Journal of Geochemical Exploration 71 (2000) 1–12

www.elsevier.nl/locate/jgeoexp

Leachability of major elements from minerals in strong acids

S. Snäll*, T. Liljefors
Geological Survey of Sweden, Box 670, S-75128 Uppsala, Sweden
Received 28 December 1998; accepted 24 March 2000

Abstract
Leachability of major elements from common silicate and oxide minerals has been determined experimentally. K-feldspar,
plagioclase, biotite, muscovite, chlorite, hornblende, epidote, sphene, rutile anatase and ilmenite were leached in aqua regia,
nitric acids, and hydrochloric acids in varying concentrations.
The results revealed appreciable differences in the leachability of major elements from different minerals. Biotite was
completely soluble and chlorite almost completely soluble in the strong acids used in this study, and the yield of elements
was nearly complete. For less soluble minerals, such as feldspars and hornblende, grain size is of great importance for the
resulting yield. By using ⬍63 mm instead of ⬍2 mm fractions, the leachability of major elements can be increased from 2–3 up
to 7-fold. In some cases, strong acids with a hydrochloric component are more effective than pure nitric acid for leaching major
elements from the minerals. Leaching of a ⬍2 mm fraction rather than finer grain size fractions is recommended if the chemical
analyses are intended to provide data for mineralogical interpretations.
The results obtained in this study are suitable for geochemistry, environmental geology and forestry, where the elemental
concentration in till, based on acid leaching, is frequently presented. The varying solubility of the minerals can be exploited in
many ways, e.g. for determining the content of readily soluble minerals in the samples and as a starting point for calculating the
total mineralogical composition of soil samples, provided that the data on total concentrations of elements and acid-leached
fractions are available. A high content of readily soluble minerals and consequently with high leachability of elements indicates
high resistance to acidification and high soil fertility. 䉷 2000 Elsevier Science B.V. All rights reserved.
Keywords: leaching studies; leachability of elements; strong acids; minerals; silicates; oxides

1. Introduction The results are stored in a database and some have

been published in the Geological Survey (series
In addition to total chemical analyses, various Rapporter och meddelanden). Also, at the Geological
partial analyses (various leaching procedures) are Survey of Finland (Salminen, 1995) leaching by aqua
used to analyse the geological samples. Within the regia is a routine procedure. Since 1997, the SGU also
field of geology, strong mineral acids are commonly applies HNO3 (7 M) as a leachant, in addition to aqua
used as leachants. In the geochemical mapping regia leaching of samples. This is done: (1) for analy-
programme at the Geological Survey of Sweden tical reasons; and (2) to make the procedure conform
(SGU), aqua regia leaching of samples has been more to that of a Swedish standard method for deter-
applied since the start of the programme, in 1983. mination of metals in sediments (Swedish Standard
SS 028150, 1980).
* Corresponding author. Fax: ⫹46-18-17-92-10. The proportions (percentages of total content) of
E-mail address: [email protected] (S. Snäll). the elements soluble by the leachant (leachability)
0375-6742/00/$ - see front matter 䉷 2000 Elsevier Science B.V. All rights reserved.
PII: S0375-674 2(00)00139-4
2 S. Snäll, T. Liljefors / Journal Geochemical Exploration 71 (2000) 1–12

Table 1
Mineral specimens used for the leaching experiments; chemical composition expressed as oxides

Mineral SiO2 (wt%) Al2O3 (wt%) TiO2 (wt%) Fe2O3 (wt%) MgO (wt%) CaO (wt%) Na2O (wt%) K2O (wt%) LOI (wt%)

Microcline a n.d. 18.8 0.13 4.1 10.9

Plagioclase b 69.1 20.4 0.006 0.01 0.02 1.32 10.3 0.57
Hornblende b 48.6 6.8 0.06 16.3 14.1 11.7 1.13 0.32
Epidote b 38.1 23.1 0.1 14.8 0.11 23.3 ⬍0.05 0.07
Biotite b 35.5 13.4 2.86 29.4 6.56 ⬍0.1 0.18 8.7 0.6
Mg-Chlorite b 35.8 14.8 0.68 8.63 30.3 0.36 0.19 0.48 9
Fe-Chlorite b,c 55.1 11.6 1.47 14.8 9.64 0.24 0.36 0.36 5.2
Muscovite d 46.01 34.21 0.18 0.74 1.55 0.11 0.74 10.13 5.6
Sphene e 30.7 1.3 37.97 0.71 0.01 28.6
Ilmenite f,g 58.8 13
Anatase h 98.6 1.4
Rutile h 98.77 0.61
a
The chemical composition was determined by ICP-MS after dissolution in HF-HNO3.
b
The chemical composition was determined by ICP-AES after fusion with LiBO2 and dissolution in dilute HNO3.
c
Contaminated by quartz.
d
The chemical composition was determined by classical wet chemical analysis (Asklund et al., 1966).
e
The chemical composition was determined by microprobe.
f
The chemical composition was determined by XRF.
g
Contaminated by other Ti minerals.
h
Chemical composition, taken from Deer et al. (1962).

vary according to type of leachant, leachant concen- leached titanium. Information about the mineralogical
tration, leaching time, temperature, grain size and composition of Quaternary deposits has many fields of
mineralogical composition of the samples, and experi- application, e.g. in mineral exploration, in forestry, for
mental conditions (Fällman and Aurell, 1996). For the environmental studies and for modelling of weather-
routine analyses applied in mapping programmes of ing rates (Sverdrup and Warfvinge, 1991).
SGU, as many of these parameters, as possible, are It is well known that calcite is completely dissolved
held constant and the major variation relates to the by strong acids, whereas quartz is almost insoluble.
mineralogical composition of the samples. In geolo- The solubility of other minerals in the acids used for
gical samples, the leachability of the elements is geochemical mapping is not well known.
extremely variable as was illustrated by the leachabil- The purposes of the present study were to ascertain:
ity values reported by Koljonen and Malisa (1991): Si, (1) the effectiveness of the acids to leach elements
0.5; Na, 0.6; Ca, 9; Al, 17; Ti, 28; Mg, 42; Fe, 52; and from various selected minerals; (2) the importance
P, 89% (median values of aqua regia leaching of of the grain size of the analysed fraction regarding
Finnish till samples). However, the leachability of leachability of the elements; and (3) to identify the
each element varies according to compositional varia- fraction that is preferable for leaching, in order to
tions in the mineralogy of the tills (Andersson, 1992; obtain the best mineralogical characterisation of the
Koljonen et al., 1992). samples.
Knowledge of the solubility of common minerals in
tills is important because the acid-soluble fraction,
together with the total content of elements, is often 2. Materials and methods
utilised to calculate the mineralogical composition of
tills and other Quaternary deposits (Räisänen et al., 2.1. Mineral specimen
1995; Snäll and Ek, 1996). Here, for example, the
biotite content in a sample is determined from aqua For the leaching experiments, 12 mineral species
regia leached potassium and the sphene content from (Table 1) commonly present in the glacial tills of
 S. Snäll, T. Liljefors / Journal Geochemical Exploration 71 (2000) 1–12 3

the Precambrian bedrock areas of Scandinavia were 2.3. Leaching

chosen. The most accurate results would be obtained
if the individual minerals had been collected from the In the present study, the following leachants were
tills. However, it is very difficult to obtain pure used: 1 M HCl, 4 M HCl, 1 M HNO3, 7 M HNO3 and
mineral specimens from natural deposits. In every aqua regia (HNO3 and HCl, 1:3). The routine of Sjös-
case, time-consuming separation work is needed. tröm and Qvarfort (1992) was followed for the aqua
To make the study possible within a reasonable regia leaching.
time framework, most of the minerals were gathered Leaching with HNO3 was done in three different
from the collections of the SGU and Uppsala Univer- ways: (1) according to the Swedish Standard (Swedish
sity. The feldspars, the muscovite and the biotite origi- Standard SS 028150, 1980) where 1 g of the sample is
nated from different pegmatite deposits in Sweden, leached in 7 M HNO3 in an autoclave; (2) according to
hornblende and epidote from skarn iron ores in the routine now in use at the SGU for samples in the
Sweden, the Fe-chlorite from a locality near geochemical mapping programme, with 2 g of the
Himmerfjärden (Södermanland county, Sweden), the sample leached in 10 ml 7 M HNO3 (boiling for
Mg-chlorite from Maaninka (Finland), the ilmenite 30 min) on a hot plate in a vessel covered with a
from Sabaragamurva (Sri Lanka) and the anatase watch glass; and (3) in a weaker acid, 1 M HNO3,
from Greina Pass (Switzerland). The rutile has an using the same procedure in order to ascertain how
unknown origin. The sphene (from Capelinha, much the leachability is reduced by using a less
Minas Gerais, Brazil) was purchased from a commer- concentrated acid. Some of the minerals were also
cial company (Geocity AB, Stockholm). leached in HCl for a comparison of the effectiveness
of different acids. HCl leaching is a common method
for selective dissolving of minerals for clay mineral
2.2. Preparation of samples for leaching identification, e.g. chlorite (Brindley, 1961). Two
grams of each sample was leached in boiling 1 M
Three main size fractions were used in the study: HCl for 2 h and in boiling 4 M HCl for 30 min.
⬍2 mm, ⬍63 mm, and ground material of grain size During the laboratory work it was observed that the
⬍10–20 mm (powder). The mineral species were pressure prescribed in Swedish Standard (Swedish
crushed in an Ellis mortar and passed through two Standard SS 028150, 1980) was not wholly achieved.
nylon sieves, dry-sieved through a 2 mm sieve (to This means that the “SS 028150” leaching of this
get material ⬍ 2 mm) and wet-sieved (in distilled study was somewhat similar to that in 7 M HNO3,
water) through a 63 mm sieve (to get material which is carried out at atmospheric pressure.
⬍63 mm). The material was ground in a McCrone
Micronizing Mill in alcohol, using agate cylinders 2.4. Measurement of element concentration
as grinding elements.
The micas and chlorites were prepared in a different 2.4.1. Total concentration
manner, as they could not be crushed in a mortar. Total concentration of elements (expressed as
Plates of these minerals were cut with scissors into oxides, cf. Table 1) in the minerals has been deter-
small pieces, which were then ground repeatedly for a mined with different methods. Analyses of plagio-
few seconds in a Micronizing Mill, subsequently wet- clase, hornblende, epidote, biotite and the chlorites
sieved (in alcohol) through nylon sieves. were done at SGAB, Luleå (Sweden). Samples were
Before leaching, the pure minerals were diluted fused with LiBO2 and dissolved in 5% HNO3. The
with quartz (5 or 10% admixture) (with particle size elements in the solution were determined using induc-
in the range 0.02–0.2 mm). Dilution was necessary tive coupled plasma atomic emission spectrometry
because only very small samples of some of the (ICP-AES). The analysis of elements in microcline
minerals were available for the experiments. Thus, was performed at SGU. Samples were dissolved in
crystals of sphene were handpicked under a micro- hot HF and HNO3 and the elements in the solution
scope to obtain the pure mineral specimen and only were determined using inductive coupled plasma
minor amounts of anatase was found. mass spectrometry (ICP-MS). The sphene was
4 S. Snäll, T. Liljefors / Journal Geochemical Exploration 71 (2000) 1–12

Table 2
Leachability of potassium from biotite, muscovite, microcline and hornblende

Mineral K Leachability (% of total wt%) in

Fraction Total (wt%) 1 M HCl, 2 h 4 M HCl, 0.5 h 1 M HNO3 7 M HNO3 SS 028150 Aqua regia

Biotite 0.5–2 mm 7.3 117 113 29 60 58 105

Powder 7.3 104 107 106 106 109 114

Muscovite Powder 8.4 6.5 5.6 6.2 8.6 8.7 6.1

Microcline ⬍2 mm 9 1.3 0.6 0.4 0.3 0.4 0.2
⬍63 mm 9 3 1.7 1 0.9 1.1 0.5
Powder 9 9.6 6.1 3.2 3.5 3.5 2.5
Hornblende ⬍2 mm 0.27 ⬍0.1 ⬍0.1 ⬍0.1 ⬍0.1 ⬍0.1 ⬍0.1
⬍63 mm 0.27 14 9.7 ⬍0.1 ⬍0.1 3.6 4.2
Powder 0.27 21 21 2.6 14 20 14

analysed with a CAMECA SX50 microprobe, for the powder. In hydrochloric acid, leachability of
equipped with wavelength spectrometers (WDS), at potassium increased to 6–10%. From muscovite
the Department of Earth Sciences, Uppsala University powder, the leachability of potassium was in the
(Sweden). The analysis of the ilmenite (determination range of 6.1–8.6%.
of titanium and iron contents) was performed at Leaching of larger grain sizes reduced the leach-
Analytica AB, Täby (Sweden) by X-ray fluorescence ability of potassium considerably. In the ⬍63 mm
(XRF) technique. For the muscovite, an old analysis fraction, leachability was 2–3 times less for potassium
was available, made according to classical wet chemi- bound in microcline and hornblende, compared with
cal methods (Asklund et al., 1966). the yield from powder, while in the ⬍2 mm fraction,
leachability was approximately 10 times less for
2.4.2. Leached fraction potassium in microcline. The leachability of potas-
sium bound in a coarser fraction (0.5–2 mm) of biotite
2.4.2.1. Potassium. Feldspars, micas, and to some was reduced to approximately 60% in 7 M nitric acid.
extent amphiboles (represented here by hornblende) In hydrochloric acids (including aqua regia), however,
are the major potassium-bearing minerals of the Scan- there was no decrease in leachability. The biotite
dinavian tills and sediments in the Precambrian dissolved completely and all potassium was put into
bedrock areas (Snäll et al., 1992). Table 2 shows the the solution.
leachability of potassium from these minerals (of
different grain size) for the various leachants used in 2.4.2.2. Sodium. Plagioclase is the major source of
the study. sodium in the Scandinavian Quaternary deposits.
Three grain size fractions for microcline and horn- Lesser amounts of sodium are bound in amphiboles.
blende were leached, two for biotite and only ground Table 3 reports the leachability of sodium from pure
sample (powder with grain size ⬍10–20 mm) for plagioclase (with albite content approximately 90%),
muscovite. From Table 2 it is evident that potassium albite in microcline, and from hornblende (of different
is most easily leached from biotite. The finest fraction grain sizes) for the various leachants. The figures for
(powder) of this mineral dissolved completely in all leachability of sodium from the plagioclases are
the acids used in the study. From hornblende, about similar to those of potassium from microcline.
20% of the available potassium was leached from the Sodium leachability may be somewhat higher in the
powder by the most effective acids, but only 2.6% by finest fraction (powder) and somewhat lower in the
1 M nitric acid. Muscovite and microcline yielded the coarsest fraction, than for potassium from
least potassium, with lowest yield for microcline in microcline. As expected, sodium leachability from
nitric acids (including aqua regia), namely 2.5–3.5% hornblende was about the same as for potassium.
 S. Snäll, T. Liljefors / Journal Geochemical Exploration 71 (2000) 1–12 5

Table 3
Leachability of sodium from microcline, plagioclase and hornblende

Mineral Na Leachability (% of total wt%) in

Fraction Total (wt%) 1 M HCl, 2 h 4 M HCl, 0.5 h 1 M HNO3 7 M HNO3 SS 028150 Aqua regia

Albite in microcline ⬍2 mm 3 0.9 0.4 0.3 0.4 0.3 0.3

⬍63 mm 3 1.9 1.0 0.9 0.8 0.9 0.6
Powder 3 6.8 3.8 2.8 2.8 3.6 2.7
Plagioclase ⬍2 mm 7.6 0.2 0.1 0.2 0.2
⬍63 mm 7.6 0.8 0.7 0.9 0.9
Powder 7.6 3.8 4.1 4.5 4.5
Hornblende ⬍2 mm 0.84 1.0 1.1 ⬍0.1 0.4 0.8 0.5
⬍63 mm 0.84 15 11 1.9 3.0 7.8 6.2
Powder 0.84 24 22 7.9 20 27 17

Nitric acid was a more efficient leachant of sodium sphene, approximately 70% leachability in nitric
from powders, than of potassium. acid, and somewhat less, 61% in aqua regia.
Calcium is more readily leached from epidote than
from hornblende. From powdered epidote the stron-
2.4.2.3. Calcium. In non-calcareous tills and gest acids leach 30–34% of the available calcium.
sediments in Scandinavia, calcium is found mainly From hornblende 23–29% was leached by hydrochlo-
in plagioclases, pyroxenes, amphiboles, epidotes and ric acid and the stronger nitric acids (including SS
sphene. In this study, sodium-rich plagioclase (cf. 028150). Aqua regia yielded somewhat less calcium,
Table 1), hornblende, epidote and sphene were all 18%. 1 M nitric acid proved to be a less good
leached in nitric acid and in aqua regia. In addition, leachant. From epidote 15, and from hornblende
hornblende was leached in hydrochloric acid (cf. 8.8% was leached, respectively.
results in Table 4). Calcium is least leachable from the plagioclase
Comparing the results for leaching from powders, it (albite-oligoclase), only 7–10% being leached by
is evident that calcium is most easily leached from the acid used.

Table 4
Leachability of calcium from plagioclase, hornblende, epidote and sphene

Mineral Ca Leachability (% of total wt%) in

Fraction Total (wt%) 1 M HCl, 2 h 4 M HCl, 0.5 h 1 M HNO3 7 M HNO3 SS 028150 Aqua regia

Plagioclase ⬍2 mm 0.94 0.9 1.0 1.0 0.8

⬍63 mm 0.94 2.3 1.8 2.5 2.3
Powder 0.94 10.0 9.7 7.4 7.8
Hornblende ⬍2 mm 8.4 1.4 1.1 0.6 1.0 1.4 1.1
⬍63 mm 8.4 16 11 2.8 3.1 8.5 6.9
Powder 8.4 27 25 8.8 23 29 18
Epidote ⬍2 mm 16.7 0.3 0.7 1.2 0.7
⬍63 mm 16.7 0.9 4.1 4.7 4.8
Powder 16.7 15 30 34 30
Sphene(titanite) ⬍2 mm 20.4 3.6 3.1 2.7
⬍63 mm 20.4 37.0 38.0 42.0
Powder 20.4 71.0 72.0 61.0
6 S. Snäll, T. Liljefors / Journal Geochemical Exploration 71 (2000) 1–12

Table 5
Leachability of magnesium from chlorite, biotite and hornblende

Mineral Mg Leachability (% of total wt%) in

Fraction Total (wt%) 1 M HCl, 2 h 4 M HCl, 0.5 h 1 M HNO3 7 M HNO3 SS 028150 Aqua regia

Mg-Chlorite 0.5–2 mm 18.3 41 18 6 80 52 48

Powder 18.3 82 79 66 88 91 81

Fe-Chlorite 30–500 mm 5.8 73 65 57 83 89 73

Powder 5.8 74 69 66 88 88 76
Biotite 0.5–2 mm 4 97 88 25 53 54 82
Powder 4 93 95 89 93 100 91
Hornblende ⬍2 mm 8.5 3.3 2.8 1.6 2.3 3 2.3
⬍63 mm 8.5 17 13 5.3 4.6 11 9.9
Powder 8.5 26 24 9.6 22 28 18

The leachability of larger grain sizes is less effi- than from the powdered fraction, in hydrochloric
cient. For example, when comparing the ⬍2 mm frac- acids (including aqua regia) and when leaching is
tion with the powder fraction of plagioclase, done according to the Swedish Standard (Swedish
leachability of calcium was reduced to about one- Standard SS 028150, 1980). The same was true of
tenth. For the other minerals studied, leachability the weaker nitric acid (1 M HNO3).
was 20–40 times less.
2.4.2.5. Iron and aluminium. These elements were
2.4.2.4. Magnesium. Dark micas, chlorites, various present as major or minor components of most
clay minerals, pyroxenes and amphiboles are the minerals used for the leaching experiments in this
major magnesium-bearing silicate minerals in the study (cf. Table 1). Leachability results, from the
tills and sediments of Scandinavia (Snäll et al., minerals in which the elements are found in higher
1992). In this study, specimens of biotite, chlorite concentrations, are given in Tables 6 and 7. Generally,
and hornblende were leached in the various acids they correspond to those of the previously treated
(cf. Table 5). elements of the individual minerals.
The greatest leachability of magnesium was from
biotite (up to 100% for the powdered fraction) and for 2.4.2.6. Titanium. Sphene, ilmenite, rutile and anatase
the chlorites (up to 91%). Mg leachability from the are the major titanium minerals in the Scandinavian
powder fractions of hornblende lay between 18 and tills and sediments in the Precambrian bedrock areas.
28% in the strongest acids. Titanium is also present in biotite in small amounts.
Leaching from coarser fractions of hornblende was The leaching of biotite and the typical titanium
considerably less effective than leaching from horn- minerals yielded the leachability percentages
blende powder. For example, the ⬍2 mm fraction reported in Table 8. Titanium in biotite dissolves
yielded ten times less magnesium than the powder. completely in the acids, indicating that the element
From biotite, the leachability of magnesium was was present within the structure of the biotite—and
about the same from the coarser as from the powdered not as inclusions of a separate titanium mineral (e.g.
fraction in hydrochloric acids (including aqua regia). rutile needles) within the biotite crystal. When
The pure nitric acids yielded less magnesium from the comparing the results from leaching of typical
coarser fractions. From iron-rich chlorite (Fe- titanium minerals it is evident that most titanium is
chlorite), leachability was about the same from the obtained from sphene—more than 50% (55–63%)
coarser fraction (30–500 mm) as from the powder. of the available titanium is leached by the acids
Magnesium-rich chlorite (Mg-chlorite) yielded less from the powdered material. Leaching of a ⬍63 mm
magnesium from the coarser fraction (0.5–2 mm) fraction yielded 33–42% and from a ⬍2 mm fraction,
 S. Snäll, T. Liljefors / Journal Geochemical Exploration 71 (2000) 1–12 7

Table 6
Leachability of aluminium from microcline, plagioclase, hornblende, epidote, chlorite, biotite and muscovite

Mineral Al Leachability (% of total wt%) in

Fraction Total (wt%) 1 M HCl, 2 h 4 M HCl, 0.5 h 1 M HNO3 7 M HNO3 SS 028150 Aqua regia

Microcline ⬍2 mm 9.9 1.1 0.5 0.4 0.3 0.4 0.3

⬍63 mm 9.9 2.5 1.4 1.0 0.8 1.0 0.5
Powder 9.9 8.2 5.0 3.1 3.1 3.2 2.2
Plagioclase ⬍2 mm 10.8 0.3 0.2 0.3 0.3
⬍63 mm 10.8 0.9 0.7 0.9 0.7
Powder 10.8 3.6 3.6 3.8 3.3
Hornblende ⬍2 mm 3.6 4.3 3.8 1.8 2.8 3.8 3.1
⬍63 mm 3.6 22 16 6.9 5.9 14 13
Powder 3.6 31 28 12 25 31 22

Epidote ⬍2 mm 12.2 0.2 0.4 0.8 0.6

⬍63 mm 12.2 1.0 4.2 4.6 5.2
Powder 12.2 15 30 34 33
Mg-Chlorite 0.5–2 mm 7.8 46 21 8 96 55 61
Powder 7.8 101 96 78 106 103 106
Fe-Chlorite 30–500 mm 6.1 62 50 47 72 78 63
Powder 6.1 65 55 58 80 79 66
Biotite 0.5–2 mm 7.1 101 95 27 57 52 94
Powder 7.1 107 110 100 102 102 104
Muscovite Powder 18.1 6.5 5.6 6.2 8.6 8.7 6.1

Table 7
Leachability of iron from hornblende, epidote, chlorite, biotite and ilmenite

Mineral Fe Leachability (% of total wt%) in

Fraction Total (wt%) 1 M HCl, 2 h 4 M HCl, 0.5 h 1 M HNO3 7 M HNO3 SS 028150 Aqua regia

Hornblende ⬍2 mm 11.4 5.4 6.3 1.3 2.4 3.6 4

⬍63 mm 11.4 21 16 5.1 5.7 13 14
Powder 11.4 34 31 10 29 35 26
Epidote ⬍2 mm 10.4 0.5 1.4 1.6 1.6
⬍63 mm 10.4 1 4.7 5.1 5.5
Powder 10.4 15 31 36 32
Mg-Chlorite 0.5–2 mm 6.0 32 18 6 63 40 50
Powder 6.0 75 74 57 75 76 81
Fe-Chlorite 30–500 mm 10.4 70 64 59 83 89 77
Powder 10.4 73 68 72 87 88 80
Biotite 0,5–2 mm 20.6 104 98 25 54 54 82
Powder 20.6 96 97 90 94 99 90
Ilmenite Powder 9.1 17 n.d. 57
8 S. Snäll, T. Liljefors / Journal Geochemical Exploration 71 (2000) 1–12

Table 8
Leachability of titanium from biotite, sphene, ilmenite, rutile and anatase

Mineral Ti Leachability (% of total wt%) in

Fraction Total (wt%) 1 M HCl, 2 h 4 M HCl, 0.5 h 1 M HNO3 7 M HNO3 SS 028150 Aqua regia

Biotite 0.5–2 mm 1.7 113 110 26 55 60 91

Powder 1.7 105 108 91 96 105 96

Sphene(titanite) ⬍2 mm 22.8 3.3 3 2.7

⬍63 mm 22.8 33 36 42
Powder 22.8 55 63 58
Ilmenite Powder 35.3 15 n.d 63
Rutile Powder 59.2 0.7 1 1.2

Anatase Powder 59.1 0.2 0.4 0.4

2.7–3.3%. From the powdered ilmenite, the however, indicate that all aluminium was leached
leachability of titanium is approximately the same from the mineral by all the acids, except the 1 M
as from sphene (cf. Table 8) if the mineral is nitric acid (78% leachability).
leached in aqua regia, but less in nitric acid (only Among the calcium-bearing silicate minerals
15%). The other titanium-bearing minerals studied, the highest leachability was from sphene.
yielded very little titanium. Leaching of powdered Some 60–70% of the available calcium (and some-
rutile yielded 0.7–1.2% and from anatase, only what less titanium) was leached by the acids used for
0.2–0.4%. this mineral.
In the stronger acids, 30–36% of each analysed
2.4.2.7. Leachability of minerals. To facilitate element was leached from epidote. The weaker acid,
comparison of the difference in leachability of the 1 M HNO3, yielded only 15%.
individual minerals studied, they have been arranged The leachability of elements from hornblende
in order of decreasing leachability from top to bottom, was somewhat lower than from epidote. In hydro-
in Table 9. Leachability values of the major elements chloric acids and the stronger nitric acids the leach-
in the minerals (except silica) are given. To achieve ability of the elements was between 20 and 35%;
the most accurate comparison, only the results of somewhat lower, 14–18%, for sodium, potassium,
leaching from powder fractions are shown. It is calcium and magnesium in aqua regia and for potas-
evident from these results that biotite is the most sium in 7 M HNO3. In the weaker 1 M HNO3 acid,
easily leached of the minerals studied. The powder the leachability of the elements decreased to less
of the biotite specimen appeared to dissolve than half of that in the stronger acids. For potas-
completely in all the acids used in the study. Some sium, leachability was 5–8 times less, compared
elements, however, were not completely leachable with the stronger acids.
(e.g. iron and magnesium); other elements yielded Leachability of elements from the mica minerals
more than 100% (e.g. aluminium and titanium). studied varied widely. As previously shown, biotite
Analytical errors account for these differences. dissolved completely in the acids used in this study,
The leachability of elements from chlorite indi- but from muscovite, only 5.6–8.7% of the available
cates that this mineral specimen did not dissolve elements leached.
completely in all acids. About 80–90% leachability From plagioclase, calcium was more easily leached
was obtained for magnesium and iron in the hydro- than sodium (7.4–10 and 3.8–4.5%, respectively, in
chloric acids and in the strongest nitric acids. 1 M nitric acids and aqua regia). Leachability of aluminium
nitric acid yielded less iron and magnesium, 60– is 3.3–3.8%. From microcline, the same acids yield
70%. The leachability figures for aluminium, rather less, 2.5–3.5% of potassium and 2.2–3.2% of
 S. Snäll, T. Liljefors / Journal Geochemical Exploration 71 (2000) 1–12 9

Table 9
Leachability of major elements from the minerals studied (powdered fraction)

Mineral Element Total conc. (wt%) Leachability (% of total wt%) in:

1 M HCl, 2 h 4 M HCl, 0.5 h 1 M HNO3 7 M HNO3 SS 028150 Aqua regia

Biotite K 7.3 104 107 106 106 109 114

Mg 4.0 93 95 89 93 100 91
Fe 20.6 96 97 90 94 99 90
Ti 1.7 105 108 91 96 105 96
Al 7.1 107 110 100 102 102 104
Mg-Chlorite Mg 18.3 82 79 66 88 91 81
Fe 6.0 75 74 57 75 76 81
Al 7.8 101 96 78 106 103 106
Sphene(titanite) Ca 20.4 71 72 61
Ti 22.8 55 63 58
Ilmenite Fe 17 57
Ti 15 63
Epidote Ca 16.7 15 30 34 30
Fe 10.4 15 31 36 32
Al 12.2 15 30 34 33
Hornblende K 0.27 21 21 2.6 14 20 14
Na 0.84 24 22 7.9 20 27 17
Ca 8.4 27 25 8.8 23 29 18
Mg 8.5 26 24 9.6 22 28 18
Fe 11.4 34 31 10 29 35 26
Al 3.6 31 28 12 25 31 22

Muscovite K 8.4 6.5 5.6 6.2 8.6 8.7 6.1

Al 18.1 6.5 5.6 6.2 8.6 8.7 6.1

Plagioclase Na 7.6 3.8 4.1 4.5 4.5

Ca 0.94 10.0 9.7 7.4 7.8
Al 10.8 3.6 3.6 3.8 3.3
Microcline K 9.0 9.6 6.1 3.2 3.5 3.5 2.5
Al 10.8 8.2 5.0 3.1 3.1 3.2 2.2
Rutile Ti 59.2 0.7 1.0 1.2
Anatase Ti 59.1 0.2 0.4 0.4

aluminium. Hydrochloric acids leaches more of the 3. Discussion

elements, 5.0–9.6%.
More than half of the available titanium is leached This study has shown that factors of considerable
from sphene, in both aqua regia and nitric acid. The uncertainty (low precision and accuracy) may be
same yield of titanium is obtained from ilmenite if present in the per cent figures of leachability. This is
aqua regia is used as leachant, but nitric acid yields demonstrated by minerals that dissolve completely in
only 15% (cf. Table 9). The other titanium minerals acid, e.g. leachability of elements from biotite.
studied yielded very little of the element: from Figures both above and below 100% were obtained
rutile, 0.7–1.2% and from anatase, only 0.2–0.4%, (89–110%, indicating a relative error of 10–15%).
respectively. The way of expressing leachability as a quotient
10 S. Snäll, T. Liljefors / Journal Geochemical Exploration 71 (2000) 1–12

between two determinations, each of them with an minerals were in the range 0.5–2.0 mm. From the
analytical error, is an explanation of the relative less soluble potassium-rich and magnesium-rich
large uncertainties of leachabilty figures. The use of minerals, microcline and hornblende, only 0.2–0.4%
different methods, one for analysis of total content and of the total content was leached by the acids. This
another for analysis of the leached elements, is means that leached potassium from the ⬍2 mm frac-
another factor of uncertainty. tion of the mineral soil reflects the content of biotite
In geochemical mapping programmes and for and illite (muscovite is difficult to dissolve, cf. Tables
geochemical exploration, the ⬍63 mm grain size frac- 2 and 9). Leached magnesium reflects the content of
tion of tills is usually analysed with various chemical chlorite, provided it is corrected for magnesium
methods, total analyses of elements, acid leachable assigned to biotite.
elements, etc. In forestry and in many environmental Leaching of finer fractions, e.g. ⬍63 mm fraction,
studies, the ⬍2 mm fraction is also used. When inter- increased the leachability of potassium from micro-
preting the results it is important to be aware of the cline by a factor in the range 2.5–3 (cf. Table 2) and
fact that the two grain size fractions often have differ- increased the leachability of magnesium from horn-
ent chemical composition (cf. Snäll and Ek, 1996), the blende by a factor of 2–4.3 (cf. Table 5). Leaching of
reason being that the minerals forming the tills are ground samples (powders) increased the leachability
usually enriched in certain grain fractions (‘terminal of these elements some 8.8–12.5-fold and 7.8–9.6-
grade’ of minerals, Dreimanis and Vagners, 1971). fold for microcline and hornblende, respectively.
The terminal grade of individual minerals depends This means that the contrasts in leachability emanat-
partly on the major parent rock of the till (cf. Snäll ing from the different solubilities of the minerals are
and Ek, 1996). increasingly levelled out as the grain size is decreased.
The two-grain size fraction shows differing compo- For aerial presentation of results, on maps of the finer
sitions when subjected to total chemical analyses. grain material (e.g. ⬍63 mm fraction) this is disadvan-
Analyses of acid extractable fractions will also differ, tageous because the pattern of distribution of the
the explanation being that some minerals vary in solu- leachable elements will become increasingly similar
bility in acids. As shown here, biotite dissolved to the pattern of elements from total analyses of
completely in some of the strong acids used. Chlorite, samples.
too, is largely dissolved, but from other minerals only In models, for the calculation of normative miner-
smaller fractions of the elements are leached by the alogy (Räisänen et al., 1995; Snäll and Ek, 1996)
acids. For the less acid-soluble minerals, grain size is some input parameters are the aqua regia fractions
of great importance for the leachability of the of elements. To get the most reliable results from
elements. Leaching of ⬍63 mm fraction (vis-a-vis these calculations it is very important to have a selec-
the ⬍ 2 mm fraction) increases the leachability of tive dissolution of the minerals. The present study has
the base cations from feldspars, hornblende and shown that this will be best fulfilled if acid leaching is
epidote by a factor between 2 and 7. Leaching of applied to the ⬍2 mm fraction (whole matrix) of the
the powder fraction (instead of ⬍2 mm) increases samples. Leaching of finer fractions will lead to an
the leachability of base cations from 8- up to about overestimation of, for example, the biotite content,
40-fold. because a larger proportion of the potassium leach
Thus, for the interpretation of results obtained by from inter alia K-feldspar, if a finer fraction is used.
acid leaching of samples, it is of great importance to The results also show that acids are varyingly able
know what grain size fraction has been used in the to leach elements from minerals. This is obvious from
analyses. The best indication of the mineralogy of leaching of the ilmenite (cf. Table 9). Aqua regia
samples is obtained by leaching the ⬍2 mm fraction. leaches 63% of the available titanium from the
It is clear from Tables 2 and 5 (leachability of potas- mineral, whereas nitric acid leaches only 15%.
sium and magnesium, respectively) that the most However, from the chemical analyses of the ilmenite
easily leached common minerals, biotite and chlorite, (Table 1) it is suspected that the ilmenite specimen
are largely dissolved (biotite completely dissolved) by was not a pure mineral. A TiO2 content of 58.8% is too
some of the acids used, although grain sizes of the high; the maximum TiO2 content of ilmenite is
 S. Snäll, T. Liljefors / Journal Geochemical Exploration 71 (2000) 1–12 11

52.66% (Kostov, 1968). It is therefore assumed that (i.e. analytical error) the following conclusions may
the specimen was contaminated by other titanium be drawn.
minerals, most of which have a very low solubility. Minerals in tills and sediments are more or less
The measured leachability of titanium from ilmenite soluble in the classical strong acids used for leaching
may, for this reason, be slightly too low. of geological samples. Consequently, leachabilities of
The effectiveness of different acids to leach elements differ. Yields of individual elements depend
elements was also observed in some of the other on the overall mineralogy of the samples, chemical
minerals. For example, aqua regia and the hydrochlo- composition of the minerals and their grain size. For
ric acids dissolved biotite more or less completely the minerals leached in strong acids in this study, the
(including the most coarse-grained fraction, cf. following ranking of solubility was obtained for
Table 2), whereas the 7 M nitric acid did not. On fractions of equal size:
the other hand, 7 M nitric acid leached Mg-chlorite
more effectively. The more easily dissolved Fe-chlor- biotite ⬎ chlorite ⬎ sphene ⬎ ilmenite ⬎
ite, however, does not reveal these differences in epidot ⬎ hornblende ⬎ muscovite ⬎ plagioclase
leachability. (albite-oligoclase) ⬎ microcline ⬎ rutile ⬎ anatase
The leachability of the elements also depends on
the duration of leaching. For instance, 2-h leaching in Generally speaking, leachability of the elements is
1 M hydrochloric acid can yield twice as much of an highly dependent on the grain size of the leached
element as a 30-min leaching in 4 M hydrochloric minerals, because surface area increases with decreas-
acid (cf. Tables 2 and 3, potassium and sodium from ing grain size; the finer the fraction of the samples
microcline). used for the analyses, the greater the leachability of
A non-stoichiometric dissolution of biotite and the elements. In the present study, this has been
chlorite has been observed by Malmström et al. demonstrated for a number of minerals.
(1995), but because of the use in this study of much The order of solubility of minerals given above is
stronger acids (biotite and chlorite are more or less correct, given that the minerals have a similar distri-
completely dissolved), this could not be verified in the bution in all grain size fractions. In most geological
present study. A tendency to non-stoichiometric samples, however, this is seldom the case. Some
dissolution may, however, be observed for the less minerals are enriched in certain grain size fractions.
soluble minerals, e.g. for plagioclase. Calcium As a consequence, a less soluble mineral may make a
seems to be more easily leached than other elements larger contribution to the total quantity of a leached
from this mineral. element than does a more soluble mineral, if it is
In natural surficial Quaternary deposits, the chemi- present mainly in the finer fractions.
cal composition of the minerals may differ from these Most geological samples also contain minerals that
pure specimens from mineral collections. The most are completely soluble in the leachants (the strong
easily leachable varieties have disappeared as a result acids). The results of this study show that this is true
of exogenic processes. Mineral grains still present in of biotite in aqua regia and of chlorite (almost comple-
the deposits are coated by a weathering crust and/or tely dissolved). Selective dissolution of these minerals
by precipitated iron and aluminium oxides. Such coat- can thus be used for mineralogical interpretation.
ings reduce the efficiency of leachants when working Potassium is a major element of biotite. K-feldspar
with natural sediments. This means that the applica- and muscovite are other potassium-rich minerals
tion of the results obtained in the present study should commonly present in the geological samples. As the
be treated with caution. Figures for leachability are leachability of potassium from the latter is low (espe-
maximum values for optimal leaching conditions. cially from larger grain sizes) leachable potassium
largely reflects the content of biotite in the samples.
Leachable magnesium may be used to estimate chlor-
4. Conclusions ite content, provided that correction is made for
magnesium bound in biotite. A contribution of
Although our results contain some uncertainties magnesium from pyroxenes and amphiboles may be
12 S. Snäll, T. Liljefors / Journal Geochemical Exploration 71 (2000) 1–12

present but this is relatively low for coarser grain size assessment of inorganic substances in wastes. The Science of
fractions. Therefore leaching of ⬍2 mm instead of the Total Environment 178, 71–84.
Koljonen, T., Malisa, E., 1991. Solubility in aqua regia of selected
much finer fractions (⬍63 mm or powder) is chemical elements occurring in the fine fraction of till. In: Pulk-
suggested if the purpose is an interpretation of miner- inen, E. (Ed.), Environmental Geochemistry in Northern
alogy of the samples. Europe. Geological Survey of Finland. Special Paper, 9, 49–52.
Leaching of powdered samples rather than ⬍2 mm Koljonen, T., Gustavsson, N., Noras, P., Tanskanen, H., 1992. The
fraction increases the leachability of elements consider- Geochemical Atlas of Finland, Part 2: Till. Geological Survey of
Finland, Espoo, 1992, 218pp.
ably, for some elements up to 40-fold. The consequence Kostov, I., 1968. Mineralogy, Oliver and Boyd Ltd, Edinburgh
is that the contrasts produced by the differing solubility (587pp.).
of the minerals are to a large extent evened out. MalmstrMalmström, M., Banwart, S., Duro, L., Wersin, P., Bruno,
J., 1995. Biotite and chlorite weathering at 25⬚C. SKB technical
report 95-01, January 1995, 128pp.
Acknowledgements Räisänen, M.L., Tarvainen, T., Aatos, S., 1995. NORMA—a
program to calculate a normative mineralogy for glacial tills
We would like to thank C.M. Backman and K. Lax and rocks from chemical analysis. GFF 117, 215–224.
Salminen, R. (Ed), 1995. Regional geochemical mapping in Finland
at the Geological Survey of Sweden for their in 1982–1994. Geological survey of Finland, Report of Inves-
comments on this paper. tigation, 130, 47pp.
Sjöström, J., Qvarfort, U., 1992. Long-term changes of soil chem-
istry in central Sweden. Soil Science 154, 450–457.
References Snäll, S., Brusewitz, A.M., Ålinder, C., 1992. Mineralogisk analys
av moräner. SGU (Geological Survey of Sweden) BRAP 92005.
Andersson, M., 1992. Från Falkenberg till Blomstermåla; nuläge Unpublished report, 16pp.
och framtida effekter av syrabelastning. Geological Survey of Snäll, S., Ek, J., 1996. Moräners mineralogi och kemi över vissa
Sweden, Rapp. and medd., 70, 21pp. bergartstyper i Sverige. SGU (Geological Survey of Sweden).
Asklund, A.M., Grundulis, V., Rönnholm, B., 1966. Våtkemisk Unpublished report. 117pp.
analys av silikatbergarter. Geological Survey of Sweden, C Sverdrup, H., Warfvinge, P., 1991. On the geochemistry of chemi-
616, 55pp. cal weathering. In: Rosén, K. (Ed.), Chemical weathering under
Brindley, G.W., 1961. Chlorite minerals. In: Brown, G. (Ed.). The field conditions. Reports in Forest Ecology and Forest Soils 63.
X-ray Identification and Crystal Structures of Clay Minerals, Swedish University of Agricultural Sciences. Uppsala, 1991,
Mineralogical Society, London, pp. 242–296. 79–118.
Deer, W.A., Howie, R.A., Zussman, J., 1962. Rock-forming Miner- Swedish Standard SS 028150, (Svensk Standard SS 028150), 1980.
als, , Non-Silicates, vol. 5, Longmans, London (371pp.). Metal content of water, sludge and sediment determined by
Dreimanis, A., Vagners, U.J., 1971. Bimodal distribution of rock atomic absorption spectroscopy, atomization in flame—General
and mineral fragments in basal tills. In: Goldthwait, R.P. (Ed.). principles and guidelines (Vattenundersökningar—Metallhalt i
Till/a Symposium, Ohio State University Press, Columbus, pp. vatten, slam och sediment—Bestämning med atomabsorptions-
237–250. spektrofotometri i flamma—Allmänna principer och regler).
Fällman, A.M., Aurell, B., 1996. Leaching tests for environmental SIS—Standardiseringskommissionen i Sverige. 8 pp.