- Email: [email protected]
Fabrication of a three-electrode battery using hydrogen-storage materials
Journal of Power Sources 280 (2015) 125e131
Contents lists available at ScienceDirect
Journal of Power Sources journal homepage: www.elsevier.com/locate/jpowsour
Fabrication of a three-electrode battery using hydrogen-storage materials Chi-Woo Roh, Jung-Yong Seo, Hyung-Seok Moon, Hyun-Young Park, Na-Yun Nam, Sung Min Cho, Pil J. Yoo, Chan-Hwa Chung* School of Chemical Engineering, Sungkyunkwan University, Suwon 440-746, Republic of Korea
h i g h l i g h t s
g r a p h i c a l a b s t r a c t
We design new-concept three-electrode battery using hydrogen as energy carrier. Charging process is electrolysis of alkaline electrolyte. Discharging process is operation of alkaline fuel cell. Hydrogen is stored at bifunctional electrode as metal hydride with charging process. Bifunctional electrode acts as electrolysis cathode and fuel cell anode at same time.
a r t i c l e i n f o
a b s t r a c t
Article history: Received 28 October 2014 Received in revised form 26 December 2014 Accepted 16 January 2015 Available online 16 January 2015
In this study, an energy storage device using a three-electrode battery is fabricated. The charging process takes place during electrolysis of the alkaline electrolyte where hydrogen is stored at the palladium bifunctional electrode. Upon discharging, power is generated by operating the alkaline fuel cell using hydrogen which is accumulated in the palladium hydride bifunctional electrode during the charging process. The bifunctional palladium electrode is prepared by electrodeposition using a hydrogen bubble template followed by a galvanic displacement reaction of platinum in order to functionalize the electrode to work not only as a hydrogen storage material but also as an anode in a fuel cell. This bifunctional electrode has a sufficiently high surface area and the platinum catalyst populates at the surface of electrode to operate the fuel cell. The charging and discharging performance of the three-electrode battery are characterized. In addition, the cycle stability is investigated. © 2015 Elsevier B.V. All rights reserved.
Keywords: Bifunctional electrode Palladium hydride Dynamic hydrogen-bubble template Dendrite structures Three-electrode battery
1. Introduction Recently, along with high-edge technologies for portable electric devices, energy storage systems with a high volumetric energy density are highly demanded. Currently, electric energy is generally
* Corresponding author. E-mail address: [email protected] (C.-H. Chung). http://dx.doi.org/10.1016/j.jpowsour.2015.01.101 0378-7753/© 2015 Elsevier B.V. All rights reserved.
stored in secondary batteries or capacitors. Whereas capacitors have lower volumetric energy densities than secondary batteries as well as somewhat unsuitable discharge characteristics for portable electronic devices, secondary batteries possess slightly more suitable discharge characteristics and higher volumetric energy densities [1]. Therefore, secondary batteries are more advantageous than capacitors for configuring small scale energy storage systems. Among secondary batteries, Li-ion batteries are currently the most commonly used for energy storage applications. However, in
126
C.-W. Roh et al. / Journal of Power Sources 280 (2015) 125e131
the case of thin film lithium ion batteries which have a small volume, it has been found that there is a stability issue which limits increasing the film thickness because of the increased channel distance through which Li ions come in and out. As a result, it is hard to form a structure thicker than a few mm [2]. On the other hand, in the case of metal hydride, hydrogen atoms penetrate into the metal lattice structure, which is much larger than the size of the hydrogen atoms being stored. Therefore, the channels for hydrogen atoms are not plugged so that there are fewer limitations concerning the electrode thickness. In addition, metal hydride can theoretically have a volumetric energy density of 2000 Wh/l or higher, which is larger than that of thin film Li-ion batteries [1]. Metal hydrides can be utilized in fuel cell (FC) type secondary batteries. In this FC-type battery, the oxidation of hydrogen stored as a hydride, and the reduction of oxygen occur in an alkaline electrolyte in the discharging process. In this case, an active type fuel cell, into which pure oxygen is injected, as well as a passive type fuel cell, to which oxygen in air is supplied through a breathing process, can be formed for discharging. This FC-type secondary battery which employs a metal hydride has many advantages. The fuels are only hydrogen and oxygen which are nontoxic and completely environmentally friendly. In addition, the electrolyte is
an aqueous alkaline solution instead of an explosive organic solution. Also, it charges faster or requires no charging time, has a longer service life as well as a lack of self-discharge [3]. As a result, FC-type secondary batteries are considered as a good alternative secondary battery and diverse studies of these battery systems have been carried out. For example, G. Erdler and C. Bretthauser [4e7] et al. designed chip integrated FC-type secondary batteries using polymer electrolyte membranes. In this study, based on these recent technical trends, we designed a 3-electrode battery, as shown in Fig. 1, and adopted palladium as hydrogen storage materials among the many potential hydrogen storage metals and alloys. Since the present study focuses on the performance of batteries in relation to the structure of the metal hydride electrodes, palladium, which is easy to handle and possesses excellent ability for hydrogen storage at ambient condition [8,9], is a suitable material. The batteries designed in the present study are in the form of three electrodes immersed in an alkali electrolyte solution, as shown in Fig. 1. The electrodes are an electrolysis anode, a bifunctional electrode, and a fuel cell cathode. Fig. 1(a) shows a mimetic diagram of the operation of the charging process used to store hydrogen in the bifunctional electrode. In this hydrogen charging process, protons meet equivalent electrons at the bifunctional electrode and are stored as palladium hydride (PdHx) while water is electrolyzed and O2 is evolved at the electrolysis anode. The chemical reactions of this process are as follows.
c c Anode ðelectrolysis anodeÞ : cOH / H2 O þ O2 þ ce 2 4 (1) Cathode ðbifunctional electrodeÞ : Pd þ cH2 O þ ce / PdHc þ cOH
Overall reaction : Pd þ
c c H O / O2 þ PdHc 2 2 4
(2)
(3)
Fig. 1(b) shows the discharging process which utilizes the oxidation reaction of the hydrogen stored in the bifunctional electrode. In this hydrogen discharging process, the fuel cell is operated and each hydrogen atom stored in the form of palladium hydride meets an OH ion to form a water molecule and discharge one electron. As a result, an electric current flows in this process. The process can be expressed by the following chemical formulas.
Anode ðbifunctional electrodeÞ : cOH þ PdHc /cH2 O þ Pd þ ce
Cathode ðfuel cell cathodeÞ :
Overall reaction : PdHc þ
Fig. 1. Schematic diagram of the (a) charging and (b) discharging processes of the 3electrode battery. In the charging process, protons are stored as palladium hydride (PdHx) in the bifunctional electrode. In the discharging process, hydrogen atoms stored in the form of PdHx meet OH ions to become water and discharge electrons.
(4)
c c H O þ O þ ce / cOH 2 2 4 2 (5)
c c O / H2 O þ Pd 4 2 2
(6)
The characteristics of the individual electrodes of the 3electrode batteries developed using this concept were analyzed by SEM (scanning electron microscope), EDS (energy-dispersive Xray spectroscopy), and ICPeAES (inductively coupled plasmaeatomic emission spectroscopy). In addition, the chargeedischarge characteristics and cycling stability of the fabricated 3-electrode batteries were analyzed through electrochemical experiments.
C.-W. Roh et al. / Journal of Power Sources 280 (2015) 125e131
2. Experimental The electrolysis anode was made to possess a small resistance by depositing gold on 100 mesh size Ni mesh (Nilaco, Japan). Before being deposited with gold, the Ni mesh was pretreated as follows to facilitate the gold deposition. First, the Ni mesh was treated with acetone, iso-propyl alcohol (IPA), and ethanol at 50 C for 20 min each to remove organic matter. Second, the Ni mesh was treated at 50 C with 6 M HCl for 30 min to remove the oxide film. The above process was implemented in an ultrasonic cleaner (SAEHAN, Republic of Korea). The gold was deposited via electroless deposition. The electroless deposition solution was prepared as follows. A 100 mL/L GoBright® TWX-40-M10 (C. Uyemura, Japan) solution, 12 mL/L gold solution (Au 100 g/L, C. Uyemura, Japan), and 0.5 mL/L potassium cyanide solution (KCN 100 g/L, C. Uyemura, Japan) were mixed. The deposition process then proceeded as follows. First, the electroless deposition solution was added to a thermostatic tank and the temperature was fixed at 60 C using a water bath. Then, the pretreated Ni mesh was immersed in the solution for 40 min so that the entire surface was sufficiently deposited with gold. We made sure that the Ni mesh was completely immersed in the solution such that no part of the Ni mesh was in contact with air. When the entire surface was evenly deposited with gold, the Ni mesh was washed with D.I. water and dried in a vacuum dryer. The fuel cell cathode was prepared by applying 5 mg/cm2 Pt black catalyst (Alfa Aesar, UK) on GDL pre-coated carbon paper (CNL Energy, Republic of Korea) using a spray method. In this case, the catalyst ink contained catalyst and Nafion contents of 85 wt.% and 15 wt.%, respectively, in the solvent (D.I. water:IPA ¼ 1:4, volume ratio) while they were dispersed for at least one hour in an ultrasonicator. As mentioned above, the bifunctional electrode plays the role of both an electrolysis cathode and a fuel cell anode. To determine the effect of the structure of this electrode on the performance of the battery, three electrodes, PdeCu, PdeCu/Pt, and PdeCu/Pt/PdNP, were prepared. PdeCu was made by placing the pretreated Ni mesh as a working electrode and a platinum plate as a counter electrode into an electrolyte composed of 5 mM CuSO4, 15 mM PdCl2, and 1 M H2SO4 while inducing electrodeposition for five minutes at a very high voltage of 4 V without any reference electrode. The voltage of 4 V lies in the hydrogen evolution reaction (HER) region where hydrogen is actively generated. When electrodeposition is induced in this voltage region, structures with a high surface area are made because of dynamic hydrogen bubble templates [10]. The PdeCu/Pt electrode was made by immersing the PdeCu electrode fabricated using the above mentioned deposition process in a platinum displacement solution composed of 5 mM H2PtCl6 and 0.1 M HClO4 for 90 min to displace the copper with platinum through galvanic displacement reactions, thereby forming platinum on the surface. The PdeCu/Pt/PdNP electrode was made by additionally applying palladium nano-powder (PdNP) on PdeCu/Pt. PdNP was made by utilizing galvanic displacement reactions to take advantage of the reduction potential differences between aluminum and palladium. The aluminum plate was immersed in 1 M NaOH for five minutes to remove the oxide layer followed by soaking in a solution composed of 10 mM PdCl2 and 1 M H2SO4. Then, PdNP with a high surface area was obtained with hydrogen generation [11]. Then, a mixture of the obtained PdNP and Nafion was prepared at a PdNP:Nafion weight ratio of 9:1 in the solvent (D.I. water:IPA ¼ 1:4, volume ratio) where it was evenly dispersed for at least one hour in an ultrasonicator. Then, the obtained ink was applied on PdeCu/Pt using a spray method to make PdeCu/Pt/PdNP.
127
A 3-electrode battery was fabricated by immersing the as-made electrolysis anode, bifunctional electrode, and fuel cell cathode in a 5 M KOH solution. Three batteries were fabricated: AujKOHjPdeCujKOHjPt, AujKOHjPdeCu/PtjKOHjPt, and AujKOHjPdeCu/Pt/PdNPjKOHjPt. To charge the 3-electrode batteries, a 200 mA current was applied in the constant current mode and the voltage change was observed during charging. When the batteries were discharged, the alkaline fuel cell with the bifunctional electrode as its anode was operated with the injection of 50 ccm of pure oxygen. Unlike conventional fuel cell performance measurement methods, chronopotentiometry should be applied at a current of 20 mA. Cycling stability tests were conducted by repeating the charging and discharging cycles a total of 50 times. 3. Results & discussion 3.1. Electrolysis anode & fuel cell cathode An electrolysis anode was used to electrolyze KOH aqueous solutions where the current flows should not be obstructed by anything other than the limiting current generated by the bifunctional electrode. Therefore, this electrode should have a small resistance and sufficient surface area for electrolysis. Therefore, a mesh with a wider surface area than the metal plate was used as the electrode and the surface was deposited with gold to additionally reduce the electric resistance. Fig. 2(a) shows an SEM image of the electrolysis anode on which gold was evenly applied throughout the surface. The fuel cell cathode plays a role in the process of reducing oxygen in a KOH aqueous solution to generate OH ions. This electrode locates between the electrolyte and oxygen inlet, preventing electrolyte in the battery from flowing out through the oxygen inlet while, at the same time, passing on oxygen supplied through the oxygen inlet to the electrolyte well. The electrode satisfies the above properties because it was made by spraying Pt catalyst on GDL pre-coated carbon paper, which is waterproof. If this carbon paper is used as a wall between the electrolyte and oxygen inlet, only oxygen is allowed to penetrate into the electrolyte without any outflow of the electrolyte. Fig. 2(b) shows an SEM image of the fuel cell cathode, in which the catalyst was applied evenly throughout the electrode. 3.2. The structures and compositions of each bifunctional electrode Fig. 3 shows SEM images of the bifunctional electrodes (PdeCu, PdeCu/Pt, and PdeCu/Pt/PdNP) and PdNP. In the image shown in Fig. 3(a), it can be seen that PdeCu contains highly rough and dendritic structures. The voltage of 4 V, which was adopted to the electrode in the preparation step, lies in the hydrogen evolution reaction (HER) region in which hydrogen is actively generated. If electrodeposition is induced in this voltage region, structures with a high surface area will be made because of electrodeposition with dynamic hydrogen bubble templates [10]. One of the most general methods used to make materials with a high surface area is to synthesize materials using templates with small-structured spaces between them and removing the templates using a method such as etching. However, if electrodeposition with dynamic hydrogenbubble templates is used, not only can materials with a high surface area can be formed without making templates because the generated hydrogen plays the role of a template, but no additional process is necessary to remove the template because the hydrogen will be removed by natural mixing with air [12]. In addition, the diffusion limitation in solutions can be overcome because the generated hydrogen continuously plays the role of a stirrer while
128
C.-W. Roh et al. / Journal of Power Sources 280 (2015) 125e131
Fig. 2. SEM images of (a) an electrolysis anode and (b) a fuel cell cathode.
Fig. 3. SEM images of (a) PdeCu, (b) PdeCu/Pt, (c) PdNP, and (d) PdeCu/Pt/PdNP.
being attached and detached on the electrode [13]. If an alloy is formed with copper in addition to electrodeposition with dynamic hydrogen bubble templates, the electrode will grow into a dendritic structure because of the properties of copper. As a result, the electrode that is grown possesses a high surface area. We previously formed large surface area electrodes consisting of NieSn [11], Ag [14], Pb [15], AueCu [16], NiO [17], or RuO [18] using this method and applied the electrodes to diverse applications. Fig. 3(b) shows PdeCu/Pt in which Pt surrounded the dendritic branches of PdeCu. The PdeCu with a high surface area grown through electrodeposition with dynamic hydrogen bubble templates consist of palladium along with copper. If this electrode is immersed in a platinum displacement solution to displace the copper with platinum using galvanic displacement reactions, the electrode will have platinum layers on its surface. Despite the platinum placed on the surface, the high surface area is maintained. Fig. 3(c) shows a surface SEM image of PdNP and Fig. 3(d) shows an image of PdeCu/Pt/PdNP. The PdNP was made using galvanic displacement reactions. If an aluminum plate, which has a low standard reduction potential, is immersed in an acidic solution where metal ions, which have high standard reduction potentials, are dissolved, protons and metal ions will be reduced together on
the aluminum surface. The generated hydrogen will act as dynamic hydrogen-bubble templates and the reduced metal will have a high surface area [10]. Since the PdeCu/Pt/PdNP was made by spraying PdNP on PdeCu/Pt, the surfaces of PdNP and PdeCu/Pt/PdNP were identified to have the same forms, as expected. Table 1 shows the compositions of the individual electrodes determined by EDS and ICPeAES. Since the EDS analysis collects surface information more sensitively, relatively more information of the surfaces can be obtained, whereas overall information of the specimens can be obtained in the case of ICPeAES without distinction between the surface and the inside. Concerning PdeCu, as can be seen in the table, the composition determined from the
Table 1 EDS and ICPeAES data of each bifunctional electrode. Bifunctional electrodes
PdeCu PdeCu/Pt PdeCu/Pt/PdNP
EDS (wt.%)
ICPeAES (wt.%)
Pd
Cu
Pt
Pd
Cu
Pt
82.53 2.38 88.86
17.47 4.29 5.34
e 93.33 5.80
82.93 63.24 79.52
17.07 10.27 5.72
e 26.49 14.76
C.-W. Roh et al. / Journal of Power Sources 280 (2015) 125e131
EDS data is almost the same as that in the ICPeAES data. It can be concluded that the compositions of the surface and inside are not different. On the other hand, in the cases of PdeCu/Pt and PdeCu/ Pt/PdNP, there are differences in the EDS and ICPeAES data. Among them, the differences of the PdeCu/Pt data are larger. Whereas more than 90% platinum was detected by EDS, the platinum content was determined to be about 26.5% in the ICPeAES data. The results obtained by the two analysis methods support the fact that an electrode with Pd inside and a Pt surface was made since PdeCu was electrodeposited and Pt was grown on the surface by displacing Cu. In the case of PdeCu/Pt/PdNP, the ratio of palladium is higher in the EDS data, contrary to PdeCu/Pt. This indicates that the surface of PdeCu/Pt is covered by Pd powder. 3.3. Charge and discharge characteristics of individual batteries Fig. 4(a)e(c) displays the charge characteristics of AujKOHjPdeCujKOHjPt, AujKOHjPdeCu/PtjKOHjPt, and AujKOHjPdeCu/Pt/PdNPjKOHjPt, respectively, when they were charged by applying a constant current of 200 mA. The charging process of these batteries is electrolysis. During general electrolysis, water is decomposed and hydrogen gas is generated at the cathode. However, in these batteries, hydrogen atoms are stored at the bifunctional electrode in the form of a palladium hydride instead of becoming hydrogen gas. These three graphs show that the voltage increases as charging progresses and is then maintained at a certain level. This phenomenon occurs because the distances between the lattices increase as palladium is changed into palladium hydride. As a result, the electric resistance increases so that the level of voltage which is necessary to apply a constant current increases [8]. Therefore, the high charging voltage indicates that a large volume of hydrogen is stored. Fig. 4(a) shows the charge characteristics of the first and tenth charges of AujKOHjPdeCujKOHjPt. In the graph of the first charge, the curve increases twice, possibly because the hydrogen is stored as two different palladium-hydride phases: a a-phase and b-phase.
129
The a-phase palladium hydride has hydrogen atoms in the lattice of the palladium as solid solution and when the concentration of hydrogen in palladium increases, the a-phase palladium hydride starts to nucleate and becomes a b-phase palladium hydride [19]. Comparison of these two curves shows that the starting voltage of the tenth charge is higher than that of the first charge. This means that hydrogen is left in PdeCu when discharging is finished. In the case of AujKOHjPdeCujKOHjPt, since the starting voltage of the tenth charge is the same as the voltage when charging progressed for approximated 20 s in the first charge, the hydrogen amounting to the volume corresponding to charging for 20 s should remain when discharging finished, demonstrating that the hydrogen use efficiency is low. The charge characteristics of AujKOHjPdeCu/PtjKOHjPt are shown in Fig. 4(b). The first and tenth charge characteristics are almost the same. This means that only a very small amount of hydrogen is left in the bifunctional electrode. From this result, it can be seen that in the case of AujKOHjPdeCu/PtjKOHjPt, the hydrogen use efficiency is high during discharging. The charge characteristics of AujKOHjPdeCu/Pt/PdNPjKOHjPt are shown in Fig. 4(c). Similar to the first charge characteristics of AujKOHjPdeCujKOHjPt, there are a-phase and b-phase palladium hydride curves. Given that the charging voltage rises faster in the tenth charge characteristic graph, the charging can be regarded to progress faster than in the first charge. Careful observation reveals that there are wave-like graphs from 70 to 120 s in the first charging characteristic curve. This may result from activation of the electrode. During activation, the 3-phase boundary, which is the contact surface of Nafion, electrolyte, and metal electrode, is prepared and activated [20]. The discharge characteristics of the tenth discharges of the charged AujKOHjPdeCujKOHjPt, AujKOHjPdeCu/PtjKOHjPt, and AujKOHjPdeCu/Pt/PdNPjKOHjPt when the batteries were discharged at 20 mA are shown in Fig. 4(d). This battery discharging process represents the operation of fuel cells. When a bifunctional electrode is used in a fuel cell system, it is utilized as a fuel cell
Fig. 4. Graphs comparing the 1st and 10th charge characteristics of (a) AujKOHjPdeCujKOHjPt, (b) AujKOHjPdeCu/PtjKOHjPt, and (c) AujKOHjPdeCu/Pt/PdNPjKOHjPt, and the (d) 10th discharge of each battery.
130
C.-W. Roh et al. / Journal of Power Sources 280 (2015) 125e131
anode. Conventionally, at a fuel cell anode, hydrogen gas is decomposed to become hydrogen ions, discharging two electrons. Unlike conventional fuel cell anodes, at the bifunctional electrode, instead of the decomposition of hydrogen gas, the hydrogen in palladium hydride becomes a hydrogen ion and discharges one electron. Upon comparison of the voltage levels in the sections where discharge is stabilized first, the voltage level of AujKOHjPdeCu/PtjKOHjPt is higher than that of AujKOHjPdeCujKOHjPt or AujKOHjPdeCu/Pt/PdNPjKOHjPt. In the case of AujKOHjPdeCujKOHjPt and AujKOHjPdeCu/Pt/ PdNPjKOHjPt, the surface of the bifunctional electrode is covered by palladium and the surface of the bifunctional electrode of AujKOHjPdeCu/PtjKOHjPt is covered by platinum. Because palladium, which is on the surface of the bifunctional electrodes of AujKOHjPdeCujKOHjPt and AujKOHjPdeCu/Pt/PdNPjKOHjPt, has a lower hydrogen oxidation activity than platinum [21], the initial discharge voltage level of these batteries is low. On the other hand, AujKOHjPdeCu/PtjKOHjPt is installed with PdeCu/Pt where platinum exists on the surface where the voltage drop due to the hydrogen oxidation rate is less and as a result, the initial discharge voltage level is higher. Next, upon reviewing the section between the stabilization of discharge and voltage drop due to hydrogen shortage, it can be seen that, unlike AujKOHjPdeCu/PtjKOHjPt and AujKOHjPdeCu/Pt/ PdNPjKOHjPt, AujKOHjPdeCujKOHjPt shows remarkable voltage drops instead of maintaining the voltage. This phenomenon occurs because copper exists on the surface of the bifunctional electrode with palladium. When hydrogen oxidation reactions occur on the surface of the bifunctional electrode, Cu oxidation can also occur [22]. In this case, while being oxidized, the Cu can be oxidized to Cu(OH)x. As discharging progresses, the amount of copper oxidized on the surface increases and the copper hydroxide species simultaneously cover the surface of the electrode, reducing the surface area and eventually increasing the resistance of the electrode so that a voltage drop occur. This phenomenon can be observed in the graph of the discharge characteristics. On the contrary, on AujKOHjPdeCu/Pt/PdNPjKOHjPt installed with PdeCu/Pt/PdNP, of which the entire surface is composed of palladium, copper hydroxide species are not generated even if the reactions progress because copper does not exist on the surface. Therefore, a constant discharge voltage can be maintained. This phenomenon also occurs in AujKOHjPdeCu/PtjKOHjPt installed with PdeCu/Pt, of which the surface is covered by platinum. The amounts of electric power discharged per 1 g of palladium contained in the AujKOHjPdeCujKOHjPt, AujKOHjPdeCu/PtjKOHjPt, and AujKOHjPdeCu/Pt/PdNPjKOHjPt electrodes are 18.73 mWh, 31.70 mWh, and 24.40 mWh, respectively. To express the values as electrical charge densities (mAh), the electrical energy densities (mWh) are divided by the average discharge voltages of approximately 0.63 V, 0.75 V, and 0.68 V to result in electrical charge densities of 29.7 mAh, 42.3 mAh, and 35.9 mAh per 1 g of palladium, respectively. Theoretically, considering that hydrogen atoms can be stored up to 41.2 at % as palladium hydride under ambient conditions [8], the maximum electrical charge density is about 176.7 mAh per 1 g of palladium. Based on this theoretical value, the discharge efficiencies of three different bifunctional electrodes are 16.8%, 23.9%, and 20.3%, respectively. The bifunctional electrode should be in a form that can be used in both systems. First, when this electrode is evaluated from the viewpoint of fuel cell anodes, since oxidation reactions will occur on the surface of the catalyst when hydrogen atoms become hydrogen ions, the surface area of the catalyst should be large [23]. Second, when this electrode is evaluated from the viewpoint of hydrogen storage, since hydrogen penetrates into the lattices of metals for storage, hydrogen storage is greatly affected by not only
the surface area but also the total mass or total volume [8]. Therefore, for hydrogen storage, palladium, which is a hydrogen storage medium, should have a large mass or volume. Platinum was used as a hydrogen-oxidizing catalyst in this study and it should have a high surface area and be in contact with palladium over a large area. Conceptually, it was the most similar to PdeCu/Pt and showed the highest performance in AujKOHjPdeCu/PtjKOHjPt. Fig. 5 shows a conceptual diagram of the PdeCu/Pt electrode, which consists of two layers. One layer is an internal palladiumecopper alloy layer and the other is a surface platinum catalyst layer. As can be seen in Fig. 5, the platinum catalyst is located on the surface of the alloy layer. In this case, if the platinum catalyst layer formed on the alloy layer is thick, it may prevent hydrogen from being absorbed from the surface to the inside. As a result, the platinum catalyst layer should be formed with an appropriate thickness on the surface of the alloy layer. Since the hydrogen storage capacity of Cu is lower than that of palladium, if Cu can be removed from the alloy layer of the electrode, the entire battery could show better performance.
3.4. Cycling stability test The results of cycling stability tests of AujKOHjPdeCu/PtjKOHjPt and AujKOHjPdeCu/Pt/PdNPjKOHjPt used to demonstrate performance to some extent are shown in Fig. 6 as the amount of electrical energy discharged as a function of the number of cycles. From the cycling stability test graph for AujKOHjPdeCu/PtjKOHjPt, it can be seen that as the number of cycles increased, the amount of discharged electric power gradually increased. The distances between metal atoms increase due to the hydrogen atoms which are charged in the bifunctional electrode through repetitive charging and discharging. As a result, the spaces for hydrogen storage are widened such that larger amounts of hydrogen can be used. Therefore, the amount of discharged electric power gradually increases. Continuous increases of the amount of the discharged electric power of AujKOHjPdeCu/PtjKOHjPt were identified over 50 repeated charging/discharging cycles, indicating that AujKOHjPdeCu/PtjKOHjPt is stable. In the cycling stability test graph for AujKOHjPdeCu/Pt/ PdNPjKOHjPt, it can be seen that the amount of discharge electric power increased gradually, similar to AujKOHjPdeCu/PtjKOHjPt in the 5th-20th cycle region. However, in the initial 5 cycles, fluctuations including drastic increases and decreases of the energy density were obtained. The increasing power density is due to the effect of 3-phase boundary activation which eases the charging as well as discharging of hydrogen. After several cycles, PdNPs that are loosely contacted with Nafion seem to lose contact with the Nafion completely, resulting in a decrease of the electrical energy density in the 3rd-5th cycles. In particular, on the PdeCu/Pt/PdNP electrode, an additional PdNP layer is formed on top of the PdeCu/Pt electrode. Therefore, it requires more charging and discharging cycles to be stabilized by the activation. The amount of discharge electric power was maintained in the 21ste40th cycle region and decreased in the 40the50th cycle region. The reason why the amount of discharged electrical energy decreased over the 40the50th cycles is that since the volume of palladium is repeatedly increased and reduced due to the continued charging and discharging, the binding between Nafion, which plays the role of an adhesive, and PdNP is loosened and the palladium contained in the electrode is lost. In the 21ste40th cycle region, a constant amount of discharged electric power is assumed to be maintained because the increase of the amount of stored hydrogen due to repetitive charging and discharging and the loss of the palladium contained in the electrode act together.
C.-W. Roh et al. / Journal of Power Sources 280 (2015) 125e131
131
Fig. 5. Mimetic diagram of PdeCu/Pt, showing a platinum layer with a high surface area on a palladiumecopper alloy layer.
will become attractive devices in fabricating small volume energy storage systems. Acknowledgments This research was supported by the Basic Science Research Program through the National Research Foundation of Korea funded by the Ministry of Science, ICT and Future Planning (NRF2013R1A2A2A01014102) and the GRRC program of Gyeonggi province [(GRRCSKKU2014-B04), Fabrication of coreeshell structured metal powder for environmental friendly automobiles]. References Fig. 6. Cycling stability test results of AujKOHjPdeCu/PtjKOHjPt and AujKOHjPdeCu/ Pt/PdNPjKOHjPt.
4. Conclusion In this research, palladium hydride bifunctional electrodes that can be used as hydrogen storage electrodes as well as fuel cell anodes were made using electrodeposition with dynamic hydrogen-bubble templates and galvanic displacement reactions. In addition, 3-electrode batteries were fabricated to evaluate various electrochemical properties. The 3-electrode battery is a hybrid system that has two types of electrochemical systems: an electrolysis cell and a fuel cell. From the 3-electrode battery charge characteristics tests, the hydrogen use efficiency was the highest in AujKOHjPdeCu/PtjKOHjPt, which is a battery applied with an electrode made by forming a platinum catalyst layer on a palladiumecopper alloy layer. In addition, in the discharge characteristics tests, the same battery showed the best performance of 31.70 mWh/g Pd. Therefore, among the bifunctional electrodes made in this study, PdeCu/Pt was demonstrated to have the most ideal structure. In the cycling stability tests, AujKOHjPdeCu/ PtjKOHjPt was identified to be the most stable. In this study, bifunctional electrodes made by forming a platinum catalyst layer on palladium, which has a high surface area, were found to be effective in 3-electrode batteries. Future studies of the electrodes and batteries are required to demonstrate that they
[1] C. Bretthauer, C. Müller, H. Reinecke, Electrochim. Acta 54 (2009) 6094e6098. [2] A. Patil, V. Patil, D.-W. Shin, J.-W. Choi, D.-S. Paik, S.-J. Yoon, Mater. Res. Bull. 43 (2007) 1913e1942. [3] E. Slavcheva, G. Ganske, U. Schnakenberg, Sci. World J. 2014 (2014). Article ID 146126. [4] G. Erdler, M. Frank, M. Lehmann, H. Reinecke, C. Müller, Sensors Actuators A 132 (2006) 331e336. [5] M. Frank, G. Erdler, H.-P. Frerichs, C. Müller, H. Reinecke, J. Power Sources 181 (2008) 371e377. [6] C. Bretthauer, C. Müller, H. Reinecke, Electrochim. Acta 54 (2009) 6094e6098. [7] C. Bretthauer, C. Müller, H. Reinecke, J. Power Sources 196 (2011) 4729e4734. [8] F.D. Manchester, A. Sam-Martin, J.M. Pitre, J. Phase Equilib. 15 (1994) 62e83. [9] E.-W. Shin, S.-I. Cho, J.-H. Kang, W.-J. Kim, J.-D. Park, S.-H. Moon, Korean J. Chem. Eng. 17 (2000) 468e472. [10] K. Zhuo, M.-G. Jeong, C.-H. Chung, J. Power Sources 244 (2013) 601e605. [11] K. Zhuo, M.-G. Jeong, C.-H. Chung, RSC Adv. 3 (2013) 12611e12615. [12] Y. Li, W.-Z. Jia, Y.-Y. Song, X.-H. Xia, Chem. Mater. 19 (2007) 5758e5764. [13] N.D. Nikoli c, K.I. Popov, in: S.S. Djokic (Ed.), Modern Aspects of Electrochemistry, Springer, New York, 2010, pp. 1e70. [14] S. Cherevko, C.-H. Chung, Electrochim. Acta 55 (2010) 6383e6390. [15] S. Cherevko, X. Xing, C.-H. Chung, Appl. Surf. Sci. 257 (2011) 8054e8061. [16] S. Cherevko, N. Kulyk, C.-H. Chung, Langmuir 28 (2012) 3306e3315. [17] M.-G. Jeong, K. Zhuo, S. Cherevko, C.-H. Chung, Korean J. Chem. Eng. 29 (2012) 1802e1805. [18] M.-G. Jeong, K. Zhuo, S. Cherevko, W.-J. Kim, C.-H. Chung, J. Power Sources 244 (2013) 806e811. [19] M. Hirscher (Ed.), Handbook of Hydrogen Storage, WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim, 2010, pp. 279e340. [20] S.J. Lee, S. Mukerjee, J. McBreen, Y.W. Rho, Y.T. Kho, T.H. Lee, Electrochim. Acta 43 (1998) 3693e3701. [21] J.K. Nørskov, J. Rossmeisl, A. Logadottir, L. Lindqvist, J. Phys. Chem. B 108 (2014) 17886e17892. [22] L.-P. Xia, P. Guo, Y. Wang, S.-Q. Ding, J.-B. He, J. Power Sources 262 (2014) 232e238. [23] H.A. Gasteiger, N.M. Markovi c, Science 324 (2009) 48e49.
Contents lists available at ScienceDirect
Journal of Power Sources journal homepage: www.elsevier.com/locate/jpowsour
Fabrication of a three-electrode battery using hydrogen-storage materials Chi-Woo Roh, Jung-Yong Seo, Hyung-Seok Moon, Hyun-Young Park, Na-Yun Nam, Sung Min Cho, Pil J. Yoo, Chan-Hwa Chung* School of Chemical Engineering, Sungkyunkwan University, Suwon 440-746, Republic of Korea
h i g h l i g h t s
g r a p h i c a l a b s t r a c t
We design new-concept three-electrode battery using hydrogen as energy carrier. Charging process is electrolysis of alkaline electrolyte. Discharging process is operation of alkaline fuel cell. Hydrogen is stored at bifunctional electrode as metal hydride with charging process. Bifunctional electrode acts as electrolysis cathode and fuel cell anode at same time.
a r t i c l e i n f o
a b s t r a c t
Article history: Received 28 October 2014 Received in revised form 26 December 2014 Accepted 16 January 2015 Available online 16 January 2015
In this study, an energy storage device using a three-electrode battery is fabricated. The charging process takes place during electrolysis of the alkaline electrolyte where hydrogen is stored at the palladium bifunctional electrode. Upon discharging, power is generated by operating the alkaline fuel cell using hydrogen which is accumulated in the palladium hydride bifunctional electrode during the charging process. The bifunctional palladium electrode is prepared by electrodeposition using a hydrogen bubble template followed by a galvanic displacement reaction of platinum in order to functionalize the electrode to work not only as a hydrogen storage material but also as an anode in a fuel cell. This bifunctional electrode has a sufficiently high surface area and the platinum catalyst populates at the surface of electrode to operate the fuel cell. The charging and discharging performance of the three-electrode battery are characterized. In addition, the cycle stability is investigated. © 2015 Elsevier B.V. All rights reserved.
Keywords: Bifunctional electrode Palladium hydride Dynamic hydrogen-bubble template Dendrite structures Three-electrode battery
1. Introduction Recently, along with high-edge technologies for portable electric devices, energy storage systems with a high volumetric energy density are highly demanded. Currently, electric energy is generally
* Corresponding author. E-mail address: [email protected] (C.-H. Chung). http://dx.doi.org/10.1016/j.jpowsour.2015.01.101 0378-7753/© 2015 Elsevier B.V. All rights reserved.
stored in secondary batteries or capacitors. Whereas capacitors have lower volumetric energy densities than secondary batteries as well as somewhat unsuitable discharge characteristics for portable electronic devices, secondary batteries possess slightly more suitable discharge characteristics and higher volumetric energy densities [1]. Therefore, secondary batteries are more advantageous than capacitors for configuring small scale energy storage systems. Among secondary batteries, Li-ion batteries are currently the most commonly used for energy storage applications. However, in
126
C.-W. Roh et al. / Journal of Power Sources 280 (2015) 125e131
the case of thin film lithium ion batteries which have a small volume, it has been found that there is a stability issue which limits increasing the film thickness because of the increased channel distance through which Li ions come in and out. As a result, it is hard to form a structure thicker than a few mm [2]. On the other hand, in the case of metal hydride, hydrogen atoms penetrate into the metal lattice structure, which is much larger than the size of the hydrogen atoms being stored. Therefore, the channels for hydrogen atoms are not plugged so that there are fewer limitations concerning the electrode thickness. In addition, metal hydride can theoretically have a volumetric energy density of 2000 Wh/l or higher, which is larger than that of thin film Li-ion batteries [1]. Metal hydrides can be utilized in fuel cell (FC) type secondary batteries. In this FC-type battery, the oxidation of hydrogen stored as a hydride, and the reduction of oxygen occur in an alkaline electrolyte in the discharging process. In this case, an active type fuel cell, into which pure oxygen is injected, as well as a passive type fuel cell, to which oxygen in air is supplied through a breathing process, can be formed for discharging. This FC-type secondary battery which employs a metal hydride has many advantages. The fuels are only hydrogen and oxygen which are nontoxic and completely environmentally friendly. In addition, the electrolyte is
an aqueous alkaline solution instead of an explosive organic solution. Also, it charges faster or requires no charging time, has a longer service life as well as a lack of self-discharge [3]. As a result, FC-type secondary batteries are considered as a good alternative secondary battery and diverse studies of these battery systems have been carried out. For example, G. Erdler and C. Bretthauser [4e7] et al. designed chip integrated FC-type secondary batteries using polymer electrolyte membranes. In this study, based on these recent technical trends, we designed a 3-electrode battery, as shown in Fig. 1, and adopted palladium as hydrogen storage materials among the many potential hydrogen storage metals and alloys. Since the present study focuses on the performance of batteries in relation to the structure of the metal hydride electrodes, palladium, which is easy to handle and possesses excellent ability for hydrogen storage at ambient condition [8,9], is a suitable material. The batteries designed in the present study are in the form of three electrodes immersed in an alkali electrolyte solution, as shown in Fig. 1. The electrodes are an electrolysis anode, a bifunctional electrode, and a fuel cell cathode. Fig. 1(a) shows a mimetic diagram of the operation of the charging process used to store hydrogen in the bifunctional electrode. In this hydrogen charging process, protons meet equivalent electrons at the bifunctional electrode and are stored as palladium hydride (PdHx) while water is electrolyzed and O2 is evolved at the electrolysis anode. The chemical reactions of this process are as follows.
c c Anode ðelectrolysis anodeÞ : cOH / H2 O þ O2 þ ce 2 4 (1) Cathode ðbifunctional electrodeÞ : Pd þ cH2 O þ ce / PdHc þ cOH
Overall reaction : Pd þ
c c H O / O2 þ PdHc 2 2 4
(2)
(3)
Fig. 1(b) shows the discharging process which utilizes the oxidation reaction of the hydrogen stored in the bifunctional electrode. In this hydrogen discharging process, the fuel cell is operated and each hydrogen atom stored in the form of palladium hydride meets an OH ion to form a water molecule and discharge one electron. As a result, an electric current flows in this process. The process can be expressed by the following chemical formulas.
Anode ðbifunctional electrodeÞ : cOH þ PdHc /cH2 O þ Pd þ ce
Cathode ðfuel cell cathodeÞ :
Overall reaction : PdHc þ
Fig. 1. Schematic diagram of the (a) charging and (b) discharging processes of the 3electrode battery. In the charging process, protons are stored as palladium hydride (PdHx) in the bifunctional electrode. In the discharging process, hydrogen atoms stored in the form of PdHx meet OH ions to become water and discharge electrons.
(4)
c c H O þ O þ ce / cOH 2 2 4 2 (5)
c c O / H2 O þ Pd 4 2 2
(6)
The characteristics of the individual electrodes of the 3electrode batteries developed using this concept were analyzed by SEM (scanning electron microscope), EDS (energy-dispersive Xray spectroscopy), and ICPeAES (inductively coupled plasmaeatomic emission spectroscopy). In addition, the chargeedischarge characteristics and cycling stability of the fabricated 3-electrode batteries were analyzed through electrochemical experiments.
C.-W. Roh et al. / Journal of Power Sources 280 (2015) 125e131
2. Experimental The electrolysis anode was made to possess a small resistance by depositing gold on 100 mesh size Ni mesh (Nilaco, Japan). Before being deposited with gold, the Ni mesh was pretreated as follows to facilitate the gold deposition. First, the Ni mesh was treated with acetone, iso-propyl alcohol (IPA), and ethanol at 50 C for 20 min each to remove organic matter. Second, the Ni mesh was treated at 50 C with 6 M HCl for 30 min to remove the oxide film. The above process was implemented in an ultrasonic cleaner (SAEHAN, Republic of Korea). The gold was deposited via electroless deposition. The electroless deposition solution was prepared as follows. A 100 mL/L GoBright® TWX-40-M10 (C. Uyemura, Japan) solution, 12 mL/L gold solution (Au 100 g/L, C. Uyemura, Japan), and 0.5 mL/L potassium cyanide solution (KCN 100 g/L, C. Uyemura, Japan) were mixed. The deposition process then proceeded as follows. First, the electroless deposition solution was added to a thermostatic tank and the temperature was fixed at 60 C using a water bath. Then, the pretreated Ni mesh was immersed in the solution for 40 min so that the entire surface was sufficiently deposited with gold. We made sure that the Ni mesh was completely immersed in the solution such that no part of the Ni mesh was in contact with air. When the entire surface was evenly deposited with gold, the Ni mesh was washed with D.I. water and dried in a vacuum dryer. The fuel cell cathode was prepared by applying 5 mg/cm2 Pt black catalyst (Alfa Aesar, UK) on GDL pre-coated carbon paper (CNL Energy, Republic of Korea) using a spray method. In this case, the catalyst ink contained catalyst and Nafion contents of 85 wt.% and 15 wt.%, respectively, in the solvent (D.I. water:IPA ¼ 1:4, volume ratio) while they were dispersed for at least one hour in an ultrasonicator. As mentioned above, the bifunctional electrode plays the role of both an electrolysis cathode and a fuel cell anode. To determine the effect of the structure of this electrode on the performance of the battery, three electrodes, PdeCu, PdeCu/Pt, and PdeCu/Pt/PdNP, were prepared. PdeCu was made by placing the pretreated Ni mesh as a working electrode and a platinum plate as a counter electrode into an electrolyte composed of 5 mM CuSO4, 15 mM PdCl2, and 1 M H2SO4 while inducing electrodeposition for five minutes at a very high voltage of 4 V without any reference electrode. The voltage of 4 V lies in the hydrogen evolution reaction (HER) region where hydrogen is actively generated. When electrodeposition is induced in this voltage region, structures with a high surface area are made because of dynamic hydrogen bubble templates [10]. The PdeCu/Pt electrode was made by immersing the PdeCu electrode fabricated using the above mentioned deposition process in a platinum displacement solution composed of 5 mM H2PtCl6 and 0.1 M HClO4 for 90 min to displace the copper with platinum through galvanic displacement reactions, thereby forming platinum on the surface. The PdeCu/Pt/PdNP electrode was made by additionally applying palladium nano-powder (PdNP) on PdeCu/Pt. PdNP was made by utilizing galvanic displacement reactions to take advantage of the reduction potential differences between aluminum and palladium. The aluminum plate was immersed in 1 M NaOH for five minutes to remove the oxide layer followed by soaking in a solution composed of 10 mM PdCl2 and 1 M H2SO4. Then, PdNP with a high surface area was obtained with hydrogen generation [11]. Then, a mixture of the obtained PdNP and Nafion was prepared at a PdNP:Nafion weight ratio of 9:1 in the solvent (D.I. water:IPA ¼ 1:4, volume ratio) where it was evenly dispersed for at least one hour in an ultrasonicator. Then, the obtained ink was applied on PdeCu/Pt using a spray method to make PdeCu/Pt/PdNP.
127
A 3-electrode battery was fabricated by immersing the as-made electrolysis anode, bifunctional electrode, and fuel cell cathode in a 5 M KOH solution. Three batteries were fabricated: AujKOHjPdeCujKOHjPt, AujKOHjPdeCu/PtjKOHjPt, and AujKOHjPdeCu/Pt/PdNPjKOHjPt. To charge the 3-electrode batteries, a 200 mA current was applied in the constant current mode and the voltage change was observed during charging. When the batteries were discharged, the alkaline fuel cell with the bifunctional electrode as its anode was operated with the injection of 50 ccm of pure oxygen. Unlike conventional fuel cell performance measurement methods, chronopotentiometry should be applied at a current of 20 mA. Cycling stability tests were conducted by repeating the charging and discharging cycles a total of 50 times. 3. Results & discussion 3.1. Electrolysis anode & fuel cell cathode An electrolysis anode was used to electrolyze KOH aqueous solutions where the current flows should not be obstructed by anything other than the limiting current generated by the bifunctional electrode. Therefore, this electrode should have a small resistance and sufficient surface area for electrolysis. Therefore, a mesh with a wider surface area than the metal plate was used as the electrode and the surface was deposited with gold to additionally reduce the electric resistance. Fig. 2(a) shows an SEM image of the electrolysis anode on which gold was evenly applied throughout the surface. The fuel cell cathode plays a role in the process of reducing oxygen in a KOH aqueous solution to generate OH ions. This electrode locates between the electrolyte and oxygen inlet, preventing electrolyte in the battery from flowing out through the oxygen inlet while, at the same time, passing on oxygen supplied through the oxygen inlet to the electrolyte well. The electrode satisfies the above properties because it was made by spraying Pt catalyst on GDL pre-coated carbon paper, which is waterproof. If this carbon paper is used as a wall between the electrolyte and oxygen inlet, only oxygen is allowed to penetrate into the electrolyte without any outflow of the electrolyte. Fig. 2(b) shows an SEM image of the fuel cell cathode, in which the catalyst was applied evenly throughout the electrode. 3.2. The structures and compositions of each bifunctional electrode Fig. 3 shows SEM images of the bifunctional electrodes (PdeCu, PdeCu/Pt, and PdeCu/Pt/PdNP) and PdNP. In the image shown in Fig. 3(a), it can be seen that PdeCu contains highly rough and dendritic structures. The voltage of 4 V, which was adopted to the electrode in the preparation step, lies in the hydrogen evolution reaction (HER) region in which hydrogen is actively generated. If electrodeposition is induced in this voltage region, structures with a high surface area will be made because of electrodeposition with dynamic hydrogen bubble templates [10]. One of the most general methods used to make materials with a high surface area is to synthesize materials using templates with small-structured spaces between them and removing the templates using a method such as etching. However, if electrodeposition with dynamic hydrogenbubble templates is used, not only can materials with a high surface area can be formed without making templates because the generated hydrogen plays the role of a template, but no additional process is necessary to remove the template because the hydrogen will be removed by natural mixing with air [12]. In addition, the diffusion limitation in solutions can be overcome because the generated hydrogen continuously plays the role of a stirrer while
128
C.-W. Roh et al. / Journal of Power Sources 280 (2015) 125e131
Fig. 2. SEM images of (a) an electrolysis anode and (b) a fuel cell cathode.
Fig. 3. SEM images of (a) PdeCu, (b) PdeCu/Pt, (c) PdNP, and (d) PdeCu/Pt/PdNP.
being attached and detached on the electrode [13]. If an alloy is formed with copper in addition to electrodeposition with dynamic hydrogen bubble templates, the electrode will grow into a dendritic structure because of the properties of copper. As a result, the electrode that is grown possesses a high surface area. We previously formed large surface area electrodes consisting of NieSn [11], Ag [14], Pb [15], AueCu [16], NiO [17], or RuO [18] using this method and applied the electrodes to diverse applications. Fig. 3(b) shows PdeCu/Pt in which Pt surrounded the dendritic branches of PdeCu. The PdeCu with a high surface area grown through electrodeposition with dynamic hydrogen bubble templates consist of palladium along with copper. If this electrode is immersed in a platinum displacement solution to displace the copper with platinum using galvanic displacement reactions, the electrode will have platinum layers on its surface. Despite the platinum placed on the surface, the high surface area is maintained. Fig. 3(c) shows a surface SEM image of PdNP and Fig. 3(d) shows an image of PdeCu/Pt/PdNP. The PdNP was made using galvanic displacement reactions. If an aluminum plate, which has a low standard reduction potential, is immersed in an acidic solution where metal ions, which have high standard reduction potentials, are dissolved, protons and metal ions will be reduced together on
the aluminum surface. The generated hydrogen will act as dynamic hydrogen-bubble templates and the reduced metal will have a high surface area [10]. Since the PdeCu/Pt/PdNP was made by spraying PdNP on PdeCu/Pt, the surfaces of PdNP and PdeCu/Pt/PdNP were identified to have the same forms, as expected. Table 1 shows the compositions of the individual electrodes determined by EDS and ICPeAES. Since the EDS analysis collects surface information more sensitively, relatively more information of the surfaces can be obtained, whereas overall information of the specimens can be obtained in the case of ICPeAES without distinction between the surface and the inside. Concerning PdeCu, as can be seen in the table, the composition determined from the
Table 1 EDS and ICPeAES data of each bifunctional electrode. Bifunctional electrodes
PdeCu PdeCu/Pt PdeCu/Pt/PdNP
EDS (wt.%)
ICPeAES (wt.%)
Pd
Cu
Pt
Pd
Cu
Pt
82.53 2.38 88.86
17.47 4.29 5.34
e 93.33 5.80
82.93 63.24 79.52
17.07 10.27 5.72
e 26.49 14.76
C.-W. Roh et al. / Journal of Power Sources 280 (2015) 125e131
EDS data is almost the same as that in the ICPeAES data. It can be concluded that the compositions of the surface and inside are not different. On the other hand, in the cases of PdeCu/Pt and PdeCu/ Pt/PdNP, there are differences in the EDS and ICPeAES data. Among them, the differences of the PdeCu/Pt data are larger. Whereas more than 90% platinum was detected by EDS, the platinum content was determined to be about 26.5% in the ICPeAES data. The results obtained by the two analysis methods support the fact that an electrode with Pd inside and a Pt surface was made since PdeCu was electrodeposited and Pt was grown on the surface by displacing Cu. In the case of PdeCu/Pt/PdNP, the ratio of palladium is higher in the EDS data, contrary to PdeCu/Pt. This indicates that the surface of PdeCu/Pt is covered by Pd powder. 3.3. Charge and discharge characteristics of individual batteries Fig. 4(a)e(c) displays the charge characteristics of AujKOHjPdeCujKOHjPt, AujKOHjPdeCu/PtjKOHjPt, and AujKOHjPdeCu/Pt/PdNPjKOHjPt, respectively, when they were charged by applying a constant current of 200 mA. The charging process of these batteries is electrolysis. During general electrolysis, water is decomposed and hydrogen gas is generated at the cathode. However, in these batteries, hydrogen atoms are stored at the bifunctional electrode in the form of a palladium hydride instead of becoming hydrogen gas. These three graphs show that the voltage increases as charging progresses and is then maintained at a certain level. This phenomenon occurs because the distances between the lattices increase as palladium is changed into palladium hydride. As a result, the electric resistance increases so that the level of voltage which is necessary to apply a constant current increases [8]. Therefore, the high charging voltage indicates that a large volume of hydrogen is stored. Fig. 4(a) shows the charge characteristics of the first and tenth charges of AujKOHjPdeCujKOHjPt. In the graph of the first charge, the curve increases twice, possibly because the hydrogen is stored as two different palladium-hydride phases: a a-phase and b-phase.
129
The a-phase palladium hydride has hydrogen atoms in the lattice of the palladium as solid solution and when the concentration of hydrogen in palladium increases, the a-phase palladium hydride starts to nucleate and becomes a b-phase palladium hydride [19]. Comparison of these two curves shows that the starting voltage of the tenth charge is higher than that of the first charge. This means that hydrogen is left in PdeCu when discharging is finished. In the case of AujKOHjPdeCujKOHjPt, since the starting voltage of the tenth charge is the same as the voltage when charging progressed for approximated 20 s in the first charge, the hydrogen amounting to the volume corresponding to charging for 20 s should remain when discharging finished, demonstrating that the hydrogen use efficiency is low. The charge characteristics of AujKOHjPdeCu/PtjKOHjPt are shown in Fig. 4(b). The first and tenth charge characteristics are almost the same. This means that only a very small amount of hydrogen is left in the bifunctional electrode. From this result, it can be seen that in the case of AujKOHjPdeCu/PtjKOHjPt, the hydrogen use efficiency is high during discharging. The charge characteristics of AujKOHjPdeCu/Pt/PdNPjKOHjPt are shown in Fig. 4(c). Similar to the first charge characteristics of AujKOHjPdeCujKOHjPt, there are a-phase and b-phase palladium hydride curves. Given that the charging voltage rises faster in the tenth charge characteristic graph, the charging can be regarded to progress faster than in the first charge. Careful observation reveals that there are wave-like graphs from 70 to 120 s in the first charging characteristic curve. This may result from activation of the electrode. During activation, the 3-phase boundary, which is the contact surface of Nafion, electrolyte, and metal electrode, is prepared and activated [20]. The discharge characteristics of the tenth discharges of the charged AujKOHjPdeCujKOHjPt, AujKOHjPdeCu/PtjKOHjPt, and AujKOHjPdeCu/Pt/PdNPjKOHjPt when the batteries were discharged at 20 mA are shown in Fig. 4(d). This battery discharging process represents the operation of fuel cells. When a bifunctional electrode is used in a fuel cell system, it is utilized as a fuel cell
Fig. 4. Graphs comparing the 1st and 10th charge characteristics of (a) AujKOHjPdeCujKOHjPt, (b) AujKOHjPdeCu/PtjKOHjPt, and (c) AujKOHjPdeCu/Pt/PdNPjKOHjPt, and the (d) 10th discharge of each battery.
130
C.-W. Roh et al. / Journal of Power Sources 280 (2015) 125e131
anode. Conventionally, at a fuel cell anode, hydrogen gas is decomposed to become hydrogen ions, discharging two electrons. Unlike conventional fuel cell anodes, at the bifunctional electrode, instead of the decomposition of hydrogen gas, the hydrogen in palladium hydride becomes a hydrogen ion and discharges one electron. Upon comparison of the voltage levels in the sections where discharge is stabilized first, the voltage level of AujKOHjPdeCu/PtjKOHjPt is higher than that of AujKOHjPdeCujKOHjPt or AujKOHjPdeCu/Pt/PdNPjKOHjPt. In the case of AujKOHjPdeCujKOHjPt and AujKOHjPdeCu/Pt/ PdNPjKOHjPt, the surface of the bifunctional electrode is covered by palladium and the surface of the bifunctional electrode of AujKOHjPdeCu/PtjKOHjPt is covered by platinum. Because palladium, which is on the surface of the bifunctional electrodes of AujKOHjPdeCujKOHjPt and AujKOHjPdeCu/Pt/PdNPjKOHjPt, has a lower hydrogen oxidation activity than platinum [21], the initial discharge voltage level of these batteries is low. On the other hand, AujKOHjPdeCu/PtjKOHjPt is installed with PdeCu/Pt where platinum exists on the surface where the voltage drop due to the hydrogen oxidation rate is less and as a result, the initial discharge voltage level is higher. Next, upon reviewing the section between the stabilization of discharge and voltage drop due to hydrogen shortage, it can be seen that, unlike AujKOHjPdeCu/PtjKOHjPt and AujKOHjPdeCu/Pt/ PdNPjKOHjPt, AujKOHjPdeCujKOHjPt shows remarkable voltage drops instead of maintaining the voltage. This phenomenon occurs because copper exists on the surface of the bifunctional electrode with palladium. When hydrogen oxidation reactions occur on the surface of the bifunctional electrode, Cu oxidation can also occur [22]. In this case, while being oxidized, the Cu can be oxidized to Cu(OH)x. As discharging progresses, the amount of copper oxidized on the surface increases and the copper hydroxide species simultaneously cover the surface of the electrode, reducing the surface area and eventually increasing the resistance of the electrode so that a voltage drop occur. This phenomenon can be observed in the graph of the discharge characteristics. On the contrary, on AujKOHjPdeCu/Pt/PdNPjKOHjPt installed with PdeCu/Pt/PdNP, of which the entire surface is composed of palladium, copper hydroxide species are not generated even if the reactions progress because copper does not exist on the surface. Therefore, a constant discharge voltage can be maintained. This phenomenon also occurs in AujKOHjPdeCu/PtjKOHjPt installed with PdeCu/Pt, of which the surface is covered by platinum. The amounts of electric power discharged per 1 g of palladium contained in the AujKOHjPdeCujKOHjPt, AujKOHjPdeCu/PtjKOHjPt, and AujKOHjPdeCu/Pt/PdNPjKOHjPt electrodes are 18.73 mWh, 31.70 mWh, and 24.40 mWh, respectively. To express the values as electrical charge densities (mAh), the electrical energy densities (mWh) are divided by the average discharge voltages of approximately 0.63 V, 0.75 V, and 0.68 V to result in electrical charge densities of 29.7 mAh, 42.3 mAh, and 35.9 mAh per 1 g of palladium, respectively. Theoretically, considering that hydrogen atoms can be stored up to 41.2 at % as palladium hydride under ambient conditions [8], the maximum electrical charge density is about 176.7 mAh per 1 g of palladium. Based on this theoretical value, the discharge efficiencies of three different bifunctional electrodes are 16.8%, 23.9%, and 20.3%, respectively. The bifunctional electrode should be in a form that can be used in both systems. First, when this electrode is evaluated from the viewpoint of fuel cell anodes, since oxidation reactions will occur on the surface of the catalyst when hydrogen atoms become hydrogen ions, the surface area of the catalyst should be large [23]. Second, when this electrode is evaluated from the viewpoint of hydrogen storage, since hydrogen penetrates into the lattices of metals for storage, hydrogen storage is greatly affected by not only
the surface area but also the total mass or total volume [8]. Therefore, for hydrogen storage, palladium, which is a hydrogen storage medium, should have a large mass or volume. Platinum was used as a hydrogen-oxidizing catalyst in this study and it should have a high surface area and be in contact with palladium over a large area. Conceptually, it was the most similar to PdeCu/Pt and showed the highest performance in AujKOHjPdeCu/PtjKOHjPt. Fig. 5 shows a conceptual diagram of the PdeCu/Pt electrode, which consists of two layers. One layer is an internal palladiumecopper alloy layer and the other is a surface platinum catalyst layer. As can be seen in Fig. 5, the platinum catalyst is located on the surface of the alloy layer. In this case, if the platinum catalyst layer formed on the alloy layer is thick, it may prevent hydrogen from being absorbed from the surface to the inside. As a result, the platinum catalyst layer should be formed with an appropriate thickness on the surface of the alloy layer. Since the hydrogen storage capacity of Cu is lower than that of palladium, if Cu can be removed from the alloy layer of the electrode, the entire battery could show better performance.
3.4. Cycling stability test The results of cycling stability tests of AujKOHjPdeCu/PtjKOHjPt and AujKOHjPdeCu/Pt/PdNPjKOHjPt used to demonstrate performance to some extent are shown in Fig. 6 as the amount of electrical energy discharged as a function of the number of cycles. From the cycling stability test graph for AujKOHjPdeCu/PtjKOHjPt, it can be seen that as the number of cycles increased, the amount of discharged electric power gradually increased. The distances between metal atoms increase due to the hydrogen atoms which are charged in the bifunctional electrode through repetitive charging and discharging. As a result, the spaces for hydrogen storage are widened such that larger amounts of hydrogen can be used. Therefore, the amount of discharged electric power gradually increases. Continuous increases of the amount of the discharged electric power of AujKOHjPdeCu/PtjKOHjPt were identified over 50 repeated charging/discharging cycles, indicating that AujKOHjPdeCu/PtjKOHjPt is stable. In the cycling stability test graph for AujKOHjPdeCu/Pt/ PdNPjKOHjPt, it can be seen that the amount of discharge electric power increased gradually, similar to AujKOHjPdeCu/PtjKOHjPt in the 5th-20th cycle region. However, in the initial 5 cycles, fluctuations including drastic increases and decreases of the energy density were obtained. The increasing power density is due to the effect of 3-phase boundary activation which eases the charging as well as discharging of hydrogen. After several cycles, PdNPs that are loosely contacted with Nafion seem to lose contact with the Nafion completely, resulting in a decrease of the electrical energy density in the 3rd-5th cycles. In particular, on the PdeCu/Pt/PdNP electrode, an additional PdNP layer is formed on top of the PdeCu/Pt electrode. Therefore, it requires more charging and discharging cycles to be stabilized by the activation. The amount of discharge electric power was maintained in the 21ste40th cycle region and decreased in the 40the50th cycle region. The reason why the amount of discharged electrical energy decreased over the 40the50th cycles is that since the volume of palladium is repeatedly increased and reduced due to the continued charging and discharging, the binding between Nafion, which plays the role of an adhesive, and PdNP is loosened and the palladium contained in the electrode is lost. In the 21ste40th cycle region, a constant amount of discharged electric power is assumed to be maintained because the increase of the amount of stored hydrogen due to repetitive charging and discharging and the loss of the palladium contained in the electrode act together.
C.-W. Roh et al. / Journal of Power Sources 280 (2015) 125e131
131
Fig. 5. Mimetic diagram of PdeCu/Pt, showing a platinum layer with a high surface area on a palladiumecopper alloy layer.
will become attractive devices in fabricating small volume energy storage systems. Acknowledgments This research was supported by the Basic Science Research Program through the National Research Foundation of Korea funded by the Ministry of Science, ICT and Future Planning (NRF2013R1A2A2A01014102) and the GRRC program of Gyeonggi province [(GRRCSKKU2014-B04), Fabrication of coreeshell structured metal powder for environmental friendly automobiles]. References Fig. 6. Cycling stability test results of AujKOHjPdeCu/PtjKOHjPt and AujKOHjPdeCu/ Pt/PdNPjKOHjPt.
4. Conclusion In this research, palladium hydride bifunctional electrodes that can be used as hydrogen storage electrodes as well as fuel cell anodes were made using electrodeposition with dynamic hydrogen-bubble templates and galvanic displacement reactions. In addition, 3-electrode batteries were fabricated to evaluate various electrochemical properties. The 3-electrode battery is a hybrid system that has two types of electrochemical systems: an electrolysis cell and a fuel cell. From the 3-electrode battery charge characteristics tests, the hydrogen use efficiency was the highest in AujKOHjPdeCu/PtjKOHjPt, which is a battery applied with an electrode made by forming a platinum catalyst layer on a palladiumecopper alloy layer. In addition, in the discharge characteristics tests, the same battery showed the best performance of 31.70 mWh/g Pd. Therefore, among the bifunctional electrodes made in this study, PdeCu/Pt was demonstrated to have the most ideal structure. In the cycling stability tests, AujKOHjPdeCu/ PtjKOHjPt was identified to be the most stable. In this study, bifunctional electrodes made by forming a platinum catalyst layer on palladium, which has a high surface area, were found to be effective in 3-electrode batteries. Future studies of the electrodes and batteries are required to demonstrate that they
[1] C. Bretthauer, C. Müller, H. Reinecke, Electrochim. Acta 54 (2009) 6094e6098. [2] A. Patil, V. Patil, D.-W. Shin, J.-W. Choi, D.-S. Paik, S.-J. Yoon, Mater. Res. Bull. 43 (2007) 1913e1942. [3] E. Slavcheva, G. Ganske, U. Schnakenberg, Sci. World J. 2014 (2014). Article ID 146126. [4] G. Erdler, M. Frank, M. Lehmann, H. Reinecke, C. Müller, Sensors Actuators A 132 (2006) 331e336. [5] M. Frank, G. Erdler, H.-P. Frerichs, C. Müller, H. Reinecke, J. Power Sources 181 (2008) 371e377. [6] C. Bretthauer, C. Müller, H. Reinecke, Electrochim. Acta 54 (2009) 6094e6098. [7] C. Bretthauer, C. Müller, H. Reinecke, J. Power Sources 196 (2011) 4729e4734. [8] F.D. Manchester, A. Sam-Martin, J.M. Pitre, J. Phase Equilib. 15 (1994) 62e83. [9] E.-W. Shin, S.-I. Cho, J.-H. Kang, W.-J. Kim, J.-D. Park, S.-H. Moon, Korean J. Chem. Eng. 17 (2000) 468e472. [10] K. Zhuo, M.-G. Jeong, C.-H. Chung, J. Power Sources 244 (2013) 601e605. [11] K. Zhuo, M.-G. Jeong, C.-H. Chung, RSC Adv. 3 (2013) 12611e12615. [12] Y. Li, W.-Z. Jia, Y.-Y. Song, X.-H. Xia, Chem. Mater. 19 (2007) 5758e5764. [13] N.D. Nikoli c, K.I. Popov, in: S.S. Djokic (Ed.), Modern Aspects of Electrochemistry, Springer, New York, 2010, pp. 1e70. [14] S. Cherevko, C.-H. Chung, Electrochim. Acta 55 (2010) 6383e6390. [15] S. Cherevko, X. Xing, C.-H. Chung, Appl. Surf. Sci. 257 (2011) 8054e8061. [16] S. Cherevko, N. Kulyk, C.-H. Chung, Langmuir 28 (2012) 3306e3315. [17] M.-G. Jeong, K. Zhuo, S. Cherevko, C.-H. Chung, Korean J. Chem. Eng. 29 (2012) 1802e1805. [18] M.-G. Jeong, K. Zhuo, S. Cherevko, W.-J. Kim, C.-H. Chung, J. Power Sources 244 (2013) 806e811. [19] M. Hirscher (Ed.), Handbook of Hydrogen Storage, WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim, 2010, pp. 279e340. [20] S.J. Lee, S. Mukerjee, J. McBreen, Y.W. Rho, Y.T. Kho, T.H. Lee, Electrochim. Acta 43 (1998) 3693e3701. [21] J.K. Nørskov, J. Rossmeisl, A. Logadottir, L. Lindqvist, J. Phys. Chem. B 108 (2014) 17886e17892. [22] L.-P. Xia, P. Guo, Y. Wang, S.-Q. Ding, J.-B. He, J. Power Sources 262 (2014) 232e238. [23] H.A. Gasteiger, N.M. Markovi c, Science 324 (2009) 48e49.