Porous, One dimensional and High Aspect Ratio Mn3O4 Nanofibers: Fabrication and Optimization for Enhanced Supercapacitive Properties

Porous, One dimensional and High Aspect Ratio Mn3O4 Nanofibers: Fabrication and Optimization for Enhanced Supercapacitive Properties

Accepted Manuscript Title: Porous, One dimensional and High Aspect Ratio Mn3 O4 Nanofibers: Fabrication and Optimization for Enhanced Supercapacitive ...

760KB Sizes 2 Downloads 23 Views

Accepted Manuscript Title: Porous, One dimensional and High Aspect Ratio Mn3 O4 Nanofibers: Fabrication and Optimization for Enhanced Supercapacitive Properties Author: Jai Bhagwan Asit Sahoo Kanhaiya Lal Yadav Yogesh Sharma PII: DOI: Reference:

S0013-4686(15)01444-9 http://dx.doi.org/doi:10.1016/j.electacta.2015.06.073 EA 25198

To appear in:

Electrochimica Acta

Received date: Revised date: Accepted date:

23-1-2015 16-6-2015 18-6-2015

Please cite this article as: Jai Bhagwan, Asit Sahoo, Kanhaiya Lal Yadav, Yogesh Sharma, Porous, One dimensional and High Aspect Ratio Mn3O4 Nanofibers: Fabrication and Optimization for Enhanced Supercapacitive Properties, Electrochimica Acta http://dx.doi.org/10.1016/j.electacta.2015.06.073 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Porous, One dimensional and High Aspect Ratio Mn3O4 Nanofibers: Fabrication and Optimization for Enhanced Supercapacitive Properties Jai Bhagwan,a,b Asit Sahoo,a Kanhaiya Lal Yadav,b and Yogesh Sharmaa,b,* a

Energy Storage Laboratory, Department of Applied Science & Engineering, IIT Roorkee Saharanpur Campus, Saharanpur-247001, India b Centre of Nanotechnology, IIT Roorkee, Roorkee-247667, India *Corresponding author: [email protected]


Hollow Mn3O4 is fabricated by versatile electrospinning technique.

The effect of solution properties on the porosity, alignment and aspect ratio of Mn3O4 is studied.

The best optimized nanofiber composed of interconnected nanoparticles forming hollow structure.

This hollow structure enables the reduced diffusion path of ions and increases the contact area of active material


Morphology of Mn3O4 is tuned to the nanoparticles / nanorods / nanofiber if ratio of metal precursor to polymer is varied from 0.33:1 to 2:1 in electrospinning solution. The best optimized nanofiber of Mn3O4 in terms of surface area, pore size and its distribution, and aspect ratio are obtained when equal amount of metal precursor and polymer (MN1:1) in electrospinning solution is taken, and sintered precisely at 1


C min-1. The structural,

morphological and thermal characterizations are carried out by XRD, FESEM, TEM, SAED, BET surface area and TG analysis. Further, these morphologies of Mn3O4 are subjected to the electrochemical characterization for evaluating the supercapacitive performance. The value of

specific capacitance of MN1:1 is found to be 210 (±5) F g-1 and 155 (±5) F g-1 at 0.3 A g-1 in 1M KCl and 1M Na2SO4, respectively. Improved supercapacitive performance of MN1:1 in both electrolytes is attributed to the unique nanofibric morphology where small nanoparticles are interconnected with good amount of open pores and forms a porous, one dimensional and high aspect ratio nanofibers. Electrochemical impedance spectroscopy (EIS) shows very low charge transfer resistance in MN1:1 favorable for fast and facile transportation of electrolyte ions to electrode and vice versa. Keywords: Supercapacitor, Electrospinning, Nanofiber, Spinel-Mn3O4, Impedance.

1. Introduction Global warming and increasingly worsened environmental pollution have stimulated the urge to develop energy storage devices that can store energy obtained from renewable sources like solar, wind and hydropower and supply the same whenever required [1,2]. Supercapacitor, an energy storage device, placed between batteries and conventional capacitor is planned to use in high end applications where both high energy and power densities are required [3]. Presently available supercapacitor demonstrate only high power density, however to meet the increasing energy density for next generation, number of binary and ternary transition metal oxides are being explored as prospective electrode material [4-13]. Among them, RuO2 is found to be most promising as far as high Cs (>1500 F g-1) is concerned [6]. However, high cost and toxicity of ruthenium are the main hindrance towards its commercialization. In recent years, Mn-based oxides due to varying oxidation of Mn, low cost, non-toxic and abundant in nature with good environmental compatibility are receiving gigantic attention as functional materials [7-13]. In particular, spinel-Mn3O4 consisting of voids/gaps in its crystal structure may be a good choice

for investigating its supercapacitive properties. Since these voids/gaps present in spinel structure may facilitate the redox reaction due to intercalation/de-intercalation of electrolytic ions and thereby improving the pseudocapacitive performance of Mn3O4. Recently, a number of studies on supercapacitive performance of Mn3O4 and/or its composites with conducting species prepared by different synthetic procedures have been reported [8-13]. But, most of these studies have not been found so encouraging probably due to poor electrical conductivity of Mn3O4 (even in the composites with conducting species) and/or improper morphologies or surface properties. The supercapacitive performance of Mn3O4 may be improved by creating the nanostructures of appropriate morphology of high surface area with suitable pore size, alignment and high aspect ratio. The optimized morphology would enable the reduced diffusion path of ions and increase the contact area of active material to the aqueous electrolyte. This will lead to increase mobility of foreign ions and reduction in the charge transfer resistance at electrode/electrolyte interface and hence improved supercapacitive properties of Mn3O4 may be obtained. Morphological properties such as surface area, pore size and its distribution, and aspect ratio are not easy to tune and control with high degree of reproducibility using many of available cost effective and easy methods. In view of this, electrospinning technique is considered as a simple, versatile and fast emerging technique which may enable one-dimensional, porous and high aspect ratio nanofibers with good degree of reproducibility [14-18]. Although, electrospinning technique seems simple but tuning and controlling the morphology to obtain high surface area, appropriate pore size and their distribution, and good aspect ratio is still a challenging task. Seeing the various important morphology dependent functionality of Mn3O4 as catalytic [19], magnetic [20-22], gas sensing [23] and energy storage material [8-13], creation and optimization of above mentioned morphological features by easy electrospinning technique

need to be carried out thoroughly, and then its functionality in respective fields may be examined. Presently, solution properties of electrospinning solution are tuned by varying the metal content over polymer and then its effect on the porosity, alignment and aspect ratio of Mn3O4 is investigated. Further, correlation of these morphological features with supercapacitive properties is studied for the first time. Metal precursor to polymer ratio of 0.33:1, 0.5:1, 1:1 and 2:1 produce nanoparticles (MN0.33:1), nanorods (MN0.5:1) and nanofibers (MN1:1 & MN2:1), respectively. Metal precursor to polymer in the ratio of 1:1 is found to be most appropriate to get aligned, one dimensional and high aspect ratio nanofiber of Mn3O4. These nanofibers are composed of interconnected nanoparticles that form the hollow structure of high surface area (~24 m2 g-1) and meso-pores (majority of 10–30 nm) with almost homogeneous distribution. MN1:1 exhibits specific capacitance (Cs) of 210 (±5) F g-1 in the range of 0.0V–0.9V in aqueous solution of 1M KCl and 155 (±5) F g-1 in 1M Na2SO4 in the range of 0.0V–1.0V at current density of 0.3 A g-1 and shows improved rate capability. These results are well supported by ex-situ FESEM and electrochemical impedance spectroscopy (EIS) studies. 2. Experimental details 2.1 Fabrication of Mn3O4 nanofibers To prepare Mn3O4 nanofibers, manganese acetate tetrahydrate (Mn(CH3COO)2.4H2O, ≥ 99.5%, Merck), polyvinylpyrrolidone ((C6H9NO)x, Mw=360,000, Sigma Aldrich), ethanol (C2H5OH, Merck, 1L=0.790 kg) and acetic acid (CH3COOH, ≥ 99%, Merck) were used as starting material without further purification. In typical process, 0.5 g of Mn(CH3COO)2.4H2O was dissolved in 0.8 ml of distilled water, and separately 0.5 g polyvinylpyrrolidone (PVP) was dissolved in 8.4 ml ethanol to make 7% concentration of polymer solution. Finally, both the

solutions containing equal amount of metal precursor and polymer (mentioned as MN1:1) were mixed thoroughly and allowed to stir for 30 minutes. Later, 200 µl of acetic acid was added to resulting solution which was further stirred for 20 h to get homogenous and clear solution. For electrospinning process, the hypodermic syringe connected to spinneret (stainless steel, 25G) was hosted with prepared solution and loaded with syringe pump as shown in Fig. 1. The separation between the syringe tip and collector was kept at 10 cm. Solution pumping rate of 2 ml h-1 and voltage of 15 kV was fixed. The temperature and humidity inside the electrospinning chamber were kept at 43 (±2) 0C and 15 (±5)%, respectively. As-spunned nanofibers were dried for 12 h at 70 0C in vacuum oven and then sintered at 350 0C for 4 h in a muffle furnace using precise heating rate of 1 0C min-1. Three other solutions with metal precursor to polymer ratios of 0.33:1 (MN0.33:1), 0.5:1 (MN0.5:1) and 2:1 (MN2:1) were also prepared and processed following same procedure. The dark brown powder of sintered Mn3O4 was obtained and preserved in desiccator for further characterization. 2.2 Material characterization To get the information about thermal stability, thermal gravimetric analysis (TGA) and differential thermal analysis (DTA) were performed by thermal analyzer (SII 6300 EXSTAR) using ramping rate of 5 0C min-1 in air. Structural analysis was carried out by X-ray diffractometer (Rigaku Ultima IV) equipped with Cu-Kα radiation (λ=1.541Å). The fouriertransform infrared (FTIR) spectra of the precursors, as-spun and sintered nanofibers were recorded using spectrometer (Perkin-Elmer C91158) in the wave number ranging from 4000 cm-1 to 400 cm-1. Surface area of samples was determined by Brunauer–Emmet–Taller (BET) principle and the pore parameters by Barrett–Joyner–Halenda (BJH) method using surface area analyzer (Quantachrome ASIQWIN at 77K). Viscosity of solution was measured by DV-E

viscometer by keeping the spindle’s speed of 60 rpm (Brookfield, spindle-3 (RA/HA/HB series)). Microstructural images were obtained using field emission scanning electron microscope (FESEM; QUANTA FEG-200) and transmission electron microscope (TEM; FEI TECHNAI G2 20 S-TWIN). 2.3 Electrode fabrication and electrochemical characterization Active material, conducting carbon (super P) and polyvinylidene difluoride (PVDF) were taken in the weight ratios of 70:15:15, and mixed thoroughly using N-methyl 2-pyrrolidone (NMP) as solvent. The mixed solution was further stirred for 12 h to get homogeneous slurry which was then coated on graphite sheet (1cm×2cm) and dried at 80 0C for 12 h to evaporate the solvent. The active material loading for all samples was kept in the range of 3–4 mg. Galvanostatic charging-discharging (GCD), cyclic voltammetry (CV) and electrochemical impedance spectroscopy (EIS) were performed by multichannel Poteniostat/Galvanostat (MAC80039 AUTOLAB) in three electrode configuration where Mn3O4 nanofiber, Ag/AgCl and Pt wire were used as working, reference and counter electrode, respectively. The aqueous solution of two different electrolytes namely 1M KCl and 1M Na2SO4 were used. Specific capacitance of the active material from the CV and GCD curves are calculated using the following equations 1 and 2, respectively [24-27].

Cs =

Ι mv


where I is average current (A) and v is scan rate (V s-1) for cyclic voltammetry. Cs =

Ι  ∆t    m  ∆V 


where C s is the specific capacitance (F g-1), Ι is the fixed current (A) for charge/discharge analysis, m represents the mass of active material (g) and

is the slop of the discharge

curve. Since pseudocapacitor also possess redox reaction similar to battery so, it is always debatable whether the unit of charge storage should be F g-1 or mAhg-1. The superiority of redox reaction over EDLC behavior is crucial to determine the suitability of F g-1 or mAhg-1. If prominent redox peaks in CV and long flat plateau in GCD studies are observed then writing mAhg-1 is more justified. However, in present case almost rectangular CV and triangular GCD curves are seen, hence, F g-1 is used as the unit of charge storage. Energy density (W h kg-1) and power density (W kg-1) of the material are also calculated using GCD curves as per equations 3 and 4, respectively [28].


1 2 C s (∆V ) 2 × 3.6



E × 3600 T


where ∆V is the potential window,

represents discharging time (s). Electrochemical

impedance spectroscopy (EIS) was performed between the frequency of 1 MHz to 10 mHz employing bias voltage of 5 mV, and the data are shown as Nyquist plot (Z' and -Z''). 3.

Results and discussion

3.1 Thermal, structural and morphological studies Thermal stability of as-spun nanofiber in the range of room temperature (RT) to 700 0C was studied by TGA to confirm the calcination /sintering temperature of as-spun nanofiber to obtain pure phase of Mn3O4 retaining the nanofibric morphology. As seen in Fig. 2a, the weight loss (~21%) in the range of 26 0C to 278 0C is due to the loss of adsorbed moisture, water molecules from the metallic precursor and trapped solvents (water, ethanol etc) present in asspun nanofiber. The major weight loss in the range of 278 0C to 391 0C corresponds to the decomposition of metallic precursor (manganese acetate) and PVP, thereby removal of polymer

leaves behind the pores (shown in FESEM image, Fig. 6). Thereafter till 700 0C, no significant weight loss is observed. Thus, slightly lower temperature (350 0C) than the decomposition temperature (~391 0C) with precise and slow heating rate of 1 0C min-1 was fixed as sintering temperature to retain the nanofibric morphology with high aspect ratio of end product [14,29]. Slow heating rate allow us to keep sintering temperature slightly lower than the complete decomposition temperature. Similarly other reports in literature [14,29] emphasized on the importance of sintering conditions to tune the morphology even at lower temperature than the complete decomposition temperature of PVP. The removal of PVP may result the formation of carbon traces inside the sintered sample which may further affect the supercapacitive properties of Mn3O4 (MN 1:1) nanofibers, therefore, to quantify the weight of carbon, the TGA analysis of sintered MN 1:1 from room temperature to 900 0C is carried out and shown in Fig. 2b. For comparison, other three samples were also subjected to TGA (along with DTA) and results are shown in Fig. S1 (supporting information). As can be seen in figures, weight losses of (about 2.5 wt %) till 300 0C attributable to the removal of adsorbed moisture/water molecules from the surface of material is observed. Interestingly, slight increase in the weight of all four samples from 400 0C -600 0C (evident hump in case of nanoparticles (MN0.33:1) and nanorods (MN0.5:1)) is observed. This minute increment in the weight may be ascribed to the transformation of tetrahedral-Mn3O4 to cubicMn2O3 as it is well supported by X-ray patterns recorded at different temperature (Fig. S2). The amount of weight loss in the range of 600 0C -900 0C becomes important to quantify the carbon content (if any) in all the samples. Since, this is the temperature range where left over C traces react with oxygen to form CO2 or CO gaseous product and thereby the weight loss in the sample appears. However, presently, none of the samples shows any significant changes in this range

except 1-2 wt % loss. Based on TGA analysis, the contribution of left over carbon to the specific capacitance (particularly to the value) may be neglected (since weight of carbon in the active material is negligible), but dispersion of carbon, even if it is too less (~ 1-2 wt %), will definitely facilitate the electronic/ionic activity of the active material (Mn3O4) to the electrolyte and vice versa, and thereby contribute to overall storage performance of material. To observe the dispersion of left over carbon in all the samples (heat treated at 350 0C for 4h), elemental mapping through energy dispersive X-ray analysis (EDAX) is shown in Fig S3a-d (supporting information). The observed weight % of carbon in EDAX analysis is found to be in the range of 1.5-2.5 wt % for all the samples which further supports the findings of TG analysis. The inhomogeneous dispersion of carbon and its poor mapping density (Fig. S3a-d) further indicate that carbon content in present study may not have any significant effect, however, its contribution in providing the electronic wiring between individual particles of Mn3O4 and thereby enhancing the participation of active material cannot be ruled out. X-ray diffraction (XRD) is a reliable and non-destructive technique to determine crystal structure of Mn3O4. The corresponding XRD pattern of sintered Mn3O4 nanofiber MN1:1 is shown in Fig. 3. XRD patterns of other sintered samples are also shown as supporting information (Fig. S4). The refined XRD pattern of MN1:1 using PDXL-2 (Rigaku) software is also shown as inset. All characteristic peaks present in XRD pattern correspond to the tetragonal structure of Mn3O4 (PDF file 24-0734). The lattice parameters (a = b = 5.77 (±0.01) Å and c = 9.47 (±0.01) Å) calculated from refined data are found to be in good agreement with the reported values [30]. Crystallographic structure of Mn3O4 generated using refined data is shown (Fig. 4), where a spinel structure that contains Mn2+ at tetrahedral sites and Mn3+ at octahedral sites is displayed.

The FTIR spectrum of Mn3O4 nanofiber (MN1:1) is shown in Fig. 5. For reference, spectra of precursors: PVP and manganese acetate are also shown in Fig. 5a and Fig. 5b, respectively. As can be seen in Fig. 5c, the various bands present in FTIR spectrum of as-spun nanofiber is the union of spectrum of the individual Mn-acetate and PVP. For instance, two bands at 3478 cm-1 and 2957 cm-1 are assigned to the stretching vibration of hydroxyl group O–H and C–H, respectively [31]. The band at 1664 cm-1 is attributed to the C=O functional group [32]. Two peaks i.e. one at 1445 cm-1 and another at 1285 cm-1 are due to the scissoring of CH2 and stretching of C–N group, respectively [33]. The intensity of peaks corresponding to CH2, C–N and C=O group are reduced in the as-spun nanofiber of Mn3O4 where Mn-precursor and PVP are thoroughly mixed. The FTIR spectrum for sintered sample is found to be entirely different, since PVP is completely degraded and hence no bands related to either of reactants are observed. The FTIR spectrum of sintered Mn3O4 nanofiber in the range of 400 to 650 cm-1 is shown in Fig. 5d. The first two peaks at 630 cm-1 and 511 cm-1 may be assigned to the Mn–O stretching in tetrahedral sites and distorted vibration of Mn–O in octahedral environment, respectively [34]. Third peak at 422 cm-1 may be attributed to the vibration of manganese species (Mn3+) in the octahedral site [34]. The results obtained from FTIR analysis concludes that sintered nanofibers of Mn3O4 are highly crystalline. This results further complements the TGA results where 350 0C was found suitable temperature to sinter the as-spun nanofiber to get well crystalline Mn3O4 retaining nanofibric morphology with high aspect ratio (as depicted in Fig. 7a–d). The morphology of Mn3O4 nanofibers is altered by a systematic study of varying the metal precursor to polymer weight ratio of electrospinning solution. The as-spun nanofibers of MN0.33:1, MN0.5:1, MN1:1 and MN2:1 are shown in Fig. 6. By increasing the weight of metallic precursor over polymer from 0.33:1 to 2:1, a steady increase in viscosity (172 cP to 235

cP) is obtained. This variation in viscosity and dispersion of metal precursor into the PVP do not affect the shape and size of as-spun nanofiber, significantly. However, a drastic change in the morphology of sintered nanofibers is observed (Fig. 7). If the metal precursor is kept one third of PVP (MN0.33:1), severe destruction of fibrous morphology is found that results the formation of individual nanoparticles (Fig. 7a). This is probably due to wider dispersion of metallic precursor into the polymer solution and thereby the larger separation between individual metal precursor particles. Further, if weight of the metallic precursor slightly increases to half of polymer (MN0.5:1) nanofibers are broken down into shorter segments and forms nanorod kind of morphology (Fig. 7b). Further increase of metal precursor equal to polymer (MN1:1) (viscosity 220 cP) provides suitable dispersion between two individual metal particles mediated by polymer and then slow removal of PVP assisted by precisely controlled heating rate creates gaps or pores between two individual Mn3O4 nanoparticles and facilitates unidirectional growth of Mn3O4 nanofibers. If metal precursor is then increased twice of the polymer (MN2:1), a closer dispersion of metal particles into polymer is expected that leads to form a densed nanofiber as depicted in schematic diagram (Fig. 7). Hence, equal weight ratio of metallic precursor to PVP is found to be appropriate to fabricate the hollow and high aspect ratio nanofiber. These nanofibers are made up of interconnected small nanoparticles (20–50 nm) that provide easy and facile transportation to the foreign species, K+ and Na+ and enable improved supercapacitive performance (as discussed later). Transmission electron microscope (TEM) is used to further support the finding of FESEM. The TEM (Fig. 8) displays an aligned, high aspect ratio and one dimensional nanofiber of Mn3O4 (MN1:1) where open active sites/pores (encircled 30–40 nm) favorable for good supercapacitive properties are shown in Fig. 8. Selected area electron diffraction (SAED) pattern (shown as inset of Fig. 8) depicts the concentric rings with contrast

bright spot indicative of well polycrystallinity of Mn3O4 nanofiber. The interplanar distance (d) corresponding to these rings (inner to outer from the center) are 4.92, 3.07, 2.88, 2.76 Å, respectively that corresponds to the (101), (112), (200) and (103) plane of tetragonal-Mn3O4. These results complement the finding of X-ray analysis. Typically, the pseudocapacitive performance of an electrode is closely related to the distribution of the pores and the specific surface area of the electroactive material [35-38]. Thus, Mn3O4 nanoparticles / nanorods / nanofibers are further characterized by surface area analyzer using nitrogen adsorption and desorption techniques at 77K and results are shown in Fig. 9 (for MN1:1) and Fig. S5 (supporting information for rest of the systems). The BET surface area of hollow, aligned and one dimensional (MN1:1) is found to be highest (24 m2 g-1). Controlled and slow removal of PVP during sintering may leave behind the pores (meso) that are beneficial for fast ions transfer at electrode/electrolyte interface in MN1:1 (as discussed later). Mesopores (majority in the range of 10–30 nm) assessed by the Barrett–Joyner–Halenda (BJH) method are shown in inset of Fig. 9. The values of surface area and pore volume for other systems are tabulated in Table 1. 3.2 Electrochemical studies To investigate the performance of all morphologies of Mn3O4 as supercapacitor electrode, electrochemical characterization such as CV and GCD are carried out. Detailed results of only MN1:1 are discussed here, however the CV curves and GCD graphs for other systems are given in Fig. S6 (supporting information). The CV curves recorded at different scan rates i.e. 3, 10, 30 and 50 mV s-1 for MN1:1 are shown in Fig. 10. The value of specific capacitance using cyclic voltammetry are found to be 190 (±5) F g-1, 160 (±5) F g-1, 140 (±5) and 130 (±5) F g-1 in 1M KCl electrolyte. The value of Cs at different scan rate for different morphologies is also tabulated

in Table 1. The shape of the cyclic voltammograms for MN1:1 is found to be independent of scan rates indicating unrestricted motion of electrolyte ions in the pores/voids of interconnected nanofiber (Fig. 10). The Cs value remains almost constant at 140 (±10) F g-1 even if the scan rate increases five-fold from 10 mV s-1 to 50 mV s-1. This high performance of MN1:1 nanofiber is ascribed to the unique nanofibric morphology where small nanoparticles are interconnected with good amount of open pores (Fig. 7–9). Further, percolation of electrolyte ions even to interior of nanofiber provides the maximum active site (even at high scan rate of 50 mV s-1) increases its importance for high rate performing supercapacitor. To test further the importance of the nanofibric morphology of MN1:1, the sample was cycled in another electrolyte, 1M Na2SO4 which gives the Cs value of 155 (±5) F g-1, 145 (±5) F g-1, 135 (±5) F g-1 and 130 (±5) F g-1 at the scan rates of 3, 10, 30 and 50 mV s-1, respectively. Further, the size of sulphate anions (258 pm) is bigger than Cl anion (190 pm) [39,40]. From the literature, we notice that 1M Na2SO4 is not as good as 1M KCl owing to lower ionic conductivity and poor mobility of Na+ ions as compared to the K+ [39,40]. From this result, we can conclude that if morphology of electrode material is favorable to the ionic movement, the supercapacitive performance of given material may also be improved in poor electrolyte. However, it is to be noted that not only the morphology of electrode material but also the crystal structure of the material is important for improved supercapacitive properties [5,9,36-38]. Further, the electrolyte’s properties (such as redox behavior, ionic size, conductivity and electrochemical voltage window) also play vital role in improving the supercapacitive properties [39,40]. Presently employed Mn3O4 nano-fibers have both structural features (voids/gaps in crystal structure) as well as morphological features (surface area, porosity, alignment, interconnected nanoparticles forms a porous nanofiber)) which justified it suitability as prospective supercapacitor electrode material. Further, the linear

relationship between average peak current and the square root of scan rate (I vs ʋ1/2) shown as inset of Fig. 10c and Fig. 10d indicate a rapid redox reaction as a consequence of mass transport controlled process occurring in MN1:1 [41,42]. To test the practical performance of MN1:1 as supercapacitor, the GCD analysis is also carried out in both the electrolytes and corresponding charge-discharge curves are shown in Fig. 11. The discharge curves are made up of mainly two regions: initial part (straight) may be ascribed to double layer charge storage process and the lower part (diffusion part) may be attributed mainly to the redox process [43]. Further, the voltage drops (15.6 mV in 1M KCl and 11.7 mV in 1M Na2SO4 at the current density of 0.3A g-1) are found to be small probably due to the small internal resistance of electrode material as discussed above and evident from CV analysis. The Cs value of 210 (±5) F g-1 and 155 (±5) F g-1 at 0.3 A g-1 are obtained in 1M KCl and 1M Na2SO4, respectively employing equation 2. Further, the cycling stability of MN1:1 nanofiber is also examined and results are shown in Fig. 12 where a stable Cs value of 210 (±5) F g-1 with almost 100% coulombic efficiency at least upto 500 cycles is obtained. In case of 1M Na2SO4, slightly lower Cs value of 155 (±5) F g-1 is observed. The rate capability of nanofibric MN1:1 at different current densities using 1M KCl and 1M Na2SO4 is also examined and results are shown in Fig. 12. A stable and high capacitance of 160 (±5) F g-1 even at higher current density of 0.6 A g-1 is realized, however, for 1M Na2SO4, Cs value of 140 (±5) F g-1 at current density of 0.6 A g-1 is obtained. It may be noted that MN 1:1 exhibits large difference in Cs values at lower current density in respective electrolytes, however this difference is reduced significantly at higher current density. This behavior of Mn3O4 may be attributed to rate dependent charge storage mechanism. Presently studied nanofiber Mn3O4 has both the beneficial features: structural as well as morphological. Spinel structure having voids/gaps will provide

extra active site for charge storage as a consequence of intercalation/de-intercalation or redox reaction [37-39]. Since ionic size of K+ is less than the ionic size of Na+ [39,40], so most likely K+ will go deeper into material (voids /gaps present in Mn3O4) and access maximum active sites available to charge storage. Further, KCl has higher mobility too and good conductivity, therefore at slow scan rate of 0.3 A g-1, the observed capacitance is more [44]. At the same rate, since Na2SO4 electrolyte is not as appropriate as KCl (already discussed in literature, in terms of conductivity, mobility and ionic size), less contribution to capacitance from redox reaction is realized [39,40]. This hypothesis may also be supported by seeing the GCD curves where longer sloppy region (in case of KCl (encircled) as compared to Na2SO4) is observed (Fig. 11). Hence, at the slower rate, both the redox reaction and EDLC do contribute to storage properties, however at higher rate, redox based storage properties drops significantly as compared to EDLC. This is the reason why Mn3O4 nanofiber shows big difference in capacitance value for different electrolyte at slow rate and will behave almost identical at higher rate (if given). Presently obtained results in terms of capacitance are well compared with literature reports. For example, Li et al. observed the Cs value of 70 F g-1 at 0.2 A g-1 for Mn3O4 [10]. Lee et al. shows very low specific capacitance (25 F g-1) of Mn3O4 at the current density of 0.5 A g-1 [11]. Subramani et al. reported the Cs value of 65 F g-1 of Mn3O4 at 0.5 A g-1 [12]. However, Zhang et al. demonstrated the specific capacitance of 147 F g-1 for graphene/Mn3O4 nanocomposite at the current density of 0.1 A g-1 [13]. Chen et al. synthesized nanostructured fibers of Mn3O4–C and obtained the capacitance of 80 F g-1 [45]. The capacitive performance of MN1:1 is also compared with other three systems: MN0.33:1, MN0.5:1 and MN2:1 (Table 1 and Fig. S6). The improved supercapacitive performance of MN1:1 evident from CV and GCD analysis is attributed to the smaller contact/interparticle resistance in unique nanofibric

morphology consisting of the pores and their homogenous distribution. The importance of creating the pores employing expensive RuO2 in Mn3O4 has also been shown by Youn et al. where they have shown improved capacitance (> 250 Fg-1) in porous RuO2–Mn3O4 nanocomposite [46]. Electrode integrity of electrode material for prolonged cycles is found to be responsible for good rate capability and cyclability [47]. Similar findings have also been observed in the present study. Ex-situ FESEM image of composite electrode (after 500 cycles) shows good integrity of electrode where nanofibric morphology is retained even after prolonged (Fig. 13). This justifies the importance of porous, one dimensional and aligned nanofiber. The energy density and power density are also calculated which are found to be 23 W h kg-1 and 120 W kg-1, respectively at 0.3 A g-1. Electrochemical impedance spectroscopy (EIS) has been used to complement the findings of CV and GCD studies, and to find the reason of improved performance of MN1:1 as compared to other samples. The Nyquist plots of the MN0.33:1, MN0.5:1, MN1:1 and MN2:1 electrodes in 1M KCl are shown in Fig. 14 and its inset indicates a small semi circle in the higher to medium frequency region (1 MHz–50 kHz) followed by a slopping line making an angle (~450) with xaxis. To distinguish the individual process and corresponding impedance parameters, an equivalent circuit comprising series and parallel combination of equivalent series resistance (Rs), electric double layer capacitance (Cdl), pseudocapacitance (Cp), charge transfer resistance (Rct) and frequency dependent Warburg impedance (W) is fitted [48,49]. The fitted values of Rs and Rct of MN1:1 are found to be lowest 0.2 Ω and 1.0 Ω, respectively. The value of Rs is found to be independent of morphology. However, diameter of semicircle along x-axis increases indicating the variation in charge transfer resistance. In case of dense nanofiber (MN2:1), nanorods (MN0.5:1) and nanoparticles (MN0.33:1), the values of Rct increases to 3, 5 and 6 Ω,

respectively. Further, sloppy region (~450) i.e. Warburg region in these samples is clearly seen. This sloppy region (~450) in low frequency regime represents frequency dependence ion diffusion/transportation of the electrolyte [50,51]. Higher value of Rct and significant Warburg region are found detrimental to good supercapacitive performance in these samples [17]. However, in case of MN1:1, the lowest value of Rct (1.0 Ω) and unnoticeable Warburg impedance facilitates the smooth and facile movement of electrolyte ions into electrode material and vice versa. Similar findings are obtained in GCD analysis where negligible voltage drop due to small internal resistance is realized (Fig. 11). Lower Rs and unnoticeable Warburg impedance are considered to be good for improving the energy density and power density of supercapacitor material which further justifies the suitability of aligned and high aspect ratio nanofiber of Mn3O4. 4.

Conclusions In the present work, we have demonstrated the simple, cost effective and potentially

scalable way to synthesize the nanofiber of Mn3O4. The morphology of Mn3O4 nanofiber is tuned employing various process and system parameter of electrospinning process. Metal to polymer ratio of 1:1 was found to be appropriate to produce porous, aligned and high aspect ratio nanofiber of Mn3O4. Further, the reduction of metal to polymer ratio changes the fiber shape into nanorods or nanoparticles. The best optimized system in terms of surface area, alignment and aspect ratio is further characterized as supercapacitor electrode using both CV and GCD techniques. Cyclic voltammetry was performed with different scan rate in two electrolytes. The best electrochemical performance was measured for 1M KCl where a specific capacitance of 210 (±5) F g-1 at the current density of 0.3 A g-1 is obtained. Slightly lower value of specific capacitance i.e. 155 (±5) F g-1 is obtained in case of 1M Na2SO4. Higher energy density of 23 W

h kg-1 at the power density of 120 W kg-1 was obtained in 1M KCl. EIS measurement revealed that active material of nanofiber of Mn3O4 has low Rct value and unnoticeable Warburg impedance which facilitates the ionic movement from electrolyte to electrode and vice versa. Further, tuning and optimization of morphology in terms of pores, aspect ratio and dimensionality by simple and cost effective electrospinning method for Mn3O4 opens new opportunity not only to tune the supercapacitive properties but also to the other allied applications. Acknowledgments We are thankful to Department of Atomic Energy (DAE), BRNS, Government of India (Grant No. 2012/34/44/BRNS) for providing the financial support to complete this work.

References [1] P. Simon, Y. Gogotsi, Materials for electrochemical capacitors, Nature Mater. 7 (2008) 845. [2] W. Wei, X. Cui, W. Chen, D. G. Ivey, Manganese oxide-based materials as electrochemical supercapacitor electrodes, Chem. Soc. Rev. 40 (2011) 1697. [3] D. Pech, M. Brunet, H. Durou, P. Huang, V. Mochalin, Y. Gogotsi, P.-L. Taberna, P. Simon, Ultrahigh-power micrometre-sized supercapacitors based on onion-like carbon, Nature Nanotechnol. 5 (2010) 651. [4] L. Niu, Z. Li, Y. Xu, J. Sun, W. Hong, X. Liu, J. Wang, S. Yang, Simple synthesis of amorphous NiWO4 nanostructure and its application as a novel cathode material for asymmetric supercapacitors, ACS Appl. Mater. Interfaces 5 (2013) 8044.

[5] G. Q. Zhang, H. B. Wu, H. E. Hoster, M. B. Chan-Park, X. W. Lou, Single-crystalline NiCo2O4 nanoneedle arrays grown on conductive substrates as binder-free electrodes for high-performance supercapacitors, Energy Environ. Sci. 5 (2012) 9453. [6] C.-C. Hu, W.-C. Chen, Effects of substrates on the capacitive performance of RuOx·nH2O and activated carbon–RuOx electrodes for supercapacitors, Electrochim. Acta 49 (2004) 3469. [7] L. Lian, J. Yang, P. Xiong, W. Zhang, M. Wei, Facile synthesis of hierarchical MnO2 submicrospheres composed of nanosheets and their application for supercapacitors, RSC Adv. 4 (2014) 40753. [8] S. Guo, M. Zhang, G. Zhang, L. Zheng, L. Kang, Z.-H. Liu, Synthesis of novel Mn3O4 microsphere and its distinctive capacitance change during electrochemical cycling, Powder Technology 228 (2012) 371. [9] D. P. Dubal, R. Holze, All-solid-state flexible thin film supercapacitor based on Mn3O4 stacked nanosheets with gel electrolyte, Energy 51 (2013) 407. [10]

L. Li, K. H. Seng, Z. Chen, H. Liu, I. P. Nevirkovets, Z. Guo, Synthesis of Mn3O4-

anchored graphene sheet nanocomposites via a facile, fast microwave hydrothermal method and their supercapacitive behavior, Electrochim. Acta 87 (2013) 801. [11]

J. W. Lee, A. S. Hall, J.-D. Kim, T. E. Mallouk, A facile and template-free hydrothermal

synthesis of Mn3O4 nanorods on graphene sheets for supercapacitor electrodes with long cycle stability, Chem. Mater. 24 (2012) 1158. [12]

K. Subramani, D. Jeyakumar, M. Sathish, Manganese hexacyanoferrate derived Mn3O4

nanocubes–reduced graphene oxide nanocomposites and their charge storage characteristics in supercapacitors, Phys. Chem. Chem. Phys. 16 (2014) 4952.


X. Zhang, X. Sun, Y. Chen, D. Zhang, Y. Ma, One-step solvothermal synthesis of

graphene/Mn3O4 nanocomposites and their electrochemical properties for supercapacitors, Mater. Lett. 68 (2012) 336. [14]

P. F. Teh, Y. Sharma, Y. W. Ko, S. S. Pramana, M. Srinivasan, Tuning the morphology

of ZnMn2O4 lithium ion battery anodes by electrospinning and its effect on electrochemical performance, RSC Adv. 3 (2013) 2812. [15]

D. Li, Y. Xia, Electrospinning of nanofibers: Reinventing the wheel, Adv. Mater. 16

(2004) 1151. [16]

E. Lee, T. Lee, B.-S. Kim, Electrospun nanofiber of hybrid manganese oxides for

supercapacitor: Relevance to mixed inorganic interfaces, J. Power Sources 255 (2014) 335. [17]

S. Chaudhari, Y. Sharma, P. S. Archana, R. Jose, S. Ramakrishna, S. Mhaisalkar, M.

Srinivasan, Electrospun polyaniline nanofibers web electrodes for supercapacitors, Appl. Polym. Sci. 129 (2013) 1660. [18]

P. F. Teh, Y. Sharma, S. S. Pramana, M. Srinivasan, Nanoweb anodes composed of one-

dimensional, high aspect ratio, size tunable electrospun ZnFe2O4 nanofibers for lithium ion batteries, J. Mater. Chem. 21 (2011) 14999. [19]

K. A. M. Ahmed, Q. Zeng, K. Wu, K. Huang, Mn3O4 nanoplates and nanoparticles:

Synthesis, characterization, electrochemical and catalytic properties, J. Solid State Chem. 183 (2010) 744. [20]

Z. H. Wang, D. Y. Geng, Y. J. Zhang, Z. D. Zhang, Morphology, structure and magnetic

properties of single-crystal Mn3O4 nanorods, J. Cryst. Growth 310 (2008) 4148.


S. Sambasivam, G. J. Li, J. H. Jeong, B. C. Choi, K. T. Lim, S. S. Kim, T. K. Song,

Structural, optical, and magnetic properties of single-crystalline Mn3O4 nanowires, J. Nanopart Res. 14 (2012) 1138. [22]

A. Vazquez-Olmos, R. Redon, G. Rodriguez-Gattorno, M. E. Mata-Zamora, F. Morales-

Leal, A. L. Fernandez-Osorio, J. M. Saniger, One-step synthesis of Mn3O4 nanoparticles: Structural and magnetic study, J. Colloid and Interface Sci. 291 (2005) 175. [23]

T. Yamashita, A. Vannice, Temperature-programmed desorption of NO adsorbed on

Mn2O3 and Mn304, Appl. Catal. B: Environmental 13 (1997) 141. [24]

Y. Jin, H. Chen, M. Chen, N. Liu, Q. Li, Graphene-patched CNT/MnO2 nanocomposite

papers for the electrode of high-performance flexible asymmetric supercapacitors, ACS Appl. Mater. Interfaces 5 (2013) 3408. [25]

S. Nagamuthu, S. Vijayakumar, G. Muralidharan, Synthesis of Mn3O4/amorphous carbon

nanoparticles as electrode material for high performance supercapacitor applications, Energy Fuels 27 (2013) 3508. [26]

Q. Lu, M. W. Lattanzi, Y. Chen, X. Kou, W. Li, X. Fan, K. M. Unruh, J. G. Chen, J. Q.

Xiao, Supercapacitor electrodes with high-energy and power densities prepared from monolithic NiO/Ni nanocomposites, Angew. Chem. Int. Ed. 50 (2011) 6847. [27]

D. Cai, D. Wang, B. Liu, Y. Wang, Y. Liu, L. Wang, H. Li, H. Huang, Q. Li, T. Wang,

Comparison of the electrochemical performance of NiMoO4 nanorods and hierarchical nanospheres for supercapacitor applications, ACS Appl. Mater. Interfaces 5 (2013) 12905. [28]

S. He, W. Chen, High performance supercapacitors based on three-dimensional ultralight

flexible manganese oxide nanosheets/carbon foam composites, J. Power Sources 262 (2014) 391.


G. Wee, H. Z. Soh, Y. L. Cheah, S. G. Mhaisalkar, M. Srinivasan, Synthesis and

electrochemical properties of electrospun V2O5 nanofibers as supercapacitor electrodes, J. Mater. Chem. 20 (2010) 6720. [30]

J. Duan, Y. Zheng, S. Chen, Y. Tang, M. Jaroniec, S. Qiao, Mesoporous hybrid material

composed of Mn3O4 nanoparticles on nitrogen-doped graphene for highly efficient oxygen reduction reaction, Chem. Commun. 49 (2013) 7705. [31]

C. Qizheng, D. Xiangting, W. Jinxian, L. Mei, Direct fabrication of cerium oxide hollow

nanofibers by electrospinning, J. Rare Earths 26 (2008) 664. [32]

S. Maensiri, P. Laokul, V. Promarak, Synthesis and optical properties of nanocrystalline

ZnO powders by a simple method using zinc acetate dihydrate and poly(vinyl pyrrolidone), J. Crystal Growth 289 (2006) 102. [33]

Y. Borodko, S. E. Habas, M. Koebel, P. Yang, H. Frei, G. A. Somorjai, Probing the

interaction of poly(vinylpyrrolidone) with platinum nanocrystals by UV-Raman and FTIR, J. Phys. Chem. B 110 (2006) 23052. [34]

R. Song, S. Feng, H. Wang, C. Hou, Effect of organic solvents on particle size of Mn3O4

nanoparticles synthesized by a solvothermal method, J. Solid State Chem. 202 (2013) 57. [35]

Z.-F. Li, H. Zhang, Q. Liu, L. Sun, L. Stanciu, J. Xie, Fabrication of high-surface-area

graphene/polyaniline nanocomposites and their application in supercapacitors, ACS Appl. Mater. Interfaces 5 (2013) 2685. [36]

P. Sen, A. De, Electrochemical performances of poly(3,4-ethylenedioxythiophene)–

NiFe2O4 nanocomposite as electrode for supercapacitor, Electrochim. Acta 55 (2010) 4677. [37]

A. Sahoo, Y. Sharma, Synthesis and characterization of nanostructured ternary zinc

manganese oxide as novel supercapacitor material, Mater. Chem. Phys. 149 (2015) 721.


Q. Wang, B. Liu, X. Wang, S. Ran, L. Wang, D. Chen, G. Shen, Morphology evolution

of urchin-like NiCo2O4 nanostructures and their applications as psuedocapacitors and photoelectrochemical cells, J. Mater. Chem. 22 (2012) 21647. [39]

R. N. Reddy, R. G. Reddy, Sol–gel MnO2 as an electrode material for electrochemical

capacitors, J. Power Sources 124 (2003) 330. [40]

K. V. Sankar, D. Kalpana, R. K. Selvan, Electrochemical properties of microwave-

assisted reflux-synthesized Mn3O4 nanoparticles in different electrolytes for supercapacitor applications, J. Appl. Electrochem.42 (2012) 463. [41]

S. K. Meher, P. Justin, G. R. Rao, Microwave-mediated synthesis for improved

morphology and pseudocapacitance performance of nickel oxide, ACS Appl. Mater. Interfaces 3 (2011) 2063. [42]

Y. Wang, Q. Chen, Dual-layer-structured nickel hexacyanoferrate/MnO2 composite as a

high-energy supercapacitive material based on the complementarity and interlayer concentration enhancement effect, ACS Appl. Mater. Interfaces 6 (2014) 6196. [43]

L. Li, H. Song, Q. Zhang, J. Yao, X. Chen, Effect of compounding process on the

structure and electrochemical properties of ordered mesoporous carbon/polyaniline composites as electrodes for supercapacitors, J. Power Sources 187 (2009) 268. [44]

X. Wen, D. Zhang, L. Shi, T. Yan, H. Wang, J. Zhang, Three-dimensional hierarchical

porous carbon with a bimodal pore arrangement for capacitive deionization, J. Mater. Chem. 22 (2012) 23835. [45]

L.-L. Chen, X.-L. Wu, Y-G. Guo, Q.-S. Kong, Y.-Z. Xia, Synthesis of nanostructured

fibers consisting of carbon coated Mn3O4 nanoparticles and their application in electrochemical capacitors, J. Nanosci. Nanotechnol. 10 (2010) 8158.


D.-Y. Youn, H. L. Tuller, T.-S. Hyun, D.-K. Choi, I.-D. Kim, Facile synthesis of highly

conductive RuO2-Mn3O4 composite nanofibers via electrospinning and their electrochemical properties, J. Electrochem. Soc. 158 (2011) A970. [47]

D. Bhattacharjya, J.-S. Yu, Activated carbon made from cow dung as electrode material

for electrochemical double layer capacitor, J. Power Sources 262 (2014) 224. [48]

M. Ghaemi, F. Ataherian, A. Zolfaghari, S. M. Jafari, Charge storage mechanism of

sonochemically prepared MnO2 as supercapacitor electrode: Effects of physisorbed water and proton conduction, Electrochim. Acta 53 (2008) 4607. [49]

S. E. Moosavifard, J. Shamsi, S. Fani, S. Kadkhodazade, 3D ordered nanoporous

NiMoO4 for high performance supercapacitor electrode materials, RSC Adv. 4 (2014) 52555. [50]

Z. Lei, F. Shi, L. Lu, Incorporation of MnO2-coated carbon nanotubes between graphene

sheets as supercapacitor electrode, ACS Appl. Mater. Interfaces 4 (2012) 1058. [51]

Q. Wu, Y. Xu, Z. Yao, A. Liu, G. Shi, Supercapacitors based on flexible

graphene/polyaniline nanofiber composite films, ACS Nano 4 (2010) 1963.

Table 1. Surface area, pore volume and electrochemical performance of nanoparticles (MN0.33:1), nanorods (MN0.5:1), porous nanofibers (MN1:1) and densed nanofibers (MN2:1)

Sample Id





Surface area (m2 g-1) Pore volume (cm3 g-1) Scan rate ( mV s-1)

13 17 24 15 0.098 0.118 0.027 0.078 -1 Specific capacitance Cs (F g ) using CV analysis

3 10 30 50 Current density (A g-1)

95 107 190 126 66 76 160 101 54 60 140 80 48 49 130 70 -1 Specific capacitance Cs (F g ) using GCD analysis

0.3 0.6

58 52

65 54

210 160

86 75 Jai Bhagwan et al

Fig. 1. Schematic diagram of electrospinning set-up. Fig. 2. (a) TGA curve of as-spun nanofibers and (b) sintered Mn3O4 nanofibers (MN1:1). The TGA experiments of as-spun nanofibers and sintered Mn3O4 nanofibers are conducted in air from room temperature (RT) to 700 0C and to 900 0C, respectively. Fig. 3. XRD analysis of sintered Mn3O4 nanofiber (MN1:1). Inset shows the rietveld refined XRD data. Fig. 4. Unit cell of spinel Mn3O4 generated by refined XRD data. Fig. 5. FTIR analysis of (a) PVP, (b) Mn(CH3COO2)2.4H2O, (c) as-spun nanofiber of Mn3O4/PVP, and (d) Mn3O4 nanofiber (MN1:1) sintered at 350 0C. Fig. 6. FESEM image of as-spun nanofiber (a) MN0.33:1, (b) MN0.5:1, (c) MN1:1 and (d) MN2:1 shows the high aspect ratio with the diameter in the range of 200 to 300 nm. Fig. 7. FESEM images of sintered Mn3O4 nanofiber for polymer to metal precursor ratio (a) MN0.33:1, (b) MN0.5:1, (c) MN1:1 and (d) MN2:1. (e) Schematic scheme for the fabrication of nanoparticles and nanofibers/nanorods. Fig. 8. TEM image showing aligned and high aspect ratio and inset shows corresponding SAED pattern. The rings/bright spots show the various plane of tetragonal Mn3O4. Fig. 9. N2 adsorption-desorption isotherm of the obtained MN1:1 nanofiber with the polymer concentration of 7%. The inset shows pore size distribution of the obtained Mn3O4. Fig. 10. Cyclic voltammetry (CV) curves (a) and (b) at different scan rate whereas (c) and (d) specific capacitance as function of scan rate in the electrolyte system of 1M KCl and 1M Na2SO4, respectively. Inset shows the relationships between average current and the square root of the scan rate ʋ1/2.The polymer concentration was kept 7% (w/w). Fig. 11. Charge-discharge curves in two different electrolytes of 1M KCl and 1M Na2SO4 in the potential window of 0.0V to 0.9V vs Ag/AgCl and 0.0V to 1.0V vs. Ag/AgCl, respectively. Fig. 12. Specific capacitance as a function of cycle number. Fig. 13. Ex-situ FESEM image shows that the Mn3O4 nanofiber (MN1:1) maintain their mechanical integrity in terms of nanofabric morphology dispersed into carbon with pores even after 500 cycles. Fig. 14. Nyquist plots of MN0.33:1, MN0.5:1, MN1:1 and MN2:1 in 1M KCl electrolyte. Inset equivalent circuit and high frequency region (after zooming).

Jai Bhagwan et al

Fig. 1

Jai Bhagwan et al

Fig. 2

Jai Bhagwan et al

Fig. 3

Jai Bhagwan et al

Fig. 4

Jai Bhagwan et al

Fig. 5

Jai Bhagwan et al

Fig. 6

Jai Bhagwan et al

Fig. 7

Jai Bhagwan et al

Fig. 8

Jai Bhagwan et al

Fig. 9

Jai Bhagwan et al

Fig. 10

Jai Bhagwan et al

Fig. 11

Jai Bhagwan et al

Fig. 12

Jai Bhagwan et al

Fig. 13

Jai Bhagwan et al

Fig. 14

Jai Bhagwan et al