Nanoporous copper: fabrication techniques and advanced electrochemical applications

2.4.1 Chemical dealloying

Chemical dealloying has been extensively used for the fabrication of a number of NPMs (Jiang et al., 2013; Liu et al., 2011; Luo et al., 2013; Qi et al., 2009a,b; Sun et al., 2004; Wang et al., 2009; Zhang et al., 2009a,b; Zhao et al., 2009; Zhao et al., 2010). NPC has been synthesized by chemical dealloying a long ago. For example, Raney Cu was synthesized by the chemical dealloying of CuAl₂ binary alloys in NaOH solution (Smith et al., 1999). Through numerous studies on the formation process of NPC, it has been found that dealloying conditions have a major effect on the final porous structure of NPC (Hayes et al., 2006; Jiang et al., 2013; Liu et al., 2011; Luo et al., 2013; Qi et al., 2009b; Zhao et al., 2009). However, there is still a lack of systematic investigation on the effect of dealloying conditions on the formation and evolution of nanoporosity of NPC. Erlebacher et al. (2001) described in detail the physical mechanism behind dealloying and the evolution of nanoporosity in metals and suggested that an inherent dynamical pattern formation mechanism was involved in it.

To produce a uniform nanoporous structure by dealloying, two basic requirements of Cu-containing precursors are needed to be fulfilled: (1) homogeneous single-phase solid solution and (2) sufficient electrochemical difference between Cu and other alloying elements (Erlebacher, 2004). Chen et al. (2009) presented a systematic study on the formation of NPC with tunable nanoporosity. They carried out the selective dissolution of Mn out of Cu₃₀Mn₇₀ binary alloy precursor in HCl solution. Figure 8 represents the formation of uniform NPC obtained by dealloying Cu₃₀Mn₇₀ alloy in 0.025 m HCl solution. It can be seen from the cross-sectional SEM image that the ligament-channel structure was uniformly arranged throughout the sample. The uniform nanoporosity was further revealed by the transmission electron microscopy (TEM) images in Figure 8B.

Figure 8:

(A) Cross-sectional SEM image and (B) TEM micrograph of NPC produced by dealloying Cu₃₀Mn₇₀ ribbons in 0.025 m HCl solution for 12 h.

Reproduced from Chen et al. (2009), with permission from Wiley.

Chen et al. (2009) also described the effect of dealloying time on the formation nanoporosity. The coarsening of nanopores occurred with the increase in dealloying time. Figure 9 shows NPC formed with different pore sizes according to the subjected dealloying time. The average nanopores size could be tailored from ~15 to ~120 nm by monitoring the dealloying time. In particular, the bicontinuous structure of all samples was necessarily the same apart from the variation in the specific length of pores and ligaments, which is same as that of NPG.

Figure 9:

SEM images of NPC produced by dealloying Cu₃₀Mn₇₀ in 0.025 m HCl for (A) 0.5 h, (B) 2 h, (C) 6 h, (D) 12 h, (E) 20 h, and (F) 32 h.

Reproduced from Chen et al. (2009), with permission from Wiley.

Asghari et al. (2016) recently reported on NPC fabrication through dealloying for application in catalysis. Cu sheets were coated with a layer of Cu-Zn alloy. The zinc content of the alloy was decreased during immersion in HCl solution and an NPC electrode with lower zinc content was formed. Tin oxide nanoparticles were synthesized through a galvanostatic pathway on polypyrole (PPy)-coated NPC. In methanol solution, a significant enhancement in the catalytic action of PPy was observed after decoration of tin oxide nanoparticles. This enhancement is attributed to the effect of tin oxide on the adsorption of intermediates of methanol oxidation as well as the oxidation of by-products such as CO.

It has been reported that dealloying of one single-phase solid solutions in different electrolytes generated nanoporous structures with considerably different morphologies and length scales; however, all of the electrolytes produced bicontinuous ligament-channel structures. Hayes et al. (2006) studied the effect of different electrolytes on the morphology of NPC obtained by dealloying Mn_0.7Cu_0.3 in HCl, citric acid, H₂SO₄, MnSO₄, and (NH₄)₂SO₄+MnSO₄ aqueous solutions. Figure 10A–D shows the SEM images of NPC obtained from dealloying Mn_0.7Cu_0.3 in (A) pH1.3 HCl, (B) 1 m citric acid, (C) 0.01 m H₂SO₄+0.001 m MnSO₄, and (D) 1 m (NH₄)₂SO₄+0.01 m MnSO₄. Hayes et al. suggested that the NPC obtained with significant different ligament sizes and structures is attributed to the difference in dealloying rates and surface motilities of Cu in different electrolytes. Thus, the selection of suitable electrolyte is highly important to generate a nanoporous structure with desirable characteristics.

Figure 10:

Dealloying electrolytes affect both the ligament size and morphology.

These SEM micrographs show the structure resulting from dealloying Mn_0.7Cu_0.3 in (A) pH 1.3 HCl, (B) 1 m citric acid, (C) 0.01 m H₂SO₄+0.001 m MnSO₄, and (D) 1 m (NH₄)₂SO₄+0.01 m MnSO₄. Reprinted from Hayes et al. (2006), with permission from Materials Research Society.

2.4.2 Electrochemical dealloying

One of the promising strategies to fabricate NPMs is electrochemical dealloying (Luo et al., 2013; Sun et al., 2004; Zhang et al., 2009a,b; Zhao et al., 2010). The morphology of NPMs can be adjusted by fine-tuning the initial alloy composition and typical dealloying conditions. Numerous researchers studied the effect of initial precursor alloy composition on the final microstructure of NPMs (Aburada et al., 2011; Dan et al., 2013; Liu et al., 2011; Mao et al., 2012).

A number of reports have been published on the fabrication of NPC through the electrochemical dealloying of binary alloys (Dan et al., 2012, 2013; Hayes et al., 2006; Li et al., 2015; Luo et al., 2013; Qi et al., 2009b; Zhao et al., 2010). Tang et al. (2016) reported on the fabrication of well-ordered NPC by dealloying Cu-Mn. Wang et al. (2016) reported on the fabrication of nanosheets Co₃O₄ by oxidation-assisted dealloying method for high-capacity supercapacitors. Luo et al. (2013) described the nucleation and growth of NPC ligaments during the electrochemical dealloying of Mg-based metallic glasses. Zhang et al. (2009a,b) reported on the fabrication of a number of NPMs including Cu through the electrochemical dealloying of Al-based alloys. Zhang et al. found that interactions among the coexistent phases take place during the dealloying process of a two-phase alloy.

Zhao et al. (2010) described the electrochemical dealloying of Mg-Cu binary alloys in 0.2 m NaCl aqueous solutions and suggested that the electrochemical activities of this binary alloy were dependent not only on the initial alloy compositions but also on the phase constitutions. The critical potential and open circuit potential for MgCu₂ phase are much higher than those of Mg₂Cu phase by about 450 mV. Furthermore, the dissolution of less noble Mg₂Cu determines the critical and open circuit potential of biphasic alloy rather than being proportional to the amounts of constitutive elements or phases. The dealloying manners of single-phase and biphasic Mg-Cu alloys depend on the electrochemical properties of constitutive phases.

Figure 11 shows the cyclic voltammetry (CV)-based investigation of the dealloying behavior of Mg₅₀Cu₅₀ biphasic alloy under the applied potential of −1.0 to −0.3 V versus Ag/AgCl (Zhao et al., 2010). Both quenched alloys were composed of Mg₂Cu and MgCu₂ phases; however, the dealloying behaviors of the two alloys were different. In the first 10 cycles, a type of pitting corrosion was observed for the dealloying of biphasic Mg₅₀Cu₅₀ alloy (Figure 11A). That is, the current density of the forward scan is lower than that of the reverse scan. It was suggested that the dealloying process of Mg₂Cu is dominant at this initial stage. A crossover appeared in the 18th cycle but quickly disappeared with increasing cycles (Figure 11B). At lower potential section, the dealloying of more active Mg₂Cu phase took place because the current density of the reverse scan is higher than that of the forward scan (pitting corrosion), whereas at higher potential higher current density can be observed for the forward scan, which is linked with the dealloying of the less active MgCu₂ phase. Upon the dissolution of Mg₂Cu, the crossover disappears. The dealloying of the less active MgCu₂ phase is dominant at this stage. In addition, it can be observed that the dissolution rate of Cu is higher during the dealloying of MgCu₂ than that during the dealloying of Mg₂Cu with in the same potential zone.

Figure 11:

Evolution of CV curves of the rapidly solidified Mg₅₀Cu₅₀ alloy subjected to 80 cycles of CV measurements.

(A) Indicates 1, 3, 10 cycles of CV measurements and (B) indicates 18, 25, 50 and 80 cycles of CV measurements. Reprinted from Zhao et al. (2010), with permission from Elsevier.

Figure 12 represents the CV plots of Mg₃₃Cu₆₇ alloy subjected to the CV measurements between −0.6 and −0.2 V versus Ag/AgCl. A pitting corrosion was observed at higher potential during the initial dealloying stage, which indicates that the dissolution of Cu is unavoidable even at the start. The current density values during the initial 25 cycles (Figure 12A) are at least one order smaller than that of Mg₅₀Cu₅₀ alloy. It was noticed that current density gradually increases with the increase in CV cycles up to 200, even at a lower potential regime. In addition, a crossover was observed on all loops after the 50th cycle and reducing peak becomes indiscernible. Furthermore, with increasing cycles, the crossover transfers to the higher potential, suggesting that the dealloying process is converted from pitting-type corrosion to bulk corrosion (Figure 12B; Zhao et al., 2010).

Figure 12:

Evolution of CV curves of the rapidly solidified Mg₃₃Cu₆₇ alloy subjected to 200 cycles of CV measurements.

Reprinted from Zhao et al. (2010), with permission from Elsevier.

Figure 13 shows the section-view SEM images of as-dealloyed Mg₅₀Cu₅₀ and Mg₃₃Cu₆₇ alloys in 0.2 m NaCl solution at an applied potential of −0.3 V versus Ag/AgCl. The microstructure of as-dealloyed Mg₅₀Cu₅₀ consists of a bicontinuous interpenetrating ligament-channel structure with an average size of 188±45 nm (Figure 13A). Although two phases were present in the starting precursor Mg₅₀Cu₅₀, the final microstructure is homogeneous throughout the sample.

Figure 13:

Section-view SEM images showing the microstructures of NPC through potentiostatic dealloying of the rapidly solidified Mg-Cu (A) Mg₅₀Cu₅₀, (B) Mg₃₃Cu₆₇ alloys in the 0.2 m NaCl aqueous solution at the potential of −0.3 V versus Ag/AgCl.

Inset in (B): the entire section-view SEM of the Mg₃₃Cu₆₇ alloy ribbons. Reprinted from Zhao et al. (2010), with permission from Elsevier.

Figure 13B shows the cross-section-view microstructure of the as-dealloyed Mg₃₃Cu₆₇ alloy. A typical nanostructure is obtained after electrochemical dealloying. It is worth noting that ligaments are flake like, with sizes much smaller than those of the channels (120±30). On the contrary, ligaments observed in Figure 13A are rod-like and similar in size to that of channels. These microstructures are the result of typical electrochemical dealloying mechanism described in detail in the above section (Zhao et al., 2010).

McCue et al. (2016a) presented a comprehensive review of dealloying and dealloyed materials. They discussed the fundamental material principles underlying the formation of dealloyed materials and then looked at two major applications: catalysis and nanomechanics. McCue et al. (2016b) also studied the size effects in the mechanical properties of bulk bicontinuous Ta/Cu nanocomposites made by liquid metal dealloying.

3.2.1 Supercapacitors

Supercapacitors, also called electrochemical capacitors, have been recognized for several decades and are known as one of the promising potential energy storage devices in addition to batteries (Conway, 1999). Based on the electrochemical response, there are two types of electrochemical supercapacitors. One category of capacitor materials is that showing high surface reactivity resulting in the formation of a double layer, also basically known as electrical double-layer (EDL) capacitors (EDLC). The second ones are those exhibiting Faradic electrochemical reactions taking place on the surface (Lokhande et al., 2011). Supercapacitors can store considerably more energy per unit volume compared to conventional capacitors due to two reasons: (1) the charge separation occurs over a very little distance in the EDL, which is the distance between the electrolyte and the adjacent electrode, and (2) the substantially enlarged specific surface area of the electrode generated by the huge number of nanopores offers the ability to store an increased amount of charge. The process of energy storage is inherently rapid due to the simple movement of ions to and from electrode surfaces. The supercapacitors display a quite high degree of reversibility in charging-discharging processes and exhibit a cycle life of more than 500,000 cycles (Lokhande et al., 2011). These excellent properties make the supercapacitors promising for a number of applications in portable electronic market, power quality systems, and particularly low-emission buses, cars, and trucks (Winter & Brodd, 2004).

Supercapacitors consist of two nonreactive porous plates suspended in an electrolyte under an applied voltage across the two porous plates (Divyashree & Hegde, 2015; Khan et al., 2000). When the potential is applied on two plates, the positive plate (anode) attracts positive charges from the electrolyte, whereas the negative plate (cathode) attracts electrons as displayed in Figure 16. This phenomenon efficiently provides sufficient space to the two layers for storing charge (Brenna et al., 2009). Therefore, supercapacitors are well suited for the electrostatic storage of charges effectively.

Figure 16:

Schematic diagram of charge distribution in a supercapacitor.

Reproduced from Divyashree and Hegde (2015), with permission from The Royal Society of Chemistry.

The charging and discharging mechanism of EDL supercapacitors is presented in Figure 17 (Divyashree & Hegde, 2015). It represents the accumulation of the charges around an electrode during the charging/discharging cycle of the supercapacitors.

Figure 17:

Structure of double-layer capacitor during charging and discharging.

Reproduced from Divyashree and Hegde (2015), with permission from The Royal Society of Chemistry.

To better understand the specific power density and energy density of supercapacitors in comparison to batteries and electrochemical energy conversion systems, a Ragone plot is presented in Figure 18 (Simon & Gogotsi, 2008). It can be observed clearly that supercapacitors occupy a specific and significant position in the Ragone plot. Today’s need of energy storage is met by supercapacitors because of its extraordinary energy density and huge power capability. Supercapacitors can work in a broad temperature range; they exhibit long cycle life and can supply high-power density.

Figure 18:

Ragone plot of energy storage devices.

Reproduced from Simon and Gogotsi (2008), with permission from Nature Publishing Group.

A large number of articles and reviews have been published on the high capacitance electrode materials for supercapacitors (Burke, 2000; Chmiola et al., 2010; Conway, 1999; Divyashree & Hegde, 2015; Kotz and Carlen, 2000; Lang et al., 2011; Mahon & Drummond, 2001; Miller & Burke, 2008; Nishino, 1996; Shukla & Martha, 2001; Simon & Burke, 2008; Wang et al., 2012; Zhang & Zhao, 2009; Zheng et al., 1997). Numerous reports have been presented on the fabrication of nanoporous materials for capacitors applications. For example, Nguyen et al. (2016) recently reported on hydrogen bubbling-induced microporous/nanoporous MnO₂ films prepared by electrodeposition for pseudocapacitor electrodes. Capacitance depends largely on the features of electrode materials, especially surface areas and pore size distribution. Cupric oxide (CuO) has been widely used for high capacitance supercapacitors. NPC in the form of composite materials has been successfully used for high-energy density supercapacitors.

Prasad et al. (2016) recently reported on the fabrication and characterization of nanoporous carbon electrodes embedded with CuO nanoparticles for supercapacitors. Zhang et al. (2011b) demonstrated the porous CuO nanobelts for high-energy density pseudocapacitors. Moosavifard et al. (2015) recently reported on the fabrication of 3D highly ordered nanoporous CuO electrode for high-performance asymmetric supercapacitors. They described that CuO with highly ordered and interconnected nanopores and nanowalls have a specific surface area of about 149 m² g⁻¹ and exhibited a specific capacitance of 431 F g⁻¹ at 3.5 mA cm⁻². They assembled the electrode in an asymmetric cell that demonstrated energy density of about 19.7 W h kg⁻¹ and a very high cycle life.

3.2.2 Lithium ion batteries

Batteries are electrochemical rechargeable energy storage devices that convert chemical energy into electrical energy as a result of a redox reaction taking place at anode and cathode (Becker, 1957; Linden, 1984; Samuel, 1964). This principle was first presented by Becker (1957). When an external load is applied to the battery, a flow of electrons takes place from anode to cathode resulting in the discharging of the battery (Appleby, 1988; Henne, 2007). A schematic of a conventional battery with cathode, anode, and electrolyte is presented in Figure 19. When a battery has to be recharged, the electrons’ flow direction is reversed using an external power source. This involves the restoring of the original states of cathode and anode for further redox reactions (Ogasawara et al., 2006).

Figure 19:

Cross-section of a conventional rechargeable battery with anode, electrolyte, and cathode connected using an external electrically powered device.

Reproduced from Divyashree and Hegde (2015), with permission from The Royal Society of Chemistry.

Nanostructured electroactive materials can effectively enhance the rate of lithium ion insertion/extraction to increase the power output of batteries. However, sluggish ion and electron transport kinetics combined with conventional approaches for the fabrication of nanostructured metal oxide results in considerable reductions in the energy capacities. Hou et al. (2013) reported on the 3D bicontinuous NPC/MnO₂ hybrid electrode for high-performance lithium ion batteries. The solid/nanoporous hybrid enhanced the electron transport of MnO₂, facilitated the fast ion diffusion, and exhibited ultrahigh charge/discharge rates and a stable capacity of 1100 mAh g⁻¹ for 1000 cycles. Figure 20 shows the fabrication process of seamlessly integrated solid/NPC/MnO₂ bulk electrode.

Figure 20:

Scheme showing the fabrication of seamlessly integrated S/NP Cu/MnO₂ bulk electrode: (A) cleaned Cu foil substrate; (B) Cu₃₀Mn₇₀ alloy film deposited on the Cu foil by magnetron sputtering; (C) NPC layer on the Cu foil produced by chemically dealloying Cu₃₀Mn₇₀ in diluted HCl solution; (D) nanocrystalline MnO₂ directly grown onto S/NP Cu skeleton using electroless plating; (E) photograph of a flexible S/NP Cu/MnO₂ hybrid bulk electrode being bent (2×3 cm²); (F) four batteries assembled with S/NP Cu/MnO₂ bulk electrodes power blue, red, and green LEDs; (G) schematic battery constructed with S/NP Cu/MnO₂ and lithium foil as electrodes, 1 m LiPF₆ in EC/EMC/DMC as electrolyte, and porous polymer as separator.

Reprinted from Hou et al. (2013), with permission from Nature Publishing Group.

Liu et al. (2013) reported on the fabrication of NPC-supported cuprous oxide (Cu₂O) electrode for high-performance lithium ion batteries. The hybrid electrode was fabricated by the selective etching of Cu₅₀Al₅₀ alloy followed by in situ thermal oxidation. During the first cycle, a high lithium storage capacity of about 2.35 mA h cm⁻² and a high reversible capacity of 1.45 mA h cm⁻² were attained after 120 cycles.

Figure 21A presents the fabrication mechanism of NPC-supported Cu₂O schematically. Ingots of Cu₅₀Al₅₀ alloy were first prepared by electron beam melting and Al was selectively dissolved in NaOH aqueous solution to produce NPC (Figure 21B and C). Finally, to prepare a Cu₂O film on the surface of NPC, thermal treatment was carried out for 3 min in 140°C in air (Liu et al., 2013).

Figure 21:

(A) Schematic showing the fabrication process of 3D NPC@Cu₂O anodes, (B) and (C) SEM images of NPC fabricated by dealloying Cu₅₀Al₅₀, and (D) a digital photograph showing a Cu₅₀Al₅₀ ingot and NPC fabricated by dealloying Cu₅₀Al₅₀.

Reprinted from Liu et al. (2013), with permission from The Royal Society of Chemistry.

3.3.1 CO₂ reduction

The production of hydrocarbons and liquid fuels from the electrochemical reduction of CO₂ has been extensively investigated because CO₂ is the huge and sustainable carbon feedstock and the conversion of CO₂ into hydrocarbons and liquid fuels could provide an incentive for the CO₂ capture (Gattrell et al., 2007; Hori, 2008). However, efficient and durable electrocatalysts for the electroreduction of CO₂ and its derivatives into desirable fuels are not available at the present (Benson et al., 2009; Costentin et al., 2012; Costentin et al., 2013). Many catalysts can reduce the CO₂ to CO (Chen et al. 2012; DiMeglio & Rosenthal, 2013; Ebbesen & Mogensen, 2009; Sen et al., 2014; Tornow et al., 2012), but the synthesis of liquid fuels requires that CO should be further reduced using H₂O as H⁺ source. CO is a derivative of CO₂, which acts as an intermediate during the production of hydrocarbons.

Since the discovery of Hori et al. in 1985, Cu has been known as a decent catalyst for the electrochemical reduction of CO₂/CO to hydrocarbons (Hori et al., 1985, 1986). However, in bulk form, its efficiency and selectivity for the production of higher hydrocarbons and liquid fuels are far too low for commercial applications. A mixture of compounds is produced on the polycrystalline Cu foil in CO saturated aqueous solutions, which at low overpotentials are dominated by H₂ and at high overpotentials are dominated by CO and HCO₂ and by multicarbon oxygenates and hydrocarbons at very extreme potentials (Hori et al., 1989; Kuhl et al., 2012). Particularly, H₂O reduction to H₂ outcompetes CO reduction on the Cu electrode at low overpotentials, and at higher overpotentials, gaseous hydrocarbons are the major products of CO₂ reduction (DiMeglio & Rosenthal, 2013; Ebbesen & Mogensen, 2009). Hence, there is a high need of a catalyst that can work efficiently at low overpotentials.

NPC has been reported as an exclusive novel material for the electrochemical reduction of CO₂ into hydrocarbons with high Faraday efficiency and selectivity (Jia et al., 2014; Li & Kanan, 2012; Li et al., 2014; Manthiram et al., 2014; Sen et al., 2014; Tang et al., 2012). Ponnurangam et al. (2016) recently reported on a poly(4-vinylpyridine)-Cu electrode as a robust catalyst for the electroreduction of CO₂. The cathodic current relating to the production of hydrocarbons is significantly improved on NPC during the electroreduction of CO/CO₂ at very low overpotentials (Jia et al., 2014). The onset potential for the reduction of CO₂ at porous Cu foam was −1.0 V versus Ag/AgCl with the formation of formic acid (HCOOH) initially (Figure 22). The Faraday efficiency of HCOOH rises from 4% to 26% at −1.1 V on porous Cu foam, which is appreciably higher than that for smooth Cu (i.e. <1% at −1.1 V). This value increases to 37% at −1.5 V, which is the highest value of Faraday efficiency reported to date for HCOOH at a Cu electrode (Sen et al., 2014). Furthermore, the production of propylene from CO₂ reduction has been observed for the first time on a high surface area nanostructured Cu foam electrode (Figure 22).

Figure 22:

Product distribution as a function of applied potential obtained during the electrochemical reduction of CO₂. The working electrode was Cu nanofoam electrodeposited for 15 s. Data for the electrochemical reduction of CO₂ to formate at a smooth Cu electrode was also included for comparison.

Reproduced from Sen et al. (2014), with permission from American Chemical Society.

The main primary reactions occurring during the electrochemical reduction of CO₂ at a Cu electrode are given below. A series of steps (1) to (6) are predicted for the electroreduction of CO₂ at the Cu electrode (Sen et al., 2014). The asterisk in any reaction indicates either a surface-bound species or a vacant active site.

From the above reaction steps, we can conclude the complete reactions for the formation of hydrocarbons and liquid fuels.

The formation of HCOOH, CO, and alcohols during the electroreduction of CO₂ has been reported for various Cu electrodes; however, the porous Cu nanofoam is the only catalyst that generates propylene in detectable quantities. Although propylene has been measured at very low yields on porous Cu nanofoam, it has not been observed on any other Cu electrodes reported to date (Sen et al., 2014). The production of ethylene and propylene suggests that the NPC foam can provide both the nanostructured surfaces and cavities that promote the reaction between CO₂ and hydrogen to produce the higher hydrocarbons (ethylene, propylene, etc.) during the electroreduction of CO₂.

It has been reported by many researchers that that electroreduction of CO₂ on Cu is very sensitive to its surface structure (Goncalves et al., 2013; Hori et al., 2002, 2003; Montoya et al., 2013). Hori et al. (2002, 2003) described that the formation of C₂H₄ occurs preferentially at Cu(100) facets at potentials lower than that for Cu foil, whereas the formation of CH₄ took place selectively at Cu(111) at same potentials. The most negative potentials were required for the reduction of CO₂ at Cu(110) facets, and as a result, C₂ and C₃ products were obtained. These investigations indicate that the surface structure of the electrode plays a crucial role in the electroreduction of CO₂.

Sen et al. (2014) suggested that the mechanism of CO₂ electroreduction was changed by porous Cu foam based on the observation of following points:

1. Enlarged Faraday efficiency of formate at all potentials

2. Reduced Faraday efficiency of CO and CH₄

3. Formation of saturated hydrocarbons (e.g. C₂H₆)

4. Generation of novel higher hydrocarbons, namely, C₃H₆.

Insight into the importance of I to III was provided in detail by Norskov’s group (Durand et al., 2011; Peterson et al., 2010). They reported the detailed description of chemical processes occurring during the CO₂ electroreduction at the Cu-water interface. Finally, point IV, the generation of novel C₃-hydrocarbons (namely, propylene) might be caused by the hierarchical porous structure of Cu foam.

The production of saturated hydrocarbons, such as ethane, and C₃ products, such as propylene, on hierarchical porous Cu foam suggested that the mechanism of CO₂ electroreduction on Cu nanofoam might be changed, as these products were not observed on smooth Cu surface by previous studies. The generation of the saturated hydrocarbons and C₃ products may be caused by the hierarchical porous structure of the Cu foam because the residence time of the various intermediates within these confined spaces of pores may increase, allowing for the generation of products that were not observed on smooth Cu electrodes. The idea that nanoporous electrodes favor the reaction pathways that are different from those observed at smooth electrodes via a confinement mechanism has been studied in detail in the electroreduction of O₂ at nanoporous Pt electrodes (Bae et al., 2012; Boo et al., 2004).

Reaction pathways and intermediates: Formate and CO are formed at first before the production of hydrocarbons at higher cathodic potentials. Cu nanoparticles prepared by traditional vapor condensation produce about 96% CO from the reduction of CO₂ (Rosen et al., 2011). This might be due to the reason that CO is first produced before the production of alcohols. Although the conversion mechanism of CO₂ to liquid fuels is not very clear at present, but it is widely accepted that CO₂ is first reduced to CO and then it is further converted to alcohols and other hydrocarbons through multistep electron transfer pathways. Earlier studies have found that CO reduction on Cu results in similar products as obtained from CO₂ reduction, proposing that CO was obtained as an intermediate during the electroreduction of CO₂ (DeWulf et al., 1989; Hori et al., 1987; Laitar et al., 2005; Sullivan et al., 2012). Further reactions of formate do not produce any quantifiable products (Cook, et al., 1989).

Two voltage-dependent pathways were predicted for the electrochemical reduction of CO₂ at a Cu electrode (Durand et al., 2011; Peterson et al., 2010).

1. The formate (OCHO) or F-intermediate pathway and

2. The carboxyl (COOH) or C-intermediate pathway

The F-intermediate pathway leads exclusively to formic acid and the C-intermediate pathway leads to formic acid and high-order hydrocarbons. The calculations of the products obtained from both pathways have been performed on (111), (100), and (211) surfaces of Cu. The F-intermediate pathway dominates on (111) and (100) surfaces, because it has the lowest change in free energy of the two pathways (Durand et al., 2011; Schouten et al., 2012). It was proposed that (100) and (200) surfaces of Cu are assumed to be equivalent in density functional theory calculations, and the production of HCOOH via the F-intermediate pathway is expected to be enhanced at Cu nanofoam as 22% more (200) surface was observed comparatively (Sen et al., 2014).

These earlier studies signify that CO is adsorbed on the electrode surface during CO₂ reduction and this CO acts as an intermediate during the hydrocarbons production. The second point was well documented in the report of Smith et al. (1997), where saturated solutions of both CO and CO₂ showed similar spectra. Due to high reportage of CO during the reduction of CO₂, the adsorbed CO reduction was proposed to be the rate-determining step for the entire reaction to hydrocarbon formation (Kim et al., 1988). Therefore, many researches have been carried out on the CO reduction, as this takes a benefit of only a few reaction steps to be considered (Hori et al., 1997; Li et al., 2014; Schouten et al., 2013; Verdaguer-Casadevall et al., 2015; Zhang et al., 2014).

3.3.2 CO reduction

CO reduction at Cu electrode was studied in detail by Hori’s group (Hori et al., 1997). CO reduction reactions were conducted at different pH values using borate buffer or phosphate buffer. Watanabe et al. (1994) evaluated the state of adsorbed CO anion radical on Cu using ab initio calculations. They predicted a little decrease in Cu-C bonds and an uplift of the C-O bonds to about 1.25 Å; hence, they estimated a mostly double-bond character (Watanabe et al., 1993, 1994). When Cu is supplied with only CO in the absence of CO₂, higher hydrocarbons and multicarbon oxygenates are produced; however, to compete with H₂ evolution, much negative potentials are still required to stimulate CO reduction (Hori et al., 1987, 1997). Energetic efficient electrolysis is impeded by large overpotentials and favor the production of hydrocarbons over multicarbon oxygenates.

Recently, it was discovered that oxide-derived nanocopper has a very high selectivity for CO reduction over H₂ evolution than that of polycrystalline Cu foil (Li & Kanan, 2012). It was reported in a current letter to Nature (Li et al., 2014) that CO can be reduced with a very high Faraday efficiency (57%) over oxide-derived Cu at a very low overpotential of −0.3 V versus reversible hydrogen electrode (RHE), which is >0.6 V lower compared to polycrystalline Cu foil. It was proposed that high CO reduction activity was due to the grain boundaries participating in catalysis. The major products obtained from the reduction of CO were ethanol and acetate, which related to 8e⁻ and 4e⁻ reductions using H₂O as H⁺ ion source (Li et al., 2014).

The activity of CO reduction was measured using steady-state conditions through constant potential electrolysis in CO saturated (1 atm) 0.1 m KOH solution at ambient temperature. Polycrystalline Cu foil showed lower current density and H₂ was the only measureable product under these similar conditions. Another study has shown 22% overall Faraday efficiency for the reduction of CO at very negative potentials (about −0.7 V vs. RHE) and its Faraday efficiency for oxygenates was only 7% (Hori et al., 1987). A maximum of 65% Faraday efficiency from CO reduction have been stated for polycrystalline Cu foil at −0.9 V versus RHE, but Faraday efficiency of oxygenates was only 10% (Hori et al., 1997).

CO reduction over oxide-derived Cu exhibited quite higher geometric current densities than that of over Cu nanoparticles because of their roughness factor (Figure 23). Faraday efficiency was ≥94% for H₂ evolution at all potentials for Cu nanoparticle electrode, and small remaining current was related to ethanol, acetate, and ethylene formation. On the contrary, a much higher tendency of CO reduction was shown by oxide-derived Cu electrodes. Oxide-derived Cu-1 electrode showed an overall CO reduction efficiency of ~57% at −0.3 V versus RHE, whereas oxide-derived Cu-2 showed 48% Faraday efficiency at −0.4 V versus RHE. With increase in the negative potentials, these values decreased because of catalyst reaching mass-transport-current density for the CO reduction. Oxide-derived Cu electrodes represented higher intrinsic CO reduction activity due to its higher Faraday efficiency compared to Cu nanoparticles (Li et al., 2014).

Figure 23:

Faraday efficiencies for CO reduction products ethanol (EtOH), acetate (AcO-), n-propanol (n-PrOH), ethylene (C₂H₄), and ethane (C₂H₆) at specific potentials versus RHE.

Reprinted from Li et al. (2014), with permission from Nature Publishing Group.

Electrolysis data (Figure 24) give some insights into the CO reduction mechanism on oxide-derived Cu. The absence of C₁ products indicates that C-C coupling was very fast during the start of CO reduction (Montoya et al., 2013) or maybe because initial electron transfer was combined to the C-C bond formation between a CO from solution and surface-bound CO (Calle-Vallejo & Koper, 2013). It is estimated that acetate formation results from the attack of HO⁻ on a carbonyl-containing intermediate or a surface-bound ketene after the C-C bond formation. This argument was signified from the observation of enlarged acetate formation when electrolysis was executed in 1 m KOH solution (Figure 24).

Figure 24:

Faraday efficiency of various products for CO reduction bulk electrolysis on oxide-derived Cu-1 in 1 m KOH solution saturated with 1 atm CO.

Reprinted from Li et al. (2014), with permission from Nature Publishing Group.

Detailed mechanism can be observed vividly from Figure 25. In our previous report (Abbas et al., 2016), we suggested that CO₂ is first converted into CO (ads) active species at Cu or possibly at another electrode and binds to the catalyst. This CO (ads) may dimerize in the presence of an alkaline media and form C₂O₂ (ads). The hydrogenation process of these species would produce CH₂CHO (ads). Two pathways originate from this intermediate toward the formation of ethylene and ethanol. (1) The hydrogenation of the CH₂ group leads to the formation of acetaldehyde, which is eventually reduced to ethanol. (2) The hydrogenation of C in the carbonyl group leads to the formation of ethylene and O (ads), which quickly forms water.

Figure 25:

Schematic of the mechanism for the electroreduction of CO₂ into hydrocarbons at NPC electrode.

Reprinted from Abbas et al. (2016), with permission from Taylor & Francis Group.

Influence of electrolyte, temperature, and pressure on the CO/CO₂ reduction: The Gouy-Chapman theory describes the relation between the Debye length of EDL and electrolyte concentration (Boo et al., 2004). They proposed that with the increase in electrolyte concentration the EDL thickness decreases and at a critical value the EDL becomes thin enough that the electric fields of adjacent pores do not overlap anymore. In this way, the electric field maps the exact shape of the pore and hence provides the extra surface area to participate in electrochemical reactions.

In Figure 26, chronoamperometric measurements show that at a smooth Cu electrode the current density gradually increases from 7 to 31 mA cm⁻² (~4.5×) with the increase of electrolyte concentration from 0.1 to 1 m. On the contrary, at NPC foam (60 s), the current density increases sharply from 10 to 82 mA cm⁻² (~8×) above a critical concentration of 0.5 m KHCO₃ (Sen et al., 2014). Thus, this data shows that nanoscale pores present inside the Cu foam are only accessible at a concentration above 0.5 m KHCO₃ (i.e. at this point, the thickness of the double layer is minimum). These 3D small electroactive areas enable the production of C₂ and C₃ products (e.g. propylene due to increase of residence time of surface intermediates; Sen et al., 2014).

Figure 26:

Chronoamperometric measurements of CO₂ electroreduction at smooth Cu and NPC foam plotted versus electrolyte concentration.

Reproduced from Sen et al. (2014), with permission from American Chemical Society.

The product distribution from CO₂ reduction is also influenced by the reaction temperature. It was reported in galvanostatic measurements being carried out over the range of 0°C–40°C that lower temperature results in a variation in current density for formate, ethylene, and methane. Current efficiencies for ethylene formation and hydrogen evolution were decreased and the efficiency of methane was increased, reaching a value of 65% at 0°C (Hori et al., 1986). Some other workers have reported, in addition to the increase in current efficiency for methane, a shift in Tafel slop from 93 mV decade⁻¹ at 22°C to 520 mV decade⁻¹ at 0°C (Kim et al., 1988).

It was confirmed by Kanan’s group (Li et al., 2014) that geometric current densities of CO reduction increase with the increase of CO pressure. The geometric current density was 1.8–2.4-fold higher at 2.4 atm CO compared to 1 atm CO at potentials between −0.3 and −0.5 V (Figure 27). The Faraday efficiency for CO reduction was also significantly improved at potential E<−0.3V under these conditions. These results point out that the practical current densities may be possibly obtained with the more increase in the CO transport to the catalyst.

Figure 27:

Comparison of CO reduction in 0.1 m KOH saturated with 1 atm CO versus 2.4 atm CO.

Reprinted from Li et al. (2014), with permission from Nature Publishing Group.

3.3.3 Organic synthesis using NPC

NPC-catalyzed click reaction: Click reaction has recently emerged as a promising method for organic synthesis (Hu et al., 2013; Markiewicz et al., 2010; Milles et al., 2012; Yamamoto, 2014). The Huisgen [3+2] cycloaddition of terminal alkynes and organic azides, namely, click reaction, has been widely catalyzed by homogeneous Cu(1) salts for organic synthesis. Heterogeneous Cu(1) has also been used to catalyze the click reaction, such as Cu-in-charcoal nanoparticles, immobilized Cu nanoclusters and CuO nanoparticles, Cu₃N nanoparticles, and Cu zeolites. Researchers also reported that click reaction also be catalyzed by Cu metal; however, the reaction was very slow, which required high catalytic loading.

Jin et al. (2011) carried out the click reaction using NPC as a catalyst. Click reaction catalyzed by NPC progressed very smoothly and catalyst efficiency was greatly dependent on its nanoporosity (Figure 28). The highest trizole yield was obtained when cat-3 with a ligament size of 40 nm was used. The specific surface area of NPC-cat-3 measured by BET method was 14 m² g⁻¹ and the turnover frequency (TOF) extended to 0.26/s.

Figure 28:

Click reaction dependence on nanoporosity.

Reproduced from Jin et al. (2011), with permission from Wiley.

Yamamoto (2014) carried out click reaction using cat-3 catalyst for a wide range of organic azides and its scope is presented in Figure 29. Triazoles substituted by various functional groups could be synthesized in high chemical yields. Catalyst loading could be reduced to 0.1 mol% under clean conditions, mean without using toluene solvent, and the reaction efficiency was still similar to that of 2 mol% NPC in toluene solvent and TOF was lifted up to 5.2 s⁻¹. Moreover, cat-3 NPC can be reused more than 10 times under conditions presented in Figure 28.

Figure 29:

Scope of click reaction using cat-3 NPC.

Reproduced from Yamamoto (2014), with permission from Elsevier.