Soil Science Society of America Journal 63:923-929 (1999)
© 1999 Soil Science Society of America
DIVISION S-5-PEDOLOGY
Role of Free Silicon, Aluminum, and Iron in Fragipan Formation
M.M. Duncana and
D.P. Franzmeierb
a Crop Production Services, P.O. Box 43, Ferris, IL 62336-0043 USA
b Dep. of Agronomy, Purdue Univ., W. Lafayette, IN 47906-1150 USA
dfranzme{at}purdue.edu
 |
ABSTRACT
|
|---|
Fragipans form by various physical and chemical processes. The objective of this study was to elucidate the role of free Si, Al, and Fe in weak cementation of fragipan soils. We studied a toposequence of soils in which natural drainage ranged from poor to good on a dissected pre-Wisconsinan plain covered with loess in southern Indiana. Soils were sampled in 10 to 15 cm layers. In the laboratory, strength was determined by a rupture test on disturbed samples, and Si, Al, and Fe were determined on citrate–bicarbonate–dithionite (CBD) extracts (with d subscripts). Strength reached a profile maximum in the fragipan. With molar Sid/(Sid + Ald) ratio of 
0.5 there was no correlation between this ratio and strength, but in horizons in which the ratio was >0.5, mainly in fragipans, the correlation was strong. Strength was also correlated with exchangeable Mg. In fragipans the molar Ald/(Ald + Fed) ratio was 
0.20, and above the fragipan the ratio was greater. We propose that in the fragipan most of the Ald is derived from Al substituted for Fe in Fe oxides, but in horizons above the fragipan there is more Al than the Fe oxides can accommodate. The extra Ald combines with free Si in these horizons to form aluminosilicates. In the fragipan, however, most free Al is tied up in the Fe oxides and cannot react with Si, so the free Si that is translocated from upper horizons bonds to clay minerals, and especially to Fe oxides, to cause cementation.
Abbreviations: CBD, citrate–bicarbonate–dithionite
|
INTRODUCTION
|
|---|
SEVERAL THEORIES have been proposed to explain the formation of fragipans (Smalley and Davin, 1982; Smeck and Ciolkosz, 1989; Franzmeier et al., 1989). Some entail relict processes that were mainly active soon after deposition of parent materials, such as in periglacial activity, desiccation, and initial consolidation of loess. Others involve active processes that continue today, such as clay bridging and chemical cementation. A combination of processes may be invoked to explain the development of many fragipan soils. Some theories explain why fragipans form in certain soils, but they do not explain why fragipans do not form in nearby soils. For example, in southern Indiana and other parts of the north central region of the USA, soils formed in moderately deep loess have fragipans, but those formed nearby in deep loess lack them. If relict processes were primarily important, one would expect fragipans to form under both conditions.
We believe that a primary process that is responsible for the strength of fragipan horizons is bonding by Si-rich material. Briefly, soil solutions accumulate in lower subsoil horizons in the winter and spring. Then, in the summer and early fall, tree roots extract water from this zone and concentrate silica compounds that serve as binding materials (Franzmeier et al., 1989). Soil solutions are held up mainly by slowly permeable horizons below those in which fragipans form. They can also be held up by particle-size discontinuities (Smeck et al., 1989), or the moisture gradient may be due to the depth of penetration of precipitation.
We also showed that extractable Si was positively correlated with soil strength, extractable Al was negatively correlated with strength, and the Si/Al ratio was more strongly correlated with strength in soils with high-strength horizons other than fragipans. They included soils with hard-setting horizons in Australia (Franzmeier et al., 1996) and soils formed in dense till in Indiana (McBurnett and Franzmeier, 1997). Norfleet and Karathanasis (1996) related soil strength to Si/(Al + Si) in fragipan soils of Kentucky. Thus, there is evidence for cementation by silica in several kinds of soils.
The above findings demonstrate the relationship of Si and Al to soil strength, but Fe may also be important. The objective of this study was to determine the relationship of extractable or free Si, Al, and Fe to soil strength in a toposequence of soils with fragipans in southern Indiana.
|
Site and soils studied
|
|---|
The study was conducted on Chestnut Ridge in the Muscatatuck National Wildlife Refuge, 6 km east of Seymour, IN. Jenkinson (1998) studied soil water relationships and oxidation–reduction patterns on the same toposequence. It was thought that this region is a till plain of Illinoian age that was covered by loess during late Wisconsinan time. Gamble et al. 1990, however, suggested that the older deposit is older than Illinoian, much of it is outwash rather than till, and the landform represents an erosion surface rather than a deposition surface as implied by the name till plain. These investigators suggested that the buried surface may be an Iowan erosion surface. The physiographic region, from a soils perspective, may best be called a loess plain because it is a plain and most of the soil on it formed in loess.
The soils on the Refuge formed in Peorian loess, deposited about 20000 years ago, over Sangamon paleosols in pre-Wisconsinan glacial drift. The loess is 2 m thick on stable surfaces, and is divided into two materials (Steinhardt and Franzmeier, 1979). Parent Material 1, silty loess, contains 9 to 18% sand, and overlies Parent Material 2, gritty loess, which contains 20 to 29% sand. The gritty loess could be a separate deposit, it could be due to the mixing in of sand as the loess was deposited, or it could be a result of mixing with the underlying material by fauna after the loess was deposited (Gamble et al., 1990). Parent Material 3 is the paleosol (3Btb horizon).
Soil horizons A, E, EB or BE, and Bt formed in Parent Material 1. A fragipan horizon formed mainly in Parent Material 2, but in some soils, part of it is in Parent Material 1. Within the fragipan, clay content increases downward, so the upper part of the fragipan is designated a Bx horizon, and the lower part, a Btx horizon. The Btx horizons also have clay films on prism surfaces. Both the fragipan and paleosol restrict downward water movement. Jenkinson (1998) identified three piezometric surfaces in these soils. The upper one is held up by the fragipan, the middle one by the paleosol, and the lowest one represents the regional water table.
Four soil members of a toposequence were sampled. The soil names used here are from the published soil survey of the local county (Nickell, 1976). Some soil series names, rules of classification, and their interpretations have changed since the survey was published. The Clermont soil (poorly drained) is on the interior of the loess plain, Avonburg (somewhat poorly drained) is on the edge of the plain, Rossmoyne (moderately well drained) is on the upper shoulder of the loess plain bevel, and Cincinnati (well drained) is on the upper backslope of the bevel (Fig. 1) . Classification of the pedons sampled and other reference information is in Table 1
. Clermont, Avonburg, and Rossmoyne soils have fragipans, but their upper boundary is below 1 m, so the fragipan is recognized at the subgroup level instead of the great group level. Cincinnati has a B horizon with some brittle characteristics (Btx), but the prisms are not large enough nor well enough developed to qualify as a fragipan, so the fragic character is also recognized at the subgroup level.
View larger version (45K):
[in this window]
[in a new window]
|
Fig. 1 Block diagram of the area in which the toposequence was sampled. Cr = Clermont, Av = Avonburg, Rs = Rossmoyne, Cn = Cincinnati (Jenkinson, 1998). Easting and northing numbers are Universal Transverse Mercator coordinates. Add 604000 m to easting numbers in graph and 4311000 m to northing numbers to give actual coordinates
|
|
Mean annual temperature in Seymour is 22.5°C in summer and -1.1°C in winter. Mean annual precipitation is 1080 mm. The area is in original forest vegetation. The dominant tree species are American beech (Fagus grandifolia Ehrh.), sugar maple (Acer saccharum Marsh.), yellow poplar (Liriodendron tulipfera L.), and northern red oak (Quercus rubra L.). There is no evidence that the site has ever been plowed, but it was probably selectively logged before 1930. It may have been grazed before 1965, but it has been protected since then.
|
Methods
|
|---|
The soils were sampled by horizons from pits in 1992 at the initiation of Jenkinson's (1998) soil water study. Clay content (pipette method), pH (1:1 H2O; 1:2, 0.01 M CaCl2 suspensions), exchangeable cations (NH4OAc extraction), and bulk density (Saran-coated clods) were determined by the National Soil Survey Laboratory (Soil Survey Staff, 1991) on samples collected from pits.
We resampled the soils in 1996 by collecting samples with an auger within 1 to 2 m of where the pit samples were collected. Strength and extraction determinations were done in the Purdue labs on samples air dried and passed through a 2-mm sieve. Soil strength was determined by a procedure proposed by Cochrane and Aylmore (1992). Twelve holes 20 mm in diameter were bored in a 10-mm-thick 100 x 150 mm Delron plastic plate. The plate was placed on Kimwipe tissue (Kimberly-Clark, Neenah, WI) supported by stainless steel screen. A soil sample was poured into each hole, and the plate was lightly tapped to settle the sample into the holes, but the samples were not tamped. The samples were wet from the bottom, drained, and dried first at room temperature and then in a 40°C oven. Strength was determined on the resulting discs, removed from the plate, by slowly lowering the flat end of a 6.4-mm diam. steel rod into the disc and recording the force in kg at which it shattered. They all failed abruptly; i.e., all were brittle. Two replicate plates of 12 discs each were used for each soil sample.
Extraction with CBD done was according to Jackson (1969). Extracts were analyzed for Fe and Al by atomic absorption, and for Si by Weaver et al. (1968), as adapted by Franzmeier et al. (1977). Results are represented by subscript d (Fed).
|
Results and discussion
|
|---|
Clay, pH, and Exchangeable Ions
The pH-depth curves are similar for all four soils (Fig. 2)
. Water pH is mainly around 4.2 to 4.6 in E and Bt horizons, which increases to 5 in the fragipans, below 100 cm in depth. The pH in CaCl2 is about 0.5 to 1.0 units less than in water. Bulk density of moist (0.03 MPa) samples averaged 1.57 Mg m-3 in the B horizons above fragipans and 1.73 Mg m-3 in fragipan horizons.
In the Clermont, Avonburg, and Rossmoyne soils, the clay content reaches a maximum in the Bt horizon and decreases somewhat in the fragipan (Fig. 3 and 4)
. In the Cincinnati soil, the loess is thinner, and the clay content reaches a plateau in the 2Btx1 and 2Btx2 horizons, and then it increases somewhat in the paleosol. The convergence of the 2Btx2 and 3Btb1 horizons represent welded horizons (Ruhe and Olson, 1980).
View larger version (38K):
[in this window]
[in a new window]
|
Fig. 3 Depth functions of contents of clay, exchangeable Ca and Mg, Fed, Ald, Sid, strength, molar Feo/Fed ratio, and molar Sid/(Sid + Ald) ratio in the Clermont and Avonburg soils. xxx indicates the upper and lower (if above 250 cm) boundaries of fragipans
|
|
View larger version (38K):
[in this window]
[in a new window]
|
Fig. 4 Depth functions of contents of clay, exchangeable Ca and Mg, Fed, Ald, Sid, strength, molar Feo/Fed ratio, and molar Sid/(Sid + Ald) ratio in the Rossmoyne and Cincinnati soils. xxx indicates the upper and lower boundaries of the fragipan in Rossmoyne, and xx, the upper and lower boundaries of a fragic horizon in Cincinnati
|
|
The cation-exchange sites are dominated by Ca2+, Mg2+, and H+ (or Al ions). The surface horizons are rich in base cations, probably as a result of cycling by the forest vegetation. In the acid upper subsoil horizons, the contents of Ca2+ and Mg2+ are relatively low, and H+ is high. In the upper horizons, exchangeable Ca is more abundant than exchangeable Mg, but this relationship is reversed in lower Bt horizons, and in Bx horizons Mg usually dominates over Ca (Fig. 3 and 4).
Extractable Iron, Aluminum, and Silicon
Citrate–bicarbonate–dithionite solution extracts practically all secondary Fe oxides and organic complexes in soils (Schwertmann and Taylor, 1989). Citrate–bicarbonate–dithionite also extracts Al from non-crystalline Al hydrous oxides and organic complexes, and Al and Si from allophane-like material (Wada, 1989). In most Fe-oxide minerals, some Al substitutes for Fe in the crystal structure. In synthetic goethites the maximum Al-for-Fe substitution involves 1/3 of the Fe sites (Schwertmann and Taylor, 1989; Cornell and Schwertmann, 1996); in hematite, 1/6 of the sites, and in lepidocrocite there is no Al substitution. In other Fe-oxide minerals substitution is possible, but the maximum amounts are not known. In natural goethites maximum Al substitution is less than in synthetic ones. Cornell and Schwertmann (1996) suggested that it is 0.07 to 0.15 Al/(Al + Fe) (mol/mol) in Alfisols, Inceptisols, and Pseudogley soils of temperate climates. Roth et al. (1969) found a fairly constant 0.10 Al/Fe weight ratio, or 0.17 Al/(Al + Fe) (mol/mol) in several vermiculite and saprolite samples, which could represent Al substitution in Fe-oxide minerals. From these studies it appears that the maximum Al substitution in natural samples is about 0.15 to 0.17 mol/mol.
Curves for Fed tend to follow the clay curves in all four soils (Fig. 3 and 4), and the two are significantly correlated in all but the Rossmoyne soil (Table 2)
. The association of clay and Fe is sometimes represented as a coating of Fe oxides on a silicate mineral, but electron micrographs often show small equidimensional crystals of Fe oxides scattered around on the relatively large flat surfaces of a clay mineral.
View this table:
[in this window]
[in a new window]
|
Table 2 Correlation coefficients showing relationships among properties of soil horizons formed in loess (Parent Materials 1 and 2)
|
|
The contents of Ald and Sid are about one-tenth that of Fed. In upper horizons Ald content is greater than Sid, but the two curves cross in the Bt horizons, and Sid contents remain high throughout the lower Bt horizons and fragipans (Fig. 3 and 4). Sid is significantly correlated with clay in all soils, but Ald is not (Table 2).
We noted that there tended to be more Ald in upper horizons than in lower ones and represented the relationship by plotting moles of Ald vs. moles of Ald plus moles of Fed (Fig. 5)
for the fragipan horizons, and for all horizons above the fragipan except A horizons, which may have a significant content of Al–organic complexes. The slope of the lines in this graph represents the Ald/(Ald + Fed) molar ratio. The two populations were fairly distinctive. The two regression lines had similar slopes of 0.20 and 0.22, but the line representing the upper horizons had a greater intercept. If we accept previous work that soil Fe-oxide minerals have around 0.15 to 0.17 (molar) Al substitution, i.e.,
, then 75% of the Ald in the fragipan horizons could be in the Fe oxides.
View larger version (22K):
[in this window]
[in a new window]
|
Fig. 5 Relation of Ald and Ald + Fed for fragipan horizons (lower line) and subsoil horizons above the fragipan (upper line) for the three soils with fragipans
|
|
Soil strength generally increases with depth (Fig. 3 and 4). In horizons formed in loess, strength reaches a maximum in the fragipan. Below the pan, strength may increase because of increased clay in the paleosol, as shown by the Cincinnati soil (Fig. 4). Strength is positively correlated with Sid in all soils (significantly correlated in the Clermont and Cincinnati soil, Table 2). Strength is negatively correlated with Ald in three soils (significantly correlated in the Avonburg and Rossmoyne soils). Combining those two trends, strength is positively correlated with Sid/(Sid + Ald) (0.01 level) in all four soils. From this relationship it appears that Sid contributes to strength, and Ald counteracts strength. We examined this relationship in more detail. In the three soils with fragipans, with molar Sid/(Sid + Ald) ratio 0.5 there is no relationship between strength and this ratio, but when the ratio exceeds 0.5, the ratio is significantly correlated with strength (Fig. 6)
.
In the three soils with fragipans, strength is also significantly correlated with exchangeable Mg (Table 2), which has been linked to fragipan formation (Hutcheson et al., 1959).
|
Conclusions
|
|---|
These relations lead us to the following conclusions. The three soils with fragipans are saturated and reduced during the winter and spring but are oxidized most of the summer and fall (Jenkinson, 1998). These conditions provide an optimum environment for incorporation of Al into Fe oxides. Every year some Fe-oxide minerals dissolve in the spring and reprecipitate in the fall. When they precipitate, they may incorporate any available free Al into their structures. Horizons above the fragipan have a higher Ald/(Ald + Fed) ratio than fragipan horizons (Fig. 5). We propose that in fragipans, most of the Al is in the Fe-oxide structures, but in upper horizons there is more Al than the Fe oxides can hold, and these Al oxides are in some free form. Weathering of silicate minerals is intense in upper, acid horizons in which soil acidity increases during oxidation cycles (Ransom and Smeck, 1986). The weathering minerals are mainly 2:1 clay minerals and feldspars, so more Si is released than Al. Silica is leached downward, and probably laterally (Jenkinson, 1998). In Bt and Bx horizons it combines with free Al oxides to form amorphous or weakly crystalline aluminosilicates, such as allophane, and clay minerals. This is the origin of part of the clay maximum in Bt horizons; some clay is also translocated to them. Some silica leaches down to the Bx and Btx horizons where most of the Al is tied up in the Fe oxides and little remains to combine with silica. There the Si/(Si + Al) ratio exceeds 0.50, and the extra silica (in addition to that combined with Al) is responsible for greater strength (Fig. 6). Silica may be adsorbed on the surfaces of aluminosilicates or Fe-oxide crystals that lie on the aluminosilicate clays. With continued additions of silica, silica polymers could form and eventually bridge to other Fe oxides or silicate minerals (Chadwick et al., 1987). This bridging may be responsible for the strength and brittle nature of fragipans. When pressure is applied to a piece of the soil, it fails suddenly when the bonds are broken. Alternatively, weak cementation may be due to an aluminosilicate material (Norfleet and Karathanasis, 1996). Exchangeable Mg may also contribute to fragipan formation, possibly through its tendency to disperse soil materials.
|
NOTES
|
|---|
Purdue Agric. Res. Programs journal no. 15812.
Received for publication August 19, 1998.
|
REFERENCES
|
|---|
- Chadwick O.A., Hendricks D.M., Nettleton W.D. Silica in duric soils: I. A depositional model. Soil Sci. Soc. Am. J. 1987;51:975-982.[Abstract/Free Full Text]
- Cochrane, H.R., and L.A.G. Aylmore. 1992. Measuring soil structural stability on the farm. Abstr. of Poster Presentations, Australian Soc. of Soil Sci., National Soils Conference, Adelaide.
- Cornell R.M., Schwertmann U. The iron oxides. Germany and New York: VCH. Weinheim, 1996.
- Franzmeier D.P., Chartres C.J., Wood J.T. Hardsetting soils in southeast Australia: Landscape and profile processes. Soil Sci. Soc. Am. J. 1996;60:1178-1187.[Abstract/Free Full Text]
- Franzmeier D.P., Norton L.D., Steinhardt G.C. Fragipan formation in loess of the Midwestern United States. In: Smeck N.E., Ciolkosz E.J., eds. Fragipans: Their occurrence, classification, and genesis. Soil Sci. Soc. Am. Spec. Publ. 24. Madison, WI: SSSA, 1989:69-98.
- Franzmeier D.P., Steinhardt G.C., Crum J.R., Norton L.D. Soil characterization in Indiana: I. Field and laboratory procedures. Res. Bull. 943. West Lafayette, IN: Purdue Univ. Agric. Exp. Stn, 1977.
- Gamble E.E., Mausbach M.J., Sinclair H.R., Franzmeier D.P. A reevaluation of the "Illinoian Till Plain" and the origin of the gritty loess substratum of the Avonburg and Clermont soils in Southeastern Indiana. SSIR 41. Lincoln, NE: USDA-SCS, Natl. Soil Surv. Center, 1990.
- Hutcheson T.B., Lewis R.J., Seay W.A. Chemical and clay mineralogical properties of certain Memphis catena soils of Western Kentucky. Soil Sci. Soc. Am. Proc. 1959;23:474-478.
- Jackson, M.L. 1969. Soil chemical analysis: Advanced course. 2nd. ed. Dep. of Soil Sci., Univ. of Wisconsin, Madison, WI (published by the author).
- Jenkinson, B.J. 1998. Wet soil monitoring project on two till plains in south and west central Indiana. M.S. Thesis. Purdue University, West Lafayette, IN.
- McBurnett S.L., Franzmeier D.P. Pedogenesis and cementation in calcareous till in Indiana. Soil Sci. Soc. Am. J. 1997;61:1098-1104.[Abstract/Free Full Text]
- Nickell A.K. Soil Survey of Jennings County Indiana. Washington, DC: USDA-SCS in cooperation with the Purdue Agricultural Experiment Station. U.S. Gov. Print Office, 1976.
- Norfleet M.L., Karathanasis A.D. Some physical and chemical factors contributing to fragipan strength in Kentucky soils. Geoderma. 1996;71:289-301.
- Ransom M.D., Smeck N.E. Water table characteristics and water chemistry of seasonally wet soils of southwestern Ohio. Soil Sci. Soc. Am. J. 1986;50:1281-1289.[Abstract/Free Full Text]
- Roth C.B., Jackson M.L., Syers J.K. Deferration effect on structural ferrous–ferric iron ratio and CEC of vermiculites and soils. Clays and Clay Minerals 1969;17:253-264.
- Ruhe R.V., Olson C.G. Soil welding. Soil Sci. 1980;130:132-139.
- Schwertmann U., Taylor R.M. Iron oxides. In: Dixon J.B., Weed S.B., eds. Minerals in soil environments. 2nd ed. SSSA. Book Ser. 1. Madison, WI: SSSA, 1989:379-438.
- Smalley, I.J., and J.E. Davin. 1982. Fragipan horizons in soils: A bibliographic study and review of some of the hard layers in loess and other materials. New Zealand Soil Bureau Bibliographic Rep. 30. Dept. of Sci. and Industrial Res. New Zealand.
- Smeck, N.E., and E.J. Ciolkosz (ed.) 1989. Fragipans: Their occurrence, classification, and genesis. SSSA Spec. Publ. 24. SSSA, Madison, WI.
- Smeck N.E., Thompson M.L., Norton L.D., Shipatalo M.J. Weathering discontinuities: A key to fragipan formation. In: Smeck N.E., Ciolkosz E.J., eds. Fragipans: Their occurrence, classification, and genesis. SSSA Spec. Publ. 24. Madison, WI: SSSA, 1989:99-112.
- Soil Survey Staff. Soil survey laboratory methods manual. SSIR 42, version 1.0. Lincoln, NE: Natl. Soil Surv. Center, 1991.
- Soil Survey Staff. Keys to soil taxonomy, 7th ed Washington, DC: USDA-NRCS. U.S. Gov. Print. Off, 1996.
- Steinhardt G.C., Franzmeier D.P. Chemical and mineralogical properties of the fragipans in the Cincinnati catena. Soil Sci. Soc. Am. J. 1979;43:1008-1013.[Abstract/Free Full Text]
- Wada K. Allophane and imogolite. In: Dixon J.B., Weed S.B., eds. Minerals in soil environments, 2nd ed Madison, WI: SSSA Book Ser. 1. SSSA, 1989:1051-1088.
- Weaver R.M., Syers J.K., Jackson M.L. Determination of silica in citrate–dithionite–bicarbonate extracts of soils. Soil Sci. Soc. Am. Proc. 1968;32:497-501.
TOP
ABSTRACT
INTRODUCTION
