clays and clay minerals. vol. 37, no. 6, 515-524, 1989.coweeta.uga.edu/publications/347.pdf ·...

Clays and Clay Minerals. Vol. 37, No. 6, 515-524, 1989.

WEATHERING OF HORNBLENDE TO FERRUGINOUS PRODUCTS

BY A DISSOLUTION-REPRECIPITATION MECHANISM:

PETROGRAPHY AND STOICHIOMETRY

MICHAEL ANTHONY VELBELDepartment of Geological Sciences, 206 Natural Science Building

Michigan State University, East Lansing, Michigan 48824

Abstract—Hornblende of the Carrol Knob mafic complex (southern Blue Ridge Mountains, North Car-olina) has weathered under humid, temperate conditions. Hornblende weathering appears to have beena dissolution-reprecipitation reaction, in which hornblende dissolved stoichiometrically, and the ferru-ginous and aluminous weathering products (goethite, gibbsite, and kaolinite) precipitated from solution(neoformation). During the earliest stage of alteration, ferruginous weathering products formed as liningsof fractures within and around crystals and cleavage fragments of hornblende. Side-by-side coalescenceof lenticular etch pits during more advanced weathering produced characteristic "denticulated" termi-nations on hornblende remnants in dissolution cavities bounded by ferruginous boxworks. Dissolutioncavities are devoid of weathering products. Small "pendants" of ferruginous material project from theboxwork into void spaces. Because these products are separated from the hornblende remnants by voidspace, they must have been produced by dissolution-reprecipitation reactions. Complete removal of theparent hornblende left a ferruginous microboxwork or "negative pseudomorph." Only Al and Fe wereconserved over microscopic distances; alkali and alkaline-earth elements were stoichiometrically removedfrom the weathering microenvironment during the weathering process.Key Words—Gibbsite, Goethite, Hornblende, Iron, Kaolinite, Neoformation, Petrography, Scanningelectron microscopy, Weathering.

INTRODUCTION

Recent geochemical studies have suggested that min-eral-water interactions involving ferromagnesian sili-cates (especially pyroxenes and amphiboles, or "pyri-boles") play a significant role in base-cation andhydrogen ion budgets during the weathering of land-scapes underlain by silicate rocks (e.g., Katz et al, 1985;April et al., 1986; Bricker, 1986; Drever and Hurcomb,1986). The mechanisms of these reactions in naturalsystems, however, are not well understood. During pet-rographic studies of hornblende weathering in thesouthern Blue Ridge Mountains, the present authorobserved crystallographically aligned aggregates of fer-ruginous weathering products of hornblende. Becausequantitative studies of the stoichiometry of the weath-ering process in natural systems are still lacking despitenumerous studies of pyribole alteration textures andreplacement minerals, and because the mechanism andstoichiometry of hornblende weathering are importantin neutralization of acid deposition, the aggregates andtheir textural relations to the parent mineral were ex-amined in detail and are reported here.

PREVIOUS WORK

Cation depleted (leached) layers have often been invokedto explain apparent non-stoichiometric dissolution of silicateminerals in laboratory systems and surface compositions onnaturally weathered soil pyriboles. Schott et al. (1981), Bernerand Schott (1982), and Schott and Berner (1983,1985) studiedexperimentally and naturally weathered surfaces of iron-free

and iron-bearing pyroxenes by X-ray photoelectron spectros-copy (XPS). They found that a thin layer (a few tens of Ang-stroms thick) depleted in Ca and Mg relative to silica formedon iron-free pyriboles. Berner et al. (1985) pointed out thatthe XPS data were equally well explained by localized Caremoval along dislocation cores, rather than uniform deple-tion across the entire mineral surface, and that the formerinterpretation was consistent with the existence of etch pitson the surfaces of altered pyroxenes. Schott and Berner's (1983,1985) results on iron-bearing silicates experimentally weath-ered in an oxidizing environment demonstrated the presenceof a hydrous ferric oxide layer superimposed on a Mg-de-pleted, protonated ferric silicate layer.

Petit etal. (1987) examined experimentally weathered diop-side by a resonant nuclear reaction and found direct evidencethat diopside dissolution proceeded by surface hydration todepths of about 1000 A. They also found notable (and stoi-chiometric) depletion of primary elements, signifying in-creased microporosity over the thickness analyzed. Petit etal. (1987) interpreted these results to suggest the diffusion ofwater along defects, followed by stoichiometric dissolution ofthe diopside. Using Auger electron spectroscopy (AES), Mogkand Locke (1988) found preferential depletion of Mg, Ca, and,to a lesser degree, Si and Fe (relative to Al) over thicknessesof as much as 1200 A in naturally weathered hornblende.Their observations indicate non-stoichiometric dissolutionsimilar to that reported by Berner and his coworkers, but overthicknesses similar to those observed by Petit et al. (1987).

Previous petrographic studies of naturally weathered pyr-iboles have noted what appeared to be early-stage weatheringproducts in fractures and cleavages (see, e.g., Basham, 1974;Cleaves, 1974;Colman, 1982; Proust, 1982,1985). Alterationapparently originated at grain boundaries and intragranularfractures and grew inward to the present contact. If numerouscleavage fragments of the same parent crystal exhibit suchalteration and the relict parent-mineral cores have been re-

Copyright © 1989, The Clay Minerals Society 515

516 Velbel Clays and Clay Minerals

moved, the texture has been described as boxwork' (Stoopsetal., 1979).

The initial weathering products of ferromagnesian silicateminerals may be iron oxides, iron oxyhydroxides, or clayminerals, depending on the intensity of leaching in the weath-ering profile (Cleaves, 1974). Continued weathering produceseither pseudomorphs of clay minerals after pyribole (e.g., Eg-gleton, 1975; Cole and Lancucki, 1976; Delvigne, 1983; Eg-gleton et al., 1987) or a boxwork structure in which fracturefillings are preserved, but the pyribole itself is dissolved (Bas-ham, 1974; Cleaves, 1974; Delvigne, 1983).

Most pyribole weathering products exhibit preferred crys-tallographic orientation (e.g., Basham, 1974; Wilson, 1975;Delvigne, 1983), butnot all (Wilson andFarmer, 1970;Bernerand Schott, 1982). Eggleton (1975) and Cole and Lancucki(1976) suggested that the crystallographic orientation of theproducts with respect to parent structures resulted from "sol-id-solid" transformation involving minimal rearrangement oftetrahedra and octahedral cations. Weathering products formedby solid-solid transformation inherit a significant portion oftheir silicate skeleton from the parent mineral, in contrast toneoformation, in which the weathering products form by ac-tual precipitation from solution (Duchaufour, 1982; Eslingerand Pevear, 1988). Recent studies of weathered and/or oth-erwise altered pyriboles using high-resolution transmissionelectron microscopy (HRTEM) have supported the transfor-mation interpretation (e.g., Eggleton and Boland, 1982; Eg-gleton, 1986).

The studies summarized above can be assigned to two dis-tinct, non-overlapping categories: studies of the stoichiometryof pyribole weathering, and petrographic studies of parent-product textural relationships. In the present investigationweathering textures have been used to make inferences re-garding reaction stoichiometry.

GEOLOGY OF THE STUDY AREA

The study area is the Coweeta Hydrologic Labora-tory of the U.S. Department of Agriculture Forest Ser-vice, located in the Nantahala Mountains 15 km south-west of Franklin, North Carolina (Figure 1).Physiographic, climatic, and lithologic backgroundmaterial was summarized by Velbel (1985a). Bedrockin the area consists of high-rank (amphibolite facies)metasedimentary schists and gneisses, with local podsof manc-ultramafic rocks (Hatcher, 1979, 1980). Slopesare steep (average 27°), annual average rainfall is high(~2 m), and temperatures are moderate (mean annualtemperature is 12.8°C). All crystalline rocks of the studyarea have been extensively weathered to saprolite, whichranges from 0 to 20 m in thickness (avg. 6 m). Thesoils are Ultisols and Inceptisols; the former occur chief-ly on slopes, and the latter are developed on trans-ported colluvial parent materials (Velbel, 1988). A sin-gle logging road transects the study watershed. Theroadcut on the upslope side provides the only expo-sures. Most exposed material consists of reddish friable

1 The American Geological Institute Dictionary of Geolog-ical Terms (1976) (p. 53-54) defines boxwork as: "Limoniteand other minerals which originally formed as blades or platesalong cleavage or fracture planes and then the interveningmaterial dissolved leaving the intersecting blades or plates asa network."

saprolite; only one partially altered corestone was re-covered.

The hornblende examined in this study is from theCarrol Knob complex in watershed 3 of the CoweetaHydrologic Laboratory (Figure 1). In the immediatestudy area, only the plagioclase-hornblende-gneiss(metagabbro; McSween and Hatcher, 1985) facies ofthe Carrol Knob complex is exposed (Hatcher, 1980).These rocks consist of hornblende and plagioclase, withminor quartz, epidote, and other accessory minerals(McSween and Hatcher, 1985). McSween and Hatcher(1985) showed that the Carrol Knob complex is com-positionally very similar to the Laurel Creek complexin Georgia, less than 25 km southeast of the CoweetaHydrologic Laboratory. Microprobe analyses andstructural formulae of calcic amphiboles (hornblendeto ferroan pargasite in composition) from the LaurelCreek complex (Helms et al., 1987) were used in thepresent study.

MATERIALS AND METHODS

A single corestone (~16 x 12 x 7.5 cm) having awell-developed weathering rind (2.5-5 cm thick) wascollected from the saprolite matrix at a depth of about0.5 m within the boundaries of watershed 3 (Figure 1).Suites of epoxy-impregnated petrographic thin sectionsfrom zones sampled outward from the fresh core to theextensively weathered rind were examined on a LeitzOrtholux II-pol petrographic microscope.

Microboxwork areas containing hornblende relics andisolated hornblende grains were exhumed from the handsamples with dissecting needles and forceps under abinocular microscope; some were cleaned ultrasoni-cally. These samples were then affixed to SEM stubswith double-sided adhesive tape and coated for SEMexamination. The samples were examined using ETECAutoscan and JEOL JSM T-20 scanning electron mi-croscopes (SEMs).

The brittle nature of the boxwork making up therind precluded fine subdivision of the weathered por-tion of the sample. The transition between fresh coreand weathered rind was abrupt (<0.3 cm), and littleheterogeneity in the weathered rind was visible. Thus,the weathering rind was characterized by a single bulksample of fragments, which were crushed and groundfor X-ray powder diffraction (XRD). The <10-/wnfraction was prepared using the method of Drever (1973)and scanned from 2° to 30°20 on a Rigaku X-ray pow-der diffractometer at a scan rate of l°20/min, using Ni-filtered CuKa radiation. Divergence, receiving, andscatter slits were V6°, 0.3 mm, and 2°, respectively.

RESULTS

Hand specimens

The corestone consisted of a 2.5 x 5 cm core of fresh(not visibly oxide-stained) hornblende amphibolite-

Vol. 37, No. 6, 1989 Neoformation of hornblende weathering products 517

LEGEND, Watershed47 number & boundaryC4* Core locality

Shape Fork Fault'////' Corro/ Km>6 Complex

Figure 1. Map of study area, southern Blue Ridge Moun-tains, North Carolina.

metagabbro, surrounded by a thick (2.5 cm thick alongshort axis of corestone, 5 cm thick along long axis)zone consisting of a complex orange "microboxwork".Black hornblende relics were visible within the micro-boxwork near the core. The abundance of hornblenderelics decreased abruptly with increasing distance fromthe fresh core, so most of the weathered rind consistedof microboxwork containing no hornblende.

Petrography and SEM morphology

Fresh hornblende occurred in several different pe-trographic forms, including anhedral poikiloblasts con-taining inclusions of other silicates (e.g., quartz) andeuhedral prisms. Thin sections transected the latter invarious orientations, giving both longitudinal sections(seen perpendicular to the z-axis) and cross-sections(viewed essentially parallel to the z-axis; Figure 2). Mi-croboxworks were developed after all forms and ori-entations.

In incipiently altered materials, ferruginous weath-ering products appear to have lined fractures withinand around hornblende crystals and cleavage frag-ments. Little hornblende appeared to have been re-moved, leaving only a few small (< 20 /^m) visible voidsbetween the hornblende remnant and the microbqx-work. The thickness (<40 jtm) of the ferruginous septaat this stage was greater than the width (<5 /mi) offractures in the fresh hornblende.

More advanced weathering produced "sawtooth" or"denticulated" terminations (Figures 4 and 5), in which

Figure 2. Photomicrograph of pseudohexagonal prismaticcross-section of hornblende, showing characteristic 56° and124° cleavages (C, arrows). Plane-polarized light; field of viewis 1.5 mm across.

the "teeth" were parallel to the z-axis of the horn-blende. These denticulated terminations coexisted withalmond-shaped, lenticular etch pits on prism faces,elongated parallel to the z-axis (Figure 3). In still moreadvanced stages, hornblende remnants occurred in dis-solution cavities bounded by ferruginous boxworks(Figure 6). Figure 6 shows such a boxwork separatedfrom denticulated hornblende by void space. The septaof the microboxwork (Figures 5 and 6) actually con-sisted of a symmetrical pair of layers, separated by acentral parting. The layers themselves appear to havea vaguely defined fibrous character (Figure 6), in whichthe fibers are perpendicular to the orientation of thegrain-transecting fractures. The dissolution cavities aredevoid of weathering products; they contain no non-crystalline intermediary products, nor are the etchedhornblende surfaces covered by any well-crystallizedclay minerals. Figure 6 shows, however, oriented fer-

Figure 3. Scanning electron micrograph of cleavage-paralleletch-pits in weathered hornblende. i


Figure 4. Scanning electron micrograph showing denticu-lations resulting from etching of hornblende.

ruginous pendants, which, like the microboxwork, arealso separated from the denticulated margin by voidspace.

Thin-sections of areas equivalent in scale and weath-ering stage to that shown in Figure 6 also showed den-ticulated hornblende remnants separated from micro-boxwork by void space. Subtle changes in thebirefringence of the boxwork were apparent in cross-polarized light. If the field of view was rotated so thatthe hornblende z-axis was parallel to the polarizingfilter, individual boxwork segments in prism-parallelcleavage pseudomorphs exhibited simultaneous ex-tinction under cross-polarized light. If the same fieldof view was rotated 45°, weak birefringence colors ap-peared simultaneously in several different portions ofthe boxwork, all of which were previously at extinction.

Complete removal of the parent hornblende pro-

Figure 5. Scanning electron micrograph showing (1) incip-ient microboxwork (B), occurring as paired layers of ferru-ginous products having a central parting, and (2) incipientformation of denticulated margins (D) by congruent disso-lution of hornblende.

Figure 6. Scanning electron micrograph of weathered horn-blende showing ferruginous microboxwork (B) consisting ofpaired layers having a central parting, separated from dentic-ulated hornblende (D) by void space. Note oriented pendants(P) projecting from the microboxwork into the void space.Note also that pendants are separated from the denticulatedhornblende by void space.

duced a "negative pseudomorph," i.e., a boxwork con-sisting of ferruginous materials. Figures 7 and 8 showadvanced stages of hornblende weathering, in whichthe hornblende has been completely removed. All thatremains is ferruginous boxwork, which mimics theprism cleavages of the parent hornblende (cf. Figure2). Here also, the septa of the microboxwork showbilateral symmetry with respect to the central parting(cf. Figure 6 and discussion thereof). Small pendantsof similar material project from the boxwork into thevoid spaces.

Figure 9 is a photomicrograph taken in plane-po-larized light of pendants, which project from the box-work into the void space. Examined in the same ori-entation in cross-polarized light, the same field of viewexhibited parallel extinction of the material making upthe pendants. If the microscope stage was rotated 45°counterclockwise, birefringence colors appeared simul-taneously in several different portions of the boxworkand, especially, the pendants, all of which were at op-tical extinction in the previous orientation (Figure 9).


Figure 7. Photomicrograph showing advanced stage of horn-blende weathering, in which hornblende has been completelyremoved, leaving only ferruginous microboxwork (open ar-rows right of center), which mimics the 56° and 124° cleavageof the parent hornblende (compare with Figure 2). Plane-polarized light; field of view is 1.5 mm across.

Powder X-ray diffraction

An XRD pattern of the <10-/im fraction of bulkboxwork fines (Figure 10) shows strong gibbsite andkaolinite peaks, weak goethite peaks, and a weak quartzpeak.

DISCUSSION

Mineralogy of the ferruginous material

Both optical and XRD examination of the bulk finesindicated the presence of goethite, gibbsite, and ka-olinite. Goethite exhibited parallel extinction (Kerr,1977), as did the pendants, and gibbsite exhibited highbirefringence, one of the characteristics of the micro-boxworks and pendants. Microprobe analyses and

Figure 9. Photomicrograph of weathered hornblende show-ing ferruginous microboxwork (B) and delicate pendants (P).Plane-polarized light; field of view is 600 Mm across.

structural formulae of calcic amphiboles (hornblendeto ferroan pargasite in composition) from the nearbyand petrologically similar Laurel Creek complex (Helmset al., 1987) showed appreciable Al, consistent withthe aluminous nature of the weathering products of theCarrol Knob hornblende. Thus, the aluminous prod-ucts (gibbsite and kaolinite) probably formed from theweathering of Carrol Knob hornblendes.

Origin of the observed alteration textures

Incipient stage of weathering. During incipient altera-tion, ferruginous weathering products formed an in-cipient boxwork, lining fractures within and aroundcrystals and cleavage fragments of hornblende. Theboxwork occurred as paired layers having a centralparting. Texturally similar materials were previouslyinterpreted as chemically precipitated microfracturefillings by Berner and Schott (1982); however, thethickness of the ferruginous boxwork at this stage (as

Gi

Figure 8. Scanning electron micrograph of weathered horn-blende showing ferruginous microboxwork (B) consisting ofpaired layers. Note oriented pendants (e.g., P) projecting frommicroboxwork into void space.

°26Figure 10. X-ray powder diffraction pattern of microbox-work-rich outer portion of cobble C79-40, indicating the pres-ence of kaolinite (K), gibbsite (Gi), goethite (Go), and quartz(Q). < 10-jnm fraction, Mg-saturated; Ni-filtered CuKa radia-tion.


thick as 40 /am in petrographic thin sections) was muchgreater than the width of fractures (< 5 iaa) in the freshhornblende. The paired layers were too thick to bechemically precipitated (neoformed) fracture fillings;therefore, they probably were replacement materials(either transformation products, or products of highlylocalized dissolution-neoformation reactions) emanat-ing symmetrically from the fracture. The ferruginousalteration products appear to have originated at thehornblende grain boundaries and intragranular frac-tures and grew inward to the present contact.

Textures identical to the observed weathering pat-tern have been reported in serpentinized pyroxenes andamphiboles and described as mesh textures [nomen-clature of Wicks et al, 1977; specifically, the (retro-grade) type-3 serpentinization texture of Wicks andWhittaker, 1977]. Cressey (1979) suggested that suchserpentinization textures resulted from localizeddissolution-reprecipitation reactions rather than fromtopotactic replacement. In serpentinized pyriboles,mesh-textured replacements across the interior of theserpentine pseudomorphs always have central partings(Wicks and Whittaker, 1977). Such bilaterally sym-metrical mesh components have been interpreted asalterations along pre-existing grain-transecting frac-tures and cleavages (Wicks and Whittaker, 1977). Cen-tral partings in boxwork septa have been widely ob-served and illustrated in weathering studies of ferro-magnesian silicates (e.g., BemerandSchott, 1982; Velbel,1984); however, unlike serpentinization studies, re-ports of weathering studies seldom mention centralpartings as a significant petrographic feature. All mi-croboxwork septa observed in the present study havecentral partings (Figures 5 and 6), consistent with thesuggestion that grain-transecting fractures (both ran-dom fractures and internal cleavage planes) are theprimary loci of initial attack during hornblende weath-ering.

Intermediate stages of weathering. "Sawtooth" or"denticulated" terminations characteristic of pyriboleremnants at intermediate stages of weathering (Figures4-6) have been observed by numerous workers in otherstudy areas (e.g., Cleaves, 1974; Delvigne, 1983; Colinetal, 1985; Wilson, 1986). Berner e£al. (1980)showedthat etch pits (see Figure 3) reflect selective attack atdefects and dislocations; the sharp-pointed denticula-tions (see Figures 4-6) probably resulted from side-by-side coalescence of lenticular etch pits (Berner et al.,1980; Berner and Schott, 1982).

"Denticulated" margins are commonly empty (i.e.,devoid of clay minerals or other weathering products)in nature (Cleaves, 1974; Glasmann, 1982); all areempty at Coweeta (Figures 3-6). The formation of thedissolution voids involves stoichiometric dissolution—the entire volume of mineral that formerly occupiedthe void has been removed. The expression for stoi-

chiometric dissolution of hornblende at Coweeta, usinga slightly idealized parent-mineral composition basedon the data of Helms et al. (1987), is:

Na0.5Ca2(Fe1.3Mg2.6Al1.1)(Al1.6Si6.4)022(OH)2

+ 15H+ + H2O - 0.5 Na+ + 2Ca2+ + 2.6Mg2+

+ 1.3Fe2+ + 2.7A1(OH)2+ + 6.4 H4SiO4(aq) (1)

The fact that pendants of ferruginous material werenoted, separated from hornblende remnants by voidspace (Figure 6) suggests that the pendants must haveformed by dissolution-reprecipitation (neoformation)reactions. During these reactions, the hornblende dis-solved stoichiometrically (reaction (1)), the dissolvedmaterials were transported through the void space, andFe, Al, and Si were incongruently reprecipitated ontothe earlier formed boxwork, according to the followingreactions:

A1(OH)2+ + H20 « A1(OH)3 + H+ (2)

2 A1(OH)2+ + 2 H4SiO4 <->

Al2Si2O5(OH)4 + 3 H2O + 2 H+ and (3)2 Fe2+ + '/2 O2 + 3 H2O « 2 FeO(OH) + 4 H+. (4)

The absence of 2:1 clay minerals in saprolites of Co-weeta is apparently due to intense leaching in the studyarea, which prevented solutions in the weathering pro-files from becoming saturated with respect to 2:1 clayminerals. Previous workers (Wilson and Farmer, 1970;Cleaves, 1974; Nahon and Colin, 1982; Colin et al.,1985) reported the formation of 2:1 clays only frompyribole weathering in poorly drained weathering en-vironments and/or microenvironments, whereas 1:1clays and/or oxyhydroxides have been reported fromwell-leached macro- and microenvironments (Cleaves,1974; Nahon and Colin, 1982). Natural solutions inpreviously studied watersheds in the study area (Figure1) are in equilibrium with kaolinite, but not with 2:1clays (Velbel, 1985b).

Advanced stages of weathering. Complete removal ofthe parent hornblende left a "negative pseudomorph,"i.e., a boxwork (Figures 7-9) consisting of goethite,gibbsite, and kaolinite. In his study of pyroxene weath-ering, Delvigne (1983) also described crystallographi-cally aligned "porous pseudomorphs" identical to themicroboxwork-pendant assemblages reported here.

The pendants were generally optically parallel to oneanother and to the relict prismatic cleavage of the par-ent hornblende (Figures 7-9). The pendants most likelyinherited this optic orientation from the parent horn-blende by epitactic nucleation and growth upon the(topotactic?) boxwork that formed during earlier stagesof hornblende weathering. The petrographic and SEMdata suggest that crystallographically oriented altera-tion products of chain-silicate weathering formed byboth neoformation (dissolution-reprecipitation) andtransformation.


Scales of elemental redistribution during weatheringofferromagnesian silicates

To examine the possible consequences of the petro-graphically inferred dissolution-reprecipitation mech-anism for the mobility of Al, Fe, and Si during horn-blende weathering, the volume of boxwork and pendantgoethite, gibbsite, and kaolinite that would be formedby the weathering of a unit volume of parent horn-blende was calculated, assuming that Fe, Al, and Siwere conserved. The total number of moles of a givenelement (e) in parent mineral (p) is given by

me,]_ ce,pne,kDpVp

Mk(5)

where me>p = total number of moles of element e inthe parent mineral p; ce p = weight fraction of the oxideof element e in parent mineral p (oxide wt. % x 10~2);nek = stoichiometric coefficient of element e in analyteoxide of element e; Dp = density of parent mineral p;Vp = volume of parent mineral p; and Mk = gram-formula weight of analyte oxide k. The total numberof moles of element e per unit volume of daughtermineral d can be expressed as

(6)

where me d = total number of moles of element e inthe daughter mineral d; neid = stoichiometric coefficientof element e in daughter mineral d; Vd = volume ofdaughter mineral d; and V°d = molar volume of daugh-ter mineral d. Rearranging Eq. (6) to solve for med,setting meip = me d (that is, conserving element e, lettingall of element e present in the parent mineral be in-corporated into the daughter mineral), combining Eqs.(5) and (6), and rearranging gives

V°dce,pne,KDp _ Vd

Mkne,d Vp(7)

where Vd/Vp is the volume of daughter mineral pro-duced per unit volume of parent mineral, if element eis conserved.

Averaged microprobe data from Laurel Creek horn-blendes (hornblende to ferroan pargasite in composi-tion; Helms et al., 1987) were used to estimate thecomposition of the parent hornblende, using molarvolumes for the three daughter minerals from Smythand Bish (1988). The density of pargasite (Dp = 3.165)as reported by Smyth and Bish (1988) was also used;the value chosen is well within the range of densitiesfor compositionally similar hornblendes, as tabulatedby Deer et al. (1963). Values used in the calculationsand the results are shown in Table 1.

If Al and Fe are conserved on the scale of the parent-mineral crystals, and goethite and kaolinite are formed,the maximum possible volume of alteration products

is (0.1096 + 0.4880 = 0.5976; Table 1) x the originalvolume of hornblende. Thus, the total volume of al-teration products formed by weathering of hornblendeto hydroxides and/or aluminosilicates cannot exceed60% of the original volume of hornblende, if Al andFe are conserved. If some of the Al is incorporatedinto gibbsite instead of kaolinite, the total volume ofmaterial produced will be less, as little as (0.1096 +0.3169 = 0.4265) x the original hornblende volume(here also, complete conservation of Fe and Al is as-sumed).

Modal abundances of boxwork and void space wereestimated by point counting Figures 7 and 9 (856 and1051 points, respectively). The volumes of the box-work and pendant materials are 59.7% and 60.9%, re-spectively. Within experimental error (±3.5%, Petti-john et al., 1987, their Figure A-3), the measured valuescorrespond exactly to the volume calculated, assumingthat Al and Fe are conserved in goethite and kaolinite.To whatever extent the point counts failed to detectmicroporosity, the actual solid volume of products maybe somewhat smaller, which is consistent with the pres-ence of gibbsite. The petrographic data strongly sup-port the hypothesis that Al and Fe were conservedduring weathering of the hornblende to microboxworkand pendants.

Dissolution-reprecipitation mechanism andthe stoichiometry offerromagnesiansilicate weathering

Dissolution-reprecipitation mechanisms have beeninvoked on petrographic grounds to explain severaloccurrences of pyribole weathering (e.g., Berner andSchott, 1982; Glasmann, 1982), serpentinization (e.g.,Cressey, 1979), feldspar weathering (e.g., Grant, 1963,1964; Koppi and Williams, 1980; Anand and Gilkes,1984; Anand et al., 1985), and pseudomorphous clayreplacements after plagioclase feldspar, which appearto preserve twinning in clay pseudomorphs (Velbel,1983). Transport of ostensibly "immobile" elementsin dissolved form is also required to account for thecomposition of pseudomorphous weathering productsof ferromagnesian silicates. The pseudomorphs com-monly contain elements not found in the parent min-eral (e.g., Al in products of pyroxene weathering; Na-hon et al., 1982; Nahon and Bocquier, 1983; Fontanaudand Meunier, 1983; Parisot et al, 1983). The presentdata suggest that crystallographically oriented productsof pyribole alteration formed by dissolution-reprecip-itation (neoformation) as well as by previously pos-tulated solid-state (transformation) reactions that in-volved minimal rearrangement of tetrahedra andoctahedral cations.

The stoichiometric expression for the overall horn-blende weathering reaction at Coweeta results by add-ing reactions (1H4) (reactions (2)-(4) have been mul-


Table 1. Volume of alteration products formed per unit volume of hornblende weathered, assuming conservation of Fe, Al,and Si.

Product/element conserved Mk (g/mole) VV (cmVmole)

Goethite/FeGibbsite/AlKaolinite/SiKaolinite/Al

12.0215.8444.3715.84

71.8464101.9612860.0843

101.96128

1212

1122

20.69332.22299.23699.236

0.10960.31691.1600.4880

1 Concentration of element in parent mineral; average of seven values from Helms et al. (1987).2 Molar volume of daughter mineral; data from Smyth and Bish (1988).3 Data from Smyth and Bish (1988). The value chosen is well within the range of densities reported for hornblendes

compositionally similar to the Laurel Creek/Carrol Knob hornblendes, as tabulated by Deer et al. (1963).Other symbols are denned in text (Eqs. (5)-(7)).

tiplied by appropriate stoichiometric coefficients forFe, Al, and Si from Eq. (1)):

Nao.5Ca2(Fe, ,3Mg2.6 Al ,. , )(A1 , .6Si6.4)O22(OH)2

+ 9.7 H+ + 11.65H2O + 0.325 O2 -'0.5 Na+ + 2Ca2+ + 2.6Mg2+ + 1.3FeO(OH)

+ xA!2Si2O5(OH)4 + (2.7 - 2x)Al(OH)3

+ (6.4-2x)H4Si04(aq), (8)

where some indeterminate amount (x) of kaolinite isformed. The value x could have been determined if theSi content of the boxwork-pendant assemblage had beenknown.

Both petrographic photomicrographs and SEMs sug-gest that the combined volume of boxwork + pendantsformed from the neoformation reaction discussed above(as a fraction of original hornblende volume) is exactlythat predicted by assuming the conservation of Fe andAl; so, no export or loss of these elements is indicated.Some Si was also retained in the weathering products;however, all other elements (Na, Ca, and Mg) wereremoved from the hornblende-weathering microen-vironment. The overall incongruent stoichiometry ofthe hornblende weathering reaction at Coweeta (Eq.(8)) resulted from stoichiometric dissolution of thehornblende (Eq. (1)), followed by non-stoichiometricreprecipitation of Fe, Al, and some dissolved Si duringneoformation of gibbsite, goethite, and kaolinite (Eqs.(2)-(4)). No "leached layer" is required to explain thenon-stoichiometry of the overall hornblende weath-ering reaction. The stoichiometric nature of the dis-solution step is consistent with the findings of Petit etal (1987).

SUMMARY AND CONCLUSIONS

Hornblende of the Carrol Knob mafic complexweathered under humid, temperate conditions, accom-panied by intense leaching of the weathering profile.Petrographic and SEM observations reveal the com-plete transition from fresh hornblende to ferruginousand aluminous products (goethite, gibbsite, and ka-olinite). During the earliest stage of alteration, ferru-

ginous weathering products lined fractures within andaround hornblende crystals and cleavage fragments.More advanced weathering produced characteristic"sawtooth" or "denticulated" terminations on horn-blende remnants within dissolution cavities boundedby ferruginous boxworks. These sharp pointed dentic-ulations resulted from side-by-side coalescence of len-ticular etch pits. Dissolution cavities were found to bedevoid of weathering products; no noncrystalline in-termediary was noted, nor were well-crystallized clayminerals present on the etched hornblende remnants.The absence of such material indicates that the volumeof mineral which formerly occupied the volume nowrepresented by the etch pits was removed stoichio-metrically from the etch-pit microenvironment. Theabsence of 2:1 clay minerals was probably the resultof intense leaching in the study area, which preventedsolutions in the weathering profiles from becoming sat-urated with respect to 2:1 clay minerals.

Intermediate and advanced weathering resulted inboxwork texture, generally as paired layers having cen-tral partings, many of which were too thick to be sym-metrical, chemically precipitated fracture fillings, andtherefore, probably represented thin replacements em-anating symmetrically from the fracture. Simultaneousoptical extinction of individual boxwork segments inprism-parallel cleavage pseudomorphs suggests that thegoethite-gibbsite-kaolinite products were crystallo-graphically controlled by (topotactic?) replacement ofthe parent hornblende. Small "pendants" of ferrugi-nous material grew from the boxwork into the voidspaces. Because they are separated from the hornblenderemnants by void space, these pendants must havebeen formed by dissolution-reprecipitation. Duringthese reactions, the hornblende dissolved stoichio-metrically, the dissolved materials were transportedthrough the void space, and Fe, Al, and Si were in-congruently reprecipitated onto the earlier formed box-work. Pendants were generally optically parallel to theprismatic cleavage, and most likely inherited this ori-entation from the parent hornblende by epitactic nu-


cleation and growth upon the earlier-formed boxwork.Thus, crystallographically oriented products of pyri-bole alteration probably formed by dissolution-repre-cipitation, as well as by solid-state reactions involvingminimal rearrangement of tetrahedra and octahedralcations. Complete removal of the parent hornblendeleft a "negative pseudomorph," a microboxwork con-sisting of goethite, gibbsite, and kaolinite.

The dissolution-reprecipitation mechanism has sig-nificant consequences for the distance-scales of elementmobilization during hornblende weathering. The over-all incongruent stoichiometry of the hornblende weath-ering reaction at Coweeta (reaction (8)) resulted fromstoichiometric dissolution of the hornblende, followedby non-stoichiometric reprecipitation of Fe, Al, andsome dissolved Si to form the gibbsite, goethite, andkaolinite. No "leached layer" was required to explainthe incongruent nature of the overall hornblendeweathering reaction.

ACKNOWLEDGMENTS

I am grateful to J. D. McKee, D. F. Sibley, F. J.Wicks, A. B. Taylor, N. L. Romero, D. S. Brandt andT. Kremer, for helpful discussions at various stages ofthis project. Reviews by D. B. Nahon, F. A. Mumpton,D. S. Brandt, D. F. Sibley, and an anonymous reviewerare greatly appreciated. This research was supportedby NSF Grant BSR 85-14328.

REFERENCES

American Geological Institute (1976) Dictionary of GeologicTerms: Revised ed., W. H. Matthews, 3rd and R. E. Boyer,eds., Anchor Press/Doubleday, Garden City, New York,472 pp.

Anand, R. R. and Gilkes, R. J. (1984) Weathering of horn-blende, plagioclase and chlorite in meta-dolerite, Australia:Geoderma34, 261-280.

Anand, R. R., Gilkes, R. J., Armitage, T. M., and Hillyer, J.W. (1985) Feldspar weathering in lateritic saprolite: Clays& Clay Minerals 33, 31^3.

April, R. A., Newton, R. M., and Coles, L. T. (1986) Chem-ical weathering in two Adirondack watersheds: Past andpresent-day rates: Geol. Soc. Amer. Bull. 97, 1232-1238.

Basham, I. R. (1974) Mineralogical changes associated withdeep weathering of gabbro in Aberdeenshire: Clay Miner.10, 189-202.

Berner, R. A. and Schott, J. (1982) Mechanism of pyroxeneand amphibole weathering—II. Observations of soil grains:Amer. J. Sci. 282, 1214-1231.

Berner, R. A., Sjoberg, E. L., Velbel, M. A., and Krom, M.D. (1980) Dissolution of pyroxenes and amphiboles dur-ing weathering: Science 207, 1205-1206.

Berner, R. A., Holdren, G. R., Jr., and Schott, J. (1985)Surface layers on dissolving silicates (Comments on "Studyof the weathering of albite at room temperature and pres-sure with a fluidized bed reactor" by L. Chou and R. Wol-last): Geochim. Cosmochim. Acta 49, 1657-1658.

Bricker,O.P. (1986) Geochemical investigations of selectedeastern United States watersheds affected by acid deposi-tion: J. Geol. Soc., London 143, 621-626.

Cleaves, E. T. (1974) Petrologic and chemical investigation

of chemical weathering in mafic rocks, eastern piedmontof Maryland: Maryland Geol. Surv. Kept. Invest. 25, 38 pp.

Cole, W. F. and Lancucki, C. J. (1976) Montmorillonitepseudomorphs after amphibole from Melbourne, Australia:Clays & Clay Minerals 24, 79-83.

Colin, F., Noack, Y., Trescases, J.-J., and Nahon, D. (1985)L'alteration laterique debutante des pyroxenites de Jacuba,Niquelandia, Bresil: Clay Miner. 20, 93-113.

Colman, S. M. (1982) Chemical weathering of basalts andandesites: Evidence from weathering rinds: U.S. Geol. Surv.Prof. Pap. 1246,51 pp.

Cressey, B. A. (1979) Electron microscopy of serpentinitetextures: Can. Mineral. 17, 741-756.

Deer, W. A., Howie, R. A., and Zussman, J. (1963) Rock-Forming Minerals, Vol.2, Chain Silicates: Longmans, Lon-don, 379 pp.

Delvigne, J. (1983) Micromorphology of the alteration andweathering of pyroxenes in the Koua Bocca ultramafic in-trusion, Ivory Coast, West Africa: in Petrologie des Altera-tions et des Sols, Vol. 2: D. Nahon and Y. Noack, eds., Sci.Geol. Mem. 72, 57-68.

Drever, J. I. (1973) The preparation of oriented clay mineralspecimens for X-ray diffraction analysis by a filter-mem-brane peel technique: Amer. Mineral. 58, 553-554.

Drever, J. I. and Hurcomb, D. R. (1986) Neutralization ofatmospheric acidity by chemical weathering in an alpinedrainage basin in the North Cascade Mountains: Geology14, 221-224.

Duchaufour, P. (1982) Pedology, 1. R. Paton, transl.: GeorgeAllen & Unwin, London, 448 pp.

Eggleton, R. A. (1975) Nontronite topotaxial after heden-bergite: Amer. Mineral 60, 1063-1068.

Eggleton, R. A. (1986) The relation between crystal structureand silicate weathering rates: in Rates of Chemical Weath-ering of Rocks and Minerals, S. M. Colman and D. P. De-thier, eds., Academic Press, New York, 21-40.

Eggleton, R. A. and Boland, J. N. (1982) Weathering ofenstatite to talc through a sequence of transitional phases:Clays & Clay Minerals 30, 11-20.

Eggleton, R. A., Foudoulis, C., and Varkevisser, D. (1987)Weathering of basalt: Changes in rock chemistry and min-eralogy: Clays & Clay Minerals 35, 161-169.

Eslinger, E. and Pevear, D. (1988) Clay Minerals for Petro-leum Geologists and Engineers: Soc. Econ. Paleo. Mineral.Short Course Notes 22, 422 pp.

Fontanaud, A. and Meunier, A. (1983) Mineralogical faciesof a weathered serpentinized Iherzolite from the Pyrenees,France: Clay Miner 18, 77-88.

Glasmann, J. R. (1982) Alteration of andesite in wet, un-stable soils of Oregon's western Cascades: Clays & ClayMinerals 30, 253-263.

Grant, W. H. (1963) Weathering of Stone Mountain Gran-ite: in Clays and Clay Minerals, Proc. llth Natl. Conf.,Ottawa, Ontario, 1962, Ada Swineford, ed., Pergamon Press,New York, 65-73.

Grant, W. H. (1964) Chemical weathering of biotite-plagio-clase gneiss: Clays & Clay Minerals 12, 455-463.

Hatcher, R. D. (1979) The Coweeta Group and Coweetasyncline: Major features of the North Carolina-Georgia BlueRidge: Southeastern Geol. 21, 17-29.

Hatcher, R. D. (1980) Geologic map and mineral resourcessummary of the Prentiss Quadrangle, North Carolina, in-cluding geologic map of Coweeta Laboratory: North Car-olina Dept. Nat. Resources and Community DeveL, Geol.Surv. Sect., GM-167-SW and MRS 167-SW, 10 pp.

Helms, T. S., McSween, H. Y., Jr., Labotka, T. C., and Ja-rosewich, E. (1987) Petrology of a Georgia Blue Ridgeamphibolite unit with hornblende + gedrite + kyanite +staurolite: Amer. Mineral. 72, 1086-1096.


Katz, B. G., Bricker, O. P., and Kennedy, M. M. (1985)Geochemical mass-balance relationships for selected ionsin precipitation and stream water, Catoctin Mountains,Maryland: Amer. J. Sci. 285, 931-962.

Kerr, P. F. (1977) Optical Mineralogy, 4th ed., McGraw-Hill, New York, 492 pp.

Koppi, A. J. and Williams, D. J. (1980) Weathering anddevelopment of two contrasting soils formed from grano-diorite in south-east Queensland: Aust. J. Soil Res. 18,257-271.

McSween, H. Y., Jr. and Hatcher, R. D., Jr. (1985) Ophio-lites (?) of the southern Appalachian Blue Ridge: in FieldTrips in the Southern Appalachians, N. B. Woodward, ed.,Univ. Tennessee Dept. Geol. Sci. Studies in Geol. 9, 144-170.

Mogk, D. W. and Locke, W. W., III. (1988) Application ofauger electron spectroscopy (AES) to naturally weatheredhornblende: Geochim. Cosmochim. Acta 52, 2537-2542.

Nahon, D. and Bocquier, G. (1983) Petrology of elementstransfers in weathering and soil systems: in Petrologie desAlterations et des Sols, Vol. 2: D. Nahon and Y. Noack,eds., Sci. Geol. Mem. 72, 111-119.

Nahon, D. B. and Colin, F. (1982) Chemical weathering oforthopyroxenes under lateritic conditions. Amer. J. Sci. 282,1232-1243.

Nahon, D., Colin, F., and Tardy, Y. (1982) Formation anddistribution of Mg,Fe,Mn-smectites in the first stages of thelateritic weathering of forsterite and tephroite: Clay Miner.17, 339-348.

Parisot,J. C, Delvigne, J., andGroke, M. T. C. (1983) Pet-rographical aspects of the supergene weathering of garnetin the Serra dos Carajas (Para, Brazil): in Petrologie desAlterations et des Sols, Vol. 2, D. Nahon and Y. Noack,eds., Sci. Geol. Mem. 72, 141-148.

Petit, J.-C, Delia Mea, G., Dran, J.-C, Schott, J., and Berner,R. A. (1987) Mechanism of diopside dissolution fromhydrogen depth profiling: Nature 325, 705-707.

Pettijohn, F. J., Potter, P. E., and Siever, R. (1987) Sandand Sandstone, 2nd ed., Springer-Verlag, New York, 553PP.

Proust, D. (1982) Supergene alteration of hornblende in anamphibolite from Massif Central, France: in Proc. Int. ClayConf., Bologna, Pavia, 1981, H. van Olphen and F. Veniale,eds., Elsevier, Amsterdam, 357-364.

Proust, D. (1985) Amphibole weathering in a glaucophane-schist (He de Groix, Morbihan, France): Clay Miner. 20,161-170.

Schott, J. and Berner, R. A. (1983) X-ray photoelectronstudies of the mechanism of iron silicate dissolution duringweathering: Geochim. Cosmochim. Acta 47, 2233-2240.

Schott, J. and Berner, R. A. (1985) Dissolution mechanismsof pyroxenes and olivines during weathering: in The Chem-istry of Weathering, J. I. Drever, ed., Reidel, Dordrecht,The Netherlands, 35-53.

Schott, J., Berner, R. A., and Sjoberg, E. L. (1981) Mech-anism of pyroxene and amphibole weathering—I. Experi-mental studies of iron-free minerals: Geochim. Cosmochim.Acta 45, 2123-2135.

Smyth, J. R. and Bish, D. L. (1988) Crystal Structures andCation Sites of the Rock-Forming Minerals: Allen & Unwin,Boston, 332 pp.

Stoops, G., Altemuller, H.-J., Bisdom, E. B. A., Delvigne, J.,Dobrovolsky, V. V., Fitzpatrick, E. A., Paneque, G., andSleeman, J. (1979) Guidelines for the description of min-eral alterations in soil micromorphology: Pedologie 29, 121-135.

Velbel, M. A. (1983) A dissolution-reprecipitation mecha-nism for the pseudomorphous replacement of plagioclasefeldspars by clay minerals during weathering: in Petrologiedes Alterations et des Sols, Vol. 1, D. Nahon and Y. Noack,eds., Sci. Geol. Mem. 71, 139-147.

Velbel, M. A. (1984) Natural weathering mechanisms ofalmandine garnet: Geology 12, 631-634.

Velbel, M. A. (1985a) Geochemical mass balances andweathering rates in forested watersheds of the southern BlueRidge: Amer. J. Sci. 285, 904-930.

Velbel, M. A. (1985b) Hydrogeochemical constraints onmass balances in forested watersheds of the southern Ap-palachians: in The Chemistry of Weathering, J. I. Drever,ed., Reidel, Dordrecht, The Netherlands, 231-247.

Velbel.M.A. (1988) Weathering and soil forming processes:in Forest Hydrology and Ecology at Coweeta, W. T. Swankand D. A. Crossley, Jr., eds., Springer-Verlag, New York,93-102.

Wicks, F. J. and Whittaker, E. J. W. (1977) Serpentine tex-tures and serpentinization: Canad. Mineral. 15, 459-488.

Wicks, F. J., Whittaker, E. J. W., and Zussman, J. (1977)An idealized model for serpentine textures after olivine:Canad. Mineral. 15, 446-458.

Wilson, M. J. (1975) Chemical weathering of some primaryrock-forming minerals: Soil Sci. 119, 349-355.

Wilson, M. J. (1986) Mineral weathering processes in pod-zolic soils on granitic materials and their implications forsurface water acidification: J. Geol. Soc. London 143, 691-697.

Wilson, M. J. and Farmer, V. C. (1970) A study of weath-ering in a soil derived from a biotite-hornblende rock. II.The weathering of hornblende: Clay Miner. 8, 435^44.(Received 16 May 1988; accepted 29 March 1989; Ms. 1791)

clays and clay minerals. vol. 37, no. 6, 515-524, 1989.coweeta.uga.edu/publications/347.pdf ·...

Documents