A CLOSE LOOK AT ELECTROLYTIC MANGANESE DIOXIDE … · ISSN Figure 5. XtalDraw© [20] rendition of a 2:1 De Wolff regular-interstratified EMD with Prr = 0.5. and broadenings in reasonable

A CLOSE LOOK AT ELECTROLYTIC MANGANESE DIOXIDE (EMD) AND THE γ-MnO2 & ε-MnO2 PHASES USING RIETVELD MODELING

D. E. Simon, R. W. Morton, and J. J. Gislason

DES Consulting, 5561 Chickering Court, Bartlesville OK 74006

Route 1 Box 343, Bartlesville OK 74003 4733 Dartmout Drive, Bartlesville OK 74006

ABSTRACT Electrolytic Manganese Dioxide (EMD) material was analyzed by Rietveld refinement of x-ray diffraction patterns to answer the question, “Is EMD composed of gamma- or epsilon-MnO2?” An electron diffraction study of EMD using a 20 nanometer spot size electron beam reported observing only the ε-MnO2 structure and no γ-MnO2 structure. However, a Transmission Electron Microscopy (TEM) study of EMD observed only γ-MnO2 structure as revealed by atom planes with a 0.4 nanometer spacing along with crystal twinning. Rietveld refinement results of EMD x-ray diffraction patterns indicate that EMD can be adequately described using both the gamma- and epsilon-manganese dioxide (γ-MnO2 and ε-MnO2) phases with an occasional occurrence of pyrolusite (β-MnO2). It is proposed that the ε-MnO2 structure observed in both electron and x-ray diffraction patterns is only a signature of a disordered manganese occupancy of the long range hexagonal oxygen framework and not a discrete phase, and EMD material predominately composed of short range ordered γ-MnO2. INTRODUCTION X-ray diffraction (XRD) has been one of the physicochemical properties frequently used to probe the secrets of the excellent battery activity exhibited by EMD. Much of the voluminous literature on this subject is cited in several reviews [1-4]. The EMD’s employed in alkaline cells typically exhibit poor quality powder patterns, which are described at best as a small number of broad peaks on top of an undulating background (Figure 1). The peak positions and widths vary among samples deposited under different conditions. The pattern characteristics observed signify disorder, which has been thought to be the origin of the battery activity, especially since the closely related polymorphs of EMD, i.e., pyrolusite (β-MnO2) and ramsdellite (an uncommon mineral), possess a high degree of order but poor alkaline battery activity. Battery activity has also been associated with chemical non-stoichiometry, in which Mn3+ ions and protons substitute for Mn4+ ions in the MnO2 lattice [5].

Copyright ©JCPDS - International Centre for Diffraction Data 2004, Advances in X-ray Analysis, Volume 47. 267 ISSN 1097-0002

http://www.icdd.com/

This document was presented at the Denver X-ray Conference (DXC) on Applications of X-ray Analysis. Sponsored by the International Centre for Diffraction Data (ICDD). This document is provided by ICDD in cooperation with the authors and presenters of the DXC for the express purpose of educating the scientific community. All copyrights for the document are retained by ICDD. Usage is restricted for the purposes of education and scientific research. DXC Website – www.dxcicdd.com

ICDD Website - www.icdd.com

ISSN 1097-0002

http://www.dxcicdd.com/

http://www.icdd.com/

The crystal structure of EMD is closely related to the beta-, epsilon-, and gamma- polymorphs of MnO2, all comprised of a hexagonally close packed lattice of O2- anions with the Mn4+ cations filling one-half the octahedral sites in the oxygen lattice. The difference in the above polymorphs lies in the arrangement of the Mn4+ within the octahedral sites. In both forms, the [MnO6] octahedrons link to other [MnO6] octahedrons so as to produce [MnO6] chains parallel to the c-axis and forms tunnels between these chains[1]. In pyrolusite (Figure 2), the [MnO6] units form (1x1) tunnels, whereas, in ramsdellite (Figure 3), characterized by Bystrom [6], the [MnO6] octahedral units form (1x2) tunnels. De Wolff [7] analyzed some line-rich γ-MnO2 samples and suggested that γ-MnO2 could be described as a random intergrowth of layers of ramsdellite (Figure 4) and pyrolusite (Figure 5) where Prr is the probability of two pyrolusite being adjacent to each other in the structure. With certain assumptions, De Wolff accounted for the line shifts. Figure 1. Typical X-ray diffraction pattern of an EMD material.

0

45

90

135

180

225

270

315

360

405

450

Inte

nsity

(cps

)

10 20 30 40 50 60 702theta (degrees )

80 90 100 110 120

SSA --- 29.6 m2/gm

Inte

nsity

, c/s

2Θ, Degrees


Figure 2. XtalDraw© [20] rendition of pyrolusite – 1x1 tunnels. Figure 3. XtalDraw© [20] rendition of ramsdellite – 1x2 tunnels. Figure 4. XtalDraw© [20] rendition of a 2:1 De Wolff regular-interstratified EMD with Prr = 0.0.


Figure 5. XtalDraw© [20] rendition of a 2:1 De Wolff regular-interstratified EMD with Prr = 0.5. and broadenings in reasonable agreement with experimental data (cf. Ref. 3 for a lucid analysis). The single chains (pyrolusite-like) of [MnO6] octahedrons in the ramsdellite matrix are often termed “De Wolff Disorders”. In other studies, a number of investigators found that EMD’s exhibited hexagonal symmetry rather than the orthorhombic symmetry of γ-MnO2. In the first of these studies, De Wolff, Visser, Giovanoli and Brutsch [8] interpreted EMD deposited at a relatively high current density to be ε-MnO2, where the Mn4+ are distributed randomly in one-half of the octahedral sites. The broad diffraction line at d ~ 0.42 nanometer could not be indexed, and was rationalized as a consequence of a partial ordering between Mn4+ ions and vacancies within the octahedral sites (avoidance of simultaneous occupation by Mn4+ ions at the center of octahedrons sharing faces). Similarly, Heuer, He, Hughes, and Feddrix [9] interpreted transmission electron microscopy and electron diffraction patterns of an EMD material to be ε-MnO2 and they proposed a model for ε-MnO2. Again, the broad diffraction line at 0.42 nanometers could not be indexed and it was postulated that this peak was due to possible vacancy ordering in the ε-MnO2. Chabre & Pannetier [3] showed in a theoretical study that the micro-twinning of γ-MnO2 along the twin planes {021} and {061} may demonstrate features not explained by De Wolff Disorders. The net effect of this work was to formulate a phase diagram with two variables: percent De Wolff disorder versus percent micro-twinning. This phase diagram accommodates ramsdellite, all chemically deposited manganese dioxide (CMD), EMD and heat-treated EMD (which also have β-MnO2 peaks). Using numerical simulation methods, Chabre and Pannetier [3] showed that De Wolff Disorders did not explain the lack of resolution that characterizes the XRD pattern of most EMD’s and CMD’s. They simulated the effect ascribed to giving the 0.42 nanometer line in the case of ε-MnO2, and found that no line of the observed magnitude resulted; all that resulted was a step in the diffuse background and not a peak. Therefore, they discredited ε-MnO2 as a structure in EMD material.


The γ-MnO2 and ε-MnO2 concepts were coupled, well before Chabre and Pannetier’s work, giving a useful correlation between the XRD pattern of EMD and the deposition current density. This was done through the introduction of the “Q ratio” of the peak heights at 2Θ = 22.0o and 37.0o (Cu Kα radiation), these being the most characteristic peaks of γ- and ε-MnO2, respectively [9]. Preisler [10] and later investigators [11,12] found that the Q ratio (referred to as the γ-/ε- character) decreases as the deposition current density increases. The B.E.T. surface area and other related features of porosity (e.g., pore volume & bulk density) also monotonically change as the deposition current density increases, and hence these properties correlated with the Q ratio [10]. Simon, Andersen, and Elliott [13] described an EMD model using Rietveld refinement analysis of EMD x-ray diffraction patterns where it was proposed that EMD material is composed of γ-MnO2, and ε-MnO2 plus or minus β-MnO2 phases. They showed that their model worked very well with samples having specific surface areas ranging from 10 to 86 m2/gm. Their model assumes that EMD is characterized as a binary mixture of γ-MnO2 and ε-MnO2 crystallites with different crystallite domain sizes based on Rietveld refinement analysis. The major conclusion reached was that the crystallite domain size of the ε-MnO2 is approximately 3 times that of the γ-MnO2. Also, the ratio of ε-MnO2 to γ-MnO2 was essentially constant at a value of 1.5 throughout the surface area range. In a second paper, Simon, Andersen, and Elliott [14] described a revised structural model for EMD material composed of only small crystallite domain sized crystals of γ-MnO2 in the range of 15 to 50 angstroms. The ε-MnO2 diffraction pattern portion was described as the signature of the oxygen framework with a crystallite domain size approximately 3 times that of the γ- MnO2

and probably not a discrete phase in EMD materials. The goals of this study were (a) to apply Rietveld refinement by assuming a binary mixture of manganese dioxide phases and thus to treat such a model quantitatively and (b) to describe a physical model of the EMD structure (“in light the reported” isn’t good English) that explains the reported contradiction between transmission electron microscopy and electron diffractometry. EXPERIMENTAL Sample Preparation: Eight commercially available EMD samples were chosen for this study and ground to less than 2.5 nanometer particle size for use in x-ray diffraction. Experimental X-Ray Patterns: Patterns were obtained with a Siemens D-500 diffractometer, with Cu Kα radiation from a long, fine-focus tube. A curved graphite monochrometer served to suppress the Kβ and background radiation. Settings were: tube voltage = 40 kV; tube current = 35 mA; detector voltage = 1010 V; 0.1 degree receiving slit, 1 degree scatter slit; 0.02 degree 2Θ /step; 2 sec/step time interval; scan range = 4-120 degrees 2Θ . The collected data were directly used in the Rietveld refinement.


Rietveld Procedure: The instrumental background profile was determined from an x-ray scan of a “low background holder” without a mounted sample. The background profile is similar to the background encountered for the pyrolusite diffraction pattern (Figure 6). Once the background shape is numerically defined, the net x-ray intensity above background is assumed to be contributions from the phases in the sample. Crystallographic descriptions of the phases were as follows: Bystrom’s [6] data for ramsdellite were used for γ-MnO2 and the data of De Wolff et al. [8] were used for ε-MnO2. These references provided the space group assignments and initial input data for the unit cells, atom positions and thermal parameters. Rietveld refinement was applied to each sample, and the lattice parameters and peak width parameters were refined to fit the pattern. The thermal parameters for the Mn and O were set equal in both phases. Output from the refinement, in addition to the XRD patterns for the component MnO2 phases, includes the lattice parameters, weight percent, and crystallite domain size of each phase. The peak width at half-maximum height is related to the crystallite domain size through the Scherrer equation [16]. Williamson Hall [21] calculations were perfomed using the peak width relationship with respect to 2Θ obtained from the Rietveld refinement to separate the strain contribution to observed peak width from the crystallite domain contribution to the observed peak width. Both γ-MnO2 and ε-MnO2 structures (and β-MnO2 when present) contribute to x-ray diffraction intensity, and together with background, make up the entire pattern. The intensity contribution from each phase and background are additive for every 2Θ data Table 1. Data obtained from the Rietveld refinement of x-ray diffraction patterns on eight commercially available EMD materials.

1 2 3 4 5 6 7 8

Gamma – MnO2 Weight % 50 ± 1 50 ± 1 51 ± 1 48 ± 1 52 ± 1 47 ±1 52 ±1 48 ± 1 CDSs - Ǻ 26 35 35 36 30 37 34 31

CDSwh – Ǻ 29 37 36 35 34 42 38 33 Strainwh – ∆L/Lx10-3

< 0.01

< 0.01

< 0.01

< 0.01

< 0.01

< 0.01

< 0.01

< 0.01

Epsilon – MnO2 Weight % 50 ± 1 50 ± 1 49 ± 1 52 ± 1 48 ± 1 53 ± 1 48 ± 1 46 ±1 CDSs - Ǻ 113 119 119 104 113 101 106 103

CDSwh – Ǻ > 2000 > 2000 > 2000 > 2000 > 2000 > 2000 > 2000 > 2000Strainwh – ∆L/Lx10-3

1.59

1.58

1.46

1.86

1.67

1.77

1.87

1.86

Beta – MnO2 Weight % ND ND ND ND ND ND ND 6 ± 0.5


point. The lattice parameters of both phases are allowed to vary along with the peak width parameters from which the crystallite domain sizes of both phases are estimated. The background contribution is modeled as a smooth declining curve with increasing 2Θ angle and is based on the characteristics similar to x-ray diffraction patterns of β-MnO2 (Figure 6). RESULTS AND DISCUSSION General Pattern Features Figure 7 shows a typical Rietveld refinement of an EMD material and includes the raw and refined XRD patterns, the difference pattern (raw – refined), and the refined patterns for the two individual phases modeled. Results of the Rietveld refinement are tabulated in Table 1. Noteworthy features from the refinement plot and data in Table 1 are as follows:

a. The calculated model defines the experimental patterns very well. b. The background is low and uniform, in contrast to what was considered to be

a wavy undulating background is actually the lower relative intensity diffraction peaks from the small crystallite domain γ-MnO2 phase.

c. The γ-MnO2 peaks are much broader than the ε-MnO2 peaks. d. The reported weight ratio of γ-MnO2 and ε-MnO2 is roughly 1:1. When

present, β-MnO2 does not appear to change this relationship. e. The γ-MnO2 crystallite domain size (CDSs) using the Scherrer Equation varies

from 30 to 40 Å. This is in agreement with crystallite domain sizes using Williamson-Hall calculations (CDSwh), based on the refined peak width parameters, show that the γ-MnO2 CDSwh is 30 to 40 angstroms and are strain free crystals.

For ε-MnO2, the CDSs is approximately 120 Å. However, the CDSwh for ε-MnO2 calculates at greater than 2000 angstroms with high strain values evident (1.2 – 1.9 x 10-3 ∆ L/L values). Typical Williamson-Hall plots for both γ-MnO2 and ε-MnO2 phases are shown in Figures 8 and 9. Transmission Electron Microscope Image Study Transmission Electron Microscopy (TEM) was performed on several EMD samples and reported earlier by Simon, Andersen, and Elliott [21]. The samples were ultrasonically dispersed and then transferred to a carbon grid and lattice fringe images obtained. Figure 10 is an example of an EMD sample from this work. This TEM indicates that the ε-MnO2 structure is not observed; i.e., the TEM image reveals only γ-MnO2 structure, as


Figure 6. β-MnO2 XRD pattern showing smooth declining background. Figure 7. Contribution of background, ε-MnO2, and γ-MnO2 portions to the XRD pattern of EMD.

0

4 5

9 0

1 3 5

1 8 0

2 2 5

2 7 0

3 1 5

3 6 0

4 0 5

4 5 0

Inte

nsity

(cp

s)

1 0 2 0 3 0 4 0 5 0 6 0 7 02 t h e t a ( d e g r e e s )

8 0 9 0 1 0 0 1 1 0 1 2 0

S S A - - - 2 9 . 6 m 2 / g mR e f in e d P a t t e r nB a c k g r o u n dD if f e r e n c e P a t t e r nP t n : M n O 2 - e p s ilo nP t n : M n O 2 - g a m m a

2Θ, Degrees

Inte

nsity

, c/s

0

5 0

1 0 0

1 5 0

2 0 0

2 5 0

3 0 0

3 5 0

4 0 0

4 5 0In

ten

sity

(cp

s)

1 0 2 0 3 0 4 0 5 0 6 0 7 02 t h e t a ( d e g r e e s )

8 0 9 0 1 0 0 1 1 0 1 2 0

I B A # 6R e f i n e d P a t t e r nB a c k g r o u n d

2Θ, Degrees

Inte

nsity

, c/s


Figure 8. Williamson Hall plot of PWHM data for the γ-MnO2 phase in EMD showing no slope to the curve, indicating no strain in the structure. Figure 9. Williamson Hall plot of PWHM data for the ε-MnO2 phase in EMD showing significant slope to the data, indicating strain in the structure.

0.000

0.005

0.010

0.015

0.020

0.025

0.030

0.035

0.040

0.045

0.050

0.055

0.060

0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00

4 sin(theta)

PWH

M C

os(t

heta

)

0.000

0.005

0.010

0.015

0.020

0.025

0.030

0.035

0.040

0.045

0.050

0.055

0.060

0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00

4 sin(theta)

PWH

M C

os(t

heta

)


Figure 10. TEM lattice fringe image of EMD sample [14] showing crystal lattices of 0.4 nanometers typical of γ-MnO2 crystal structure. No evidence of ε-MnO2 stucture is observed.

2 nm


Figure 11. Theoretical long range ordering of γ-MnO2. Figure 12. Short-range order crystallites of γ-MnO2 due to micro-twinning and De Wolff disorder. Viewing of the whole area is representative of the ε-MnO2 disordered superstructure.


By definition, ε-MnO2 is composed of a framework of hexagonally close-packed O-2 anions with one-half of the octahedral sites randomly filled with Mn4+ cations. In contrast, γ-MnO2 has one-half of the octahedral sites filled with Mn4+ cations in an ordered configuration creating 1x2 tunnels between the MnO6 octahedrons. Likewise, (β-MnO2) has Mn4+ cations arranged in one-half of the octahedral sites in another ordered configuration, creating 1x1 tunnels between the MnO6 octahedrons. This ordering or disordering of the MnO6 octahedrons can lead to different diffraction patterns depending on the area being observed (greater or less than 10 nanometers). Figure 11 show a schematic of the long range order expected in a γ-MnO2 crystal having a large crystallite domain size of hexagonal close packed oxygen framework (this schematic shows only the Mn4+ ions at one-fourth above and below the plane of O2- ions at zero and one-half, respectively). X-ray diffraction patterns of this material would typically have narrow high intensity peaks. However, in contrast, Figure 12 shows how micro-twinning by Mn4+ creates small ordered crystallite domains of γ-MnO2 within the hexagonal close-packed O-2 ion framework. X-ray diffraction patterns of this material would typically have broad low intensity peaks typical of short range order. However, since long range order of the O2- ions exists in EMD material, as evidenced by the CDSwh > 200 nanometers, one could suspect that a superlattice signature may be present in the x-ray diffraction pattern. In EMD x-ray patterns, the ε-MnO2 pattern observed is considered to be this signature of the Mn4+ ion disorder throughout the long range oxygen framework (Figure 12). In other words, the ε-MnO2 phase found in EMD x-ray diffraction patterns results from the appearance of Mn+4 disorder within the long range ordered oxygen framework. However, the γ-MnO2 pattern is the result of the short range ordering of the Mn+4 ions with a pattern similar to that of ramsdellite (Figure 3) with micro-twinning being the boundaries of the short range order crystallites. These small ordered Mn4+ domains put strain on the O2- ion long range ordering as evidenced by the Williamson-Hall strain values ranging from 1.5 to 1.9 ∆L/Lx10-3 for the ε-MnO2 structure. Another criteria for this reasoning is the fact that the ε-MnO2 to γ-MnO2 concentration ratio is constant at 1:1. Thus mass is conserved between the two structures when refining the same crystalline arrangement with two different structures at the same time. CONCLUSION From the nature of electro-crystallization, we suggest that the large CDSwh disordered ε-MnO2 structure found in EMD X-ray diffraction patterns represents the dimensions of the hexagonal close-packed O2- framework, possibly due to the intersection of growth sites, which emanate from the nucleation sites. Within each growth site, the γ-MnO2 domain sizes represent the average distance between micro-twin boundaries.


Therefore, we propose that although the ε-MnO2 structure interpreted to be present in x–ray diffraction patterns of EMD, it is not a discrete phase for quantitative analysis. It is rather a superstructure signature of long range disordered Mn4+ ions in the ordered hexagonal close-packed oxygen framework. We consider EMD materials to be composed only of small ordered crystallites of γ-MnO2 crystallites that are related to each other by either micro-twinning within the large hexagonal close-packed oxygen framework. Therefore, depending on your point of view, EMD material must be composed of either the short range ordered γ-MnO2 or the long range disordered ε-MnO2 phase and not both phases. ACKNOWLEDGEMENTS

The authors gratefully acknowledge Kerr McGee Chemical LLC for supplying the samples used in this study. REFERENCES

1. R. G. Burns, in Battery Material Symposium, Vol. 1, Brussels 1983 ed. By A.

Kozawa and M. Nagayama, IBA, Cleveland, Ohio (1984) pp. 341-356. 2. J. B. Fernandes, B. D. Desai and V. N. K. Dalal, J. Power Sources 15 (1985) 209-237. 3. Y. Chabre and J. Pannetier, Progress in Solid State Chemistry 23 (1995) 1-130. 4. T. N. Andersen, Modern Aspects of Electrochem., Vol. 30, ed. by R. E. White, B. E.

Conway and J. O’M. Bockris, Plenum Press, New York (1996) 313-413. 5. P. Ruetschi, J. Electrochem. Soc. 131(1984) 2737-2744. 6. A. Bystrom, Acta Chem. Scand., 3 (1949) 163; cf. ICDD # 39-375. 7. P. M. De Wolff, Acta Crystallogr., 12 (1959) 341-345. 8. P. M. De Wolff, J. W. Visser, R. Giovanoli and R. Brutsch, Chimia 32 (1978) 257-

259. 9. A. H. Heuer, A. Q. He, P.J. Hughes, and F. H. Feddrix, ITE Letters on Batteries, New

Technologies & Medicine, 1 (2000) 76-80. 10. E. Preisler, J. Applied Electrochem. 19 (1989) 540-546. 11. T. N. Andersen, Prog. Batteries & Battery Materials 11 (1992) 105-129. 12. R. Williams, R. Fredlein, G. Lawrance, D. Swinkels and C. Ward, Progress in Battery

& Battery Materials, 13 (1994) 102-112. 13. D. E. Simon, T. N. Andersen, and C. D. Elliott, ITE Letters on Batteries, New

Technologies & Medicine, 1, (2000) 10-19. 14. D. E. Simon, T. N. Andersen, and C. D. Elliott, Poster presented at the 50th Annual

Denver X-ray Conference, August 2001, Steamboat Springs, CO. 15. T. N. Andersen and R. G. Moody, in Prog. Batteries & Battery Materials, 13 (1994)

1-11. 16. B. D. Cullity, Elements of X-ray Diffraction, Addison-Wesley Publishing Co., Inc.,

Reading, Mass, 1959


17. A. Kozawa, Batteries, Vol. 1, Manganese Dioxide, Ed. byi K. V. Kordesch, Marcel Dekker, New York, 1974, Chapt. 3.

18. E. Preisler and G. Mietens, Dechema Monogr. 109, Verlag Chemie, Weinheim (1987) 123-137.

19. M. Mauthoor, The Effect of Operating Parameters on the Properties of Electrolytic Manganese Dioxide, Ph.D. Thesis, Dept. of Eng., University of the Witwatersrand, Johannesburg, So. Africa (1995).

20. B. Downs and K. Bartelmehs, XTALDRAW, (1997). 21. G. K. Williamson and W. H. Hall, Acta Met. 1(1953) 22.


Documents

A CLOSE LOOK AT ELECTROLYTIC MANGANESE DIOXIDE … · ISSN Figure 5. XtalDraw© [20] rendition of a 2:1 De Wolff regular-interstratified EMD with Prr = 0.5. and broadenings in reasonable