Read 0702.pdf text version

CHEMISTRY AND ORIGIN OF FRESHWATER FERROMANGANESE CONCRETIONS Rob&

Department of Oceanography,

C. Hawks

State University, Tallahassee 32306

Florida

Arthur

Department of GeJogy,

and G. Troup

University, ETamilton, Ontario

McMaster

ABSTRACT

Ferromanganese concretions from Grand Lake and Ship IIarbour Lake in Nova Scotia and Mosque Lake in Ontario are most common in water 0.5 to 2 m deep. X-ray diffraction studies show the ferromanganese portions of the concretions to be amorphous. Petrographic and electron probe studies of the ferromanganese material reveal chemical banding of iron and manganese. Bulk chemical analyses indicate that the Fe : Mn ratios of concretions from different sites within a single lake are similar, whereas concretions from different lakes have characteristic Fe : Mn ratios. Trace element concentrations are different in different lakes and are generally several orders of magnitnde less than those of oceanic nodules.

This paper presents the results of a detailed field and laboratory study of freshwater ferromanganese concretions1 from several lakes in eastern Canada. Although oceanic ferromanganese concretions have been investigated extensively, literature on freshwater concretions is scarce. They have been observed in small lakes in Nova Scotia, Ontario, Sweden, Finland, northern England, New York, New Hampshire, Minnesota, Vermont, and Michigan and in three of the Great Lakes-Michigan, Ontario, and Erie. To date there have been no reports of ferromanganese concretion occurrences in freshwater lakes in tropical regions. One of the earliest papers is by Honeyman (1881) who discussed lacustrine hematite concretions collected from Grand Lake, Nova Scotia. The physical features of freshwater ferromanganese precipitates are highly variable, ranging from thin crusts and concretions described by Gorham and Swaine ( 1965) in the English Lake District and by L junggrcn ( 1953 ) in northern Sweden, to

nodular and disk-shaped concretio,ns studied by Kindle (1935) in Nova Scotia lakes. Ljunggren ( 1955b) was the only investigator to consider the mineralogy of freshwater ferromanganese precipitates in detail. X-ray diffraction analyses of various Swedish lake and stream concretions suggested the presence of the minerals goethitc, &Mn02 , and manganous manganite. A few chemical analyses have been reported for ferromangancsc concretions from a variety of geological and hydrological environments. The most noteworthy characteristic of published analyses is the large variation in the iron and manganese concentrations. Gorham and Swaine (1965) compared the composition of freshwater and marine ferromanganesc precipitates and demonstrated that iron generally predominates in freshwater concretions and manganese in marine nodules. This study was designed to obtain detailed information on the geological environment, mineralogy, petrographic charactcristics, and chemistry of ferromanganese concretions from Grand Lake and Ship Harbour Lake (Nova Scotia) and Mosque r Throughout this paper the term "concretion" is Lake ( Ontario). used to refer to the ferromanganese oxide covering This research was supported by the Naplus the enclosed nucleus; the term "ferromantional Research Council of Canada and ganese precipitate" refers only to the ferromangaGeological Survey of Canada. WC thank ncse oxide portions of the concretions. 702

INTRODUCTION

FRESHWATER TABLE

FERROMANGANESE

CONCRETIONS

703

sites

Benthic vcgctation Thickness of zone (cm) Type Algae Aquatic grass Algae Aquatic grass Algae 5 Location of fcrromanganese precipitate* 2.5-cm zone above substrate 3-cm zone above substrate At scdimentwater interface

1.

Genera2 features

of the lakes and concretion

Location Grand Lake 32 km due N of Halifax, N.S. Ship Harbour Lake 51 km N 70" E of IIalifax, N.S. Mosque Lake 96 km NNW of Kingston, Ont. * The substrate

NO. of Length sites (km) 4 14.4

Avg width (km) 1.6

Distance Inlet

from concretion location Outlet 3.2 km S

Underlying bcclrock Carbonifcrous shales and sandstones Cambrian shales and sandstones Grenville gneisscs sites.

Collcction "(2': 0.6-2

12.8 km S

OX-l.5

5

1

19.2

1.6

19.2 km NW

2 km NE

0.6-2

0.5

1 consisted

2.4

1.2

nil

1.2 km S sand at all concretion

of coarse-graincd

D. Church, J. R. Kramer, and G. V. Middleton for comments, K. H. McNutt for assistance with the X-ray fluorcsccnce analyses, and J. Grootcnbocr for assistance in preparation of polished sections for electron pro,be studies. Dr. E. Gorham and Mr. J. Mcro reviewed the manuscript and offered many excellent suggestions, particularly regarding the mechanism of formation of the concretions.

PHYSICAL AND GEOLOGICAL ENVIRONMENT

Concretions were collected from Grand Lake (Lake Shubenacadie) and Ship Harbour Lake (Lake Charlotte) in Nova Scotia and Mosque Lake in Ontario. The location and physical features of these lakes and descriptions of the concretion sites are outlined in Table 1. The fcrromanganese oxide portion of lacustrine concretions occurs only in a narrow zone in these lakes, usually immcdiately above the sediment-water interface. This zone is from 0.5 to 3.0 cm thick in the different lakes and is associated with a zone of abundant benthic vegetation covering the bottom sediments. In all of our concretion sites, the substrate consisted of a coarse-grained sand; we found no concretions above a clay substrate. In both Grand Lake and Mosque Lake, concretions occur at depths >0.6 m in areas immediately adjacent to deep portions of the lake in which a hypolimnion develops

during summer stratification. This suggests that proximity to deep water is an important factor in concretion development. However, in Ship Harbour Lake concretions occur in a partially enclosed bay far from deep portions of the lake. Thermal investigations of Mosque Lake indicate that no underwater springs enter the lake in the vicinity of the concretion site. The investigations were made at a time when such springs should have been most active, so it is unlikely that the concretions result from a discharge of manganese- and iron-rich groundwater into the more oxidizing lake water.

PHYSICAL CIIARACX'ERISTICS OF THE AND INTERNAL

STRUCXJFiE

CONCRETIONS

Grand Lake All Grand Lake concretions have a pcbble or cobble nucleus that may be a fragment of slate, sandstone, or granite; no preference among these has been observed. Rock fragments <2.0-cm diam do not have a ferromanganese oxide crust, and pebbles <4.0-cm diam seldom have more than a ferromanganese stain, Some Grand Lake concretions are illustrated in Fig. 1. The ferromanganese oxide rim forms around the nucleus in a plane parallel to the bottoam sediments. It starts at the scdimcnt-water interface, but attains its maximum width ( up to 3.0 cm) about 1.2 cm above the interface. The thickness of the oxide rims is always <2.5

ROBERT

C. EIARRISS

AND

ARTIIUI~

G. TROUP

Samples 1 and 2 arc nodules from Ship Harbour Lake. Samconcretions. FIG. 1. Ferromanganesc ples 3 and 4 are upper and lower surfaces respectively of typical Ship Harbour Lake concretions. Specimen 5 displays the upper surface of an unusual form of Grand Lake precipitate. Samples 6 and 7 exhibit upper and lower surfaces respectively of typical Grand Lake concretions, Specimen 7 is unusual in having an oxide crust over the entire lower surface; most do not. Samples 8 and 9 represent typical upper and lower surfaces of Mosque Lake concretions.

cm. Occasionally, on smaller nuclei, the oxide covering extends over the entire upper surface of the pebble; this portion of the crust is seldom more than 0.2 cm thick. The upper surfaces of large nuclei, more than 2.5 cm above the sediment-water interface, may be entirely free of any oxide covering or at most have only a ferromanganese stain. We polished cross sections of concretion samples from Grand Lake and studied them by reflected-light microscopy (Zeiss Ultraphot II ) . Concretions consis ted of alternating concentric bands under inSubsequent electron clined white light. probe studies indicate that these correspond to manganese- and iron-rich bands. Under direct monochromatic light, the iron oxide bands appear white and the manganese-rich oxide appears gray. Although many bands are continuous about the concretion rim, most are discontinuous and

irregular, indicating that they -have been broken after precipitation of the b'and. The distance between the centers of any two adjacent bands is extremely variable, ranging from 0.05 to 0.25 mm. We also studied the polished concretion sections with dark-field illumination. The observations indicated that the manganeserich oxide portions of the concretions have a much higher porosity than the iron-rich oxide bands. Ship Harbour Lake

The Ship Harbour Lake -concretions ( Fig. 1) are similar in color and texture to the Grand Lake concretions but arc not saucer-shaped. On many of the concrctions, the ferromanganesc oxide complctcly encloses the pebble nucleus, making the concretion appear nodular or lenticular. Such concretions forming about a fragment of another concretion are also found

FRESHWATER

FERROMANGANESE

CONCRETIONS

705

at this site. The nodules are invariabiy <3-O-cm diam; lenticular concretions without a pebble nucleus are found up to 7.5,cm diam, but seldo'm over 1.5 cm thick. The maximum diameter is invariably attained 0.5 cm above the interface, suggesting that growth occurs most rapidly at or only slightly above the interface. Oxide rims formed about pobblcs and cobbles are between 0.5 and 3.0 cm thick and up to 4.0 cm wide. The maximum thickness is at the sediment-water interf act, giving the concretions a domed appearance rather like the cap of a mushroom. These concretions are quite distinct from those at Grand Lake and may have a somewhat different mode of precipitation. Ship Harbour L'ake concretions are more porous8 than Grand Lake concretions and contain large quantities of diatom tests within the ferromanganese oxide. In many samples certain layers arc so rich in diatomm tests that they appear in hand spccimcn as concentric white bands, conformable with the concentric oxide banding of the concretions. Mosque Lake Mosque Lake concretions consist of a ferromangancse oxide rim 0.2 to 0.5 cm thick and up to 0.5 cm wide about a pobble or cobble nucleus. Nuclei are granitic and sandstone pebbles >2.0-cm diam. Generally both the upper and lower surface of the nuclei are free of oxide. In a few cases cobbles are found exhibiting several o,xide rims oriented at slightly different angles, suggesting that they have been disturbed several tim.es during concretion development; such o'xide rims at an oblique angle to the current position of the sediment-water interface do not extend below the oxide rim currently at the interfact. When oblique rims extend 1 cm or more above the interface, they are broken and discontinuous, indicating that they arc subject to abrasion, Broken cross sections of Mosque Lake concretions show that each ferrolmangancse oxide rim has several individual oxide bands, up to 2 mm thick, which can be

distinguished by color. A detailed study of a num.ber of ferromanganese rims about oae nucleus showed, in all casts, an alternating sequence of manganese oxide and iron oxide bands. Seven of the nine rims were initiated by precipitation of iron oxide.

MINERALOGY

No crystalline structures could bc identified o,ptically in the fcrrosmanganese rims. There were peaks in X-ray diffraction traces from only two Ship Harbour Lake samples-both identified as alpha quartz peaks. Quartz and feldspar grains are the main constituents of the underlying sand in Ship Harbour Lake. The iron and manganese precipitate in our lakes is apparently in an X-ray amorthermal analysis phous state. Differential of Swedish concretions (Ljunggren 1955b) has shown most of the iron and manganese to bc in the form of hydrated amorphous oxides.

CONCRETION CHEMISTRY

Analytical

techniques

An electron probe (MS-64 Acton) was used to investigate the distribution and relationship of various elements within individual ,concretions. Polished cross-sectional surfaces cut from concretions from Grand Lake were studied. The technique of Burns and Fuerstenau (1966) was followed closely; details of the analytical procedure were outlined by Troup ( 1969 ) . All probe intensity-point-countings were repeated 5 times and the maximum variation in number of counts was 1+8.8%. Several concretion samples were analyzcd for the major elements Fe and Mn, and the minor elements Co, Ni, Cu, Zn, H+g, and Pb. A colmplcte chemical analysis was not attempted since concretions contain a large, variable percentage of enclosed sand grains. The major elements, Fe and Mn, were determined by X-ray fluorescence techniques. All trace elements except Hg were detcrmincd by atomic absorption techniques; a mercury vapor detector was used for Hg.

706

ROBERT

C. HARRISS

AND

ARTHUR

G. TROUP

B. S. E.

Fe

FIG, 2. Electron-beam scanning photograph of back-scattered electrons (B.S.E. ), Mn KBI, and Fe KaI radiation from a 300- x 300-p section of a concretion From Grand Lake, site 2. The X-ray intensitypoint-counting path, B-B', is shown.

Internal distribution of elements Back-scatter and X-ray photograpl~s of a sam,ple from Grand Lake (Fig. 2) show manganese concentrations to be inversely related to iron concentrations. Iron is concentratcd in relatively narrow co,ncentric

bands corresponding to arcas of high reflcctivity in back-scatter electron photographs. Detailed profiles of X-ray intensity for a traverse along the line A-A' (Fig. 2) arc in good agreement with the secondary X-

FRESHWATER

FERROMANGANESE

CONCRETIONS

707

on the order of 45% Fe, and the iron-poor portion contains about 2% Fe. Bulk chemical analysis

1100 -

0

900

0 0

700 `;i is 2 = sooa 23 lz!

-

300 -

Chemical analysts for 11 individual and 7 composite samples of Canadian concrctions are .shown in Table 2. The two1 elements Fe and Mn account for 37.2 to 55.8 weight percent of the concretions. If it is assumed that these elements occur as the oxides FeaOs and MnOz, they may account for 57.1 to 85.1 weight percent. Physical examination of the concretions suggests that the variation in major element abundances probably is due to differences in the amount of silicate contaminant present. Thus it is more meaningful to use the Fc : Mn ratio than the ab,solute weight percent of these clcments when comparing diffcrcnt concretions, since this will not be appreciably altered by variations in the percentage of silicate contaminant. Fe : Mn ratio The Fe : Mn ratio varies from 0.43 to 2.56 for samples from the three lakes (Table 2). Probe studies rcvcal that the Fe : Mn ratio within a single concretion may show even greater fluctuations, ranging from 0.04 to as high as 5. The plot in Fig. 3 illustrates the wide variations in the Fc : Mn ratio, even between consecutive irono,r manganese-rich bands. Chemical analyses of 41 concretions from Grand Lake ( all sites ) give an avcragc Fe : Mn ratio of 0.50, while the Fc : Mn ratio for 7 concretions analyzed individually fell within the range of 0.43' to 0.59. Nineteen Ship Harbour Lake concretions have an average Fc : Mn ratio of 0.63. Only three concretions from this lake were analyzed individually, but the Fc : Mn ratio of these ranged from 0.45 to 0.71a significantly greater range. Seventeen Mosque Lake concretions have an average Fe : Mn ratio of 2.5. Unfortunately, only one concretion from this lake was large enough for individual analysis, but its Fc : Mn ratio was close to the mean. There is a significant diffcrcncc in the average Fc : Mn ratio of concretions from

100 -

I

fi 100

I

1

I

1

I

I

300 500 Mn (countsiseo)

700

FIG. 3. A plot of the X-ray intensity-pointcount data for traverse B-B' shown in Fig. 2.

radiation photographs. Two very intense Fe peaks with corresponding low values in the Mn curve arc found in the position of the two iron-rich bands shown in the photographs. The X-ray intensity-point-counting traversc B-B' is shown in Fig. 2; the results are shown in Fig. 3. Between the lowest Fe intensity (54 counts/set) and the highest (1,207) the variation is 22-fold, although there is only a 5-fold variation between the lowest Mn intensity (123 counts/set) and the highest (610). A pyrite standard, known to have an Fe content close to 46.6%, gave an Fe count of 1,219 counts/see. Since the maximum Fe count of the section of concretion examined is very close to that of the standard, this particular band contains

708

TABLE 2. ChemistrtJ

ROBERT

C. I-IARRISS

AND

ARTIIUR

G. TROUP

of concretions

from Grand Lake (GL) and Ship Ha&our tia, and Mosque Lake (ML), Ontario

co

Ni Cu (mm)

Lake

(%I),

Nova

Sco-

Zn

Pb

Fc:Mn

% Fe

% Mn

CL-l-W CL-l-X GL-l-XA GL-1-Y CL-l-YA GL-1-Z GL-l-Cl GL-l-C2 GL-2-Y GL-2-Z GL-B-Cl GL-2-ClB GL-4-Z GL-4-Cl SH-1-x SH-1-Y w-1-z SH-l-ZA SH-l-Cl ML-l-Z ML-l-Cl

( 10) ( 10) (9) (9) (5)

(16) ( 16)

0.433 0.599 0.599 0.552 0.549 0.577 0.530 0.549 0.468 0.434 0.449 0.449 0.470 0.457 0.458 0.709 0.560 0.560 0.636 2.52 2.56

15.5 20.9 20.9 18.6 18.7 19.3 18.1 17.8 16.8 14.6 14.2 14.2 16.3 14.7 11.7 20.0 15.7 15.7 16.9 39.8 40.2

35.9 34.9 34.91 33.7 34.0 33.4 34.2 32.4 35.9 33.6 31.7 31.7 34.4 32.2 25.5 28.3 28.1 28.1 26.5 15.7 15.7

220 202 192 192 192 192 192 202 212 183 192 202 230 230 212 220 220 135

338 243 230 215 215 215 373 272 373 238 238 316 125 149 149 149 107 95

13 12 8 10 10 10 16 16 14 11 10 13 9 8 8 8 6 10

1,940 1,575 1,500 1,575 1,575 1,340 1,850 1,650 1,875 1,575 1,650 1,500 536 500 518 536 467 250

25 24 28 26 26 25 26 28 13 28 29 19 19 19 28 10 28 24

50 34 37 46 46 40 64 40 44 34 50 56 50 67 40 14 21 30

* Concretion samples are numbered to indicate the lake, the site within that lyzed; for example, No. GL-1-W indicates specimen W, collected from site 1 in carried out on four samples to determine analytical precision and are indicated men numbers. Composite samples are indicated by the letter "C" followed by number of concretions cornnosing the sample: for example, No. GL-l-Cl (10) - posed of 10 concretions from site 1 in Grand. Lake.

lake, and the particular specimen anaGrand Lake. Duplicate analyses were by the letter "A" following the specithe composite sample number and the indicates composite sample No. 1 com-

different lakes, suggesting that the bulk of the Fe : Mn ratio for samples within a particular lake is to a large extent determined by the chemistry of the local watershed. Sedimentary rocks underlying the Nova Scotia lakes contain manganese minerals and recent precipitates in the lakes have relatively low Fe : Mn ratios, However, the fluctuating Fe : Mn ratio within individual concretions and the relatively large variations in Fe : Mn ratios in concretions collected from an individual site, such as Ship Harbour Lake, suggest that local variations in pH-Eh and consequently in dissolved iron and manganese concentrations may be affecting the Fe : Mn ratio of the ferromanganese precipitates. Trace elements The Co, Cu, Ni, and Zn contents ferromanganese material from the lakes investigated show distinct ences, although Pb and Hg show of the three diffcrsimilar

ranges in abundance in concretions from all three lakes (Table 2). The 89 samples analyzed from these lakes show no significant correlations between trace element and Fe (or Mn) concentrations. The major and some trace elements in selected freshwater ferromanganese prccipitates from different localities and in manganese nodules from the Pacific Ocean are compared in Table 3. The data on large oxidate crusts from the English Lake District are used for comparison due to their similar morphology to Nova Scotia concretions and minimum silicate contamination as indicated by low concentrations of zirconium, There is considerable overlap in the Fe : Mn ratio of freshwater and oceanic fcrromanganese precipitates, although Fe : Mn ratios greater than 1 are more common in freshwater precipitates (Gorham and Swaine 1965; Manheim 1965). The Zn content of freshwater precipitates is extremely variable, while that of oceanic

FRESHWATER

FERROMANGANESE

CONCRETIONS

709

precipitates

Zn Pb

TABLE 3.

Summary

of data available

on the chemistry co

of ferromanganese

Ni Cu (mm)

% Fe

%Mn

Grand Lake, Avg 41 This study Ship Harbour Avg 19 This study Mosque Lake, Avg 17 This study English Lake Gorham and

N.S. Lake, Ont. District Swainc crusts 1965 N.S.

Nova Scotia concretions Kindle 1932, 1935 New Hampshire Kindle 1935 Swedish bog arcs Ljunggren 1955a Swedish bog ores Manheim 1965 Karaluii-Finnish Manhcim 1965 Pacific nodules Avg Mero 1965 lake ores concretion

Max Min Am Max Min Aw Max Min Avg Max Min Avg Max Min Avg Max Min Avg Max Min Avg Max Min Avg Max Min Avg Max Min As

20.9 14.2 16.6 20.0 11.7 16.7 40.2 22.1 11.1 15.2 15.5 11.6 13.6 26.5 66.0 0 32.5 35.6 22.5 26.6 2.4 14.0

35.9 31.7 33.0 28.3 25.5 26.6 15.7 13.2 8.0 11.1 2.2.8 22.3 22.5 14.5 37.0 0 17.0 4.7 14.0 41.1 8.2 24.2

212 183 196 230 212 221 135 60 20 40

373 215 296 149 107 112 95 40 10 20

16 10 14 9 t

1,940 1,340 1,665 536 467 475

29 13 26 28 10 27 24

64 34 47 67 14 26 30

250 10 20 20,000 3,000 10 11,000

230

40

50

27

10

80

40

40

130 40 23,000 20,000 16,000 140 1,600 280 3,500 9,900 5,300

800 400 470

3,600 200 900

ferromanganese nodules is less so. The amount of Co and Ni is generally 2-10 times higher in C,anadian fcrromangancsc precipitates than in English Lake District crusts, although Cu content is similar. The Canadian freshwater precipitates have Co, Ni, Cu, and Pb contents one and two orders of magnitude below the levels in oceanic nodules (Tab,le 3). The abundance of most trace elements in freshwaters is variable but in unpolluted lakes seldom exceeds oceanic values. For example, the Co, Zn, Ni, and Cu contents of Sierra Lakes (Bradford, Bair, and Hunsaker 1968) and the Columbia River (Silker 1964) are similar to those in scawater. However, the total tract elem,ent content of freshwater ferromanganesc precipitates is generally several orders of magnitude below that of oceanic ferro-

manganese nodules. Gorham and Swaine (1965) and Manheim (1965) have suggested that the much lower growth rates of oceanic ferromanganese concretions result in higher trace element concentrations. Another possible factor in controlling the trace element content of natural ferromangancse precipitates is the effect of pII on the surface chemical properties o,f hydrated iron-manganese oxides. Natural iron hydroxides commonly have a minimum surface charge in the pH range of &8 (Parks 1967). Data on the zero point of charge of natural manganese oxides do not seem to be available. Measurements on synthetic samples indicate a minimum surface charge for manganese oxides in the pH range of 3-7. The pH of freshwater lakes commonly varies over the range of about 5-7 and that of seawater is 8.1. Ferroman-

710

ROBERT

C. IIARRISS

AND

ARTHUR

G. TROUP

ganese precipitates at the lower pH values characteristic of lake water would thus be closer to the zero point of charge and have a lower capacity for adsorbing trace elements. Jenne (1968) has pointed out that pH is of major importance in controlling both the rate and capacity of trace element adsorption by fcrromangancsc oxides.

DISCUSSION

Two mechanisms of lacustrine ferromanganese concretion formation have been proposed. The theory discussed by Kindle ( 1935) attributed concretion formation to the action of iron- and manganesc-precipitating bacteria; iron-precipitating bacteria have been found in Nova Scotia lakes ( Gorham 1957). The second theory was proposed by Manhcim ( 1965), who suggested that iron and manganese diffuse through the interstitial water of bottom sediments into the oxidizing aquatic environment where they are prccipitatcd around suitable nuclei at the sedimentwater interface. Neither of these can adequately explain the chemical banding revealed by electron probe studies or the restricted environmental conditions under which the concretions investigated occur. There are two primary sources of iron and manganese to the epilimnion: runoff from the watershed in the form of surface drainage and shallow groundwater discharge and diffusion from a reducing hypolimnion, Fractionation of iron and manganese in either environment depends on the presence of an oxidation-reduction cycle. Stratified lakes with a reducing hypolimnion generally have high concentrations of dissolved iron and manganese, which precipitate as hydrated oxides as they reach the oxidizing epilimnion. Fractionation of iron and manganese takes place in lakes bccausc of differential oxidation kinetics. Iron is rapidly oxidized and settles back into the hypolimnion to bc recycled, while manganese oxide particles form higher in the water column and can be dispcrscd through the epilimnion (Tanaka 1965). environIn temperate and subarctic

mcnts, shallow groundwater systems may become deoxygenated during winter ice conditions. In spring, large quantities of shallow groundwater enriched in dissolved iron are discharged into the lake where the iron rapidly oxidizes. If iron and manganese for concretion formation are supplied from the groundwater and from a reducing hypolimnetic environment, the manganese-rich bands would form during summer and possibly winter stratification and the iron-rich bands would form during spring groundwater discharge into the lake and at times when the lake stratification breaks down. Dr. E. Gorham (personal communication) has pointed out that in some lakes where concretions arc common a reducing hypolimnion is not present at any time of the year, while lakes with a highly reducing hypolimnion rich in dissolved iron and manganese may not have concretions in the epilimnion. He suggests the following alternativc hypothesis for the origin of the concretions :

During the period when the vegetation is inactive, the oxidizing potential of the surrounding aquatic environment will be lower than when the vegetation is photosynthesizing actively, and releasing large amounts of oxygen at an elevated pH. Moreover, it is much well established . . . that Fe precipitates more rapidly than Mn in oxidizing environments. (The soluble Fe and Mn may perhaps enter the lake largely in combination with protective organic colloids or as chelates.) Knowing these facts, one might expect that while soluble Fe will precipitate rapidly in the epilimnion during all seasons, soluble Mn will precipitate rapidly only when plants are most active in producing oxygen, i.e., in the summer. Probably it will only precipitate rapidly in close proximity to, the photosynthesizing plants, and the elevated pH presumably promotes oxidative precipitation. If the above hypothesis is correct, then banding results from the differential sensitivity of Fe and Mn to oxidation, with Mn precipitating in abundance only when photosynthesis brings about exceptionally oxidizing conditions in summer. Perhaps Fc will also precipitate more rapidly then, and Mn may precipitate slowly throughout the year; but all that is needed is a differential sensitivity to oxidation to bring about handing, since the data show that (Fig. 1) both Mn and

FRESHWATER

FERROMANGANESE

CONCRETIONS

711

Fe are present throughout the concretions, although in greatly differing abundance (i.e., the Fe/Mn ratio never reaches either infinity or zero ) .

Rate of concretion form&on Kindle ( 19:35), who found concretions in the shallow waters of an artificial lake, concluded that many of the manganiferous crusts reprcscntcd a growth period of no more than 25 years. Ljunggren (1953, p'. 288 quoting Naumann) stated that orescraping for economic use could take place about once every 51) years in certain Swedish lakes ( Gorham and Swaine 1965). If the chemical banding WC found is connected with seasonal variations in water chemistry, and if we assume that the growth of one season can produce only two oxide bands (one manganese-rich and one iron-rich) we can then calculate a hypothctical growth rate for the precipitates. In Grand Lake samples the distance between the centers of two adjacent similar bands averages 0.1 mm, and in Mosque Lake concretions this distance averages 1.5 mm. A Grand Lake concretion having an oxide rim 1 cm wide would then form in a period of 100 years, whereas in Mosque Lake such a concretion could form in only 7 years. Although the growth rate for Grand Lake concretions is quite low, both values are compatible with Kindle's estimate of 25 years. This suggests that the initial assumption is valid; if so, it provides a method for determining concretion age and growth rate.

SUMMARY AND CONCLUSIONS

The results of our investigation can bc summarized as follows. 1. The ferromanganese portion of the lacustrine concretions studied occurs only in a narrow zone immediately above the sediment-water interface. On concretions that have been disturbed, that portion of the ferromanganese rim which was buried beneath the interface has been rcmovcd by solution and that portion raised above its original zone of occurrence shows evidence of being rcmovcd by abrasion. It is

apparent that active precipitation of the ferromangancse portion of the concretions occurs only within a narrow zone immediately above the sediment-water interface. 2. In all the concretion sites we visited, the thickness of the zone of concretion fo,rmation corresponds closely with the thickness of a zone of abundant benthic vcgctation. This suggests that the vcgetation is responsible for the establishment of a microenvironment in which oxidation and precipitation o,f iron and manganese may occur because of a locally higher pH and Beals (1966) has oxygen concentration. described concretions collected from a depth of 5 m in Ship Harbour Lake as nodular or bun-shaped, suggesting that in decpcr water concretion formation results from precipitation of manganese and possibly iron directly from lake water supersaturated with respect to these elements. 3. All of these concretions contain a nucleus, which in mo,st cases is a pebble or small stone but may be a broken fragment of ano,ther concretion. This nucleus is required to support the fcrromanganese precipitate above the underlying, reducing substrate whcrc re-solution would rapidly take place. 4. X-ray diffraction studies reveal that the ferro,mangancse portions of the concretions are amorphous. Recent investigations show that marine nodules contain only weakly developed ferromanganese minerals. The amorphous nature of lacustrine concretions may be a result of a much greater rate of growth. 5. Electron probe and microscope studies show the concretions to display a concentric chemical banding across which the Fc : Mn ratio may vary up to two orders This suggests that concrcof magnitude. tion formation is the result of a cyclic process, possibly involving a fluctuation in the Fc : Mn ratio of the overlying lake watcr. Gorham and Swaine ( 1965) also noted layered structure in concretions from the English Lake District; concretion growth by the addition of iron- and manganeserich bands may be a common phenomenon. 6. Bulk chemical analyses show large

712

ROBERT

C. HARRISS

AND

ARTHUR

G. TROUP

variations in the Fe : Mn ratio of concretions from similar depth environments in different lakes, indicating that the chcmistry of the local watershed has a major influence on concretion chemistry.

REFERENCES BEALS, H. L. 1966. Manganese-iron concretions in Nova Scotia lakes. Mar. Sediments 2: 70-72. BRADFON), G. R., F. BAKR, AND V. HUNSAICER. 1968. Trace and major element content of 170 High Sierra lakes in California. Limnol. Oceanogr. 13 : 526-529. BURNS, R. G., AND D. W. FLJERSTENAU. 1966. Electron probe determination of intcrelement relationships in manganese nodules. Amer. Mineral. 51: 895-902. GOMM, E. 1957. The chemical composition of lake waters in Halifax County, Nova Scotia. Limnol. Oceanogr. 2: 12-2,l. AND D. J. SWAINE. 1965. The influ-> ence of oxidizing and reducing conditions upon the distribution of some elements in lake sediments. Limnol. Oceanogr. 10: 268279. HONEYMAN, D. 1881. Nova Scotia geology (superficial). Proc. Trans. N.S. Inst. Natur. Sci. 5: 328. JUNE, E. A. 1968. Controls on Mn, Fe, Co, Ni, Cu, and Zn concentrations in soils and waters: The dominant role of hydrous Mn and Fe oxides. Advan. Chem. Ser. 73, p. 337-387.

KINDLE, E. M. 1932. Lacustrine concretions of manganese. Amer. J. Sci. 24: 496-504. -. 1935. Manganese concretions in Nova Scotia lakes. Trans. Roy. Sot. Can. Sect. 4 29 : 163-180. L JUNGGREN, P. 1953. Some data concerning the formation of manganiferous and ferriferous bog ores. Geol. Foren. Stockholm Forh. 75: 277-298. -. 1955a. Geochemistry and radioactivity of some Mn and Fe bog ores. Geol. Foren. Stockhohn Forh. 77: 33-44. -. 1955b. Differential thermal analysis and X-ray examination of Fe and Mn bog ores. Geol. Foren. Stockholm Forh. 77: 135-147. MANHIXM, F. T. 1965. Manganese-iron accumulations in the shallow marine environment, p. 217-276. In Mar. Geochem., Proc. Symp., 1964 ( 1965 ) . MERO, J. L. 1965. The mineral resources of the sea. Elsevier. 312 p. PARKS, G. 1967. Aqueous surface chemistry of oxides and complex oxide minerals. Isoelcctronic point and zero point of charge. Advan. Chem. Ser. 67, p. 121-160. SILKER, W. 1964. Variations in elemental concentrations in the Columbia River. Limnol. Oceanogr. 9: 540-545. TANAKA, M. 1965. Formal discussions [of paper presented by J. J. Morgan and W. Stumm], p. 118-123. In 0. Jaag [ed.], Advances in water pollution research, v. 1. Pergamon. TROUP, A. 1969. Geochemical investigations on lacustrine ferromanganese concretions. M.S. thesis, McMaster cniv., Hamilton, Ontario. 82 p.

Information

11 pages

Report File (DMCA)

Our content is added by our users. We aim to remove reported files within 1 working day. Please use this link to notify us:

Report this file as copyright or inappropriate

1231857


You might also be interested in

BETA