Ethylenediamine-Enabled Sustainable Synthesis of Mesoporous

Subscriber access provided by University of Newcastle, Australia

Ethylenediamine-enabled sustainable synthesis of mesoporous nanostructured Li2FeIISiO4 particles from Fe(III) aqueous solution for Li-ion battery application Huijing Wei, Xia Lu, Hsien-Chieh Chiu, Bin Wei, Raynald Gauvin, Zachary Arthur, Vincent Emond, De-Tong Jiang, Karim Zaghib, and George P. Demopoulos ACS Sustainable Chem. Eng., Just Accepted Manuscript • DOI: 10.1021/ acssuschemeng.8b00090 • Publication Date (Web): 22 Apr 2018 Downloaded from http://pubs.acs.org on April 22, 2018

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 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 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.

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 25 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

ACS Sustainable Chemistry & Engineering

Ethylenediamine-enabled sustainable synthesis of mesoporous nanostructured Li2FeIISiO4 particles from Fe(III) aqueous solution for Li-ion battery application Huijing Wei1, Xia Lu1, 2, Hsien-Chieh Chiu1, Bin Wei1, Raynald Gauvin1, Zachary Arthur3, Vincent Emond3, DeTong Jiang3, Karim Zaghib4, and George P. Demopoulos1,*

1. Materials Engineering, McGill University, 3610 University Street, Montréal, Québec H3A 0C5, Canada. 2. State Key Laboratory of Organic-Inorganic Composites, College of Energy, Beijing University of Chemical Technology, Beijing 100029, China. 3. Department of Physics, University of Guelph, Guelph, Ontario N1G 2W1, Canada. 4. Centre d’excellence-ETSE, Hydro-Québec, Varennes, Québec J3X 1S1, Canada. *Corresponding author: Prof. G. P. Demopoulos, Email: [email protected]

Abstract: Engineering of nanostructured lithium iron silicate (LFS) particles is pursued via a novel benign synthesis approach seeking to understand the crystalline particle formation process and its impact on energy storage capacity. Specifically, mesoporous Li2FeSiO4 nanostructured particles are synthesized via a novel dual-step process involving organic-assisted hydrothermal precipitation from concentrated Fe(III) (1 mol/L) aqueous solution followed by reductive (5 vol.% H2) thermal transformation of the precipitate at 400°C (LFS400) and 700°C (LFS700). Scanning and transmission electron microscopy revealed the formation of secondary submicron sized porous agglomerates of unitary primary nanocrystals (~ 50nm for LFS400 and ~200nm for LFS700). Both ethylene glycol and ethylenediamine are used as crystallization control additives. It is demonstrated that formation of LFS from Fe(III) precursor is made possible only by the action of ethylenediamine.

The obtained LFS particles are found to be predominantly

monoclinic as per X-ray diffraction and Rietveld refinement and bear an in situ formed N-doped carbon coating layer as characterized by X-ray photoelectron spectroscopy. TEM coupled with selected area electron diffraction (SAED) analysis confirmed the Rietveld Refined XRD phase compositions. The reductive annealing-induced phase transformation sequence leading to LFS crystallization is characterized and the enabling role of ethylenediamine is discussed. Initial galvanostatic charging-discharging and cyclic voltammetry measurements indicate the annealing

ACS Paragon Plus Environment

1

ACS Sustainable Chemistry & Engineering 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 25

temperature of LFS formation to influence the Li-ion storage profile as it shifts from two-phase reaction in LFS700 to solid solution in LFS400-this being attributed to nanostructural changes. Keywords: Li2FeSiO4, Hydrothermal synthesis, Reductive annealing, Crystallization control, Ethylenediamine, Li-ion batteries Introduction Rechargeable lithium-ion batteries (LIBs) have powered over the past two decades the tremendous proliferation of portable electronics and now have started playing a key role in powering a shift towards electric vehicle transportation.1-3 This shift imposes new LIB requirements, namely high energy density (for longer driving range), good rate performance (for higher power), long cycle life, but also high safety factor, and low cost scalable and sustainable synthesis.3-4 Towards this end, lithium metal orthosilicates are of great interest because of their two lithium-ion storage per formula unit (Li2FeSiO4) that could provide a theoretical capacity that is double (330 mAh/g) of that of LiFePO4 (LFP),5-6 corresponding to a specific energy of 1,120 Wh/kg surpassing all current cathode materials.7

Although Li2FeSiO4 has similar

polyoxyanion framework as that of LFP,8 the existence of different polymorphs of lithium metal silicates have hampered their development and performance as cathode materials in Li-ion batteries.6 For example, in the case of Li2FeSiO4 (LFS) the different polymorphs are predominantly associated with two main crystal structures, monoclinic and orthorhombic, which due to very close formation energies tend to coexist,9-10 not to mention their significant degree of crystal disorder and defects by which they are plagued from.11 As such their synthesis and subsequent phase-dependent evaluation of their electrochemical performance is highly complicated.10, 12-15 For making progress with the development of these potentially high energy density Li-ion storage materials thus it becomes important to understand the underlying crystallization mechanism of LFS phase formation and ultimately control via the advancement of new synthesis routes. Equally important in this pursuit is that the new synthesis routes are sustainable in nature to allow for ultimate scalable clean production. Over the last 10 years, different methods have been applied to synthesis of lithium iron silicates among which notable are solid state synthesis14, 16 and several solution methods.8, 13, 17-18 Conventional solid-state synthesis,16 although simple, tends to produce rather non-homogeneous micrometer-sized particles,7 a problem partly solved by mixing the precursor chemicals by high-


2

Page 3 of 25 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


energy planetary ball milling prior to calcination.14 Solution synthesis (or “soft chemistry”) methods are better suited to achieve the necessary nanocrystal control.7 It is important however as argued by Zaghib et al.5 to ensure the feasibility of the synthesis method in producing “…many tons of these nanostructured materials to prepare batteries for electric vehicles at a reasonable price”. In this context complex solution synthesis routes like supercritical sol-gel reaction producing nanosheet crystals19 may not offer a practical option. Alternatively, hydrothermal synthesis5, 8 as discussed by Toyota’s J. Chen20 offers a truly sustainable process option. Use of functional organic additives is a powerful way to regulate the nucleation, aggregation and growth of the hydrothermally formed particles.21 Furthermore combining solution precipitation with thermal transformation/annealing offers flexibility in producing different phases,13 and eliminating Li-Fe anti-site defects.8

Also, because of the very low

conductivity of lithium metal silicate crystals, carbon coating is commonly pursued via the use of organic chemicals during the solution synthesis22 or during post heat-treatment.23. The properties of resultant particles, such as crystal structure, size, habit, purity, and porosity are greatly influenced by the path of crystallization. As a consequence, the method of synthesis–particle formation, can directly impact the electrochemical performance, i.e. intercalation properties of resultant Li2FeSiO4 cathode material. To date, most of the previous solution synthesis studies have made use of ferrous (Fe(II)) salts as the iron precursor followed by high-temperature annealing in an inert atmosphere. The use of Fe(II) as precursor, necessitates oxygen-free handling that greatly complicates the synthesis process. Developing a method starting with ferric (Fe(III)) salts as precursor–ferric salts are significantly less expensive than their ferrous counterparts, can greatly simplify the process rendering it more amenable to large scale production. To our knowledge no solution based method for the synthesis of LFS from a ferric salt as precursor has been reported other than some rather complex processing routes involving the use of tri-block copolymers as in situ template that are not suitable for large scale synthesis.22, 24

It is the scope of the present work to report on a novel ferric salt-based synthesis method for lithium iron silicate involving ethylenediamine-controlled hydrothermal precipitation of an intermediate, which upon annealing in a reducing atmosphere converts to mesoporous C-coated nanostructured LFS particles suitable for fabrication of Li-ion battery cathodes.


3


Page 4 of 25

Experimental Hydrothermal precipitation: The synthesis method involved hydrothermal preparation of an intermediate precipitate in the presence of ethylene glycol and ethylenediamine that was subsequently transformed thermally to LFS via reductive annealing. The method was adapted from the work of Goodenough’s group on the synthesis of LiFePO4 microspheres.25 The precursors were initially added in deionized water (194 mL) in stoichiometric amounts, typically in the following concentrations: 0.1936 mol of Fe(NO3)3.9H2O (Sigma-Aldrich 99.95%), 0.4065 mol of CH3COOLi (Sigma-Aldrich: 99.95%), and 0.1935 mol of fumed silica (Sigma-Aldrich, 0.2-0.3 µm average particle size). After the solution was well homogenized the two organic reagents were added, namely, 10.4 mL (0.1865 mol) of ethylene glycol (EG, Sigma-Aldrich, 99.8% anhydrous) and 10.4 mL (0.1557 mol) of ethylenediamine (EN, Sigma-Aldrich 99.5%). Upon their addition and thorough mixing in ultrasonic bath, a brownish suspension with pH of 8 – 9 was obtained, which was then transferred to a Teflon-lined stainless-steel autoclave (Parr 4567) stirred at 300 rpm. The detailed precursor solution preparation procedure is given in the Supplementary Information. Hydrothermal reaction was carried out at 180°C for 12 hours. The resultant slurry was dried in air (i.e. water got evaporated) on an 80°C stirring hotplate to obtain an amorphous/poorly crystalline intermediate product that was subsequently used in the annealing experiments. Reductive Annealing: The annealing step was carried out in a tube furnace (Carbolite GHA) under a constant flow of reducing Ar/H2 gas mixture (95:5). The initial step involved purging with the reducing gas for one hour at room temperature. Then the temperature was raised to 200°C at a rate of 3°C/min, where it was kept for 3 hours before the temperature was increased again at 3°C/min to a target temperature of 400°C or 700°C (namely, LFS400 and LFS700) and kept for a pre-selected amount of time (typically 6 hours) followed by natural cooling process to room temperature. Characterization: The characterization analyses included: X-ray diffraction (Bruker D8 Discover) with Co Kα source of radiation (λ ~ 1.78901 Å). Additionally, some XRD spectra were collected with a Philips diffractometer with Cu source and some were collected using synchrotron hard X-rays at Canadian Light Source Ltd. (λ ~ 0.6886 Å). The XRD spectra were


4

Page 5 of 25 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


analyzed using Rietveld refinement by TOPAS Total Pattern Analysis for determination of the phase composition (i.e. the weight fraction of monoclinic vs. orthorhombic phase). Other characterization methods included scanning electron microscopy (Hitachi SU-8230) and transmission electron microscopy (FEI F20); X-ray photoelectron spectroscopy (K-Alpha XPS, Thermal Fisher Scientific Inc.) and Raman spectroscopy (Bruker Senterra dispersive Raman microscope) for surface characterizations; thermogravimetric analysis for determining the carbon weight content (TA500 thermogravimetric analyzer) in a mixture of N2/O2 (60:40) atmosphere; inductively coupled plasma optical emission spectroscopy (ICP-OES, Thermo Scientific iCAP 6500 ICP Spectrometer) for solution analysis of the acid-digested powders; and Fe2+/Fe3+ redox titration. Electrode preparation and cell testing: The electrode paste was prepared by mixing the active material (LFS), carbon black and binder polyvinylidene fluoride (PVDF) in a typical ratio of 8:1:1, using N-Methyl-2-pyrrolidone (NMP) as a solvent. The pristine LFS400 nanomaterial was used as-is for making the paste. By contrast the LFS700 material because of its micron-sized particles was firstly ball milled for three hours at 250 RPM in a high energy planetary ball milling machine (Fritsch, Pulverisette 7) in the presence of 10 wt.% of carbon black, using ZrO2 beads as grinding media and isopropanol as a solvent. The ball milled LFS700 material was obtained after filtration and drying under vacuum (namely, LFS700BM). After mixing the LFS with carbon black and PVDF, the slurry paste was then cast on aluminum foil and dried in vacuum oven overnight before being punched for further drying and coin cell assembly. In the half cells, lithium metal was used as the anode, LiPF6 in EC/DMC (1:1) was used as the electrolyte, and polypropylene film (Celgard2200) was used as the separator between two electrodes. The coin cells were then tested using the Arbin multi-channel cycler for cycling and Bio-Logic VSP potentiostat for cyclic voltammetry (CV). Charging-discharging of the two electrodes was done at a rate of C/10 (1C=166 mAh/g) and CV at a scan rate of 0.04 mV/s over the first three cycles between 1.50 to 4.67 V. Electrochemical testing was done at elevated temperature 45 to 55°C instead of room temperature because of LFS’ poor intrinsic electronic (6x10-14 S/cm)8 and ionic26 conductivities. Results and Discussion LFS crystallization sequence


5


Page 6 of 25

Hydrothermal precipitation: We investigated the reactions occurring during the hydrothermal step by analyzing the formed precipitate and final solution. It was found that essentially all of Si (> 99.3%) and Fe (100%) had precipitated during hydrothermal reaction step but only about 25% of the Li had precipitated. The remaining 75% of the Li staying in solution (Table S1 in Supplementary Information) precipitated as salt during subsequent solution evaporation/drying treatment. The solution pH remained essentially the same (pH 8-9) before and after the hydrothermal treatment. Titration analysis of the acid-digested precipitate, done in glovebox to ensure no re-oxidation of any Fe2+ present (refer to Supporting Information), showed negligible signs of reduction from Fe3+ to Fe2+ (< 0.2%). This finding differs from earlier work25 involving the lithium iron phosphate system, where it was reported Fe3+ to have been reduced to Fe2+ during hydrothermal treatment-the reduction been attributed to the action of ethylene glycol (EG) but with no direct analysis been provided. It is quite likely this difference between the two systems to reflect stronger Fe-O bond strength in silicate vs. phosphate anion. After the evaporative drying step at 80°C, the solids were subjected to XRD analysis. The XRD pattern (Figure S1a in Supplementary Information) shows a rather amorphous material with no sharp diffraction peaks but only two broad peaks at around ~ 30° and ~ 63°, respectively. These broad peaks seem to overlap with the ferric oxy-hydroxide known as 2-line ferrihydrite (FeOOH·H2O);27 in this case though at all likelihood they correspond to amorphous hydrated III

iron (III) silicate, Fe (Si O )(OH) ·2H O, known as hisingerite that also exhibits two similar 2

broad peaks.

28

2

5

4

2

As for the 25% of Li that co-precipitated with the amorphous iron(III)-silica

precipitate, it seemed to be in an adsorbed state favored by the moderately alkaline pH environment. Following the evaporation of the solution (drying at 80°C) the remaining soluble fraction (75 %) of Li crystallized out as lithium acetate into the hydrous ferric silicate matrix (hisingerite) that serves as LFS precursor during subsequent thermal reductive annealing. In other words no direct crystallization of Li2FeSiO4 had occurred during the hydrothermal processing step when ferric iron was used as the source of iron, in contrast to the case of hydrothermal synthesis of ferrous iron-derived LFS.13, 17, 29 It was also found from FTIR analysis (Figure S2 in Supplementary Information) that as indicated by the presence of characteristic CO, C-H, N-H and O-H bending/stretching modes, the organic additives EN and EG are integral building components of the LFS intermediate via their known action as bidentate complexing ligand30 and surface active agent31 respectively.


6

Page 7 of 25 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


Reductive annealing-nucleation and transformation: After water evaporation, the organic additive-assembled hydrous ferric silicate intermediate was subjected to reactive thermal transformation/annealing in a 5% H2 reducing atmosphere at increasing temperature and time to determine its crystallization pathway. First XRD characterization after 3 hours holding time at 200°C was done (Supplementary Information of Figure S1b). As it can be seen, the pattern remained predominantly amorphous, but this time weak peaks seemed to start emerging in the region 24° to 37°, where key peaks of crystalline LFS reside. This can be interpreted as evidence of the early stage organization of Li-FeII-SiO4 constituents into nuclei of LFS by the reducing action (FeIII FeII) of H2. TEM analysis (image shown in Figure S3a in Supplementary Information) revealed indeed the onset of LFS nucleation, as judged from some barely-seen broken diffraction rings at the centre of the SAED pattern (shown in Figure S3b in Supplementary Information). Following this early annealing step at 200°C, the temperature was further increased to 400°C where it was held for different times, 1 min, 1 hr, 3 hrs and 10 hrs. The XRD patterns of Li2FeSiO4 annealed at different times at 400°C were shown in Figure 1. By examining the XRD pattern of the material annealed at 400°C for 1 minute, it can be seen that crystalline Li2FeSiO4 had already formed out of the nucleation cluster observed at 200°C (Supplementary Information of Fig. S1b). Comparing the XRD patterns at different annealing times it is seen that the full width half maximum (FWHM) decreases as the annealing time increases (Fig 1b), which is an indication of improved crystallinity as a result of diffusion-dependent crystal growth and refinement (crystal ordering). It is also noted that the main crystal phase did not change with increasing annealing time. The monoclinic (112) peak was observed for all samples, as labeled in Figure 1, although due to nanocrystalline nature some small peak features were not fully resolved. By examining the TEM images and selected area diffraction patterns (Figure S4-S7 in Supplementary Information), the predominant presence of monoclinic crystalline LFS phase was observed in all samples annealed from 1 minute up to 10 hours at 400°C. As annealing time increases the small nuclei grow and bond into each other to form polycrystalline particles. A thin amorphous layer with about less than 3-5 nanometer thickness was observed in all samples (Figure S4-S7 in Supplementary Information), which most likely represents the in situ formed carbon layer from the decomposition of the organics as further discussed in the next section.


7


From this first XRD analysis the co-presence of the LFS orthorhombic phase cannot be ruled out. It is also worth noticing that small impurity peaks (possibly Fe3O4 and Li2SiO3, labelled with stars) were present at shorter annealing time that disappeared as the annealing time increased to 10 hours, indicating the formation of impurity-free LFS nanocrystals. (a)

(b) (112)

(112)

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 25

* *

* *

Fig. 1. (a) XRD patterns (synchrotron source, λ=0.6886 Ǻ) for Li2FeSiO4 samples annealed at 400°C for 1 min, 1 hour, 3 hours and 10 hours. (b) A zoomed in view of the XRD patterns in (a) focusing on the range from 10-16°. * indicates presence of minor impurities.

It must be underlined that during annealing not only a physical transformation of the amorphous precipitate into LFS crystals occurs but also co-current chemical reduction of ferric iron to ferrous iron by hydrogen. To monitor this reaction, LFS samples at different annealing times were digested and analyzed by titration to determine the percentage content of ferric to ferrous reduction. Thus it was found after the 1 min annealing point at 400°C only 60.2% of Fe was reduced to ferrous state. After 3 hrs, the degree of reduction was 69%, after 6 hrs was 97%, with no apparent further increase after 10 hrs. Therefore, 6 hours was selected as the standard annealing time. Crystalline properties of LFS400 and LFS700 particles

Phase composition: LFS has been reported to exhibit different crystal phase polymorphs depending on the temperature of synthesis such as low-temperature orthorhombic (Pmn21 or βII) obtained at 200°C, high-temperature monoclinic (P21/n or γs) obtained at 600-800°C, and hightemperature orthorhombic (Pmnb or γII) obtained at 900°C.6, 13 Previous studies have mostly focused on the electrochemistry of the monoclinic (P21/n) LFS synthesized by annealing of


8

Page 9 of 25 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


ferrous-derived precursor phases at around 700°C.14, 16, 32-33 Here we prepared LFS samples at both 400°C and 700°C to evaluate the impact of the ferric-derived precursor and annealing temperature on crystal phase properties and ultimately electrochemical response. The XRD spectra (Figure 2) of both 400°C and 700°C (LFS400 and LFS700) samples (annealed for 6 hours) were fitted using Rietveld refinement to quantify the LFS phases present.

The

predominant phase in both annealed samples was determined to be the high temperature monoclinic (P21/n) phase, but to co-exist with a small fraction of orthorhombic LFS of different space group. Thus the LFS400 material comprised 10% low temperature orthorhombic (Pmn21), while the LFS700 one comprised 25% high temperature orthorhombic (Pmnb) with the bulk in both cases being the monoclinic (P21/n, 75%) phase. These results show a gradual phase transition with increasing annealing temperature from Pmn21 P21/n Pmnb. The majority of previously reported studies on high temperature monoclinic LFS make no mention of the presence of secondary orthorhombic phase except perhaps the work of Bini et al.34 despite theoretical predictions of phase co-existence.10,

35

Interestingly in an annealing study by

Armstrong et al.36 involving as starting material hydrothermally precipitated LFS from Fe(II)containing solution the obtained phase at 400 °C was Pmn21 and not P21/n that was produced here. It becomes evident therefore that the ethylenediamine-Fe(III)-silicate intermediate of the present work makes possible the formation of P21/n monoclinic LFS phase at a lower annealing temperature by opening a lower activation energy crystallization pathway. Upon electrochemical cycling, the high temperature (~700°C) monoclinic LFS phase (P21/n or γs) is well established to undergo a phase transition to low temperature orthorhombic one (Pmn21 or βII).7, 14-15, 33, 37 However, due to the favorable energetic phase-coexisting behavior10, 35 and the experimental findings of this work, the presence and type of the secondary orthorhombic phase should not be ignored, as it may play an important role in triggering the monoclinic to orthorhombic phase transformation during electrochemical cycling.15 In other words, the secondary phase (Pmn21) nanograins may alter the phase transformation reaction (P21/n Pmn21) by acting as nucleation (seed) sites and in doing so affecting the long-term cycling performance of LFS cathodes.


9


(a)

Page 10 of 25

(b)

Figure 2. The Rietveld refined XRD spectra (Co Kα, λ ~ 1.78901 Å) of LFS400 (a) and LFS700 (b) materials. The black lines indicate the experimentally collected spectra, the red bubbles indicate the Rietveld fitted patterns assuming the coexistence of LFS monoclinic and orthorhombic phases, the blue lines indicate the discrepancies between the experimental and calculated results. The XRD pattern of the LFS400 sample was fitted as a 10:90 mixture of Pmn21 and P21/n phases and that of the LFS700 sample as a 25:75 mixture of Pmnb and P21/n phases (fitting error within 7%).

Morphology: TEM coupled with selected area electron diffraction (SAED) analysis confirmed the Rietveld Refined XRD phase compositions of the LFS400 (Figure 3a, b, c) and LFS700 (Figure 3d,e,f) samples. Both monoclinic and orthorhombic regions were visualized distinctly in LFS400 and LFS700 nano/submicron particles. In the LFS400 sample, (102) and (31 1 ) planes were detected with interplanar d-spacing of 0.35 nm and 0.24 nm, which correspond to a monoclinic region of the sample; (001) and (210) planes were detected with interplanar d-spacing of 0.52 nm and 0.26 nm, which correspond to an orthorhombic region of the sample. Similarly, in the LFS700 sample, the (010) and (102) planes were detected with interplanar d-spacing of 0.52 nm and 0.38 nm, corresponding to a monoclinic region of the sample; the (100) and (010) planes were detected with interplanar d-spacing of 0.31 nm and 0.52 nm, corresponding to an orthorhombic region of the sample. From the TEM images (Figures 3a, d), the crystallite size in LFS400 sample was deduced to be around 50 nm while for the LFS700 sample it was close to 200 nm due to apparent grain coarsening effect. It can be further noticed that the crystallite clusters were covered with carbon and had some degree of porosity (Figure S8, S9 in Supplementary Information), which are desirable features for a cathode material as they allow for electrolyte infiltration and enhanced conductivity.7


10

Page 11 of 25 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. TEM characterization of LFS400 (a, b, c) and LFS700 (d, e, f) samples. (a, d) Typical crystal morphology and size. (b) A monoclinic region of the LFS400 sample, where the diffraction spots of the (102) and (311) planes are observed with interplanar d-spacing of 0.35 nm and 0.24 nm and (c) an orthorhombic region, where the diffraction spots of the (001) and (210) planes are seen with interplanar d-spacing of 0.52 nm and 0.26 nm. The incident electron beam was parallel to the [120] zone axis. (e) A monoclinic region of the LFS700 sample, where the diffraction spots of the (010) and (102) planes are observed with interplanar d-spacing of 0.52 nm and 0.38 nm, indexed to the zone axis [201] and (f) an orthorhombic region of the same sample, where the diffraction spots of the (100) and (010) planes are seen with interplanar d-spacing of 0.31 nm and 0.52 nm. The incident electron beam was parallel to the [001] zone axis.

The size and morphology of LFS400 and LFS700 nanostructured particles were also examined using scanning electron microscopy (SEM) (Fig. 4). The LFS400 image (Fig 4a) shows that the material was in the form of secondary submicron sized porous agglomerates of unitary primary nanoparticles (typically around 40-60 nm). In contrary, the LFS700 image (Fig. 4b) shows that the agglomerates hierarchically organized into a 3D framework of rings of unitary


11


Page 12 of 25

crystallites. The primary crystallites this time are seen to have grown to bigger size (around 100200 nm) as a result of the coarsening effect of the elevated annealing temperature (700°C). The rings form openly connected porous cages of submicron size. Such porous structure can be advantageous for the LFS cathode as it would allow for unhindered electrolyte infiltration and thereby enhanced lithium ion transport.7 Additional EDS analysis coupled with SEM imaging (Figure S10 and S11 in Supplementary Information) revealed the key constituent elements (C, Si, Fe, O) to be uniformly dispersed in the selected region with no obvious signs of impurities (i.e. SiO2 where Fe was deficient or iron oxides where Si was deficient) in both LFS400 and LFS700 samples. Furthermore, the BJH pore size distribution derived from the BET adsorption measurement showed that the LFS400 material had an average pore width around 50 nm, which corresponds to Type I mesoporous material. (Supplementary Information Fig. S12) For LFS700 pristine sample it was mainly Type II macroporous however after ball milling it became Type I mesoporous with pore size around 50nm as well (Supplementary Information Fig. S13). This confirmed their microscopically observed porous structure.

Figure 4. SEM secondary electron images of LFS400 (a) and LFS700 (b) samples.

Carbon coating and iron speciation: In addition to the abovementioned phase and morphology characteristics of the two annealed LFS samples, the in situ coated carbon layer (evident in the TEM images given in Figure S8, S9 in Supplementary Information) and the chemical states of the surface Fe species, are also important material attributes for LFS in Li-ion battery cathode application. The LFS particles were coated with carbon as a result of the decomposition of the surface-anchored organic additives (EN and EG) during annealing. The carbon content was quantified using thermal gravimetric analysis (TGA) to be about 8% for LFS400 (Figure S14 in Supplementary Information.) and slightly less than 5% for LFS700. Furthermore, XPS spectra confirmed the presence of carbon on the surface as shown in Figure 5.


12

Page 13 of 25 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


Various carbon bonds were detected, including C-C, C-O, and C=O, located at 284.8 eV, 286.6 eV, and 289.9 eV respectively38-39 in both LFS400 and LFS700 samples. Furthermore, it is interesting to notice that besides carbon, nitrogen signals were also detected (as -C-N(quaternary), -C=C-NH (pyrrolic) and -C-N=C- (pyridinic) bonds-as shown in Figure S15 in Supplementary Information), strongly suggesting the decomposition of ethylenediamine rendering the carbon layer N-doped. The presence of sp2-typed unsaturated carbon bonds together with the N bonding made the carbon layer potentially more conductive due to the additional electrons donated to the conduction band,40-41 an important aspect in enhancing the functionality of LFS that has very poor electronic conductivity (10-12 – 10-16 S • cm-1 at room temperature).42 XPS Fe spectra analysis revealed the LFS400 nanoparticles to have some Fe3+ species at the outmost surface but not the LFS700 particles (Figure 5 b,d). The Fe 2p3/2 bonding at 711.4 eV of the LFS400 spectrum corresponds to the ferric oxidation state while the Fe 2p3/2 binding at 710.4 eV of the LFS700 spectrum to the Fe2+ oxidation state.25 The presence of the ferric species in LFS400 was ascribed in part to the inevitable air oxidation of the nanoparticles (smaller size than that of LFS700-refer to Figure 3 a,d). The presence of ferric species in the bulk of the samples was quantified by chemical titration following digestion. The corresponding quantities in LFS400 and LFS700 was 4% and 1% ferric respectively. In addition to surface oxidation incomplete reduction of ferric to ferrous during reductive annealing is held responsible for the higher ferric fraction in LFS400 in agreement with the analysis presented in the previous section. Other than quantifying the ferric amount, the chemical analysis of the digested solids yielded deficient in Li content, namely the Li: Fe ratio is 1.69 and 1.72 in LFS400 and LFS700 samples. (Table S2 in Supplementary Information) Such ratio corresponds to 15% of the iron as ferric in pure LFS. Considering the fact that very little ferric (

Ethylenediamine-Enabled Sustainable Synthesis of Mesoporous

Recommend Documents