American M ineralogist, Volume 62, pages 278-285, I977

Characterizationof a marinebirnessite

E. D. Gt-ovrn Departmentof Geologyand Geophysics Uniuersityof Wisconsin,Madison, Wisconsin

Abstract

Birnessite,more crystallinethan has beenheretofore described, has beenfound in micro- nodulesfrom a depth of 2-4 metersbelow the top of a Caribbeancore in foraminiferal sediments.The associatedsediments have been carbon-dated at 33,640t 5340years. The mineral has been studied by scanningelectron microscopy, X-ray diffraction,and electronmicroprobe analysis. The surfacestructure is madeup of curved,intersecting plates between500-1000 A thick.The X-ray data agreereasonably well with the dataon synthetic birnessitecompounds. The chemicalformula, by analogywith chalcophanite,is (Na, Ca, K)o.rrMg,onMnu ru(Fe, Co, Ni, Cu)o,.O,, u(HrO)r.. The assignment of thisformula with Mn in the * 4 valencestate leansheavily on the findingsfor the syntheticcompounds since no analysiswas made for the Mn valencestate(s) or the amountof waterpresent.

Introduction and general description material. The clusters at this level contain mainly Birnessiteand todorokite are two of the most im- birnessiteand todorokite, as indicated by X-ray dif- portant manganeseminerals comprising ferromanga- fraction, both minerals occurring in the samecluster. nesenodules. Their content of first transition-group At the 300cm level and below,the clustersare almost metals in certain regions of the sea floor, often at exclusively birnessite. They are a few tenths of a economicallyimportant levels,has led to mushroom- millimeter or lessin diameter. To date a number of ing researchon nodule genesis.Because the manga- birnessiteclusters have been collected and studied, neseoxide phasesare typically intermixed on a fine but no clusters are exclusivelytodorokite. Photo- scale with iron oxides (Burns and Burns, 1976), as micrography, study of the X-ray diffraction effects, well as with other authigenic and detrital minerals, and chemical analysisof the birnessiteclusters have they are difficult to characterize. been carried out in order to characterizethis marine I have found an occurrenceof thesetwo minerals birnessiteas to structureand composition. where one of them (birnessite)occurs with a more electron rnicroscope ordered crystal structure than has hitherto been re- Observations with the scanning (SEM) ported for the mineral. The material is in a core from the Gulf of Mexico just south of the northwesttip of Birnessitewas originally describedby Jones and Cuba in 2244 meters of water in hemipelagic, fora- Milne (1956) as composedof black grains. More miniferal sediments(Atlantis cruiseA-185-23, May recently scanning electron microscopeexamination 15,1953; Lat.2l"32'N, Long.85o4.5'W,Ewinget al., has shown a profusion ofsurface structureson a scale 1958).The zoneof occurrenceof manganeseminerals of a few microns or less:plates, disks, box-work of is from 200-400 cm below the top of the core. The intersecting plates or disks, box-work of curved sedimentshave been carbon-dated at the 240-245 cm plates with crenulate edges (Woo, 1973); sponge- level,giving an apparent age of 33,640t 5,340years work, aggregatesor botryoidal clustersof spherulites (H. F. Nelson, personal communication). (Fewkes, 1973;Sorem and Fewkes, 1976);and clus- The manganeseoxides are found in microclusters, ters with a cellular structureof bent plates(Brown e/ usually on or in foraminifer tests.The most abundant al., l97l). occurrenceis in a zone atthe 240 cm level, where they The present material shows many of these same may be noticed with the naked eye as a scattered, features,but the most characteristicsurface structure black discolorationof the calcium carbonateskeletal is that of intergrown curved platesor surfaceswhich, 278 GLOVER: CHARACTERIZATION OF A MARINE BIRNESSITE seenend-on, appear vermiform. Figure I showsSEM Fe radiation. Becauseof their roughly spherical photos of birnessiteclusters Nos. I and 2 at two growth,the clustersgave excellent powder patterns, different magnifications.In the original photographs, even without rotation. The birnessiteclusters also at the highestmagnification (30,000 X ), one can seea containsmall amountsof calcite(foram shellpar- banding or lineation on the termination of the indi- ticles,coccoliths, possible secondary calcite), quarlz, vidual curved plates of cluster No. l. This seemsto and euhedralzeolite crystals, probably phillipsite. indicate substructure smaller than the 600-1200A The accessorymineral lines were usually easily distin- thicknessof the platelets.However, the revelationof guishedfrom the birnessitelines on the basisof their this apparent substructuremay be the result of evac- spottycharacter or greatersharpness. uation in the SEM or the heat of the carbon arc in In Table I X-ray diffractionpatterns of clustersI vacuo while carbon coating. and 2 are comparedwith publisheddata. The refer- encepatterns are from (synthetic)manganese (III) X-ray diffraction results manganate(IV) (Giovanoliet al., 1970b)(Cols. 5, 6, To observe the X-ray diffraction effects,the cluster and 7), which is presentlyaccepted as beingequiva- to be examined was glued to a freshly-madeglass lent to birnessiteby the PowderDiffraction File edi- fiber, placed in a 57.3 mm or I14.6 mm diameter tors (for a completecritical discussion of X-ray data powder camera, and rotated in unfiltered or filtered seeBurns and Burns, 1976).The reflectionsfound

Ftc. l. SEM picturesof birnessiteclusters No. I and 2 Surfaceof No. 1, (a) and (b); surfaceof No. 2, (c) and (d). Length of bar in microns:(a) 2, (b) 0.5, (c) 50, (d) L 280 GLOVER: CHARACTERIZATIONOF A MARINE BIRNESSITE

Tnsle l. X-ray diffraction data from birnessite

Manganese Oxide HydraEe, Giovanoli et al. - Powder Diffraction File Birnessite, This Study Card 23-1239

Cluster No. 1* Cluster No, 2** (r) Q) (3) (4) (s) (6) (7) d-spacing I/Io+-l'fMeas. d-spacing Calc, d-spacing d-spacing I/ Io hkl

7.08 r00 7.06 7.ll 7.2L 100 001 3.547 28 3.556 (3.5s6) i 3.61 80 002 2.468 L7 2.464 2.467 2.46 r00 100 2.333 43 2.330 2.330 2.33 100 l0l 2.03r 24 2.O27 2.OZl 2.04 80 r02 r.7l5 I L.775 r.1 18 1. 802 10 004 r. 7ll 29 l. 710 L.709 I.723 80 103 I r.441 L.442 L.454 20 LO4 L.426 tl L.424 (r.424)tt 7.422 60 110 1.398 l0 L.397 L.396 1.391 40 111 4 r.322 L.322 T.32L 20 l_L2 1.234 6 L.231 L.233 L.229 20 200 L.219 6 1.215 l. 215 L.2I2 20 zor 1.185 4 1.182 l.185 006 L. L67 4 l.163_ 1.1651 L.L64 10 202 l l.105 5 r. 093, r. 0940 L.094 10 203 o 5 1.014^ 1.0133 1,015 l0 204 I 0 .929 l0 2L0 o. 905 l0 008 0 .867 IO 2L3 0 .819 10 300

* Unfiltered Fe radiation, 114.6 nm camera, not corrected for shrinkage' Three K$ lines from birnessite plus 3 (Ko) lines from calcite omitted' ** Filtered Fe radiation, 57.3 nm camera, corrected for shrinkage' Barely perceptible main calcite reflection omiEted. f Used to estimate c, ti Used to estimate a. -fft Estimates of intensitles of lines with I/I- : 10 were made by visual comparison of lines with a timed exposure series oE ilots under a low-power microscope. These values were multiplied by estimates of line widths nade by direct measure- ment to give the reported intensity values. The remaining weak lines were estimated visually. Thus, the measurement of the strong lines gives numbers approximating lntegrated intensities, agreewith those of Giovanoli et al., with the limita- value of c in the subject specimen.The intensities, tion that Fe radiation was used (not molybdenum), however,differ markedly. This is not surprising,since except that the (006) reflection was present.'It was the composition of the natural mineral differs from quite easily distinguishedfrom the nearby (202) re- the synthetic reference compound (see below). In flection and measured,even with a film made in the films made with Fe radiation, wherethere is sufficient small camera.In fact, the relativelygood line resolu- resolution of the diffraction pattern,one can seewith tion in patternsfrom the natural materialwould seem the naked eye that (/rft0) and (00/) reflections are to make the present specimen the best crystallized relatively sharp, while (ftkl) reflections are noticeably example of the birnessitestructure yet found, syn- broadened. The internretation here is that there is thetic or natural. reasonablygood order within the layers of the pro- The d values agreefairly well with the Giovanoli er posed structure (Giovanoli et al., 1970b)and a rela- al. data, the most obvious difference being a lower tively uniform stacking sequence of -7A layers. GLOV ER' CH A RACTERIZATIO N OF A MARINE BIRNESSITE 281

There is, however,much lessorder in the registration Discussion of the adjacentlayers, and the situation is quite com- parable, as Giovanoli et al. (1970a,b)have noted, to Formula of birnessite the layer structureof some of the 7A clay minerals. If the data are cast into the formula corresponding to the structure proposedby Giovanoli et al .,19'70b Chemical analysis (assuming 7 Mn atoms) then the oxygen content is Qualitative analysis was carried out in both the high (20.8 calculated by differenceus. l8 for the refer- SEM and the electronprobe. Presentin the birnessite enceformula). If one assumes7 (Mg * Mn) then the are: Mn (main constituent);Na, Mg, K, and Ca, each averageformula is: in sizeableamounts (0-7% level); Fe, Co, Cu, and Ni, Nao.nuCao , Ko ru Mg, o, Mnu.ru(Feo ou Coo.o, Nio.o, each ( I percent,with usually Ni ) Fe ) Cu ) Co. Cuoor) Ora..(H2O)B 8. Thus the presentmaterial showsthe characteristicof 'scavenging' manganesenodule mineralsof transition Written more compactly the formula may be com- metals. Other elements revealed by their X-ray pared to the two previously proposed formulae for spectra are Cl, S, and Si. They occur in small and birnessite(Jones and Milne, 1956;Brown et al. 1976), variableamounts and are thought to be adventitious. the two proposed forms of Giovanoli et al ., and the A complete quantitative analysis of these small proposed(Giovanoli et al. l970a,b) relatedstructure, specimensis limited by the fact that neither the per- chalcophanite(Wadsley, 1955): centageof Mn reducedbelow valence4, the oxygen This study content, nor the water content can very easily be (Na, Ca, K)o,, Mg, o. Mflu ru(Fe, Co, Ni, Cu)o,u O* u determined. Also, the vacuum used in microbeam (H,O)r.' analysismay causeloss of some of the water of crys- Jones and Milne tallization before the analysis can be made. Never- (Nao' Caor) MnrO,a (HrO), t theless,valuable information may be obtained,espe- Brown et al. cially using the low beam current of a SEM equipped (NaounCaou,) (Mn31.u, Mnilrn) O1n.1(H2O)2e with the highly sensitive EDS system (energy dis- Manganese oxide hydrate, Powder Dffiaction File persivespectrometer). In this way a completeanalysis Card 23-1239 may usually be carried out for elementsabove fluor- (Mnl*, Mnl*) O,,o (H,O)n.,u ine, where concentrationsare above a few tenthsof a Sodium mqnganeseoxide hydrate, Powder Dffiaction percent, with beams of very small radius with high File Card 23-1046 spatial resolution. Point by point concentrationsof Na, (Mn!+.' Mn?10Mnf*) O1s.u(H2O)4.u all major and minor elementsabove fluorine are ob- Chalcophanite tained. This has been done on the two clusters of Zn, (Mnt+) O,n (H,O)6. birnessiteseen in Figure I which were usedpreviously for X-ray diffraction, the transition elementsbeing This comparisonshows the apparentoccurrence of analyzedby the wavelengthdispersive system (WDS) Mg in place of some of the lower-valenceMn of the of the probe. An example of the results' for one maganeseoxide hydrate. It shows a Na * K * Ca samplepoint of clusterNo. 2 (point No. l) are given content of 0.83 atoms, which is intermediatebetween in Table 2. the two referencecompounds of Giovanoli et al. lt shows a total of about 0.16 atoms of transition met- als,which are not presentin the syntheticcompounds ' For complete data and a discussionof the analysissee Appendix but are presentin the related marine mineral; and it

Tnslr 2 Analysis at one samplepoint of clusterNo. 2

Nar0 MCO Si02 Cl K-0 CaO MnO- FeO CoO NiO CuO Total zt

1.9* 6.2 0.9 0.1 0.2 1.8 0.39 15.8 0.55 0.14 0.80 0.33 88.9

* A11 numbers in weighE percent. GLOVER,'CHARACTERIZATION OF A MARINE BIRNESSITE indicatesa variablewater content,which may be With syntheticbirnessites Feitknecht and Marti, expectedwhen analyzing hydrates in a vacuum.(All and Buseret al. (reviewedin Burnsand Burns, 1976) probework is consistentwith this loss.Whereas the found that the higherthe oxidationstate of Mn the ratio of other elementsto Mn is fairly consistent- poorer wasthe stackingof the layers,giving a diffuse someelements, such as K and Mg, more than oth- or absent7A spacing.Also, structuralarguments ers-the ratio of Mn to oxygen[by difference]varies (Burnsand Burns,1976) that Mn2+substituting for considerablywith the analyticalconditions during Mna+ in the octahedralMn-O sheetgives necessary analysisof the sample:beam voltage, current, and stabilitywould seemto associatethe higheroxidation areaof specimensampled by the beam).Thus this statewith poor crystallinity.For thesereasons and compositionagrees well with the Giovanol\et ql. becausemost chalcophanitesthemselves contain compoundsbut hasa verysimilar formula overall to someMn2+, Mn is probablypresent in this relatively thoseof Jonesand Milne and of Brown et al.; the well-crystallizedspecimen partly as Mn2+,in either striking differenceis the high Mg and K content, the Mn-O octahedralsheet, the alternatingsheet, or whichis consistentwith the marineorigin andpossi- both. bly may be characteristicof it. The fitting of the data to the aboveformula could The ideal structureof chalcophanite(Wadsley, not havebeen accomplished with anyassurance with- 1955),to which Giovanoli et al. (1970a,b)believe out the basic data of Giovanoli et al., becausea birnessiteis related,is a layerstructure composed of a completechemical analysis could not be done.If one Mno+-O octahedralsheet alternating with a sheet acceptstheir estimateof the probablestructure, then composedof Zn atomscoordinated to oxygenand the comparisonof the presentsample to theirmanga- water.In mostexamples of chalcophanitepart of the neseoxide hydrate is recapitulatedas follows: The Mn is presentas Mn2+ (Wadsley,1955). The pro- present sample gives an X-ray diffraction pattern posedformula (basis, Mn t Mg :7) of the present agreeingreasonably well with the patternof the hex- sampleis equivalentto scalingthe numbersof atoms agonalsynthetic compound within the limitationof presenton the basisof 6 Mna+(see analysis results, the rangeof d spacingscovered by the Fe radiation Tables3, 4 and 5 in the Appendix).This could be used,with one additionalbasal order (006) visible in written:(Na, Ca, K, Fe, Co, Ni, Cu)'.r, Mgr.onMnu the patternof the presentsample. If one caststhe O.',n(HrO) whereMg andthe sumof Na, Ca, K, Fe, chemicalanalysis into their form, oneobtains a for- Co, Cu, and Ni formallycorrespond to the two Zn mula intermediatebetween their two compounds, atomsof chalcophanite. with other ions (K, Ca, Mg) replacingsome of the

Tesle 3. Number of atoms in formula for cluster No. l+

IKNaCa f,K NaCa INaCa Si lig Mn IMgMn +T.M,** c1 H20

0.49 0.25 0. 06 0. 80 0.13 0. 04 l. 01 6.00 7.01 0. 96 0.l8 13.5 1.5 '0.1I 0.29 0,24 0.15 0. 68 0.05 r.00 6.00 7.00 0. 84 o.27 13.5 1.0 0.58 0.26 0. 06 0.90 0.75 0.02 1, 05 6. 00 7.05 1.06 0.1r 13,6 1 r 0.47 0.24 0. 1r 0.82 0.69 0.00 1.01 6.00 7.OL 0. 98 0.11 13.5 r. r 0,39 0.25 0.15 0,79 0.68 0.03 1.07 6.00 1.Ol 0. 95 0.14 13,6 3.5 0.40 0.26 0. 11 0.77 0.63 0.05 l. 16 6.00 7.t6 0. 93 0. 13 13.6 3.3

0.45 0.27 0.26 0.9B 0.87 0.15 1.10 6.00 1.ro L.14 0.17 t3.7 3.7 0.46 0.27 0.23 0.96 0.83 0.05 1.06 6.00 1.06 L.L2 0. 14 r3,7 L.7 0.2L 0.24 0.36 0. 81 0.14 0.15 r.05 6.00 t.o5 0.91 o.L7 r3.7 5.8 0.23 0.25 0,23 0.7L 0,60 0.15 r,o2 6.00 1.02 0.87 0.14 13.5 4,4 0.28 0.21 0.10 0.65 0.51 0.07 0.99 6.00 6.99 0. 81 0. 13 L3.4 3.0

* Each line in body of table represents analysis at one sample point. Analysis of transition elements by WDSsystem of electron nlcroprobe, all other analyzed elements by EDS system of SEM, O and H"O by differ- ence . ** Transitlon MeLals: average of 3 points on cluster No. 1. Fe 0.050(3), Co 0.0L0(J.), Ni 0.070(4), Cu 0.030(3) I Fe, Co, Ni, Cu = T.M. = 1.60(6) GLOVER: CHARACTERIZATION OF A MARINE BIRNESSITE 283

Tnslr 4. Number of atomsin formula for clusterNo. 2*

IKNaCa IKNaCa INaCa Si i4n IMglin +T.M.x* Cl HzO 0+H2O

0.43 0.27 0.05 0.75 0.48 0.1 1.06 6. 00 7.06 0.91 0.02 0.03 13.6 4.2 17.7 0.64 0.24 0.07 0.95 0.7l c.02 0.99 6.00 6.99 1.11 0.02 0.06 L3.7 3.1 17.4 0.45 0,24 0,06 0,75 0.5r 0.3 1,06 6.00 7.06 0.91 0.0 0.0 L3,4 5.8 r9.2 0.55 0.23 0.06 0.84 0.61 0.2 r.10 6.00 7.10 r.00 0.0 0.02 r3.3 5.6 18.9 0.57 0.26 0.07 0.90 0.64 0.07 r.00 6.00 7.OO r.06 0.0 0.04 13.5 6.4 18.9

* Each line in body of table represenEs analysis at one sample point, Analysis of transition elements by WDS system of electron microprobe, all other analyzed elements by EDS system of SEM, O and HrO by difference. ** Transition lletals: average of 3 points on cluster No. I Fe 0.050(3),co 0.010(1),Ni 0.070(4),Cu 0.030(3) I Fe, Co, Ni, Cu= T,t{.=0.160(6)

sodium and manganeseof the postulatedinterlayer, a formula agreeing well, overall, with previous pro- (l) rhe birnessireJ#il:L"re differsfrom pre- posalsfor birnessiteand with chalcophanite. viouslydescribed materials in severalrespects. (a) Its contentof Na, Mg, K, and Ca is differentfrom that Signifcance o"f MS and K of anypreviously documented specimens, but its bulk formula is analogouswith previouslydescribed bir- The presenceof high Mg and moderate K concen- nessitesand resemblesclosely the chalcophanitefor- trations, showinglittle variability with respectto Mn, mula on which its structuremay be based.(b) It is strongly suggestsconsideration of this compositional betterordered, structurally, than any previouslyde- relationship as a possible characteristicof marine scribedbirnessite mineral. (c) It is the first natural birnessite.In fact, this observationagrees well with birnessitein which the X-ray diffraction pattern the correlation made by Burns and Furstenau( 1966), agreeswith the presentlyaccepted diffraction data for using X-ray maps of K, Mg, and Mn in a marine 7A the syntheticanalog in all importantdetails, and with Mn-O mineral phase, presumably a birnessite-like a proposedformulae bracketed by the synthetic(bir- form. Other analysesof ferromanganesenodule ma- nessite)and Na (birnessite)forms. terial have not shown this correlation,either because (2) The findingand characterization of thismineral (1) Mg and K were not measured,or (2) the Mg or K lends supportto the proposalthat the hexagonal level was more or lessuniform and/or not strongly structureof manganeseoxide hydrate of Giovanoliet associatedwith any one phase, or (3) no identi- a/. (PowderDiffraction File Card 23-1239)is a valid fication was made of the mineral speciesbeing ana- referencefor the naturallyoccurring birnessite miner- lyzed. als. In the study of Burns and Furstenau(1966) Mn-O (3) The K and Mg contentsuggests a possibly phases '6 thought to be a more disordered Mn Or' important marineassociation and demandsconsid- (Burns species and Burns, 1976)did not show the K, erationof the role of K and Ms in ferromansanese Mg, Mn correlation. nodulegenesis. Since it has not been possible as yet to separatethe supposedtodorokite phase(9.6.{) in the presentma- terial, it is not known with any certainty whether or Appendix not this todorokite contains K and Mg. Other analy- ses have not revealeda Mg, K, Mn associationfor Chemicalanalysis marine todorokites. After SEM photographyand X-ray diffractionthe It is hoped that techniquessimilar to those de- approximately0.1 mm diametergrains were flattened scribed here, plus electron diffraction studies applied on the glyptal-smearedsurface of a l" epoxyplug to the present material and other nodule material, and vacuum-coated-with carbon. Sincethe major may clarify further the Mg, K, Mn association. contaminantwas CaCO, (from the admixedcoccol- GLOVER CHARACTERIZATIONOF A MARINE BIRNESSITE

Tnslr 5. Final hypothesizedformulae for eachsample point of clustersNos. I and 2: Basis,Mna+ = 6 (Si, Cl and equivalentNa. and S deleted)

N1 H20

Cluster No. 1 0.49 0.06 0.25 1.01 6.00 0.05 0.01 0.07 0.03 13.5 1.5 o.29 o,24 1.00 6.00 0.05 0.01 0.07 0.03 13.5 1.0 0.58 0.06 o.26 1.05 5.00 0.05 0.01 0.07 0.03 13.6 1.1 0.11 o.24 1.01 6.00 0.05 0.01 0.07 0.03 13.5 1.1 0. 39 0.15 v.z) L.07 6.00 0.05 0.01 0.07 0.03 13.6 3.5 0.40 0.L2 o.26 1.16 6.00 0.05 0.01 0.07 0.03 13.5 3.3

0.45 o.26 0.27 1.L0 6.00 0.05 0.01 0.07 0.03 L3.7 3.7 0.46 o.23 0.27 1.06 6.00 0.05 0.01 0.07 0.03 L3.7 1.7 U.JI o.24 1.05 6.00 0.05 0.01 0.07 0.03 73.7 5.8 0.23 o.23 n tq r.o2 6.00 0.05 0.01 0.07 0.03 13.5 4.4 0.28 0.L0 0.99 6.00 0.05 0.01 0.07 0.03 t3.4 3.0

Average 0.40(13) * o. 170.0) 0.2s(1) r.os(5) 6.00 0.os 0.0r. 0.07 o.03 13.6(1) 2.7(16)

CLuster No. 2 0.43 0.05 n ,1 1.05 6.00 0.05 0.01 0.07 0.03 l-3.6 4.2 0.64 0.07 0.24 0.99 6.00 0.05 0.0r- 0.07 0.03 73.7 3.7 0.45 0. 06 o.24 1.06 6.00 0.05 0.01 0.07 0.03 13.5 5.8 0.55 0. 06 6 Aa 1.10 6.00 0.05 0.01 0.07 0.03 13.6 5.7 0.57 0.07 0.26 1.00 6.00 0.0s 0.01 0.07 0.03 13.5 s.4

0.s3(7) 0.05(1) 0.25(1.) 1.04(5) 6.00 0.0s o.01 o.07 0.03 13.5(1) s.0(10)

Average of 2 Clusters 0.45 (8) 0.12(s) 0.251D r.-a4(t 6,_AA. 0.!: A._A! 0.-AZ q.93 13.-61_1) 3.8(10)

*Parenthetical figures represent the estinated standard deviatlon (esd), in terms of least unlts cited for the value to their intoedlate left, thus 0.40(13) indicates an esd of 0.13.

iths and, often, intergrown foraminifer chamber in Tables 3, 4, or 5. Typical analysisconditions for walls), it was necessaryto selectEDS analysisregions the SEM and the probe were as follows: where the Ca X-ray level was at a minimum. Points SEM: l5kV, 0.5-1.0 X lO-'gamperes, absorbed acceptedfor analysishad a Ca peak no greaterthan current, 50'-300" X-ray collection time. that of K, and low or absent Si and S Peaks.The Standards: NaCl, KH,PO4, CaCOr, MgO, SiOr, deliberatechoice of samplepoints, then, givesresults Mn metal,CuS. that should be qualified in that Si and S may not be Electron Probe:25kV,6 X 10-8absorbed current. negligible,and Ca valuesmay be somewhatdifferent Standards:Fe, Co, Ni, Cu metals;average of sam- than those shown (Tables 3,4, and 5). Sincethe Ca pled areas,-20 micron diameter circle. values found in cluster No. I were variable and in some casescomparable to clusterNo.2, a vgry low or Correction of data zero value may be closeto the true composition. The actual routine for correction of the X-ray in- That part of the analysisdone in the SEM was for tensitiesto concentrationswas as follows.Data taken all the elementsexcept the transition group metals at l5kV, using the EDS system,was multiple least- Fe, Co, Ni, and Cu. Becauseof the difficulties of squaresfitted by standard spectra(Schamber, 1973), analyzing for low concentrations in the EDS system, then corrected using the theoretical correction pro- the analysesof the transition metalswere made in the gram ('ZAF' correctionsalso by Schamber,1974),in electronprobe with its WDS system,where all the Ka the PDP 1ll05 computer which is part of the X-ray lines could be resolved and analyzed with greater analysis system. Schamber'sprogram is similar to sensitivity. Three points in cluster No. I were ana- other widely used theoretical correction programs lyzed, with resultsand standard deviationsas shown such as J. W. Colby's "Magic 4" (1971). From the CLOVER: CHARACTERIZATION OF A MARINE BIRNESSITE 285

'output of concentrationsthe numbers for Na, Mg, Si, Geological Observatory, Palisades,New York, have kindly given Cl, S, Ca, K, and Mn wereaccepted as final as far as their permission to use the ,4-185 cores in research on cathodo- correctionof datais concerned.Then the probedata luminescence of carbonates, of which this study is a part. H. F. and Development Corporation, pres- (K ratios)for Nelson of Mobil Research Fe,Co, Ni, and Cu wereentered along ent on the A- I 85 cruise, made available the cuts of the cores, which with the knownconcentrations for Na, Mg, Si, Cl, S, are stored at Mobil's Field Research Laboratory, Dallas, Texas, K, Ca and Mn in the sameprogram for correctionof and generouslysupplied the C-14 date on core A-185-23. Fe,Co, Ni andCu; correctionswhich are minimal in this case.The two correction runs were necessary References since the takeoff anglesare differentin the two in- Brown,F. H., A. Pabstand D. L. Sawyer(1971) Birnessite on struments.Completg analysis data for l l pointson colemaniteat Boron.California. Am. Mineral.,56, 1057-1064. V. Burns(1976) Mineralogy of ferro- I points given Burns,R. C. and M. cluster and 5 on cluster2 are in Tables manganesenodules. In G. P. Glasby,Ed., Marine Manganese 3 and4. The finalassessment of numbersof atomsin Deposits.Elsevier Publishing Company, Amsterdam, in press. the formula, basedon 6 Mna+ atoms,is given in - and D. W. Furstenau(1966) Electron probe determination Table5 for eachsample point. In this tabulationSi, of inter-elementrelationships in manganesenodules. Am. Min' S, and Cl are deleted,since they are thought to be eral..54. 895-902. Colby, J. W. (1971)MAGIC IV, A new improvedversion of adventitious. MAGIC. Proc. 6th Natl Conf. Electron ProbeAnalysis, Pitts- burgh, Pennsylvania,July 27-30,p. 17A-178. Methodsof calculation Ewing,M., D. B. Ericsonand B. C. Heezen(1958) Sediments and The methodof obtainingnumbers of atomsof O topographyof the Gulf of Mexico.In L. G. Weeks,Ed., Habitat and HrO wasas follows:a first estimateof the O plus of Oil. AAPG, Tulsa,Oklahoma. (1973) marine (or Fewkes,R. H. External and internal featuresof HrO OH) contentwas obtainedby subtracting manganesenodules as seen with the SEM and their implications the total of analyzedelements from 100percent. This in noduleorigin. ln M. Morgenstein,Ed.,Papers on the Origin number was adjustedappropriately when SiOr, and and Distribution of ManganeseNodules in the Pacifc and Pros- Na equivalentto Cl and SOnwere subtractedfrom peclsfor Exploration.Honolulu, Hawaii, p.2l-29. the analysis.This adjustednumber was converted to Giovanoli, R., E. Stahli and W. Feitknecht(1970a) Uber Oxidhydroxidedes vierwertigen Mangans mit Schichtengitter.l. relativenumber of O atomsby dividingby 16,and Natrium-mangan(II, III) manganat(lY). Helu Chim.Acta, 53, then multiplied by the factor necessaryto give 6 209-220. atomsof Mn (or 7 atomsof Mg + Mn). Thisgave the -, E. Stzihliand W. Feitknecht(1970b) Uber Oxidhydroxide total O atomsin the formula,assuming the difference des vierwertigenMangans mit Schichtengilter.2. Mangan of the sum of the analyzedelements from 100percent (III)-manganat(lV). Helu Chim Acta, 53,453-4U. Jones,L. H. P. andA. A. Milne (1956)Birnessite, a new manga- is oxygen.When the numbersof oxygenatoms asso- nese oxide mineral from Aberdeenshire,Scotland. Mineral. ciatedwith the analyzedelements were added up (Mn Mag.,3l, 283-288. asMnOr), and subtractedfrom the total of O atoms, Schamber,F. H. (1973)A new techniquefor deconvolutionof the remaindergave the numberof watermolecules on complex X-ray spectra. Proc. 8th Natl. Conf. Electron Probe an oxygenbasis. The final figurefor numberofwater Analysis,New Orleans,La., Aug. l3-17,p.85A-85C. - (1974)ZAF correctionprocedure, Tracor Northern Techn. moleculeswas obtained,then, by multiplying by Lit., TracorNorthern, lnc., Middleton,Wisconsin. 16/18.A somewhatmore accuratevalue might be Sorem,R. K. and R. H. Fewkes(1976) Internal characteristics of obtainedby iterationof thiscalculation, but this was marinemanganese nodules. In G. P. Glasby,Ed., MarineMan- not done, sincethe high variabilityof this number ganeseDeposits. Elsevier Publishing Company, Amsterdam, in overthe samplepoints indicates,a poor accuracyfor press. (1955) chalcophanite, (analysis Wadsley,A. D. The crystal structureof reasonsalready suggested in vacuum). ZnMn,O,.3HzO.Acta Crystallogr.,8, 165-172. Woo,C. C. (1973)Scanning electron micrographs of marineman- ganesemicronodules, marine pebble-sizednodules, and fresh Acknowledgements water manganesenodules. In M. Morgenstein,Ed., Paperson I wishto thank S. W. Bailey,University of Wisconsin,Madison, the Origin and Distribulion of ManganeseNodules in the Pacific Wisconsin,for hisadvice and criticism of the X-ray dataand R. G. and Prospectsfor Exploralion Honolulu, Hawaii, p. 165-171. Burns,Massachusetts Institute of Technology,Cambridge, Mas- sachusetts,for criticalcomment on an early versionof the manu- Manuscript receiued,October 28, 1975;accepled script. J. D. Hays and (the late) M. Ewing, Lamont-Doherty for publication,Nooember 2, 1976.