Cent. Eur. J. Geosci. • 2014 • 6(2) • 170-181
DOI: 10.2478/s13533-012-0170-3
Central European Journal of Geosciences
SEDMIN - Microsoft ExcelTM spreadsheet for
calculating fine-grained sedimentary rock mineralogy
from bulk geochemical analysis
Research Article
Uwe R. Kackstaetter1∗
1 Department of Earth and Atmospheric Sciences, Metropolitan State University of Denver, Denver, CO 80217, USA
Received 12 October 2013; accepted 26 March 2014
Abstract: Normative mineralogical calculations from bulk geochemistry of sedimentary rocks are problematic because
of variable depositional environments, particle hydraulics and sedimentary source systems. The development
of SEDMIN, a Microsoft ExcelTM spreadsheet solution, is a practical attempt for a computational routine
focusing specifically on smectite, chlorite, kaolinite, illite and the ambiguous sericite within various pelitic
sedimentary lithologies. While in essence a mathematical approach, the use of statistical evaluation of empirical
lithogeochemical data combined with modal analytical procedures yields reasonable geochemical associations,
more precise chemical phases and revised procedural allotment paradigms. Thus, an algorithm using TiO2 as
a key to the normative calculation of kaolinite is proposed. Incorporating additional parameters, such as LOI
(Loss-on-ignition) in conjunction with carbon, sulfur, carbonate and sulfate, provides that clay phases can be
more accurately determined than from bulk oxides alone. Even when presented with atypical sample data, the
spreadsheet solution is able to accurately predict predominant clay minerals. Besides some drawbacks, the likely
benefit from SEDMIN is the incorporation of results in classification norms and diagrams indicative of sedimentary
lithologies. The ”SEDMIN Sedimentary Mineral Calculator.xlsx” spreadsheet can be freely downloaded from
http://earthscienceeducation.net/SEDMINSedimentaryMineralCalculator.xlsx.
Keywords: major elements • geochemistry • normative calculation • sedimentary rocks • fine-grained • clay • mineralogy
© Versita Sp. z o.o.
1.
Introduction
While the normative calculation of aphanitic mineralogies
in igneous systems from bulk geochemical data is
now a well accepted procedure, assessing sedimentary
lithologies in a similar manner is problematic and rarely
attempted. The limitations appear to be obvious. Igneous
∗
mineralogy follows a reasonably predictable pattern
bound by the laws of magma chemistry. In contrast, clastic
sedimentary rocks as a whole exhibit large variations
in depositional environments, particle hydraulics and
sedimentary source systems. Despite many assumed
random probabilities in the mineralogical composition
of sediments, certain premises are commonly applicable.
High analytical results of SiO2 in a sedimentary sample
would most likely point to the presence of quartz.
Increased levels of Al2 O3 may be indicative of clays
and/or feldspars. If elevated amounts of CaO and CO2
E-mail: [email protected]
170
Unauthenticated
Download Date | 6/17/17 8:00 PM
U. R. Kackstaetter
are measured, calcite would be a logical conclusion.
Rosen, Abbyasov, and Tipper [1] indicate that there are
”significant statistical regularities in the mineralogical
compositions of sedimentary rocks, regularities that can
be used to provide pointers to the likely mineralogical
compositions of most of the common types of sedimentary
rock”.
The underlying problem is to quantify such
assumptions using a meaningful algorithm that would be
able to estimate plausible sedimentary mineralogies from
geochemical-analytical results, especially in fine-grained,
clay-bearing rocks.
The development of SEDMIN introduced here was
specific to such clay-bearing samples. This Microsoft
ExcelTM spreadsheet solution was designed particularly
to calculate clay phases within fine-grained sedimentary
lithologies in an attempt to aid in the investigation
of diffusive pollutant transport through selected natural
geologic barriers of Southern Germany [2].
While
in essence a mathematical approach, the use of
the statistical evaluation of empirical lithogeochemical
data combined with modal-analytical procedures yields
reasonable geochemical associations, more precise
chemical phases and revised procedural allotment
paradigms.
An additional parameter-LOI (Loss-onignition), absent in other computational approaches-has
also been incorporated. Together with currently available
routine analysis for C and S, LOI provides valuable data
on hydrated minerals such as hydrated clays and those
phases decomposing at 1,000◦ C temperatures. Thus the
carbonate, sulfate and clay phases can be more accurately
determined than with bulk oxide computations alone.
2.
MATERIAL AND METHODS
2.1.
Lithologic Material
Subsurface claystone samples from drill cores at four
different locations in northern Bavaria, Germany, as
summarized in Table 1, were used to establish the
normative calculation routine.
Varying depositional
environments
qualified
for
enough
lithological
differentiation to make the attempt for normative
calculations meaningful.
Core samples eliminated
adjustments due to chemical alteration from surface
weathering. Selected core segments were subjected to
geochemical whole-rock analysis and optical petrographic
investigation using thin sections. In addition, XRD (X-ray
diffraction) procedures were included to establish the
predominant clay mineralogy using a 2 µm size fraction to
avoid interference with coarser non-clay species. These
analytical base data were used to develop the normative
calculation algorithms discussed below.
Whole rock analysis for major rock forming elements was
performed on core cuttings pulverized to 60-mesh grain
size (0.15 mm). A 200 mg sample split was fused with 1.2
g of LiBO2 at approximately 925◦ C for about 45 minutes.
Loss on ignition (LOI) was also recorded. The resulting
material was then dissolved in 100 ml 5% HNO3 and
analyzed by ICP-MS (inductively coupled plasma mass
spectrometry) for SiO2 , Al2 O3 , Fe2 O3 , MgO, CaO, Na2 O,
K2 O, TiO2 , P2 O5 , MnO, Cr2 O3 , and BaSO4 , as well as for
oxides of Ni, Sr, Zr, Y, Nb, and Sc. In addition, carbon and
sulfur content was examined using the LECOTM method.
Graphite, organic carbon, and CO2 , as well as sulfide and
sulfate sulfur were distinguished.
Two representative samples from each lithologic unit
were subjected to X-ray diffractive (XRD) studies to
ascertain clay mineralogies. Samples were dried and iron
oxide and organic materials removed using the Mehra
and Jackson method [7] and the 10% H2 O2 process,
respectively. Calcareous cement was extracted through
a 0.1 m EDTA (ethylenediaminetetraacetic acid) or an
acetate buffer solution. For quantitative work the material
was segregated into a grain size fraction of smaller than
2 µm.
Expanding clays (e.g. smectites) were identified by
the ethylene glycol solvation method and XRD. A
prepared sample mount was placed for one week into
a desiccator next to a dish of ethylene glycol [8, 9].
Especially smectites show a rather uniform response to
this treatment, yielding an XRD-detectable basal spacing
of ∼ 17 Å. Vermiculite clays are also susceptible to
this procedure but with different resulting spacings of
14.3 to 16.3 Å [8]. Mixed-layered clays can also be
distinguished and quantified by a combination of various
solvation methods, heat treatment and mathematical
approximations [10].
Identification of kaolinite in a mixture with other clay
minerals was accomplished by heat treating the sample
at 550-600◦ C for 1 hour.
This method destroys
the crystallinity through dehydroxylation in nearly all
kaolinites. Comparing XRD patterns before and after
heating indicates a missing basal reflection at 7 Å for
kaolinite clays after the treatment [8]. Problems only arise
in the presence of chlorite with 002 reflection at 7 Å, which
is not effected by heat.
Additional information was obtained by applying IRspectrometry to the identification of mixed clay samples.
The material was combined with KBr (potassium
bromide), pressed into pellets and subjected to IR
investigation.
Quantitative differentiation between
kaolinite, chlorite, and illite is made according to IR
absorption patterns [8, 11]. Kaolinite displays indicative
171
Unauthenticated
Download Date | 6/17/17 8:00 PM
SEDMIN - Microsoft ExcelTM spreadsheet for calculating fine-grained sedimentary rock mineralogy
Table 1.
Selected claystone lithologies of Northern Bavaria. Data modified after Dobner [3], Haarländer [4], Rutte [5], Schwarzmeier [6].
Relative
Stratigraphic
Descriptive
age
Name
Jurassic
Amaltheen Clay
mudshale
siltshale
to uniform marine
Triassic
Feuerletten
mudstone
siltstone
to oscillating deltaic 55 m (11 m)
Creußen
Triassic
Lehrberg Layers
siltstone
siltshale
to fluviomarine
Langenzenn
Triassic
Lower
Röttonsteine
siltstone
siltshale
to limnic - terrestrial 20 m (13 m)
absorption bands at 3695-3700 cm−1 and 3620-3625
cm−1 . Illite is somewhat variable but has a characteristic
maximum at 3625 cm−1 . The greatest variations are found
in the IR-spectra of chlorite.
In addition, total carbonate content was determined
according to DIN 18 129 [12] using a 10% HCl solution
to liberate and assess CO2 . Individual calcite and
dolomite were quantified by XRD. As cross reference,
Ca and Mg ions in a sample leachate were measured
and corresponding dolomite and calcite contents were
calculated. Illite content was calculated by multiplying
the K2 O values from the geochemical analysis with an
empirical factor of 12.6 as determined by Kohler, Heimerl,
and Czurda [10]. Their basic assumptions include that
particle sizes below 2 µm should have geochemical K
exclusively attributed to illite clays.
However, this
assumption was not always congruent with the results
from mineral calculations. Hence the potassium-bearing
mineral sericite, a coarser grained alternative to illite,
was added to mitigate excess K2 O and Al2 O3 during
calculation, which worked remarkably well.
Kohler,
Heimerl, and Czurda [10] also give the following equation
to estimate illite content from a mixture of illite, kaolinite,
chlorite and montmorillonite in percent using XRD
patterns:
%Illite =
100 · 1.0Aillite
,
(1.0Aillite · 0.24Akaol · 1.07Achlorite · 0.22Amont )
where A = planimetric intensity.
A good approximation of A follows the peak height
(intensity) multiplied by half of the peak width. The
numerics given in the equation are peak correction factors
of intensities established by Tributh [9] and Laves and
Jähn [13].
Three representative samples from each drill core were
selected for point counting analysis from thin sections.
Depositional
approx. thickness Location
Environment
(coring depth)
40 m (9 m)
35 m (14 m)
Kalchreuth
Marktheidenfeld
The material was vacuum impregnated with blue resin to
contrast pore spaces and voids. Specimens were then
cut parallel to the coring direction using oil to avoid
dissolution and leaching and sections were ground to a
standard thickness of 0.03 mm. During point counting
procedures all materials too small to be distinguished
(usually particles ¡ 0.03 mm) were allocated as clay.
Discolored reddish, brown to dark-brown fine-grained
material was interpreted as being iron-oxide-stained and
thus was further subdivided into Fe-stained clay. While
other iron minerals are plausible, intense red staining
was allotted as hematite because of appearance and
prevalence in sedimentary systems. The term sericite was
used for all identifiable phyllosilicates resembling mica
grains. It was attempted to resolve carbonate mineralogy
from thin sections without staining techniques. Calcite
is often coarser with fewer inclusions, while dolomite is
finer, showing more inclusions and frequently changes
relief when rotated under plain polarized light. Since
this determinative method is weak, mistaken identities
are acknowledged and accepted.
Therefore X-ray
determinative techniques were preferred over thin section
analysis in order to distinguish the main carbonates.
2.2.
Normative Calculation Method
The suite of minerals included in the computational
method was selected as follows. Minerals definitively
recognized through x-ray and optical determinative
methods were quartz, kaolinite, chlorite, illite, swelling
clays (smectites), dolomite and calcite. Additionally,
minor minerals identified through thin sections and
optical microscopy were fine grained muscovite or sericite,
hematite (blood red staining with or without opaque
core), K-feldspar (visible tartan twinning), and minor
apatite. Muscovite/sericite were also shown by Dobner [3]
to be present in the Lower Röttone, Lehrberg Layers
172
Unauthenticated
Download Date | 6/17/17 8:00 PM
U. R. Kackstaetter
and Amaltheen Clay samples. Other minerals, while
not directly identified, were assumed because of their
common occurrence in sedimentary rocks as indicated
through geochemical associations. The calculation model
followed the suggestions of early pioneers in mineral
calculations like Imbrie and Poldervaart [14] listing pyrite,
gypsum, rutile, albite and ferrodolomite as candidates.
The presence of pyrite and gypsum in some of the
samples could be assumed because of elevated total
sulfur and sulfide sulfur values.
Furthermore, both
minerals were mentioned by Dobner [3] for the Lower
Röttone and Amaltheen Clay. Rutile and albite could
be ascertained because of geochemical associations with
measured TiO2 and Na2 O, respectively. Ferrodolomite
is probable when carbonate minerals and Fe2 O3 coexist
in sufficient quantities and was included as such into
the computational algorithm. The term ferrodolomite is
used to describe any Fe- and Mg-containing carbonate,
collectively.
Certain assumptions needed to be made in order to
calculate mineralogy from measured elemental oxide
constituents. Simplified ideal compositions for each
mineral listed in oxide format were established. Care
was taken to find relationships between minerals to be
calculated and the minor oxides indicated. The results
are summarized in Table 2.
While most of the idealized compositions are
straightforward and can be derived directly from the
respective chemical formulas of the minerals, clay
mineralogy is more complex. In order to accomplish
the most truthful compositional representation, the
empirical chemical formula was used as a base. Whole
rock geochemical analysis both published [15–17] and
measured was then used to allot oxide mole fractions in
such a manner that the summative molecular weight from
these respective mole fractions corresponded closely with
the accepted average molecular weight of the mineral.
The value for H2 O in the representative oxide formulas
was deduced from published values and the LOI data
established during the geochemical analysis. Table 3
compares calculated versus true molecular weights of the
5 assessed representative clay minerals.
Greater details for deriving the idealized composition
of the above clays will be given in the discussion of
mineral calculation procedures. A step by step outline
of the procedures is summarized within the SEDMIN
spreadsheet program on the incorporated PROC tab. The
usual approach requires one to find a rock forming oxide
with a unique association to a certain mineral. For
example, P2 O5 can be clearly identified with the mineral
apatite (3 CaO - P2 O5 ), since no other mineral is affiliated
with the same phosphorus oxide. Oxides shared in a
great variety of minerals, such as SiO2 or Al2 O3 , are less
helpful in determining mineral assemblages. In general
equation (1) is used to compute the percentage of a certain
mineral from rock forming oxide data.
%min =
%xO
,
(MWxO · MWmin · mfxOmin )
(1)
where %min = percentage of calculated mineral, %xO =
percentage of rock forming oxide associated with mineral
of interest, MWxO = molecular weight of rock forming
oxide, MWmin = molecular weight of mineral of interest,
and mf xOmin = mole fraction of rock forming oxide in the
idealized mineral formula.
Step 1: Calculating percent Kaolinite using TiO2
Titanium correlates well with Al2 O3 , as indicated by
Pearson statistics and scatter plots using 2 µm kaolinite
concentrations (Figure 1). Correns and Tillmanns [18]
state that some kaolinites exhibit significant TiO2 contents
with Weaver and Pollard [19] giving a TiO2 range of 0.41%
to 2.48% and an average content of 1.43%. This value is
substantiated by leaching experiments from Dolcater et
al. [20] showing 1.5% Ti. Thus kaolinite contains one of
the highest titanium concentrations of the sheet silicates.
Weaver [21] states that most of the TiO2 in kaolinite
is in the form of 0.1 µm anatase pellets. Additional
impurities such as Fe2 O3 and MgO may also be present.
Rengasamy [22] gives the structural formula of a Georgiatype commercial kaolinite after removal of all impurities
as Si4.04 Al3.78 Fe0.08 Ti0.10 O10 (OH)8 . Thus, using TiO2 for
calculating percentages of kaolinite appears to be a valid
approach. The indicated idealized kaolinite formula of
2 SiO2 -Al2 O3 -0.05 TiO2 -2 H2 O is proportionate to the
data given by Rengasamy [22]. For simplification the trace
of iron oxide occasionally present in the kaolinite is not
considered because of the small quantities.
By using a scatterplot of TiO2 versus kaolinite for the
sample suite smaller than 2 µm, certain relationships are
evident (see Figure 1). Samples containing no kaolinite
consist of up to 0.63% TiO2 . Titanium oxide concentrations
of 0.64% or more show a second order polynomial
relationship with kaolinite. Regression analysis and curve
fitting give an intersection of calculated and observed
graphs at 0.639% TiO2 . This value serves as an estimated
cut-off between samples containing kaolinite and those
lacking the silicate in mineral calculations. Instead of
inferring the 0.639% TiO2 cut-off to the larger whole
rock assemblage outside the 2 µm range, a 0.82% cutoff was computed using total clay percentages from point
counting. Thus the differentiation of kaolinite versus rutile
is set at 0.82% TiO2 and applied to the calculation by
subtracting this value from the analytical measurement of
173
Unauthenticated
Download Date | 6/17/17 8:00 PM
SEDMIN - Microsoft ExcelTM spreadsheet for calculating fine-grained sedimentary rock mineralogy
Table 2.
Idealized chemical compositions for minerals to be calculated. Oxide formulas for the more complex phyllosilicates are simplified as
representative for the group rather than individual specific minerals (e.g. chlorite, mont./smectite).
Quartz
SiO2
Calcite
CaO - CO2
Gypsum
CaO - SO3 - 2 H2 O
Dolomite
CaO - MgO - 2CO2
Pyrite
FeS2 (Fe2 O3 ×0.6994 = Fe) Ferrodolomite
CaO - 0.5 Fe2 O3 - 2 CO2
Apatite
3 CaO - P2 O5
Illite
3.7 SiO2 - 0.7 Al2 O3 - 0.1 Fe2 O3 - 0.3 MgO -0.3 K2 O - 2.7 H2 O
Albite
Na2 O - Al2 O3 - 6 SiO2
Sericite
3 SiO2 - 1.5 Al2 O3 - 0.5 K2 O - 1 H2 O
K-spar
3 SiO2 - 0.5 Al2 O3 - 0.5 K2 O Chlorite
Rutile
TiO2
Mont./Smectite 4 SiO2 - 1Al2 O3 - 0.1 Na2 O - 0.1 CaO - 10.9 H2 O
Hematite
Fe2 O3
Kaolinite
Table 3.
3 SiO2 - 1Al2 O3 - 0.6 Fe2 O3 - 3.7 MgO - 3.9 H2 O
2 SiO2 - 1Al2 O3 - 0.05 TiO2 - 2 H2 O
Comparison of calculated vs. standard molecular weights of clays used in calculating sedimentary minerals.
Calculated
Standard
Clay
Idealized Composition
Illite
3.7 SiO2 - 0.7 Al2 O3 - 0.1 Fe2 O3 - 0.3 MgO -0.3 K2 O - 2.7 H2 O 398.64
389.341
Sericite
3 SiO2 - 1.5 Al2 O3 - 0.5 K2 O - 1H2 O
398.30
398.712
Chlorite
3 SiO2 - 1Al2 O3 - 0.6 Fe2 O3 - 3.7 MgO - 3.9 H2 O
597.41
595.22
Mont./Smectite
4 SiO2 - 1Al2 O3 - 0.1 Na2 O - 0.1 CaO - 10.9 H2 O
550.46
549.073
Kaolinite
2 SiO2 - 1Al2 O3 - 0.05 TiO2 - 2 H2 O
262.15
258.163
Molecular Weight
Molecular Weight
3
Sources for molecular weight and whole rock geochemistries: 1 Gaines et al. [15]; 2 O’Donoghue [16]; and 3 Duda and Rejl [17],
who used Clinochlore for Chlorite and Montmorillonite for Mont./Smectite
The amount of other oxides used in the calculation of
kaolinite, such as SiO2 , Al2 O3 , and H2 O are computed as
generalized by equation (2). These other oxide quantities
are tracked in a tally within the SEDMIN spreadsheet.
Water is allotted to LOI.
%xO = mfxOmin · %min · MWxO · MWmin ,
(2)
where %xO = percentage of rock forming oxide associated
with mineral of interest, mfxOmin = mole fraction of rock
forming oxide in the idealized mineral formula, %min =
percentage of calculated mineral, MWxO = molecular
weight of rock forming oxide, and MWmin = molecular
weight of mineral of interest.
Figure 1.
Scatterplot of TiO2 vs. Kaolinite concentrations in
samples <2 µm. Second order polynominal curve for TiO2
concentrations of 0.64% or greater calculated as: [Kaol] =
233.5 + (-617.9)·[TiO2 ] + 442.9·[TiO2 ]2 .
titanium oxide. If the result is less than zero, no kaolinite
is believed to be present and all TiO2 is allotted to rutile.
If the result is greater than zero, 0.82% TiO2 is assigned
to rutile and the remainder is calculated as kaolinite.
Step 2: Establishing percent Rutile using TiO2
Since no other mineral of interest contains TiO2 , all
titanium oxide below 0.82% is calculated as rutile as
explained in Step 1. Allotment of TiO2 is now complete.
Step 3: Calculating percent Apatite from P2 O5
All of the phosphorous oxide is assigned to apatite. The
remaining CaO is tracked according to equation (2) in the
tally segment of the SEDMIN spreadsheet.
174
Unauthenticated
Download Date | 6/17/17 8:00 PM
U. R. Kackstaetter
Step 4: Computing Gypsum from S if data are available
Sulfur analysis by LECOTM was performed for several
samples of Feuerletten, Amaltheen, Rötton and Lehrberg
Layer clays yielding values for total S, sulfide S and SO3 .
Gypsum is calculated using SO3 and the idealized formula
CaO-SO3 -2 H2 O, allotting all SO3 . The remaining CaO
and H2 O are again tallied as described above.
Step 5: Calculating Pyrite from S if data are available
Analytical results for sulfur are applied to the calculation
of pyrite, thus S is now entirely distributed. Since
Fe2 O3 is not used in the idealized chemical formula for
pyrite, Fe in the iron sulfide is converted to Fe2 O3 by
FePyrite/0.6994 for tracking purposes.
Step 6: Figuring Smectite Clay from Na2 O
Only two minerals of interest show Na2 O in their idealized
chemical formula, smectite (4 SiO2 -Al2 O3 -0.1 Na2 O 0.1
CaO-10.9 H2 O) and albite (Na2 O-Al2 O3 -6 SiO2 ). In
order to differentiate between sodium oxide in these two
minerals, some basic generalizations are made. It is
assumed that all smectite resides in the ¡2 µm fraction
while albite likely exists in the larger grain sizes. By
knowing the percent of clay in the total sample and the
amount of smectite within the 2 µm clay fraction, the
amount of Na2 O needed to satisfy the smectite mineralogy
can be estimated. The remainder of the sodium oxide
concentration would then reside in albite. In this manner
probable Na2 O distributions for albite and smectite
can be estimated and used in a regression analysis.
Figure 2 plots the association of sodium oxide to smectite
concentrations in whole rock sample. The observed
correlation shows two distinctively different relationships
for Na2 O concentrations above and below 0.158%, thus
necessitating two distinct equations (equation (3) and
equation (4)) to determine smectite concentrations from
Na2 O amounts.
%Smectite = − 3.7133 + (101.719 · %Na2 O)
− (283.92 · (%Na2 O)2 )
(3)
If % Na2 O ¿ 0.158%.
%Smectite = − 281.72 + (4636.26 · %Na2 O)
− (17840 · (%Na2 O)2 )
(4)
If % Na2 O ¡ 0.158%.
After calculating smectite, the other individual oxide
concentrations inherent to the clay are computed and
tallied. Any remaining Na2 O is assigned to albite.
Figure 2.
Relationship of Na2 O vs. smectite concentrations with
calculated curves and associating formulas for mineral
concentration modeling.
Step 7: Calculating Albite from remaining Na2 O
The persisting Na2 O from Step 6 is now used to solve for
percent albite. If no Na2 O is remaining or has a negative
number then albite is equal to zero. Step 7 completely
allots all sodium oxide.
Step 8: Computing Calcite using CaO and CO2
The relationships of CaO, MgO, and CO2 to specific
carbonate mineral concentration and computing details
in sedimentary rocks are well established by Imbrie and
Poldervaart [14]. Their advice for calculating procedures
has been adopted as follows:
Moles of calcium oxide remaining after allowing for
association in gypsum, apatite, and smectite are
subtracted from the molar proportion of CO2 . The balance
represents the amount of dolomite necessary to use up
any remaining CO2 . By subtracting this balance from the
available calcium oxide, the molar proportion of calcite is
determined. When total MgO and CaO are insufficient
to use up CO2 , ferrodolomite is calculated by subtracting
the MgO molecular weight ratio from the above-mentioned
balance and multiplying the answer by the molecular
weight of ferrodolomite.
The key for precise calculation of carbonates, such as
calcite, lies in the accurate determination of the CO2
concentration, which can be challenging. The employed
LECOTM and LOI methods proved to be insufficiently
accurate to yield meaningful results in several samples,
especially those of greater carbonate latitudes. While
the greatest agreement was found in most cases when
175
Unauthenticated
Download Date | 6/17/17 8:00 PM
SEDMIN - Microsoft ExcelTM spreadsheet for calculating fine-grained sedimentary rock mineralogy
LECOTM data were used, the results from mineral
calculations were often still erroneous. Therefore a
modified approach is employed.
Since the amount of CO2 used is tallied during
mineral calculations, the difference between total carbon
dioxide available and allotted should give a measure of
computational reliability. The greater the difference, the
more erroneous the results. If, however, tallied CO2
zeroes out the available carbon dioxide, the computational
results are most believable. To achieve the state of
desired accuracy despite variations in CO2 measurements,
the complete mineral calculations package must be
established first. Initially calculations are started using
measured carbon dioxide values. If discrepancies between
measured and tallied CO2 are observed, the measured (i.e.
input) data are adjusted until the resulting calculations
yield the smallest possible difference between input and
tallied values. In most cases the difference between these
shows zero (or at most very small) discrepancies. Thus
carbon dioxide concentrations established through mineral
calculation appear to be more reliable for particular
sample suites than the measured values.
Step 9: Calculating Dolomite and Ferrodolomite from MgO
First the MgO mole fraction is determined by dividing the
molecular weight of magnesium oxide into percent MgO.
The oxide values allotted for dolomite in Step 8 are now
subtracted from the MgO mole fraction. If the difference is
smaller than zero, ferrodolomite is calculated as indicated
above. Percent dolomite is now computed by utilizing any
remaining CaO, dividing it by the molecular weight for
calcium oxide and multiplying the answer by the molecular
weight of dolomite. Any remaining MgO is assigned to
chlorite and illite. Both CO2 and CaO are now completely
allotted.
Figure 3.
Relationship of K2 O and MgO concentrations to whole
rock illite content showing observed and calculated
values. MgO curve to be used if 3MRK2 O /2MRMgO < 1;
otherwise K2 O computed as illite.
Using this approach unmodified yielded only partial
satisfactory results for the computation of clays in varied
samples. However, by combining the ratio method present
in equation (5) with a regression analysis involving
MgO and K2 O and whole rock illite concentrations (see
Figure 3), the following preliminary algorithms were
established, indicated in equation (6) and equation (7).
%Illite = −18.095 + (10.0391 · %MgO) − (0.7462 · (%MgO)2 )
(6)
If 3 MRK2 O/2 MRMgO¡1.
Step 10: Calculation of percent Illite from K2 O and MgO data
Imbrie and Poldervaart [14] suggested the use of
a potassium/magnesium mole ratio to establish
relationships between illite and sericite.
First the
mole ratios for potassium oxide (MRK2 O ) are established
by taking percent K2 O and dividing it by the molecular
weight of the same oxide. Secondly, moles of MgO
(MRMgO ) are determined by applying the equivalent
computation to percent MgO remaining after carbonate
computations. A ratio is then established as indicated by
equation (5).
3MRK2 O
2MRMgO
(5)
If this ratio equals or exceeds 1, all of the MgO is
calculated as illite and the balance of K2 O as sericite.
%Illite = −40.777 + (20.5364 · %K2 O) − (1.7464 · (%K2 O)2 )
(7)
If 3 MRK2 O /2 MRMgO ≥ 1.
If the before-mentioned ratio is smaller than 1, percent
illite can be satisfactorily calculated using percent MgO
as demonstrated with equation (6). Otherwise equation (7)
and percent K2 O are to be used.
Step 11: Compute Chlorite from remaining MgO
Any remaining MgO from previous calculations is now
allotted to the clay mineral chlorite, according to
the idealized formula 3 SiO2 -Al2 O3 -0.6 Fe2 O3 -3.7 MgO3.9 H2 O. If tallied magnesium oxide is less than zero, no
chlorite is computed. MgO is now completely assigned.
176
Unauthenticated
Download Date | 6/17/17 8:00 PM
U. R. Kackstaetter
Step 12: Estimating percent Sericite and Potassium Feldspar
(orthoclase, microcline) from remaining K2 O and Al2 O3
Oxides of aluminum and potassium are very common
constituents of many minerals. While most of the oxides
have been allotted at this point, Al2 O3 and K2 O are major
building blocks in sericite and potassium feldspar, the
last remaining minerals to be calculated. Both minerals
have been observed during thin-section and point-count
analysis to various degrees with certainty. In order to
see which of the two minerals is predominant and what
oxide quantity is to be assigned to each mineral, the
following approach is used. According to their idealized
chemical make-up, sericite and K-spar exhibit different
molecular ratios of Al2 O3 to K2 O. For sericite this ratio
equals 3.247, while for K-feldspar it is equivalent to 1.082,
as seen in Table 4. For Step 12, the remaining Al2 O3
from the previous mineral calculations is divided by the
remaining K2 O. When comparing the resulting ratios to
those for sericite and potassium feldspar, the following is
observed. Values close to 3.2 or higher are most likely
associated with predominating sericite with little or no
K-spar present. Ratios of 1.1 or smaller would point
to potassium feldspar as the main mineral. A sliding
scale of allotment of oxides can be established for values
between 1.1 and 3.2 as represented by equation (8) and
equation (9).
%K2 Osericite = 0.01(%K2 O·(46.1926·Al2 O3 /K2 O−50)) (8)
K2 O to be used in sericite calculations.
%K2 OKspar = 0.01(%K2 O · (−46.1926 · Al2 O3 /K2 O + 150))
(9)
K2 O to be used in potassium feldspar calculations.
If the Al2 O3 /K2 O ratio for remaining oxides is greater than
3.247, 100% K2 O is calculated as sericite. For ratios of
1.082 or below, Al2 O3 is completely assigned to K-spar.
Intermediate values between 3.247 and 1.082 necessitate
the remaining potassium oxide to be split between sericite
and Kspar according to equation (6) and equation (7).
The amount of both minerals is then calculated and all
K2 O should now be allotted except for ratios smaller than
1.082.
Step 13: Remaining Fe2 O3 assigned to Hematite
In the last and final step, the remaining iron (III) oxide
is assigned to hematite. Since the idealized formula is
the same as the oxide formula, no special calculation
is necessary. All oxides analyzed during whole rock
geochemical investigation are utilized during mineral
computations except for two rock forming oxides, Cr2 O3
and MnO. Their relative minor amounts do not fit in
any particular idealized formula for the mineral suite
observed. Manganese oxide may be most likely present
as a dark mineral stain not easily distinguished from
iron oxides during optical mineralogy. Chromium, on the
other hand, could occur in small amounts as very stable
detrital chromite in the heavy mineral fraction of some
sedimentary materials. Levinson [23] also reports Cr as
possible substitution in micas or clay minerals.
2.3. SEDMIN - Description and use of the
spreadsheet
SEDMIN is a prepopulated Microsoft ExcelTM
spreadsheet that contains the introduced mathematical
algorithms and macros for calculating mineralogies in
fine-grained sedimentary systems.
Freely download
the ”SEDMIN Sedimentary Mineral Calculator.xlsx”
spreadsheet
from
http://earthscienceeducation.net/
SEDMINSedimentaryMineralCalculator.xlsx.
After opening the file in Microsoft ExcelTM several tabs
are present:
DATA tab: This sheet is the geochemical data input
and mineralogy output pane. Sample values are present
and should be overwritten by typing or pasting your
own values into the pink shaded area. Other fields are
protected and will not allow user input. Calculation
is automatic as new values are added. The fields ”%
Clay (calc.)”, ”SO3 (calc.)”, and ”CO2 (calc.)” within the
input column are computed automatically. This additional
information may be used for other applications as desired.
The pink input for ”CO2 ” can be populated with externally
measured data or the value from the computed ”CO2
(calc.)” field can be entered. Pink areas without data
must be filled with zeros.
CLAY tab: This panel houses the algorithm to compute the
minerals kaolinite, smectite, illite, chlorite and sericite.
The latter is indicated as undifferentiated phyllosilicate
resembling fine-grained muscovite. Calculated values are
automatically transferred to the front DATA panel.
MIN tab: Amounts for the carbonates calcite, dolomite,
and ferrodolomite, minor mineral constituents of gypsum,
pyrite, apatite, albite, rutile and hematite as well as
potassium feldspar and quartz are solved here. Sheet and
field references are given in addition to the procedural
sequence reference explained in detail in the PROC tab.
Computed values are again transferred to the front DATA
panel.
D tab: This segment is left blank for future use and is
currently not part of the computation.
PROC tab: The displayed outline details the procedural
chronological sequence of the computation. Steps and
sub-steps are numbered accordingly. This list is for
177
Unauthenticated
Download Date | 6/17/17 8:00 PM
SEDMIN - Microsoft ExcelTM spreadsheet for calculating fine-grained sedimentary rock mineralogy
Table 4.
Mole ratio comparison of Al2 O3 and K2 O in Sericite and K-spar. (MW = Molecular Weight).
MW Al2 O3
Mineral
Idealized Formula
MW Al2 O3
MW K2 O
in mineral
MW K2 O
in mineral
MW
Al2 O3 (mineral) /
MW K2 O(mineral)
Sericite
3 SiO2 - 1.5 Al2 O3 - 0.5 K2 O - 1H2 O
K-Feldspar
3 SiO2 - 0.5 Al2 O3 - 0.5 K2 O
101.9600
94.1956
152.9400
47.0978
3.2473
50.9800
47.0978
1.0824
reference only and has no bearing on the computational
algorithm of SEDMIN.
TALLY tab: Here, the bulk geochemical values are
assigned to the respective minerals according to a
predefined algorithm.
All other panes requiring
computation rely on the outcome of this sheet.
CLAYCOMPARE tab: An additional output pane showing
distribution of clay minerals and generating pie graphs of
the sedimentary mineralogy for export. This tab is not
required for computation.
Advanced users wishing to modify sheets and associated
computational algorithms or the presented input or output
may do so by turning off the protection for selected tabs
under FILE in Microsoft ExcelTM 2010. Once disabled,
changes can be made.
3.
Discussion
In order to validate the results of the mineral calculations,
several procedures were employed. A simple and rapid
validation method is the summation of minerals calculated
in each sample. Values should equal 100 percent if
the calculation error is zero and all oxides from the
geochemical survey are allotted with no remainders.
Deviation from this target are construed as percent error in
the overall calculation algorithm. Results are summarized
in Figure 4 which displays the 100% accuracy as a red line
and the calculated mineral totals as bars approaching this
target. As indicated, the Lower Rötton, Lehrberg Layer,
and Feuerletten samples fall within ±5% of the anticipated
100% mark.
The Amaltheen clay, however, falls almost 10% short in
samples used from the upper and middle segment of the
drill core, with lower core sample showing a 7% error.
The reason for this discrepancy becomes obvious when
the TALLY sheet in SEDMIN is summoned. The two
samples with the greatest error still show a remainder
of Al2 O3 between 6.5% and 7.5% after calculation and
allotment of all other oxides. The possibility therefore
exists that (a) either some of the clay minerals within
the Amaltheen lithology have a slightly different structural
Figure 4.
Quick check for accuracy of method. Sum of calculated
minerals from geochemical oxides in upper, central, and
lower area of drill cores for respective stratigraphic units.
Hundred percent mark (red) indicates all oxides precisely
allotted with calculated mineral quantities adding up to
exactly 100%.
formula, containing more Al2 O3 than indicated, or (b) an
additional unidentified mineral consisting predominantly
of aluminum oxide, such as diaspore, might be present.
However, adjusting calculating parameters to satisfy
Al2 O3 allotment gives erroneous results with other
mineralogies. Since the deviation is limited to 10% and
general mineralogical results are plausible, no further
attempt was made to correct for this discrepancy.
To show accuracy between actual minerals quantitatively
identified and respective individual minerals calculated, a
bivariate Pearson’s Correlation analysis was performed.
Established XRD mineral quantities were matched
against computed minerals as indicated in Figure 5.
Most calculated mineral species plot close or on the
100% correlation diagonal target line in the graph,
demonstrating a significant correlation with measured
XRD results.
Stray chlorite values can be traced
to excess of MgO without significant presence of
dolomite. Additional computed correlation coefficients
place kaolinite at 99.2% significant correlation followed
by swelling clays (montmorillonites/smectites) with 90.5%.
The lower correlation for smectite clay concentrations of
178
Unauthenticated
Download Date | 6/17/17 8:00 PM
U. R. Kackstaetter
sericite + illite compared to XRD-measured illite are to
be expected.
Figure 5.
Plot of measured XRD values vs. calculated mineral
percentages for selected clay minerals and quartz.
Illite and sericite combined. Diagonal indicates 100%
correlation.
less than 10% to computed values may be a limitation of
the Na2 O differentiation between albite and smectite, as
the oxide is used to derive both mineral concentrations.
While the assumption is made that the clay usually
resides in the ¡2 µm fraction of the sample and albite
is allotted to the coarser portion, a decrease in smectite
may reveal an inflated drift toward albite, thus falsifying
the results. Doubt is also indicated by an apparent
calculation limit of approx. 5% concentration for smectite.
Smectite concentrations are presumed higher in many fine
grained samples, and no resolution was attempted during
this study.
Illite with a lower correlation at the 83.2 percentile
deserves special consideration because of association
with sericite.
Folk [24] describes sericite as fine
grained muscovite, somewhat coarser than illite. Both
minerals show extreme similarities and are impossible
to distinguish by optical methods. X-ray diffractive
techniques are also inept when identifying illite versus
sericite, except the latter has moderately sharper peaks.
While sericite was omitted from the XRD analysis in
favor of illite, the geochemistry of the lithologic materials
pointed to the presence of at least some sericite and
potassium feldspar in selected samples. Also, high
residual amounts of Al2 O3 and K2 O are encountered
during calculations if sericite is not included. Because of
their tight resemblance, both minerals are grouped when
plotted in Figure 5. Hence certain deviations of calculated
Correspondence in the carbonate mineralogy can be found
in the correlation of calculated results versus point count
data from thin sections. Strangely, computed calcite
corresponds 100% with point-counted dolomite at a high
statistical significance. This could be the result of a
mistaken identity during point count analysis, allotting
dolomite instead of calcite, since no distinguishing
staining techniques were employed, as explained above.
However, only one investigated sample showed dolomite
as well as calcite during thin-section analysis, based
on appearance under the optical microscope, while
mineral calculations indicated solely calcite. A plausible
explanation for the discrepancy would be, again, a
mistaken identity, or a change in mineralogy on a small
scale. Note that while the thin section may indeed exhibit
dolomite as the main carbonate, the sample processed for
geochemical analysis, even though only centimeters away,
could truly contain predominantly calcite.
Other interesting verifying correlations are found in the
minor mineral assemblages. Potassium feldspar matches
to 79% with illite identified during the XRD investigation
which is congruent with similar trends described by
Kohler, Heimerl, and Czurda [10]. In addition, point
counted K-feldspar from thin sections matches to 81.4%
with calculated mineral amounts.
4.
Conclusion
SEDMIN performs reasonably well within the latitude
of the investigated fine-grained sedimentary lithologies
from Bavaria, Germany, as indicated. Most clay minerals
demonstrate a substantial correlation in their calculation
when compared to XRD analysis. Trial runs with data
from other localities around the world show promise, but
have not been specifically evaluated.
Moreover, SEDMIN displays some ability in handling
unusual sedimentary geochemistry, such as exotic soils
found above kimberlites.
The prepopulated data
example in the downloadable SEDMIN calculator consists
of geochemical data from heavily weathered material
above a kimberlite pipe in northern Colorado. While
unconsolidated soil does not necessarily qualify as
sedimentary lithology, nevertheless the software was able
to predict the two predominant clay minerals in this
atypical sample with remarkable accuracy. The computed
chlorite matches the XRD value by 97.5%, while SEDMIN
calculated the expected illite as sericite corresponding
99.1% between measured and calculated values. Keeping
in mind that SEDMIN was not written for soils with such
179
Unauthenticated
Download Date | 6/17/17 8:00 PM
SEDMIN - Microsoft ExcelTM spreadsheet for calculating fine-grained sedimentary rock mineralogy
exotic geochemistries, it is encouraging that the program
fared well in this challenge.
The SEDMIN spreadsheet algorithm is by no means
perfect and has drawbacks, notably when unusual high
amounts of Al2 O3 or carbonates are present. The software
also appears to fall short when predicting smectites at
low concentrations (¡10%). Testing with larger data
sets from other areas will most likely uncover additional
weaknesses in the present approach. Yet several workable
relationships were indicated during this study and have
been incorporated into SEDMIN. The possibility of an
universally applicable computation of kaolinite from the
presence of TiO2 should be further investigated.
The strength of the approach introduced here lies in
computing smectite, chlorite, kaolinite, illite and the
ambiguous sericite from a variety of pelitic sedimentary
lithologies. Users are encouraged to modify the software
in order to make the calculation more applicable to their
specific needs and to identify limitations. The most
likely and progressive benefit from SEDMIN may be the
incorporation of the results in classification norms and
diagrams indicative of sedimentary lithologies.
References
[1] Rosen O. M., Abbyasov A. A., Tipper J. C., MINLITHan Experience-based Algorithm for Estimating the
Likely Mineralogical Compositions of Sedimentary
Rocks from Bulk Chemical Analyses, Computers and
Geosciences, 30(6), 2004, 647-661, http://dx.doi.org/
10.1016/j.cageo.2004.03.011
[2] Kackstaetter U. R., Contaminant Diffusion and
Sorption of an Artificial Leachate in Selected
Geologic Barriers of Frankonia, Bavaria, Germany,
PhD Thesis, Universitätsbibliothek der Universität
Würzburg,
2005,
http://www.opus-bayern.de/
uni-wuerzburg/volltexte/2005/1615/
[3] Dobner A., Tone, Mergel, Lehme. In: Oberflächennahe
mineralische Rohstoffe von Bayern, Clays, Marls,
Loams. In: Near surface mineral resources of Bavaria,
Geol. Bavarica, GLA, Munich, Germany, 86, 1984,
441-494
[4] Haarländer W., Geologische Karte von Bayern
1:25000, Erläuterungen zum Blatt 6432 ErlangenSüd, Geologic map of Bavaria 1:25000, Explanation
for sheet 6432 South Erlangen, GLA, Munich,
Germany, 1966
[5] Rutte E. Einführung in die Geologie von
Unterfranken, Introduction to the geology of
lower Franconia, Würzburg: Laborarztverl., 1957,
Print
[6] Schwarzmeier J., Geologische Karte von Bayer,
Erläuterungen zum Blatt Nr. 6123 Marktheidenfeld,
Geologic map of Bavaria, Explanation for sheet
6123 Marktheidenfeld, Bayerisches Geologisches
Landesamt, 1979
[7] Mehra O. P., Jackson M. L., Iron oxide removal from
soils and clays by a dithionite-citrate system buffered
with sodium bicarbonate, Clays and Clay Minerals, 7,
1960, 317-327, DOI: 10.1346/CCMN.1958.0070122
[8] Wilson M. J., X-ray powder diffraction methods. In:
Wilson, M.J. (Ed.), A handbook of determinative,
methods in clay mineralogy, Blackie, Chapmann &
Hall, 1987, 26-98
[9] Tributh,
H.,
Qualitative und ”quantitative”
Bestimmung der Tonminerale in Bodentonen.
In: Tributh, H., and Lagaly, G. (Eds.), Identifizierung
und Characterisierung von Tonmineralen, Qualitative
and quantitative determination of clay minerals in
soil clays. In: Identification and characterization
of clay minerals, Berichte der Deutschen Ton-u.
Tonmineralgruppe e.V., DTTG-Convention, Gießen,
May, 10.-12, 1989
[10] Kohler E. E.,
Heimerl H.,
Czurda K.,
Quantitative
Mineralanalyse,
Sonderdruck.
In:
Methodenhandbuch für tonmineralogische
Untersuchungen, Quantitative mineral analysis,
special edition, In: Methods for clay mineral
examination, Bundesanstalt f. Geowiss. u. Rohstoffe,
Hannover, 1994
[11] Köster H. M., and Schwertmann U., Beschreibung
einzelner Tonminerale, In: Jasmund, K. and Lagaly,
G. (Eds.), Tonminerale und Tone:
Struktur,
Eigenschaften, Anwendungen und Einsatz in
Industrie und Umwelt, Description of individual clay
minerals. In: Clay minerals and clays: structure,
properties, purpose and uses in indurstry and
environment, Steinkopff, Darmstadt, 1993, 33-88,
DOI: 10.1007/978-3-642-72488-6_2
[12] DIN.18.129,
Baugrund,
Versuche
und
Versuchsgeräte: Kalkgehaltsbestimmung, Building
sites, testing and testing devices: determining
carbonate concentrations, Beuth, Berlin, 1990
[13] Laves D. u. Jähn,
G.,
Zur quantitativen
röntgenographischen
Bodenton-Mineralanalyse,
Concerning quantitative x-ray diffractive soil clay
mineral analysis, Arch. Ackeru. Pflanzenbau u.
Bodenkde., 16, H. 10, 1972, 735-739
[14] Imbrie J., Poldervaart A., Mineral Compositions
Calculated from Chemical Analyses of Sedimentary
Rocks, J. Sediment. Petrol., 1959, 29, No.
4, 588-595, DOI: 10.1306/74D709A2-2B21-11D78648000102C1865D
180
Unauthenticated
Download Date | 6/17/17 8:00 PM
U. R. Kackstaetter
[15] Gaines R. V., Skinner H. C. W., Foord E. E.,
Rosenzweig A., Dana’s New Mineralogy, 8th Edition,
John Wiley and Sons, New York, 1997
[16] O’Donoghue M., American Nature Guides - Rocks
and Minerals, Gallery Books, New York, 1990
[17] Duda R., Rejl L., Minerals of the World. Arch Cape
Press, New York, 1990
[18] Correns C. W., Tillmanns, Titanium: Ti(22). In:
Handbook of geochemistry Vol II/2. Ed. by K.H
Wedepohl, Springer Verlag: Berlin Heidelberg, 1978,
ISBN 3-540-04840-5
[19] Weaver C. E., Pollard L. D., The Chemistry of Clay
Minerals, Elsevier, 1973
[20] Dolcater D. L., Syers, J. K., Jackson M. L., Titanium as
free oxide and substituted forms in kaolinite and other
[21]
[22]
[23]
[24]
soil minerals, Clays and Clay Minerals, 1970, 18,
71-79, http://www.clays.org/journal/archive/volume%
2018/18-2-71.pdf
Weaver C. E., Clays, Muds, and Shales, Elsevier,
1989
Rengasamy P., Substitution of iron and titanium in
kaolinite, Clays and Clay Minerals, 24, 1976, 265266
Levinson A. A., Introduction to exploration
geochemistry, 2nd ed., Applied Publishing Ltd.,
1980
Folk R. L., Petrology of Sedimentary Rocks. Hemphill
Publishing Company, Austin, 1980
181
Unauthenticated
Download Date | 6/17/17 8:00 PM