Topotactic Relations Among Pyrolusite, Manganite, and Mnror

American Mineralogist, Volume 71, pages 805-814, 1986

Topotactic relations among pyrolusite, manganite,and MnrOr: A high-resolution transmissionelectron microscopyinvestigation

J*ms H. Rlsx Department of Geology, Arizona State University, Tempe, Araona 85287 Pnrnn R. Busncx Departments of Geology and Chemistry, Aizora State University, Tempe, Arizona 85287

Ansrnrcr Pyrolusite (MnOr, tetragonal) is commonly created in the natural environment by the oxidation of manganite (MnOOH, monoclinic). Such secondarypyrolusite typically displays anomalous, nontetragonalcharacteristics. In this high-resolution transmission electron microscopy (Hnrrvr) study, MnrO, (monoclinic) was found intergrown with pyrolusite that had formed from manganite.This occurrenceof MnrO, is ofinterest for severalreasons: (l) MnrO, has been identified in a natural manganite-pyrolusitemixture. Assuming that it is not an artifact of nnreu observation, this is the first report of naturally occurring MnrOr. (2) Aligned monoclinic MnrO, intergrown with pyrolusite may be partly responsible for the nontetragonalcharacter of secondarypyrolusite. (3) MnrOr-pyrolusite intergrowths occur adjacentto manganite.This occurrencesuggests that the oxidation of manganitemay form both pyrolusite and MnrOr. Another new Mn oxide has been identified through selected-areaelectron difraction. We hypothesizethat this phaseis a structural modification of pyrolusite. Ordering of OH groups and Mn atoms of different oxidation statesis a possibleexplanation for this modification.

fNrnooucrroN translations, a, and ar. Either of these should have an equal probability of becoming the manganite a (or b) Pyrolusite (MnOr), a tetragonal mineral with the rutile translation when pyrolusite is reduced to manganite. structure, is the most stable form of manganeseoxide in However, that is not what is observed. In the sequence manyterrestrialenvironments.Distinctionshavelongbeen manganite (primary) * pyrolusite (secondary) + man- recognized between the relatively rare primary form of ganite (secondary), primary and secondary manganite in- pyrolusite and the much more common secondaryform variably have the same orientations. A memory of the that occursas pseudomorphicreplacements of other man- original manganite orientation is conveyed by the pyro- ganeseoxide minerals, particularly manganite (MnOOH, lusite intermediate, but there is a question as to how this monoclinic). Primary pyrolusite has a hardness of 7, memory is transmitted. whereas secondary pyrolusite displays variable, much The nontetragonal characteristicsof secondary pyro- lower hardness values. Secondary pyrolusite also pos- lusite have been attributed to microstructures formed in sessesseveral characteristics suggestiveof a symmetry pyrolusiteuponitscreationfrommang;anite(Strunz, 1943; lower than tetragonal. For example, reflectedJight mi- Champness,l97l).Pyrolusiteandmanganitehave similar croscopy shows only one cleavagedirection as well as structures.The manganitea and c translationsare halved optical anisotropy in the (001) plane of secondarypyro- to form the a and c pyrolusite unit-cell translations,while lusite. Until Strunz (1943) showedthrough single-crystal D of manganite contracts from 5.28 Arc q.q0 A to form X-raymeasurementsthatthetwoformsofpyrolusitehave the other a translation of pyrolusite (Fig. l). This 150/o identicalcrystalstructures,primarypyrolusitewastermed contraction along b presents the possibility that micro- polianite and considereda distinct mineral. Later studies scopic cracksparalleling the manganite (010) planes sep- (de Wolfl 1959; Potter and Rossman,1979) have found arate newly made crystallites of pyrolusite. Images ob- that some secondarypyrolusites are actually orthorhom- tained by transmission electron microscopy confirm the bic. existenceof lamellar micropores in secondarypyrolusite Another anomalous property of secondarypyrolusite, (Champness, l97l). Such microcracks may explain the termed the memory effect,was noted by Dent Glasserand aberrant optical properties, the decreasedhardness, and Smith ( 1968).Ideally, pyrolusite hastwo equivalent lattice the great chemical activity and adsorptivity of secondary 0003{04xl86/0506-0805$02.00 805 806 RASK AND BUSECK:HRTEM STUDY OF Mn OXIDES

Table l. Unit-cell dimensions of pyrolusite, manganite,and MnrO,

Pyrol usi te Mangani te l'4n-0^ D6 a = 4.3999A a=8.98A a = 10.347A

c = 2.8140A b=5.28A b= 5.72A

c = 5.71A c = 4.852A

I = i09'25' = (#136) = (#i4) = (#12) SG P4^./mnnz sG B2rld SG czlm Source : Source l Source : Gl,*illr% Bauer,1976 Buerger, 1936 oswald and Hanpetich, 1967

for 20 h at 5 kV andfor 2 h at 1.5kV. A reor--rernr200CX 200- kV anda Philips400T 120-kVmicroscope were the instruments usedin this study.

Fig. l. A drawingto illustratethe dimensionsof the pyro- Cnysrl.r-r-ocRApHrc RELATToNSAMoNG lusite,manganite, and MnrO, unit cellsand their relativeori- MANGANTTE,PYROLUSTTE, AND MnrOt entationsin the topotacticreactions. The unit cells of pyrolusite, manga.nite,and MnrO, are closelyrelated (Baur, 1976;Buerger, 1936; Oswald et al., 1967), and in their topotactic transformations,their crys- pyrolusite, but probably cannot explain observed slight tallographic axes remain in nearly the same relative ori- deviations of such pyrolusite from tetragonal symmetry. entations. The a, b, and c axes ofmanganite correspond The memory efect may also be a result of these aligned directly to the a axes and the c axis ofpyrolusite. The a lamellar microcracks. axis of MnrO, is at an angle of 19'from one a axis of Here we report the frndings of high-resolution trans- pyrolusite and a of manganite; b of MnrO, corresponds mission electronmicroscopy (nnrnrvr) examinations of un- to the c translation of pyrolusite and manganite;and c of heated natural mixtures of manganite and pyrolusite and MnrO, has the same orientation as b of manganite, and similar examinationsof portions of thesemixtures heated a of pyrolusite (Dent Glasserand Smith, 1968).Pertinent in air to 300.C.We have observedmicropores, as reported crystallographic data are given in Table 1, and the ori- by Champness(1971), but we have also determined that entation relations among the unit cells of these minerals an intermediate phase, MnrO, (Mn;+Mn1+Or), plays an are illustrated in Figure L important role in the anomalous behavior of some sec- Sincemanganite has perfect (010) cleavage,crushed grain ondary pyrolusite. Our discovery of MnrO, in unheated, mounts commonly contain manganite particles oriented ion-milled, natural samples is of additional interest be- with the zone axis nearly parallel to the electron causeMnrO, in a natural occurrencehas not previously [010] beam. The ion-thinned sample also was examined only been reported. Furthermore, since MnrO, is an interme- in this orientation. The preponderanceofthis orientation diate phase in oxidation and reduction reactions of Mn has advantagesand disadvantages.Although the changes oxides,it should be given considerationin studiesof Mn- from manganite to pyrolusite and MnrOr are clearly ev- oxide phaseequilibria. ident in the selected-areaelectron diffraction (seeo) pat- ExpnnrvrnNTAl DETArrs terns taken from this orientation, micropores that parallel the manganite (010) planes are not observable. The specimensstudied are from the StanfordUniversity min- patterns in eral collection.They are labeledas manganiteoriginating from sAED of these minerals, all the orientation the Iake Superiorregion (sample no. 5l ll0) and from Ilfeld, correspondingto [0 l0] of parent manganite,are illustrated Harz,Germany (sample no. 7l 52).Powder X-ray diftaction (xno) in Figure 2. The reflections along the a* direction are revealedthese specimens to be mixturesof pyrolusiteand man- virtually identical in the manganite [010] and the pyro- ganite.To continuethe reaction ofmanganite to plnolusite,por- lusite [010] diffraction patterns. However, the manganite tionsofthe specimenswere powdered and heatedin air for 1.5 (101) spacingis double that ofpyrolusite, so the 101 and to 3 h. An xno powderpattern ofa samplethat hadbeen heated 101 manganite reflections appear halfuay between the at 300"Cfor 3 h showedno manganitereflections. This pattern origin and the more intense spots that correspond to the pyrolusitepeaks, contained,along with strong lessintense re- pyrolusite 101 and l0I reflections (in the manganitepat- flectionsattributable to MnrOr. Bothheated and unheated sam- terns theseare the 2OZana 202 reflections).The l0l, 10I, pleswere examined by nnrerraas grains mounted on holey-carbon in manganite film. To obtainbetter observations of the relationshipsamong 200, and 200 reflections do not appear the intergrownphases, an ion-thinnedsample ofthe LakeSuperior pattern becauseof the ll + l: 4n requirement for allowed specimenwas preparedand observedbefore and after heating. ftOl reflectionsimposed by the diamond glide in the man- In themilling process, ionized argon bombards the samplein an ganite structure. evacuatedchamber. The lake Superiormanganite was ion-milled The [001] MnrO, pattern was commonly observed to RASK AND BUSECK: HRTEM STUDY OF MN OXIDES 807

MnO o o202 58 o o 101 o oo 000 o o 101 o o

MANGANTTE [o1o]

oo Mnn. ooMn'. OO o Fig. 3. A schematic illustration of a MnrOr-pyrolusite inter- o o growth as seenlooking down the pyrolusite [001] axis (MnrOt r01 [010]). Unit-cell boundariesare shown by dotted lines. The rep- resentationsof pyrolusite and MnrO, are adapted from Dent Glasserand Smith (1967). In the illustration ofpyrolusite, filled circles represent oxygen and Mn4+ at t/zc; open circles represent oxygenand Mna+ at c:0. In the MnrO, illustration, open circles o Oooo oroo representoxygen at D : 0, 0.500; Mna+ at D : 0; and Mn2+ at D: 0; whereas filled circles represent oxygen at 0.25b,0.75b; Mna+at 0.25b.0.75b:and Mn2+at 0.500.Note that b of MnrO. is approximately twice the length of c in pyrolusite. Therefore, in the third dimension of this illustration, the oxygens are at similarheights in the two structures,as are the Mn4+ ions' Hnruv o o images of MnrOr-pyrolusite intergrowths shown in subsequent figuresare looking down the pyrolusite [010]-MnrO, [001] axis. The bold arrow to the right ofthe illustration indicates this view- ing direction. PYRoLUsTTE[oro]

be superimposed on a more intense pyrolusite [010] pattern, indicating an intergrowth relationship between these o2ot"tn-o^ lolouro two phases. In such patterns, the 200 reflection of MnrO' ooo is located slightly less than halfway between the central spot and the pyrolusite 200 reflection (MnrOr droo: 4.88 oooo A and pyrolusite d,* : 4.40 A;. Thus, the MnrO, 200 1l otrnro, reflection is readily distinguished from the 100 reflection of pyrolusite that, though kinematically forbidden, ap- OooOooO pears in some diffraction patterns of these intergrowths. ooo 2oorunuo, 2ooou'o In intergrowthsof pyrolusite and well-crystallizedMn'Or, diffraction spotsrepresenting MnrO, (l l0), (II0), (110), oooo and (110) planes appear approximately halfway between the origin and diffraction spots representing pyrolusite o o o (l0l), (101),(101), and (101)planes. MnrOg: A Ntw MTNERAL? Klingsberg and Roy (1960), in a study of the Mn-O pyRoLUSrrE [oro] ana unro, [oor) system, recognizeda new compound in several of their products. It is apparent from their published xRD Fig. 2. Illustrations of pyrolusite,manganite, and MnrO, s.orro run (1965) patterns. The orientation of each of these sAEDpatterns corre- data that the compound is Mn'O'. Oswald et al. pattern spondsto the [010] orientation of parent manganite.This is the used CdrMnrO, as a model to index the xno of orientation of all subsequentHRrEM images of pyrolusite, man- MnrO.. The structureofMnrOr, asdeterminedby Oswald ganite, and MnrOr. and Wampetich (1967), is basedon layers of edge-shared 808 RASK AND BUSECK: HRTEM STUDY OF Mn OXIDES

MnO. octahedra.Thus, in the classificationscheme pro- xno, and it is not evident in any ofseveral published xno posedby Turner and Buseck( I 98 l), MnrO, is a I x oo Mn patterns of secondary pyrolusite (Potter and Rossman, oxide. One quarter of the octahedral positions in these 1979;de Wolfl 1959).Even if present,MnrO, may elude layersare vacant. Mn2+ atoms occupy interlayer sitesad- xnp detection becauseof poor crystallinity or scarcity. jacent to the vacant octahedra.Several methods by which However, given the easy synthesis of MnrO, from the MnrO, may be synthesizedhave been reported (Klings- relatively common mineral manganite,it seemslikely that bergand Roy, | 9 59; Oswaldet al., I 965; Feitknecht, 1964; MnrO, does have natural occunences. Dasgupta,1965; Davis, 1967).Burns and Burns (1979) included MnrO* in their review of Mn oxides. OnrnNron TNTERGRowTHSoF Identification of MnrO, in unheated, natural samples PYRoLUSrrE,rNo MnrO, is significant becausesuch an occurrence has not been Figure 4 showselectron-diffraction patterns and images previously reported. We have identified poorly crystal- taken from the unheated ion-milled sample of the Lake lized MnrO, intergrown with pyrolusite in the unheated, Superior manganite. This sample consistsof domains of ion-milled Lake Superior manganite sample. seeo pat- manganite and domains of pyrolusite intergrown with ternsand high-resolutionimages from the pyrolusite [010], poorly crystalline MnrOr. The domains are at slightly dif- Mn,O, [00 1] orientation suggestthat the octahedrallayers fering (+51 orientations, all near to the [010] pyrolusite- of MnrO, are intermeshedwith the I x I octahedraltunnel [0 l0] manganit+[00 I ] MnrO, orientation.Figure 4a shows structureof pyrolusite, as illustrated schematicallyin Fig- an image and saep pattern of manganite. In the image ure 3. The 4.40-4, tunnel dimension of pyrolusite, cor- mode, manganite can be recognized by a uniform ap- respondinglo droo,is significantly smaller than its coun- pearance,with little evidenceof crystal strain or defects. terpart in MnrOr, the 4.88-A interlayer dimension that Pyrolusite-MnrO, intergrowths typically have a mottled correspondsto droo.Accommodations for this misfit must appearance,probably as a responseto crystal strain in- be present in these intergrowths. Becauseno high-reso- curred during the manganite-to-pyrolusitetransition. lution imageshave been obtained from the necessaryori- Intergrowthsofpyrolusite and poorly crystalline MnrO, entation ([001] pyrolusite, [010] MnrOr), the nature of occur adjacent to manganite. Figure 4b shows a high- such accommodationsis unknown. resolution image of such an intergrowth and its sAED Our identification of MnrO, from unheated samplesis pattern, which is a strong pyrolusite [010] pattern with basedtotally on HRTEM,which hasthe inherent possibility additional weaker reflections along a*. The weak 4.9-A that samples may be altered upon interaction with the reflections and the corresponding lattice fringes of the electron beam. Also, alteration of manganite could have image are interpreted as representingthe (200) planes of occurred during ion-milling. Thus we cannot conclude MnrO, and provide the evidence for MnrO. in this un- unequivocally that MnrO, occursin nature. In fact, when heatedsample. The lack of MnrO, I l0 spots (as occur in an electron beam of maximum intensity (achievedby re- the saeo patternsin Fig. 5) suggestsMnrO, is orderedwell moval of the condenseraperture) is focusedon a particle enoughto produce distinct sepo reflectionsonly along a*. of manganite, the manganite is quickly transformed to The ordering along b, involving vacanciesin the layers MnrOr. and interlayer Mn2+, is poorly developed. We believe, however, that most of the Mn5O, that we MnO, with the diaspore structure (ramsdellite in its observedwas not formed through such decomposition of pure form) forms intergrowths on the unit-cell scale with manganite in the electron beam. Experiencehas shown pyrolusite(de Wolfl 1959;Giovanoli et al., 1967).Nat- that manganite is realtively stable under a beam in the urally occurring specimensof this intergrowth are called usual configuration (condenseraperture in). AIso, MnrO, nsutite, while synthetic preparations of it are termed created from manganite by an intense electron beam is 7-MnOr. We are confident that the phaseintergrown with pure, whereasmost MnrO, observedunder normal beam pyrolusite in Figure 4b is MnrO, and not ramsdellite. The conditions is intergrown with pyrolusite. Finally, MnrO, 4.9-A spacing of the weak spots along a* in our saep unquestionablyincreased in abundanceafter the samples patternsis too largeto attribute to the only plausiblerams- were heated outside the microscope.This increase,noted dellite spacing,dozo: 4.64 A. Also, the a* direction in in xno as well as HRTEMresults, proves that the MnrO, sAEDpatterns of unheated samplesis identical to the a* under consideration is not solely an artifact of Hnrevr direction seenin seeo patterns of definite MnrOr-pyro- observation. lusite intergrowths from heated samples. We have yet to detect MnrO, in unheated samplesby Intergrowths of pyrolusite and well-crystallized MnrO,

--) Fig. 4. sAEDpatterns and micrographs ofthe unheated,ion-thinned Lake Superior specimenofpyrolusite and manganite. (a) SAEDpattern and image of manganitein the [010] orientation. The setsof fringes that run diagonally acrossthe image correspond to manganite(202) and (202) spacings(2.4 A) O) CorrespondingsAED pattern and image of intergrown pyrolusite (zone [010]) and MnrO. (zone [001]). Reflectionsthat representMnrO, (200) planes(marked by arrows) have appearedat 4.9 A-, along a*. Streaking along a* in this s,lro pattern indicates that disorder is present along the a direction of MnrO* or pyrolusite. In the nnrErr,rimage, horizontal fringes representingMnrO, dr*are superimposedupon the diagonal pyrolusite (l0l) and (l0I) fringes. i i.r

)'tl

608 saclxo uw co ^cnrs nlaruH ilJgsns oNV)svu sacxo uI r do A(InJs wiIJuH ryJgsng (INv)svu 0r8 'se8uuJ 8I eseqlJo sEuDBdspoJnsseru eqlJo Jora plueur ecueprlepepp€ epr ordcA emEtd ut seEuu;gtTrry.:J+,^ -uedxa ulqlly' sr onp^ slql '(cS 'Erg;o urelled cavs eq1 o'ul I Jo ur lueserd,{.1enpulnq'purlxe,{.pesoddns - e1rsn1or,(drog uorlueJc'flqeurnserd'pue elruetueru go EutleeqseAIoûI oo'p) oo?) eq,{.eur V ov'v pue ('o'uw roJ Y 88'? ere ca emErg ur qcrqrn'1cege,{:orusureql JoJlueuodrur,(letcedse 8gsutr41go 'ettsn1or,{d saEuuggJrol i ern JoJelqrsuodser sEurceds EurddelJoô er{l 1ueruufi1eeqa fiepuocesJo scqsua1ce '6I -reqc lBql eunssB e1{'JIv e Jo Eurcedse se,rrE1 uo4enbg leuo8erlsluou oql esnec dleq ,(eu leql sqlmortrel 'uorlelueuosrtu 8gsuy41-elrsnyor,(d 3o epue eql sr p eJeqa -ur eql olur ,(dorlosrue uE secnpor1ln tO'uIlrI cruqcouorugo qcns 'uoq4uolro (d '(p -'p)/'pa:lu 1ueuru61p II00] eql ur sr (suortetueuo [OtO] rteql u1 elrsnlodd pue eltuut 'e1tue :11 rre Jo srx€ tuuler,r eq1 lnoq? -ueur .{leurtuo seznqcrqrn) eldurus peuulql-uol pelesq totu4 setuug gJroINequo uorlsloJ ? ur lFseJ secr1el Eurddepeô srll w elrsnlor,{d qlul rurorErelur pe^Jesqo I1y 'eJouueggnd 'secqlel '(qS 'Etg) egl Jo suorlelueuosrtu oATIBIOJ peseeJculst Eurleoqeql s? eseeJcur Eurddulra,roeq1 ur seueld;o s8urcedseJe zp pue Ip ereqan ol sreeddesqyrrorErelur 8gsul4i-s1rsn1or[d eq1 ol eÎluIeJ '8OtuW 8gsu141pezrlep,{-rc-11e,tgo lunotue eql pIrE4rsnl (t) '(p -'p)/('p Ip) : ds + -or[d umorErelurJoseoJe JaErq ur spuelsl sBJncco seteurr :uorlenbe eql fq ue,rr8sr wsruH a4q-qruoc,buoq'puorsuetulp-orl\l ecnpord 1eq1 'ds 'Eurcuds 'OtuIrrI '(qS 'eg 'strg) setuul qcns;o eq1 ..'secrllel EurddEIJeo o/r\l ur Jo seerepez1lu1sf-rc-1le1\ seldures sruoleJo seueldpelceford eqtgo ocueprcuroc,,ruo{ {nseJ pel€equn uI ueql luspunqE eJou pu? psJepJorelleq sr '(/S6I) 'le totutr{ 'seldures 'seldures seEuu;grlol I ts felqse4 o18urprocry'eôqs qcns uI peteeq I1uw uortque 8gcu141-e1rsn1orf.d to'un-[gl6] peqlrcsep se JncJo sqlnorErelur leql -lro [I00] e1rsnlor,(deqt uI pe^resqo ore^\

'le{c?rq e ,{q pelecrpur ere *'p totul/{ pu? mlp elrsn1or,{duoo^ugq lgsru oql 'C.O6Z 'c) '(OOZ)tgtuf'I ruo{t lpser leql se8uuggrlol I f? q g'I roJ peleoq elduresur sqlnor8relur elrsnlorfd-8o'ul I (e,roqe rueserder seEuu; pluozuoq eql seerega'(0lI) totu6 lueserder se3uugpuoSerp eqt'ure8y'C.00€ 13 q g'Z I€uoF1pP?ue pel?eg oldtu?s 'seueld 'olrsnlorfd 8OsuI etueseqt ur sqlnorSrelur snoEopuy (UeI'O (O0Z)totuw lueserderseSuug I"luozuoq eqJ pu? I pexllrrelul Jo eereue slueserderflqeqord et?tur eqlJo uorUodrei$ol eql'totufrlJo (Olt) pue (911) lueserderse8uu; leuoEerp eq1,'8OsuW 'ornd 'J"00ZP pezqplsdrclort s? peleJfuelur sr ecuereeddee{I-quoc,(euoq e s?q leql e3eurreqlJo uoluod Joddn eqJ II 9'I roJ pepeq eldues ur sqlnorfuolur olrsnloJfd-sotufrl (tJoI'e)'rre ur Sqteeq rege elruetueu pue e1rsn1or,{d;oueulceds Jouedns e)P'I 'S 'tlC peuurql-uor er$ rn pe^Jesqoere,n sqlmorfuelu eseqJ 'sqlnor?relur 8gcu141-e1rsn1or,(dgo sqdaSoJcrlu pu? su.relledorvs

II8 saclxo uw co ^cnrs waJuHryJasns (INv )sYu -uI '(11ggu1,q) uB punoJ (0S0t) ,(oU pue EreqsEurpl otl relJe ueurcsds rouedns e)F I eqlJo suo4B^Jesqo'rlaruH -1nort 'slueruuoJr^ 'J"00€ o1 elqlepsrueJJo uorlruoJ peleler erll uI e^oqe Jo lB €OzuIAIpue zgu141ueel$eq Euruuog -ue '(596I ''le crleueEerp uregeo ur e1rsn1oild pue olruetueu qlr,r\ to ppzns6) uorsse-doJdsrql tn es€qdel?Ipeuuelul 8gsutr41 tO'uI/{ '}O'UIAI lsueoc {eu leql slseE:tns elrueEueu puu elrsnl uE se peqsqq4se IIe^\ sI o1 {1pug pue -orfd 'turleeq go eJnlxrru l€rnleu pel?equn ue ur elrsnlord.d qtytr tO.uIAI o1 seot zgu14 se pecnpeJsr u141 JoqunJ sgsu;lq '1nq-uorlrsueJl'Ouru u,norErelur euq1e1s,fucflroodgo uorle^Jesqo JnO uodn o1 HOOUW eql ur-pezrp ''O -rxo sr uf[ ur ernl€eJEurlserelur u? /hogs rru uI (g) + tO*lUIAI*luW: zO*ruIAIg lsJU l?ql Eu4eequodn elruetueuJo suorl?ulroJsu?Jleseqd eq1 uenu \ eq u€c tOtuI/rI o1 elrsnlord.d go uorlce 'elluetueur 'elrsnlorr{.d r^rtrrs^s tr -eJ eql ruor; lou drepuoces uror; ozH-o-utr 86su;41 'C.00€ Nr sNorrJYsu uo.{ sNoIrYJrTdI^II srurog eloqe sernleJedurel lu Jr? ur lerll 'ql1rus (896I pue rsssBIC lueq) turpug eql srrrrguoc elrsnl 'seueld (961) errsnlorfd eqt pue seueld -or,(d urog tOtuIAI go qpnor8 srql 'peseercur sr turleeq (00I) 'O'unt eql uee^\leq uorl€loJ ou sr eJeql 'pelcedxe '1eql 'saEuu; tO'uI se a1rsn1or,(d ruorg Eurrnort pue elrsnlor{d qlvrr u,rorErel se selecrpur6 uonenbg ec41e1(697) I -ur tOtuI/rI ',(leuopppv 'y Eul.rrne3s pez4p6frclle^\ ^\oqs rre ur turleeq eqt lelpred seEuuggrrol{ esoql ecurs

'141 eseqdgoelnecrput sl reqt *[IOI] Buo1z 3u11dpflee^\ oql ql1,raure1ed davs u? pue dtoloqdrour erodorc -rur tuquls e s^\oqsleql uorupeds elpe8ueru .zreH ,pIeJlI aql go ure:8 poleequn uV (c) '"9 'trg ur se ,sl\o!? qll^r pe{reru ore 'q W oseqdturlouep suortoeuerer{I Z roJ C.00€ le rre ur pel?oq ueurceds elru?tueu rouedng eIs-I eqlJo lunou urerS-peqsruce 'eldtues .s^\ous Iernleu B ruo{ urelted uo4cqgrp sno8o1eq7(O qllr pe{reu er? pu? Ilacqns elrsnlor,{d eqlJo *[1gy] nesuf yq eseqdelouep lerp suorpeger eql'urelled [6tO] eflsnlorfd eql uI .9 .ou ,o1d osoql ol Iecrluaprere suoncegel esuelw lsoru eq1 31 -tues elrsnlori(dpj?pu?ls eJo lunour urerS-peqsuce ur alcnred e tuorg u:e4ed oavs uy (e) 'ilecqns elrsnlor{d srql ol reJer suJeil?d (avs eseql uo peceld sox€aqJ 'ellsnlor,(dJo urelled cavs ([gyg] ro) [00I] eqt ol spuodseroc sureu?d oavs eql ur uoes [ecqns 'I .l?ql esuelu er{J I eseqdpeu:e1 e^eq ea\'ecuerue^uocJoJ eprxo uW u^\oulun uego qdertorcrru B prr? suroped oavs .9 'AIC

saolxo uwco A(InJs I ftIIlIH rycgsns qNv)svu zr8 RASKAND BUSECK:HRTEM STUDY OF Mn OXIDES 8r3 termediate phase (not MnrOr) and tentatively named it orderedentry of another element(except, perhaps, H) into groutellite. In their study of the manganite-to-pyrolusite the pyrolusite structure.Electron energyJossspectroscopy reaction, they found no such intermediate phase. How- and energy-dispersiveX-ray emission spectroscopyde- ever, their identifications relied solely on xno, which we tected no elements other than Mn and O in this phase. also found to provide no evidenceof intermediate MnrOr. Nor do they indicate large tunnel structures,common in Our sarn patterns of an unheated pyrolusite-manganite some Mn oxides (Turner and Buseck, 1981). Such large mixture provide evidencethat MnrO. is presentwith py- tunnel dimensions would be made evident by additional rolusite and manganite at low temperatures. reflectionsalong a*, not along [01]. The reaction of manganite to pyrolusite may be writ- We believe that the 7.2-A spacingof phaseM may be ten as a result of ordering along the [01] direction involving OH groups and Mn atoms of different oxidation states. 4Mn3+OOH* Or:4Mna+O, + 2H2O. (4) Analogous doubling along [01] marks the differencein Strunz (1943) stated that the most likely mechanism for Mn oxidation betweenmanganite and pyrolusite; tripling this reaction is migration of H out of manganite, rather along [01] in phaseM may also signal a different average than O influx into manganite. However, in the absence Mn oxidation state. of mechanistic evidence, we have chosen to write this oxidation reaction in the more conventional manner. Our CoNcr,usroNs finding of low-temperature MnrOr-pyrolusite inter- HRrEM study of natural mixtures of pyrolusite and growths in close proximity to manganite suggeststhat manganite before and after heating has resulted in the MnrO, may also be formed from manganite decomposi- following findings: (l) MnrO, has been identified through tion by a reaction such as electron diftaction and imaging in unheated natural pyrolusite-manganitemixtures. In such specimens,MnrO, 20Mn3+OOH* Oz:4Mnl+Mn!+O8 + l0H2O. (5) occurs as a poorly crystalline phase intergrown with pyro- Reactions 4 and 5 may operate simultaneously, the pro- lusite. This representsthe first reported identification of portion of pyrolusite to MnrO, depending on local vari- MnrO, in natural samples.X-ray confirmation ofthe iden- ations in oxygenfugacity. Crystallite sizeis a further factor tification is neededsince it is possiblethat MnrO, formed in controlling such reactions,as was shown by Giovanoli upon ion-bombardment during sample preparation or and Leuenberger(1969) for the oxidation of a-MnOOH upon exposure to the electron beam. (2) Aligned inter- (groutite). The micropores that develop parallel to man- growth of monoclinic MnrO, in secondarypyrolusite may ganite (010) as a result of these reactions may provide be responsible,along with oriented micropores, for the channelsfor the escapeof HrO. They may also facilitate nontetragonal characteristicsdisplayed by such pyrolu- migration of oxygen into the crystal and thus oxidation site. (3) Intergrowths of MnrO, and pyrolusite were found and elimination of MnrO, by the reverseof Reaction 3. in heatedand in unheatedsamples. The low-temperature occunenceof MnrO, with pyrolusite and manganite sug- Nrcw Mn oxrDE geststhat MnrO, may be formed togetherwith pyrolusite In the course ofthis study, we have discovered a pre- during the diageneticdecomposition of manganite.Well- viously unreported Mn oxide phase henceforth termed crystallized MnrOr, seenin heated samples,corresponds "phase M." This phaseis identified by diffraction patterns to a previously documented occurrenceof MnrO. as a that have strongreflections exactly corresponding to those compositionally intermediate phase between MnO, and in the [010] pattern ofpyrolusite and weak reflectionsthat MnrO, above 300'C. (4) A new phasehas been found that trisectthe (l0l) and (l0I) spacingsofthe dominant pyro- has a diftaction pattern related to the pyrolusite [010] lusite pattern (Figs. 6a, 6b). The weak reflections denote pattern, but that contains reflections corresponding to a periodicity of 7.2 A in the pyrolusite [01] direction. spacingsthree times that of the pyrolusite (l0l) planes. Pyrolusite, manganite, and MnrO, do not have spacings Ordering of OH groups and Mn atoms with diferent ox- close to 7.2 A, and we have found no other Mn oxide idation statesacross the (l0l) planesof the structure may structure that appears appropriate. explain these reflections. If so, this phase has a stoichi- Phase M has been found in crushed grain mounts of ometry intermediate between MnOOH and MnOr. the heated Lake Superior sample; of the unheated Ilfeld, pyrolusite Harz, sample; and of standard sarnple IC no. AcxNowlnncMENTs 6 (Kozawa, l98l). In unheatedIlfeld manganite, the unknown phase occurs in a grain that shows a pronounced Theauthors would like to thankG. E.Brown, Jr., for providing micropore morphology (Fig. 6c). Assuming that thesemi- uswith samples.We also thank Barbara Miner, Reinhard Neder, Gerard Spinnler,Theresa Keegan, Wheatley,and Shirley cropores parallel (010) of a parent manganite, the 4.7-A John Turnerfor theirassistance in variousaspects ofthis study.Help- spacing of the lattice fringes paralleling the micropores ful reviewsof this paperwere provided by R. G. Burns,Rudolf presumably representsthe lattice translation of phase M Giovanoli,and R. J. Reeder.Funding was providedby NSF that formed from the 5.28-A b fanslation of manganite. GrantsOCE-82001l8 and OCE-8401107. Electron microscopy Two obvious explanationsfor theseextra reflectionscan wasperformed at the HRrEMRegional Facility in the Centerfor be immediately discounted. They are not a result of an SolidState Science at ArizonaState University. 814 RASKAND BUSECK:HRTEM STUDY OF MN OXIDES

RBrnnrNcns convertibility of phasesin the system Mn-O-OH. American Mineralogist, 44, 8 19-838. Baur, W.H. (1976) Rutile-type compounds. V. Refinementsof -(1960) Solid-solid and solid-vapor reactions and a new MnO, and MgFr. Acta Crystallographica,B.32, 2200-2204. phasein the systemMn-O. American Ceramic SocietyJournal, Buerger,M.J. (1936)The symmetry and crystal structureof man- 43.620-626. ganite, Mn(OH)O. Zeitschrift fiir Kristallographie, 95, 163- Kozawa, Akiya. (1981) Data, useful information, articles,etc. on 174. MnOr collectedby the IC MnO, ofrce. In B. Schumm, H.M. Burns, R.G., and Burns, V.M. (1979) Manganeseoxides. In R.G. Joseph,and A. Kozawa, Eds. Manganesedioxide symposium, Burns, Ed. Marine minerals. Mineralogical Societyof America Tokyo, 1980,2,559-599. The ElectrochemicalSociety, Cleve- Reviews in Mineralogy, 6, I-46. land, Ohio. Champness,P.E. (1971) The transformation manganite - pyro- Oswald, H.R., and Wampetich, M.J. (1967) Die Kristallstruk- Iusite. Mineralogical Magazine, 38, 245-248. turen von MnrO, und CdrMnrOr. Helvetica Chimia Acta, 50, Dasgupta, D.R. (1965) Oriented transformation of manganite 2023-2034. during heat treatment. Mineralogical Magazine,35, 131-139. Oswald, H.R., Feitknecht, Walter, and Wampetich, M.J. (1965) Davis, R.J. (1967) Somemanganese oxide pseudomorphs.Min- Crystal data of MnrO, and CdrMnrOs. Nature, 207, 72. eralogicalMagazine, 36, 274-27 9. Pashley,D.W., Menter, J.W., and Bassett,G.A. (1957) Obser- Dent Glasser, L.S., and Smith, LB. (1968) Oriented transfor- vation of dislocations in metals by means of Moir6 patterns mations in the systemMnO-O-HrO. Mineralogical Magazine, on electron micrographs.Nature, l7 9, 752-7 55. 36,976-987. Potter, R.M., and Rossman, G.R. (1979) The tetravalent man- de Wolff, P.M. (1959) Interpretation of some7-MnO, diffraction ganeseoxides: Identification, hydration, and structural rela- patterns.Acta Crystallographica,12, 34 I-345. tionshipsby infrared spectroscopy.American Mineralogist, 64, Feitknecht, Walter. (1964) Einfluss der Teilchengrosseauf den I 199-1218. Mechanismus von Festkiirperreaktionen. Journal of Pure and Strunz, Hugo. (1943) Beitrag zum Pyrolusitproblem. Naturwis- Applied Chemistry, 9, 423-440. senschaften.Wochenschrift liir die Fortschritte, der Naturwis- Giovanoli, Rudolf, and Leuenberger,U. (1969) Uber die Oxy- senschaft,der Medizin und der Technik., 31, 89-91. dation von Manganoxidhydroxid. Helvetica Chimia Acta, 52, Turner, Shirley, and Buseck, P.R. (1981) Todorokites: A new 2333-2347. family ofnaturally occurring manganeseoxides. Science,212, Giovanoli, Rudolf, Mauer, R., and Feitknecht, Walter. (1967) 1024-1027. Zur Struklur des .y-MnOr. Helvetica Chimia Acta,5Q, lO72- 1080. MlNuscnlpr REcETvEDJuNe I, 1985 Klingsberg,Cyrus, and Roy, Rustum (1959) Stability and inter- MeNuscnrpr AccEprEDJllulnv 17, 1986