Allophane and imogolite in Swedish

soils

or why small, previously unknown, fibres influence the water quality in forests

Jon Petter Gustafsson Erik Karltun

Prosun Bhattacharya

Stockholm 1998 Research Report TRITA-AMI 3046 Division of Land and Water Resources Department of Civil and Environmental Engineering Royal Institute of Technology (KTH) Stockholm, Sweden

1

Table of contents

Preface 3

1. Review - What are allophane and imogolite? 4

2. Occurrence of allophane and imogolite in Sweden 10

3. The dissolution chemistry of allophane / imogolite 15

4. What mechanisms are responsible for the aluminium release? 18

5. Surface properties of synthetic imogolite 21

6. Possible engineering applications 24

7. Implications and suggestions for future research 26

8. References 27

Appendix: X-ray spectra 30

Infrared spectra 31

Svensk sammanfattning av forskningsprojektet 33

2

Preface

This report presents the work carried out within the research project ”Non-crystalline

aluminosilicates in Swedish forest soils”, sponsored by the Swedish Environmental Protection

Agency from 1995 to 1997.

We wish to express our thanks to the following people, who have all made essential

contributions in various parts of the project: Dr. David Lumsdon, Dr. Tony Fraser, Dr. Bill

McHardy and Dr. Derek Bain (all at the Macaulay Land Use Research Institute, Aberdeen,

UK), MSc Magnus Simonsson and Dr. Dan Berggren (at the Department of Soil Sciences,

SLU). We also acknowledge the laboratory assistance by the following people: Donald

Duthie, Britta Hedengran, Christina Stålhandske, Gunilla Hallberg and Kenth Andersson.

Graduate students from the Aquatic and Environmental Engineering programme of the

Uppsala University made valuable contributions. The senior author would also like to thank

his grandfather Yngve Gustafsson for illuminating discussions concerning the long history of

the theories on Podzol formation.

Apart from the Swedish Environmental Protection Agency, financial support has been

gratefully received from KTH, SLU and from the Royal Academy of Sciences.

As an outcome of the project three research papers have been, or are being, produced.

These three are not included in this report, but referred to by their roman numerals:

I. Gustafsson, J.P., Bhattacharya, P. and Karltun, E. 1998. Mineralogy of poorly crystalline aluminium phases

in the B horizon of Podzols in southern Sweden. Applied Geochemistry, accepted manuscript. II. Gustafsson, J.P., Lumsdon, D.G. and Simonsson, M. 1998. Aluminium solubility characteristics of spodic B

horizons containing imogolite-type materials. Clay Minerals 33, 77-85. III. Gustafsson, J.P., Berggren, D., Simonsson, M., Zysset, M. and Mulder, J. 1998. Aluminium solubility

controls in podzolised soils. Manuscript in preparation.

3

1. Review - What are allophane and imogolite?

The first discoveries

In some of the volcanic ash soils of Japan and New Zealand, the occurrence of allophane

has been known for a long time. These soils were (and are) very difficult for farmers to

manage because of their very anion adsorption capacities. As a consequence, phosphorus (P)

fertilisers were simply fixed very strongly by the soils instead of being taken up by the plants.

The chemical structure of allophane was long unknown, although soil chemists knew that it

was an X-ray aluminosilicate, i.e. a compound composed of aluminium (Al), silicon (Si),

oxygen (O) and hydrogen (H) in chemical combination which was not crystallised enough to

be detected by X-ray diffraction techniques.

In 1962, the Japanese researchers Yoshinaga and Aomine68 discovered a more crystalline

material in one such volcanic ash soil. This compound had a fixed chemical structure, i.e. an

Al:Si atomic ratio of 2:1 and a fibre-like appearance under the transmission electron

microscope (TEM). The new mineral was called imogolite (the prefix imogo is Japanese for

‘glassy volcanic ash’).

At first soil scientists believed that imogolite, as well as allophane, was a unique thing for

volcanic ash soils. Already in 1972, however, Byron and Shimoda7 found evidence for the

existence of allophane in acid forest soils of Nova Scotia, Canada. And in 1978, Tait et al.59

identified imogolite in the B horizon of Scottish Podzols. Since then imogolite has been

identified in the B horizon of many Podzols around the world9, 66, 69 , including Sweden

(chapter 2).

From the 1970s and onwards intense research activities in this subject were carried out by

V.C. Farmer and others at the Macaulay Land Use Research Institute in Scotland2, 16-20, 37-38.

Farmer, originally a trained spectroscopist, described the structure and properties of allophane

and imogolite, as well as discussing the possible role of these minerals in the formation of

Podzols. One of Farmer’s early achievements was the finding that the allophane occurring in

Podzols could be described as a precursor or a poorly crystallised form of imogolite19.

4

Fig. 1. Podzol formation according to Farmer2, 16 .

The connection between imogolite and the formation of Podzols

Farmer developed a view of Podzol formation2, 16 (sketch in Fig. 1), which proved to be

controversial. His view was in fact a new version of the ideas put forward by the Swedish

researchers Sante Mattson and Yngve Gustafsson as early as 193741. Farmer questioned the

‘classical’ podzolisation theory, advocated by Ponomareva50, Petersen49 and others, which

centred around the role of organic acids as carriers for weathered Al and Fe from the leached

E horizon to the B horizon, in which the organo-metallic complexes were arrested and

degraded. Farmer instead argued that within the leached E horizon, released Fe and Al

combined with Si to form positively charged iron-containing ‘proto-imogolite sols’, i.e. high-

molecular weight, polymeric Fe-Al-Si complexes. These sols migrated downwards to the B

horizon in which the higher pH caused the arrest of the proto-imogolite sols and the

subsequent formation of allophane, imogolite and Fe oxide.

OE

B

C

pH = 4

pH = 5

weatheringFe, Al, Si proto-imogolite

sols

Fe oxideAllophaneImogolite

organic acids

Adsorption oforganic acids

In 1984, Buurman and van Reeuwijk8 refuted this concept suggesting that allophane and

imogolite may instead have formed inside the B horizon once the organo-metallic complexes

had been degraded. Later soil solution studies carried out by, among others, Ugolini and

Dahlgren in the U.S.62 and by Van Hees et al. in northern Sweden64, support this view as they

showed that most dissolved Al in the E and B horizons is present either as organic complexes

or as monomeric ionic species despite the occurrence of allophane and imogolite in the B

horizon of these soils.

5

Fig. 2. Current view of the process for Podzol formation.

Today most soil scientists agree that ‘proto-imogolite sols’ do not migrate to any significant

extent and that allophane and imogolite are formed directly in the B horizon where the ion

activity product for these minerals, in contrast to the situation in the E horizon, is sufficiently

high to allow their formation. We also have a clearer overall picture of podzol formation. The

recent state-of-the-art reports by Courchesne and Hendershot12 and by Lundström et al.39, to

which the interested reader is referred, allow us to present the generalised picture in Fig. 2.

Although many details remain unclear, the following stages are almost certainly involved:

OE

B

C

pH = 4

pH = 5

weatheringFe, Al, Si organo-Fe

organo-Al

Fe oxideAllophaneImogoliteAl hydroxide

organic acids

silicic acid

Fe

Al-organiccomplexesAl

1. The formation of the white / grey E horizon is induced by intense weathering, which is, at

least partly, mediated by the excretion of low-molecular weight organic acids from

mycorrhizal hyphae.

2. The released Al and Fe are complexed with organic acids and these complexes migrate in

the soil solution to the reddish B horizon where they are arrested mainly because of

adsorption to previously formed allophane / imogolite and Fe oxides. Iron, however, is

probably quickly removed from the organic complexes by precipitation of Fe oxide once

the complexes reach the less acid B horizon.

3. Microbial activity degrades the organic components of the organo-Al complexes causing

an increase of the soil water concentrations of inorganic Al. This in turn causes aluminium

hydroxide and imogolite / allophane to precipitate.

6

Chemical properties of imogolite and allophane

The chemical formula of imogolite is (HO)3Al2O3SiOH. Its equilibrium with dissolved Al3+

ions and silicic acid is described with the following solubility reaction:

0.5(HO)3Al2O3SiOH + 3H+ ⇔ Al3+ + 0.5H4SiO4 + 1.5H2O *Ks

According to Farmer and Fraser17 the *Ks value at 25oC was about 106.0. Later more precise

determinations suggest a *Ks value of about 106.5 for imogolite55 and slightly more for Al-rich

(see below) allophane37-38, 58. Imogolite is a paracrystalline mineral, which means that it

crystallises only in one direction. As a result imogolite appears as long thread- or fibre-like

structures which may be up to 1 µm long while the outer diameter is not more than 2 nm65.

Fig. 3 shows a network of imogolite fibres isolated from a Japanese pumice bed. In 1972,

Cradwick et al.13 presented a structural analysis of imogolite which, with minor

modifications, remains valid (see Fig. 4). According to their analysis, an imogolite fibre is a

tube. Orthosilicate groups face the inner side while the outer sheet consists of gibbsite units.

Fig. 3. Transmission electron micrograph of imogolite from a Japanese pumice bed.

Fig. 4. Cross-section of an imogolite tube. From Cradwick et al.13

Important properties of imogolite are the high surface area and the high anion adsorption

capacity, see chapter 6. The high anion adsorption capacity is a consequence of the gibbsite

groups of the outer sheet, which causes imogolite to behave much like gibbsite. As in the case

7

of gibbsite extra hydrogen ions are easily incorporated into the hydroxyls at the surface.

Because of this the outer surface of imogolite, in its pure form, possesses a net positive charge

below pH 1057, which favours the adsorption of anions in the ‘natural’ pH range.

Allophane is different depending on the environment in which it is formed. In Si-rich

volcanic deposits the Al:Si ratio of allophane can be as low as 1. The Al coordination

chemistry in such allophanes is quite different from that of imogolite in that a significant

proportion of the Al is in tetravalent coordination21, 46. This leads to the presence of structural

permanent negative charge, in addition to the variable charge of the hydroxyl groups. Silica-

rich allophane is not known from Scandinavia, however, in which we only know of the

existence of Al-rich, or ‘proto-imogolite’ allophane. The structure of aluminium-rich

allophane has been described as ‘fragments of imogolite tubes’ 46, 47. The Al:Si ratio is

normally between 2 and 3; the reason for the higher Al:Si ratio compared to imogolite has to

do with the substitution of the O3SiOH tetrahedra of the inner surface for hydroxyls46, 47. This

causes the material to form small (4 - 5.5 nm thick) spherules instead of fibres47.

Methods of detection

Imogolite and allophane are normally referred to as X-ray-amorphous materials, but this is

strictly not correct. They do give weak XRD (X-ray diffraction) traces and Kodama and

Wang32 developed a special method for detecting these minerals by XRD. Imogolite, but not

allophane, is also possible to identify using differential thermal analysis (DTA)65.

Still however, by far the best sensitivity for anyone interested in identifying imogolite or

allophane is obtained using infrared (IR) spectroscopy52. Fig. 5 shows the IR spectra recorded

for synthetic imogolite and Al-rich allophane. The most characteristic band is the 348 cm-1

band which is due to Si-O stretching. Other characteristic bands appear at about 428, 506, 577

and 967 cm-1 while the band in the -OH stretching region (i.e. around 3500 cm-1) is very

broad and diffuse. Imogolite generally has sharper peaks than allophane and it also has

characteristic doublets of the bands at 577 and 967 cm-1; this permits a semi-quantitative

calculation of the relative abundance of the two minerals in a clay sample. The spectra of

allophane and imogolite contrast sharply with the bands of kaolinite and gibbsite, which have

sharp bands in the -OH stretching region.

8

Fig. 5. Infrared spectra of ‘proto-imogolite allophane’ and imogolite. From Gustafsson et al.25

Transmission electron microscopy (TEM) is also quite often used to identify imogolite

(allophane is not seen by this method). Although it has a characteristic fibrous morphology,

just seeing imogolite is not enough for identification purposes since there are other ‘fibrous’

minerals. However, if the amount of imogolite fibres is sufficient, characteristic electron

diffraction patterns can be recorded43.

To aid the identification of these minerals it is necessary to avoid the presence of too many

different clays in the isolated clay fraction, as they interfere with the identification. To do this

Farmer and coworkers developed an acid dispersion method to isolate a clay fraction in which

allophane and imogolite occur in a relatively high concentration20.

Indications of the occurrence of imogolite and allophane in soils can be obtained using

oxalate (0.2 M ammonium oxalate / oxalic acid buffer) and pyrophosphate (0.1 M Na4P2O7)

extractants25. Oxalate dissolves all Al and Si bound to imogolite / allophane plus other

reactive Al forms (mainly organically complexed Al). Pyrophosphate dissolves organically

complexed Al and almost no Si. If the ratio of the extracted ‘inorganic’ Al to Si is close to 2,

it indicates that most oxalate-extractable inorganic Al consists of allophane or imogolite.

9

2. Occurrence of allophane and imogolite in Sweden

Previous research

Podzolised soils comprise about 50 % of the total land surface in Sweden. They are

typically associated with coniferous forest vegetation (spruce, pine), coarse-textured parent

materials (normally glacial till) and dry nutrient-poor conditions. Over the years there has

been a lot of research focusing on the genesis and properties of Podzols.

In the B horizons of Swedish Podzols, aluminium phases, and to a lesser degree, iron

phases, have accumulated29, 44, 61 (see chapter 1). The accumulated aluminium is composed of

organically complexed Al and inorganic Al compounds. However, for a long time the identity

of the accumulating inorganic Al compounds was unknown. Already in 1931 Tamm61 (who

was the inventor of the oxalate extraction method) observed the presence of substantial

amounts of Si in the oxalate extracts of B horizons from Swedish Podzols. This observation

was repeated in 1988 by Nömmik et al.44 who also noted that there was a relationship

between the oxalate-extracted Al and Si in their B horizons and they speculated that this

might have been due to the presence of imogolite.

Fig. 6. Relationship between oxalate-extractable Si (Sio) and oxalate minus pyrophosphate-extractable Al (Alo - Alp) in 67 B horizon samples from northern Sweden, Finland and Norway25.

0 50 100 150 200 250 300Sio (mmol / kg)

0

100

200

300

400

500

600

700

Alo

- Alp

(mm

ol /

kg)

R2 = 0.975y = -4.25 + 2.48x

In 1995 we showed that there was a highly significant relationship between oxalate-

extractable Si and the difference between oxalate- and pyrophosphate-extractable Al in a

collection of podzolised B horizons from northern Sweden, Finland and Norway25 (Fig. 6).

The slope of the regression, 2.48, is close to what would be expected if allophane

predominated among the precipitated Al compounds of the B horizon. Three B horizon

samples were sent to the Macaulay Land Use Research Institute for the identification of

10

allophane and imogolite. The presence of these minerals (predominantly allophane) was

confirmed for all three soils.

0

100

200

300

400

500

Extra

cted

Al (

mm

ol /

kg)

AloAlp

0

100

200

300

400

500

600

700

800

900

Extra

cted

Al (

mm

ol /

kg)

AloAlp

Fig. 7. Oxalate- and pyrophosphate-extractable Al in; (left) 67 soil samples from the top 20 cm of the B horizon in the north25 (same as Fig. 6); (right) 40 soil samples from the top 40 cm of the B horizon in Halland, SW Sweden, which were sampled in an eariler investigation23.

Accumulated aluminium phases in southern Sweden

Whereas the presence of allophane and imogolite of the B horizons seems to be the rule

rather than the exception in the North25, 30, this situation does not always prevail in southern

Sweden. The reasons are that the Podzols of southern Sweden are generally more acid and

that they contain more organic matter. These conditions are unfavourable for the

accumulation of inorganic Al phases such as allophane. Instead most Al is retained by organic

matter complexation. This geographical difference is amply illustrated in Fig. 7, which shows

pyrophosphate- and oxalate-extractable Al in the upper B horizons of (a) 33 Podzols from

northern Scandinavia25 (same as Fig. 6) and (b) another 18 Podzols from Halland, SW

Sweden23. In Halland, the pool of pyrophosphate-extractable Al pool is often of the same

magnitude as the oxalate-extractable pool, meaning that most Al is organic, while in the

North, a considerable proportion of the accumulated Al is inorganic.

However, most of any inorganic Al found in the South is also in the form of allophane and

imogolite, as has been revealed by our most recent study on the subject (I). Seven Podzols of

southern and central Sweden were selected for an in-depth mineralogical analysis (Table 1).

As Fig. 8 shows the corresponding plot to Fig. 6 looks very similar and the slope of the

regression is only somewhat higher, i.e. 2.86.

11

Table 1. Site characteristics for the seven soils of (I)

Site Location MAAT

(oC) Soil texture* Classification O + A horizons

(cm)

Asa 57o11'N;14o49'E 6.1 Gl. till, sl Typic Haplorthod 6.7 Ed 58o55'N;11o55'E 5.1 Gl. till, ls Typic Haplorthod 6.4 Finnedal 59o12'N;18o16'E 5.6 Gl. till, ls Typic Endaquod 9 Gullringen 57o49'N;15o37'E 5.3 Gl. outwash, s Typic Haplorthod 5.1 Heda 56o54'N;15o16'E 5.8 Gl. till, sl Typic Haplorthod 7.9 Kvibille 56o47'N;12o53'E 6.8 Gl. outwash, ls Typic Haplorthod 10.6 Lövkullen 60o13'N;14o35'E 3.5 Gl. till, ls Typic Haplorthod 7.2

MAAT = Mean Annual Air Temperature

*Gl. = Glacial; s = sand; ls = loamy sand; sl = sandy loam.

Fig. 8. Relationship between Sio and (Alo-Alp) for 39 samples from 7 profiles in southern and central Sweden. From (I).

0 50 100 150 200 250Sio (mmol/kg)

-1000

100200300400500600700800

Alo-

Alp

(mm

ol/k

g)

r2 = 0.962y = -6.22 + 2.86x

n = 39

The clay fractions of the B horizons from all seven soils were studied using X-ray

diffraction. As could be expected no evidence for imogolite or allophane was found using

this method. X-ray diffraction plots of untreated samples from two profiles (Gullringen and

Heda) are shown in the Appendix. These plots of course show the presence of quartz (peaks at

3.3, 4.2 Å) and feldspars (two major peaks near 3.2 Å). However, XRD also showed the

common occurrence of hydroxy-interlayered vermiculite (HIV) in the upper B horizon which

has probably formed as a result of the podzolisation process. As HIV generally decreased

with increasing depth in the profile, mica (as in Heda) or chlorite (as in Lövkullen) increased,

suggesting that HIV formed upon the weathering of these minerals. This is in accordance with

previous research by, among others, Lång and Stevens33.

12

The occurrence of allophane and imogolite was studied using IR spectroscopy and TEM.

Using these techniques, allophane and imogolite was identified in all seven soils, although not

in all B horizon subsamples. As one example we show unheated IR spectra for the acid-

dispersable clays from the Lövkullen profile (Fig. 9). The spectra show the presence of

allophane in all horizons except for the uppermost one. In fact the spectra from the deeper

parts of the B horizon of Lövkullen are very similar to the Al-rich allophane spectrum shown

in Fig. 5. Spectra from some of the other soils are shown in the Appendix. In some profiles

heat treatments, in addition to unheated treatments, had to be done to secure the identification

of allophane / imogolite as other clay minerals were present. This methodology has been

described by Russell and Fraser52.

4300 3300 2300 1300 300

Wavenumber / cm-1

Tran

smitt

ance

3460

3525

3620

3696

2345

1000

567-

592

428Lövkullen

Bhs

Bs1

Bs2

Bs3

BC1

BC2

Fig. 9. Infrared spectra from the acid-dispersible clays of the Lövkullen B horizons.

In addition to allophane / imogolite, IR spectroscopy showed the presence of kaolinite and

gibbsite in some of the soils. Both minerals were found in the B horizon of the Gullringen

profile and in the deeper parts of the Heda soil. Kaolinite (peaks at 3620 and 3700 cm-1) was

also found in the uppermost B horizon from Lövkullen. Kaolinite and gibbsite are mostly

believed to have formed during pre-glacial periods67 (see also I) and they are thus inherited

from earlier parent materials.

Transmission electron microscopy detected imogolite in the B horizon of all seven soils,

but again not in all B horizon subsamples. The most well developed imogolite fibres were

found in Heda (TEM micrograph on title page) and in Finnedal (Fig. 10). The reason why

13

these soils, rather than any other ones, possessed the most well developed imogolite fibres

are not clear, however.

Fig. 10. Transmission electron micrograph of the acid-dispersible clay from the Bhs2 horizon of Finnedal.

Summary:

Most accumulated inorganic Al in the B horizon of Swedish Podzols is in the

form of allophane and imogolite. These minerals are, however, unstable in very

acid soils, where most of the accumulated Al is instead present in organic

complexes.

14

3. The dissolution chemistry of allophane and imogolite

The release of aluminium from forest soils is one of the most serious consequences of

acidification. The toxic effects of aluminium on fish is well documented. Despite all these

years of acidification research, we still lack a complete understanding of the mechanisms

involved, as will also be discussed in the next chapter. The role of imogolite (for simplicity

we include Al-rich allophane in the term ‘imogolite’ in chapters 3 and 4) in releasing soil Al

to solution has not previously been addressed. To do this we shall consider the dissolution

chemistry of imogolite.

To study the dissolution rate of soil Al we decided to treat the Bs2 horizon of an imogolite-

containing soil from Lövkullen, Dalarna, Sweden, with a dilute salt (1 mM NaCl) solution for

various time periods at 8oC (paper II). By measuring pH, quickly reacting aluminium11,

silicon and major ions in the resulting extracts, we could determine ion activity products

(IAPs) for the Al-containing minerals of interest using the chemical equilibrium program

MINTEQA21.

IAP (imogolite) = {Al }{H SiO }

{H }

34 4

0.5

3

+

+ ; IAP (gibbsite) = {Al }{H }

3

3

+

+

0 10 20 30 40Time (days)

6,8

7

7,2

7,4

7,6

7,8

log

IAP

(imog

olite

)

1:1 ratio

1:10 ratio

40 10 20 30 0

Time (days)

8,8

9

9,2

9,4

9,6

9,8

log

IAP

(gib

bsite

) 1:10 ratio

1:1 ratio

Fig. 11. Development of the logarithm of the ion activity products for gibbsite and imogolite with time for the Lövkullen Bs2 horizon subjected to 1 mM NaCl in batch experiments. Results are displayed from experiments conducted at a 1:1 (g soil : cm3 solution) as well as at a 1:10 soil:solution ratio.

15

As Fig. 11 shows, the IAP (gibbsite) quickly reached a more or less constant value while,

on the other hand, IAP (imogolite) increased more slowly with time, especially at low

soil:solution ratios. This was interpreted in the following way:

1. The suspensions quickly reached equilibrium with an Al(OH)3 (‘gibbsite’) phase

2. Equilibrium with imogolite was not reached for at least three weeks

In other words two minerals were involved: an Al(OH)3 phase and an imogolite phase,. The

first one reached equilibrium quicker than the second. This was confirmed in a number of

other experiments (II, III). The two following solubility equations were established:

Al(OH)3(s) + 3H+ ⇔ Al3+ + 3H2O log *Ks = 8.29 at 25oC

log *Ks = 9.40 at 8oC

0.5(HO)3Al2O3SiOH(s) + 3H+ ⇔ Al3+ + 0.5H4SiO4 + 1.5H2O log *Ks = 6.60 at 25oC

log *Ks = 7.65 at 8oC

As the constants imply the temperature dependence of Al release in imogolite-containing

soils is quite important. To predict Al release it is therefore necessary to correct for

temperature. We showed (III) that this can be done quite successfully using the Van’t Hoff

approximation using the heats of reaction for these minerals as stated by other researchers37, 45

.

Despite its importance any Al(OH)3 mineral phase is rarely identified (chapter 2). This

contradiction may be due to the fact that the pool of Al(OH)3 is generally small and non-

crystalline, making direct identification difficult. The reason why Al(OH)3 does not

accumulate may be due to the slow transformation to imogolite which could be a more stable

phase thermodynamically (II).

Whatever the reason, it is clear that an Al(OH)3 phase determines Al solubility in many

imogolite-containing soils. This also agrees with the results from some previous studies in

spodic B horizons5, 14 . Fig. 12 shows the results from an experiment with 15 different

imogolite-containing B horizons. The slope between the logarithm of the Al activity and the

pH was remarkably consistent with a log *Ks value for Al(OH)3 of 9.4 at 8oC. However when

16

treated with HCl some of the suspensions deviated from the solubility line as can be seen in

the bottom part of the figure. It is quite probable that large acid additions causes the reactive

Al(OH)3 phase to be completely dissolved, which would explain this deviation5, 6.

Fig. 12. The minus logarithm of the Al activity (pAl) as a function of the pH for different imogolite-containing soils at 8oC. The solid line represents an Al(OH)3 phase with a log *Ks value of 9.4, while the dashed line represents crystalline gibbsite with log *Ks = 8.8545.

For imogolite the results indicated that the solubility constant is more variable. The value

for Lövkullen Bs2 (log *Ks = 7.65 at 8oC) was the highest recorded among the investigated

soils. The Finnedal B horizons, for instance, had log *Ks values of about 7.4 (II). Probably

soil imogolites do not have a fixed crystallinity nor a fixed composition; they continuously

undergo transformations which might explain the different solubilities in different soils56.

Not all imogolite-containing B horizons behave in the fashion described above. As

Simonsson and Berggren53 have pointed out, in some top B horizons Al activities are

undersaturated with respect to imogolite and gibbsite (i.e. less Al is dissolved at the same pH

than we would expect from the above equations). The reasons will be discussed in Chapter 4.

Summary:

Aluminium solubility in imogolite-containing soils is often determined by an Al

hydroxide phase, rather than by allophane or imogolite. Furthermore the solubility

constant for Al(OH)3 is the same in different soil environments while the one for

allophane / imogolite is variable.

4,2 4,4 4,6 4,8 5 5,2 5,4pH

3

4

5

6

7

pAl

NaCl, 1:1 ratioHCl, 1:1 ratioInitial oversaturationNaCl, 1:10 ratioHCl, 1:10 ratio

17

4. What mechanisms are responsible for the aluminium release?

It is now agreed that there is not one single mechanism that determines Al solubility in

soils5, 34, 60. We can divide the Al solubility mechanisms into two categories:

I. Complexation by organic matter

Aluminium has a large tendency to form complexes with carboxylic and phenolic groups of

the soil organic matter. This process can be described by the following reaction53:

AlX + nH+ ⇔ HnX + Al3+

where X represents an organic ligand residing in the solid phase while n represents the

number of protons that are exchanged for each Al3+. Thus there is an equilibrium between the

Al3+ concentration in the water phase and Al3+ in the complexed phase. The less Al3+ in the

complexed phase, the lower is also the Al3+ concentration in solution.

II. Precipitation / dissolution of Al-containing minerals

As we could see in the previous chapter Al solubility may also be governed by equilibria

with Al(OH)3 and (less likely) imogolite. For this mechanism to be operating, however, there

must have been a certain amount of Al3+ present (at a given pH) to initiate the precipitation of

Al(OH)3. This is mirrored by the ion activity product.

When does what mechanism regulate the Al solubility?

To illustrate the different mechanisms that are operating, we can consider measured Al and

calculated ion activity products for the different soil horizons of the Finnedal soil27 (Fig. 13),

which is an acid forest soil situated in the Tyresta National Park, south of Stockholm.

18

5 6 7 8 9

log IAP (gibbsite)10

50

40

30

20

10

0

-10

Dep

th (c

m)

OA

Bhs1

Bhs2

Bs

C

0 10 20 30 40 50 60 70 80 90 100 110

Al (uM)

-10

0

10

20

30

40

50

Dju

p (c

m) Al3+

Al-OHAl-SO4Al-FAl-org

Fig. 13. Concentrations of Al3+ and of Al complexes in the soil water of different horizons in the Finnedal soil (left)27 and calculated ion activity products for gibbsite / Al(OH)3 (right). The dashed line represents the solubility line for an Al(OH)3 phase with a log *Ks of 9.4 at 8oC.

In the surface horizons, large amounts of acidity is generated by the production of organic

acids during the microbial breakdown of organic matter. As the pH is low the Al3+ activity

has to be quite high to precipitate reactive inorganic Al(OH)3. However the weathering of soil

minerals is usually not sufficiently high as to allow an ample supply of Al. As a consequence

these horizons are undersaturated with respect to Al(OH)3 (see Fig 13). In such horizons the

Al solubility is determined solely by complexation to organic matter.

Further down into the B horizon the pH increases. This is in itself a consequence of the

microbial breakdown of the organic acids, as well as cumulative effects of weathering along

the water flow path. Now much less Al is required to precipitate Al(OH)3. In the Finnedal soil

the weathering of Al from soil minerals is sufficient to cause an Al(OH)3 equilibrium already

in the upper B horizon. Although the organic complexation for Al is still in operation the Al3+

activity will be fixed by the Al(OH)3 solubility relationship. It is then said that Al(OH)3,

rather than organic complexation, determines Al solubility.

As was mentioned in the previous chapter some soils maintain the Al-organic complexation

equilibrium at greater depths than in the Finnedal soil. We believe that this should be seen in

the contexts of pedogenesis and acidification26, 53. In many cases the upper B horizons, and

sometimes the whole of strongly acidified soils may be seen as zones of eluviation since more

Al is leached from such horizons than is supplied from above horizons and from weathering

(qout > qin, see Fig. 14). Since the pool of Al(OH)3 is small, this pool may soon be depleted,

and as a consequence the Al solubility control is handed over to organic complexation.

19

Although some imogolite may still persist, the dissolution of imogolite is not fast enough to

control Al concentrations.

Fig. 14. Simplified sketch of the mass balance of aluminium in a B horizon of an acid forest soil. Biological mechanisms have been disregarded. From paper III.

By contrast in zones of illuviation, with an ample supply of Al, the Al(OH)3 pool is

constantly renewed which causes Al solubility to be governed by Al hydroxide. However Al

hydroxide will not accumulate to a significant extent because the Al(OH)3 will slowly

redissolve to form imogolite which is more thermodynamically stable.

It is also important to consider the water flow patterns and soil heterogeneity. If preferential

flow occurs (i.e. fast water flow along stones etc. after major rain events) the drainage water

from B horizons may be undersaturated with respect to Al(OH)3 as the contact time with the

soil has been minimal42. It is also quite likely that different Al solubility controls will prevail

in different soil microenvironments because of soil heterogeneity. This makes the overall

picture for, say, a forested watershed quite complex and difficult to model.

Summary:

The mechanisms controlling the Al release from soils are complex. Commonly

organic complexation is important in the surface horizons of acid soils, while Al

hydroxide dissolution may be important in deeper horizons. The balance between

the amounts of supplied and leached Al, as well as water flow patterns, are

important factors in determining the overall Al solubility.

Weathering

Inflow

Leaching

qin

qout

[Al3+]

Primaryminerals

Al-organiccomplexes

Al(OH)3

Imogolite

υ1

υ-1

20

5. Surface properties of synthetic imogolite

Probably the most significant property of allophane and imogolite is their ability to bind

anions. Ion binding can occur on the reactive hydroxyls of the outer gibbsite sheet of the

imogolite tube. These hydroxyls have the ability to lose, or to gain, protons from solution and

hence to change the surface charge. At low pH the net charge of the surface is positive while

it gets negative at very high pH. This can be shown by the following two chemical reactions:

≡AlOH + H+ ⇔ ≡AlOH2+ Ka, 1

≡AlOH ⇔ ≡AlO- + H+ Ka, 2

The pH value where the net charge is zero, i.e., when [≡AlOH2+ ] = [≡AlO- ], is called the

pH of point-of zero charge (pH(PZC)). For pure imogolite this pH is about 10 or maybe even

somewhat higher57. Below pH 10 the positively charged surface species [≡AlOH2+ ]

predominates. This enables anions to be adsorbed by a combination of electrostatic attraction

and specific chemical (‘surface complexation’) forces. Anions such as phosphate, arsenate

and fulvic acids are easily adsorbed by these surfaces since they form a strong surface

complex with the Al central atom10, 48, while chloride and nitrate, which do not form surface

complexes, are hardly adsorbed at all. Anion adsorption is further enhanced by the large

specific surface area of imogolite and allophane; BET (N2) surface areas of c. 150 m2 / g for

imogolite, and of c. 580 m2 / g for allophane, have been reported46, 65. Due to the high

pH(PZC) most cations do not sorb to these surfaces at natural pH, unless they (like lead) form

strong surface complexes.

Therefore soils containing allophane and imogolite may be expected to possess very high

anion adsorption capacities. In other studies focusing on sulphate and arsenate adsorption in

imogolite-containing soils we have also found this to be the case22, 24-25, 29.

To study anion adsorption to pure imogolite, we synthesised imogolite according to the

procedure of Farmer et al.18 (displayed in the Recipe box). The resulting imogolite suspension

contained 2.5 mM Al as imogolite. Freeze-dried specimens of this suspension showed that

this imogolite had a BET(N2) surface area of 133 m2 / g. In various experiments, we added

low concentrations of arsenate and lead. After reacting these suspensions the solid imogolite

21

phase was separated from the water phase with centrifugation and filtration through 0.1 µm

nylon filters. Despite these precautions it was difficult to obtain a colloid-free water phase at

low pH because of the small size of imogolite.

Recipe - How to mix your own imogolite

1. Prepare a 2 mM H4SiO4 stock solution by vigorous stirring for 16 h of tetraethyl-o-silicate

diluted to 2 mM Si with water.

2. Prepare a solution containing 2.5 mM Al(ClO4)3 and 1.55 mM H4SiO4; adjust to pH 5 and

then reacidify immediately to give final added concentrations of 1 mM HClO4 and 2 mM

acetic acid. Let this mixture stand overnight before the next step.

3. Reflux the above solution at 95-100oC for 1-3 days

4. Dialyse the obtained imogolite suspension to reduce salt content.

Fig. 15. Adsorption of arsenate (AsV) to synthetic imogolite as a function of pH and added As concentration.

3 4 5 6 7 8 9 10 11

pH

0

20

40

60

80

100

% A

sV a

dsor

bed TOTAs = 2 uM

TOTAs = 5 uMTOTAs = 7.5 uMTOTAs = 10 uMTOTAs = 15 uM

As Fig. 15 shows the

adsorption of arsenate was very strong at low pH, but above pH 7 the adsorption of arsenate

dropped steadily. At higher As concentrations the adsorption was less efficient, simply

because the surface was close to saturation. The maximum arsenate adsorption to synthetic

imogolite was in the order of 0.015 mol / mol imogolite-bound Al, which agrees well with

previous studies with phosphate10. This value is higher for allophane, probably as a result of a

higher surface area and/or more reactive hydroxyls10. The As adsorption behaviour of

imogolite is similar to that of Al hydroxide / gibbsite3, 40 , while it is different from arsenate

adsorption to Fe oxide15, 40 as the latter adsorbs As more strongly at high pH values.

22

Fig. 16. Adsorption of lead to synthetic imogolite as a function of pH and added Pb concentration.

For Pb (lead) the adsorption was strong above pH 7 but it dropped to almost zero below pH

5. Again Pb adsorption was similar to that of Al hydroxide31 but different (lower) than that of

Fe oxide15, 31. In fact the Pb adsorption to imogolite was so weak that inert teflon materials

had to be used throughout in order not to obscure lead adsorption with lead uptake by the

plastic container walls. In nature the naturally occurring imogolites and allophanes are partly

covered with surface complex-forming ions which render the pH(PZC) lower. In the field

therefore Pb adsorption to imogolite is stronger, but As adsorption weaker, than to pure

imogolite.

Summary:

Because of a high pH(point-of zero charge) imogolite and allophane may adsorb

anions strongly, while the adsorption of metal cations is comparably weak. The

arsenate and lead adsorption patterns are similar to those of gibbsite / Al

hydroxide, but unlike those of Fe oxide.

4.5 5 5.5 6 6.5 7 7.5pH

0

20

40

60

80

100

% P

b ad

sorb

ed

TOTPb = 10 µMTOTPb = 1 µM

23

6. Possible engineering applications

Environmental Engineering

Thanks to their surface characteristics, allophane and imogolite can be used as sorbents for

contaminants and nutrients. Their capacity to remove anionic contaminants such as arsenic

and cyanide, and anionic nutrients such as phosphate, is excellent. As described in the

previous chapter the removal of cationic heavy metals, such as lead, copper and cadmium, is

poor at low pH, but it may be significant at higher pH especially for naturally occurring

allophane and imogolite due to their lower pH(PZC) compared to their pure counterparts.

��������������������������������������������������������������������������������������������������������������

������������������������������������������������������������������������������������������������������������������������������������

Contami- nated soil

B-horizon

Sand

Sand

Peat/sand

Arsenic concentration in the double-barrier column

1 36 62 890

0.5

1

1.5

2

2.5

3 ���������������������������������������������

�������������������������

������������������������������

������������������������� �����

Time (days)

As c

onc

(mg/

dm3)

����������Before barriers

After B-horizon barrierAfter peat/sand barrier

Fig. 17. Removal of arsenic from a contaminated water by a spodic B horizon in a column experiment. From Lindberg et al.35

A rapidly emerging research area is the construction of filters or ‘reactive barriers’ (also

referred to as ‘permeable walls’) for the passive removal of contaminants / nutrients from

waters51, 54. The low cost of this technology makes it an attractive alternative to: (i) traditional

remediation technologies in moderately polluted environments, and (ii) other P removal

options for wastewaters in rural areas. At our laboratory we have previosuly shown that

imogolite-containing B horizons may be used for these purposes, with excellent results for

arsenic35 and cyanide63.

In some applications (i.e., the removal of phosphate from wastewaters) it would be

desirable to use filter materials with a higher concentration of reactive surfaces since the

concentration of allophane / imogolite in a B horizon soil matrix is often only about 1-2 %28.

In theory this could be done by using synthetic allophane / imogolite. Another benefit of

24

using synthetic materials is that there is no need for the ‘mining’ of natural spodic B horizons.

However according to our judgment it appears to be easier and cheaper to synthesise large

and concentrated batches of synthetic Fe oxide for this purpose. As Fe oxide has similar

adsorption properties it is likely that the application of iron-oxide-coated sand4 and similar

media will have the greatest development potential.

Materials engineering

As Farmer and coworkers realised18, synthetic imogolite may have applications as a

catalyst. This is due to its unique properties (i.e., its low surface acidity, absence of

exchangeable cations, its regular, internal and external surface and its defined porosity).

Imogolite suspensions may also be used as a support for solids in suspension. Liz-Marzán and

Philipse36 recently found dispersed imogolite fibres to very efficient stabilisers for platinum

colloids. The sorption of platinum by the dispersed fibres prevents significant platinum

aggregation. According to the authors the use of imogolite fibres can be extended to a general

preparation technique for metallic nanoparticles in a wide range of concentrations.

Summary:

Allophane- and imogolite-containing soils can be used for the removal of

contaminants and nutrients from waters, particularly for arsenic and cyanide.

Imogolite is also an excellent catalyst and a possible support for metallic

nanoparticles, properties which can be used in industrial applications.

25

7. Implications and suggestions for future research

We now recognise allophane and imogolite as important constituents of the B horizon of

forest soils. Because of their common occurrence, and because of their properties, it is clear

that these minerals strongly influence the quality of the water that infiltrates the soil to form

groundwater and surface water. This is due to:

• The ability of dissolving allophane / imogolite to mobilise aluminium in acid soils.

Although allophane / imogolite do not control Al concentrations in acidified waters they

nevertheless comprise large sources of easily weatherable Al in many soil systems.

• The adsorption properties of allophane / imogolite. As has been recognised elsewhere the

sulphate adsorption capacity of soils is closely tied to their allophane / imogolite content25,

29. As a result the response in stream water chemistry to changes in acidic deposition can

be delayed for decades. Also the strong adsorption of arsenic and phosphorus to allophane

/ imogolite makes these elements almost immobile22, 24.

There are still many questions that remain to be answered:

• What is the ecosystem function of allophane / imogolite and other oxidic phases in the B

horizon of forest soils? Possibly they might be important in recycling phosphate and

anionic micronutrients (such as molybdate, selenite, borate) in nutrient-poor soils. This

could be important to consider whenever the original soil cover is to be removed.

• How should we model ion adsorption to allophane and imogolite? To predict the

solubility of contaminants in soils mechanistic adsorption models have to be used. Because

of the enormous complexity of natural systems this is a difficult task.

• To what extent can we use allophane / imogolite and other oxidic phases for environmental

engineering purposes? See Chapter 6.

26

8. References 1. Allison, J.D., Brown, D.S., and Novo-gradac, K.J. 1991. MINTEQA2 / PRODEFA2. A geochemical

assessment model for environmental systems. EPA/600/3-91/021 U.S. EPA, Athens, GA 30613, USA. 2. Anderson, H.A., Berrow, M.L., Farmer, V.C., Hepburn, A., Russell, J.D. and Walker, A.D. 1982. A

reassessment of podzol formation processes. J. Soil Sci. 33, 125-136. 3. Anderson, M.A., Ferguson, J.F. and Gavis, J. Arsenate adsorption on amorphous aluminium hydroxide. J.

Colloid Interface Sci. 54, 391-399. 4. Benjamin, M.M., Sletten, R.S., Bailey, R.P. and Bennett, T. 1996. Sorption and filtration of metals using

iron-oxide-coated sand. Wat. Res. 30, 2609-2620. 5. Berggren, D. and Mulder, J. 1995. The role of organic matter in controlling aluminum solubility in acidic

mineral horizons. Geochim. Cosmochim. Acta 59, 4167-4180. 6. Berggren, D., Mulder, J. and Westerhof, R. 1998. Prolonged leaching of mineral forest soils with dilute HCl

solutions: the solubility of Al and organic matter. Eur. J. Soil Sci. 49, 305-316. 7. Brydon, J.E. and Shimoda, S. 1972. Allophane and other amorphous constituents in a podzol from Nova

Scotia. Can. J. Soil Sci. 52, 465-475. 8. Buurman, P. and Van Reeuwijk, L.P. 1984. Proto-imogolite and the process of podzol formation: a critical

note. J. Soil Sci. 35, 447-452. 9. Byzova, Y.V., Sokolova, T.A. and Sinani, T.I. 1990. Amorphous and weakly crystallized components of

fine fractions of the southern Far East. Soviet Soil Sci. 22, 98-111. 10. Clark, C.J. and McBride, M.B. 1984a. Cation and anion retention by natural and synthetic allophane and

imogolite. Clays Clay Miner. 32, 291-299. 11. Clarke, N., Danielsson, L.-G. and Sparén, A. 1992. The determination of quickly reacting aluminium in

natural waters by kinetic discrimination in a flow system. Int. J. Environ. Anal. Chem. 48, 77-100. 12. Courchesne, F. and Hendershot, W.H. 1997. La genèse des Podzols. Géographie physique et Quaternaire

51, 235-250. 13. Cradwick, P.D.G., Farmer, V.C., Russell, J.D., Masson, C.R., Wada, K. and Yoshinaga, N. 1972. Imogolite,

a hydrated aluminium silicate of tubular structure. Nature Phys. Sci. 240, 187-189. 14. Dahlgren, R.A., Driscoll, C.T. and McAvoy, D.C. 1989. Aluminum precipitation and dissolution rates in

Spodosol Bs horizons in the northeastern USA. Soil Sci. Soc. Am. J. 53, 1045-1052. 15. Dzombak, D.A. and Morel, F.M.M. 1990. Surface complexation modeling - Hydrous ferric oxide. John

Wiley & Sons, New York. 16. Farmer, V.C. 1982. Significance of the presence of allophane and imogolite in podzol Bs horizons: a review.

Soil Sci. Plant Nutr. 28, 571-578. 17. Farmer, V.C. and Fraser, A.R. 1982. Chemical and colloidal stability of sols in the Al2O3-Fe2O3-SiO2-H2O

system: their role in podzolization. J. Soil Sci. 33, 734-742. 18. Farmer, V.C., Adams, M.J., Fraser, A.R. and Palmieri, F. 1983. Synthetic imogolite: Properties, synthesis,

and possible applications. Clay Miner. 18, 459-472. 19. Farmer, V.C., Fraser, A.R., Russell, J.D. and Yoshinaga, N. 1976. Recognition of imogolite structures in

allophanic clays by infrared spectroscopy. Clay Miner. 12, 55-57. 20. Farmer, V.C., Russell, J.D. and Berrow, M.L. 1980. Imogolite and proto-imogolite allophane in spodic

horizons: evidence for a mobile aluminium silicate complex in podzol formation. J. Soil Sci. 31, 673-684. 21. Goodman, B.A., Russell, J.D., Montez, B., Oldfield, E. and Kirkpatrick, R.J. 1985. Structural studies of

imogolite and allophanes by aluminium-27 and silicon-29 nuclear magnetic resonance spectroscopy. Phys. Chem. Miner. 12, 342-346.

22. Gustafsson, J.P. 1998. Arsenate adsorption to spodic B horizons: Rate of reaction, effect of competing ions and surface complexation modelling. Geochim. Cosmochim. Acta, submitted manuscript.

23. Gustafsson, J.P. and Jacks, G. 1993. Sulphur status in some Swedish Podzols as influenced by acidic deposition and extractable organic carbon. Environ. Poll. 81, 185-193.

24. Gustafsson, J.P. and Jacks, G. 1995. Arsenic geochemistry in forested soil profiles as revealed by solid-phase studies. Appl. Geochem. 10, 307-315.

25. Gustafsson, J.P., Bhattacharya, P., Bain, D.C., Fraser, A.R. and McHardy, W.J. 1995. Podzolisation mechanisms and the synthesis of imogolite in northern Scandinavia. Geoderma 66, 167-184.

26. Gustafsson, J.P., Simonsson, M., Berggren, D. and Lumsdon, D.G. 1997. Aluminium solubility in the spodic B horizon and its relationship with pedogenesis and acidification. J. Conf. Abs. 2, 190.

27. Gustafsson, J.P., Simonsson, M. and Karltun, E. 1998. Podzolisation in the shallow soils of Tyresta, Sweden. Conference proceedings, 16th International Congress on Soil Science, Montpellier, France.

27

28. Johansson, L. 1998. Industrial by-products and natural substrata as phosphorus sorbents. Environ. Technol., submitted.

29. Karltun, E. and Gustafsson, J.P. 1993. Interference by organic complexation of Fe and Al on the SO42-

adsorption in spodic B horizons in Sweden. J. Soil Sci. 44, 625-632. 30. Karltun, E., Bain, D.C., Gustafsson, J.P., Mannerkoski, H., Murad, E., Wagner, U., Fraser, A.R., McHardy,

W.J. and Starr, M. 1998. Surface reactivity of poorly ordered minerals in podzol B horizons. Geoderma, accepted manuscript.

31. Kinniburgh, D.G., Jackson, M.L. and Syers, J.K. 1976. Adsorption of alkaline earth, transition, and heavy metal cations by hydrous oxide gels of iron and aluminum. Soil Sci. Soc. Am. J. 40, 796.

32. Kodama, H. and Wang, C. 1989. Distribution and characterization of noncrystalline inorganic componenets in Spodosols and Spodosol-like soils. Soil Sci. Soc. Am. J. 53, 526-534.

33. Lång L.-O. and Stevens R.L. 1996. Weathering variability and aluminium interlayering: clay mineralogy of podzol profiles in till and glaciofluvial deposits, SW Sweden. Appl. Geochem. 11, 87-92.

34. LaZerte, B.D. and Findeis, J. 1995. The relative importance of oxalate and pyrophosphate-extractable aluminum to the acidic leaching of aluminium in Podzol B horizons from the Precambrian Shield, Ontario, Canada. Can. J. Soil Sci. 75, 43-54.

35. Lindberg, J., Sterneland, J., Johansson, P.-O. and Gustafsson, J.P. 1997. Spodic material for in situ treatment of arsenic in ground water. Ground Water Mon. Remed. 17, 125-130.

36. Liz-Marzán, L.M. and Philipse, A.P. 1995. Stable hydrosols of metallic and bimetallic nanoparticles immobilized on imogolite fibers. J. Phys. Chem. 99, 15120-15128.

37. Lumsdon, D.G. and Farmer, V.C. 1995. Solubility characteristics of proto-imogolite sols: How silicic acid can de-toxify aluminium solutions. Eur. J. Soil Sci. 46, 179-186.

38. Lumsdon, D.G. and Farmer, V.C. 1997. Solubility of a proto-imogolite sol in oxalate solutions. Eur. J. Soil Sci. 48, 115-120.

39. Lundström, U., van Breemen, N. and Bain, D.C. 1998. The podzolisation process. A review. Geoderma, submitted.

40. Manning, B. and Goldberg, S. 1996. Modeling competitive adsorption of arsenate with phosphate and molybdate on oxide minerals. Soil Sci. Soc. Am. J. 60, 121-131.

41. Mattson, S. and Gustafsson, Y. 1937. The electro-chemistry of soil formation. I. The gel and the sol complex. Annals of the Agricultural College of Sweden 4, 1-54.

42. Matzner, E., Pijpers, M., Holland, W. and Manderscheid, B. 1998. Aluminum in soil solutions of forest soils: Influence of water flow and soil aluminum pools. Soil Sci. Soc. Am. J. 62, 445-454.

43. McHardy, W.J. 1971. Imogolite. In The electron-optical investigation of clays (ed. J.A. Gard), pp. 359-364. Mineralogical Society, London.

44. Nömmik, H., Popovic, B., Larsson, K., Nilsson, Å. and Clegg, S. 1988. Sulphate stores and sulphate adsorption capacity of the spodic B horizon of Swedish forest soils. Report SNV Pm 1969, Swedish Environmental Protection Agency, Stockholm, Sweden.

45. Palmer, D.A. and Wesolowski, D.J. 1992. Aluminum speciation and equilibria in aqueous solution: II. The solubility of gibbsite in acidic sodium chloride solutions from 30 to 70oC. Geochim. Cosmochim. Acta 56, 1093-1111.

46. Parfitt, R.L. 1990. Allophane in New Zealand - A review. Aust. J. Soil Res. 28, 343-360. 47. Parfitt, R.L. and Henmi, T. 1980. Structure of some allophanes from New Zealand. Clays Clay Miner. 28,

285-294. 48. Parfitt, R.L., Fraser, A.R. and Farmer, V.C. 1977. Adsorption on hydrous oxides. III. Fulvic acid and humic

acid on goethite, gibbsite and imogolite. J. Soil Sci. 28, 289-296. 49. Petersen, L. 1976. Podzols and podzolisation. Ph.D. thesis, Royal Veterinary and Agricultural University,

Copenhagen. 50. Ponomareva, V.V. 1969. Theory of podzolisation. Nauka, Moskva-Leningrad. Israel Program for Scientific

Translations, Jerusalem. 51. Ptacek, C.J., Blowes, D.W., Robertson, W.D. and Baker, M.J. 1994. Adsorption and mineralization of

phosphate from septic system effluent in aquifer materials. In: Proceedings of the Waterloo Centre for Groundwater Research, Annual Septic System Conference, Waterloo Nutrient Removal Technologies and Onsite Management Districts, Waterloo, Ontario, June 6 1994, pp. 25-44.

52. Russell, J.D. and Fraser, A.R. 1994. Infrared methods. In Clay Mineralogy: Spectrocopic and chemical determinative methods (ed. M.J. Wilson), pp. 11-67. Chapman & Hall, London.

53. Simonsson, M. and Berggren, D. 1998. Aluminium solubility related to secondary solid phases in upper B horizons with spodic characteristics. Eur. J. Soil Sci. 49, 317-326.

54. Starr, R.C. and Cherry, J.A. 1994. In situ remediation of contaminated ground water: The funnel-and-gate system. Ground Water 32, 465-476.

28

55. Su, C. and Harsh, J.B. 1994. Gibbs free energies of formation at 298 K for imogolite and gibbsite from solubility measurements. Geochim. Cosmochim. Acta 58, 1667-1677.

56. Su, C. and Harsh, J.B. 1998. Dissolution of allophane as a thermodynamically unstable solid in the presence of boehmite at elevated temperatures and equilibrium vapor pressures. Soil Sci. 163, 299-312.

57. Su, C., Harsh, J.B. and Bertsch, P.M. 1992. Sodium and chloride sorption by imogolite and allophanes. Clays Clay Miner. 40, 280-286.

58. Su, C., Harsh, J.B. and Boyle, J.S. 1995. Solubility of hydroxy-aluminium interlayers and imogolite in a Spodosol. Soil Sci. Soc. Am. J. 59, 373-379.

59. Tait, J.M., Yoshinaga, N. and Mitchell, B.D. 1978. The occurrence of imogolite in some Scottish soils. Soil Sci. Plant Nutr. 24, 145-151.

60. Takahashi, T., Fukuoka, T. and Dahlgren, R.A. 1995. Aluminum solubility and release rates from soil horizons dominated by aluminium-humus complexes. Soil Sci. Plant Nutr. 41, 119-131.

61. Tamm, O. 1931. Studier över jordmånstyper och deras förhållande till markens hydrologi i Nordsvenska skogsterränger. Meddelanden från Statens Skogsförsöksanstalt 26, 163-355.

62. Ugolini, F.C. and Dahlgren, R.A. 1987. The mechanisms of podzolization as revealed by soil solution studies. In Podzols et Podzolisation (eds. D. Righi and A. Chauvel), pp. 195-203. Association Francaise de l’Étude du Sol, Paris.

63. Ullrich, J. 1995. In-situ treatment of cyanide contaminated groundwater with filters of podzolic B-horizon material. Studienarbeit Verfahrenstechnik. Division of Land and Water Resources, Royal Institute of Technology, Stockholm.

64. Van Hees, P., Lundström, U., Starr, M. and Giesler, R. 1998. Modelling of factors influencing aluminium in soil solution of podzolic soils. Geoderma, submitted manuscript.

65. Wada, K. 1989. Allophane and imogolite. In Minerals in Soil Environments (eds. J.B. Dixon and S.B. Weed), pp. 603-638. 2nd ed., SSSA, Madison, WI.

66. Wang, C., McKeague, J.A. and Kodama, H. 1986. Pedogenic imogolite and soil environments: Case study of Spodosols in Quebec, Canada. Soil Sci. Soc. Am. J. 50, 711-718.

67. Wilson, M.J., Bain, D.C. and Duthie, D.M.L. 1984. The soil clays of Great Britain. II. Scotland. Clay Miner. 19, 709-735.

68. Yoshinaga, N. and Aomine, S. 1962. Imogolite in some Ando soils. Soil Sci. Plant Nutr. 8, 22-29. 69. Zysset, M., Blaser, P., Luster, J. and Gehring, A.U. (1998) Evidence for aluminum solubility control by

imogolite in different horizons of a Podzol. Manuscript in review.

29

Appendix Figures A1-A2: X-ray diffraction spectra from two of the soils of (I) Figures A3-A6: Infrared spectra from four of the soils of (I).

0 10 20 30 42-theta (Cu K-alpha)

00

1000

2000

3000

4000

5000

Cou

nts

Bhs

Bs1

Bs2

Bs3

Bs4

Bs5

14.1

Å

7.0

Å

4.84

Å

4.2

Å

3.3

Å3.

2 Å

Fig. A1. X-ray diffraction spectra from clays from the Gullringen soil (untreated). Apart from quartz and feldspars (peaks at 3.2, 3.3 and 4.2 Å) appreciable amounts of vermiculite (7.0 and 14.1 Å) were also observed. This vermiculite was partially hydroxy-interlayered, as evidenced by heat treatments. A small amount of gibbsite (4.84 Å) was seen most clearly in the Bs2 horizon.

0 10 20 30 42-theta (Cu K-alpha)

00

1000

2000

3000

4000

5000

Cou

nts Bs1

Bs2

Bs3

BC2

14.1

Å

10.0

Å

4.2

Å

3.3

Å3.

2 Å

2.95

Å

Fig. A2. X-ray diffraction spectra from the B horizon clays of the Heda soil (untreated). Apart from quartz and feldspars, a small amount of hydroxy-interlayered vermiculite (peak at 14.1 Å) was found in the two uppermost B horizons. At greater depth a mica peak (10 Å) appeared.

30

4300 3300 2300 1300 300

Wavenumber / cm-1

Tran

smitt

ance

3460

3525

3620

3696

2345

1000

567-

592

428

Bs1

Bs2

Bs3

Bs4

Bs5

Asa

Fig. A3. Infrared spectra of the acid-dispersible clays of the Asa B horizons. The presence of allophane and imogolite is seen by the peaks at 348, 428 and 567 cm-1. The presence of a small amount of kaolinite (peaks at 3620, 3700 cm-1) is evident for the Bs1 horizon.

4300 3300 2300 1300 300

Wavenumber / cm-1

Tran

smitt

ance

3460

3525

3620

3696

2345

1000

567-

592

428

Bs1

Ed

Bs2

Bs3

Bs4

Fig. A4. Infrared spectra for the acid-dispersible clays from the Ed B horizons. Peaks at 348, 428 and 567 cm-1 show the presence of allophane and imogolite. No kaolinite or gibbsite could be identified.

31

4300 3300 2300 1300 300

Wavenumber / cm-1

Tran

smitt

ance

3460

3525

3620

3696

2345

1000

567-

592

428Finnedal

Bhs1

Bhs2

Bs

Cg

Fig. A5. Infrared spectra for the acid-dispersible clays from the Finnedal B and C horizons. The B horizons contained allophane and imogolite but no gibbsite or kaolinite. The C horizon only contained a trace amount of allophane and imogolite.

4300 3300 2300 1300 300

Wavenumber / cm-1

Tran

smitt

ance

3460

3525

3620

3696

2345

1000

567-

592

428Heda

Bs1

Bs2

Bs3

BC1

BC2

Fig. A6. Infrared spectra for the acid-dispersible clays from the Heda B horizons. All samples except the BC1 horizon contained allophane and imogolite. The deepest horizons also contained small amounts of

kaolinite (3620 and 3700 cm-1) and gibbsite (peaks at 3460 and 3520 cm-1).

32

Icke-kristallina aluminosilikater i svenska skogsjordar

Slutrapport för projekt finansierat av Naturvårdsverket. Dnr. 802-244-95-Fr

Projektledare: Dr. Jon Petter Gustafsson, avd. f. mark- och vattenresurser, KTH, 100 44 Stockholm. E-mail: gustafjp@aom.kth.se

Podsolen är Sveriges vanligaste jordmånstyp och utgör mer än 50 % av Sveriges landyta. I mineraljorden i podsoler återfinns ett urlakat blekjordsskikt, även benämnt E-horisont, och under den ett anrikat rostjordsskikt, B-horisonten, där järn- och aluminiumföreningar anrikats. Medan man sedan länge vet att järnföreningarna är järnoxider så har det länge varit oklart vad aluminiumföreningarna egentligen består av. Det var först 1995 som klara tecken erhölls på att en viktig del av dessa föreningar i Nordkalottens podsoler utgörs av allofan och imogolit, som är en slags okristallina aluminosilikater. I det här projektet har vi noga studerat huruvida allofan och imogolit även förekommer i södra Sverige och om dessa mineral har någon betydelse för aluminiumupplösning från försurade jordar och för ämnestransport.

Sju podsoler i södra och mellersta Sverige utvaldes för studien. Rostjordarna i dessa

podsoler karakteriserades noga. Samtliga rostjordar befanns innehålla allofan och imogolit, om än inte i samtliga underhorisonter. Av det aluminium som anrikats i rostjorden var en stor del organiskt bundet men det mesta av den resterande delen var allofan och imogolit. Således var allofan / imogolit den dominerande oorganiska Al-formen som anrikats. En annan men mindre viktig del var Al bundet som polymerer mellan vermikulitskikt. Smärre mängder kaolinit och gibbsit påträffades också i vissa horisonter.

Genom en serie skakförsök har vi erhållit kunskap om vilka processer som styr

aluminiumupplösning i rostjordar. Trots att allofan / imogolit dominerar tycks dessa mineral sällan kontrollera Al-koncentrationerna i markvattnet, detta beroende på att de nybildas / upplöses alltför långsamt. I stället tycks ofta någon form av Al-hydroxid kontrollera Al-koncentrationen. Denna Al-hydroxid ackumuleras aldrig i någon större mängd i marken förmodligen beroende på att den är termodynamiskt instabil jämfört med allofan / imogolit och långsamt omvandlas till denna fas istället. Inom projektets ram bestämdes jämviktskonstanter för de berörda Al-hydroxid- och imogolitfaserna. I jordar som utsätts för nettoutlakning av Al, till exempel i rostjordar utsatta för försurning, löses Al-hydroxiden upp och Al-koncentrationerna kontrolleras då istället av komplexbildning till organiskt material.

Allofan och imogolit har en mycket stor förmåga att adsorbera anjoner på grund av att de

utvecklar positiv laddning vid naturliga pH-värden. I experiment med syntetisk imogolit bekräftades att adsorptionen av arsenat var mycket stark vid pH < 7. För metallkatjonen bly var däremot adsorptionen svag vid pH < 7 men blev stark vid höga pH-värden. Troligen har allofan och imogolit en central betydelse i regleringen av B-horisonternas vattenkemi på grund av sina starka adsorptionsegenskaper.

33