ARTICLE IN PRESS

Computers & Geosciences 35 (2009) 1540–1551

Contents lists available at ScienceDirect

Computers & Geosciences

journal homepage: www.elsevier.com/locate/cageo

CALCMIN – an EXCELTM Visual Basic application for calculating mineral

structural formulae from electron microprobe analyses$
Andreas Brandelik
Institut für Mineralogie und Kristallchemie, Universität Stuttgart, Azenbergstr. 18, D-70174 Stuttgart, Germany

a r t i c l e in f o a b s t r a c t

Article history: CALCMIN, an open source Visual Basic program, was implemented in EXCELTM. The program was
Received 5 April 2007 primarily developed to support geoscientists in their routine task of calculating structural formulae of
Received in revised form minerals on the basis of chemical analysis mainly obtained by electron microprobe (EMP) techniques.
1 September 2008
Calculation programs for various minerals are already included in the form of sub-routines. These
Accepted 10 September 2008
routines are arranged in separate modules containing a minimum of code. The architecture of CALCMIN
allows the user to easily develop new calculation routines or modify existing routines with little
Keywords: knowledge of programming techniques. By means of a simple mouse-click, the program automatically
Structure formula generates a rudimentary framework of code using the object model of the Visual Basic Editor (VBE).
End member content
Within this framework simple commands and functions, which are provided by the program, can be
Fe2+/Fe3+ estimation
used, for example, to perform various normalization procedures or to output the results of the
Electron microprobe
Normalization computations. For the clarity of the code, element symbols are used as variables initialized by the
program automatically. CALCMIN does not set any boundaries in complexity of the code used, resulting
in a wide range of possible applications. Thus, matrix and optimization methods can be included, for
instance, to determine end member contents for subsequent thermodynamic calculations.
Diverse input procedures are provided, such as the automated read-in of output files created by the
EMP. Furthermore, a subsequent filter routine enables the user to extract specific analyses in order to
use them for a corresponding calculation routine. An event-driven, interactive operating mode was
selected for easy application of the program. CALCMIN leads the user from the beginning to the end of
the calculation process.
& 2009 Elsevier Ltd. All rights reserved.

1. Introduction
Perhaps the most widely used instrument for full chemical
analyses (except H, Li, and noble gases) of minerals and solid
inorganic compounds is the electron microprobe (EMP). In
contrast to other analytical techniques, the EMP allows one to
obtain reliable non-destructive analyses of solid phases in the
micrometer range with a high degree of accuracy. After analytical
work, often producing a large amount of data, the geoscientist is
confronted with the task of recalculating the analyses arranged in
data files into cationic formula units of minerals that can
occasionally have a complex composition. Such structural for-
mulae are needed, for instance, for thermobarometric calculations
or simply for classification purposes and mineralogical character-
ization of a rock sample. Most published computer programs for
calculating a mineral structural formula using EMP data are aimed
at specific minerals. Many such programs are designed for
amphiboles (e.g., Rock and Leake, 1984; Spear and Kimball,
1984; Rock, 1987; Mogessie et al., 1990; Richard and Clarke,
$
Code available from server at http://www.iamg.org/CGEditor/index.htm.
E-mail address: andreas.brandelik@loop.de
1990; Tindle and Webb, 1994; Rao and Rao, 1996; Yavuz, 1996,
1999; Mogessie, 2001; Esawi, 2004). Programs for the structural
recalculation of other rock-forming minerals (e.g., garnets,
Knowles, 1987; micas, Yavuz, 2003; pyroxenes, Sturm, 2002)
from EMP data are less frequent. In contrast to such ‘‘specific
mineral calculators’’, only a small number of general mineral
formulae calculation programs are presented in the literature
which are, for instance, PASFORM (De Bjerg et al., 1995),
HYPER-FORM (De Bjerg et al., 1992), MINTAB (Rock and Caroll,
1990), and MINFILE (Afifi and Essene, 1988). Furthermore, sub-
routines exist, such as COMP.EXE and AX, which are associated
with or included in programs for calculating phase equilibria
(TWQ: Berman, 1991; Thermocalc: Holland and Powell, 1998,
respectively). These sub-routines are mainly designed to create
composition files compatible with subsequent phase equilibria
computations.
Except for the latter computer programs, there are only a few
citations in the literature of the programs mentioned above. On
the author’s experience, many scientists use unpublished software
solutions for recalculating mineral formulae. In many cases, these
programs are internal software developments of the correspond-
ing scientific departments or individually coded solutions. The
main reason for using such individual software solutions is that

0098-3004/$ - see front matter & 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.cageo.2008.09.011
 ARTICLE IN PRESS

A. Brandelik / Computers & Geosciences 35 (2009) 1540–1551 1541

the flexibility of the above programs is insufficient. Most of the
published programs deal, for example, with a limited number of
elements and ignore the possible input of minor or trace elements,
or determine end members not suitable for subsequent thermo-
dynamic calculations. In addition, the flexibility of the afore-
mentioned programs is also limited due to the difficulty or even
impossibility of achieving modifications in the original algorithms
of the authors, especially for sequentially coded programs. The
high degree of flexibility required is predominantly caused by the
fact that only some aspects of the calculation procedure to
determine mineral structural formulae are straightforward. In
many cases, the final structural formula can only be approximated
because of analytical uncertainties. In addition to possible
measurement errors, such uncertainties can arise from the fact
that the lightest elements (Zo5) cannot be analyzed. Further-
more, elements with Z between 5 and 10, among the most
abundant element in silicates, namely oxygen, are rarely analyzed,
because not all EMPs in service today are featured with synthetic
multi-layer diffraction devices enabling analyses of light ele-
ments. An additional problem is the inability of EMP techniques to
determine the oxidation and coordination states of elements, such
as Fe. Vacancies are also not analyzable, although these kinds of
point defects are probably frequent. Thus, in many cases the final
structural formula of a mineral can only be approximated by
means of assumptions primarily based on mineral-specific knowl-
edge about crystal chemistry. From the above restrictions, it
follows that an algorithm for calculating a mineral structural
formula must be mineral-specific and in some respect variable. A
number of specific algorithms have been described for the
estimation of ferrous–ferric iron in minerals (Stout, 1972; Lindsley,
1983; Droop, 1987; Jacobson, 1989; Schumacher, 1991; Holland
and Blundy, 1994).
In addition to the lack of flexibility of the computer programs
mentioned above, another important point is the user’s con-
frontation with the task of extracting data manually from an
output file and preparing these data for a specific input format
before running the calculation routines. Only the sequential
program MINFILE (Afifi and Essene, 1988) enables an automated
read-in of output files of EMP data. MINFILE is thus, in this regard,
more user-friendly than others.
The aim of this work was to develop both a user-friendly and a
flexible software tool supporting primarily geoscientists in their
routine task of calculating mineral structural formulae on the
basis of chemical analysis obtained by EMP. An event-driven,
interactive operating mode is coded to automatically read-in
output files of the CAMECA SX, a widely used EMP, via network.
The program can also be used with other EMP data (e.g., JOEL). In
this case, the data have to be transferred into the worksheet
MANUAL INPUT (see Section 4.1.2). Moreover, a complete dialogue
and event-driven filter routine allows users to easily extract
specific analyses in order to use them for a corresponding
calculation routine. Furthermore, the presented program CALC-
MIN provides a high-performance Editor and allows the develop-
ment of individual calculation procedures with a minimum of
knowledge of programming techniques. In contrast to previously
published programs, CALCMIN allows the assignment of addi-
tional quantitative data to each analysis (e.g., ferric/ferrous iron
ratios determined by Mössbauer spectroscopy or estimates of
CALCMIN was also designed to simplify the input procedure of
the thermodynamic phase equilibrium calculation program
PTGIBBS (Brandelik and Massonne, 2004). In contrast to other
published mineral structural formulae calculation programs, the
calculation modules already included in CALCMIN provide, for the
first time, consistent site and mole fractions for the solution
models integrated into PTGIBBS.
2. Calculation procedure
In general, the recast of an analysis into a mineral formula is
based upon charge neutrality and knowledge of crystal chemistry
of the phase to be calculated and proceeds in the following three
steps.
Due to the fact that microprobe analyses are usually reported
in mass units, the first step is to determine the atomic proportions
of the measured elements by conversion from wt% to molar
proportions.
The second step is the normalization of the atomic proportions,
to calculate the respective number of measured elements nim in
the structural formula. This can be done in several ways.
For instance, the maximum number of analyzed atoms
substituting in a known number of sites in the mineral structure
can be taken for normalization (Afifi and Essene, 1988). Atoms
that are found at substituting structural sites or associated with
hydrogen atoms (often related to OH anions) as well as vacancies
should therefore be excluded from the normalization procedure.
Both calculation steps, the conversion from wt% to molar
proportions and the normalization procedure to a specific number
of atoms Nn, can be expressed by the equation
n
wm i N
nm
i ¼ P z (1)
mm
i
n n
j¼1 ðwj =mj Þ
with nm i ¼ number of measured element i in formula,
wm m
i ¼ concentration of element i in wt%, mi ¼ corresponding the
atomic weight, z ¼ number of elements used for normalization,
wnj ¼ concentration in wt%, and mnj ¼ atomic weight of element j
on which normalization is based. Another method of normalization
is based on the charges of elements in the mineral structure and is,
for instance, done on the basis of the number of oxygen anions in
the formula, which determines the necessary number of positive
charges to achieve charge neutrality. When normalizing to a
number of oxygen atoms Noxygen, the two first calculation steps in
allocating the formula can be summarized by the equation
wm ð2N oxygen Þ
nm
i ¼ Pi z (2)
mm
i
n n n
j¼1 ðwj cj =mi Þ
with cnj ¼ charge of cation j. In this equation, the summation
includes all measured cations and requires a predefinition of
oxidation states of the considered elements.
For full analyses of hydrous minerals (including anions such as
OH, F, and Cl), Noxygen can be expressed by
Noxygen ¼ Oanhydrous
!
atot ðwm m
H =mH Þ
þ þ 1 (3)
2 ðwm m m m m m
H =mH Þ þ ðwF =mF Þ þ ðwCl =mCl Þ

cation site occupancy determined by X-ray techniques) which can with atot ¼ total number of anions in the hydrous structural site. If
very easily be addressed in the freely programmable module used the H2O content is unknown, normalization may be based on
for subsequent calculation. This extends the functionality of the equivalent oxygen anions by
program far beyond pure mineral structural formulae calcula-
tions. Using site fractions of components in associated minerals as atot
Noxygen ¼ Oanhydrous þ (4)
additional data, for instance, a variety of thermobarometric 2
computations can be simultaneously carried out for the entered The third and final step, which has to be undertaken to
analyses. determine the desired structural formula, is the eventual
 ARTICLE IN PRESS
1542 A. Brandelik / Computers & Geosciences 35 (2009) 1540–1551
determination of various oxidation states of an element, and
the subsequent assignment of elements to structural positions
based on information about cation preference sequences within or
between various sites. As already mentioned above, the microp-
robe technique is unable to detect light elements, vacancies,
oxidation and coordination states and thus, the determination
of the final structural formula is a process of approximation
using simplifications. A frequently made simplification is, for
example, the assumption that only two various oxidation
states for one element exist and vacancies do not occur.
Then the proportions of the oxidation states can be directly
calculated by means of the simultaneous normalization to charges
and cations. For instance, Droop’s (1987) method of estimating
Fe3+/Fe2+ proportions in a mineral phase is based on such
assumptions.
3. Basis of CALCMIN
Most computer programs developed in recent years to
calculate structural formulae of minerals are based on spread-
sheet solutions using spatial relations between a set of cells
defined within a grid. These programs therefore generally use the
readily available Microsoft Office program EXCELTM, which
features an intuitive interface and capable calculation tools. One
advantage of this flexible environment is the simplicity with
which data can be changed, copied, selected, and arranged. In
addition, a high degree of compatibility to other text processing
computer programs exists. By means of the large number of built-
in functions (e.g., mathematical and statistical functions) and
available optimization methods it is also possible to undertake
complex computations within the grid. However, spreadsheet
solutions developed so far to calculate structural formulae are
limited to specific phases (see above), due to the fact that spatial
cell relations within a grid enabling the calculation of various
mineral phases would be extremely complex. Such an application
would forfeit the flexibility generally associated with this
environment because the modification and change of already
defined cell relations is difficult. In order to create a mineral
structural calculation program, which (1) is not restricted to a
specific phase and (2) uses the extreme flexibility of the EXCELTM
environment without limitations, the powerful programming
language Visual Basic for Application (VBA) was applied for
CALCMIN. By means of VBA, a completely event- and dialogue-
driven interactive operating system was implemented into the
EXCELTM environment. The complex mineral structural calcula-
tions required are not readily handled within the structure of the
grid cell calculations used by EXCELTM. The open source code is an
integral part of the workbook of CALCMIN and can be easily edited
by a high-performance Editor (VBE) included in the EXCELTM
environment. Each specific mineral calculation procedure is
stored in a separate module containing a minimum of code. The
architecture of CALCMIN enables users, who do not have
experience in programming, to easily develop new calculation
routines or modify already existing routines. By mouse-click, the
framework of code and the necessary modules are automatically
created by CALCMIN using the VBE technique. Within the code
framework simple commands can be used, for example, to
perform various normalization procedures. For the clarity of the
code element symbols are used as variables initialized by the
program automatically.
The database of CALCMIN is arranged in a worksheet and
includes the atomic weights of the elements published by the
International Union of Pure and Applied Chemistry (IUPAC) (Loss,
2003).
4. Program description
CALCMIN consists of 13 worksheets. After starting CALCMIN,
the main worksheet ‘‘START’’ is automatically activated (see
Fig. 1). In the upper portion of this worksheet, actual information
on the program status is given. Underneath, various command
buttons are arranged allowing a simple navigation within the
entire program, the start of different computation and input
routines, and the coding or editing of sub-routines to calculate
structural formulae of minerals. The atomic weights and the
possible oxidation states of each element are arranged in the
worksheet ‘‘DATABASE’’. The names of already included mineral
calculation routines and additional information in the form of
comments are listed in the worksheet ‘‘PROGRAM_DEFINITION’’.
In order to carry out a data input within the grid, the worksheet
‘‘MANUAL_INPUT’’ and, for users who want to develop their own
worksheet applications using the available functions built into
CALCMIN, a blank worksheet named ‘‘DEVELOPMENT’’ are
provided. The results of computation are presented in the
worksheet ‘‘RESULT’’ and ‘‘PRINT_OUTPUT’’. The remaining seven
internal worksheets are primarily used by CALCMIN as temporary
storage of data input. Table 1 concisely describes all worksheets.
The program is completely driven by dialogues leading the
user from the beginning to the end of the calculation process
(Fig. 2). Depending on the type of data input, a filter routine is
automatically started or has to be started manually. This routine
enables the user to easily extract specific analyses in order to
apply them in a corresponding calculation routine. By means of
this filter procedure, mineral analyses to be calculated are
transferred into the worksheet ‘‘FILTERDATA’’. The subsequent
routine computing the final structural formula retrieves data only
from this sheet. After running the appropriate sub-routine, the
calculation results are automatically presented in the worksheet
‘‘PRINT_OUTPUT’’. In the following sections, the four various input
procedures supported by CALCMIN and the filter process are
described in more detail.
4.1. Data input
4.1.1. Manual input by dialogue
By clicking on the button MANUAL INPUT BY DIALOGUE an
interactive routine will be started allowing input of mineral
analyses (in wt% of oxides). After starting the routine, a dialogue is
displayed, in which elements can be selected for subsequent data
input (Fig. 3). The calculation of the corresponding oxide formulae
is based on default valences which are arranged for each element
in the worksheet ‘‘DATABASE’’ (column 3). Other valences can be
defined by clicking on the button SPECIFIC VALENCES. A window is
opened which shows the selected elements and two combo boxes
listing all possible valences of respective atoms. By selecting two
different valences for an element two oxide formulae will be
generated for input purposes (e.g., Fe: Fe2O3 and FeO). After
confirming with the OK-button, the interactive input procedure
will proceed by clicking on the button NEXT. The subsequent
dialogue shows the generated oxide formulae in a list box. For
defining the order of oxides used in the input procedure the
button CHANGE ORDER has to be clicked. The sequence of double-
clicks on the oxides in the left list box determines the order of the
subsequent input. The selection procedure can be started at any
time by reclicking the button CHANGE ORDER. After confirming
with the OK-button, the input dialogue is displayed (Fig. 4). For
each analysis, a comment can be entered in a text box at the top of
the window. If more than one analysis should be calculated, the
button INPUT FURTHER ANALYSIS has to be clicked after finishing
the input of an analysis. The data input is finished by clicking on
 ARTICLE IN PRESS
A. Brandelik / Computers & Geosciences 35 (2009) 1540–1551
Fig. 1. Screenshot of CALCMIN after starting program.
the button OK and will be automatically displayed in the
worksheet ‘‘MANUAL_INPUT’’ (see Fig. 5). The command buttons
at the top of this sheet allow the data input to be saved in a cmi-
file (CALCMIN MANUAL INPUT FILE), procedures to be run to filter
and calculate the analyses, and permit the loading of already
existing cmi-files.
4.1.2. Manual input within worksheet
By clicking on the button MANUAL INPUT WITHIN WORKSHEET,
the worksheet ‘‘MANUAL INPUT’’ is automatically activated. In this
worksheet, analyses can be directly copied from the clipboard or
entered manually. An example for an input is given in Fig. 5. Oxide
names are provided in row (1) and must be finished with the
entry ‘‘Total’’. Anions such as F or Cl must be indicated with a
minus () at the end (e.g., F). Column 1 (A) and column 2 (B) are
reserved for comments and numbers, respectively. The CALCMIN
calculation routines allow the input of different oxides of an
element. Before running the calculation routine, one should save
the data input by clicking on the button SAVE AS MANUAL INPUT
FILE.
4.1.3. LOAD CMI-file
CMI-files are produced by CALCMIN containing data of mineral
analyses given in wt% of oxides and are arranged in a table as
described before. By clicking on the button LOAD CMI-File a
window appears, which enables users to search for this type of file
in the network. After opening a file the worksheet ‘‘MANUAL
INPUT’’ is automatically activated containing the transferred
1543
mineral analyses (see Fig. 5). The calculation routines can be
immediately started.
4.1.4. LOAD .COR file (UNIX)
COR-files are output files of the CAMECA SX EMP running on a
UNIX platform. This kind of file contains element analyses in
weight or atom percent. The routine subsequently described
enables users to calculate in best time mineral structure formulae
in an interactive environment. By clicking on the button
LOAD .COR FILE (UNIX), CALCMIN starts a dialogue to search for
this kind of file in the network. After selecting a file and clicking
the button OPEN, a new window opens informing about the
CAMECA SX output file (Fig. 6). A change of default settings for
element valences for the subsequent calculation of oxides can be
achieved by selecting the option NEW VALENCES FOR ATOMS. After
clicking the button NEXT and using this option an additional
window displays available valences for each atom listed in
corresponding combo boxes. The calculation of oxides at this
time is only used for displaying the analyses in the subsequent,
automatically started filter procedure. CALCMIN imports the
analyses into the worksheets ‘‘WTPERCENT_ATOMS’’, ‘‘CONPER-
CENT_ATOMS’’, and ‘‘READIN_DATA’’. Additional data such as XYZ-
coordinations of analyses are also imported. The recalculated
oxides are located in the worksheet ‘‘WTPERCENT_OXIDES’’. The
above worksheets can be viewed simply by clicking the button
SHOW SHEET, arranged in the worksheet ‘‘START’’, and selecting
the specific sheet.
 ARTICLE IN PRESS

1544 A. Brandelik / Computers & Geosciences 35 (2009) 1540–1551

Table 1
Short description of worksheets included in CALCMIN.

Name of worksheet Short description

START Is the main worksheet. It is automatically activated after starting CALCMIN. The start of the computation procedures, the editing or coding of
calculation routines and the navigation in the workbook can be achieved from here
INPUT_MANUAL Enables the direct input of mineral analyses within the grid. If an interactive manual data input is made the data will also be arranged in this
sheet. The saving of input data in a cmi-file and the loading of such a file can also be achieved from this sheet
DATABASE Lists atomic weights, default valences, and other possible oxidation states
PROGRAM_DEFINITION Lists names of already included mineral structural calculation programs and corresponding comments
PRINT_OUTPUT Shows the result of a structural calculation in an automatically formatted table. The result can be saved in an external workbook or printed out
from here. The sheet is automatically activated after calculation
RESULT Presents the result of calculation. Each calculated analysis is arranged in a row
DEVELOPMENT Is an empty worksheet for user who wants to develop pure spreadsheet applications based on available functions built in CALCMIN
FILTERDATA Lists the analyses given in wt% of atoms which are filtered for subsequent calculation. The included mineral calculation programs take the data
from this sheet
CONPERCENT_ATOMS Presents input data in atomic concentration percentages
WTPERCENT_ATOMS Presents input data in atomic weight percentages
WTPERCENTS_OXIDES Presents input data in oxide weight percentages
READIN_DATA Shows the original read-in microprobe measurement file in a tabular form
WAIT Is an internal worksheet of CALCMIN. It will be displayed during calculation process

Fig. 2. Schematic overview of CALCMIN’s program structure.

Fig. 3. Screenshot of dialogue allowing selection of elements for subsequent manual data input.

4.2. Filter routine event-driven and interactive filter procedure allows the extraction
of specific analyses in order to apply them in a corresponding
The filter routine is the connecting link between data input and calculation routine. CALCMIN’s filter routine can only be started
the calculation of mineral structural formulae. The completely if data exist in the form of a manual input (within
 ARTICLE IN PRESS

A. Brandelik / Computers & Geosciences 35 (2009) 1540–1551 1545

Fig. 4. Screenshot of CALCMIN’s manual input window after input of a clinopyroxene analysis in wt% (see working example 1). Number, type, and order of oxides of input
data were defined interactively (see Section 4.1.1).

Fig. 5. Screenshot of worksheet ‘‘MANUAL INPUT’’ after loading cmi-file of working example 2. In this worksheet, analyses can also be directly inserted from clipboard or
entered manually.

worksheet ‘‘MANUAL INPUT’’) or in the form of loaded CAMECA
SX analyses arranged in the worksheets ‘‘WTPERCENT_ATOMS’’,
‘‘CONPERCENT_ATOMS’’, and ‘‘READIN_DATA’’. By means of the
filter procedure applied once or several times, mineral analyses
are selected and transferred into the worksheet ‘‘FILTERDATA’’.
The analyses located in this sheet are always given in wt% of
elements. Thus, mineral analyses, which were manually entered in
wt% of oxides, are calculated back to element wt% of elements.
The filter routine can be manually started by clicking the button
RUN FILTER in the worksheet ‘‘START’’ or the button FILTER AND
CALCULATE arranged in the upper portion of the worksheet
‘‘MANUAL_INPUT’’. In the case of transferred EMP data, the filter
routine is run automatically. After the first start of the filter
procedure a dialogue is displayed showing all analyses given in wt%
of oxides (see Fig. 7). By means of three option buttons arranged at
the top of the window analyses can be displayed also in atom% or
atom wt%. CALCMIN creates automatically for each oxide/element
and the ‘‘Total’’ two corresponding text boxes to define maximum
and minimum values. Appropriate selections enable the user to
filter the data for a specific type of mineral. The filter process is
started by clicking the button FILTER and subsequently displays the
data fulfilling the conditions of the range definitions. The number
of filtered data is given in the status frame. Range definitions can be
arbitrarily modified to repeat the filter procedure. If no maximum
 ARTICLE IN PRESS

1546 A. Brandelik / Computers & Geosciences 35 (2009) 1540–1551

Fig. 6. Screenshot of CALCMIN’s information window after loading Cameca SX output file monazite.cor of working example 3.

Fig. 7. Screenshot of filter dialogue displaying monazite analyses of working example 3. Filter procedure allows extraction of specific analyses in order to apply them in a
corresponding calculation routine.

or minimum values are defined all analyses are taken, and thus, are available. The analyses to be calculated later can be selected by a
released for subsequent calculations. By selecting the option button mouse-click. The procedures for both filter techniques are finished
FILTER BY MOUSE SELECTION an additional filter technique is by clicking the OK-button.
 ARTICLE IN PRESS

A. Brandelik / Computers & Geosciences 35 (2009) 1540–1551 1547

5. Calculation routines
5.1. Starting of calculation routines
After finishing the filter process, CALCMIN automatically opens
a new window showing the mineral structure calculation
programs which are included in CALCMIN by the author and/or
the corresponding user (see Table 2). A sub-routine to calculate
the final structural formula can be selected. The calculation
programs can also be run by clicking on the button RUN
CALCULATION PROGRAM in the worksheet ‘‘START’’, provided that
mineral analyses already exist in the worksheet ‘‘FILTERDATA’’.
Thus, various routines to allocate the structure formulae can be
consecutively run using the same set of filtered analyses. To view
the code of a specific sub-routine the button EDIT CALCULATION
ROUTINE has to be clicked.
5.2. Coding of calculation routines
New codes can be easily produced by clicking the button
CREATE A NEW ROUTINE arranged in the worksheet ‘‘START’’. In the
opening dialogue, a name for the new program and a comment
can be entered. The program name should have more than five
letters (numbers are only allowed after the first position). Spaces
and other signs such as commas, points, brackets except (‘‘_’’) are
not allowed in the program name. After confirmation, CALCMIN
creates a new module in the programming environment contain-
ing a sub-routine with the name entered. The module name
differs only from the corresponding sub-routine name by the
added string ‘‘_code’’. The automatically generated framework of
the code looks as shown in Fig. 8.
The first command ‘‘init’’ within the routine counts for the
number of analyses to be calculated and initializes necessary
variables. In the next line a ‘‘For...Next’’ loop is started. The public
variable ‘‘all’’ corresponds to the number of analyses to be
calculated and, thus, determines how often the code, placed
between the loop key words ‘‘For analysis ¼ 1 To all’’ and ‘‘Next
analysis’’, is repeated. The first command ‘‘load_element_value
(analysis)’’ within the loop provides element data of correspond-
ing analysis by transforming the data given in wt% into atom
percent and storing the values in variables. The names of these
variables match exactly the symbols of elements except for
arsenic, iodine, and indium. For these three elements, the full
names have to be used with upper case letter at the beginning. In
the subsequent code section (between the command ‘‘load_ele-
ment_value (analysis)’’ and the key words ‘‘Next analysis’’), the user
has to implement the main calculation routine. In this section
element calculations can be simply undertaken by using the
element symbols. If a separation into various oxidation states of
an element is required CALCMIN provides additional public
variables. The valency of a cation is represented by an integer
after the element symbol (e.g., Fe2, Fe3, Mn5). Anions are denoted
by an underscore between the element symbol and an integer
(e.g., S_2, As_2). It is important to know that variables associated
with valences are all set to empty by the command ‘‘load_ele-
ment_value (analysis)’’ at the beginning of a new loop. Thus, if
input data contain information about various oxidation states of
an element (e.g., Fe2O3 and FeO) only the total of the correspond-
ing element (in atom%) is stored in the element variable.
To avoid loss of information ten additional public variables
(variant type) are available within the code namely VAR1,
VAR2,y,VAR10. These variables can contain additional informa-
tion entered together with an analysis in the worksheet ‘‘MANUAL
INPUT’’ (e.g., ferric/ferrous iron ratios determined by Mössbauer

Table 2
Description of structural formula calculation routines already included in CALCMIN.

Name of routine Phase Normalization and short description

Ti_garnet_B Garnet [6]- and [8]-coordinated cations ¼ 10, Fe2+/Fe3+-calculation by cation difference, end member calculation for mixing
model Ti_garnet (Brandelik and Massonne, 2004)
garnet_MAS Garnet [6]- and [8]-coordinated cations ¼ 10, Fe2+/Fe3+-calculation by cation difference, end member calculation for mixing
model garnetmas (Massonne, 1995; Brandelik and Massonne, 2004)
white_mica Mica Valences ¼ 42, cations ¼ 12.1, Fe2+/Fe3+-calculation, end member calculation for mixing model WhiteMica (Massonne
and Szpurka, 1997; Brandelik and Massonne, 2004)
titan_mica Mica Valences ¼ 42, cations ¼ 12.1, Fe2+/Fe3+-calculation, end member calculation for mixing model Ti-muscovite
(Massonne et al., 1993; Brandelik and Massonne, 2004)
alumosilicat Al2SiO5 Cations ¼ 3
feldspar Plagioclase Oxygens ¼ 8, end member calculation for plagioclase mixing model PLAG Brandelik and Massonne (2004)
chlorite Chlorite O ¼ 28, calculation of XSi, XMg, and XFe
titanite Titanite Si ¼ 1
stilp1_mas Stilpnomelane O ¼ 192, without Fe2+/Fe3+-calculation
epidote Epidote O ¼ 12.5
ilmenite_TT Ilmenite Cations ¼ 2, Fe2+/Fe3+-calculation by O ¼ 3
talc_mas1 Talc Cations ¼ 7, Fe2+/Fe3+-calculation by valences ¼ 22
spinel1_MAS Spinel Cations ¼ 3, Fe2+/Fe3+-calculation by valences ¼ 8
olivin1_MAS Olivine Cations ¼ 3, Fe2+/Fe3+-calculation by valences ¼ 8
carbonat1_MAS Calcite Valences ¼ 2, set C ¼ 1
epidote_Sm Epidote O ¼ 12.5, addition of Sm by iteration
orthopyroxen_MAS Orthopyroxene Valences ¼ 12, cations ¼ 4, Fe2+/Fe3+-calculation, end member calculation for mixing model OPX (Massonne, 1995;
Brandelik and Massonne, 2004)
clinopyroxen_MAS Clinopyroxene Valences ¼ 12, cations ¼ 4, Fe2+/Fe3+-calculation, Na adjustment by iteration, end member calculation for mixing
model cpx C2/c and cpx P2/n (Massonne, 1995; Brandelik and Massonne, 2004)
amphibol_MAS Amphibole Cations ¼ 13, Fe2+/Fe3+-calculation, end member calculation for mixing model PLAG (Massonne, 1992a; Brandelik and
Massonne, 2004)
cpx_lindsley Clinopyroxene Cations ¼ 4, valences ¼ 12, Fe2+/Fe3+-calculation according to Lindsley (1983)
cpx_droop Clinopyroxene Cations ¼ 4, valences ¼ 12, Fe2+/Fe3+-calculation according to Droop (1987)
chloritoid3 Chloritoide Cations ¼ 6, Fe2+/Fe3+-calculation, XMn, XMg, and XFe
monazite Monazite Valences ¼ 8, calculation of Th, Pb, U age
garnet_droop Garnet Cations ¼ 8, valences ¼ 24, Fe2+/Fe3+-calculation according to Droop (1987)

Architecture of CALCMIN allows user to develop easily new calculation routines or modify already existing routines with little knowledge of programming techniques.
 ARTICLE IN PRESS
1548 A. Brandelik / Computers & Geosciences 35 (2009) 1540–1551
spectroscopy, estimates of cation site occupancy determined by
X-ray techniques or mole fractions of components in associated
minerals). For these data, 10 cells are provided starting after the
‘‘Total’’ entry and one empty column in the corresponding row of
specific analysis (for example, total is located in column 16 then
VAR1 and so forth are located in column 18, etc.).
To store intermediary results, the user is free to introduce
additional variables within the code segment. If variable names
and their stored values should appear in the output, a declaration
in the form of a public type variable in the module ‘‘specific
declaration’’ has to be made in advance. Already defined public
variables, which can be used for output purposes, are listed in the
above-named module.
Within the user defined code section, the user can implement
code not limited in size and complexity. The complete code syntax
of the powerful VBA language can be used. Various matrices and
Fig. 8. VBE editor of CALCMIN showing complete code of sub-routine ‘‘droopexample’’ created in working example 1. Framework of code is automatically generated by
CALCMIN (see Section 5.2).
Table 3
Built-in functions and commands in CALCMIN.
Name of function/commands
GET_ACTUAL_VALENCY (‘‘element’’)
GET_DEFAULT_VALENCY ‘‘element’’
SET_ACTUAL_VALENCY ‘‘element’’, valency
SET_DEFAULT_VALENCY ‘‘element’’
GET_OXVAL_END (‘‘oxide_string’’, number of analysis)
load_element_value (number of analysis)
norm_ions X, Array (element 1, element 2, element 3, etc.) i
write_oxides analysis, Array(‘‘element 1’’,‘‘element 2’’, ‘‘element
3’’, etc.) i
out_structure analysis, Array (‘‘element 1’’,‘‘element 2’’, ‘‘element
3’’, ‘‘variable’’, ‘‘element 4’’, ‘‘variable’’, etc.) i
sum_val (Array(‘‘element 1’’,‘‘element 2’’, ‘‘element 3’’, etc.)) i
norm_val X, Array((‘‘element 1’’,‘‘element 2’’,’’element 3’’, )
Get_U_Th_Pb_age (wt% ThO2, wt% UO2, wt% PbO)
Init
Arrays where elements associated with oxidation state such as ‘‘Fe2’’ and ‘‘Mn4’’ are permitted are each identified by an ‘‘i’’ at the end of the bracket.
optimization methods, for instance, can be implemented without
any difficulty. In order to shorten the code, CALCMIN provides
additional commands and functions within the freely program-
mable code section framed by the loop (see Table 3). For instance,
the command ‘‘norm_ions 5, Array(Ca, Si, Mg)’’ allows a normal-
ization to the sum of Ca, Si, and Mg equal to 5 (see Eq. (1)).
Variable names associated with the oxidation state such as Fe2
and Mn4 are not permitted here. The elements in brackets have to
be separated by commas and placed without quotation marks.
The function ‘‘sum_val(Array())’’ determines the sum of
valences and the command ‘‘norm_val 2 Noxygen, Array()’’ executes
a normalization on the basis of a specific number Noxygen of
oxygens (see Eq. (2)). In the array of the ‘‘sum_val’’ function
element variables associated with oxidation states (e.g., Fe2, Mn3)
can be used. In contrast to the ‘‘norm_ions’’ command described
above the array elements have to be placed in quotation marks.
Short description
Returns the actual valency of an element
Returns the default valency of an element
Sets a valency of an element. The function can be used in the code section within the basic loop (‘‘For
analysis 1 to all’’ and ‘‘Next analysis’’)
Sets the default valency of an element
Returns the wt% of an oxide within a sub-routine. The function can only be used after using the
command write_oxides
Provides element data of corresponding analysis by transforming the data given in wt% into atom%
and storing the values in variables. The names of these variables match exactly the symbols of
elements except for arsenic, iodine, and indium
Allows a normalization to the sum of elements given in the array equal to X
Leads to the output of oxides and the corresponding sum. Oxide formulae are automatically
computed on the basis of the valences associated to the element variables used in the array.
Causes the output of the calculated structural formula. All variables applied in the code section can be
used
Determines the sum of valences of given elements in the array
Executes a normalization on the basis of a specific number Noxygen of oxygens
Determines the age of a monazite analysis on the basis of the Th, U, and Pb content according to
Montel et al. (1996)
The command ‘‘init’’ within the routine counts for the number of analyses to be calculated and
initializes necessary variables
 ARTICLE IN PRESS

A. Brandelik / Computers & Geosciences 35 (2009) 1540–1551 1549

For output purposes, two further commands are provided by Table 4

CALCMIN, which can be placed at the end of the main code Clinopyroxene analysis calculated by Droop (1987) and CALCMIN (see working
example 1).
section. The command ‘‘write_oxides analysis, Array()’’ leads to the
output of oxides and the corresponding sum. Oxide formulae are Droop (1987) CALCMIN
automatically computed on the basis of the valences associated to
the element variables used in this array. SiO2 56.21 56.21
If no integer is added to the element symbol, the default TiO2 0.00 0.00
Al2O3 10.56 10.56
valency of the element determines the corresponding oxide Fe2O3 3.39 3.39
formula. The order of output is defined by the array list. FeO 3.05 3.05
The command ‘‘out_structure analysis, Array()’’ causes the MnO 0.01 0.01
output of the calculated structural formula. All variables applied MgO 7.33 7.33
CaO 12.13 12.13
in the code section can be used. It is important to know that a
Na2O 7.64 7.64
successful output requires variables, which were declared as K2O 0.04 0.04
public variables, and corresponding values also initialized during Total 100.36 100.36
run time.
Si 1.998 1.9976
Ti 0.000 0.0000
6. Printing and saving of results Al 0.442 0.4423
Fe3+ 0.091 0.0907
6.1. Printing Fe2+ 0.091 0.0906
Mn 0.000 0.0003
Mg 0.388 0.3883
After successful calculation, the result is automatically pre- Ca 0.462 0.4619
sented in the worksheet ‘‘PRINT_OUTPUT’’ and can be printed out Na 0.526 0.5264
by using the button PRINT. Before sending the data to the system’s K 0.002 0.0018
Total 4.000 4.0000
default printer CALCMIN runs formatting procedures to optimize
the print-out automatically. Original EMP analysis (in wt% oxide: SiO2 ¼ 56.21, TiO2 ¼ 0.00, Al2O3 ¼ 10.56,
FeO ¼ 6.10, MnO ¼ 0.01, MgO ¼ 7.33, CaO ¼ 12.13, Na2O ¼ 7.64, K2O ¼ 0.04) is
6.2. Saving of data given in Droop (1987).

The calculated data can be saved in an external file with the
extension ‘‘.cmp’’ by clicking the SAVE DATA IN EXTERNAL FILE
button arranged at the top of the ‘‘PRINT_OUT’’ worksheet. Two
options are available for saving the results. By selecting the option
SAVE IN NEW PROJECT, a new workbook is created. The results are
transferred into this workbook. The respective analyses and
calculated structure formulae are arranged in the form of rows
in a worksheet, the name of which matches the name of the
program used for calculation. If the option SAVE DATA IN OLD
PROJECT is selected, CALCMIN requests opening the corresponding
*.cmp file and adds the new data to this file by creating a new
worksheet named by the calculation procedure used. After saving,
CALCMIN asks how to proceed. By using the button YES in the
opened dialogue the analyses, which were saved before, are
removed from CALCMIN’s actual input data set (e.g., microprobe
analyses). The advantage is that those analyses fail to appear in
the next filter procedure. Thus, a convenient way is given to
process a data file of an electron microprobe, containing normally
data of various phases, step by step.
7. Working examples
the VBE is started, showing a rudimentary code framework
generated automatically. Within the code section, bounded by
the command ‘‘load_element_value (analysis)’’ and the key words
‘‘Next analysis’’, 19 additional code lines have to be added as
shown in Fig. 8. The algorithm follows step by step the procedure
by Droop (1987). Using three additional variables, the commands
‘‘norm_val’’ and ‘‘norm_ions’’, two conditional statements on the
basis of the ‘‘4’’ operator, and the general equation given in Droop
(1987), a separation of Fe into Fe3+ and Fe2+ for clinopyroxene can
be carried out. For output purposes, the two commands ‘‘write_-
oxides analysis()’’ and ‘‘out_structure analysis()’’ are arranged at the
end of the freely programmable code section. The separation of Fe
in Fe2+ and Fe3+ within the array of the command ‘‘write_oxides
analysis()’’ leads to an automated recalculation of the total. After
finishing the coding, the VBE has to be closed and the mineral
analysis can be input, for instance, by clicking the button MANUAL
INPUT BY DIALOGUE (see Figs. 1, 3, and 4). Following the
instructions for manual input, the final calculation can be started
by clicking the button FILTER AND CALCULATE (see Section 4.1.1).
The results of calculations are given in Table 4 and correspond
exactly to values determined by Droop.

To demonstrate the versatility and user-friendliness of CALC- 7.2. Working example 2

MIN, three working examples are presented. In example 1, only
one analysis is computed, which has to be input manually. The
analyses used in examples 1 and 2 are included in the files
‘‘monazite.cor’’ and ‘‘erzgebirge.cmi’’, which are located in the
subdirectory ‘‘WORKING EXAMPLES’’ of the CALCMIN package.
7.1. Working example 1
Example 1 is related to the creation of a new sub-routine
allowing the computation of a clinopyroxene analysis given by
Droop (1987) using this Fe2+/Fe3+-estimation method in CALCMIN
(see Table 4). After clicking the button CREATE CALCULATION
ROUTINE and entering a name for the sub-routine to be created,
Example 2 is related to the calculation of the structural
formula and molar fractions of EMP analyses of phengite given by
Massonne and Kopp (2005). The analyses can be transferred into
the worksheet ‘‘MANUAL_INPUT’’ by clicking on the button LOAD
CMI-FILE (see Fig. 1), which is arranged in the worksheet ‘‘START’’.
The data given in wt% oxides are identical to those of Massonne
and Kopp (2005). After clicking the button FILTER AND CALCULATE
(see Fig. 5) and subsequent release of all five analyses, the sub-
routine ‘‘Whitemica_mas’’ has to be run. In this sub-routine the
EMP analyses are recalculated to a double formula unit on the
basis of 42 valences for the four- and six-fold coordinated cations.
Furthermore, the molar fractions of six white mica end members
 ARTICLE IN PRESS

1550 A. Brandelik / Computers & Geosciences 35 (2009) 1540–1551

are also determined according to Massonne (1992b). In addition, a set aside. The U–Th–Pb age is calculated by using the monazite
recalculation of oxides including H2O, a correction of the ‘‘Total’’ age equation (e.g., Montel et al., 1996). The corresponding
also considering the F content, an estimation of Fe2+/Fe3+ (not analyses, which were analyzed with a CAMECA SX50, can be
relevant here), and an assignment of elements to structural loaded into the CALCMIN environment by clicking the button
positions are automatically performed. LOAD *.COR FILE(UNIX) and selecting the file ‘‘monazite.cor’’
The calculation result (see Table 5) is automatically displayed located in the subdirectory ‘‘WORKING EXAMPLES’’ of the
in the worksheet ‘‘PRINT_OUT’’ and agrees, apart from minor CALCMIN package (see Fig. 6). After the transfer of the data,
changes caused by rounding effects, exactly with the values CALCMIN automatically opens the filter window showing all 43
calculated by Massonne and Kopp (2005). analyses included in the ‘‘monazite.cor’’ file. To exclude monazite
analyses, which have inadequate sums, the range for the total, for
instance, can be set between 99 and 101 wt%. The selection is
7.3. Working example 3 finished by clicking on the button FILTER (see Fig. 7). In the
subsequent window, the sub-routine ‘‘monazite’’ has to be
Example 3 is related to the computation of the structure selected. After confirmation, the program starts the computation
formulae and the U–Th–Pb age of monazite from Brazil. For of the 33 selected analyses. The calculation result is automatically
simplification purposes, a statistical treatment and actual, displayed in the worksheet ‘‘PRINT-OUT’’. Fig. 9 shows the result
necessary corrections to EMP analyses (e.g., Pyle et al., 2005) are of the computation in the form of an EXCEL chart, which can easily
be created navigating the worksheet ‘‘RESULT’’ and using EXCEL’S
assistant for creating charts.
Table 5 At the beginning of the computation routine ‘‘monazite’’ (see
Calculated structural formulae and end members of phengites of working Table 2), a normalization to 4 oxygens is undertaken. At the end of
example 2.
the code section, the function ‘‘Get_U_Th_Pb_age (GET_OXVA-
Comment 1802/43 1035/28 1802/42 1802/70 1035/9 L_END(‘‘ThO2’’, analysis), GET_OXVAL_END (‘‘UO2’’, analysis),
GET_OXVAL_END(‘‘PbO’’, analysis))’’ is located allowing the
X 1 2 3 4 5 computation of the U–Th–Pb age of a monazite analysis. This
Si 6.9538 6.8955 6.8374 6.7852 6.6803 function is implemented in the same module beneath the sub-
Al_T 1.0462 1.1045 1.1626 1.2148 1.3197
Sum4 8.0000 8.0000 8.0000 8.0000 8.0000
routine and solves the non-linear equation determining the age by
Al_O 2.9365 2.9840 3.0528 3.0839 3.1586 using the Newton–Raphson method.
Ti 0.0215 0.0223 0.0235 0.0272 0.0221
Cr 0.0021 0.0020 0.0021 0.0112 0.0127
Fe2 0.1655 0.1596 0.1610 0.1601 0.1634
Fe3 0.0000 0.0000 0.0000 0.0000 0.0000
Mn 0.0000 0.0000 0.0000 0.0000 0.0011
8. System requirements and availability of the program
Mg 0.9066 0.8690 0.7911 0.7501 0.6941
Sum6 4.0322 4.0369 4.0304 4.0326 4.0520 CALCMIN requires Microsoft EXCELTM installation (version 8.0
Ca 0.0112 0.0014 0.0139 0.0014 0.0000 or higher) on an IBM-compatible personal computer. The mini-
Ba 0.0128 0.0121 0.0138 0.0096 0.0105
Na 0.1035 0.1325 0.1412 0.2106 0.2361
mum system requirements are a Pentium III 1000 MHz processor
K 1.7261 1.7011 1.6758 1.5725 1.7820 (or equivalent) and 128 RAM memory. For good performance, a
Sum8 1.8535 1.8471 1.8448 1.7941 2.0285 minimum screen resolution of 1024  768 is recommended.
F 0.0124 0.0000 0.0000 0.0041 0.0127 Before using CALCMIN, one must ensure that EXCEL is able to
Cl 0.0000 0.0000 0.0000 0.0000 0.0000
access Visual Basic projects. The CALCMIN program is available
H 3.9876 4.0000 4.0000 3.9959 3.9873
from server at http://www.iamg.org/CGEditor/index.htm.

Components
MUSCOVITE 0.4140 0.4358 0.4473 0.4450 0.4929
WM_PHLOGOPITE 0.0161 0.0185 0.0152 0.0163 0.0260
AL_CELADONITE 0.4142 0.3844 0.3604 0.3286 0.2795 Acknowledgements
FE_AL_CELADONITE 0.0756 0.0706 0.0733 0.0701 0.0658
PARAGONITE 0.0558 0.0717 0.0766 0.1174 0.1164 The author wants to thank H.-J. Massonne and T. Theye for
TI_MN_CA_WM 0.0242 0.0190 0.0273 0.0226 0.0194
supporting this work. The program benefited from their sugges-
XOH 0.9969 1.0000 1.0000 0.9990 0.9968
tions for improvement.

U-Th-Pb age of monazite analyses

610
590
age (million years)

570
550
530
510
490
470
450
1 2 6 7 12 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 38 39 40 41 42 43
analysis

Fig. 9. Calculated U–Th–Pb ages of 33 monazite analyses of working example 3.

 ARTICLE IN PRESS
A. Brandelik / Computers & Geosciences 35 (2009) 1540–1551
References
Afifi, A.M., Essene, E., 1988. MINFILE: a microcomputer program for storage and
manipulation of chemical data on minerals. American Mineralogist 73,
446–448.
Berman, R.G., 1991. Thermobarometry using multi-equilibrium calculations: a new
technique, with petrological applications. Canadian Mineralogist 29, 833–855.
Brandelik, A., Massonne, H.-J., 2004. PTGIBBS – an EXCELTM Visual Basic program
for computing and visualizing thermodynamic functions and equilibria of
rock-forming minerals. Computers & Geosciences 30, 909–923.
De Bjerg, S.C., Mogessie, A., Bjerg, E., 1992. HYPER-FORM – A Hypercards program
for Macintoshs microcomputers to calculate mineral formulae from electron
microprobe and wet chemical analysis. Computers & Geosciences 18, 717–745.
De Bjerg, S.C., Mogessie, A., Bjerg, E., 1995. PASFORM – a program for IBMs PC or PC-
compatible computers to calculate mineral formulae from electron microprobe
and wet-chemical analysis. Computers & Geosciences 21, 1187–1190.
Droop, G.T.R., 1987. A general equation for estimating Fe3+ concentrations in
ferromagnesian silicates and oxides from microprobe analyses, using stoichio-
metric data. Mineralogical Magazine 51, 431–435.
Esawi, E.K., 2004. AMPH-CLASS: an Excel spreadsheet for the classification and
nomenclature of amphiboles based on the 1997 recommendations of the
International Mineralogical Association. Computers & Geosciences 30, 753–760.
Holland, T., Blundy, J., 1994. Non-ideal interactions in calcic amphiboles and their
bearing on amphibole-plagioclase thermometry. Contributions to Mineralogy
and Petrology 116, 433–447.
Holland, T.J.B., Powell, R., 1998. An internally-consistent thermodynamic data set for
phases of petrological interest. Journal of Metamorphic Geology 16, 309–343.
Jacobson, C.E., 1989. Estimation of Fe3+ microprobe analyses: observations on calcic
amphibole and chlorite. Journal of Metamorphic Geology 7, 507–513.
Knowles, C.R., 1987. A BASIC program to recast garnet end-members. Computers &
Geosciences 13, 655–659.
Lindsley, D.H., 1983. Pyroxene thermometry. American Mineralogist 68, 477–493.
Loss, R.D., 2003. Atomic weights of the elements 2001 (IUPAC Technical Report).
Pure and Applied Chemistry 75 (8), 1107–1122.
Massonne, H.-J., 1992a. Thermochemical determination of water activities relevant
to eclogitic rocks. In: Kharaka, Y.K., Maest, A. (Eds.), Water–Rock Interaction,
Proceedings of the Seventh International Symposium, Park City, vol. 2,
Moderate and High Temperature Environments, pp. 1523–1526.
Massonne, H.-J., 1992b. Thermodynamische Eigenschaften von Phasen des Systems
MgO–Al2O3–SiO2–H2O (MASH) unter besonderer Berücksichtigung von Mis-
chkristallreihen. Berichte der deutschen Mineralogischen Gesellschaft. ([Ther-
modynamic properties of phases in the system MgO–Al2O3–SiO2–H2O (MASH)
in particular consideration of solid solution series]). European Journal of
Mineralogy 4 (Beiheft 1), 186.
Massonne, H.-J., 1995. Experimental and petrogenetic study of UHPM. In: Coleman,
R., Wang, X. (Eds.), Ultrahigh Pressure Metamorphism. Cambridge University
Press, Cambridge, pp. 33–95.
Massonne, H.-J., Grosch, U., Willner, A.P., 1993. Geothermobarometrie mittels Ti-
Gehalten in Kalihellglimmern. Berichte der deutschen Mineralogischen
Gesellschaft ([Geothermobarometry using Ti-contents in potassium white
mica]). European Journal of Mineralogy 5 (Beiheft 1), 85.
1551
Massonne, H.-J., Kopp, J., 2005. A low-variance mineral assemblage with talc and
phengite in an eclogite from the Saxonian Erzgebirge, Central Europe, and its
P–T evolution. Journal of Petrology 46, 355–375.
Massonne, H.-J., Szpurka, Z., 1997. Thermodynamic properties of white micas on
the basis of high-pressure experiments in the systems K2O–MgO–Al2O3–
SiO2–H2O and K2O–FeO–Al2O3–SiO2–H2O. Lithos 41, 229–250.
Mogessie, A., Tessadri, R., Veltman, C.B., 1990. EMP-AMPH – a Hypercard program
to determine the name of an amphibole from electron microprobe analysis
according to the International Mineralogical Association scheme. Computers &
Geosciences 16, 309–330.
Mogessie, A., 2001. AMPH-IMA97: a Hypercard program to determine the name of
an amphibole from electron microprobe and wet chemical analyses. Compu-
ters & Geosciences 27, 1171–1180.
Montel, J., Foret, S., Veschambre, M., Nicollet, C., Provost, A., 1996. Electron
microprobe dating of monazite. Chemical Geology 131, 37–53.
Pyle, J.M., Spear, F.S., Wark, D.A., Daniel, C.G., 2005. Contributions to precision and
accuracy of monazite microprobe ages. American Mineralogist 90, 547–577.
Rao, D.R., Rao, T.V.S., 1996. AMPH: a program for calculating formulae for assigning
names to the amphibole group of minerals. Computers & Geosciences 22,
931–933.
Richard, L.R., Clarke, D.B., 1990. AMPHIOBOL: a program for calculating structural
formulae and for classifying and plotting chemical analyses of amphiboles.
American Mineralogist 75, 421–423.
Rock, N.M.S., Leake, R.E., 1984. The International Mineralogical Association
amphibole nomenclature scheme: computerization and its consequences.
Mineralogical Magazine 48, 211–227.
Rock, N.M.S., 1987. A FORTRAN program for tabulating and naming amphibole
analyses according to the International Mineralogical Association scheme.
Mineralogy and Petrology 37, 79–88.
Rock, N.M.S., Caroll, G.W., 1990. MINTAB: a general-purpose mineral recalculation
and tabulation program for Macintosh. American Mineralogist 75, 424–430.
Schumacher, J.C., 1991. Empirical ferric iron corrections: necessity, assumptions,
and effects on selected geothermobarometers. Mineralogical Magazine 55,
3–18.
Spear, F.S., Kimball, K.L., 1984. RECAMP – a FORTRAN IV program for estimating
Fe3+ contents in amphiboles. Computers in Geology 10, 317–325.
Stout, J.H., 1972. Phase petrology and mineral chemistry coexisting amphiboles
from Telemark, Norway. Journal of Petrology 13, 883–888.
Sturm, R., 2002. PX-NOM – an interactive spreadsheet program for the
computation of pyroxene analyses derived from the electron microprobe.
Computers & Geosciences 28, 473–483.
Tindle, A.G., Webb, P.C., 1994. Probe-AMPH – a spreadsheet program to classify
microprobe-derived amphibole analyses. Computers & Geosciences 20,
1201–1228.
Yavuz, F., 1996. Amphcal: a Quickbasic program for determining the amphibole
name from electron microprobe analysis using the IMA rules. Computers &
Geosciences 22 (2), 101–107.
Yavuz, F., 1999. A revised program for microprobe-derived amphibole analyses
using the IMA rules. Computers & Geosciences 25, 909–927.
Yavuz, F., 2003. Evaluating micas in petrologic and metallogenic aspect: I –
definitions and structure of the computer program MICA+. Computers &
Geosciences 29, 1203–1213.