity NWs was proposed by Kim et al. by optimizing the relative concentrations of Cu precursor and capping agents [35]. Another synthetic method to produce very long NWs was reported by Mayousse et al., who used octadecylamine and CuCl 2 reacting at 165 • C for 140 h, to obtain films with Rs = 55 Ω/sq at T = 94% [30]. Unlike many other reports, the authors here also study haze, intended as the amount of light that is scattered rather than transmitted. While haze is detrimental for applications like displays or LED as they decrease transmittance, this can be seen as beneficial for solar cell applications as it allows the scattered light to be reabsorbed and thus utilized for energy production. Haze evaluation is thus related to the specific application. It was found that haze is somehow related to the transmittance, and is minimum for maximum transmittance [30]. Interestingly, the authors report diameters of a few hundred nm associated with lengths lower than 50 nm. The good performance is likely justified as the length distribution also includes structures as long as several hundred µm, which may be responsible for the low Rs at such high T. A similar finding was reported by Cui et al., who obtained NWs with 17 µm length and 17 nm diameter for a performance of 35 Ω/sq at T = 90% and haze down to 1% at T = 94% ( Figure 2) [31]. The authors used a silane as reducing agent for CuCl 2 in oleylamine and oleic acid. It has been shown that haze is related to NW diameter [32] so this is a further reason to try reducing NW diameter, where haze is to be minimized, beside improving T. Another method for improving T was found to be purification [33,34]. Here, NWs were purified by three cycles of centrifugation after adding n-hexane to the aqueous solution. NWs settled on the bottom of the hexane phase separating from other particles. T up to 97% were reported, associated to Rs = 103 Ω/sq. [33]. Alternatively, a method to obtain high purity NWs was proposed by Kim et al. by optimizing the relative concentrations of Cu precursor and capping agents [35].
A method for the production of NWs as long as 250 µm was reported by Zhang et al. by using polyhydric alcohols in presence of glucose and capping agents like oleylamine and oleic acid [36]. The number of hydroxyl groups in such alcohols was also found to influence the NW diameter, with more hydroxyls decreasing diameter. Here films with Rs = 92 Ω/sq at T = 91.5% were reported [36].
A synthetic process able to tune the diameters of Cu NWs was proposed by Wang et al., who could change the diameter from 20 to 90 nm by changing the concentration of halide ions [37]. They reported Rs~60 Ω/sq at T = 90%. A method for the production of NWs as long as 250 μm was reported by Zhang et al. by using polyhydric alcohols in presence of glucose and capping agents like oleylamine and oleic acid [36]. The number of hydroxyl groups in such alcohols was also found to influence the NW diameter, with more hydroxyls decreasing diameter. Here films with Rs = 92 Ω/sq at T = 91.5% were reported [36].
A synthetic process able to tune the diameters of Cu NWs was proposed by Wang et al., who could change the diameter from 20 to 90 nm by changing the concentration of halide ions [37]. They reported Rs ~60 Ω/sq at T = 90%.
The synthesis of curved Cu NWs was also reported. Yin et al. deposited these curved structures, like those shown in Figure 3, onto glass and polyimide (PI) claiming that the curved structure improves the network properties [38,39]. However, no significant improvement in Rs and T was reported compared to other reports. Zhang et al. also reported the synthesis of curved Cu NWs and their deposition on PET and PDMS, and showed T > 95% associated with Rs = 110 Ω/sq and haze < 2% [40].
In summary, this section provided an overview of the key electrical and optical properties related to transparent conductors, how they are affected by NW morphology and a brief account of the synthetic methods proposed to tune NW morphology in order to improve network performance.  The synthesis of curved Cu NWs was also reported. Yin et al. deposited these curved structures, like those shown in Figure 3, onto glass and polyimide (PI) claiming that the curved structure improves the network properties [38,39]. However, no significant improvement in Rs and T was reported compared to other reports. Zhang et al. also reported the synthesis of curved Cu NWs and their deposition on PET and PDMS, and showed T > 95% associated with Rs = 110 Ω/sq and haze < 2% [40]. A method for the production of NWs as long as 250 μm was reported by Zhang et al. by using polyhydric alcohols in presence of glucose and capping agents like oleylamine and oleic acid [36]. The number of hydroxyl groups in such alcohols was also found to influence the NW diameter, with more hydroxyls decreasing diameter. Here films with Rs = 92 Ω/sq at T = 91.5% were reported [36].
A synthetic process able to tune the diameters of Cu NWs was proposed by Wang et al., who could change the diameter from 20 to 90 nm by changing the concentration of halide ions [37]. They reported Rs ~60 Ω/sq at T = 90%.
The synthesis of curved Cu NWs was also reported. Yin et al. deposited these curved structures, like those shown in Figure 3, onto glass and polyimide (PI) claiming that the curved structure improves the network properties [38,39]. However, no significant improvement in Rs and T was reported compared to other reports. Zhang et al. also reported the synthesis of curved Cu NWs and their deposition on PET and PDMS, and showed T > 95% associated with Rs = 110 Ω/sq and haze < 2% [40].
In summary, this section provided an overview of the key electrical and optical properties related to transparent conductors, how they are affected by NW morphology and a brief account of the synthetic methods proposed to tune NW morphology in order to improve network performance.  In summary, this section provided an overview of the key electrical and optical properties related to transparent conductors, how they are affected by NW morphology and a brief account of the synthetic methods proposed to tune NW morphology in order to improve network performance.

Oxidation Prevention
As already mentioned above, the major challenge facing Cu-based nanomaterials, including NWs, is oxidation, both as native oxide and as oxide forming over time in ambient conditions, both negatively affecting conductivity. While annealing in forming gas (5% H 2 and 95% N 2 ) is a typical way to remove the native oxide, other strategies are to be developed to protect from oxidation after synthesis [20]. For example, a UV treatment, followed by acid bath, was proposed to remove organic residues from the NW surface with subsequent performance improvement [41], though no information on the performance over time was given.
A popular method for oxidation protection is making core-shell structures with other materials, for example with Nickel [20,42,43]. This was first reported by Rathmell et al. who showed that for pristine Cu NW, when kept at 85 • C, Rs increased by two orders of magnitude within five days, while a layer of Ni kept Rs flat for at least up to 30 days (Figure 4a) [20]. The authors also reported that the presence of Ni significantly increased Rs. However, there is a clear advantage with the stability of the NWs (Figure 4b) [20]. Similar results were obtained by Stewart et al. who obtained Cu NW films deposited onto glass substrates dipped in acetic acid first to remove native oxide and then in an electroless plating Ni solution to form Cu-Ni core-shell NWs [42]. The same group later reported the synthesis of Cu-Ag core shell NWs using the same electroless plating Ag deposition [43]. The same acetic acid treatment could be repeated several times after reoxidation [34]. Unlike the case of Ni, Ag-coated Cu NWs did not show a decrease in optical and electrical performance. Xue et al. also reported Cu-Ni core-shell NWs, with findings similar to those reported above [44]. Niu et al. reported the epitaxial growth of 1-2 nm layers of Au onto Cu NWs, reporting a dramatic improvement of stability under high temperature (80 • C) and humidity (80%) [45].

Oxidation Prevention
As already mentioned above, the major challenge facing Cu-based nanomaterials, including NWs, is oxidation, both as native oxide and as oxide forming over time in ambient conditions, both negatively affecting conductivity. While annealing in forming gas (5% H2 and 95% N2) is a typical way to remove the native oxide, other strategies are to be developed to protect from oxidation after synthesis [20]. For example, a UV treatment, followed by acid bath, was proposed to remove organic residues from the NW surface with subsequent performance improvement [41], though no information on the performance over time was given.
A popular method for oxidation protection is making core-shell structures with other materials, for example with Nickel [20,42,43]. This was first reported by Rathmell et al. who showed that for pristine Cu NW, when kept at 85 °C, Rs increased by two orders of magnitude within five days, while a layer of Ni kept Rs flat for at least up to 30 days (Figure 4a) [20]. The authors also reported that the presence of Ni significantly increased Rs. However, there is a clear advantage with the stability of the NWs (Figure 4b) [20]. Similar results were obtained by Stewart et al. who obtained Cu NW films deposited onto glass substrates dipped in acetic acid first to remove native oxide and then in an electroless plating Ni solution to form Cu-Ni core-shell NWs [42]. The same group later reported the synthesis of Cu-Ag core shell NWs using the same electroless plating Ag deposition [43]. The same acetic acid treatment could be repeated several times after reoxidation [34]. Unlike the case of Ni, Ag-coated Cu NWs did not show a decrease in optical and electrical performance. Xue et al. also reported Cu-Ni core-shell NWs, with findings similar to those reported above [44]. Niu et al. reported the epitaxial growth of 1-2 nm layers of Au onto Cu NWs, reporting a dramatic improvement of stability under high temperature (80 °C) and humidity (80%) [45]. Other than metals, oxides can also be viable protective coatings onto Cu NWs. For example, Chen et al. deposited metals like Zn, Sn and In by electrodeposition, followed by their oxidation, by hydrogen peroxide, into the corresponding transparent oxides [46]. A ZnO layer deposited by atomic layer deposition was used by Tigan et al. to improve the performance of transparent thin film heaters (see applications sections for more details) [47]. An Al-doped ZnO coating on the Cu NW was also proposed by Won et al. to convey ambient stability to the Cu NW network [48]. They also reported a lactic acid treatment to remove native oxide and other organic residues with no annealing.
The use of graphene and its related materials was also reported as a viable strategy to improve ambient stability of Cu NWs. Kholmanov et al. reported hybrid thin films of Cu NWs and reduced graphene oxide (RGO) [49]. They showed that while pure RGO thin Other than metals, oxides can also be viable protective coatings onto Cu NWs. For example, Chen et al. deposited metals like Zn, Sn and In by electrodeposition, followed by their oxidation, by hydrogen peroxide, into the corresponding transparent oxides [46]. A ZnO layer deposited by atomic layer deposition was used by Tigan et al. to improve the performance of transparent thin film heaters (see applications sections for more details) [47]. An Al-doped ZnO coating on the Cu NW was also proposed by Won et al. to convey ambient stability to the Cu NW network [48]. They also reported a lactic acid treatment to remove native oxide and other organic residues with no annealing.
The use of graphene and its related materials was also reported as a viable strategy to improve ambient stability of Cu NWs. Kholmanov et al. reported hybrid thin films of Cu NWs and reduced graphene oxide (RGO) [49]. They showed that while pure RGO thin films have Rs about two orders of magnitude higher than Cu NW films (Figure 5a), when combined together the Rs is decreased by about 2× from that of Cu NW films (Figure 5b). This is likely because the presence of RGO sheets lowers the junction resistance of Cu NWs and provides additional conduction paths. More importantly, Rs of such films is flat for at least 72 h at 60 • C (Figure 5c) [49]. Dou et al. reported Cu NWs wrapped with RGO with a slightly improved electrical properties and stability over 200 days [50]. Among the RGO-Cu NW films, those with lower T are the most stable, while those with higher T are more susceptible to oxidation with almost 2× Rs increase over 200 days. Core-shell Cu-graphene structures were obtained by Ahn et al. using a chemical vapor deposition process [51]. Here the NW network was tested for 30 days in ambient conditions with little Rs change, while for pure Cu NWs the Rs increased sharply after less than five days. Composites with graphene sheets which were exfoliated in the liquid phase were recently reported by Navik et al. [52]. Here the network stability was investigated for 46 days, showing a 27% increase in ambient condition and a 92% increase at 60 • C, with a significant improvement compared to bare Cu NW networks [52]. increase in ambient condition and a 92% increase at 60 °C, with a significant improveme compared to bare Cu NW networks [52].
A different approach was recently proposed by Genlik et al., who applied a benzot azole passivation layer [53]. Here benzotriazole, an organic corrosion inhibitor, made t Cu NWs stable under ambient conditions for 370 days without affecting the optoelectric properties. Such higher stability was demonstrated also in more aggressive conditions li high temperatures p to 150 °C and humidity up to 90% [53].
In summary, this section provided an overview for the methods proposed in the l erature for the protection of Cu NWs from oxidation under ambient or adverse condition Overall, all methods have in common a protective layer around the NW, which can another metal, like Ni, metal oxides, another material like graphene, or molecules that a like corrosion inhibitors.

Joining of Nanowires
As already mentioned, junction resistance is a key factor that affects the conductivity the whole network. Over the years, methods have then been developed to join, or weld, t NWs at the junctions, in order to eliminate, or cut down, the junction resistance. Most of the methods were firstly adopted for Ag NWs and involve the use of heat and/or light [54,5 Such methods for Cu NWs also have the secondary effect of removing the oxide lay Ding et al. proposed a high intensity pulsed light to induce both deoxygenation and wel ing [56]. Figure 6a,b shows tilted SEM images before and after the pulsed light treatme and it can be observed that NWs are joint together after the treatment. Moreover, increa ing the number of pulses dramatically improves the optical and electrical performance the network (Figure 6c) [56], although no data were presented above 85% transmittanc Using the same method, the same group fabricated a polyurethane (PU) composite [5 Here, the pulsed light also softened the PU matrix allowing embedding the NWs. A different approach was recently proposed by Genlik et al., who applied a benzotriazole passivation layer [53]. Here benzotriazole, an organic corrosion inhibitor, made the Cu NWs stable under ambient conditions for 370 days without affecting the optoelectrical properties. Such higher stability was demonstrated also in more aggressive conditions like high temperatures p to 150 • C and humidity up to 90% [53].
In summary, this section provided an overview for the methods proposed in the literature for the protection of Cu NWs from oxidation under ambient or adverse conditions. Overall, all methods have in common a protective layer around the NW, which can be another metal, like Ni, metal oxides, another material like graphene, or molecules that act like corrosion inhibitors.

Joining of Nanowires
As already mentioned, junction resistance is a key factor that affects the conductivity of the whole network. Over the years, methods have then been developed to join, or weld, the NWs at the junctions, in order to eliminate, or cut down, the junction resistance. Most of these methods were firstly adopted for Ag NWs and involve the use of heat and/or light [54,55].
Such methods for Cu NWs also have the secondary effect of removing the oxide layer. Ding et al. proposed a high intensity pulsed light to induce both deoxygenation and welding [56]. Figure 6a,b shows tilted SEM images before and after the pulsed light treatment, and it can be observed that NWs are joint together after the treatment. Moreover, increasing the number of pulses dramatically improves the optical and electrical performance of the network (Figure 6c) [56], although no data were presented above 85% transmittance. Using the same method, the same group fabricated a polyurethane (PU) composite [57]. Here, the pulsed light also softened the PU matrix allowing embedding the NWs.
Wang et al. proposed a hydrogen plasma treatment to weld the NWs together, and at the same time improve their stability towards oxidation [58]. Here, the plasma treatment allowed cleaning the NW surface from organic residues from the synthesis process and oxide. At the same time, plasmonic effects allowed light intensity enhancement at the NW ends, causing thermal heating which induced melting [58]. Indeed, the Cu NWs had a plasmon resonance peak at 576 nm and a diameter of 60-100 nm, while the plasma had emission in the 570-610 nm range. Park et al. used flash light to both sinter Cu NWs, remove residues from the NW surface and induce adhesion to polyethyleneterephtalate (PET) by thermally melting the NW-PET interface (Figure 7) [59]. Similarly, plasmonic effects were exploited by Han et al. using a continuous 532 nm laser focused onto the NW network surface, for both preventing oxidation and welding the NWs [60]. Here, the Cu NWs had a diameter of 80-120 nm. Welding by electroplating was proposed by Zhao et al. who used the Cu NW film as electrode in a plating solution, then by applying voltage with a phosphor counter-electrode chemical welding was achieved [61]. Wang et al. proposed a hydrogen plasma treatment to weld the NWs together, and at the same time improve their stability towards oxidation [58]. Here, the plasma treatment allowed cleaning the NW surface from organic residues from the synthesis process and oxide. At the same time, plasmonic effects allowed light intensity enhancement at the NW ends, causing thermal heating which induced melting [58]. Indeed, the Cu NWs had a plasmon resonance peak at 576 nm and a diameter of 60-100 nm, while the plasma had emission in the 570-610 nm range. Park et al. used flash light to both sinter Cu NWs, remove residues from the NW surface and induce adhesion to polyethyleneterephtalate (PET) by thermally melting the NW-PET interface (Figure 7) [59]. Similarly, plasmonic effects were exploited by Han et al. using a continuous 532 nm laser focused onto the NW network surface, for both preventing oxidation and welding the NWs [60]. Here, the Cu NWs had a diameter of 80-120 nm. Welding by electroplating was proposed by Zhao et al. who used the Cu NW film as electrode in a plating solution, then by applying voltage with a phosphor counter-electrode chemical welding was achieved [61]. Wang et al. proposed a hydrogen plasma treatment to weld the NWs together, and at the same time improve their stability towards oxidation [58]. Here, the plasma treatment allowed cleaning the NW surface from organic residues from the synthesis process and oxide. At the same time, plasmonic effects allowed light intensity enhancement at the NW ends, causing thermal heating which induced melting [58]. Indeed, the Cu NWs had a plasmon resonance peak at 576 nm and a diameter of 60-100 nm, while the plasma had emission in the 570-610 nm range. Park et al. used flash light to both sinter Cu NWs, remove residues from the NW surface and induce adhesion to polyethyleneterephtalate (PET) by thermally melting the NW-PET interface (Figure 7) [59]. Similarly, plasmonic effects were exploited by Han et al. using a continuous 532 nm laser focused onto the NW network surface, for both preventing oxidation and welding the NWs [60]. Here, the Cu NWs had a diameter of 80-120 nm. Welding by electroplating was proposed by Zhao et al. who used the Cu NW film as electrode in a plating solution, then by applying voltage with a phosphor counter-electrode chemical welding was achieved [61].  A welding method not involving the use of any form of light was also proposed by Zhang et al., who synthesized core-shell NWs by electrodepositing Ni, Ag or Au-Ag on the Cu NW surface [62][63][64]. Here the metal shell, other than providing protection from oxidation, also provided welding by joining the NWs at the junctions, as shown in Figure 8.

Mechanical Properties
One of the main drivers for the use of n mechanical properties that, unlike ITO, guar Cu NW are no exception to this, and man bending tests performed over a large numb sults reported concerning mechanical testin cases the resistance of the NW network sho testing. In the case of stretching, it is intere In summary, methods for joining, or welding, Cu NWs mainly involve the use of light, which by thermal or plasmonic effects melt NWs at the junctions reducing the junction resistance. An electrodeposition method was also proposed, allowing welding as well as oxidation protection.

Mechanical Properties
One of the main drivers for the use of nanomaterials as transparent electrodes is their mechanical properties that, unlike ITO, guarantee flexibility upon bending and stretching. Cu NW are no exception to this, and many reports include mechanical testing, mainly bending tests performed over a large number of cycles. Table 1 summarizes the main results reported concerning mechanical testing of Cu NWs. It can be observed that in most cases the resistance of the NW network shows very little change under the length of the testing. In the case of stretching, it is interesting to note that the relative sheet resistance changes, R/R 0 , reported in Table 1, are obtained after the materials are subjected to a strain ranging from 50% to 250%.  [68] Adhesion to a substrate is often a challenge for NW, which is rarely addressed in the literature. A simple tape test can be performed to address this matter, i.e., a sticky tape is pressed onto the substrate, bearing the NW network, and then peeled off. Yin et al. carried out this test for 10 times on their PU composite noting an Rs change <10% [39]. In this case, the tape had little consequences on the material as the NWs were mainly embedded within the PU matrix. Park et al. performed peeling tests on Cu NWs on PET substrates [59]. Here, while the adhesion was significantly improved after welding, after 5 cycles of peeling the Rs already increased by 10 times. Zhao et al. also performed a tape test on Cu NWs on PET, this time noting little resistance change (hardly quantifiable from the graph shown) after 10 peeling cycles on the network exposed to welding by electroplating [61].
In summary, this section provided a brief account of the mechanical properties of Cu NW-based network under a variety of testing like bending, stretching, twisting and adhesion.

Applications of Copper Nanowires
Once demonstrated the viability of Cu NW based network as transparent conductors, a number of applications have also been demonstrated by several authors.
The first application was shown by Guo et al. who employed Cu NW transparent electrodes in a blue LED as ohmic contacts on nand p-type GaN layers (Figure 9a) [29]. The ohmic contact is proved by the I-V curves in Figure 9b,c. The use in a GaN-based LED was also reported by Yoon et al. [69]. A LED structure was also demonstrated by Zhong et al. who fabricated a Cu NW based phosphorescent organic green LED with stable operation even under flexing [70]. In the same work, the authors proposed the application as transparent heater, working simply by Joule heating by applying a voltage to the ends of a Cu NW pattern, as shown in Figure 10 [70]. A transparent heater was also reported by Tigan et al. who used core-shell NWs with ZnO and Al 2 O 3 [47]. Here, the authors employed this device for a defrosting application, reaching heating rate of 14 • C/s and maximum temperatures of 273 and 204 • C for Al 2 O 3 and ZnO, respectively [47]. Flexible transparent heaters were similarly demonstrated with high heating uniformity and stable temperature output for up to 5 h by Zhang et al. [63,64]. Flexible transparent heaters were also demonstrated by Yu et al. using a transparent flexible conductive tape, obtained by transferring a Cu NW network from glass, where it was firstly deposited, onto a transparent tape [71]. The tape obtained had Rs~20 Ω/sq at T = 80% and temperature up to 143 • C.  Another popular application that has been proposed for Cu NW networks is in solar cells. The first demonstration was reported by Sachse et al. who reported a small molecule organic solar cell with 3% efficiency [72]. Here, the authors used a poly(3,4-ethylenedioxythiophene) polystyrene sulfonate (PEDOT:PSS) layer for planarization applied on the NW network by spin-coating which, despite its hygroscopic and acidic properties, did not affect the stability of the Cu NWs. Stewart et al. reported an organic solar cell based on core-shell Ni-Cu NWs with 4.9% efficiency, again using a PEDOT-PSS layer on top of the NW network [42]. Similarly, Ohiienko et al. fabricated an organic solar cell with a very high efficiency of 7.31% [68]. Ahn et al. made a polymer solar cell with 4.04% efficiency using their Cu-graphene core-shell NWs and PEDOT:PSS as planarization layer [51]. This compared to a 1.90% efficiency for pure Cu NWs. Wang et al. reported an organic solar cell based on Cu NW network with 2.67% efficiency, where they used an in-situ polymerization method to make an acrylic polymer layer for planarization [58]. A simple scheme of the NW network-polymer assembly is shown in Figure 11. Table 2 summarizes the main information on the reported solar cell devices. Figure 11. Scheme of a NW network-polymer assembly.  Another popular application that has been proposed for Cu NW networks is in solar cells. The first demonstration was reported by Sachse et al. who reported a small molecule organic solar cell with 3% efficiency [72]. Here, the authors used a poly(3,4-ethylenedioxythiophene) polystyrene sulfonate (PEDOT:PSS) layer for planarization applied on the NW network by spin-coating which, despite its hygroscopic and acidic properties, did not affect the stability of the Cu NWs. Stewart et al. reported an organic solar cell based on core-shell Ni-Cu NWs with 4.9% efficiency, again using a PEDOT-PSS layer on top of the NW network [42]. Similarly, Ohiienko et al. fabricated an organic solar cell with a very high efficiency of 7.31% [68]. Ahn et al. made a polymer solar cell with 4.04% efficiency using their Cu-graphene core-shell NWs and PEDOT:PSS as planarization layer [51]. This compared to a 1.90% efficiency for pure Cu NWs. Wang et al. reported an organic solar cell based on Cu NW network with 2.67% efficiency, where they used an in-situ polymerization method to make an acrylic polymer layer for planarization [58]. A simple scheme of the NW network-polymer assembly is shown in Figure 11. Table 2 summarizes the main information on the reported solar cell devices. Figure 11. Scheme of a NW network-polymer assembly. Another popular application that has been proposed for Cu NW networks is in solar cells. The first demonstration was reported by Sachse et al. who reported a small molecule organic solar cell with 3% efficiency [72]. Here, the authors used a poly(3,4ethylenedioxythiophene) polystyrene sulfonate (PEDOT:PSS) layer for planarization applied on the NW network by spin-coating which, despite its hygroscopic and acidic properties, did not affect the stability of the Cu NWs. Stewart et al. reported an organic solar cell based on core-shell Ni-Cu NWs with 4.9% efficiency, again using a PEDOT-PSS layer on top of the NW network [42]. Similarly, Ohiienko et al. fabricated an organic solar cell with a very high efficiency of 7.31% [68]. Ahn et al. made a polymer solar cell with 4.04% efficiency using their Cu-graphene core-shell NWs and PEDOT:PSS as planarization layer [51]. This compared to a 1.90% efficiency for pure Cu NWs. Wang et al. reported an organic solar cell based on Cu NW network with 2.67% efficiency, where they used an in-situ polymerization method to make an acrylic polymer layer for planarization [58]. A simple scheme of the NW network-polymer assembly is shown in Figure 11. Table 2 summarizes the main information on the reported solar cell devices.
compared to a 1.90% efficiency for pure Cu NWs. Wang et al. reported an or cell based on Cu NW network with 2.67% efficiency, where they used an in-sit ization method to make an acrylic polymer layer for planarization [58]. A sim of the NW network-polymer assembly is shown in Figure 11. Table 2 summariz information on the reported solar cell devices.   A very popular field of application is in the stretchable and wearable electronics, usually when the Cu NW are coupled with polymers like PU or PDMS. The first demonstration of a flexible touch sensor, based on capacitance change, was given by Mayousse et al. who fabricated a Cu NW composite with PEDOT:PSS [30]. Park et al. showed the operation of a touch-screen panel from Cu NWs on PET [59]. Here, while the device operated correctly, the counter-electrode was made of ITO, which may hinder the stability under bending. In the same work, the authors demonstrated a transparent resistive memory, again based on an ITO counter-electrode [59]. The device showed a good ON/OFF ratio of 10 3 . Kim et al. proposed a flexible touch sensor based on their NW-PU composite [35]. Such device is able to produce a current when "touched", i.e., a small pressure is applied (Figure 12a). This ability is maintained upon bending, although the response is significantly reduced (Figure 12b) [35]. Genlik et al. also demonstrated a transparent tactile sensor that mapped the pressure applied by finger touch or other objects ( Figure 13) [53]. Zhang et al. demonstrated a wearable sensor using a NW-PDMS composite which is able to detect the movement of a finger joint with fast response [36]. A very popular field of application is in the stretchable and wearable electronics, usually when the Cu NW are coupled with polymers like PU or PDMS. The first demonstration of a flexible touch sensor, based on capacitance change, was given by Mayousse et al. who fabricated a Cu NW composite with PEDOT:PSS [30]. Park et al. showed the operation of a touchscreen panel from Cu NWs on PET [59]. Here, while the device operated correctly, the counterelectrode was made of ITO, which may hinder the stability under bending. In the same work, the authors demonstrated a transparent resistive memory, again based on an ITO counterelectrode [59]. The device showed a good ON/OFF ratio of 10 3 . Kim et al. proposed a flexible touch sensor based on their NW-PU composite [35]. Such device is able to produce a current when "touched", i.e., a small pressure is applied (Figure 12a). This ability is maintained upon bending, although the response is significantly reduced (Figure 12b) [35]. Genlik et al. also demonstrated a transparent tactile sensor that mapped the pressure applied by finger touch or other objects ( Figure 13) [53]. Zhang et al. demonstrated a wearable sensor using a NW-PDMS composite which is able to detect the movement of a finger joint with fast response [36].
Kholmanov et al. demonstrated the realization of an electrochromic device based on Prussian blue from their NW-RGO composites [49]. Such a device was fabricated depositing the dye between the NW-RGO based electrode and an ITO counter-electrode. The reduction of the Fe 2+ /Fe 3+ mixture, typical of the Prussian blue dye, to Fe 2+ upon applying a 0.6 V voltage reversibly changed the device color from blue to colorless, as shown in Figure 14 [49]. An electrochromic device was also recently demonstrated by Zhang et al. using polyaniline which was directly polymerized onto the Cu NW network [62]. Here, different colors were obtained by tuning the voltage. The device also worked for energy storage with a capacitance of 12 mF/cm 2 .
In summary, this section provided an overview of the applications proposed for Cu NW transparent electrodes. These include LEDs, solar cells, wearable electronics, energy storage and electrochromic devices. Adapted with permission from Ref. [35]. Copyright © 2021 WILEY-VCH Verlag GmbH & Co.
KGaA, Weinheim. Kholmanov et al. demonstrated the realization of an electrochromic device based on Prussian blue from their NW-RGO composites [49]. Such a device was fabricated depositing the dye between the NW-RGO based electrode and an ITO counter-electrode. The reduction of the Fe 2+ /Fe 3+ mixture, typical of the Prussian blue dye, to Fe 2+ upon applying a 0.6 V voltage reversibly changed the device color from blue to colorless, as shown in Figure 14 [49]. An electrochromic device was also recently demonstrated by Zhang et al. using polyaniline which was directly polymerized onto the Cu NW network [62]. Here, different colors were obtained by tuning the voltage. The device also worked for energy storage with a capacitance of 12 mF/cm 2 .

Conclusions and Perspectives
This review covers the fundamental properties and the main challenges facing Cu NW based transparent electrodes. After briefly discussing the target properties that a NW network should have, in terms of Rs, T, NW length and diameter, Section 2 covers some of the reported synthetic strategies proposed to obtain long and/or thin NWs, which should be able to improve the network properties. In Section 3, a range of methods to

Conclusions and Perspectives
This review covers the fundamental properties and the main challenges facing Cu NW based transparent electrodes. After briefly discussing the target properties that a NW network should have, in terms of Rs, T, NW length and diameter, Section 2 covers some of the reported synthetic strategies proposed to obtain long and/or thin NWs, which should be able to improve the network properties. In Section 3, a range of methods to In summary, this section provided an overview of the applications proposed for Cu NW transparent electrodes. These include LEDs, solar cells, wearable electronics, energy storage and electrochromic devices.

Conclusions and Perspectives
This review covers the fundamental properties and the main challenges facing Cu NW based transparent electrodes. After briefly discussing the target properties that a NW network should have, in terms of Rs, T, NW length and diameter, Section 2 covers some of the reported synthetic strategies proposed to obtain long and/or thin NWs, which should be able to improve the network properties. In Section 3, a range of methods to prevent oxidation have been summarized. Such methods generally involve the production of coreshell systems, in which the shell is another metal, a transparent oxide or graphene. Section 4 summarizes the methods used to join, or weld, the NWs together, thus reducing the junction resistance and significantly decreasing Rs. Such methods generally involved the use of light, through lasers of flash lamps, also exploiting plasmonic effects. Section 5 covers the mechanical properties of Cu NW, summarizing the findings on the bending, stretching, folding and taping tests that have been reported in the literature. Finally, Section 6 summarizes the applications that have been proposed for Cu NW transparent conductors.
From what has been reported in this paper, it can be seen that some research and development areas for Cu NWs are more mature than others. For example, there is a vast number of papers already aiming to obtain the longest and/or the thinnest NWs. There is also enough evidence of the stability of this material under mechanical stress. However, adhesion to substrates is a challenge that has been so far poorly addressed and deserves new solutions, like for example surface treatments or encapsulation. The latter might also be a further solution for preventing oxidation, beside those presented in Section 3. As for applications, while promising, these have so far been developed as single components of more traditional systems. Scalability is also a challenge that has not yet been sufficiently addressed. If the adhesion and scalability will be satisfactorily addressed then, to the author's opinion, Cu NWs may become competitive with other technologies.