Exceptional Flux Growth and Chemical Transformation of Metastable

Subscriber access provided by UNIV NEW ORLEANS

Article

Exceptional Flux Growth and Chemical Transformation of Metastable Orthorhombic LiMnO Cuboids into Hierarchically-Structured Porous H Mn O Rods as Li Ion Sieves 2

1.6

1.6

4

Fumitaka Hayashi, Shoichi Kurokawa, Hiromasa Shiiba, Hajime Wagata, Kunio Yubuta, Shuji Oishi, Hiromasa Nishikiori, and Katsuya Teshima Cryst. Growth Des., Just Accepted Manuscript • DOI: 10.1021/acs.cgd.6b00223 • Publication Date (Web): 26 Sep 2016 Downloaded from http://pubs.acs.org on September 28, 2016

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Crystal Growth & Design is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Crystal Growth & Design

Cover page Exceptional Flux Growth and Chemical Transformation of Metastable Orthorhombic LiMnO2 Cuboids into Hierarchically-Structured Porous H1.6Mn1.6O4 Rods as Li Ion Sieves Fumitaka Hayashi,†,‡ Shoichi Kurokawa,† Hiromasa Shiiba,† Hajime Wagata,† Kunio Yubuta,§ Shuji Oishi,†,ǁǁ Hiromasa Nishikiori,† Katsuya Teshima†,‡,ǁǁ,┴,* †

Department of Environmental Science and Technology, Faculty of Engineering, Shinshu University, 417-1 Wakasato, Nagano 380-8553, Japan ‡ Global Aqua Innovation Center, Shinshu University, Wakasato, Nagano 380-8553, Japan § Institute for Materials Research, Tohoku University, Sendai 980-8577, Japan ǁǁ Nagano Prefecture Nanshin Institute of Technology, 8304-190 Minamiminowa, Kamiina, Nagano 3994511, Japan ┴ Center for Energy and Environmental Science, Shinshu University, 4-17-1 Wakasato, Nagano 380-8553, Japan

ABSTRACT

Selective lithium uptake from sea water and lake brine is an important challenge in energy and environmental science. H1.6Mn1.6O4 with pseudo-spinel type structure is a highly selective adsorbent for Li ions, but it is difficult to prepare large, highly crystalline H1.6Mn1.6O4 crystals with porous structure due to its thermodynamic metastability. Herein we demonstrate simple chemical processes that transform flux-grown, idiomorphic orthorhombic LiMnO2 (o-LiMnO2) cuboids of micrometre size into hierarchically structured H1.6Mn1.6O4 rods. We have optimized the flux growth conditions such as the Mn source, holding temperature, and solute concentration, in order to yield large, single phase o-LiMnO2 particles. The use of MnO under very low solute concentration (1 mol %) and high temperature (1000 °C) is critical to obtaining the single phase, idiomorphic o-LiMnO2 cuboids. The metastability of o-LiMnO2 is confirmed by ab initio density functional theory calculation in comparison with other lithium manganates such as LiMn2O4 and Li2MnO3. The successive calcination and acid treatment allow the transformation of o-LiMnO2 into H1.6Mn1.6O4 rods with porous structure. The resultant H1.6Mn1.6O4 shows high Li+ adsorption capacity (~5.6 mmol g-1), high Li+/Na+ selectivity, and good durability compared with existing H1.6Mn1.6O4 adsorbents. * Corresponding author: Katsuya Teshima, Center for Energy and Environmental Science, Shinshu University, Wakasato, Nagano

ACS Paragon Plus Environment

1


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 24

380-8553, Japan. TEL: +81-26-269-5556; FAX: +81-26-269-5550; E-mail: [email protected]

Exceptional Flux Growth and Chemical Transformation of Metastable Orthorhombic LiMnO2 Cuboids into Hierarchically-Structured Porous H1.6Mn1.6O4 Rods as Li Ion Sieves Fumitaka Hayashi,†,‡ Shoichi Kurokawa,† Hiromasa Shiiba,† Hajime Wagata,† Kunio Yubuta,§ Shuji Oishi,†,ǁǁ Hiromasa Nishikiori,† Katsuya Teshima†,‡,ǁǁ,┴,* †

Department of Environmental Science and Technology, Faculty of Engineering, Shinshu University, 4-17-1 Wakasato, Nagano 380-8553, Japan

‡

Global Aqua Innovation Center, Shinshu University, Wakasato, Nagano 380-8553, Japan §

ǁǁ

Institute for Materials Research, Tohoku University, Sendai 980-8577, Japan

Nagano Prefecture Nanshin Institute of Technology, 8304-190 Minamiminowa, Kamiina, Nagano 399-4511, Japan

┴

Center for Energy and Environmental Science, Shinshu University, 4-17-1 Wakasato, Nagano 380-8553, Japan * Corresponding author, E-mail: [email protected]

ABSTRACT

Selective lithium uptake from sea water and lake brine is an important challenge in energy and environmental science. H1.6Mn1.6O4 with pseudo-spinel type structure is a highly selective


2

Page 3 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


adsorbent for Li ions, but it is difficult to prepare large, highly crystalline H1.6Mn1.6O4 crystals with porous structure due to its thermodynamic metastability. Herein we demonstrate simple chemical processes that transform flux-grown, idiomorphic orthorhombic LiMnO2 (o-LiMnO2) cuboids of micrometre size into hierarchically structured H1.6Mn1.6O4 rods. We have optimized the flux growth conditions such as the Mn source, holding temperature, and solute concentration, in order to yield large, single phase o-LiMnO2 particles. The use of MnO under very low solute concentration (1 mol %) and high temperature (1000 °C) is critical to obtaining the single phase, idiomorphic o-LiMnO2 cuboids. The metastability of o-LiMnO2 is confirmed by ab initio density functional theory calculation in comparison with other lithium manganates such as LiMn2O4 and Li2MnO3. The successive calcination and acid treatment allow the transformation of o-LiMnO2 into H1.6Mn1.6O4 rods with porous structure. The resultant H1.6Mn1.6O4 shows high Li+ adsorption capacity (~5.6 mmol g-1), high Li+/Na+ selectivity, and good durability compared with existing H1.6Mn1.6O4 adsorbents.

Introduction Lithium is an important element with a wide range of technological applications, including thermal resistant glass ceramics and lithium ion battery for various electric devices.1 The annual demand for lithium salts has kept increasing over the last decade, and is expected to reach 713 M ton in the near future2 mainly due to the growing need for battery manufacturing. Lithium salts are now produced from lake brine and hard minerals.3,4 Solar evaporation of lake brine is very simple but time-consuming, typically requiring 1.5 year to concentrate the lithium ions to 5000 ppm.2 In the case of minerals, after strong acid treatment the dissolved lithium ions are difficult to extract, due to the presence of interfering ions such as Mg2+. In addition, excess amount of


3


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 24

soda lime is needed to precipitate the resultant Li+ as lithium carbonate. Therefore, more efficient approaches for lithium recovery are being sought.5 One alternative is the use of a lithium ion sieve to recover Li+ from aqueous solutions. Up to now, several kinds of manganese oxides and titanates have been studied as lithium ion sieves, such as λ-MnO2,6 MnO2·0.3H2O,7 and MnO2·0.5H2O (H1.6Mn1.6O4)8,9 derived from LiMn2O4, Li1.33Mn1.67O4, and Li1.6Mn1.6O4, respectively as well as H2TiO3 (TiO(OH)2).10 The reported adsorption properties of manganate-based adsorbents are summarized in Table S1. It is well accepted that the pseudo-spinel type H1.6Mn1.6O4 shows sufficiently high selectivity and high ion exchange capacity (2.9–5.3 mmol g-1) for Li+ even at high concentrations of Na+.8,9 This remarkable property is attributed to the fact that the window size of its adsorption sites constructed by the oxygen octahedral/tetrahedral units is smaller than the ionic diameter of Na+ (0.204 nm) but larger than that of Li+ (0.152 nm). Namely, the steric size effect allows only Li+ to be trapped in the adsorption sites of H1.6Mn1.6O4. One practical problem in preparing H1.6Mn1.6O4 is that the precursor cannot be obtained through a single-step process. In other words, there is only a one-way transformation path to Li1.6Mn1.6O4 from the thermodynamically metastable orthorhombic LiMnO2 (o-LiMnO2). The reported pathways to Li1.6Mn1.6O4 are highlighted in Figure S1 in the Supporting Information.11 This limitation is due to the fact that lithium ions hardly occupy the octahedral sites of the spinel host. Among Li-Mn-O compounds, the spinel-type LiMn2O4 is the most thermodynamically stable, while o-LiMnO2 and monoclinic LiMnO2 (m-LiMnO2) are believed to be metastable.12-14 Up to now many researchers have explored preparation methods for o- and m-LiMnO2, such as solid state reaction, flux growth, ion exchange, and hydrothermal techniques.13-16 In these studies, the solid state reactions and flux growth did not yield the single phases of o- and m-


4

Page 5 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


LiMnO2. For example, Tang and co-workers reported the flux growth of o-LiMnO2 from LiClLiOH flux at 650 °C, but the products contained the impurity phase LiMn2O4 and the resultant crystals were small (~200 nm).13 Similar limitation was found in the report on the growth of oLiMnO2 from γ-MnOOH and LiCl flux at 1000 °C.16 On the other hand, the hydrothermal method enables the preparation of single phase o-LiMnO2 at approximately 140 °C.8,15 Unfortunately, the o-LiMnO2 particles were even smaller (~100 nm) and cannot be used for Li+ recovery due to adsorbent leaching problems. The hydrothermally prepared o-LiMnO2 particles were also round and less-developed, showing low crystallinity and low chemical stability. These o-LiMnO2 compounds could be transformed into pseudo-spinel Li1.6Mn1.6O4 upon calcination in air at 350-400 °C,8 followed by acid treatment to yield H1.6Mn1.6O4. However, the particle size of the resulting H1.6Mn1.6O4 crystals is still very small (~100 nm). In this study we report exceptional flux growth of micrometre-sized metastable o-LiMnO2 cuboids from chloride-based flux at 1000 °C, and their efficient transformation into macroporous H1.6Mn1.6O4 rods (HMOrods) with lithium ion sieving property. This is the first demonstration of anisotropic crystal growth of metastable mixed oxide at micrometre size using a flux method. The transformation to H1.6Mn1.6O4 involves only two simple procedures, namely calcination followed by acid treatment. The prepared H1.6Mn1.6O4 showed high Li+-exchange capacity, good Li+/Na+ selectivity, and good durability compared to existing H1.6Mn1.6O4 adsorbents.

Experimental Flux growth of o-LiMnO2 crystals. All reagents were purchased from Wako Pure Chemical Industries, Japan. Reagent grade MnO, Mn(NO3)2·6H2O, and MnCO3 were used as the Mn


5


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 24

source, while LiCl and KCl were used as the Li source and fluxes, respectively. The molar ratio of LiCl to KCl is 59.5:40.5. The solute concentration ranged from 1 to 25 mol%. The homogenous mixture of precursors and flux (~20 cm3 in volume) was tightly sealed using lids and heated under air in an electric furnace. A holding temperature of 600 or 1000 °C was applied for 0–10 h. The heating and cooling rates were 900 and 200 °C h-1, respectively. The resulting products were washed with hot water and dried at 65 °C. For comparisons, we also prepared oLiMnO2 crystals using the hydrothermal method reported by Chitrakar and coworkers.10 Conversion of o-LiMnO2 cuboids into macroporous H1.6Mn1.6O4 rods. The conversion of the o-LiMnO2 cuboid crystals to Li1.6Mn1.6O4 was performed by calcination in air at 400 °C according to literature.8 Then the Li1.6Mn1.6O4 crystals were immersed in 0.5 mol L-1 HCl solution for 1 day with shaking to prepare the porous H1.6Mn1.6O4, followed by washing with deionized water and drying at 65 °C. The hydrothermally synthesized o-LiMnO2 was calcined and acid-treated using the same procedure. Characterization. X-ray diffraction (XRD) patterns were collected with a MiniflexII diffractometer (Rigaku, Japan) with mono-chromated Cu Kα radiation (λ = 0.15418 nm, 30 kV, 20 mA). Field-emission scanning electron microscopy (FE-SEM) images with energy dispersive X-ray (EDX) spectrometry data were obtained with a JSM-7600F (JEOL, Japan). The crystal growth of o-LiMnO2 was characterized by high-resolution transmission electron microscopy (HR-TEM, TOPCON, EM-002B) operated at 200 kV. Thermogravimetric analyses were carried out using a Thermo plus EVOII (Rigaku, Japan). Ab initio calculations based on density functional theory (DFT) were performed for Li-Mn-O system using the Vienna Ab initio Simulation Package (VASP)17,18 with modified Perdew-Burke-Ernzerhof generalized gradient approximation (PBEsol-GGA)19,20 and the projector-augmented wave (PAW) method.21 The on-


6

Page 7 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


site Coulomb correction (GGA + U) was included for localized electronic states, and the U value was chosen to be 3.9 eV for Mn 3d states according to literature.22 The energy cut-off and kpoint mesh were 500 eV and >1000, respectively. The formation enthalpies of o-LiMnO2, LiMn2O4, and Li2MnO3 were calculated from following Eqns. (1)-(3):

MnO + Li O + O → orthrombic LiMnO

2MnO + Li O + O → LiMn O

(1) (2) (3)

MnO + Li O + O → Li MnO

The energy correction for O2 molecules reported by Wang and Ceder23 was used for all calculations. We evaluated the mean oxidation state of Mn species in the HMOrods as well as Li/Mn molar ratio using both the iodometric titration and ICP-AES methods, as described in the literature.24 In this experiment, 0.010 g sample was dissolved in hydrochloric acid and the evolved chlorine species was reduced in situ with potassium iodide. The liberated iodine was titrated with sodium thiosulfate solution. The proton exchange reaction under acidic conditions can be simply expressed by the following chemical reaction in Eqn. (4): Li1.6Mn1.6O4 + 1.6(1‒x) H+ → Li1.6xH1.6(1‒x)Mn1.6O4 + 1.6(1‒x) Li+ (4) The x values in Li1.6xH1.6(1‒x)Mn1.6O4 can be determined using the ICP-OES method below. The proton exchanged degree is expressed by Eqn. (5): Proton exchange degree of HMOrods (%) = 100 × (1‒x)

(5)

Li ion adsorption experiment. Li+ adsorption isotherms were measured by equilibrating 0.1 g of adsorbent sample in 100 mL of LiCl aqueous solution ([Li+]initial = 25–1000 ppm) with stirring at room temperature. The adsorbent was immersed in 20 or 50 mL of aqueous solution


7


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 24

containing alkali metal cations at 25 °C for 1 day with shaking. According to the chemical composition of “Salar de Atacama”,5 the Li, Na, Mg, and Ca ion concentrations in the solutions were controlled in the range between 0 and 73000 ppm, and the pH of the solutions was 4.6-9.6 in order to evaluate the selectivity of the adsorbents. The alkali metal ion concentrations in the supernatant were analyzed with inductively coupled plasma-optical emission spectrometry (ICPOES, SII, SPS5510) and ion chromatography (IC, Shimadzu, HIC-SP).

Results and Discussion


8

Page 9 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Flux growth of metastable o-LiMnO2. First, we explain the unique flux growth of o-LiMnO2 crystals, which is the precursor of Li1.6Mn1.6O4. We have examined the growth conditions, such as the Mn source and holding temperature. Figure 1 shows the XRD patterns of the products grown at 600 and 1000 °C using MnO, Mn carbonate, or Mn nitrate as Mn source, wherein the solute concentration is 5 mol%. At 600 °C, the main products were Mn3O4 and LiMn2O4 irrespective of the Mn source, and o-LiMnO2 was not produced at all. Increasing the holding temperature to 1000 °C did not yield o-LiMnO2 when Mn carbonate or nitrate was used. In

Figure 1. (a) XRD patterns of lithium manganate grown from MnO, Mn carbonate, and Mn nitrate together with that of reference o-LiMnO2 (PDF 00-035-0749). The holding temperatures are 600 and 1000 °C, with solute concentration of 5 mol%. (b) Variation in XRD patterns of lithium manganate with changing solute concentrations. Mn Source, MnO; temperature, 1000 °C.


9


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 24

contrast, the use of MnO at 1000 °C produced the o-LiMnO2 phase, although there were impurities such as LiMn2O4 and Mn3O4 phases. Figure 1b shows the effect of solute concentration on the XRD patterns of the product grown at 1000 °C using MnO. The XRD patterns of samples varied greatly at different solute concentrations. Notably, decreasing the concentration to 1 mol% from 5 or 25 mol% led to single phase o-LiMnO2. This finding indicates that the dissolution of MnO in the flux could be an important step in producing the pure phase of o-LiMnO2. The particle morphology and crystal growth were studied by FE-SEM and TEM analyses. Figure 2a shows the FE-SEM image of o-LiMnO2 crystals prepared at 1000 °C and 1 mol% solute concentration. The idiomorphic cuboid crystals were observed with 5-10 µm in size. Such unique particle shape has never been observed in lithium manganates.8,15 Besides, the crystals are 1-2 order(s) of magnitude larger than those prepared by the hydrothermal method (100–200 nm).8 These two characteristics demonstrate the advantages of our flux growth method. The marked difference between the present and previous studies could be attributed to the efficient crystal growth from the chloride-based flux at 1000 °C.


10

Page 11 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 2. (a) FE-SEM and (b), (c) TEM images and SAED patterns of o-LiMnO2 crystal grown at 1000 °C from MnO. The TEM image in (c) is the enlarged view of the point indicated by the arrow in (b). (d) XRD intensity ratio of {002} to {010} peak diffraction for o-LiMnO2 as a function of holding time. Solute concentration, 1 mol%. Figure 2b is the TEM image of o-LiMnO2 grown at 1000 °C and 1 mol% solute concentration, together with selected area electron diffraction (SAED) pattern, and Figure 2c shows the magnified view of TEM image in Figure 2b. The analysis was carried out using the reported crystal data: a = 0.2805 nm, b = 0.5757 nm, c = 0.4572 nm, and space group Pmnm.25 The SAED patterns were assigned to the orthorhombic system, which is observed from the [001] direction, indicating that the cuboid-like crystal is a single crystal elongated along [100] direction. The lattice fringe in the high-magnification TEM image was 0.579 nm, corresponding to the 010 spacing of o-LiMnO2 (0.5757 nm). This is a clear evidence of the high-crystallinity of flux-


11


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 24

grown o-LiMnO2 with no grain boundary. Both SEM and TEM results imply the anisotropic crystal growth of the o-LiMnO2. In order to understand such unique crystal growth, we studied the effect of holding time on the crystal growth of o-LiMnO2. The FE-SEM images are shown in Figure S2 at holding times of 0 h and 2 h. At 0 h, the particle shape was ill-defined and rather round, and the particle size was only about 1 µm. Increasing the holding time to 2 h (Figure S2b) and 10 h (Figure 2a) improved the particle morphology. Namely, the particle shape changed from spherical to rectangular cuboid, showing continued anisotropic crystal growth. Figure S3 shows the change in the XRD patterns of lithium manganates grown at 1000 °C with increasing holding time. The XRD intensity ratio of the (010) peak at 15.4° to the (002) peak at 39.4° for o-LiMnO2 is shown in Figure 2d as a function of holding time. The relative intensity of the (010) peak decreased with increasing holding time, while that of the (002) peak increased (Figure S3). As a result, the intensity ratio of the (010) to (002) diffraction peaks decreased with increasing holding time, again showing the efficient anisotropic crystal growth in LiCl-KCl flux at 1000 °C. The thermodynamic metastability of o-LiMnO is confirmed by ab initio DFT calculation, by comparison with other lithium manganates such as spinel-type LiMn2O4 and Li2MnO3. Figure 3 shows the crystal structures and formation enthalpies of o-LiMnO2, LiMn2O4, and Li2MnO3. (183.2, -395.6, and -349.0 kJ mol-1 respectively), as determined by DFT under 0 atom, -273.15 °C. The negative formation enthalpies indicate that o-LiMnO2, LiMn2O4, and Li2MnO3 can all be spontaneously formed from MnO and Li2O. The o-LiMnO2 has the highest formation enthalpy, indicating its thermodynamic metastability which is consistent with the results of XRD.


12

Page 13 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 3. Crystal structures of (a) orthorhombic LiMnO2, (b) LiMn2O4, and (c) Li2MnO3 based on DFT calculation. Purple, manganese; blue, lithium; red, oxygen. Based on the characterization results, we now discuss the possible growth mechanism of oLiMnO2 from the chloride-based flux at 1000 °C. Figure 4 schematically shows the growth pathways that yield o-LiMnO2. At the early stage of crystal growth, the starting material MnO is solvated and dissolved in liquid LiCl-KCl, resulting in the formation of Mn(II) species (Step 1, Figure 4). The generated Mn(II) is then oxidized by the dissolved oxygen gas, leading to the formation of anionic Mn(III)Ox cluster-like species. The resultant anionic MnOx species reacts with dissolved Li+ ion to form the o-LiMnO2 nucleus (Step 2). During the holding at 1000 °C, the o-LiMnO2 nucleus continues to grow through the successive reactions between anionic MnOx and Li+ along the [100] direction. There are three possible growth directions, namely [100], [010], and [001]. The [100]-direction growth is considered to be correlated with the crystal structure of o-LiMnO2. Note that the (100) and (001) faces consist of mixtures of MnO6 and LiO6 layers, while the (010) face consists of either MnO6 or LiO6 layer. In the presence of excess Li+ ion, the [010] growth of o-LiMnO2 is not feasible because the anionic MnOx and cationic Li+ would preferentially react with each other, rather than between same species due to the electrostatic repulsion. Taking into account the arrangement of oxygen ions on the top (100)/(001) surfaces of o-LiMnO2, Li+ and MnOx would prefer the (100) surface because of the


13


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 24

larger contact areas, as demonstrated in Figure S4. It should be added that the preference for the [100] growth direction agrees well with the line defect formation shown in the inset of Figure 2c.

Figure 4. Schematic illustration of possible formation mechanism for orthorhombic LiMnO2 at 1000 °C. Conversion of o-LiMnO2 cuboids to pseudo-spinel type H1.6Mn1.6O4 rods (HMOrods). Figure 5a shows the XRD patterns of o-LiMnO2 before and after the calcination at 400 °C under air.

Figure. 5 (a) XRD patterns of orthorhombic LiMnO2 before and after calcination at 400 °C under air. The patterns of o-LiMnO2 (PDF 00-035-0749) and Li1.6Mn1.6O4 (PDF 00-052-1841) are shown as references. (b) TG-DTA profile of o-LiMnO2.


14

Page 15 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


After calcination the peaks from the o-LiMnO2 phase were diminished, while the peaks assignable to pseudo-spinel type Li1.6Mn1.6O4 appeared, although there were also some weak unknown peaks. This marked transformation is well consistent with the previous results for oLiMnO2 prepared by the hydrothermal method.8 The thermogravimetry-differential thermal analysis (TG-DTA) profile for the o-LiMnO2 sample is shown in Figure 5b. Below 400 °C the sample weight monotonously decreased with increasing temperature, and then starts to increase after 400 °C as a result of the oxidation of o-LiMnO2. The amount of weight increase between 400 and 800 °C was 6.1 wt %, which could be explained by oxidation by gaseous O2: 2LiMnO2 + 1/2O2 →5/4 Li1.6Mn1.6O4

(6)

This weight increase agrees roughly with the theoretical value (8.6 wt%) from Eqn. (6). The mean oxidation state of Mn species in LMO was analyzed using both ICP and iodometric titration. For HMOrods, the Li/Mn ratios of LiMnO2, Li1.6Mn1.6O4, and H1.6Mn1.6O4 were 1.01, 1.01, and 0.05 (Table S2), which were almost the same as their stoichiometric values. The mean oxidation states of the Mn species in the LiMnO2, Li1.6Mn1.6O4, and H1.6Mn1.6O4 samples were found to be 3.05, 3.89, and 3.89 (Table S2), which agree well with the respective formal valences. Figures 6a and 6b show SEM images of the resulting Li1.6Mn1.6O4. The particle size of Li1.6Mn1.6O4 crystals changed little after calcination, but the shape of particles became round from cuboidal, due to the phase transition of o-LiMnO2 (Figures 6a and 6b vs. Figure 2a).


15


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 24

Figure 6. FE-SEM images of (a), (b) calcinated Li1.6Mn1.6O4 and (c), (d) H1.6Mn1.6O4 rods formed during acid treatment. TEM images and SAED patterns of (e) Li1.6Mn1.6O4 and (f) porous H1.6Mn1.6O4 rods. Next, the obtained Li1.6Mn1.6O4 was immersed in acid to prepare the H1.6Mn1.6O4 adsorbent (HMOrods). Figure S5 shows the XRD patterns of Li1.6Mn1.6O4 before and after the acid immersion. The diffraction peak due to the (111) facet shifted from 18.62° to 19.08° without


16

Page 17 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


significant changes in the peak shape, as a result of replacement of Li+ by H+. Elemental analysis indicated that the degree of H+ exchange was >95%. The lattice constant for the prepared HMOrods is calculated to be 0.806 nm, which is nearly the same as that (0.805 nm) for existing H1.6Mn1.6O4 adsorbents,8 indicating the pseudo-spinel structure was retained. Figures 6c and 6d show the FE-SEM images of the prepared HMOrods. The particle size (5-10 µm) of HMOrods was not changed through the immersion in the acid, but the unique microarchitecture of crystalline rods 100 nm in diameter and 1-dimensional interstitial pores was formed afterwards. This unique macroporous structure has never been observed in the reported H1.6Mn1.6O4 crystals prepared by the hydrothermal method.8,9 Nitrogen adsorption isotherms on the present H1.6Mn1.6O4 as well as Li1.6Mn1.6O4 are shown in Figure S6. The rapid increase at 0.8 of relative partial pressure was found in the isotherm of H1.6Mn1.6O4, again indicating the formation of macropore structure through the immersion in the acid. The BET surface area of the H1.6Mn1.6O4 adsorbent was 7 m2 g-1, one order of magnitude greater than that of the precursor Li1.6Mn1.6O4 (0.9 m2 g-1). This significant increase in the surface area is due to the macroporous structure formed through the acid treatment. For comparison, H1.6Mn1.6O4 crystals (abbreviated as HMOhydro) were synthesized according to the literature8, using o-LiMnO2 prepared by the conventional hydrothermal method. Figure S7 shows the XRD patterns of HMOhydro and the precursors o-LiMnO2 and Li1.6Mn1.6O4. Single phase of H1.6Mn1.6O4 was obtained, which is well consistent with the previously reported result.8 Figure S8 shows the high and low magnification FE-SEM images of HMOhydro. The particles were aggregates of H1.6Mn1.6O4 nanocrystals, 100–200 nm in size and with no ordered porosity in the bulk. This significant difference between HMOhydro and HMOrods could be due to the crystallinity and primary particle size of Li1.6Mn1.6O4.


17


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 24

Figures 6e and 6f show the TEM images of Li1.6Mn1.6O4 and H1.6Mn1.6O4 together with the SAED patterns along the [1-10] and [100] directions respectively, wherein the thin part of crystal was analyzed. Both SAED patterns were well matched with the expected patterns from the (pseudo)spinel type structure with cubic symmetry, which is in good agreement with the XRD results in Figure S6. The absence of additional spots in the patterns indicates the single crystalline nature of crystals. Therefore, the flux method enables the preparation of hierarchically structured H1.6Mn1.6O4 rods with macropores from the flux grown o-LiMnO2 cuboids. Herein we discuss the formation mechanism of the unique morphology of HMOrods. The parent Li1.6Mn1.6O4 possessed the spinel structure, while the rod-like shape of HMOrods particles was not related to the three-dimensional spinel structure. In addition, the formed pore size in H1.6Mn1.6O4 rods was ~50–100 nm. Therefore its morphology could be due to the selective dissolution of less-crystalline Li2O component of Li1.6Mn1.6O4 through the immersion in the acid. Similar phenomena have been observed in Li2MnO3 by Tang and coworkers.26

Selective adsorption of pseudo-spinel type H1.6Mn1.6O4 rods. We studied the Li+ adsorption property and chemical stability of the macroporous HMOrods. First, the Li+ exchange capacity (CECLi) was evaluated through the Li+ adsorption isotherm under equilibrium conditions. Figure 7 shows the isotherm at room temperature, which was analyzed using the Langmuir and Freundlich models. The Freundlich model fits the data better than the Langmuir model, and the simulated curve is indicated in Figure 7. The Freundlich constant was found to be 0.25, indicating the strong interaction between the HMOrods and adsorbed Li+. Based on the measured isotherm, CECLi of the HMOrods was determined to be about 5.6 mmol g-1.


18

Page 19 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 7. Equilibrium adsorption isotherm for adsorption of Li+ on HMOrods. The curve represents the fitting result with Freundlich isotherm model. Conditions: pH ~10, volume/mass ratio ~1000 mL·g−1, [Li+]initial = 25–1000 ppm, room temperature.

The kinetic curve for Li+ adsorption on HMOrods was then measured at the low Li+ concentration of 30 ppm. From the results shown in Figure S9, the Li+ ion-exchange reaction reached equilibrium in ~20 h, which showed that the duration of 1 day for obtaining the adsorption isotherm in Figure 7 is reasonable. Next, we studied the selective adsorption property of HMOrods and HMOhydro by controlling the Li+:Na+:Mg2+:Ca2+ ion ratios in the solution used for the adsorption experiments. According to the chemical composition of “Salar de Atacama,”5 the Na, Mg, and Ca ion concentrations in the aqueous solution were up to 76000, 9600, and 400 ppm, respectively. The results were summarized in Table S3, Runs 1-4. In the absence of Mg and Ca ions, the HMOrods prepared here adsorbed 5.2 and 4.9 mmol g-1 of Li+ (Runs 1 and 2), when the Na+ concentrations were 3320 and 76000 ppm, respectively, both being about 50% of the theoretical maximum (10.8 mmol g-1). The amounts of Na+ were less than ~0.5 % of that of Li+, which agrees well with previously reported results.8 When Mg2+ and Ca2+ were present in the solution, the Li+ exchange capacity


19


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 24

was reduced to 3.8 mmol g-1, mainly because Mg2+ competes for the adsorption sites on HMOrods (Run 3). Similar insertion of Mg2+ ions into the manganate crystal lattice was observed on the λMnO2 electrode,5 because the ionic diameter of Mg2+ (0.144 nm) is smaller than that of Li+ (0.152 nm). The Li+ exchange capacity of HMOhydro was almost the same as that of HMOrods (Run 4 vs. Run 3), indicating little effect of the porosity on the selective Li+ recovery. Table S1 summarizes the comparison of the present result with those from previous studies.8,9,27,28 The Li+ adsorption values are comparable to or higher than those for existing Li ion sieves (2.9–5.3 mmol g-1; Table S1, entries 1-4), indicating high Li+ adsorption capacity and selectivity. This marked selectivity for Li+ could be attributed to the small window (~0.18 nm) of the adsorption site, which is slightly smaller than the ionic size (~0.20 nm) of Na+. Last but not least, the ion uptake capacities for existing cation exchangers, such as clays and layered metalates, are generally

Exceptional Flux Growth and Chemical Transformation of Metastable

Recommend Documents