J Solid State Electrochem (2014) 18:427–433 DOI 10.1007/s10008-013-2272-3
ORIGINAL PAPER
Hollow nanodendritic nickel oxide networks prepared by dealloying of nickel–copper alloy Woo-Sung Choi & Wonyoung Chang & Heon-Cheol Shin
Received: 1 July 2013 / Revised: 22 September 2013 / Accepted: 23 September 2013 / Published online: 4 October 2013 # Springer-Verlag Berlin Heidelberg 2013
Abstract Three-dimensional hollow nanorod network of nickel has been produced by a dealloying process of the electrodeposited nickel–copper alloy with nanodendritic structure. The nanostructured nickel was subsequently heat treated to form the nickel oxide with little change in the original structure. The resulting sample was tested as the high rate anode in a rechargeable lithium battery. It shows the exceptional rate capability, far exceeding that of the counterpart of nickel–copper oxide network with normal solid branches: reversible capacity at the rate of 20.9 A g−1 is approximately 70 % of the capacity at 0.26 A g−1 rate. Keywords Dealloying . Electrodeposition . Porous structure . Nickel oxide . Lithium battery
Introduction Nickel oxide has been extensively studied as active or catalytic material for gas sensor, electrochromic device, rechargeable battery, and electrochemical capacitor applications. It has attracted particular interest within the field of battery technology as a promising anode for next-generation rechargeable lithium batteries, due primarily to its high specific capacity, low cost, and excellent cycling stability [1]. However, its high operating potential (0.8~2.0 V vs. Li/Li+) and slow lithiation/ delithiation kinetics need to be adjusted to make it viable. In W.
particular, the kinetic sluggishness is mainly attributed to the formation of the electrical insulating phase Li2O during the conversion process. One potential method to overcome this inherent limitation and thus boost energy and power output is to create a morphology that provides both an extremely large electrochemically active surface area and high accessibility to active species. In this regard, the synthesis of different shapes of nanostructured nickel oxides such as nanospheres, nanowalls, and nanowires have been prepared and tested as the anode in lithium batteries [2–6]. Nevertheless, the design and practical implementation of nickel oxide nanoarchitectures with enhanced performance remains a considerable challenge. In our recent publication [7], nickel–copper foams with nanodendritic structure were successfully prepared via electrochemical deposition under a hydrogen gas evolution condition [8, 9]. Furthermore, the copper in the nickel–copper foam was electrochemically etched to produce pure nickel foam for use in various functional devices. Similarly, Chung's group subsequently prepared nanoporous nickel oxide foam on platinum for obtaining a supercapacitor electrode [10]. Nanodendritic foam has great potential to become the favored electrode structure for high-power battery operation, as demonstrated by the excellent rate performance of metallic [11, 12] and oxide anodes [13] with solid branches. But, the proper tailoring of branch structure possibly further enhances the electrochemical performance. Here, the hollow nanorod networks of nickel(−copper) oxides were synthesized by the combination of electrodeposition, dealloying, and high temperature oxidation, and they were evaluated as the anode in a rechargeable lithium battery. Morphology change of the nickel–copper and nickel foams due to the oxidation process is suggested. Then, the electrochemical properties of nickel–copper oxide foam with solid branches and nickel oxide foam with hollow branches are compared in terms of cycling stability and rate capability.
428
Experimental procedure Three different samples were prepared in this study: (1) Nickel–copper foam was coated onto nickel foil (Alfa Aesar, 99 %) substrate in an aqueous solution of 0.4 M NiSO4, 0.04 M CuSO4, 1 M H2SO4, and 1 M NaCl, under a constant current of 2 A cm−2, and was then thermally oxidized at 350 °C in air for 10 min. (2) The copper in the nickel–copper foam of (1) was selectively etched for 5 min in an aqueous solution of 0.2 M (NH4)2S2O8 and the resulting nickel foam was subsequently oxidized under the same heat treatment conditions. (3) The nickel foam in (2) obtained right after the dealloying process was thermally oxidized at 350 °C in air for a longer time of 30 min. Processes (1), (2), and (3) produce a nanodendritic nickel–copper oxide foam (NCOF) and two nanodendritic nickel oxide foam with different degrees of oxidation [(2), nickel oxide foam (NOF)-1; (3), NOF2]. The morphologies and compositions of the samples were analyzed using a field emission scanning electron microscope (S-4800, Hitachi, Japan) equipped with an energy-dispersive X-ray spectrometer (EDS, 7593-H EMAX, Horiba, Japan) and a transmission electron microscope (TEM, Tecnai F20, FEI, The Netherlands). In order to more clearly quantify the composition of the samples, the scratched powders were analyzed with inductively coupled plasma optical emission spectrometer (ICP-OES, Optima 8300 model, PerkinElmer, USA)
Fig. 1 a Surface image of nanodendritic nickel–copper foam and b its cross-sectional view, together with the content profiles of nickel and copper across thickness. c Surface image of nanodendritic nickel–copper oxide foam (NCOF). Upper and lower insets in a and c are magnified SEM and TEM images of the branches, respectively. d X-ray diffraction patterns of nickel–copper foam and NCOF powders scratched from the substrate (Ni, JCPDS 87–0712; Cu, JCPDS 85–1326; NiO, JCPDS 78–0693; and CuO, JCPDS 80–1917)
J Solid State Electrochem (2014) 18:427–433
and the results were compared with those obtained from the EDS analysis. For structural characterization, the foams on the substrate were scratched from the substrate and the powder X-ray diffraction (XRD) patterns were recorded using X-ray diffractometer with Cu K α radiation (XRD, D8 Advance, Bruker, Germany). For electrochemical measurements, the foams on the nickel substrate were used as the working electrode without binder and conducting materials and lithium foil was used as the counter electrode. In order to compare the electrochemical property of the foams with that of thin film, the nickel substrate was thermally oxidized at 350 °C in air for 10 min to form thin nickel oxide layer on the substrate surface and the resulting sample were electrochemically characterized. One molar solution of lithium hexafluorophosphate in a 1:1 volume mixture of ethylene carbonate and diethyl carbonate was used as an electrolyte. A two-electrode stainless steel cell (Hohsen Corp., Japan) was adopted for all electrochemical measurements. In particular, a Teflon O-ring (500 μm thickness) was inserted between the sample and separator (Celgard 2400) to prevent any mechanical damage to the porous structure during cell assembly. To evaluate cycling stability, all the test cells were charged/discharged between 0.01 and 3.0 V vs. Li/Li+ for 50 cycles at the constant current density of 0.1 mA cm−2 (115 mA g−1 for NCOF, 262 mA g−1 for NOF-1, and 176 mA g−1 for NOF-2). The rate performance was analyzed
J Solid State Electrochem (2014) 18:427–433
429
at different current densities, ranging from 0.1 to 8.0 mA cm−2. All of the cells were assembled in a glove box (MBraun, Germany) filled with purified argon gas.
Results and discussion Figure 1 shows microscopic images of the as-prepared nickel– copper foam and its oxidized product NCOF, together with their X-ray diffractograms. Nickel–copper foam is characterized by micro-sized pores and nanosized branches that are uniformly distributed throughout the deposit. The proportion of nickel to copper in the deposit increased in the direction towards the substrate (Fig. 1b). Further details on the microstructure and composition of the deposits can be found in our earlier work [7]. It should be mentioned that the nickel substrate might contribute to the nickel content in the ESD analysis because of the sampling volume analysis of the technique. In this case, the nickel content of the foam near the substrate would be overestimated as it actually is. Nevertheless, it looks like such a sampling volume effect might not invalidate the composition analysis of this work at least from the qualitative perspective: the variation trend of nickel-to-copper ratio in the foam of this work is quite consistent with that in our previous study on alumina substrate/copper ultrathin film (500 nm)/nickel–copper foam [7]. The average contents of nickel and copper in the nanodendritic nickel–copper foam, NCOF, NOF-1, and NOF2 were listed in Table 1. The average contents estimated by the EDS analysis were almost identical to those obtained by the ICP-OES analysis. Thermal oxidation had little effect on the overall foam structure, although it made the nanobranches rounded and bulbous in appearance (Fig. 1c and its insets). The significant change in shape of the branches most likely originated from the volume expansion and the rearrangement of atoms during oxidation. More precisely, the X-ray diffraction pattern of the NCOF (Fig. 1d) revealed that the nickel and copper partially reacted with oxygen in air to form nickel(II) and copper(II) oxides. Despite a relatively short duration of heat treatment,
Table 1 Chemical composition of the foams Average contents (at.%)
Nickel Copper Oxygen
Sample/method Nanodendritic nickel–copper foam
NCOF
ICP-OES
EDS
EDS
61.1 38.9 –
60 40 –
36 32 32
NOF-1
NOF-2
71 – 29
59 – 41
Fig. 2 a Surface image of hollow nanodendritic nickel foam. b , c The surface images of nickel oxide foams prepared by thermal oxidation of hollow nanodendritic nickel foam at 350 °C for 10 min (NOF-1) and 30 min (NOF-2), respectively. Lower inset in c is their X-ray diffraction patterns
430
copper appeared to be oxidized to copper(II) oxides without any trace of copper(I) oxides. Fast kinetics of copper thermal oxidation [14, 15] and extremely large surface area of nickel– copper foams are responsible for the formation of copper(II) oxides. On the other hand, based on molar volume of each component, nickel(II) and copper(II) oxides have molar volumes 1.70 (11.15 cm3 mol−1) and 3.35 (12.61 cm3 mol−1) times greater than those of their original reduced forms (Ni, 6.59 cm3 mol−1; Cu, 7.11 cm3 mol−1). Volume expansion of the branches considerably reduced the free space inside the wall, as clearly seen in the microscopic images, and inhibits
Fig. 3 TEM images of the samples. (a-1, a-2), Hollow nanodendritic nickel foam; (b-1, b-2), NOF-1; (c-1, c-2), NOF-2
J Solid State Electrochem (2014) 18:427–433
facile transport of active species. This situation is not desirable for sustained high-rate operation of a battery electrode. It was previously reported [7, 10, 16] that the selective etching of copper in nanodendritic nickel–copper foam or columnar nickel–copper alloys leads to porous nickel branches. This strongly implies that the oxide branches with internal free space might be created by selective copper etching and subsequent oxidation. Figure 2a–c shows the morphology of copper-etched nickel foam, NOF-1, and NOF-2, respectively, together with their X-ray diffraction patterns (lower inset in Fig. 2c). Copper removal did not influence the overall foam structure but produced the pores within the
J Solid State Electrochem (2014) 18:427–433
branches (upper insets of the figures). The X-ray diffraction pattern showed that the heat-treated sample mainly consisted of elemental nickel and nickel(II) oxide although trace amount of copper and/or copper oxide may be still present in the sample (please see the small shoulders due to copper in the upper diffraction pattern of the inset). It is noted that the diffraction peak ratio of Ni to NiO of NOF-2 was much smaller than that of NOF-1 due primarily to the longer heat treatment time of NOF-2 and the resulting decrease in Ni content. The morphology of hollow branches of all three samples was further investigated through the TEM analysis (Fig. 3). Notably, the unique structures of the micro-sized foam and hollow nano-branches were preserved after heat treatment, and displayed an irregular branch wall caused by the oxidation-induced volume expansion of nickel. The wall and the inside free space of the hollow branches in NOF-2 were thicker and narrower, respectively, than those in NOF-1 owing to the thermal oxidation of the larger amount of nickel to nickel oxide. Figure 4a–c presents typical voltage profiles of NCOF, NOF-1, and NOF-2, respectively, for the first three cycles as
Fig. 4 Typical voltage profiles of a NCOF; b NOF-1; and c NOF-2 for the first three cycles at a current density of 0.1 mAh cm−2. d Dependence of specific discharging capacity and coulombic efficiency on the number of cycles
431
the anode in a rechargeable lithium battery. After a capacity loss during the first lithiation process, due to the formation of the solid electrolyte interphase (SEI) layer, NCOF, NOF-1, and NOF-2 reacted quite reversibly with lithium to deliver the specific (gravimetric) discharging capacities ca. 290, 420, and 860 mAh g−1, respectively, at a current density of 0.1 mAh cm−2. The specific capacity of NOF-2 is much larger than that of NOF-1 because of the larger amount of active materials (i.e., NiO). Notably, the specific capacity of NCOF is lower than that of NOF-1 in spite of the presence of the extra active phase of CuO. It is most likely due to the mass contribution of CuO and the longer lithium diffusion length through NCOF than through NOF during the cell operation. Actually, the areal capacity of NCOF (0.26 mAh cm−2) was estimated to be higher than that of NOF-1 (0.16 mAh cm−2), strongly indicating that the mass of CuO in NCOF is one of the main reasons for the reduced gravimetric capacity. In contrast, the areal capacity of NOF-2 (0.49 mAh cm−2) was higher than that of NOF-1, consistent with the tendency of gravimetric capacity. It would be interesting to compare the areal capacity of the samples with foam structures with that of thin film sample
432
J Solid State Electrochem (2014) 18:427–433
(please see “experimental procedure” for preparation details). The areal capacity of dense nickel oxide layer was estimated to be about 0.02 mAh cm−2 (the voltage profile is not presented in this work). This indicates that the measured capacities of the NCOF, NOF-1, NOF-2 are mostly attributed to the foam structures and the contribution of nickel substrate with oxidized surface is not so large. The first lithiation voltage profile of NCOF featured three inflection points associated with the reduction of copper
oxides and subsequent formation of SEI layer, whereas that of NOF showed relatively monotonous behavior. It is noted
Fig. 5 a Specific capacity and b proportional specific capacity reproduced from a as a function of discharging rate
Fig. 6 Morphologies of a NCOF; b NOF-1; and c NOF-2 after the electrochemical tests
J Solid State Electrochem (2014) 18:427–433
that both of NOF-1 and NOF-2 showed higher reduction plateau than that of typical nickel oxide (~0.5 V vs. Li/Li+ [1]). The reason for this has not been clearly understood, but the nanoporous nature of our samples is possibly responsible for such a discrepancy [2, 17]. After the first cycle, samples were characterized by two inflection points associated with the reactions of the SEI layer and their oxides [18]. Figure 4d shows the specific capacities and coulombic efficiencies of three samples as a function of the cycle number. In the first cycle, the coulombic efficiency of NOF-2 (~70 %) was much higher than those of NCOF and NOF-1 (~40 %). The enhanced coulombic efficiency or decreased irreversible capacity of NOF-2 in the first cycle is ascribed to its larger mass of the active materials (i.e., oxides) per real surface area than those of NCOF and NOF-1. Such smaller surface area-toweight ratio of NOF-2 probably results in reduced “specific” irreversible capacity. Furthermore, non-faradaic reactions arisen on extremely large surface area of NOF-1 and NOF-2 might have an effect on initial coulombic efficiency [19]. Coulombic efficiency exceeded 97 % except for the first and second cycles, and it occasionally exceeded 100 %, probably because of the partial delithiation of unstable SEI [18]. The capacity retentions of all three samples (NCOF, NOF-1, NOF-2) maintained more than 75 % of initial reversible capacity after 50 cycles. Figure 5 presents the rate performance of three samples. The NOF-1 showed the outstanding rate performance, far exceeding those of NCOF and NOF-2: about 70 % of the specific capacity at a current density of 0.1 mAh cm−2 (0.26 A g−1) was maintained at a very high current drain of 8.0 mAh cm−2 (20.9 A g−1). In sum, the thermal oxidation condition of hollow nanodendritic nickel networks plays an important role in their electrode performance. That is, the longer heat treatment leads to greater specific capacity but lower rate capability, and vice versa. This is most likely because the thicker oxide film formed during longer oxidation period would provide larger amount of active mass, but the electron conduction and solid-state lithium diffusion through it must be more limited. The overall foam structures of the samples remained unchanged after charge/discharge cycling, as shown in Fig. 6. However, the rounded nanobranches of the NCOF showed considerable agglomeration due to the cyclinginduced shape change of the oxide and SEI layer [18], and thus, the nanoporous character of the foam wall almost disappeared (inset in Fig. 6a). In contrast, such a shape change was significantly alleviated in the nanobranches of the NOF-1 and NOF-2 (insets of panel b and c of Fig. 6, respectively), possibly because of their highly open, porous structure.
433
Concluding remarks Hollow nanodendritic nickel oxide network was produced by a combination of the porous nickel–copper electrodeposition, dealloying, and heat treatment and was tested as the anode in rechargeable lithium battery. Its rate capability was found to surpass that of the counterpart of nickel–copper oxide foam with normal solid branches. The unique structure suggested in this work might be appropriate for use in a high-rate transition metal oxide anode. Acknowledgments This research was supported by the Converging Research Center Program (2013 K000212) and by the Nuclear Research and Development Program through the Ministry of Science, ICT and Future Planning, Korea. This work was partially supported by the National Research Foundation of Korea Grant funded by the Korean Government (NRF-2011-C1AAA001-0030538).
References 1. Poizot P, Laruelle S, Grugeon S, Dupont L, Tarascon JM (2000) Nature 407:496–499 2. Yuan L, Guo ZP, Konstantinov K, Munroe P, Liu HK (2006) Electrochem Solid State Lett 9:A524–A528 3. Varghese B, Reddy MV, Yanwu Z, Lit CS, Hoong TC, Rao GVS, Chowdari BVR, Wee ATS, Lim CT, Sow C-H (2008) Chem Mater 20:3360–3367 4. Wang B, Cheng JL, Wu YP, Wang D, He DN (2012) Electrochem Commun 23:5–8 5. Aravindan V, Kumar PS, Sundaramurthy J, Ling WC, Ramakrishna S, Madhavi S (2013) J Power Sources 227:284–290 6. Yan X, Tong X, Wang J, Gong C, Zhang M, Liang L (2013) J Alloys Compd 556:56–61 7. Choi WS, Jung HR, Kwon SH, Lee JW, Liu M, Shin HC (2012) J Mater Chem 22:1028–1032 8. Shin HC, Dong J, Liu M (2003) Adv Mater 15:1610–1614 9. Shin HC, Liu M (2004) Chem Mater 16:5460–5464 10. Jeong MG, Zhou K, Cherevko S, Chung C-H (2013) Korean J Chem Eng 29:1802 11. Shin HC, Liu M (2005) Adv Funct Mater 15:582–586 12. Jung HR, Kim EJ, Park YJ, Shin HC (2011) J Power Sources 196: 5122–5127 13. Chen X, Sun K, Zhang E, Zhang N (2013) RSC Adv 3:432–437 14. Hu YZ, Sharangpani R, Tay S-P (2001) J Electrochem Soc 148: G669–G675 15. Cocke DL, Schennach R, Hossain MA, Mencer DE, McWhinney H, Parga JR, Kesmez M, Gomes JAG, Mollah MYA (2005) Vacuum 79: 71–83 16. Chang JK, Hsu SH, Sun IW, Tsai WT (2008) J Phys Chem C 112: 1371–1376 17. Zhou G, Wang DW, Yin LC, Li N, Li F, Cheng HM (2012) ACS Nano 6:3214–3223 18. Débart A, Dupont L, Poizot P, Leriche J-B, Tarascon JM (2001) J Electrochem Soc 148:A1266–A1274 19. López MC, Ortiz GF, Lavela P, Alcántara R, Tirado JL (2013) ACS Sustainable Chem Eng 1:46–56