Superionic Phase Stabilization and Conductivity - ACS Publications

Jul 19, 2015 - temperature stabilizing α-phase (down to 21 °C, the lowest in state of the art temperatures) is reproducible and survives further the...
1 downloads 10 Views 6MB Size
Letter pubs.acs.org/NanoLett

Size-Controlled AgI/Ag Heteronanowires in Highly Ordered Alumina Membranes: Superionic Phase Stabilization and Conductivity Hemin Zhang,*,† Takashi Tsuchiya,† Changhao Liang,‡ and Kazuya Terabe*,† †

International Center for Materials Nanoarchitechtonics (WPI-MANA), National Institute for Materials Science (NIMS), 1-1 Namiki, Tsukuba, Ibaraki 305−0044, Japan ‡ Key Laboratory of Materials Physics and Anhui Key Laboratory of Nanomaterials and Nanotechnology, Institute of Solid State Physics, Hefei Institutes of Physical Science, Chinese Academy of Sciences, Hefei 230031, China S Supporting Information *

ABSTRACT: Nanoscaled ionic conductors are crucial for future nanodevices. A well-known ionic conductor, AgI, exhibited conductivity greater than 1 Ω−1 cm−1 in α-phase and transformed into poorly conducting β-/γ-phase below 147 °C, thereby limiting applications. Here, we report that transition temperatures both from the β-/γ- to α-phase (Tc↑) and the α- to β-/γ-phase (Tc↓) are tuned by AgI/Ag heteronanowires embedded in anodic aluminum oxide (AAO) membranes with 10−30 nm pores. Tc↑ and Tc↓ shift to correspondingly higher and lower temperature as pore size decreases, generating a progressively enlarged thermal hysteresis. Tc↑ and Tc↓ specifically achieve 185 and 52 °C in 10 nm pores, and the final survived conductivity reaches ∼8.3 × 10−3 Ω−1 cm−1 at room temperature. Moreover, the lowtemperature stabilizing α-phase (down to 21 °C, the lowest in state of the art temperatures) is reproducible and survives further thermal cycling. The low-temperature phase stabilization and enhancement conductivity reported here suggest promising applications in silver-ion-based future nanodevices. KEYWORDS: Silver iodide, heteronanowire, size dependence, phase transition, conductivity

S

room temperature (RT) by using rapid melt-quenching techniques in a glass matrix.12 Unfortunately, such stabilization easily disappears and is irreversible after thermal cycling. For example, the α-AgI would disappear at 110 °C on cooling in the system of AgI−Ag2O−B2O3.13 In particular, ionic AgI-based composite solid electrolytes, meaning heterogeneous doping with the second phase in which AgI heterogeneously mixes with an insulating matrix (e.g., an organic compound or oxide glass), are remarkably more successful in enhancing ionic conductivity than the others.14 The evolution of an interfacial phase in AgI/ Al2O3 composites, such as structures, components, and size, has demonstrated critical factors in determining the ionic conductivity of AgI. The concept of heterogeneous doping quantitatively developed by Maier15,16 comprehensively explains the ionic-conduction mechanism in such composite systems. According to the proposed concept, the enhancement of ionic conductivity is attributed to existing space-charge regions and high concentrations of compensating defects around the insulator/conductor boundaries. It has also been demonstrated that such enhanced ionic conductivity is related to the formation of 7H polytype AgI with stacking-fault

uperionic conductors are multicomponent solids that have both solid and liquid characteristics and exhibit extremely high conductivity comparable to liquid electrolytes. They have also attracted a great deal of attention due to their huge potential applications in electrochemical devices, such as batteries, fuel cells, and sensors.1 Solid-state ionic conductors are much more suitable than liquid phases in terms of device fabrication (easily shaped, patterned, integrated, and miniaturized), safety (nonexplosive), and stability (nonvolatile). Agbased superionic solids, such as AgI and Ag2S, are excellent candidates for the electrolytes in all-solid-state devices due to their good conductivity and high polarizability of Ag+ ions, resulting in a high exchange rate at the electrode.2−4 For example, all-solid-state devices based on silver ionic conductors have repeatedly been reported.5−7 Nevertheless, the total performance of such devices still needs to be significantly improved, especially to satisfy commercial applications. The extremely high ionic conductivity (larger than 1 Ω−1 cm−1) within high-temperature phase or α-phase AgI is predominantly due to Ag+ ions resulting from the large degree of disorder in the silver sublattice. Many methods have been developed to obtain high conductivity AgI at lower temperatures, such as metal halide solid solutions,8 ternary phases (RbAg4I5) derived from AgI,9 and inorganic composite materials.10,11 Stabilization of α-AgI could be successful at © XXXX American Chemical Society

Received: April 10, 2015 Revised: June 29, 2015

A

DOI: 10.1021/acs.nanolett.5b01388 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters

Figure 1. Characterization of morphology of highly ordered AAO membranes with different sizes. High resolution FESEMs of top surface (15°tilted) (a), bottom surface (b), and cross-section (c) for the 10 nm pore AAO membrane. (d−f, g−i) Corresponding FESEMs for the 20 and 30 nm pore AAO membranes. Scale bars of 30 (a−c) and 50 nm d−i).

Figure 2. Schematic of synthesis process for AgI/Ag HNWs structure using AAO membrane with electrochemical method.

arrangements in the interfacial layer.12 An alternative strategy to stabilize the superionic phase to RT is done by fabricating AgI on the nanoscale level. The temperatures of lattice fusion and of order−disorder phase transitions can be lowered with ionic conductors sized down to the nanoscale level as the proportion of surface atoms increases, which results in more mobile ions and higher defect concentration.17 Nanoscale superionics have actually demonstrated that the transition temperature is reduced at small sizes. For instance, AgI nanoplates synthesized by a water-soluble cationic polyelectrolyte displayed a lowered phase transition temperature and produced drastic enhancements in conductivity.18 Polymer surface treatments of AgI nanoparticles enabled the superionic phase of AgI to be stabilized to RT.17,19 Some work has also reported the synthesis of AgI nanorods and nanowires using anodic aluminum oxide (AAO) templates through electrochemical reactions20,21 and natural meeting reactions of Ag+ and I− ions,22 in which the pore sizes of AAO membranes have usually been larger than 50 nm. It should be noted that much smaller pore sizes of AAO membranes will exhibit considerably distinctive properties due to the significant effect of the size and interface, especially quantum confinement effects, which often occur exclusively for only less than ten nanometers or tens of nanometers. Here we systematically investigated the effect of AAO pore size on the transition phase and conductivity of AgI/Ag heteronanowires (HNWs) embedded in 10, 20, and 30 nm

pore AAO membranes. The results revealed that transition temperatures of AgI both from the β-/γ- to the α-phase (Tc↑) and from the α- to the β-/γ-phase (Tc↓) could be tuned according to the pore sizes in AAO membranes. When the pore size in AAO membranes specifically approaches 10 nm, Tc↑ achieves 185 °C and Tc↓ achieves 52 °C, leading to a large temperature difference of 133 °C. The temperature of the stabilizing α-phase in 10 nm pore AAO membranes could also be brought down to 21 °C, which is the lowest in state of the art temperatures. Moreover, the final survived conductivity for AgI/Ag nanowires in the 10 nm pore AAO membrane can reach ∼8.3 × 10−3 Ω−1 cm−1 even at RT (21 °C). More importantly, the low-temperature stabilization of the α-phase is reproducible and survives further thermal cycling. A highly ordered AAO membrane with a size of sub-10 nm can easily be obtained by using galvanostatic pretreatment before second anodization in the process of preparing AAO membranes.23 Figure 1 shows high resolution field emission scanning electron micrographs (FESEMs) of highly ordered AAO with precisely controlled pore sizes of 10, 20, and 30 nm. As shown in Figure 1a, the pore arrangement on the top surface is exceptionally ordered and uniform with a pore size of 10 ± 1 nm and a pore interdistance of 55 ± 5 nm. After the barrier layer was removed, the pores on the bottom surface exhibited perfect hexagonal arrangements and were slightly larger than those on the top surface with a size of 12 ± 3 nm (Figure 1b). B

DOI: 10.1021/acs.nanolett.5b01388 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters The field emission scanning electron micrograph (FESEM) of the cross-sectional view (Figure 1c) further demonstrates that the pore size is precisely controlled within 10 nm. Precisely controlled pore size of 20 and 30 nm (Figure 1d−f,g−i) could be obtained by using chemical etchants. Note that it was difficult to electroplate metallic nanowires into thick AAO membranes with pore sizes of less than ten nanometers or tens of nanometers because the wettability of small pores was difficult to achieve. Prior to electroplating, the AAO membrane coated with Pt layers was softly ultrasonicated in N,Ndimethylformamide for 1−3 min to improve wetting.24 Figure 2 is the schematic of the electrochemical synthesis for the structure of AgI/Ag HNWs embedded in an AAO membrane. Pulsed cathodic current deposition was adopted (see Figure S1) to improve the filling rate of Ag nanowires. The filling rate of Ag nanowires could achieve approximately 90% by combining a wettability technique and pulsed cathodic current deposition, which could be estimated from the AgI nanowire tips that protruded from the AAO membrane (see Figure S2). A pulsed anodic current was applied (see Figure S3) during electrochemical iodination to achieve the same length and homogeneous AgI nanowires. An initial AgI layer was formed at the interface of the Ag nanowires and potassium iodide (KI) solution under an applied anodic electrical field along the pore axis. Then, the oxidized Ag+ ions migrated through the solid AgI layer and reacted with I− ions at the solid/liquid interface in the following growth (the migration of Ag+ ions was more dominant inside AgI than that of I− ions). The resistance of reaction cells gradually increased (see Figure S3b) because of the continuous formation of AgI nanowires in the AAO membrane as iodination progressed. Figure 3 shows the morphology of AgI/Ag HNWs in the 10 nm pore AAO membrane. A typical top-view of AgI nanowire tips protruding from the AAO membrane can clearly be seen (Figure 3a), which indicates that highly ordered AgI/Ag HNWs are embedded in the AAO membrane, and the pore size is a

little bigger after deposition and iodination of Ag nanowire. The reason may be attributed to the increased tensile stress of highly ordered AAO membrane25 and the etching of the inner wall of AAO membrane by OH− ions in the process of iodination. We used the top surface of the AAO membrane as the Pt sputtering side to favor subsequent treatment (such as mechanical polishing and conductivity measurements). Therefore, the pore arrangement in Figure 3a is similar to that in Figure 1b. The corresponding low magnification FESEM for AgI protruding from the AAO membrane can be seen in Figure S2. A cross-sectional FESEM for AgI/Ag HNWs in the 10 nm pore AAO membrane indicates that an obvious interface exists between AgI and Ag nanowires due to the clearly different contrast of AgI and Ag caused by electron beams (Figure 3b), which was also confirmed by energy dispersive X-ray (EDX) spectra (see Figure S4). Interestingly, the AgI and Ag nanowires partially released from the 10 nm pore AAO membrane had different morphological evolutions. Compared with AgI nanowires (see Figure S5a), Ag nanowires seemed to evolve into chainlike nanorods or nanoparticles (see Figure S5b) in the 10 nm pore AAO membrane. Besides, high-angleannular-dark-field scanning transmission electron microscopy (HAADF-STEM) (Figure 3c) indicates that the original morphology of Ag nanowires embedded in the 10 nm pore AAO membrane by electroplating were straight, homogeneous, and had no breaks. Moreover, it can be seen from the STEM of AgI nanowires in the AAO membrane (Figure 3d) that the continuous nanowire shape was still maintained after electrochemical iodination. EDX element mapping for AgI nanowires in the AAO membrane (see Figure S6) shows the spatial distributions of Al, O, Ag, and I elements. Due to the instability of AgI, especially on the nanoscale level, I elements easily break away from AgI (I element signals exhibit more dispersed states than that of Ag elements, see Figure S6f), and some breakup occurs despite the protection of the AAO membrane, which results from high energy electron beam (200 kV) irradiation. However, this situation is opposite in AgI/Ag HNWs in the 20 nm pore AAO membrane. The released AgI nanowires tended to transform into nanorods or nanoparticles with respect to Ag nanowires (see Figure S7), which was similar to AgI/Ag HNWs in the 60 nm pore AAO membrane. 20 These results demonstrate that the behavior of AgI/Ag HNWs in the 10 nm pore AAO membranes is completely different with larger pore membranes (such as 20 or 60 nm pores), which further confirms the distinctive effect of the size and interface of the 10 nm pore AAO membrane. Interestingly, individual AgI nanowires released from the 10 nm pore AAO membrane exhibited significantly different performance (see Figure S8). One AgI nanowire quickly transformed into nanoparticles due to the transport of AgI and/or Ag to both ends of the nanowire during electron beam irradiation (5 kV). Surprisingly, another one became very small and had a long filament under the same conditions. This is probably related to intrinsic properties and the different orientational growth of AgI nanowires. We measured the X-ray diffraction (XRD) spectra of the Ag nanowires in different AAO membranes and after electrochemical iodination (Figure 4) to clarify the dependence of crystal structure evolution on size. All the peaks in the Ag XRD spectra (Figure 4a) can be indexed to a face centered cubic structure (Joint Committee on Powder Diffraction Standards (JCPDS) card number, 04-0783). Broadening peaks can clearly be observed with the decreasing diameter of AAO membranes, especially the main peak (220) (see Figure S9), accompanying

Figure 3. Characterization of morphology of AgI/Ag HNWs in the 10 nm pore AAO membrane. (a) High-resolution FESEM of top-view of AgI protruding from AAO (using top surface of AAO membrane as Pt sputtering side). (b) Cross-sectional FESEM of AgI/Ag HNWs in the 10 nm pore AAO membrane (dotted line denotes clearly different contrast between AgI and Ag). (c,d) HAADF-STEMs for Ag and AgI in AAO membranes. Scale bars of 30 nm (a), 1 μm (b), and 100 nm (c,d). C

DOI: 10.1021/acs.nanolett.5b01388 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters

Figure 4. (a) XRD patterns of Ag embedded in AAO with pore size of 10, 20, and 30 nm. (b) Corresponding XRD patterns of AgI/Ag HNWs after electrochemical iodination.

the decrease in their relative intensity that resulted from the lowering volume ratio for Ag and the AAO membrane and smaller crystallites. The preferential growth orientations for the three samples also remained consistent. After electrochemical iodination, AgI peaks emerged due to the formation of AgI/Ag HNWs. Apart from the three remaining Ag peaks of (111), (220), and (311), the other peaks in the XRD spectra could accordingly be indexed to the low-temperature phases of the close-packed cubic β-AgI (JCPDS card number, 09-0374) with wurtzite structures and hexagonal γ-AgI (JCPDS card number, 09-0399) with zincblende structures (Figure 4b). Interestingly, the diffraction peak intensity of plane β (002)/γ (111) weakened and that of β (110)/γ (220) strengthened with the decreasing diameter of the AAO membrane, which demonstrated that the preferential growth orientations were changed. The more the pore size decreased, the more easily homogeneous nucleation was achieved, resulting in the creation of fewer crystal planes. The reason seems more likely for AgI nanowires in the 10 nm pore AAO membrane that the growth of plane β (002)/γ (111) was suppressed due to the small pore size and few nucleations, thereby leading to increased growth in other crystal planes, especially plane β (110)/γ (220). However, direct evidence for the change of preferential growth orientation is currently not obtainable through high-resolution transmission electron microscopy observation due to significant decomposition of AgI under intense electron-beam irradiation.20 Figure 5 plots the differential scanning calorimetry (DSC) thermograms of as-synthesized AgI/Ag HNWs measured during both heating and cooling and the extracted phase transition temperature dependent on size. Two overlapping DSC peaks can distinctly be observed both when heating and cooling (the DSC thermograms for 15, 20, and 30 nm cases

Figure 5. Size-dependent phase transitions for AgI/Ag HNWs in AAO membranes. (a) DSC thermograms of AgI/Ag HNWs in different AAO membranes for the first heating cycle. (b) DSC thermograms of AgI/Ag HNWs in different AAO membranes for the first cooling cycle. (c) Evolution of the transition temperature, Tc, with changes in pore sizes. Red spheres indicate the β-/γ- → α-phase transition temperature, Tc↑, in heating cycles. Blue spheres indicate the α- → β-/γ-phase transition temperature, Tc↓, in cooling cycles. Tc↑ and Tc↓ are defined as the second DSC peak in each thermogram. The dotted line denotes the baseline, and the red arrow denotes the peak position in DSC thermograms. The data for 50−60 nm AgI nanowires have been referenced in our previous work.21

actually consist of two peaks, see Figure 5b), which is a huge difference from that in previous reports.17,21,23,26 When heating, the β-/γ- → α-phase transition temperature, Tc↑, and the region of Tc↑ increase gradually with decreasing diameters of AgI/Ag nanowires, especially for AgI/Ag nanowires in the 10 and 15 nm pore AAO membranes (Figure 5a). This phenomenon, however, is just the opposite when cooling. The α- → β-/γphase transition temperature, Tc↓, lowers gradually with decreasing diameters of AgI/Ag nanowires. The region of Tc↓ simultaneously exhibits the same tendency, obviously for AgI/ Ag nanowires in the 10 and 15 nm pore AAO membrane (Figure 5b). Thus, the suppression of phase transition when both heating and cooling leads to a progressively enlarged thermal hysteresis. Tc↓ is drastically suppressed down to ∼52 D

DOI: 10.1021/acs.nanolett.5b01388 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters °C for AgI/Ag nanowires in the 10 nm pore AAO membrane and Tc↑ is enhanced up to ∼185 °C, resulting in a huge hysteresis with a temperature width of 133 °C (Figure 5c). More importantly, the low-temperature stabilization of α-AgI is reversible and survives further thermal cycling due to AAO membrane protection (see Figure S10 and S11). The pronounced effect of suppression on phase transition seems to be attributed to the significant effect of size and the interface. The interface between the AAO inner wall and inner AgI nanowire is very unique. Particularly, many defects (such as ionized oxygen vacancies, hydroxyl groups,27 and SO42− ions28,29) may be formed on the inner wall of pores during the growth of the AAO membrane. These defects would strongly absorb Ag+ ions, whereas I− ions remain on the surface of AgI nanowires, generating a considerable number of lattice defects, which ultimately result in charge imbalance. At higher temperatures (near to or in the α-AgI phase), this effect would be greatly enhanced due to highly mobile Ag+ ions. More importantly, interfacial AgI with charge imbalance is more stable than that of inner AgI (more like bulk AgI) making it difficult to create phase transitions, probably necessitating a higher transition temperature when heating and a lower one when cooling. This is why the DSC thermograms both during heating and cooling exhibit two individual peaks. It is reasonable to speculate that the first peak results from the inner part of AgI nanowires displaying a near transition temperature of bulk AgI, especially for Tc↑. The other one should mainly originate from the interfacial part of AgI nanowires, which indicates phase transition when superheating/supercooling. Therefore, phase transition is initiated at the inner part and then shifts gradually to the interfacial part both during heating and cooling, leading to the two observed broad individual DSC peaks, which is in agreement with those of polymer coating AgI nanoparticles.16 Moreover, the experiment results demonstrated that the more the pore size decreased, the more significant the interfacial phase affected (Figure 5). For example, the Tc↑ and Tc↑ regions of interfacial AgI decreased gradually with increased pore size, disappeared when the pore size reached 50−60 nm, and only exhibited one DSC peak at higher temperatures.21 Detailed insights into the phase transition behavior of AgI/ Ag HNWs in the 10 nm pore AAO membrane were obtained from temperature-controlled X-ray diffraction measurements. Figure 6 shows consecutive XRD spectra of temperature evolution for AgI/Ag HNWs in the 10 nm pore AAO membrane from 25 to 200 °C, which were then cooled back to RT. The phase transitions are traced from the peaks of lowtemperature phases β (002)/γ (111) and high-temperature phase α (110) when heating and cooling. The peak of β (002)/ γ (111) intensity for AgI/Ag nanowires at RT in the 10 nm pore AAO membrane is much weaker than that in 20 and 30 nm pore AAO membranes (see Figure 4b). However, the intensity gradually strengthens from 75 to 130 °C with increasing temperature accompanied by an increased shift of 2θ (see Figure S12). This might be caused by the elimination of defects and structural relaxation, which decreased the distance of the crystal plane, resulting in the increase of 2θ. The main reason should be related to the gradual growth of the initial small crystallites, which thereby enhanced the peak intensity. Crystallite growth could simultaneously modify the preferences of crystallographic lattice sites, which could also lead to the change in crystal plane distance. The peak intensity continuously decreased from 130 to 200 °C due to more

Figure 6. In situ temperature-dependent XRD spectra for AgI/Ag HNWs in the 10 nm pore AAO membrane. The first thermal cycle starts from RT to 200 °C and is then followed by cooling to RT. The heating and cooling rate between measurements is 10 °C/min. The temperature controller can precisely control the set temperature at ±0.5 °C.

mobile Ag+ ions in interstitial sites, thermal expansion, and fusion/melting of crystal lattices. An obvious peak of α-AgI appeared at 160 °C, and the peak of β (002)/γ (111) completely disappeared at 180 °C (at 170 °C, the peak of β (002)/γ (111) still remained), indicating Tc↑ started at ∼160 °C and ended at ∼180 °C, which was very consistent with the resulting behavior of DSC for AgI/Ag nanowires in the 10 nm pore AAO membrane (red curve in Figure 5a). The peak intensity of α-AgI decreased at 60 °C when cooling, simultaneously the peak of β (002)/γ (111) appeared, which indicated that Tc↓ started at ∼60 °C. The peak of α-AgI still remained at 50 °C and completely disappeared at 40 °C, accompanying the occurrence of peak β (100), suggesting Tc↓ ended at ∼40 °C, which is in agreement with DSC results on cooling (red curve in Figure 5b). The detailed in situ XRD spectra and related data analysis for AgI/Ag HNWs in the 20 nm pore AAO membrane can be found in Figure S13. The results demonstrate good consistency with corresponding DSC resulting behaviors. Figure 7 plots the results of conductivity measurements for AgI/Ag HNWs in the 10, 15, and 20 nm pore AAO membranes as well as AgI polycrystals (bulk). Detailed information on conductivity measurements can be found in the schematic in Figure 7a. The huge thermal hysteretic behavior observed in DSC and in situ XRD is also well supported by the temperature-dependent conductivity, σ. We previously explained that the interfacial phase is more stable than that of the inner part and its proportion increased as the pore size decreased, generating increasing Gibbs free energy (see Figure S14) to necessitate phase transition, which led to progressively enlarged thermal hysteresis. Note that single nanowire conductivity in the 10, 15, and 20 nm pore AAO membranes was significantly enhanced to ∼3.8 × 10−3, ∼2.9 × 10−3, and ∼2.3 × 10−3 Ω−1 cm−1 at RT, which was also supported by the potential evolution in electrochemical iodination (see Figure S15). Interestingly, these values were further enhanced after first thermal cycling to ∼8.3 × 10−3, ∼6.1 × 10−3, and ∼3.9 × 10−3 Ω−1 cm−1 even cooling down to 21 °C, which were two E

DOI: 10.1021/acs.nanolett.5b01388 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters

Figure 7. Temperature-dependent conductivity, σ, of single AgI/Ag HNW in AAO membrane for the first thermal cycle. (a) Schematic of conductivity measurements for AgI/Ag HNWs embedded in AAO membranes. (b) Activation energies have been denoted, using an Arrhenius expression (estimated from ln σT vs 1/T representation). The conductivity for bulk AgI has also been included for comparison. The ac conductivity of a single AgI/Ag nanowire in AAO membrane could be estimated from the complex impedance spectrum with part of the highest frequency response, simultaneously considering the filling rate (90%, see Figure S2), pore density (approximately 1010/cm−2) in the AAO membrane, and the electrical contact area (300 μm, see Figure S16).

from ∼60, ∼80, and ∼90 °C in the 10, 15, and 20 nm pore AAO membranes, which is also consistent with both DSC and in situ XRD behaviors. Obviously, the σ in the 10 nm pore AAO on cooling was closer to that in the 15 nm pore AAO and than that in the 20 nm pore AAO. The σ in the 10 and 15 nm pore AAO membrane on heating and in the 20 nm pore AAO membrane on cooling did not exhibit very large changes at the transition temperature. This did not seem unreasonable due to the coexistence of α- and β/γ-phase AgI in the large transition region. Note that the activation energy for superionic supercooled AgI nanowire in the 10 nm pore AAO was much higher than that of the well-known superionic α-AgI (∼0.05 eV). The α-AgI phase in the 10 nm pore AAO membrane was significantly different from standard α-AgI that transformed to β-AgI at 147 °C, but supercooled by ∼95 °C and transformed back to the same preferential orientations (see Figure S17a). Interestingly, the α-AgI phase in the 10 nm pore AAO membrane seemed to have a “memory” of preferential orientations in the low temperature phase, probably due to the distribution of nonrandom silver ions near the transition temperature.32 Moreover, this further demonstrated that the low-temperature stabilization of the α-AgI phase in the 10 nm pore AAO membrane was reproducible and enabled further thermal cycling to survive. We not only stress the extremely high ionic conductivity of the stabilized α-AgI phase in AAO membranes at a practical temperature from the perspective of applications but also their facile preparation with an electrochemical method at ambient temperature and pressure. Up to now, most reported Ag+ ionbased superionic materials have needed to be processed with sintering/quenching at high/low temperatures. The results presented here indicate that the Tc of AgI in AAO membranes can be finely tuned by precisely controlling the pore size of AAO membranes on the nanoscale level. The Tc↓ in the 10 nm pore AAO membrane is suppressed by 95 °C down to 52 °C with respect to the bulk. Even though Tc↓ appears at around 52 °C, the stabilization α-phase is also able to survive at a very low temperature (21 °C) after the first thermal cycling due to strong suppression in the interfacial phase. Conductivity can simultaneously achieve ∼8.3 × 10−3 Ω−1 cm−1 at RT (21 °C).

times that before thermal cycling. The final survived σ (∼8.3 × 10−3 Ω−1 cm−1) for the 10 nm pore AAO membrane was comparable to the recently highest ionic conductivity using 11 nm AgI polymer coating nanoparticles16 (∼1.5 × 10−2 Ω−1 cm−1), which was nearly four times that of AgI nanoplates17,30 (∼2.2 × 10−3 Ω−1 cm−1), and that was almost 5 orders of magnitude greater than that of bulk AgI polycrystals (∼1.6 × 10−7 Ω−1 cm−1). The high σ might originate from the interfacial phase of AgI between the inner wall of the AAO membrane and inside the AgI nanowires. The interfacial phase of AgI with decreasing pore size in the AAO membrane accounts for increasingly larger proportion, specifically for the 10 nm pore AAO membrane. The generated space-charge region15,31 and compensating defects in the interfacial-phase AgI could provide high ionic conductivity. Furthermore, the interfacial phase AgI consists of 7H stacking fault arrangements, which significantly contribute to enhanced ionic conductivity.12 Also note that for AgI/Ag nanowires in the 10 nm pore AAO membrane, the peaks in the XRD spectrum slightly shift to smaller angles after the first thermal cycling (see Figure S17a), which indicates the generation of many interstitial Ag+ ions and compensating defects in the cooling process, leading to the increase in crystal plane distance. The generated interstitial Ag+ ions and compensating defects may contribute to further enhanced σ. More importantly, traces of the α-AgI peak (2θ in the range of 24−25°) can still be seen in the high-resolution XRD spectrum (see Figure S17b), which demonstrates that there is still slight α-AgI in the interfacial phase even after cooling down to RT. This is possibly due to almost all the AgI nanowire being involved in the space-charge region, which generates the strong suppression of phase transition. The σ gradually increased for all the AgI/Ag nanowires in the 10, 15, and 20 nm pore AAO membranes when heated, and an observed remarkable increase started at ∼160 °C, where DSC and in situ XRD displayed a thermal event. The σ seemed to demonstrate two-step transitions in the 15 and 20 nm pore AAO membrane compared with AgI in the 10 nm pore size AAO membrane, which is well correlated with two distinct thermal events in the DSC curves. The σ sustained high values until Tc↓ appeared in cooling and obviously started to decrease F

DOI: 10.1021/acs.nanolett.5b01388 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters

(14) Lee, J. S.; Adams, S.; Maier, J. Solid State Ionics 2000, 136−137, 1261. (15) Maier, J. J. Phys. Chem. Solids 1985, 46, 309. (16) Maier, J. Prog. Solid State Chem. 1995, 23, 171. (17) Makiura, R.; Yonemura, T.; Yamada, T.; Yamauchi, M.; Ikeda, R.; Kitagawa, H.; Kato, K.; Takata, M. Nat. Mater. 2009, 8, 476. (18) Guo, Y. G.; Lee, J. S.; Maier, J. Adv. Mater. 2005, 17, 2815. (19) Yamasaki, S.; Yamada, T.; Kobayashi, H.; Kitagawa, H. Chem. Asian J. 2013, 8, 73. (20) Liano, C.; Terabe, K.; Tsuruoka, T.; Osada, M.; Hasegawa, T.; Aono, M. Adv. Funct. Mater. 2007, 17, 1466. (21) Lian, C.; Terabe, K.; Hasegawa, T.; Aono, M.; Nobuo, I. J. Appl. Phys. 2007, 102, 124308. (22) Piao, Y.; Kim, H. Chem. Commun. 2003, 2898. (23) Moyen, E.; Santinacci, L.; Masson, L.; Wulfhekel, W.; Hanbücken, M. Adv. Mater. 2012, 24, 5094. (24) Tian, M.; Xu, S.; Wang, J.; Kumar, N.; Wertz, E.; Li, Q.; Campbell, P. M.; Chan, M. H. W.; Mallouk, T. E. Nano Lett. 2005, 5, 697. (25) Ko, S.; Lee, D.; Jee, S.; Park, H.; Lee, K.; Hwang, W. Thin Solid Films 2006, 515 (4), 1932−1937. (26) Liu, L. F.; Lee, S. W.; Li, J. B.; Alexe, M.; Rao, G. H.; Zhou, W. Y.; Lee, J. J.; Lee, W.; Gosele, U. Nanotechnology 2008, 19, 495706. (27) Brumlik, C. J.; Martin, C. R. J. Am. Chem. Soc. 1991, 113, 3174. (28) Thompson, G. E.; Furneaux, R. C.; Wood, G. C.; Richardson, J. A.; Goode, J. S. Nature 1978, 272, 433. (29) Thompson, G. E.; Wood, G. C. Nature 1981, 290, 230. (30) Guo, Y.-G.; Lee, J.-S.; Hu, Y.-S.; Maier, J. J. Electrochem. Soc. 2007, 154, K51. (31) Agrawal, R. C.; Gupta, R. K. J. Mater. Sci. 1999, 34, 1131. (32) Burley, G. Acta Crystallogr. 1967, 23, 1.

We attributed the suppression of phase transition not only to the size and interface effect but also to the presence of defects and the accompanying charge imbalance induced by AAO membranes. Superionic phase stabilization and the significantly enhanced conductivity in AAO membranes at a practical temperature presented here has established foundations for other superionic conductors that are applied in solid-state microbatteries, high-sensitive sensors, and fuel cells as well as high-density nanodevices. In addition, the significant effects of size and the interface are crucial in systems of superionic or mixed conductors, especially for high performance functional devices in which the structures are controlled on the nanoscale level. Moreover, the ultimate integration of highly ordered arrays in nanowire structures with tunable sizes down to 10 nm should facilitate a wide range of fundamental studies on optoelectronic and electrochemical nanodevices.



ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.nanolett.5b01388. Experimental methods and additional figures that support the characterization (PDF)



AUTHOR INFORMATION

Corresponding Authors

*Tel: +81-29-860-4383. Fax: +81-29-860-4863. E-mail: zhang. [email protected]; [email protected]. *E-mail: [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS We thank M. Nakatsu for the assistance in DSC and in situ XRD measurements. This work was supported by the World Premier International (WPI) Center for Materials Nanoarchitectonics (MANA), and the National Key Basic Research and National Nature Science Foundation of China (No. 2014CB931704, 11204308, 11174287, and 51371166).



REFERENCES

(1) Chadwick, A. V. Nature 2000, 408, 925. (2) Hull, S. Rep. Prog. Phys. 2004, 67, 1233. (3) Terabe, K.; Hasegawa, T.; Nakayama, T.; Aono, M. Nature 2005, 433, 47. (4) Liang, C. H.; Terabe, K.; Hasegawa, T.; Negishi, R.; Tamura, T.; Aono, M. Small 2005, 1, 971. (5) Takada, K.; Kanbara, T.; Yamamura, Y.; Kondo, S. Solid State Ionics 1990, 40−41, 988. (6) Murugaraj, R.; Govindaraj, G.; Ramasamy, S. J. Power Sources 2002, 112, 184. (7) Delaizir, G.; Manafi, N.; Jouan, G.; Rozier, P.; Dolle, M. Solid State Ionics 2012, 207, 57. (8) Shahi, K.; Wagner, J. B. Appl. Phys. Lett. 1980, 37, 757. (9) Bradley, J. N.; Greene, P. D. Trans. Faraday Soc. 1967, 63, 424. (10) Tatsumisago, M.; Shinkuma, Y.; Minami, T. Nature 1991, 354, 217. (11) Yamada, H.; Bhattacharyya, A. J.; Maier, J. Adv. Funct. Mater. 2006, 16, 525. (12) Lee, J. S.; Adams, S.; Maier, J. J. Electrochem. Soc. 2000, 147, 2407. (13) Shahi, K.; Wagner, J. B. J. Electrochem. Soc. 1981, 128 (1), 6−13. G

DOI: 10.1021/acs.nanolett.5b01388 Nano Lett. XXXX, XXX, XXX−XXX