LATE
-
DIAGENETIC ILLITIZATION IN MISSISSIPPIAN SANDSTONES OF THE 
MICHIGAN BASIN
 
 
By
 
 
Brian J. Ares
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
A THESIS
 
 
Submitted to 
 
Michigan State University
 
in partial fulfillment of the requirements 
 
for the degree of
 
 
Geological Sciences

Master of Science
 
 
2019
 
 
 
 
 
ABSTRACT
 
LATE
-
DIAGENETIC ILLITIZATION IN MISSISSIPPIAN SANDSTONES OF THE 
MICHIGAN BASIN
 
 
By
 
Brian J. Ares
 
Abundant authigenic clay mineral
s
 
(illite, kaolinite, chlorite), detrital mica grains (biotite, 
muscovite), and phyllosilicate
-
bearing lithic fragments (mudstone, phyllite) have been identified 
in the Mississippian age M
ar
shall and Michigan Stray
 
Sandstones of the Michigan Basin. These 
phyllosilicates were identified and analyzed with optical petrography, scanning electron 
microscopy,
 
transmission electron microscopy,
 
and X
-
ray diffraction. Many of these 
phyllosilicates exhibit morphologies and textural relationships indicating extensive alteration 
(
replacement, partial dissolution, neoformation
)
 
resulting
 
from changing pore fluid conditions. 
Illite
 
replacement of
 
prec
ursor kaolinite
 
and illite neoformation
 
are
 
observed
. 
Illite cement 
formation
 
typically
 
occurs 
at high temperatures (deep burial
), but occurs in these strata at 
depths 
shallower than most illite
-
bearing sandstones
. This cement formation 
requires an externa
l source 
of potassium. Multiple hypotheses have been proposed to account for this potassium. Here these 
hypotheses are evaluated to dete
rmine the cause of illite cement formation in the
se
 
Mississippian 
sandstones
. Illite formation is made thermodynamically
 
possible by the transport and mixing of 
deep saline brines in the Michigan Basin with Marshall pore fluids. Illite crystallinit
y data
 
indicate that authigenic illite formed 
at temperatures exceeding 200° C
.
 
Converging evidence 
suggests that
 
h
ot
 
brines 
fro
m deeper in
 
the Michigan Basin 
mixed with Mississippian formation 
waters to
 
initiate kaolinite illitization
 
and the neoformation of illite. These brines likely moved 
vertically along reactivated faults and fractures in the basin
.
iii
 
 
ACKNOWLEDGEMENTS
 
 
 
I would
 
like to sincerely thank my advisor, Dr. Michael Velbel, for his guidance and 
support throughout my entire project. I also thank him for introducing me to many other fantastic 
scientists that have helped me immensely throughout this project. I am grateful 
to my committee 
members, Dr. David Long and Dr. Susannah Dorfman
, 
for their time and effort as well as their 
roles in teaching me about some of the techniques used in this work. 
I would also like to thank 
Dr. Dalton Hardisty for providing me with 
bench space to complete my work.
 
 
Special thanks to Michigan State University and 
T
he Clay Minerals Society for providing 
financial support for this project, 
staff at the
 
Michigan Geological Repository for Research and 
Education
 
(specifically Dr. 
Bill
 
Harr
ison III
,
 
Dr. Pete Voice
, and Jenny Trout
)
 
for providing 
samples,
 
and
 
the MSU Center for Advanced Microscopy and Proto Manufacturing for assistance 
with my analytical work. 
I
t
 
truly
 
takes a village to complete a research project. 
 
 
I am thankful to all my 
friends in this department for helping make graduate school a fun 
experience, and in particular to my office mates for accepting a stray grad student into their 
office. I thank my parents for first getting me interested in science and supporting my academi
c 
goals throughout my entire education. Lastly, I would like to thank my girlfriend Vincenza for 
her continual confidence and support, and for putting up with all the long nights and 
practice 
talks
. 
 
 
 
 
 
 
iv
 
 
TABLE OF CONTENTS
 
 
LIST OF TABLES
 
vi
 
LIST OF 
FIGURES
 
vii
 
Introduction
 
1
 
 
Geologic Background
 
 
3
 
 
Clay Mineral Backgroun
d 
 
10
 
 
Preliminary Evaluation of Potassium
-
Delivery Hypotheses
 
1
9
 
 
Research Objective
 
 
2
4
 
 
Methods
 
2
6
 
 
Sampling
 
2
6
 
 
Optical Petrograph
y
 
2
6
 
 
Scanning Electron Microscopy
 
2
6
 
 
X
-
ray 
Diffraction
 
30
 
 
Transmission Electron Microscopy
 
3
5
 
 
Aqueous Geochemistry
 
3
5
 
 
Results
 
3
7
 
 
Sampling
 
3
7
 
 
Optical Petrography
 
3
7
 
 
Scanning Electron Microscopy
 
4
1
 
 
X
-
ray Diffraction
 
78
 
 
Transmission Electron Microscopy
 
8
6
 
 
Aqueous Geochemistry
 
9
2
 
 
Discussion
 
10
3
 
Well Log Interpretations
 
10
3
 
 
Paragenetic Sequence
 
10
3
 
 
Detrital Phyllosilicates
 
10
8
 
 
Diagenetic 
Chlorite and Kaolinite
 
1
1
1
 
 
Kaolinite
-
Illite Reaction
 
1
1
4
 
 
X
-
ray Diffraction
 
1
20
 
 
Transmission Electron Microscopy
 
1
2
3
 
Illite Cement 
Morphology
 
1
2
4
 
Geochemical Modeling
 
1
2
4
 
Re
-
evaluation of Potassium
-
Delivery Hypotheses
 
1
2
7
 
 
v
 
 
Recommendations for Future Work
 
1
31
 
 
Conclusions
 
1
3
4
 
 
APPENDI
CES
 
1
3
5
 
 
APPENDIX A: 
Point
-
count data
 
1
3
6
 
 
APPENDIX B: 
Thin section descriptions
 
1
40
 
 
APPENDIX C: 
Marshall aquifer groundwater data
 
1
4
9
 
 
APPENDIX D: 
Aqueous geochemical modeling well information
 
1
5
2
 
 
APPENDIX E: Pore fluid analyses used in aqueous geochemical modeling
 
1
5
3
 
 
APPENDIX F: 
XRD diffractograms
 
1
5
4
 
 
APPENDIX G: 
X
-
ray diffraction il
lite Ir number peak measurements
 
1
6
6
 
 
APPENDIX H: Transmission electron microscopy EDS data
 
16
7
 
 
REFERENCES
 
1
7
2
 
 
 
 
 
 
 
 
 
 
vi
 
 
LIST OF TABLES
 
 
Table 1: Wells sampled as part of this project
 
2
7
 
 
Table 2: M
arshall and Michigan Stray
 
Sandstone sampling depths
 
2
8
 
 
Table 3: 
Selected results from thin section point counting
 
3
8
 
 
Table 
4
: 
Selected mineral occurrences in samples observed by SEM
 
4
4
 
 
Table 
5
: 
Occurrence of dominant and minor illite morphologies based on SEM 
observations
 
4
7
 
 
Table 
6
: 
Comparison of relevant illite mineral formulae
 
5
3
 
 
Table 7: Minerals observed in powder XRD analysis
 
8
3
 
 
Table 
8
: 
Diagnostic peaks for mineral identification in
 
oriented mounts of
 
clay
-
size fractions
 
by pXRD
 
8
3
 
 
Table 
9
: Calculated Ir numbers and full width at half maximum (FWHM) values of illite 
 
(001) peaks for samples containing authigenic illite
 
8
5
 
 
Table 10
: Point
-
count data
 
1
3
6
 
 
Table 11
: 
Marshall aquifer groundwater data
 
14
9
 
 
Table 12
: Aqueous 
geochemical modeling well information
 
1
5
2
 
 
Table 13
: Pore fluid analyses used in aqueous geochemical modeling
 
1
5
3
 
 
Table 14
: X
-
ray diffraction
 
illite
 
Ir number peak measurements
 
1
6
6
 
 
 
 
 
vii
 
 
LIST OF FIGURES
 
 
Figure 1: 
Geologic map of the Michigan Basin
 
4
 
 
Figure 2: 
Michigan Lower Peninsula structural features
 
5
 
 
Figure 3: Michigan Basin stratigraphic column
 
7
 
 
Figure 4: Michigan Basin Mississippian stratigraphic column
 
9
 
 
Figure 
5
: Illite mineral structure
 
1
1
 
 
Figure 
6
:
 
Kaolinite morphology
 
1
5
 
 
Figure 
7:
 
Mi
chigan Stray/Marshall Sandstone paragenetic sequence
 
1
6
 
 
Figure 
8
: 
K
2
O
-
Al
2
O
3
-
SiO
2
-
H
2
O
 
mineral stability diagram
 
1
8
 
 
Figure 
9
: 
Potassium delivery model schematic
 
2
3
 
 
Figure 
10
: 
Sample well locations
 
2
7
 
 
Figure 1
1
: 
XRD sample preparation flowchart
 
31
 
 
Figure 1
2
: 
Kübler Index diagram
 
3
4
 
 
Figure 13: Grain size/sorting
 
3
9
 
 
Figure 1
4
: 
QFR ternary diagram
 
40
 
 
Figure 1
5
: 
Abundant lithic fragments
 
4
2
 
 
Figure 1
6
: 
Abundant mica grains
 
4
2
 
 
Figure 1
7
:
 
Pseudomatrix
 
4
3
 
 
Figure 1
8
: 
Skeletal K
-
feldspar
 
4
3
 
 
Figure 1
9
: 
Illite morphological classification images
 
4
6
 
viii
 
 
Figure 
20:
 
Detrital mudstone
 
4
8
 
 
Figure 
2
1
: 
Pseudomatrix EDS
 
4
8
 
 
Figure 2
2
: 
Pseudomatrix analysis
 
4
9
 
 
Figure 2
3
: 
Pseudomatrix SEM
 
4
9
 
 
Figure 2
4
: 
Authigenic illite fibrous morphology
 
5
1
 
 
Figure 2
5
: 
Illite
-
kaolinite relationship
 
5
2
 
 
Figure 2
6
: 
Uncoated quartz overgrowth
 
5
2
 
 
Figure 2
7
: 
Kaolinite morphology thin section
 
5
5
 
 
Figure 2
8
: 
Kaolinite morphology rock chip
 
5
5
 
 
Figure 2
9
: 
Kaolinite
-
K
-
feldspar relationship
 
5
6
 
 
Figure 
30
: 
Possible dickite 
morphology
 
5
7
 
 
Figure 
3
1
: 
Altered kaolinite crystals
 
5
9
 
 
Figure 3
2
: 
Authigenic chlorite
 
60
 
 
Figure 3
3
: 
Chlorite cement morphology
 
60
 
 
Figure 3
4
: 
Chlorite
-
carbonate relationship
 
6
2
 
 
Figure 3
5
: Chlorite
-
quartz overgrowth relationship
 
6
2
 
 
Figure 3
6
: 
Chlorite
-
kaolinite relationship
 
6
3
 
 
Figure 3
7
: 
Chlorite cement EDS
 
6
3
 
 
Figure 3
8
: 
Muscovite in lithic fragments
 
6
4
 
 
Figure 3
9
: 
Muscovite as grains
 
6
4
 
 
ix
 
 
Figure 
40
: 
Exfoliation of muscovite
 
6
5
 
 
Figure 
4
1
: 
Possible muscovite alteration
 
6
5
 
 
Figure 4
2
: 
Muscovite alteration EDS
 
6
6
 
 
Figure 4
3
: 
Abundant mica
 
6
7
 
 
Figure 4
4
: 
Deformed mica grain
 
6
9
 
 
Figure 4
5
: 
Exfoliation of biotite
 
6
9
 
 
Figure 4
6
: 
Biotite heterogeneity
 
70
 
 
Figure 4
7
: 
Possible biotite alteration to kaolinite
 
70
 
 
Figure 4
8
: 
Possible biotite alte
ration to chlorite
 
7
1
 
 
Figure 4
9
: 
Framboidal pyrite
 
7
2
 
 
Figure 
50
: 
Pore
-
filling pyrite
 
7
2
 
 
Figure 
5
1
: 
Abundant pyrite cement
 
7
3
 
 
Figure 5
2
: 
Quartz microstylolites
 
7
3
 
 
Figure 5
3
: 
Prismatic quartz overgrowth
 
7
5
 
 
Figure 5
4
:
 
Quartz Overgrowths
 
7
5
 
 
Figure 5
5
: 
Feldspar overgrowths
 
7
6
 
 
Figure 5
6
: 
Skeletal K
-
feldspar
 
7
6
 
 
Figure 5
7
: 
Preferential dissolution of K
-
feldspar overgrowths
 
7
7
 
 
Figure 5
8
: 
Authigenic carbonate morphology
 
7
9
 
 
Figure 5
9
: 
Zoned ankerite rhombs
 
7
9
 
 
x
 
 
Figure 
60
: 
Carbonate rhombs
 
80
 
 
Figure 
6
1
: 
Ti
-
O mineralization
 
80
 
 
Figure 6
2
: 
Representative diffractograms from clay sized fraction
 
8
2
 
 
Figure 6
3
: 
Michigan/Marshall illite crystallinity data
 
8
5
 
 
Figure 6
4
: 
Illite TEM crystal habit
 
8
7
 
 
Figure 6
5
: 
Illite la
th sizes
 
8
7
 
 
Figure 6
6
: 
Kaolinite TEM crystal habit
 
8
9
 
 
Figure 6
7
:
 
Selected area electron diffraction pattern #1
 
90
 
 
Figure 6
8
: 
Selected area electron diffraction pattern #2
 
9
1
 
 
Figure 6
9
: 
Marshall groundwater data
 
9
3
 
 
Figure 
70
: 
Marshall 
groundwater data with illite stability field
 
9
3
 
 
Figure 
7
1
: 
Geographic locations of relevant wells
 
9
4
 
 
Figure 7
2
: 
Modeled pre
-
illite solution
 
9
7
 
 
Figure 7
3
: 
Meteoric water mixing model results
 
9
8
 
 
Figure 7
4
: 
Devonian brine mixing model results
 
100
 
 
Figure 7
5: Silurian brine mixing model results
 
10
2
 
 
Figure 76: Representative well log
 
10
4
 
 
Figure 7
7
: Sample well and authigenic clay depth distribution
 
1
1
2
 
 
Figure 7
8
: Experimentally determined kaolinite
-
illite stability boundarie
s
 
1
1
5
 
 
Figure 7
9
: 
Marshall groundwater trend
 
11
8
 
 
xi
 
 
Figure 80
: 
W
-
1002 
XRD diffractograms
 
1
5
4
 
 
Figure 81
:
 
W
-
962 XRD diffractograms
 
161
 
 
Figure 82: W
-
1002 t
ransmission electron microscopy EDS data
 
16
7
 
 
Figure 83: W
-
962 transmission electron microscopy EDS data
 
171
 
  
1
 
 
Introduction
 
Diagenesis describes the suite of physical and chemical 
modification
 
to sediments and 
sedimentary rocks after deposition
 
(Berner, 1980)
. Creation and destruction of porosity from 
grain compacti
on, detrital grain dissolution/alteration, and secondary mineral 
precipitation/dissolution determine the present petrophysical state of the rock. Diagenesis in 
sandstones has long been the focus of extensive research due to the control these processes have
 
on reservoir quality in economically significant sedimentary basins. Understanding
 
the 
diagenetic history
 
of individual strata
 
is integral to interpreting observations and managing 
resources within 
an entire
 
basin
 
due to implications regarding resource ge
neration, migration, 
and storage.
 
Analysis of 
clastic 
diagenesis 
is a
 
critical component of overall reservoir
 
quality 
evaluation (Taylor et al., 2010)
, and therefore remains essential to perform. 
Clay cements 
specifically 
have the potential to 
control porosity and permeability within sandstone units
 
by 
occluding pore space
 
and restricting pore apertures
 
(Ehrenberg, 1990).
 
A
bundant clay 
cements
 
(illite, kaolinite, chlorite), and
 
uncommonly abundant
 
mica
 
grains
 
(biotite, muscovite) have been ident
ified within the Mississippian age 
Marshall 
and 
Michigan 
Stray 
Sandstones of the Michigan Basin (
Stearns, 1933; 
Westjohn and Weaver, 1998). 
Illite 
cement in sandstones typically 
occurs as grain
-
coating cement with a characteristic fibrous 
morphology (Wilso
n and Pittman, 1977). This cement is of interest to hydrologists and 
geologists due to the significant impact that widespread cementation of this type typically has on 
permeability 
-
 
potentially decreasing permeability by orders of magnitude 
relative
 
to no
n
-
cemented sandstones (Worden and Morad, 2003). 
 
 
 
 
2
 
 
Previous Work on Authigenic Illite in Sandstones
 
Precipitation of authigenic illite has 
a commonly present thermal property
 
-
 
heat to 
overcome the kinetic barrier for crystallization; and a
 
fundamental 
mass 
requirement
 
-
 
a supply 
of potassium ions to fill interlayer cation sites (Lanson et al., 2002). Previous studies have 
suggested the possibility of
 
illite cement precipitation
 
in sandstones
 
at temperatures as low as 
50° C (Bjorkum and Gjel
svik, 1988)
. The majority of authigenic illite, however, is believed to 
form at temperatures exceeding 100°
 
C (Lanson et al., 2002).
 
Th
e required
 
heat is often provided 
by burial 
beneath
 
overburden. The geothermal gradient increases with increasing burial,
 
so 
authigenic illite is more commonly observed in deeply buried 
strata
 
(Lanson et al., 2002). 
 
In 
many instances
 
d
iagenetic 
illite
 
is known to form in closed systems as the result of a 
reaction 
involving precursor kaolinite and
 
potassium feldspar. However, th
e destabilization of 
this kaolin
-
K
-
feld
spar assemblage is 
only 
thought
 
to be possible at 
high temperatures
 
(Lanson, 
2002). An alternative reaction pathway for the precipitation of illite 
at lower temperatures 
was 
proposed by
 
Berger et al. (199
7
) 
which involves the formation of illite from precursor kaolinite 
facilitated 
by influx into the system of
 
high K
+
/H
+
 
ratio fluid. 
For
 
this reaction to proceed a 
critical K
+
/H
+
 
ratio must be reached
 
(Berger et al., 1997)
. This ratio wou
ld need to be high 
enough that the fluid becomes highly 
supersaturated
 
with respect to illite. If this requirement is 
met then the fluid would also be 
supersaturated
 
with respect to K
-
feldspar, eliminating the 
possibility of an internal source of potassium
 
(Lanson et al., 2002)
. 
If an internal source of 
potassium is not provided 
it would be 
impossible to form
 
diagenetic illite 
without a pore
-
fluid 
compositional change. 
 
 
 
3
 
 
Previous Work on Study Area
 
Two 
models 
have been previously proposed to account for the delivery of potassium into 
these strata: (1) Downward 
percolation
 
of meteoric water interacting with potassium bearing 
aluminosilicates in the overlying Michigan Formation and transporting aqueous species in
to 
underlying porous sandstones (Zacharias, 1992). (2) Transport of potassium into porous 
sandstones via vertical cross
-
formational 
hot
 
brines rising along faults and fractures in the 
underlying units (Cox, 2015).
 
 
This study aims to evaluate the feasibili
ty of each proposed model using a combination of 
clay mineralogy, petrography, and aqueous geochemical modeling. Clastic diagenetic processes 
are results of the inter
actions
 
between mineral
s 
and 
pore fluids
;
 
therefore a study of such a 
system should aim to
 
consider both liquid and solid components
 
(Garrels and Mackenzie, 1971)
. 
Correlating aqueous chemical evolution with
 
diagenetic mineralogical changes
 
should provide 
the most comprehensive description of late diagenesis possible
 
(Worden and Burley, 2003)
. 
Additionally, 
increased
 
understanding of late diagenesis in
 
these
 
strata
 
would 
contribute to 
community understanding of basin
-
scale processes in the Michigan Basin and 
provide further 
analysis of porosity and permeability in strata with the potential for f
luid storage.
 
Geologic Background
 
 
The Michigan Basin is an intracratonic basin encompassing the 
L
ower 
P
eninsula of 
Michigan and extending into parts of Ontario, Ohio, Indiana, and Wisconsin 
(F
ig.
 
1
). The basin 
contains dominantly Paleozoic siliciclastic, carbonate, and evaporitic strata overlying 
Precambrian basement rock. The southernmost limb of the Prote
rozoic Mid
-
Continent Rift
 
 
 
 
4
 
 
Figure 1: 
Geologic map of the Michigan Basin
: Geologic map of the Michigan Basin showing 
bedrock. Mississippian units are outlined in the key to identify units of interest. Modified from 
Michigan Department of Natural Resource
s Land and Mineral Services Division (1999). 
 
 
 
 
 
 
 
 
 
5
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure 2: 
Michigan Lower Peninsula structural features
: 
Map Indicating the location of the Mid
-
Continent Rift (MCR) as well as major fault structures in the Michigan Basin
 
and the base of the 
Mississippian Marshall Sandstone.
 
From Westjohn and Weaver (1998), modified from Budai 
and Wilson (1991).
 
 
 
 
6
 
 
(MCR) runs NW
-
SE through the basin (Hinze et al., 1975). Structural features in the Paleozoic 
strata such as the Howell Anticl
ine and the Lucas and Albion
-
Scipio faults
 
(
Fig
.
 
2
) 
are thought 
 
to be associated with fault tectonics resulting from reactivation of basement faults in the MCR 
(Prouty, 1988)
.
 
 
The kinetics of illite formation 
commonly
 
require some degree of heating. Ofte
n this is 
accomplished 
by heating resulting from burial beneath overburden sediment
. In the Michigan 
Basin the thermal history is not fully understood (Velbel, 2009).
 
There is almost no
 
rock record 
during 
the 


except for
 
minor
 
Jurass
ic red
-
beds 
between
 
Carboniferous 
strata below and Pleistocene glacial till above (Velbel, 2009). Cercone (1984), and Cercone and 
Pollack (1991) reported signs of unexpectedly high temperatures in the basin based on vitrinite 
reflectance data
 
from Silurian
, 
Ordovician
, and Devonian
 
mudst
one
 
sampl
es,
 
and Pennsylvanian 
coal samples,
 
which could not be accounted for by overburden models alone
 
at normal mid
-
continent geothermal gradients
.
 
Further evidence of anomalous heating from apatite fission 
tracks
 
completed on crystalline basement core samples
 
(Crowley, 1991; Wang et al., 1994), 
dolomite fluid inclusion analysis
 
completed on Devonian samples
 
(Luczaj et al., 2006; Cox, 
2015), and geothermal modeling (
Nunn and Sleep, 1984; 
Vugrinovich, 1988; Wagenwel
t, 2015) 
suggest
s
 
that external sources of heat
 
(hydrothermal activity)
 
may be required to account for the 
observed phenomena.
 
Marshall Sandstone
 
 
The Marshall Sandstone is a
 
Lower
-
Mid Mississippian sandstone unit overlying the 
Coldwater Shale
 
(Fig. 3)
. 
S
e
dimentological features 
and sparse fossils 
as well as the presence of 
intercalated siltstone and mudstone suggest a shallow marine depositional environmental (Harrell 
et al., 1991). 
Attempts to reconstruct the subsurface geometry of the Marshall Sandstone 
f
ro
m 
 
7
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure 3: 
Michigan Basin stratigraphic 
column
: 
Stratigraphic
 
column indicating units in the 
Michigan Basin. Mississippian units of interest, the Marshall Sandstone and Michigan formation, 
are outlined by the red box. Modified from Ma et al. (2009).  
 
 
 
8
 
 
core analysis and core
-
log calibration reveal a maximum thickne
ss of 320 feet (Monnett, 1948; 
Adducci, 2015). 
Marshall thickness is at a maximum near
 
the center of the basin, 
with strata 
 
thinning to a minimum thickness of 1
30
 
feet as the strata pinch out to the north, south, and west 
(Monnett, 1948). 
The Marshall con
sists of two formally recognized, continuous sandstone 

 
(Fig. 4)
. The Lower Member ranges from 
sublitharenite
 
to 
very 
micaceous (
up to 
61%)
 
fin
e
 
grained 30
-
50 ft. basal layer of litharenite
 
based on point
-
count data from this study
.
 
This basal la
yer is not observed in every Marshall well, and 
the
 
exact subsurface geometry of 
this layer 
is
 
not well established
. 
The Napoleon Sandstone Member (or Upper Member) is 
predominantly fine
-
medium graine
d sublitharenite
 
based on point
-
count data from this study
. 
The presence of metamorphic phyllitic
-
schistose lithic fragments combined with evidence from 
detrital zircon signatures suggest metamorphic provenance of detrital grains likely originating 
from Carboniferous Appa
lachian highlands (Boothroyd, 2012). 
 
Michigan Formation 
 
 
The 
Upper
 
Mississippian Michigan Formation overlies the Marshall Sandstone. 
Lithologies include interbedded mudstone, limestone, dolomite, gypsum, anhydrite, and a basal 
layer of sandstone litholog
ically identical to the 
Napoleon Sandstone
 
and referred to informally 

 
The Michigan Stray sandstone occurs as interfingering 
deposits of sand, the maximum thickness of which is 70 feet (Monnett, 1948). The geometry of 
the 
Michigan Stray varies considerably throughout the basin, and at present there is no agreed 
upon interpretation for 
the
se
 
thickness variation
s
. 
Differentiation of the Michigan Stray from the 
Marshall is a subject of debate in the Michigan Basin
 
geological
 
c
ommunity. The basis for this 
distinction
 

band
 
of dolomite 
within
 
the Michigan 
Formation 
that
 
separates the Napoleon Sandstone from the Michigan Stray (
Monnett, 1948
).
 
The 
 
9
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure 4: 
Michigan Basin Mississippian stratigraphic column
: Detailed stratigraphic column of 
Mississippian rocks in the Michigan Basin. Not every member/unit identified in the stratigraphic 
column will be present in a given well because of the subsurface geometry o
f particular 
members.
 
 
 
10
 
 
most recent sedimentological interpretation of the Michigan Stray is that it is genetically related 
to the Marshall Sandstone
 
(Adducci, 2015)
;
 
it
 
is
 
for this reason
 
included in this study
. 
The 
occurrence of evaporites in the Michiga
n Formation indicate
s
 
a transition of the Michigan Basin 
from shallow
 
open
 
marine to a partially closed sabkha environment (Harrell et al., 1991). 
 
Clay Mineral Background
 
Illite
 
 
Illite is a 2:1 non
-
swelling dioctahedral potassium
-
bearing clay mineral (

r
odon and 
Eberl, 1984) with the general 
mineral
 
formula: 
 
(K)
0.6
-
0.9
(Al,
Fe,Mg
)
2
(Si,Al)
4
O
10
(OH)
2
 
The 2:1 structure of illite consists of two tetrahedral aluminosilicate 
sheets
 
sandwiching one 
dioctahedral 
sheet
 
with potassium ions in interlayer cation sites in order to balance the otherwise 
negative layer charge (
F
ig. 
5
). The amount of potassium in the mineral formula may vary due to 
changes in the charge of the 2:1 layer resulting from 
the 
varying 
cation substi
tution
 
in the 2:1 
layer
.
 
 
Illite is commonly represented in reactions and thermodynamic problems as muscovite 
(a
nother
 
dioctahedral phyllosilicate). 
Pure muscovite has a layer charge of one
, which comes
 
from 
the
 
1:3 Al:Si
 
ratio in tetrahedral sheets and octahedral sheets
 
that
 
are populated 
entirely 
with 
Al
3+
 
.
 
Muscovite
 
represents the trisilicic endmember of the trisilicic
-
tetrasilicic series of 
dioctahedral micas, with the tetrasilicic endmember being celadonite (Eslinger
 
and Pevear, 
1988)
, which contains 
no
 
Al
3+
 
.
 
Muscovite is chosen as the representative mineral for illite 
because of the Al
3+
 
in the crystal structure of illite. However, illite has the capability to contain 
 
 
 
11
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure 5: 
Illite mineral 
structure
: Hybrid polyhedral and ball
-
and
-
stick structural model of illite. 
Octahedral sheets are sandwiched between tetrahedral sheets. Interlayer cations show the 
location of potassium in the structure. Modified from Gualtieri et al. (2008). 
 
 
 
 
Tetrahedral
 
Octahedral
 
Tetrahedral
 
Interlayer 
 
Cation
 
12
 
 
divalent Fe
2+
 
and
 
Mg
2+
 
cations
 
in octahedral sites
 
as well
. 
Illite containing significant divalent 
substitution in octahedral sites would more closely resemble phengite,
 
the midpoint 
 
mineral in the trisilicic
-
tetrasilicic dioctahedral mica series (Eslinger and Pevear, 198
8). Phengite 
has a mineral formula (Bailey, 1980): 
 
K
(
Al
1.5
(Mg,Fe)
0.5
)
(Al
0.5
Si
3.5
)O
10
(OH)
2
 
This 
strict 
formula more closely approximates
 
illite than muscovite does. The difference between 
illite and phengite is the lower layer charge of illite, which is achievable by 
either 
decreasing the 
amount of divalent substitution in octahedral sheets or increasing the amount of 
Si
4
+
 
in 
tetrahedral s
heets. 
 
Illite is commonly a major component of detrital mudrocks, 
but
 
it 
also 
occur
s
 
in 
sandstones as a detrital component or grain
-
coating cement (Wilson and Pittman, 1977). 
Diagenetic illite forms 
by way of
 
one of two reaction mechanisms: transformation
 
or 
neoformation. Transformation is a type of alteration 
common in
 
phyllosilicate minerals in which 
a precursor mineral
 
alters to a secondary mineral through the direct incorporation and reuse of 
part of the structure of the precursor mineral
 
(
Segonzac, 1970). Neoformation refers to the 
creation of a secondary mineral via dissolution of the primary mineral followed by subsequent 
crystallization of the secondary mineral with its own unique 
structure
 
(Eberl, 1984). 
Neoformation of fibrous cements 
is common in clastic units containing abundant porosity, 
especially those with saline pore
 
fluid (Meunier and Velde, 2004). Neoformation is the dominant 
 
mechanism of illite formation in sandstones; however, both mechanisms have been found to 
occur in thes
e units. 
 
Authigenic illite of transformation origin typically forms from a precursor clay mineral, 
usually either smectite or kaolinite (Machhi, 1987; Lanson, 2002). The alteration process 
13
 
 
resulting in illite as the product mineral is referred to as illit
ization of the precursor mineral. Illite 
commonly occurs in interlayered illite/smectite (I/S) mixed layers in both detrital and 
diagenetic 
components. Pure illite is very rare; even diagenetically precipitated illite typically ha
s
 
at least 
5% smectite mix
ed layers (

rodon and Eberl, 1984). Determination of this smectite percentage is 
possible using X
-
ray diffraction (XRD) analytical techniques outlined in
 

rodon (1980)
 
and
 

rodon and Eberl (1984). 
 
 
T
he empirical observation of illite XRD peak narrowing 
wi
th increasing temperature ha
s
 
resulted in the use of illite XRD peak characteristics as a
 
rough
 
geothermometer in diagenetic
-
to
-
metamorphic
 
pelitic 
systems (Kübler, 1969). The loss of smectite layers with increasing 
temperature causes a noticeable narrowin


technique prior to its increased usage amongst clay mineralogists in the academic community 
(Kübler and Jaboy
edoff, 2000). Advances in XRD instrument accuracy and diffraction modeling 
software programs like NEWMOD
 
and 
WILDFIRE
 
have resulted in the continued use of the 
Kübler Index into the 21
st
 
century (Lee and Lee, 2001; Battaglia et al., 2004; Aldega et al., 20
07; 
Doublier et al., 2010; Laughrey et al., 2011; Mählmann and Frey, 2012; Warr and Cox, 2016).
 
Kaolinite
 
Kaolinite is a 1:1 non
-
swelling aluminosilicate clay mineral with the 
mineral
 
formula:
 
Al
2
Si
2
O
5
(OH)
4
 
The 1:1 structure consists of alternating silicon tetrahedral 
sheets
 
and aluminum bearing 
octahedral 
sheets.
 
The 3+ charge of aluminum necessitates a vacancy in every third cation site in 
the octahedral
 
sheet
 
in order to balance charge, making kaolinite a 
dioctahedral clay. It is most 
14
 
 
common as a weathering product in near
-
surface rocks rich in feldspars
 
(Nesbitt and Young, 
1989)
. 
 
Kaolinite cement
s
 
in sandstones have a distinct morphology
 
(Wilson and Pittman, 1977)
. 
Cements form as a series of pseudo
-
hexag
onal plates stacked on top of one another in a 

 
(
F
ig. 
6
)
. These booklets are pore
-
filling cements, as 
opposed to grain
-
coating cements
 
(Wilson and Pittman, 1977)
. Kaolinite cements are known to 
be precursor minerals for diagenetic 
formation of
 
gibbsite, chlorite, and illite, although instances 
of kaolinite as a product in reactions involving aforementioned minerals are 
not uncommon
 
(Bjorkum and Gjelsvik, 1988).
 
Ka
olinite
 
and Illite Thermodynamics
 
Phyllosilicate minerals exhibit a variety of morphologies and alteration textures in these 
units, the occurrences of which reflect diagenetic mineral
-
water interaction
s
. Alteration of 
primary minerals into secondary produc
t minerals is controlled by reactions in which reactant 
minerals interact with aqueous species in pore fluid to yield product minerals and new aqueous 
species (Berner, 1980). Each reaction that occurs alters the composition of pore fluid, adjusting 
the che
mical conditions within the given unit and driving new minerals to react to eventually 
reach thermodynamic stability. The sequence of events in which these reactions
,
 
as well as 
physical processes like compaction
,
 
occur is referred to as the paragenetic se
quence. A 
 
paragenetic sequence can be thought of as a solid
-
state record of geochemical alteration (in 
addition to other mechanical processes). Diagenetic formation of phyllosilicates comprises the 
last two stages of the paragenetic sequence within th
e
 
Ma
rshall/Michigan Stray Sandstones
 
(Fig. 
7
). Authigen
esis of
 
kaolinite and illite are respectively the two most recent 
diagenetic 
events in 
this
 
paragenetic
 
sequence.
 
15
 
 
Figure 6: 
Kaolinite morphology
:
 
Secondary electron (SE) scanning electron 
microscope (SEM) 
image of vermiform booklet kaolinite cement morphology. Booklets consist of a series of 
pseudohexagonal plates stacked on top of one another. From Osborne et al. (1994).
 
 
 
 
16
 
 
Figure 7: 
Michigan Stray/Marshall Sandstone paragenetic sequence
:
 
Petrographically 
determined paragenetic sequence detailing the relative timing of diagenetic modifications to the 
Michigan Stray and Marshall Sandstones. Dashed lines indicate tentative interpretation. Red 
lines indicate events that are not present throug
hout all intervals of these units
; chlorite (red) 
occurs only in intervals presently at 0 
-
 
200 m depth, whereas illite (
blue
) occurs only in intervals 
presently at 400 
-
 
450 m depth
. Modified from Zacharias (1992).
 
 
 
 
17
 
 
K
aolinite chemical precipitation from
 
a solution is heavily influenced by leaching of 
meteoric water into groundwater during shallow, early diagenesis 
(Blatt, 1979; Hurst and Irwin, 
1982)
. Weathering and recharge environments with large quantities of rainfall and snowmelt will 
result in relat
ively ion
-
diluted naturally acidic meteoric water, which will favor the formation of 
kaolinite at the expense of K
-
feldspar by the reaction (Bjorkum and Gjelsvik, 1988): 
 







































 
 
(1)
 
 
An alternative scenario 
resulting in kaolinite formation by this same reaction has been 
observed as a result of naturally acidic waters entering sandstones along contacts with adjacent 
shales due to hydrocarbon maturation and dehydration of these shales (Ehrenberg, 1991). In such
 
instances, kaolinite is observed along the contacts, but not pervasive throughout the unit. In the 
presence of
 
abundant atmospheric precipitation and
 
high flushing rates, near
-
surface leaching 
will proceed to alter diagenetic kaolinite to aluminum hydroxi
de gibbsite (Velbel, 1985). This 
reaction, as well as reaction 1 and all other K
2
O
-
Al
2
O
3
-
SiO
2
-
H
2
O
 
system reactions, are 
represented as boundary lines in
 
mineral stability diagram
s
 
(Fig. 
8
), making these diagrams 
useful for evaluating chemical changes in sy
stems containing these minerals. 
Groundwater 
composition may be plotted on these diagrams, and thermodynamic equilibria of the 
groundwater with one or more minerals may be inferred based on wh
ere
 
the data plot.
 
 
Authigenic kaolinite is common as a 
cement in sandstones, usually due to the availability of
 
aluminum from
 
feldspar in clastic rocks 
(Blatt, 1979; Hurst and Irwin, 1982)
.  
Reaction 1 is not 
kinetically inhibited by temperature (instances of the K
-
feldspar
-
kaolinite reaction at surface 
conditions are plentiful), so the reaction is free to proceed in porous rocks containing feldspar 
and dilute pore
-
fluid. 
 
 
18
 
 
Figure 8: 
K
2
O
-
Al
2
O
3
-
SiO
2
-
H
2
O
 
mineral stability 
diagram:
 
Activity
-
activity mineral stability 
diagram for the K
2
O
-
Al
2
O
3
-
SiO
2
-
H
2
O mineral system. Boundary lines constructed using data 
from Robie et al., 1979 (corrected edition).
 
 
 
19
 
 
The potential for substitution in both the 
tetrahedral and octahedral sheets and the 
resulting variable composition make illite a difficult mineral to properly treat in terms of 
thermodynamic stability.
 
Thermodynamic datasets often do not contain much information about 
illite, and those interested 
in performing calculations or writing reactions frequently use 
muscovite as a proxy. 
As previously discussed, this practice is often insufficient in representing 
the actual composition and layer charge of illite.
 
One technique for studying potassium 
alumin
osilicates involves using the 
K
2
O
-
Al
2
O
3
-
SiO
2
-
H
2
O
 
mineral stability diagram 
to 
graphically display mineral thermodynamic stability (Fig. 
8
). Attempts have been made by 
previous workers to include a discrete illite phase in th
is
 
diagram (Sass et al., 1987; A
ja et al., 
1991; Yates and Rosenberg, 1996; Aja, 1997; Noguera et al., 2011). Such attempts are based on 
experimental reactions; therefore, they do not perfectly represent illite reactions in natural 
systems. 
 
Preliminary Evaluation of Potassium
-
Delivery H
ypotheses
 
Meteoric Water 
Infiltration
 
One potential mechanism of potassium delivery into Mississippian strata of the Michigan 
Basin is the 
infiltration
 
of meteoric water enriched with potassium 
sourced 
from dissolution of 
K
-
feldspar in overlying clastic strata 
(
Zacharias
, 
1992). The feasibility of this hypothesis relies 
on the following background information: age dates of the diagenetic illite in Mississippian strata
 
(from previous work only)
,
 
the
 
prese
nce of diagenetic kaolinite in Mississippian strata
 
(previous 
work and this study)
, and the potential for a reaction in which kaolinite
/aqueous aluminum and 
silica react with potassium
 
to yield illite
 
(this study)
. 
 
20
 
 
Unpublished K/Ar age dating of diageneti
c illite from the Marshall Sandstone w
ere
 
reported
 
by Cox (2015). 
T
hree size fractions (2
-
1
 
µ
m, 1
-
0.5
 
µ
m, and <0.5
 
µ
m) from two 
sampling locations w
ere
 
dated
. Analysis of diagenetic illite is conventionally performed using 
the smallest possible size fracti
on, as detrital illite particles are typically larger (
>2
 
µm)
 
than 
diagenetically formed particles (

rodon, 198
4
)
.
 

 
(2015)
 
age dates for the smallest size 
fractions 
in the two samples 
are 275 Ma and 294 Ma. This variability could represent 
either 
multiple illitization events or a gradual illitization process. Regardless, these dates constrain the 
timing of illite formation to the Early Permian
.
 
Assuming hydr
ological
 
connectivity of unit
s 
deposited in the Michigan Basin during this interval
,
 
the possibility of meteoric water dissolving 
K
-
feldspar in clastic units (Michigan Formation, Saginaw Formation, Grand River Formation
, 
and any post
-
Grand River Group pre
-
Jurassic red
-
bed
 
units that a
re now missing
) and 
transporting that potassium downward into Mississippian strata should not be overlooked. 
 
The formation of illite requires some source of aluminum and silica as well as potassium. 
The
 
aluminum and silica
 
are provided by a precursor mine
ral in transformation reactions
, or
 
the 
aqueous 
components
 
of a dissolved precursor mineral in neoformation reactions
. Conventional 
illitization of mudrocks involves smectite functioning as the precursor mineral and altering to 
illite 
by
 
transformation (Se
gonzac, 1970). However, no discrete smectite has been observed 
by
 
petrograph
y
 
or XRD analysis in Mississippian sandstones in the Michigan Basin. Therefore, 
illitization from a precursor would require a different mineral as a reacta
nt.
 
Kaolinite 
may
 
alter t
o illite 
by
 
one of two possible reactions (Lanson et al., 2002). One 
potential reaction 
is expressed
 
below
 
(the author chose muscovite to represent illite)
:
 



















































 
(2) 
 
21
 
 
in which kaolinite reacts with 
dissolving potassium feldspar to yield illite, aqueous silica, and 
water. This reaction is only possible i
n
 
the presence of kaolinite and actively dissolving K
-
feldspar. 
Both
 
are
 
present in the Marshall Sandstone and Michigan Formation,
 
but the kaolinite
-
p
otassium feldspar assemblage is stable 
at
 
temperatures greater than what would be expected 
from a burial geotherm
 
in these strata
 
(Lanson et al., 2002)
. A second kaolinite
-
illite forming 
process is possible via the reaction
 
(the author chose muscovite to r
epresent illite)
:
 













































 
 
(3)
 
in which kaolinite reacts with aqueous 
K
+
 
ions to yield illite
, 
excess hydrogen
 
ions,
 
and water. 
This reaction
,
 
proposed by Berger
 
et al.
 
(1997)
,
 
has been suggested by Lanson et al. (2002) to 
occur in instances where kaolinite is already present within a sandstone, then K
+
-
rich
 
fluids 
invade the formation 
satisfying the thermodynamic requirement
 
for the reaction
 
to proceed
. 
Reaction 2
 
is
 
considered 
the conventional process resulting in authigenic illite formation in 
sandstones at high temperatures
 
(Lanson et al., 2002)
, although 
reaction 
3
 
is 
more 
feasible 
at 
lower temperatures 
provided invasion of K
+
-
rich
 
fluid occurs
 
(Berger et al., 1997
)
. 
 
Previous reactions have expressed 
illite formation by transformation of kaolinite
. Illite 
has also been neoformed in these strata. Neoformation of illite from ions in pore fluid would 
proceed by the reaction: 
 








































 
 
 
(4)
 
T
his reaction precipitates 
illite 
directly 
from 
aqueous ions in pore fluid. It should be noted that
 
reaction 4 requires the same reactant 
ions
 
as the previous transformation reactions, the difference 
being that reactants in reaction 4 are aqueous. Initiation of reaction 4 still requires
 
an
 
additional 
22
 
 
source of potassium
, aluminum, and silica. These reactants 
are
 
supplied by dissolution of K
-
feld
spar, or dissolution of kaolinite plus additional potassium
.
 
Mobilization of enough K
+
 
to initiate reaction
s
 
3 
or
 
4 
in the Marshall
/Michigan 
would 
require sufficiently large amounts of K
-
feldspar in the overlying strata and a large enough 
flushing rate 
of meteoric water to destabilize this K
-
feldspar and transport the K
+
 
down
 
by 
percolation. Provided those requirements are met, K
+
 
ions would certainly be introduced to the 
Mississippian strata. 
 
 
If the proposed hypothesis of potassium enriched meteoric water 
infiltration
 
were true 
then two lines of supporti
ng
 
evidence would exist: (1) aqueous geochemical mixing models 
should indica
te that thermodynamic stability of illite is 
at
tainable 
by
 
invasion of such water
, and
 
(2) the temperature of formation of the diagenetic illite would be firmly within the range of 
shallow diagenesis
.
 
Vertical Deep Brine Movement
 
 
An alternative mechanism 
for
 
potassium delivery to Mississippian
-
aged clastic strata is 
the vertical cross
-
formational movement of hot saline brines from older strata in the Michigan 
Basin. Previous researchers have suggested that deep brines migrate vertically into younger 
strata
 
along faults and fractures
 
in the Michigan Basin
 
(Girard and Barnes, 1995; Luczaj et al., 
2006; Castro et al., 2009). Such hot brine movement has been invoked as the cause of previously 
unexplained hydrothermal mineralization in some strata of the Michiga
n Basin (Machel and 
Lonnee, 2002). Evidence of some degree of brine mixing based on helium isotope concentrations 
and major element chemistry has already been observed in the Marshall 
a
quifer specifically (Ma 
et al., 2005). Movement of this brine is 
thought to be possible due to the presence of
 
basement 
 
23
 
 
Figure 9: 
Potassium delivery model schematic
: Graphic representations of two proposed K
+
 
delivery models. A: K
+
 
delivery by percolation of meteoric water enriched by dissolution of 
overlying 
K
-
feldspar. B: K
+
 
delivery by infiltration of hot brines from deep within the basin.
 
 
 
 
 
A
 
B
 
24
 
 
Mid
-
Continent Rift (MCR) faults and fractures which have been reactivated prior to the time of 
such hydrothermal mineralization. Vertical hot brine movement would also
 
provide support for 
the idea that a hydrothermal event may be responsible for the high thermal indicators
 
(vitrinite 
reflectance data, apatite fission track data, authigenic mineralization)
 
observed in the Michigan 
Basin (Cox, 2015). 
 
Helium and neon values of deep basinal brines analyzed by Castro et al. (2009) were 
found to be very similar to mid
-
oceanic ridge values, which suggests that these brines may have 
an upper mantle or deep baseme
nt source. Assuming hydrologic connectivity of Michigan Basin 
 
formations along preferential flow paths from fracturing/faulting as previous researchers suggest, 
fluid transportation to younger strata would be possible from
 
any
 
deeper formation. Aqueous 
ge
ochemical analys
e
s by Wilson and Long (1993) show that the 
Devonian 
and Silurian
-
aged 
units 
contain
 
highly concentrated brines with TDS levels ranging from 200,000
-
450,000 ppm. 
Such highly concentrated brines may be sufficiently ion rich to drive kaolinite
-
illite reaction 3
 
or 
reaction 4
 
towards thermodynamic equilibrium and sufficiently hot to overcome the kinetic 
barrier
 
of illite crystallization
.
 
 
If the deep hot brine hypothesis were true then different supporti
ng
 
evidence would be 
expected: (1) aqueous
 
geochemical mixing models should indicate that thermodynamic stability 
of illite is 
at
tainable 
by
 
mixing of Marshal
l
 
pore
-
fluid with known basinal brines
, and
 
(2) the 
temperature of formation of diagenetic illite should be 
greater than
 
the shallow diagene
tic regime
 
(anchizone or epizone).
 
Research Objective
 
This project

s objective is to combine new and existing data for solid and liquid phases in 
Mississippian sandstones of the Michigan Basin to better understand the reactions and processes 
25
 
 
influencing 
late diagenesis (
epi
diagenesis
; terminology of Worden and Burley (200
3)
) 
of th
ese
 
strata. The Marshall Sandstone and Michigan Formation contain textbook examples of 
authigenic clay mineral
s
, but with
 
present 
depths and 
corresponding expected 
thermal conditions 
that deviate from the norm. This project aims to evaluate the fe
asibility of two previously 
proposed late
-
diagenetic processes to 
better constrain
 
how the formation of authigenic illite in 
shallow
 
(modern depth)
 
sandstone 
was
 
possible. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
26
 
 
Methods
 
Sampling
 
 
Subsurface samples were obtained from cores 
housed at the Michigan Geological 
Repository for Research and Education (MGRRE) in Kalamazoo, Michigan. Marshall Sandstone 
and Michigan Formation cores of interest were identified through literature review and 
conversation with MGRRE faculty (William Harri
son, personal communication, 2018) and 
selected with the goal of encountering diagenetic facies while maximizing variation in depth and 
pore
 
fluid composition. Four cores were selected for analysis (
T
able 1, 
F
ig.
 
10
). Ten to fifteen 
evenly spaced intervals
 
of interest within each core were cut and sampled to allow broad 
coverage of the Marshall and Michigan Stray Sandstones 
from each well
 
(Table 2)
. 
 
Optical Petrography
 
 
Traditional optical petrography was completed using a petrographic transmitted
-
light 
mi
croscope. 
Forty
-
six
 
thin sections (one thin section per sample collected) were produced and 
stained with blue
-
dye epoxy to 
identify
 
open pore space. 
These thin sections were then analyzed 
and point
-
counted using 200 points per thin section to obtain quanti
tative mineral abundance
 
data
. 
Optical petrography was completed with the goal of initial identification of mineral grains, 
characterization of textures, and 
preliminary characterization
 
of diagenetic processes 
(paragenetic sequence). Petrographic analysis
 
was completed for every sample collected.
 
Scanning Electron Microscopy
 
 
Scanning electron microscopy (SEM) provides multiple advantages over optical 
petrography when studying clay minerals and micron
-
scale mineral alteration. High resolution 
 
 
 
27
 
 
Table 1: Wells sampled as part of this project
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure 
10
: 
Sample well locations
: Geographic location of sample wells. Numbers correlate to 
well numbers from Table 1.
 
The location of the Mid
-
Continent Rift
 
as well as major fault 
structures in the Michigan Basin
 
and the base of the Mississippian Marshall
 
Sandstone
 
are 
indicated by the black lines. Modified from Westjohn and Weaver (1998).
 
28
 
 
Table 2: M
arshall and Michigan Stray
 
Sandstone sample depths
 
 
29
 
 
capabilities of SEM allow for determination of clay morphologies from polished thin sections 
and rock 
chip fragments (Wilson and Pittman, 1977). Subtle co
nstituents
 
of difficult
-
to
-
 
distinguish clay matrix material, such as identification of silt sized detrital fragments and 
oriented fabric of detrital micas, are more easily visible using SEM relative to o
ptical 
microscopy. 
 
 
Polished thin sections and rock chip
 
pairs
 
were produced for 31 samples
 
(each pair 
prepared from the same sample)
 
spanning each of the four sampled cores
 
(
T
able 2
). Thin 
sections were coated using a carbon thread evaporator
, attached t
o the metal stage using sticky 
carbon tabs,
 
and then analyzed on a JEOL 6610LV SEM with backscattered electron (BSE) 
imaging and an Oxford Instruments energy dispersive X
-
ray spectroscopy (EDS) 
system
.
 
BSE 
imaging was co
nducted
 
in high vacuum mode 
using
 
an
 
accelerating voltage of 20 kV, a
 
working 
distance of 10 mm, a spot size of 53, 
and 
a 

z

 
height of 10 mm.
 
The heterogenous nature of the 
brightness observed while viewing a flat (polished) surface in BSE mode is due to differences in 
Z contrast. In electr
on microscopy Z is the average atomic number for all elements in a phase. 
Materials with larger Z values appear brighter in the image, whereas materials with low Z values 
appear darker in the image. 
BSE imaging allows for 
rapid preliminary 
identification of 
mineralogy and determination of subtle compositional differences within individual grains
 
(White et al., 1984)
. Energy dispersive X
-
ray spectroscopy (EDS) provides semi
-
quantitative 
compositional analysis at the micron scale
 
(White et al.
, 1984)
. When paired with BSE, 
identification of single
-
grain alteration and subsequent determination of product phases are 
possible. 
 
 
Secondary electron (SE) imaging of freshly fractured rock chips was completed to 
investigate in
-
situ 
mineral morphologic
al relationships. Freshly fractured rock chips were broken 
30
 
 
off from samples, coated in a gold sputter coater, 
attached to a
n aluminum stub (diameter 25 
mm)
 
using sticky carbon tabs, 
and analyzed using SE imaging and EDS. 
SE imaging was 
conducted in high va
cuum mode using
 
an accelerating voltage of 10 kV,
 
a working distance of 
25 mm, a spot size of 30,
 
and
 
a 

z

 
height of 25
 
mm
.
 
X
-
ray Diffraction
 
 
Powder X
-
ray diffraction (pXRD) has long been the primary analytical method for clay 
mineralogists concerned with 
identification, 
composition, crystallography, or quantification of 
clay minerals (Moore and Reynolds, 1989). 
 
Samples for pXRD analysis were s
elected using an illite morphology classification 
scheme to exclude samples in which dioctahedral potassium
-
bearing phyllosilicates are not 
diagenetic. Selected samples were required to contain observable illite with morphologies 
consistent with diagenetic
 
neoformation (grain coatings/pore linings) based on SEM 
observations. Samples containing pseudomatrix or ambiguous morphologies were excluded from 
pXRD analysis to avoid contamination by detrital illite. 
 
 
Sample preparation is outlined in Figure 11. 
C
lay
 
mineral XRD sample preparation 
begins with separating the clay
 
size fraction (equivalent spherical diameter of < 2
 
µm
) of bulk 
rock samples
.
 
. 
Disaggregation of bulk rock samples was completed by gentle crushing with a 
ceramic mortar and pestle. The samples were rinsed with deionized water and placed into an 
ultrasonic bath for five minutes to facilitate further disaggregation. The suspension was
 
then set 
aside to allow separation of the clay size fraction (< 2 
µ
m) using settling rates determined by 
Stokes law. The < 2 
µ
m clay size fraction was then decanted into the beakers, where the 
suspended clay size fraction was allowed to sediment onto the 
quartz disc. Sedimentation results 
 
 
 
31
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure 11: 
XRD sample preparation flowchart
:
 
Sample preparation procedure for powder X
-
ray 
diffraction of the clay size fraction.
 
 
 
32
 
 
in oriented XRD samples, as the platy crystal habit of 
clay minerals will cause them to settle out 
of suspension with their basal planes flat against the mount.
 
Clay coatings on the mounting medium are very thin when prepared using this method, 
and as a result the generated X
-
rays may interact with the crystal
 
structure of the mounting 
medium itself. To mitigate this issue low background fused
-
quartz discs were used as the 
mounting medium. These discs still exhibit some degree of background noise, which is 
expressed on XRD patterns as a slight hump from ~6
-
11° 
2

. Discs were placed at the bottom of 
250 ml beakers. 
The clay size fraction 
so
 
prepared on a quartz/glass mounting medium
 
was
 
then 
ready for 
insert
ion
 
into the diffractometer. 
 
An X
-
ray beam is generated by a cathode ray tube (
typically 
copper or chromiu
m)
,
 
then 
passed through a filter (
typically 
nickel or iron) to
 
produce a 
monochromatic 
beam
 
of 
desired
 
K

 
radiation
 
which is then aimed at a sample.
 
The X
-
rays scatter upon interaction with the sample, 
and much of the scattering 
results in
 
destructive interference. 
In the instrument configuration 
routinely used in clay mineralogy, t
he angle 
at
 
which the beam interacts with the sample (

) is 
gradually c
hanged to scan for instances when constructive interference of X
-
rays occurs, 

satisfied,
 
and diffracted X
-
rays can be recorded by the detector. 
The intensities of diffracted rays are recorded by either a point or a slit detec
tor over a range of 
angles. The 
peak 
positions 
of a sample are determined by the
 
composition
, 
geometry
, and atom 
spacing
 

. Clay minerals can be identified from diagnostic XRD 
peaks, provided peak overlap is not too severe. 
In bulk rock samples quartz typically cannot be 
separated entirely from the clay
-
size fraction, although this is not a problem as quartz may be 
used as an internal standard in pXRD analysis of clastic
 
sedimentary
 
rocks (Moore and 
Reynolds, 1989). 
 
33
 
 
Prepared
 
pXRD mounts
 
were taken to Proto Manufacturing in Taylor, Michigan for 
analysis. XRD analysis was completed using a Proto Manufacturing AXRD Benchtop 
diffractometer with a Cu tube, a nickel filter, a slit 
receiving detector (0.1 mm)
,
 
a divergence slit 
(0.5
 
mm), 
and a rotating stage to
 
reduce the effect
s of 
textural heterogeneity within 
the sample. 
No anti
-
scatter slit was used. This instrument uses a voltage of 30 kV, and a current of 20 mA. 
The diffractometer was set to record XRD patterns from 4
-
40° 2

 
with a 5 second dwell time
 
and 
0
.02° 2

 
step size 
(resulting in a scan speed of 
0
.24° 2

 
per minute) 
with binning set at 
2
.
 
The sample was then inserted into a glycolation chamber overnight at 
65
°C and analyzed again 
the following day. Ethylene glycol is 
frequently used in pXRD studies of clay minerals to 
analyze potential swelling components of a sample
 
(Moore and Reynolds, 1989). Glycolation 
results in 
expansion of interlayer sites in low layer charge smectite
-
group minerals
 
and 
consequent modification of
 
the 
(
001
)
 
peak for smectite and mixed layer illite/smectite. Th
is 
modification 
appears as a shift of the 
(
001
)
 
peak to a lower angle
 
due to expansion of the 
smectite 
layer spacing
 
and 
can be analyzed to determine the degree 
of expandable material in the 
sample.
 
 
 
The Kübler Index of illite crystallinity (IC) was used to evaluate the 
temperature of illite 
authigenesis
. The Kübler Index is an empirical metric 
determined
 
by measuring the full width at 
half the maximum (FWHM) of 
the illite 
(
001
)
 
diffraction peak in units of °2

. This value is then 
plotted on the formational regime graph (
F
ig. 
1
2
)
.
 
The boundaries of the anchizone in Kübler 
Index are intended to be adjusted based on the XRD scan rate. The boundaries of the anchizone
 
used in this diagram are after Weber (1972) and Teichmüller et al. (1979) and were determined 
using Kisch (1990).
 
34
 
 
Figure 12: 
Kübler Index diagram
: Kübler Index usage diagram indicating thermal regime of illite 
formation. Boundaries for the anchizone defined using Kisch (1990).
 
 
 
 
35
 
 
Transmission Electron Microscopy
 
 
Transmission electron microscopy (TEM) is commonly used along with pXRD to 
characteriz
e structure and crystallography of clay minerals (Mackinnon, 1990). TEM analysis 
provides a look at the crystal habit of individual illite fundamental particles (Nadeau et al., 
1984), as well as high resolution compositional analysis and the potential to d
istinguish 
authigenic illite polytypes by selected area electron diffraction (SAED) (Grubb, 1991). 
Polytypism refers to a specific type of polymorphism in phyllosilicate minerals in which the      
 
chemical formula remains the same, but the layer
-
stacking sequence of individual sheets vary in                   
 
their pattern of repetition (Smith and Yoder, 1956). Controls on illite polytypism are not yet fully 
understood, although there is a suggeste
d relationship between specific polytypes and 
hydrothermal formation of authigenic illite (Grubb, 1991; Bauluz et al., 2002).
 
TEM samples were prepared by suspending the clay size fraction of bulk rock samples in 
water as for pXRD, then pipetting one drop 
of the suspended solution onto lacey carbon coated 
Cu TEM grids. The solution was then allowed to air
-
dry. Samples were analyzed on a JEOL 
2200FS TEM equipped with an Oxford Instruments EDS. Samples were viewed in high
-
vacuum 
mode using bright
-
field imagin
g with an accelerating voltage of 10kV. Spot size was adjusted as 
needed. Images were captured using a Gatan digital camera. 
Acquisition time for EDS analysis 
was set to two minutes per sample to mitigate 
data quality reduction
 
associated with low dead 
tim
e. The duration of electron beam interaction with each sample varied.
 
Aqueous Geochemistry
 
 
Aqueous geochemical
 
mixing and forward
 
model
ing
 
w
as
 
conducted to evaluate the 
feasibility of 
achieving
 
illite
-
solution
 
thermodynamic 
equilibrium
 
under different geological 
scenarios. 
 
36
 
 
 
Groundwater chemical data for 476 wells within the Michigan Basin w
ere
 
collected as 
part of the Regional Aquifer System Analysis (RASA) USGS program and published in USGS 
 
Open
-
File Report 90
-
368 (Dannemiller and Ba
ltusis Jr., 1989).
 
Many of these sample wells were 
in the Marshall Aquifer
, so ample groundwater chemical data are available
 
for this study
. Data 
 
relevant to K
2
O
-
Al
2
O
3
-
SiO
2
-
H
2
O systems were compiled from the 90 groundwater wells 
reported for the 
Marshall aquifer in the RASA report (Appendix).
 
Pore fluid
 
compositions resulting from mineral
-
water interactions were modeled using 
the USGS software package PHREEQC
.
 
T
he Debye
-
Hückel equations
 
were used
 
for dilute
 
solution calculations, and the Pitzer eq
uations for brine solution (>10,000 ppm TDS) 
calculations. 
A d
atabase was constructed containing 
the appropriate information fro
m the
 
Debye
-
Hückel and Pitzer databases
 
that are provided as default databases in PHREEQC
. 
Real and
 
hypothesized
 
fluid compositi
ons were used as initial inputs for aqueous mixing models r
u
n in 
PHREEQC.
 
 
 
 
 
 
 
 
 
37
 
 
Results
 
Sampling
 
 
Forty
-
two
 
samples of Marshall
 
(8 basal Lower, 4 Lower, 21 Napoleon)
 
and Michigan 
Stray
 
(9)
 
Sandstone
 

 
were 
collected from four wells (
T
able 2). Thin sections were prepared from every sample collected. 
Thirty
-
one polished thin sections were prepared for BSE SEM analysis
 
(Table 2)
. Twelve XRD 
samples were prepared from two of the four wells
 
(Table 2)
. 
T
wo samples
 
were prepared for 
TEM
 
(Table 2). 
 
Optical Petrography
 
 
Optical petrographic examination was performed 
to characterize and screen all prepared 
samples, and 
to s
elect samples for 
SEM thi
n section analysis. Diagenetic components identified 
include quartz overgrowths, feldspar overgrowths, authigenic carbonate, kaolinite cement, 
chlorite cement, illite cement, and pyrite cement.
 
Individual t
hin section descriptions and raw 
point
-
count data 
are pr
ovided
 
in the Appendix. Th
ese
 
data w
ere
 
used to calculate quartz
-
feldspar
-
rock fragment (QFR) ratios for each sandstone sample 
following the methods of
 
Folk 
(
1974)
. QFR, mica abundance, and grain size/sorting data for all sandstone samples are presen
ted 
(Table 3).
 
Sandstone samples are very fine/
fine
-
grained and
 
are dominantly moderately 
well/well
-
sorted (Fig
.
 
13). 
These QFR ratios were then plotted on a ternary diagram (Fig. 1
4
) and 
used to classify each sample.
 
 
Lithologically, sandstone samples include 
sublitharenite, 
litharenite, 
feldspathic 
litharenite, 
and greywacke. 
Sublitharenite and litharenite are the dominant lithologies of these 
 
38
 
 
Table 3: Selected results from thin section point counting.
 
39
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure 13: 
Grain size/sorting
: Sandstone sample grain size and sorting. Boundaries for grain 
sorting after Folk (1974).
 
 
40
 
 
 
Figure 1
4
: 
QFR ternary diagram
: Marshall and Michigan Stray sandstone quartz
-
feldspar
-
rock 
fragment (QFR) ratio ternary diagram. Fifty percent of all sandstone samples are classified as 
litharenites, and 40.5% of all sandstone samples are classified as sublitharenites. Lithological 
cla
ssifications are after Folk (1974). 
 
 
41
 
 
units. Of the 42 sandstone samples point
-
counted, 40.5% were sublitharenites, and 50% were 
litharenites. There does not appear to be a relationship between depth and lithology, as most 
samples from all four wells plot 
within similar sublitharenite
-
litharenite ranges. Rock
 
fragments
 
consist primarily of slightly deformed mudstone and phyllite fragments as indicate
d by the
 
high 
birefringence grains (
F
ig. 
1
5
).
 
Feldspar
s
 
consist of plagioclase and K
-
feldspar
 
(microcline and
 
orthoclase)
.
 
Mica grains (biotite and muscovite) are 
present
 
throughout all four wells but 
constitute
 
between 25.5% 
-
 
61.5%
 
of thin sections in certain intervals of the RASA 5 well (
F
ig. 
1
6
). Wells W
-
1002 and W
-
962 show intervals of very little 
porosity
 
(0% 
-
 
0.5%)
, as pore space 
appears to be clogged almost entirely by deformed detrital pseudomatrix (
F
ig. 1
7
). 
Classification
 
of this material as pseudomatrix is based on SEM data and will be discussed in the SEM results 
section. Dissolution of K
-
f
eldspar grains is common and perv
asive
 
throughout all sample wells 
(
F
ig. 1
8
).
 
Scanning Electron Microscopy
 
 
Scanning electron microscopic analysis was completed for all 31 polished thin section
-
rock chip pairs 
with representation from
 
all four sample wells
 
(Table 2)
. The presence/absence of 
phyllosilicate minerals, as well as authigenic pyrite, was documented for each thin section and is 
presented (
T
able 
4)
. Identification
s
 
of minerals are based on a combination of diagnostic 
morphology/crystal habit and ED
S compositional analysis. EDS analyses of polished thin 
sections are generally considered semi
-
quantitative, while EDS analysis of minerals in rock chip 
fragments cannot be used to confidently determine mineralogy and are supportive evidence at 
best. Ident
ification of textural relationships 
by
 
scanning electron microscopy (combining rock 
chip and thin section data) is the basis for revision of the paragenetic sequence in these strata 
from Zacharias (1992)
 
(Fig. 
7
)
.
 
42
 
 
Figure 1
5
: 
Abundant lithic 
fragments
: 
Optical PPL/XPL m
icrograph of lithic fragments in 
Marshall Sandstone A: Plane polarized light view B: Cross polarized light view. Lithic fragments 
consist primarily of mudstones and phyllites. High birefringence in XPL indicates detrital mica. 
F
rom W
-
1002 well at 1356 ft. measured depth.
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure 1
6
: 
Abundant mica grains
:
 
Optical PPL micrograph of a
bundant detrital mica grains 
oriented with a preferential fabric. Deformation of mica grains, most likely due to compaction, 
nearly elimi
nates porosity. From RASA 5 well at 604 ft. (depth). 
 
43
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure 1
7
: 
Pseudomatrix
: 
Optical PPL
 
view of detrital pseudomatrix consisting of reworked 
mudstone lithic fragments. From W
-
1002 well at 1315 ft. (depth). 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure 1
8
: 
Skeletal K
-
feldspar
: 
Optical PPL view of s
keletal K
-
feldspar grain indicated by the 
red circle. Blue dyed 
epoxy represents pore space. From W
-
962 well at 1430 ft. (depth).
 
44
 
 
Ta
ble
 
4: Selected mineral occurrences in samples observed by SEM
.
 
 
 
 
 
 
 
 
 
 
45
 
 
Illite
 
 
Illite was observed as detrital and diagenetic components in two of the four sample wells. 
Illite was 
present in wells W
-
1002 and W
-
962
 
(Table 
4
)
, both of which are located in the central 
part of the basin, in Clare County
 
(Fig. 
10
)
. These two wells have nearly identical depth ranges 
and pore fluid compositions 
(Table 1) 
and are geographically close to one another. Additionally, 
these two wells are deeper than the two sample wells 
in which
 
illite was not observed. Illite was 
observed
 
with several different morphologies. A classification scheme consisting of four 
different groupings was created to differentiate between detrital and diagenetic morphologies. 
The morphologies comprising this illite classification scheme are 
(1) a constitu
ent of 
detrital 
pseudomatrix
, 
(2) illite with isotropic fabric of ambiguous origin, and 
fibrous cement
,
 
distinguishing between 
(
3
) 
complete grain
-
coating and 
(
4
) 
patchy distributions (
F
ig. 1
9
). This 
textural classification
 
was 
used to
 
identif
y
 
intervals
 
co
ntaining only diagenetic neoformed illite 
for us
e
 
in pXRD study. Table 
5
 
displays the burial depths and morphological classifications of 
observed illite. 
 
 
Detrital illite was observed in two forms: mudstone lithic fragments, and pseudomatrix. 
Mudstone lit
hic fragments 
are made up of small grains that exhibit different brightness when 
viewed in BSE mode due to differences in Z contrast, thereby showing the fine
-
grained material 
to 
consist of a mixture of multiple clay minerals and 
commonly
 
inclu
d
ing
 
quartz 
(
F
ig. 
20
).
. 
Detrital illite was also observed as pseudomatrix
 
(Fig. 
2
1
)
. This material was determined to be of 
detrital origin due to the existence of silt sized quartz
 
particles interspersed throughout the illite 
clay, and evidence of local fabric in small mica grains interspersed in the material (Fig. 22). This 
pseudomatrix is interpreted to be reworked mudstone and metasedimentary material, and 
porosity is negligible 
in intervals where it is observed (Fig. 23, Table 5).
 
 
46
 
 
Figure 1
9
: 
Illite morphological classification 
images
: Representative 
BSE/SE
-
SEM images of 
each illite textural group. A: Detrital pseudomatrix B: Isotropic illite of 
ambiguous origin. C: 
Authigenic fibrous illite (partial grain
-
coating distribution) D: Authigenic fibrous illite (full 
grain
-
coating distribution).
 
 
 
A
 
B
 
C
 
D
 
47
 
 
Table 5: Occurrence of dominant and minor illite morphologies based on SEM 
observation
.
 
 
 
Replaced kaolinite
 
48
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure 
20
: 
Detrital mudstone
: 
BSE
-
SEM thin section 
image
. Detrital mudstone lithic fragment 
surrounded by quartz grains. From W
-
1002 well at 1448 ft. (depth).
 
 
Figure 2
1
: 
Pseudomatrix EDS
: 
BSE
-
SEM. 
EDS profile of 
detrital pseudomatrix. The location of 
the EDS analysis is shown by the yellow circle. From W
-
1002 well at 1315 ft. (depth).
 
49
 
 
Figure 2
2
: 
Pseudomatrix analysis
: 
BSE
-
SEM image of d
etrital pseudomatrix containing silt 
sized quartz grains and mica grains with 
a preferential orientation. The red arrows indicate the 
direction of the mica fabric. The white circle indicates the target area of EDS analysis. EDS 
analysis reveals chemical composition resembling K
-
bearing phyllosilicate minerals. From W
-
1002 well at 13
15 ft. (depth).
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure 2
3
: 
Pseudomatrix 
SE
M
: 
BSE
-
SEM image of d
etrital pseudomatrix filling pore space and 
nearly eliminating porosity. Brighter grains indicate high atomic number (feldspar). From W
-
1002 well at 1315 ft. (depth). 
 
50
 
 
 
Diagenetic illite has the classic fibrous morphology typical of illite
 
cements
 
in sands
tones 
(Boles and Franks, 1979). Illite wisps form as grain coating cements on the surface
s
 
of detrital 
grains (
F
ig. 
2
4
). The lengths of these wisps vary from 1
-
10 
µ
m. These illite cements, although 
present, are difficult to observe in polished thin section and are more 
easily
 
viewed in 
 
rock chips using SE imaging. Illite cement can be observed nucleating on authigenic kaolinite 
plates (
F
ig. 
2
5
), indicating that illi
te precipitation is occurring after kaolinite precipitation in the 
paragenetic sequence. These cements do not always coat the outsides of quartz overgrowths (
F
ig. 
2
6
), suggesting that quartz overgrowth precipitation may occur concurrently 
with 
or later tha
n 
illite precipitation. 
 
 
EDS compositional analysis of illite was attempted. The chemical composition of detrital 
illite is of limited importance
 
to this project
;
 
therefore
,
 
only authigenic grain coatings were 
analyzed by EDS
. The effectiveness of EDS ana
lysis is dependent 
on an area within the field of 
view containing an object of interest large enough relative to surrounding features that the signal 
received by the EDS detector 
is generated only from the feature of interest. When viewing illite 
in thin s
ection the difficulty lies in finding an area of illite cement large enough that a small 
portion can be selected without including the detrital grain it is coating. When viewing illite in 
rock chips the quality of the EDS analysis
 
is considered qualitative
 
at best 
due to the effects of 
geometric modifications. Sample topography can modify the X
-
ray absorption path to the 
detector, thus deviating from the ideal absorption path required for highest quality analysis 
(Newbury and Ritchie, 2013). Therefore, rock
 
chip EDS data generated should be used 
 
cautiously. EDS analyses of grain coating illite cements were completed on 42 areas and 
revealed an average illite mineral formula of:
 
K
0
.735
(Al
1.697
Fe
(II)
0.113
Mg
0.18
)(Si
3.567
Al
0.424
)O
10
(OH)
2
 
51
 
 
 
 
 
 
 
 
 
 
 
 
Figure 24: 
Authigenic illite fibrous morphology
: SE
-
SEM image from a fractured rock chip. 
Fibrous authigenic morphology of grain
-
coating illite cements. Small prismatic quartz 
overgrowths are also observed nucleating on the grain. From W
-
1002 well at 1406 
ft. (depth).
 
 
 
52
 
 
Figure 2
5
: 
Illite
-
kaolinite relationship
:
 
SE
-
SEM image of a
uthigenic illite nucleating on and 
around authigenic kaolinite plates. From W
-
962 well at 1453 ft. (depth).
 
Mineral abbreviations 
are after Whitney and Evans (2010) and will be used for the remainder of this study. 
 
 
 
 
 
 
 
 
 
 
 
Figure 2
6
: 
Uncoated quartz 
overgrowth
: 
SE
-
SEM image of i
llite
 
cement
-
coated quartz grain. 
Prismatic quartz overgrowths are not coated by
 
illite, suggesting that they may have formed 
cogenetically or later than illite cements. From W
-
1002 well at 1406 ft. (depth).
 
 
 
ln
 
53
 
 
Table 6: Comparison of relevant illite mineral formulae. 
 
 
 
 
 
54
 
 
This formula
 
is displayed alongside other relevant illite mineral formulae (Table
 
6
). It
 
fits firmly 
within the 
generalized 
definition of illite
 

.  
Illite in this study has 
phengitic substitutions 
(tetrahedral Al:Si ratio and degree of substitu
tion divalent cations in the 
octahedral sheet) 
to a limited extent that results in 
layer charge (and corresponding interlayer site 
occupancy) firmly within the illite range.
 
Kaolinite
 
Kaolinite was observed in all four of the sample wells. 
Kaolinite occurs as pore filling 
cement (
F
ig. 
2
7
) with a classic booklet
-
like vermiform morphology consisting of a series of 
plates stacked on top of one another (
F
ig. 2
8
). Kaolinite cements were 
commonly
 
observed 
near
 
skeletal potassium feldspar grains (
F
ig. 2
9
), suggesting that formation of kaolinite cement is 
occurring at the expense of 
K
-
feldspar 
as in
 
reaction 1. This observation also suggests that 
conditions within this system do not allow the mobilization of the 
aluminum and silica
 
from
 
this 
K
-
felds
par and will result in the crystallization of kaolinite cement 
in close proximity to the 
dissolving grain
, as suggested by
 
Velbel 
(
1983)
. 
 
The thicknesses of individual kaolinite plates appeared to be 0.5
-
1 
µ
m for most cements 
observed. However, the RASA 5
 
well contained kaolin material (as determined by EDS) with 
significantly thicker plates. Individual plates comprising RASA 5 kaolin cements have 
t
hicknesses 
of
 
~
5
-
10 
µ
m
 
(
F
ig. 
30
). These observations 
correspond to
 
the morphology of dickite,
 
a high temperature kaolinite polymorph, as reported by Beaufort et al.
 
(
1998
)
. This tentatively 
classified dickite is
 
observed in the 
(
intermediate depth range
)
 
RASA 5 well. 
 
EDS analyses of kaolinite cements were conducted peri
odically to confirm the 
identification of vermiform material as kaolinite. Classification of a mineral as kaolinite from 
 
55
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure 2
7
: 
Kaolinite morphology thin section
:
 
BSE
-
SEM image of p
ore
-
filling vermiform 
morphology of kaolinite 
cements. From CH101 well at 20 ft. (depth).
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure 2
8
: 
Kaolinite morphology rock chip
: 
SE
-
SEM
 
image of pore filling kaolinite cements. 
Vermiform morphology consists of a series of pseudohexagonal stacked plates
. A
 
mineral with a 
blocky morphology (possible dickite) is present as well
. From CH101 well at 20 ft. (depth).
 
 
56
 
 
Figure 2
9
: 
Kaolinite
-
K
-
feldspar 
relationship
: 
BSE
-
SEM images of m
ultiple
 
instances of 
kaolinite cements forming adjacent to skeletal K
-
feldspar grains. From RASA 5 well at 519 ft. 
(depth).
 
 
 
Qz
 
Qz
 
Qz
 
Qz
 
Qz
 
Kfs
 
Kfs
 
Kfs
 
Kfs
 
Kfs
 
Kln
 
Kln
 
Kln
 
Kln
 
57
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure 
30
: 
Possible dickite 
morphology
: 
A
: 
SE
-
SEM
 
image of kaolin cement. Note the increase 
in thickness of particular crystals. B: 
BSE
-
SEM
 
image of kaolin cements. The yellow circle 
indicates an unusually thick kaolin crystal. From RASA 5 well at 519 ft. (depth).
 
 
58
 
 
EDS analysis is based on the atomic per
centages of O, Si, and Al. Kaolinite will have < 1% of 
any other element, and the Si:Al ratio must be exactly or very close to one. Most kaolin cements 
exactly matched the mineral formula for kaolinite. However, occurrences of kaolinite 
morphology with ano
malous composition were observed in wells W
-
1002 and W
-
962 (
F
ig. 
3
1
). 
Such 
occurrences
 
were identified initially by particular booklets appearing brighter in BSE 
imaging mode, indicating 
a
 
chemical composition 
with a higher
 
average 
ato
mic
 
number
 
than 
kaolinite. EDS analysis of 
five 
anomalous booklets revealed an average 
mineral formula
 
of
:
 
K
0.7
38
(Al
1.
8
2
9
Fe
(II)
0.
098
Mg
0.
107
)(Si
3.
212
Al
0.7
88
)O
10
(OH)
2
 
which falls within the compositional range of illite 

. The octahedral 
cations of this formula resemble the average mineral formula of authigenic grain coating illite, 
but the tetrahedral cations differ 
(Table 
6
)
.
 
Classification of t
hese crystals as illite is based on 
EDS analyses indicating values very close to kaolinite, but with less Al and an atomic percentage 
value of  > 3% for K.
 
These observations suggest that the anomalous booklets are instances of 
pseudomorphic replacement of kaolinite 
by
 
illite
, confirming illitization of kaolinite has occured.
 
 
Chlorite
 
Chlorite was observed in all four of the sample wells. Detrital chlorite 
was observed in all 
sample wells, and diagenetic chlorite was observed in wells RASA 5 and CH101 (Table 4). 
Detrital chlorite was observed as components of detrital mudstone and phyllite lithic fragments. 
Diagenetic chlorite was observed as grain coating c
ements with individual crystals displaying a 
disc
-
like morphology. These pore linings are easily visible in thin section (Fig. 32) and in rock 
chips (Fig. 33). These cements are observed coating other diagenetic features such as carbonate
 
59
 
 
Figure 3
1
: 
Altered kaolinite crystals
:
 
BSE
-
SEM images
 
showing crystals with kaolinite 
morphology but anomalous composition. Abnormal crystals are indicated by red boxes. The 
difference in brightness is due to a higher average atomic number than the surrounding kaolinite. 
A: From W
-
1002 well at 1344 ft. (dept
h) B: From W
-
962 well at 1425 ft. (depth) C: From W
-
1002 well at 1356 ft. (depth) D: From W
-
1002 well at 1444 ft. (depth).
 
 
 
 
A
 
B
 
C
 
D
 
60
 
 
Figure 3
2
: 
Authigenic chlorite
: 
BSE
-
SEM images of a
uthigenic grain
-
coating chlorite. From 
CH101 well at 85 ft. (depth).
 
 
 
Figure 3
3
: 
Chlorite cement morphology
: 
SE
-
SEM images of b
laded morphology of authigenic 
grain
-
coating chlorite. From CH101 well at 85 ft. (depth).
 
61
 
 
rhombs (Fig. 3
4
), indicating that chlorite cement precipitation occured after carbonate 
precipitiation (Fig. 
7). Quartz overgrowths occur on uncoated sections of detrital quartz grains, 
and then crystaliz
e
 
over chlorite rims (Fig. 3
5
) indicating that quartz overgrowth precipitation 
occured after chlorite
 
cement precipitation in the paragenetic sequence
 
(Fig. 
7
)
. 
Chlorite 
lines
 
pores that contain pore
-
filling kaolinite (Fig. 3
6
). The kaolinite itself is not coated by chlorite, 
indicating that chlorite precipitated before kaolinite
 
(Fig. 
7
)
. 
 
 
EDS analysis of chlorite was conducted solely to confirm the identifica
tion of chlorite 
rims once per thin section
 
(Fig. 3
7
)
. Chlorite 
was
 
not the primary focus of this project so further 

 
 
Muscovite
 
 
Muscovite is present as a detrital component in all four o
f the sample wells
 
(Table 
4
)
. 
Muscovite can be observed as components of metasedimentary lithic fragments (
F
ig. 3
8
) and 
dis
crete
 
grains (
F
ig. 3
9
). Grains of muscovite are often 
deform
ed
 
around
 
rigid detrital grains of 
quartz and feldspar, resulting in 
reduction
 
of porosity. Exfoliation of muscovite grains (
F
ig. 
40
) 
i
s
 
common. 
 
 
Exfoliation of muscovite grains 
is
 
commonly
 
observed immediately next to diagenetic 
kaolinite (
F
ig. 
4
1
). 
This figure 
d
epicts what appear to be muscovite layers altering to kaolinite 
near exfoliated edges. The change in Z contrast along with EDS analysis indicate that layers are 
becoming depleted of potassium near the edges (Fig. 42). Additionally, distinct kaolinite 
bookl
ets are observed within exfoliated layers and immediately surrounding the muscovite. 
These observations suggest the alteration of muscovite into kaolinite. 
Along with biotite, 
muscovite abundance increases dramatically in the deeper interval of the RASA 5 
well (590
-
610 
ft. depth). Muscovite and biotite comprise up to 50% of rocks in this interval (Table 3) and
 
62
 
 
 
 
 
 
 
 
 
 
Figure 34: 
Chlorite
-
carbonate 
relationship
: SE
-
SEM image of a
uthigenic
 
chlorite coating an 
authigenic carbonate rhomb. From CH101 well at 114 ft. (depth). 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure 35: 
Chlorite
-
quartz overgrowth 
relationship
: BSE
-
SEM image of q
uartz
 
overgrowth 
overlying grain
-
coating chlorite cement. The overgrowth nucleated in 
area of incomplete cement 
coverage and grew into open pore space. From CH101 well at 133 ft. (depth).
 
63
 
 
 
 
 
 
 
 
 
 
 
 
Figure 3
6
: 
Chlorite
-
kaolinite relationship
:
 
SE
-
SEM image of a
uthigenic chlorite coating a quartz 
grain. Authigenic 
kaolinite can be observed filling the pore space outside of the coated grain. 
Note that the kaolinite is not coated with chlorite. From RASA 5 well at 519 ft. (depth).
 
 
Figure 3
7
: 
Chlorite cement EDS
: 
BSE
-
SEM image of c
hlorite cement with an 
associated EDS 
chemical composition. The white circle indicates the location of the EDS analysis. From CH101 
well at 85 ft. (depth).
 
Kl
n
 
Chl
 
64
 
 
Figure 3
8
: 
Muscovite in lithic fragments
:
 
BSE
-
SEM images of l
ath shaped muscovite crystals in 
detrital metasedimentary lit
hic fragments. From W
-
1002 well at 1448 ft. (depth).
 
 
 
Figure 3
9
: 
Muscovite as grains
:
 
BSE
-
SEM imag
e
s of m
uscovite as the dominant component of 
discrete muscovite grains. From W
-
962 well at 1448 ft. (depth).  
 
 
 
 
 
 
 
 
 
65
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure 
40
: 
Exfolitaion of muscovite
: 
BSE
-
SEM image of e
xfoliation of a muscovite grain along 
interlayer plane. From W
-
962 well at 1425 ft. (depth). 
 
 
Figure 4
1
: 
Possible muscovite alteration
:
 
BSE
-
SEM images of a
uthigenic kaolinite forming 
immediately adjacent to exfoliating muscovite. A: From W
-
962 well at 1430 ft. (depth) B: From 
RASA 5 well at 598 ft. (depth).
 
 
 
A
 
B
 
66
 
 
Figure 4
2
: 
Muscovite alteration EDS
: 
BSE
-
SEM image and 
EDS analyses of different sections 
of a potentially altering 
muscovite grain. Note the difference in potassium between the two 
analyses. From W
-
962 well at 1430 ft. (depth).
 
 
 
67
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure 4
3
: 
Abundant mica
:
 
 
BSE
-
SEM image of m
ica
-
rich interval within the RASA 5 well. 
Detrital mica grains comprise >50% of 
the sample. From RASA 5 well at 607 ft. (depth).
 
 
 
68
 
 
have clear fabric in the horizontal direction parallel to bedding (Fig. 4
3
).
 
 
Biotite
 
 
Biotite is present in three of the four sample wells (not present in W
-
1002). Biotite is 
observed only as dis
crete
 
grains. As with muscovite, biotite grains are 
commonly
 
deformed
 
between rigid detrital grains (
F
ig. 
4
4
). Exfoliation of biotite grains along the 
interlayer
 
is 
common (
F
ig. 
4
5
). 
 
 
Biotite grains 
commonly
 
show compositional alteration along individual 
layers, giving 
grains an internally mottled appearance when viewed with BSE imaging (
F
ig. 
4
6
). Termination
s
 
of individual layers are 
commonly
 
associated with compositional changes within these grains. 
Some
 
grains appear to
 
have
 
alter
ed
 
(compositionally and
 
texturally) into poorly crystallized 
minerals of ambigous morphology 
with 
composition
s
 
similar to kaolinite (
F
ig. 
4
7
). Other grains 
show unus
u
ally wide (3
-
4 
µ
m) spacing between individual layers, with authigenic chlorite 
nucleating on and between 
these layers (
F
ig. 4
8
). It is unclear whether this unique appearance in 
BSE imaging is due to chlorite cements precipitating on biotite layers and growing into the 
accomodation space
 
produced by exfoliation
, or alteration of biotite to chlorite.  
 
 
Pyrite
 
 
Diagenetically formed pyrite is present in all four of the sample wells and can be 
observed in almost all samples analyzed. Pyrite is easily distinguishable in BSE imaging due to 

c pyrite occurs as 
framboidal clusters precipitating in open pore spaces (Fig. 49), and pore filling cement (Fig. 50). 
 
One locality (depth of 33 feet) was identified within the CH101 well that contains pore space 
which is entirely cemented with pyrite, co
mpletely occluding porosity (Fig. 51). 
 
 
69
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure 4
4
: 
Deformed mica grain
: 
BSE
-
SEM image of d
etrital biotite grain deforming around 
more rigid detrital grains. From RASA 5 well at 531 ft. (depth).
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure 4
5
: 
Exfoliation of 
biotite
:
 
BSE
-
SEM image of e
xfoliation of biotite grains. From W
-
962 
well at 1453 ft. (depth).
 
70
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure 4
6
: 
Biotite heterogeneity
: 
BSE
-
SEM image of c
ompositional heterogeneity of biotite 
grain. Different Z contrasts display differences in average 
atomic number. From W
-
962 well at 
1453 ft. (depth).
 
 
Figure 4
7
: 
Possible biotite alteration to kaolinite
:
 
BSE
-
SEM image of p
ossible alteration of 
detrital biotite grain to kaolin composition material. The yellow oval represents the target area of 
EDS anal
ysis. From RASA 5 well at 598 ft. (depth).
 
71
 
 
Figure 4
8
: 
Possible biotite alteration to chlorite
: 
BSE/SE
-
SEM images of a
uthigenic chlorite 
forming on and around detrital biotite grains. Note that chlorite appearts to be forming in the 
interlayer space of 
biotite. It is unclear if biotite is altering to chlorite or chlorite cements are 
nucleating directly on biotite layers. The top two images are BSE from thin sections, and the 
bottom two images are SE from rock chips. From CH101 well at 114 ft. (depth).
 
 
 
 
72
 
 
Figure 4
9
: 
Framboidal pyrite
: 
BSE
-
SEM images of di
agenetically formed pyrite clusters with 
framboidal morphology. A: From W
-
1002 well at 1356 ft. (depth) B: From CH101 well at 133 ft. 
(depth).
 
 
 
 
Figure 
50
: 
Pore
-
filling pyrite
: 
BSE
-
SEM images of d
iagenetically formed pyrite with pore
-
filling 
morphology. A: From W
-
962 well at 1328 ft. (depth) B: From CH101 at 58 ft. (depth). 
 
 
 
A
 
B
 
A
 
B
 
73
 
 
Figure 5
1
: 
Abundant pyrite 
cement
: 
BSE
-
SEM images of a
nomalously
 
abundant diagenetic 
pyrite. Pore
-
filling cement completely occludes porosity. Grains appear darker on screen due to 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure 5
2
: 
Quartz 
microstylolites
:
 
BSE
-
SEM image of m
icrostylolites along quartz grain contacts 
resulting from pressure solution. Microstylolites have no preferred orientation. From W
-
962 well 
at 1453 ft. (depth).
 
74
 
 
Quartz
 
 
Detrital quartz co
nstitutes
 
the majority of grains in these units
 
(Fig. 1
4
, Table 3)
. Quartz 
grains are primarily subangular in shape. Many quartz grains display microstylolites as a result 
of pressure solution caused by 
compact
ion
 
(
F
ig. 
5
2
). 
 
Diagenetic quartz overgrowths are commo
n in grains from all four sample wells. 
Overgrowths on these grains generally do not precipitate around the entirety of the grain surface. 
Rather, overgrowths are observed to be prismatic in shape and precipitate
 
as smaller overgrowths 
that do not engulf t
he entire grain
 
(
F
ig. 
5
3
). These overgrowths appear as euhedral blocky bulges 
with straight edges in thin section (
F
ig. 
5
4
). 
 
Potassium Feldspar
 
Potassium feldspar is present in all four sample wells. K
-
feldspar 
is
 
observed as detrital 
grains comprising between 
0
-
11.
1
%
 
of detrital grains throughout different intervals
 
(Table 3)
. 
Diagenetic overgrowths are common in grains from all wells (
F
ig. 
5
5
), indicating previous 
supersaturation of pore fluid with respect to K
-
f
eldspar. Dissolution of K
-
feldspar is also 
common in grains throughout all sample wells. Dissolution of K
-
feldspar is infe
r
red from the 
presence of skeletal K
-
feldspar grains (Fig. 56), indicating previous undersaturation of pore fluid 
with respect to K
-
fe
ldspar. 
 
Many K
-
feldspar grains were observed with diagenetic overgrowths that appear to be 
preferentially dissolving relative to the detrital grain on which they precipitated (Fig. 57). This 
indicates that pore fluid became undersatured with respect to K
-
feldspar after the precipitation of 
diagenetic overgrowths. It is also possible that there were multiple periods of pore fluid 
undersaturation with respect to K
-
feldspar and the most recent period of such conditions occured 
after the precipiation of these 
overgrowths. This preferential dissolution could be a result of 
insufficient residence time of undersaturated pore fluid, insufficient circulation of pore fluid
 
75
 
 
 
 
 
 
 
 
 
 
 
 
Figure 5
3
: 
Prismatic quartz overgrowth
: 
SE
-
SEM image of q
uartz grain displaying prismatic 
overgrowths. Overgrowths are not nucleating around the entirety of the grain, but rather as 
prisms in select areas. From W
-
1002 well at 1356 ft. (depth). 
 
 
 
Figure 5
4
: 
Quartz overgrowths
:
 
BSE
-
SEM images of e
uhedral 
quartz overgrowths with sharp, well 
defined edges. From W
-
962 well at 1338 ft. (depth).
 
 
 
 
 
76
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure 5
5
: 
Feldspar 
ove
r
growths
: 
BSE
-
SEM image of 
K
-
feldspar overgrowths forming on 
detrital grains. From W
-
962 well at 1425 ft. (depth).
 
 
 
Figure 5
6
: 
Skeletal K
-
feldspar
: 
BSE
-
SEM images of s
keletal (partially dissolved) K
-
feldspar 
grains. From RASA 5 well at 519 ft. (depth).
 
 
77
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure 5
7
: 
Preferential dissolution of K
-
feldspar 
overgrowths
: 
BSE
-
SEM images of 
K
-
feldspar 
grains 
with diagenetic overgrowths that seem to be preferentially dissolving. From W
-
962 well at 
1430 ft. (depth).
 
 
 
78
 
 
through open pore space, or a structural/compositional feature of these overgrowths that resulted 
in increased susceptibility to dissolution
 
relative to the intact detrital feldspar cores
. 
 
Additional Selected Observations
 
Diagenetically formed carbonate minerals were present in all four of the sample wells. 
Pore filling calcite cement and euhedral authigenic carbonate rhombs (
F
ig. 5
8
) were ob
served. 
These rhombs appear to be nucleating on detrital grains (Fig. 5
9
), although some of these rhombs 
appear in open pore space (Fig. 
60
), likely because the thin section did not intersect the grain 
contact. 
These rhombs were determined
 
to be 
ankerite b
ased on
 
their iron and magnesium 
content
. Figure 5
8
 
shows rhombs displaying clear zoning as indicated by the difference in the 
brightness contrast in BSE imaging and supported by EDS analysis. These zoned regions contain 
varying Fe/Mg ratios and are likely
 
caused by flu
ct
uations in the pore fluid composition as the 
rhombs were gradually precipitating. 
 
T
itanium oxide minerals with skeletal and replacement textures (
F
ig. 
6
1
) 
occured locally
 
throughout samples. The presence of these minerals is 
interpreted
 
to be associated with 
remobilization of titanium from detrital biotite and rutile in sandstone systems (Morad and 
Aldahan, 1986; 1987). 
 
X
-
ray Diffraction
 
Twelve samples of clay sized (<2 
µ
m equivalent spherical diameter) material from the 
Marshall Sandstone and Michigan Formation were analzed by pXRD. Samples for XRD analysis 
were selected using the illite morphological classification scheme (Table 5) to identify samples 
containing only fib
rous authigenic illite. Careful consideration was given to exclude samples 
containing detrital illite pseudomatrix material. Previous workers (Zacharius, 1992; Cox, 2015) 
have conducted XRD analysis of Marshall and Michigan clay sized fractions to idenify 
clay
 
79
 
 
Figure 58: 
Authigenic carbonate morphology
: SE
-
SEM images of authigenic carbonate rhombs. 
EDS analysis suggests that rhombs are ankerite. From CH101 well at 114 ft. (depth). 
 
 
 
 
Figure 59: 
Zoned ankerite rhombs
:
 
BSE
-
SEM images of 
authigenic carbonate rhombs nucleating 
on quartz grains. Rhombs are likely ankerite based on EDS analysis. Ankerite rhombs display 
zoning based on heterogeneity in Fe/Mg ratios. From CH101 well at 114 ft. (depth).
 
 
 
 
 
 
80
 
 
 
 
 
 
 
 
 
 
 
 
Figure 60: 
Carbonate 
rhomb
s
: BSE
-
SEM image of a
uthigenic
 
carbonate rhombs in open pore 
space. From W
-
1002 well at 1356 ft. (depth).
 
 
 
Figure 61: 
Ti
-
O mineralization
: BSE
-
SEM images. A: Ti
-
O mineral grain displaying a skeletal 
texture. From W
-
1002 well at 1406 ft. (depth) B: TiO 
mineral occurring as part of a 
metasedimentary lithic fragment. From RASA 5 well at 519 ft. (depth).
 
 
 
 
81
 
 
 
minerals, although none of the aforementioned work gave consideration to the seperation of 
detrital and diagenetic illite. As a result, previous XRD pa
tterns generated from these clays likely 
include components of detrital clays. Such patterns are successful at identifying  illite, but do not 
accurately represent XRD properties of the authigenic illite alone. 
 
One representative diffractogram from 
each
 
i
llite
-
bearing well 
is
 
displayed
 
(Fig. 
6
2
)
.
 
All 
other diffractograms are included in the Appendix.
 
Illite was found in all twelve of the XRD 
samples analyzed
 
(Table 
7
)
. Kaolinite was found in all twelve samples
 
(Table 
7
)
. C
hlorite was 
 
found in 
seven 
samples
 
(Table 
7
)
. Quartz was found in every sample
 
(Table 
7)
. Potassium 
feldspa
r was found 
(typically with low intensity peaks) in 
nine
 
samples
 
(Table 
7
)
.
 
Minerals were 
identified using the presence of diagnostic peaks (Table 8).
 
Sample preparation 
by
 
sedimentation 
onto the mounting medium biased the distribution of minerals at the surface of the mount in 
favor of smaller clay particles
 
(Moore and Reynolds, 1997)
. As such, peak intensities for illite 
cements are heightened, but qu
antifications of mineral abundances would not be accurate and 
therefore were not attempted. 
 
 
No discrete smectite component was found in any XRD sample analyzed. However, the 
smectitic component of highly illitic IS mixed
-
layered minerals may be determine
d based on 
differences betwen illite peaks from air
-
dried and ethylene glycol saturated samples, due to the 

determine the smectitic component of diagenetic illite sa
mples in this study using the methods of 

applying these techniques to these samples were as follows: the illite (004) peak used for one 
method was not within the ran
ge of the XRD runs in this study; the intermediate angle illite
 
(
002
)
 
and 
(
003
)
 
peaks 
used
 
with these methods were subject to quartz peak overlap (
003
) and not 
 
82
 
 
Figure 62: 
Representative 
diffractograms from clay sized fraction
: Representative diffractograms 
from both wells found to contain diagenetic illite. A: From W
-
1002 well at 1363 ft. (depth) 
scanned from 4
-
40° 2

 
B: From W
-
962 well at 1430 ft. (depth) scanned from 4
-
35° 2

. 
Mineral 
a
bbreviations are those of Whitney and Evans (2010).
 
 
 
A
 
B
 
Qtz/Sa
 
Qtz/Sa
 
100/101
 
100/101
 
83
 
 
Table 7: Minerals observed in powder XRD analysis.
 
 
 
 
 
 
 
 
Table 8: Diagnostic peaks for mineral identification in oriented mounts of clay
-
size fractions by 
pXRD.
 
 
 
84
 
 
of sufficent intensity to be 
useful (
002
)
. Weakness of the (002) peak intensity is unusual and may 
be a product of Fe substituation in octahedral sites attenuating the Cu signal. 
This test was 
initia
l
ly developed for use in analyzing bentonites and shale samples, and perhaps does not 
apply 
easily
 

quantity of clay content available was much greater considering the sample types used.
 
The illite Ir 
(intensity ratio) 
number analysis of 

rodon (1984) was compl
eted for every 
sample. This analysis approximates the degree of smectite mixed
-
layering by comparing 
intensity ratios of illite peaks from air
-
dried and ethylene glycolated samples using the equation:
 























 
 
where Ir = 1 for pure illite. Larger Ir numbers 
(
>
 
1) 
indicate that a swelling component
 
(e.g.
,
 
a 
smectite
-
group mineral)
 
is present in the sample. This technique is not intended to provide 
definitive percentages, but rather to be used as an indicator of the presence/absence of 
expandable material
 
and thus an additional consideration to the interpretation of illite 
crystalli
nity data. The calculated Ir numbers are shown in Table 9. Peak measurements used in 
 
the Ir calculation are shown in the appendix. Ir results should be interpreted with caution,
 
k
eeping in mind that the theoretical minimum Ir number possible should be one
, and yet four 
samples resulted in calculated Ir values below one. This may be due in part to error associated 
with determining the exact peak maximum intensity, or background overlap caused by the 
mounting medium background. The average Ir number from XRD
 
samples was calculated to be 
1.062 (Table 9). This value is statistically indistinguishable from 
unity, 
based on the standard 
deviation
 
of 0.171
, indicating a 
negligible
 
component of expandable material.
 
 
85
 
 
Table 9: Calculated Ir numbers and full width at half maximum (FWHM) values of illite 
(
001
)
 
peaks for samples containing authigenic illite. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure 63: 
Michigan/Marshall illite crystallinity data
:
 
FWHM values of authigenic illite 
samples 
plotted on the Kübler Index formational regime diagram. The average FWHM value is 
represented by the red line, and the range of FWHM values are represented by the yellow bar.
 
 
 
86
 
 
Illite crystallinity was investigated using the Kübler index, which 
is 
based on measuring
 
the full width at half maximum (FWHM) 
of
 
the illite 
(
001
)
 
peak. Illite 
(
001
)
 
peaks were fitted 
using 
G
aussian
 
models
 
in the software package PD Analysis, provided by Proto Manufacturing. 
FWHM measurements from all samples are displayed in 
T
able 
9
. The average value of FWHM 
from XRD samples was 0.
20
8
° 
2

 
(Table 
9
).
 
This average value, along with the upper and lower 
bound
s of the range of FWHM values, were plotted on a Kübler Index temperature regime chart 
(
F
ig. 
6
3
) with the boundaries defining the anchizone 
adjusted
 
following th
e recommendations
 
of 
Kisch (1990). FWHM values 
straddle the
 
anchizone
-
epizone boundary
 
on the diagram, 
indicating high crystallinity and suggesting high temperature of formation
 
(~265° C
 
-
 
315° C)
 
for 
authigenic illite. 
 
Transmission Electron Microscopy
 
 
Transmission electron microscopic analysis was completed on two samples, one from 
each 
of the two wells containing fibrous authigenic illite (W
-
1002 and W
-
962). The sample from 
well W
-
1002 was from a depth of 1406 feet, and the sample from well W
-
962 was from a depth 
of 1425 feet. Samples were selected using the illite morphology classificat
ion (Table 5) to 
eliminate all samples containing detrital illite. Next, samples were chosen based on SEM 
observations in an attempt to select samples that contain large amounts of fibrous authigenic 
illite. Transmission electron microscopic bright field i
magery was acquired from both samples. 
Diagenetic illite was observed as groups of laths which were commonly clumped together (Fig. 
64). These laths ranged from 30
-
90 nm in width and 200
-
1000 nm in length (Fig. 65). The 
arrangement of these laths in clumps
 
may be due to the high layer charge of illite, such that 
individual laths became attracted to one another while in suspension and remained clumped on 
the TEM grid. Diagenetic kaolinite was also observed. Kaolinite occurs as pseudohexagonal
 
87
 
 
Figure 6
4
: 
Ill
ite TEM crystal habit
:
 
TEM micrographs taken with a Gatan digital camera of 
clumps of authigenic illite crystallites displaying a lath
-
shaped crystal habit. The large darker 
shapes on the micrographs are the lacy carbon grid. From W
-
1002 well at 1406 
ft. (depth).
 
 
 
 
 
Figure 6
5
: 
Illite lath 
sizes
: 
Authigenic
 
illite laths measured using the measuring tool built into the 
TEM software. The narrow lath crystal habit suggests formation conditions that limited crystal 
growth. From W
-
1002 well at 1406 ft. (depth).
 
 
 
88
 
 
plates, significantly larger than illite laths (Fi
g. 6
6
). Larger muscovite grains were observed in 
the W
-
1002 sample. The determination of mineralogy was based on high resolution EDS 
analysis of individual grains throughout both samples (Appendix). 
 
EDS analysis was completed for four different clusters o
f illite laths to determine the 
chemical composition of fibrous illite cements. Evidence of beam damage was present on some 
illite clusters after prolonged analysis. This evidence included blurring of previously sharply 
defined cleavage planes and crystal 
edges. EDS analysis of illite laths revealed an average 
mineral formula of:
 
K
0.
645
(Al
1.
701
Fe
(II)
0.
120
Mg
0.
255
)(Si
3.
525
Al
0.
475
)O
10
(OH)
2
  
 
This mineral formula is slightly different than that of the SEM illite average (Table 6). This is 
perhaps because SEM EDS analysis included components of detrital grains, or perhaps due to 
differences between the specific SEM and TEM instruments. The illit
e laths appear to contain 
slightly less potassium than the average SEM illite compositions. 
Once again the tetrahedral 
Al:Si ratio and degree of divalent octahedral substitution are similar to phengite.
 
 
Selected area electron diffraction was used to obser
ve electron diffraction patterns for 
illite laths. Electron diffraction patterns were taken from larger illite laths in an attempt to 
 
produce the sharpest patterns as possible. Electron diffraction patterns from two grains are 
presented in Figures 67 and 
68. Clear discrete diffraction spots on the SAED pattern from both 
 
samples indicate well
-
defined single crystal patterns, instead of a powder
-
like pattern that would 
be expected from random orientation about the c* axis (Freed and Peacor, 1992). This sugg
ests 
that some degree of ordered stacking must be present in these illite crystals. 
 
 
 
 
89
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure 66: 
Kaolinite TEM crystal habit
: TEM micrograph showing pseudohexagonal plate 
crystal habit of kaolinite. The darker features are the lacy carbon grid. From W
-
1002 well at 
1406 ft. (depth). 
 
 
 
 
90
 
 
Figure 
6
7
: 
Selected area electron diffraction pattern #1
:
 
 
Clump of illite laths and the resulting 
selected area electron diffraction pattern. The combination of discrete spots and faint rings 
indicate some disorder in an otherwise ordered stacking pattern. From W
-
1002 well at 1406 ft. 
(depth).
 
 
 
 
 
 
 
 
d = 0.3772 nm
 
91
 
 
Figure 
6
8
: 
Selected area electron diffraction pattern #2
:
 
Larger illite lath and the resulting 
selected area electron diffraction pattern. Discrete spots resemble a single
-
crystal pattern rather 
than a powder
-
like pattern. From W
-
962 well at 1425 ft. (depth). 
 
 
 
 
 
 
 
 
 
 
d = 0.24
25 nm
 
d = 0.
4167
 
nm
 
92
 
 
Aqueous Geochemistry
 
 
Pore fluid compositions from all 90 Marshall Aquifer groundwater wells were compiled 
from USGS 
Open
-
File Report 90
-
368 (Dannemiller and Baltusis Jr., 1989). Data for relevant 
aqueous species were extracted and are shown in the 
Appendix. Values were plotted
 
on a log 
scale 

+
/H
+
)
 
vs
 

 
diagram representing thermodynamic mineral stability fields (Fig. 6
9
). 
Thermodynamic data used in the creation of this activity
-
activity diagram come from USGS 
Survey Bulletin 1452 (Robie et al., 1979). These same data wer
e then plotted on the stability 
diagram which now includes the approximate position of an experimentally determined illite 
thermodynamic stability field (Fig. 
70
). The illite stability field on Figure 
70
 
represents the full 
range of positions on the diagra
m where previous researchers have placed the boundaries of the 
illite stability field as determined experimentally (Aja, 1991; Aja et al., 1997; Yates and 
Rosenberg, 1997; Noguera et al., 2011). The majority of the data plot within the overlapping 
 
kaolini
te
-
illite stability fields, while some data plot in the kaolinite stability field just below the 
illite stability field. 
 
USGS aqueous geochemical software package PHREEQC (version 3) was used for
 
mixing and forward modeling. The composition of a Marshall 
aquifer well from Lake County, 
 
Michigan (Well #3 from Lake County in USGS Dannemiller and Baltusis, 1988), was selected as 
the initial input to represent the modern pore fluid composition of the Michigan Basin in 
 
intervals found to contain either fibrous illite cements or authigenic replacement illite. This well 
was selected because it most closely approximates the depth and location within the basin of 
the 
sample wells containing aforementioned illite (Fig. 7
1
). T
he representative well from Lake 
County was the deepest Marshall well near the center of the basin that was sampled as part of 
Dannemiller and Baltusis (1988). 
Using PHREEQC
,
 
the c
omposition
 
of 
groundwater 
in this well 
 
93
 
 
 
 
 
 
 
 
 
 
 
 
Figure 69: 
Marshall groundwater data
: Groundwater chemistry from the Marshall aquifer 
plotted on the K
2
O
-
Al
2
O
3
-
SiO
2
-
H
2
O system mineral stability diagram. Thermodynamic data 
used in production of boundaries from Robie et al. (1979). The least
-
squares regression line is 
displayed in blue. 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure 70: 
Marshall groundwater data with illite stability field
: Marshal
l aquifer groundwater 
chemistry plotted on illite inclusive mineral stability diagram. The illite stability field is 
represented by the blue square. This field represents all previous experimentaly determined 
positions of illite stability.
 
 
y = 
-
2.0578x 

 
4.3459
 
R
2
 
= 0.283
5
 
94
 
 
Figure 71: 
Geo
graphic locations of relevant wells
: Geographic locations of wells relevant to 
completion of aqueous geochemical modeling.
 
 
 
95
 
 
was 
brought to 
undersaturation with respect to 
muscovite
 
to represent 
(as well as possible) 
the 
composition of pore fluid 
prior to illitization
 
(Appendix)
. 
Muscovite is frequently used as a 
proxy for illite in thermodynamic calculations given the poorly defined thermodynamic 
properties of illite (Bjorkum and Gjelsvik, 1988). The 
compositional adjustment 
representing 
undersatu
ration with respect to muscovite
 
is made by fixing the saturation index of illite
 
(K
-
mica) at 
-
1 (undersaturation)
 
and a
djusting
 
the chemical composition
 
of 
relevant aqueous species 
in the
 
solution
 
to fit the fixed value. 
The s
aturation index is a value us
ed to determine whether a 
fluid is saturated, undersaturated, or supersaturated with respect to a mineral phase. This value is 
the
 
log of the
 
ratio of the ion activity product (IAP) to K
sp
 
for the 
illite (K
-
mica) dissociation 
reaction
 
displayed below:
 




































 
 
 
(5)
 
 
using the equation
s
:
 









 
where 
log
10
(
K
sp
) = 12.703 (at 25° C) (PHREEQC internal database). The K
sp
 
value 
comes from 
the equation: 
 




























 
and 
is provided in the PHREEQC internal database, and the IAP is calculated by PHREEQC 
from the equation: 
 


























 
PHREEQC will calculate the activities o
f K
+
, Al
3+
, H
4
SiO
4, 
and H
+
 
required to make the SI equal 
the user
-
defined value, and then provide an output containing the composition of this new 
solution
. The resulting pre
-
illite pore 
fluid 
should suffice to represent the
 
slight
 
chemical 
96
 
 
difference between modern pore fluid and 
pore fluid prior to fluid invasion and 
fibrous illite 
cement
 
precipitation
 
(Appendix)
. 
The precipitation of fibrous authigenic illite would have altered 
the composition of the pore fluid, and this modeled solution was created to account for the slight 
vari
ation in composition. This modeled pre
-
illite composition was then plotted on the K
2
O
-
Al
2
O
3
-
SiO
2
-
H
2
O mineral stability diagram (Fig. 
7
2
). This pre
-
illite composition was used as an 
input for aqueous mixing modeling. 
 
 
The hypothesis that illite formed by d
ownward percolation of meteoric water (Zacharias, 
1992) was tested first. The composition of meteoric water used for mixing modeling was 
selected from USGS Water
-
Supply Paper 1535 (Carroll, 1962). Using PHREEQC this solution 
was brought to thermodynamic eq
uilibrium with potassium feldspar at 25°C to represent 
enrichment of K
+
 
resulting from the interaction of potassium feldspar with meteoric water 
(Appendix). An incremental mixing model was then initiated at 25°C using modeled pre
-
illite 
pore fluid and K
+
-
enriched meteoric water as end member solutions. Potassium
-
enriched 
meteoric water was introduced at increments of 2%, 4%, 6%, 8%, 10%, 20%, 30%, 40%, 50%, 
60%, 70%, 80%, 90%, and 100%. Compositions of each of the resulting mixed solutions were 
recorded a
nd plotted along with the modern Marshall groundwater dataset (Fig. 73). Increasing 
the amount of meteoric solution being mixed with the pre
-
illite solution results in dilution of 
 
Marshall brine with respect to the K
+
/H
+
 
ratio and Si activity, and subsequ
ent movement of data 
points away from the illite stability field along both the X and Y axes and into the kaolinite 
stability field. 
 
Two deep basin brine mixing models were created to investigate the hypothesis of 
precipitating illite by mixing with diffe
rent compositions of deep formation waters. To test the 
hypothesis of upwardly mobile deep brine mixing, a Michigan Basin brine was required as an
 
97
 
 
Figure 7
2
: 
Modeled pre
-
illite
 
solution
: 
Composition
 

-

on the illite
-
inclusive mineral stability diagram. 
 
 
 
25
°
 
C
 
98
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure 7
3
: 
Meteoric water mixing model results
: A: Results from K
+
-
enriched meteoric water 
mixing 
model. B: Close
-
up on the data of interest. The arrow indicates the trend of the modeled 
pore fluid composition with increasingly larger percentages of K
+
-
enriched meteoric water being 
introduced. The blue square represents the illite stability field.
 
 
 
A
 
B
 
25
°
 
C
 
25
°
 
C
 
99
 
 
in
put. The solution used for the first deep brine input was Devonian age well ID 4069 from the 
Richfield member (part of the Lucas Formation) in the Michigan Basin, sampled and reported by 
 
Wilson and Long (1993). This well was selected because the groundwat
er solution is saline, the 
composition has been described in detail, and geographically this well is very close to the sample 
wells that contain fibrous authigenic illite. Assuming that the flow path of these brines is 
dominantly vertical
, 
groundwater from
 
this Devonian well would be the most reasonable solution 
to use as a brine end member for mixing. The location of all wells used for modeling, as well as 
the Marshall sample wells containing fibrous authigenic illite are displayed (Fig. 
7
1
). 
Additional 
information regarding these wells is presented in the Appendix. 
 
An incremental mixing model was then implemented at 25°C using modeled pre
-
illite 
Marshall pore fluid and Devonian well 4069 as end member solutions. Devonian brine was 
introduc
ed at increments of 2%, 4%, 6%, 8%, 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, 
90%, and 100%. Compositions of each of the resulting mixed solutions were recorded and 
plotted along with the modern Marshall groundwater dataset (Fig. 
7
4
). Increasing the amount o
f 
 
Devonian brine being mixed with pre
-
illite solution results in movement of the data away from 
the illite stability field and into the kaolinite stability field. This trend is similar to that of 
meteoric water mixing, although the slope of the trend is m
uch closer to zero than the meteoric 
model (Fig. 
7
3
). 
 
 
The Devonian brine mixing model failed to provide a feasible illite
-
producing scenario, 
so another mixing model was defined. The composition of the second deep brine solution was 
calculated by averagi
ng the compositions of two Silurian
-
age wells from the Niagara Group in 
the Michigan Basin as reported by Wilson and Long (1993). Preference in selection of these 
wells was given to wells located as near to the middle of the basin as possible for which val
ues
 
100
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure 7
4
: 
Devonian brine mixing model results
: A: Results from the Devonian brine mixing 
model. B: Close
-
up on the data of interest. The arrow indicates the trend of the modeled pore 
fluid 
composition with increasingly larger percentages of Devonian brine being introduced. The 
blue square represents the illite stability field.
 
A
 
B
 
25
°
 
C
 
25
°
 
C
 
101
 
 
of pH, K
+
, and H
4
SiO
4
 
were reported. This eliminated many options. The Silurian
-
age Niagara 
Group was selected due to the known existence of highy concentrated brines that have been 
described in detail by Wilson and Long (1993). Wells 2090 and 2091 were selected because they 
m
ost closely met all desired criteria (close to center of basin; pH, K
+
, H
4
SiO
4
 
reported). The 
locations of these wells in relation to the two illite
-
bearing sample wells are displayed in Figure 
7
1
. 
 
 
An incremental mixing model was then implemented at 25°C
 
using a composition 
calculated by using the average of the two Silurian brines and the pre
-
illite Marshall brine as end 
member solutions. The Silurian brine solution was introduced at increments of 2%, 4%, 6%, 8%, 
10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, 9
0%, and 100%. Compositions of each of the 
resulting mixed solutions were recorded and plotted along with the modern Marshall 
groundwater dataset (Fig. 
7
5
). Increasing the amount of Silurian brine being mixed with the pre
-
illite solution increases the K
+
/H
+
 
ratio, resulting in movement of the data up the Y
-
axis within 
the illite stability field. This result is in contrast to previous mixing models.  
 
 
102
 
 
Figure 
7
5
: 
Silurian brine mixing model 
results
: 
A
: Results from the Silurian brine mixing model. 
B: Close
-
up on the data of interest. The arrow indicates the trend of the modeled pore fluid 
composition with increasingly larger percentages of Silurian brine being introduced. The blue 
square represents the
 
illite stability field. 
 
 
A
 
B
 
25
°
 
C
 
25
°
 
C
 
103
 
 
Discussion
 
Well Log Interpretations
 
 
Well log profiles which include gamma ray
 
(GR)
, caliper, neutron porosity, and bulk 
density data were provided by MGRRE for 
the Consumers Energy Winterfield 
W
-
1002 
PN 
61260 well 
and
 
the
 
Michig
an Gas Storage Co. 25462 
-
W
-
962 
observation well 
(the illite
-
bearing 
wells). The
 
positions
 
of the Marshall and Michigan Stray sandstones are easily 
apparent
. The 
Stray Dolomite provides a
 
recognizable 
low
-
GR, 
high
-
density 
signature 
that
 
can be used to 
dete
rmine the contact between the two
 
units
 
(Fig. 76)
. 
The Marshall Sandstone and the Michigan 
Stray Sandstone have near identical signatures for all four of the included logs. Slight variation 
in density, neuron porosity, and gamma ray values exist 
within these sandstones
. These 
variations
 
occur intermittently 
and likely represent unsampled thin intervals of interbedded 
mudstone.
 
 
Paragenetic Sequence
 
 
Thin section and rock chip petrography (optical and SEM) comprise the basis for the 
paragenetic sequence (
F
ig. 
7
)
. Textural relationships between diagenetic minerals allow for the 
establishment of different stages of mineral dissolution and precipitation. The paragenetic 
sequence as presented here agrees relatively well with previous rese
archers (Zacharias
, 1992
). 
Minor modifications to previous work were made based on contrasting observations. 
 
 
Compaction
 
The 
deformation (
bending
)
 
of 
detrital mica grains demonstrates early mechanical 
compaction of rigid grains around the mica
 
(Fig. 
7
, 
4
4
)
. 
C
ompaction 
by 
pressure dissolution 
is 
observed 
in
 
quartz and feldspar grains 
(F
ig. 
5
2
) from all sample wells exc
ept
 
CH101, the 
 
104
 
 
Figure 76: 
Representative well log
: Well log profile for the Marshall/Michigan interval of the 
Winterfield W
-
1002 well. The large bulk density and low neutron porosity signatures of the 
Stray Dolomite indicate the contact between the Marshall Sandstone and the Michigan 
Formation. This log was provided by MGRRE.
 
 
105
 
 
shallowest well. Quartz grains displa
ying pressure dissolution often show grain
-
c
o
ating cements 
nucleating along the entire exterior of these grains exc
ept
 
points of contact with other grains. 
This implies that those points of contact existed prior to the formation of any type of grai
n
-
coatin
g or pore
-
filling cement. 
 
Mudstone lithic fragments and detrital mica grains in the Marshall and Michigan Stray 
Sandstones have been subjected to compressive forces resulting from overburden. In response to 
stress these grains deform
ed
 
plastically around 
rigid detrital grains like quartz and feldspar
 
(Fig. 
2
2
)
. Intervals containing large abundances of either mudstone lithic fragments or mica grains 
contain very little
 
(0.5%
 
-
 
7.5%) 
to no porosity (Fig 
1
6
, 
1
7
, 2
3
, 4
3
; Table 5
). Physical 
deformation of detri
tal phyllosilicates constitutes the largest porosity
-
reducing factor in these 
strata.
 
The mica grains and mudstone fragments deform in different ways. In response to 
compressive stresses mudstone lithic fragments appear to lose any internal structure and r
igidity, 
resulting in the formation of pseudomatrix. This pseudomatrix 
has
 
an EDS signature identical to 
illite (Fig. 
2
1
), likely because illite is the dominant component of these detrital mudstone 
fragments. As a result, localities containing this 
pseudomatrix are
 
litharenites, but also
 
pseudo
greywackes 
(
24
-
36
.6
% 
pseudo
matrix)
. Micaceous intervals in the Marshall are very 
micaceous (
up to 61.5%
). Mica grains in these intervals show less exfoliation and alteration than 
mica grains in intervals with
 
n
ormal low detrital mica abundances and
 
significant porosity. It is 
likely that reduction of porosity by compaction and grain deformation inhibit
ed
 
alteration of 
these mica grains by restricting the amount of pore fluid 
and reactant solutes 
coming into cont
act
 
with them
. 
 
 
106
 
 
 
Carbonate
 
Carbonate cements were observed primarily as diagenetic rhombs in these samples. Many 
of these rhombs were coated with chlorite cements (
F
ig. 
3
4
)
;
 
therefore
,
 
the formation of these 
rhombs had to have occurred prior to the precipitation of chlorite cements
 
(Fig 
7
)
. The later 
interval
 
of carbonate formation on the paragenetic sequence was proposed by previous 
researchers. No evidence was found 
in this study 
to 
r
efute
 
this position, so no changes were 
made
 
to the previously established episode in the paragenetic sequence
. 
The presence of Fe
-
bearing carbonates indicates
 
reducing
 
conditions
 
at the time of 
ankerite
 
formation.
 
Feldspar
 
Feldspar overgrowths are commonl
y observed preferentially dissolving (
F
ig. 
5
7
)
.
 
F
eldspar overgrowth precipitation must have occurred prior to feldspar dissolution. 
Selective 
destruction of feldspar overgrowths relative to cores is unusual
 
(Land and Milliken, 1981)
; this 
deserves further 
consideration. 
Kaolinite cements were observed precipitating adjacent to 
partially dissolved feldspar overgrowths (
F
ig. 
2
9
). This indicates that these overgrowths existed 
prior to the formation of kaolinite cements.
 
Chlorite
 
Chlorite cement 
precipitation occurred near the middle of the paragenetic sequence
 
(Fig. 
7
)
. This is based on the observation that quartz overgrowths (
F
ig. 
3
5
)
 
and 
pore
-
filling kaolinite 
cements (
F
ig. 
3
6
) 
are
 
supe
r
posed upon
 
grain coating chlorite cements. 
Zacharias (1992
) reported 
instances of authigenic illite superposed on authigenic chlorite
 
(in unspecified sample wells)
, 
although this was not observed in this study.
 
The Fe
2+
 
state of iron within the chlorite mineral 
107
 
 
formula indicates that redox conditions were reducin
g, and the environment was anoxic at the 
time of chlorite formation. 
 
Quartz Overgrowths
 
It is likely that multiple generations of quartz overgrowth precipitation occurred. Quartz 
overgrowths engulfing chlorite cements 
(Fig. 
3
5
) 
indicate that precipitation of these overgrowths 
occurred after chlorite
 
(Fig. 
7
)
. Pore filling kaolinite cements an
d pyrite cements are observe
d 
superposed upon
 
quartz overgrowths, indicating that 
an
 
earlier generation of quartz overgrowths 
occurred between chlorite formation and pyrite formation
 
(Fig. 
7
)
. A second generation of quartz 
overgrowths 
is
 
inferred based on 
multiple instances of prismatic overgrowths precipitating on 
illite
-
coated 
grains, but not coated by illite (Fig. 
2
6
). 
This suggests that a later generation may 
have formed cogenetically with illite cements
 
(Fig. 
7
)
. 
 
Pyrite
 
Based on the 
morphological vari
ation in
 
diagenetic pyrite
 
(Fig. 
4
9
, 
50
, 
5
1
)
 
it is likely that 
multiple episodes of pyrite cement formation occurred in these rocks. Framboidal pyrite 
occurs
 
in pore space directly outside of chlorite cement rims (
F
ig. 
4
9
-
B
) indicating that at least one 
generation of pyrite must have formed after chlorite cement precipitation. 
Raiswell and Berner 
(1985) suggest that pore
-
filling pyrite is Fe limited, and that significant supersaturation would be 
required for the formation of
 
this texture. Carstens (1986) proposes that framboidal pyrite 
formation is a result of direct precipitation of pyrite crystals without the existence of a precursor 
sulfide and
 
may be inhibited by the competitive growth of adjacent crystals. 
Regardless of 
texture, the o
ccurrence of authigenic pyrite indicates that redox conditions during stages of pyrite 
formation were reducing, and
 
the environment was euxinic
 
(Berner, 1980).
 
108
 
 
Kaolinite
 
Pore
-
filling kaolinite cements precipitated near the end of the 
paragenetic sequence
 
(Fig. 
7
)
. Kaolinite is observed filling pore space between chlorite
-
coated grains
 
(Fig. 3
6
)
. The 
kaolinite booklets themselves are not coated by the chlorite, indicating that kaolinite formed after 
the chlorite cements were in place. S
ome kaolinite booklets are observed 
transforming
 
to illite
 
(Fig. 
3
1
)
. 
Neoformed i
llite is also observed nucleating on and around kaolinite booklets (
F
ig. 
2
5
), indicating that illite formed after kaolinite cements were already in place
 
(Fig. 
7
)
. 
 
Illite
 
Precipitation of grain coating illite cements occurred very late in the paragenetic 
sequence
 
(Fig. 
7
)
. Illite is observed nucleating
 
(neoforming)
 
on kaolinite
 
(Fig. 2
5
)
 
and detrital 
quartz grains (Fig. 2
4
)
.
 
Illite is also observed as a transformatio
n product of kaolinite (
Fig. 3
1
). 
There is no evidence of any type of cement 
superposed upon
 
illite cements. Illite precipitation 
may have been cogenetic with a late generation of quartz overgrowth precipitation
 
(Fig. 
7, 
2
6
)
. 
The compositions of 
both varie
ties of 
authigenic illite 
(Table 6) 

 
illite of transformation, 
pseudomorphic after authigenic kaolinite, and neoformed pore
-
lining fibrous illite 

 
have mixed 
illitic
-
phengitic character, including specific tetrahedral substitutions consistent with illite
, 
phengite
-
like substitutions of divalent cations 

 
(Mg, Fe(II) 

 
in octahedral sites, and overall 
layer charge and interlayer site occupancy of illite. 
 
Detrital Phyllosilicates
 
 
Petrographic evidence suggests that many phyllosilicate minerals identified in these rocks 
either were not or are currently not in equilibrium with pore fluids. Alteration of mica grains is 
common. Alteration of detrital mica in porous intervals contribute
d 
to
 
a minor extent to the 
109
 
 
formation of 
secondary 
clay
s
. Alteration of detrital muscovite to
 
secondary
 
kaolinit
e
 
is strongly 
suggested by the formation of kaolinite cements between exfoliating muscovite layers which 
appear to be gradually depleted of potas
sium along grain edges (
F
ig. 
4
1
, 
4
2
). This
 
style of
 
alteration has been documented in soil profiles (Singh and Gilkes, 1991) and kaolinitic
 
altered
 
granites (Robertson and Eggleton, 1991) demonstrating th
at
 
this reaction
 
is common
 
in a variety 
of condition
s. TEM work 
of
 
Robertson and Eggleton
 
(1991)
 
shows evidence of a transformation 
reaction of muscovite to kaolinite
. This reaction
 
shown below
:
 








































 
 
(6)
 
proceeds 
by
 
removal of tetrahedral sheets and 
depletion of K
+
. This K
+
 
depletion is 
observed in 
Marshall/Michigan samples
 
(Fig. 
4
2
)
. Additionally, grains exhibiting this alteration show K
+
 
depletion predominantly along edges that extend into open pore space. It is likely that porosity 
and fluid flow
 
(
or, at a minimum, solute transport)
 
are critical in facilitating this reaction in 
muscovite grains in sandstones. This hypothesis would explain why alteration of grains in 
mica
-
rich
 
intervals remained minimal. This reaction is represented by the muscovite
-
kaolinite 
bo
undary
 
on the K
2
O
-
Al
2
O
3
-
Si
O
2
-
H
2
O
 
stability diagram.
 
T
his reaction could be
 
initiated by
 
the 
invasion of porosity by dilute waters. This process would dilute the Marshal
l
 
pore fluids
, 
expressed as a change in
 
the K
+
/H
+
 
ratio. If the composition of Marshall pore fluid were to 
initially plot within the muscovite stability field and then become dilute
d by the invasion of fresh 
water
, movement of the data points toward the mus
covite
-
kaolinite boundary would be expected 
and the alteration reaction may proceed. Evidence of this reaction in Marshall/Michigan strata 
may represent a previous period of meteoric water infiltration 
into
 
Mississippian strat
a. This 
meteoric infiltration 
could have occurred at any point in time between the formation of 
authigenic chlorite and the formation of authigenic illite in the Early Permian (Cox, 2015).
 
110
 
 
Biotite can alter to a variety of different secondary minerals depending on the conditions 
of the
 
system. Alteration of detrital biotite to secondary kaolinite was observed
 
(Fig. 
4
7
)
. 
Individual layers of biotite appear to lose internal structure while becoming gradually depleted of 
potassium, iron, and magnesium 
where
 
the grain extends into open pore
 
space (
F
ig. 
4
7
). The 
alteration of biotite to kaolinite is well documented
. 
The reaction can occur in soil profiles
 
(Rebertus et al., 1986)
, igneous rocks
 
(Dong et al., 1998)
, schists
 
(Ahn and Peacor, 1987)
, and 
sedimentary rocks
 
(Morad and Aldahan, 
1986)
. Multiple mechanisms have been proposed for 
this reaction, from a transformation reaction
 
of one biotite to two kaolinite layers
 
(Dong et al., 
1998) to a complete neoformation reaction in which kaolinite crystallizes at biotite layer 
boundaries (A
hn
 
and Peacor, 1987). 
This process would proceed by the reaction: 
 





































 

































  
 
 
(7)
 
 
A metastable intermediate stage consisting of altered biotite and halloysite may occur in systems 
wh
ere the reaction proceeds slowly due to low temperatures
 
(Dong et al., 1998)
. As with 
muscovite, this reaction proceeds 
in part with
 
depletion of K
+
. Biotite alteration to kaolinite
 
by 
reaction 7
 
also involves substitution of 2+ octahedral cations with alu
minum. 
 
Evidence suggesting possible alteration of biotite to chlorite was observed at 
present 
depths ranging from 70
-
140 feet. Biotite is known to alter to chlorite in sedimentary systems 
(Veblen and Ferry, 1983; Banfield and Eggleton, 1988). This alteration is typically associated 
with the metamorphism of sedimentary pelitic rocks to low grade met
amorphic greenschist 
facies. This reaction is believed to proceed 
by
 
transformation, which would result in a 
morphology similar to the primary biotite grain. The potentially altering grains observed here 
111
 
 
display 
unique
 
textural relationships
 
(Fig. 4
8
)
. Bla
ded chlorite crystals can be seen on top of and 
in between biotite layers
 
(Fig. 4
8
)
. It may be that chlorite is nucleating on and in between 
exfoliated biotite layers rather than biotite altering to chlorite 
by
 
transformation.
 
Detrital mica certainly contr
ibutes to the formation of diagenetic clay in these rocks. Th
e 
initial
 
stimulus for this project was
 
the potential 
for
 
such large amounts of detrital mica to 
provide an internal source of potassium for illitization. However, after thorough petrographic 
ana
lysis
,
 
no illite cements were observed in samples that contain abundant (> 25.5%) mica
 
(Table 4, 5)
. 
Intervals of 
abundant mica
 
have been compacted to the point of 
nearly eliminating 
porosity (maximum of 5.5% porosity in
 
mica
-
rich
 
samples)
 
(Table 4, 5)
. 
A 
sim
ilar lack of illite 
cement
 
is noted in 
phyllitic rock
-
fragment
-
rich 
litharenite pseudo
greywacke samples (maximum 
0.5% porosity)
 
(Table 4, 5)
. No illite cements are observed coating grains in samples 
with
 
minimal porosity. This suggests that significant porosity may be a requirement for illite 
formation. If true, illite is interpreted to be a product of solute transport and not internally 
sourced. 
 
Diagenetic 
Chlorite and Kaolinite 
 
A graphic representation of the depth relationships between sample wells in this study, as 
well as the depth relationships of major diagenetic clay minerals is provided 
(Fig. 
7
7
).
 
 
Diagenetic chlorite occurred as a
 
significant (> 5%) 
clay mineral in the tw
o shallower 
wells (RASA 5 and CH101). Chlorite in deep wells did not occur as pervasive grain coating 
cements, 
but 
rather as detrital components
 
of rock fragments
. Petrographic results from this 
study disagree with the previous 
observations
 
of Zacharias (1
992) that chlorite abundance 
increases with depth. 
Significant
 
diagenetic chlorite was found at 
modern 
depths as shallow as 
58
 
feet and
 
absent
 
in both deeper wells.
 

sampling
 
included more wells  
 
112
 
 
 
Figure 7
7
: 
Sample well and 
authigenic clay depth distribution
:
 
Graphic representation of the 
depth distribution of major authigenic clay cements, as well as the depth relationship between 
sample wells. Well locations and unit names are placed only within sampled intervals and may 
no
t represent the entirety of the unit(s) at each location. 
 
 
 
113
 
 
(19 
total), which spanned all the stratigraphic intervals of Mississippian Sandstones. Perhaps this 
discrepancy comes from analyzing different sample wells with different chlorite depth 
distribut
ions. Petrographic results from this study also disagree with the suggestion of Zacharias 

 
strongly
 
linked to the presence of detrital micas. Chlorite cements d
id occur in the RASA 5 well 
which contained abundant detrital mica
 
(Table 3, 4)
, but chlorite was 
also 
found in 
significant
 
abundance
s
 
in the CH101 well which contained very little 
(maximum of 3.5%) 
mica
 
(Table 3, 4)
.
 
 
Oxygen isotopic analysis of 
chlorite cements by Zacharias (1992) indicate that the 
original diagenetic fluid
 
from
 
which chlorite formed was similar in composition to sea water. To 
reconcile this information with the depth distribution of chlorite cements it should be noted that 
the d
eepest part of the Marshall Sandstone in the center of the basin has very little hyd
rological 
connectivity with the rest of Carboniferous strata (Ging et al., 1996). It is possible that evapo
-
concentrated marine formation waters from the Michigan 
F
ormation
 
percolated down into the 
Michigan 
S
tray and Marshall Sandstones and initiated the precipitation of chlorite cements. This 
 
model is attractive because it would explain the absence of 
pervasive
 
chlorite in deep wells in the 
center of the basin
 
where pore f
luids are more saline
. The presence of such abundant authigenic 
chlorite 
in shallow wells and the absence of 
similar
 
abundance in deeper wells
 
is difficult to fully 
explain without further study focusing specifically on this chlorite. 
 
Pore filling kaolini
te cements were pervasive throughout the entire depth range of all four 
sample wells (20
-
1450 feet). Kaolinite cements formed relatively late in the paragenetic 
sequence, but the final stage of paragenetic sequence (illite) has been dated to 294
-
275 Ma (Co
x, 
2015), meaning that the formation of diagenetic kaolinite 
likely occurred in the Late 
Carboniferous or Early Permian
. Kaolinite commonly forms as a result of dilute acidic meteoric 
114
 
 
water infiltration (Blatt, 1979; Hurst and Irwin, 1982). The Marshall Sa
ndstone and Michigan 
Formation were deposited in a shoreline environment, meaning that formation waters in these
 
depositional environments
 
were marine. Hydr
ological
 
connectivity of 
much of the
 
Marshall/Michigan with overlying Pennsylvanian strata
 
as establ
ished
 
by the USGS Michigan 
Basin RASA 
study
 
(Westjohn and Weaver, 1998) means that percolation of meteoric waters 
through overlying strata and into Mississippian strata was certainly possible. Dilute water 
infiltration into marine formation waters would re
sult in a mixing scenario that would favor the 
formation of diagenetic kaolinite at the expense of dissolving K
-
feldspar 
by
 
reaction 1. 
Petrographic evidence of this reaction 
is
 
observed in the form of kaolinite cements commonly 
occurring immediately adjac
ent to skeletal K
-
feldspar grains
 
(Fig. 2
9
)
. This relationship
 
is 
consistent with
 
mixing of meteoric and marine waters initiating the formation of kaolinite 
cements
 
during Late Carboniferous epidiagenesis
.
 
Kaolinite
-
Illite Reaction
 
 
Marshall 
a
quifer 
pore fluids plot on the K
2
O
-
Al
2
O
3
-
Si
O
2
-
H
2
O
 
mineral stability diagram 
in 
the kaolinite stability field 
and the
 
overlapping
 
region of the
 
experimentally determined illite 
stability field
. Furthermore, the data seem to plot along a noticeable trend line, the 
slope of 
which 
is similar to 
the experimentally determined slopes of the kaolinite
-
illite stability field 
boundaries determined by Aja et al. (1991), and Yates and Rosenberg (1996) (
F
ig. 
7
8
). This 
strongly suggests that the aqueous geochemistry in the Marshall 
a
quifer is influenced by a 
reaction which results in pore fluid compositions
 
in
 
equilibrium with kaolinite and illite. This 
observation conforms with the paragenetic sequence which pla
ces kaolinite precipitation and 
illite precipitation as the two most recent diagenetic events. SEM analysis revealed evidence of 
 
115
 
 
Figure 7
8
: 
Experimentally determined kaolinite
-
illite stability boundaries
:
 
Experimentally 
determined kaolinite
-
illite stabil
ity boundaries. A: Equilibration experiments at 25°C, from Aja 
et al., 1991 B: Equilibration experiments at 100°C, from Yates and Rosenberg, 1996.
 
 
 
 
116
 
 
kaolinite replacement by illite (Fig. 
31
), further supporting the hypothesis that a kaolinite
-
illite 
react
io
n
 
occurr
ed
 
in these strata. 
 
 
According to Aja et al. (1991), and Yates and Rosenberg (1996), the boundaries defining 
kaolinite
-
illite
-
solution equilibrium can be calculated for illite of composition:
 
K
x
Al
2
(Al
x
Si
4
-
x
)O
10
(OH)
2
 
from the reaction:
 


































 






























 
 
 
(8)
 
using the equilibrium constant expression:
 

























 
T
he slope of the kaolinite
-
illite stability boundary 
in this forma
t is then:
 









 
 
where x refers to the stoichiometry of illite
 
(specifically, the amount of Al for Si substitution in 
tetrahedral sites)
. The linear 
regression
 
of Marshall groundwater data on the K
2
O
-
Al
2
O
3
-
SiO
2
-
H
2
O mineral stability diagram has an equation of y = 
-
2.0578x 

 
4.3459 (Fig. 
6
9
). The slope
 
(m 
= 
-
2.0578)
 
of this trendline can then be used to calculate a
n x value of 0.493 for the Marshall 
data. 
When compared to x values derived from EDS analysis of illite in this study (x = 0.735, 
0.738, 0.645) (Table 6) it is app
arent that the linear regression value does not match observed 
values. The x value derived from illite lath TEM
-
EDS data most closely resembles the linear 
regression value. 
Th
e Marshall linear regression
 
x value is most similar to the calculated
 
value of 
117
 
 
x
 
= 
0.5 from the experimental results of Yates and Rosenberg (1996) at temperatures of 100° C 
and 150° C
, and x = 0.52 from Aja et al. (1991) at 150° C 
. The data of Yates and Rosenberg 
show an increase in the x value with higher temperatures, with a maximum x value of 0.88 at 
250° C. This contrasts with the results of this study. Illite crystallinity data from 
 
Marshall/Michigan samples suggest illite fo
rmation at temperatures between 250
-
300° C (Fig. 
6
3
), and yet the x value calculated for Marshall groundwater was only 0.493. 
 
 
Marshall groundwater data from Figure 69 are not tightly clustered (R
2
 
= 0.2835). 
Groundwater compositions were replotted 
reveal
ing a more apparent trend to the data defining 
the upper limit of the overall data distribution, 
and a linear regression line was calculated (Fig. 
7
9
). These data are more tightly clustered (R
2
 
= 0.58)
. The
 
equation for this 
linear regression line 
is
 
y = 
-
0.847x + 0.713 (Fig. 7
9
). The slope (m = 
-
0.847) of this trendline can be used to calculate 
an x value of 0.702. This x value is much closer to illite 
EDS values from this study (Table 6). 
Marshall groundwater data from USGS report (Dannemiller and 
Baltusi
s, 1986) included sample 
wells throughout the entire lower peninsula, not just the portion of the Marshall that may contain 
illite. 
Based on the similarity between the new calculated x value and illite EDS data, 
g
roundwater data within the compositional range 
that was 
used to calculate the linear regression 
line (Fig. 7
9
) may represent the portion of the basin where illite formation has occurred.
 
The 
linear regression line from Figure 78 would then represent the po
sition of the kaolinite
-
illite 
stability boundary for the Marshall system
, and the groundwaters plotted along the regression 
line would indicate three
-
phase (solution
-
kaolinite
-
illite) solution
-
mineral equilibrium
.
 
The 
position of a stability field boundar
y (the y
-
intercept of the line) is deter
mined by all the factors 
that influence the equilibrium constant for the reaction as written (e.g., standard Gibbs free 
energies of formation of all phases and species; solid
-
solution effects (e.g., Merino, 1975); P;
 
T). 
 
 
118
 
 
Figure 79: 
Marshall groundwater trend
: Linear regression trendline of a portion of the Marshall 
groundwater data. The blue square indicates the portion of the data used to calculate the 
trendline. 
 
 
119
 
 
However, t
he slope of a stability field boundary 
depends only on the stoichiometry of the 
reaction describing the equilibrium.
 
The (1) slope of the regression line for the highest 

K
+
/

H
+
 
Marshall Sandstone groundwaters in Figure 79 and the corresponding x value in equation (8) 
above 
for the K
-
Al
-
Si proportions of a simplified illite (specifically, the stoichiometric subscript 
for tetrahedral Al
-
for
-
Si substitution and consequent layer cha
rge and charge
-
balanced interlayer
 
K occupancy) 
inferred
 
from the slope of the regression line
 
(
Fig. 79
)
 
and (2) 
measured
 
stoichiometric proportions of K, Al, and Si in illite of both neoformation and (especially) 
transformation origin in this study 
confor
m closely. This conformity of inferred and measured 
stoichiometric values 
suggest
s
 
that a
 
three
-
phase (
illite
-
kaolinite
-
solution
 
equilibrium
)
 
for 
(phengitic) 
illite of the measured composition in the Marshall 
(Napoleon) 
and 
lower Michigan 
Stray 
Sandstone
s
 
existed. This equilibrium does not necessarily exist at present
-
day T, but it 
existed when the kaolinite transformed to illite (at temperatures of 200 

 
300 
°
C, according to the 

try has not changed upon 
cooling to present
-
day T at 400 

 
450 m depth.
 
There are m
ultiple issue
s with using this analysis
. Calculation of x using this method is 
only accurate in the case of no 
divalent cation substitution in octahedral sheet
s
. This method does 
not 
consider
 
the presence of Fe
2+
 
or Mg
2+
 
in the illite mineral structure. This is problematic 
given
 
that all EDS an
alyses in this study indicate the presence of divalent cation substitution in 
(phengitic) 
illite cements (Table 6). Additionally, x value
s
 
may 
be caused by the influence of 
other
 
factors 
on
 
Marshall groundwater chemistry. 
It is also possible that the exper
imental data of 
Aja et al. (1991) and Yates and Rosenb
e
rg (1996) do not transfer exactly to kaolinite
-
illite 
thermodynamic equilibrium in natural systems. 
 
120
 
 
Two reaction
s
 
have been proposed to describe kaolinite
-
illite
-
solution equilibrium 
in 
nature (reaction 2 and reaction 3).
 
Reaction 2 assumes a closed system and relies on an internal 
source of potassium 
by
 
dissolution of K
-
feldspar. K
-
feldspar is certainly present, but
 
petrographic evidence suggests that K
-
feldspar dissolution cease
d prior to illite precipitation
, and 
the K
-
feldspar
-
kaolinite mineral assemblage should remain stable 
to temperatures greater than 
those
 
achiev
able
 
by simple burial in th
e
s
e
 
strata 
(Lanson, 2
002). Reaction 3 is more feasible at 
low temperatures
, provided an additional potassium input 
exists
.
 
This reaction
 
requires a 
significant increase in 
pore fluid 
K
+
/H
+
 
ratio
 
to proceed (Berger et al., 1997)
. Achieving such an 
increase by mica 
dissolution alone would require large amounts of mica to be altered/dissolved.
 
Large amounts of mica are present in the Marshall/Michigan, but not in intervals containing illite 
cements. 
Illite formation would be expected in micaceous intervals if detrital
 
micas were playing 
a large role in the reaction.
 
 
X
-
ray Diffraction
 
 
Among 
the diffractograms 
acquired for this study, 
the 
(
001
)
 
(d
 
=
 
10 Å) illite peak is 
probably the only illite peak of sufficient quality to interpret with any illite
-
centered 
techniques
. 
Higher angle illite peaks, while present, are either 
obscured
 
by peak overlap or too small due to 
high 
diffraction 
intensities 
from
 
more abundant minerals (quartz and kaolinite). 
 
 
In agreement with the work of Zacharias (1992) and Cox (2015)
,
 
no smectite was 
identified in the Marshall/Michigan from XRD. The possibility of a smectitic mixed
-
layer 
component within illite cements was investigated by saturating XRD samples with ethylene 
glycol to encourage expansion of smectite layers.
 
In the pres
ence of smectite 
inter
layering one 
would expect to see shifting of the 
(
001
)
 
illite peak in the glycolated sample to 
slightly lower ° 
2

 
than
 
the 
(
001
)
 
peak in the unglycolated sample
. Such a shift was not observed in these 
121
 
 
samples
.
 
An alternative method o
f investigating smectite mixed layering is calculation of the Ir 
number
 

. Perfect 100% illite would have an Ir number of one. The larger the Ir 
number, the greater the component of smectite mixed layering. 
Difficulties
 
with the Ir number 
calc
ulation 
for
 
these samples were 
described
 
in the 
R
esults section, but the average Ir number of 
all samples was found to be 1.
0
6
2
. Extrapolation of Ir number to determine the percentage of 
smectite mixed layering is not encouraged as a confident means of det
ermination (

rodon, 1980). 
Instead, this number indicates a small
, 
minor
 
(<10 % smectite)
 
presence of smectite mixed 
layering. In combination with the lack of peak shifting upon glycolation it is inferred that the 
presence of smectite layers within these i
llite cements is minimal.  
 
 
Intensities of illite 
(
001
)
 
peaks varied considerably. This is due to a combination of 
variable illite abundance in different samples and variable degrees of efficiency in sample 
preparation. Regardless of 
(
001
)
 
peak intensity differences
,
 
the FWHM values throughout all 
samples 
span
ned a narrow range
 
(
 
= 0.
2075
° 
2

,
 

0.0
286
° 
2

). 
No trend 
of
 
FWHM as a 
function
 
of
 
increasing depth was 
identified
 
(Table 
9
).
 
 
 
Placement of the anchizone boundary positions varies 
among different workers 
within the 
illite crystallinity 
community. Original anchizone boundaries by Kübler (196
9
) were placed at 
FWHM 0.42° 2

 
(low
-
grade) and 0.25° 2

 
(high
-
grade) based on a scan speed of 1° 2

/minute. 
Thorough analysis and attempted calibration of these boundaries over various scanning 
condit
ions was conducted by Kisch (1990). Kisch determined that slower scanning speeds 
resulted in slight narrowing of the 
(
001
)
 
peak. Us
e
 
of slow scan speeds without adjustment of the 
anchizone boundaries may result in placement of IC data in a higher
-
grade the
rmal regime. 
Therefore, these boundaries 
should be
 
adjusted
 
based on scanning speeds. The scan speed in this 
study (0.24° 2

/minute) was initially selected to
 
optimize
 
signal to noise ratio and is on the 
122
 
 
slower end of the spectrum of scan speeds used in 
crystallinity studies. Us
e
 

anchizone boundaries (or anchizone boundaries used by workers 
using
 
similar scan speeds as 
Kübler) would result in conclusions that may not be re
alistic
 
due to peak narrowing from slow 
scan speeds. To adjust for the s
low scan speed used in this study anchizone boundaries were 
selected after 
Weber (1972) and 
Teichmüller et al. (1979) who used near
-
identical scan speeds of 
0.25° 2

/minute. These boundaries are 0.38° 2

 
(low
-
grade) and 0.19° 2

 
(high
-
grade). 
 
 
FWHM values
 
in this study 
plot in the anchizone and along
 
the epizone
-
anchizone 
boundary, with the average value plotting just within the anchizone range. The temperatures 
associated with these boundaries are not exact, as different researchers have different working
 
definitions of what constitutes the anchizone. The anchizone is the name assigned to the 
transition zone between diagenesis and metamorphism, and is defined by Kisch (1990) as a zone 
of weak to very weak metamorphism associated with the sub
-
greenschist fa
cies. Assignment of 
exact temperature values to the boundaries on the Kübler Index would be misleading
 
(Kisch, 
1990)
, so the boundaries are used for qualitative analysis of the thermal regime in which illite 
forms. All FWHM data from authigenic illite in t
he Marshall and Michigan plot in high thermal 
regimes on the Kübler Index. In the context of
 
the
 
two p
reviously proposed
 
models for K
+
 
delivery these results align more with the model of hot brine movement up through the basin. 
Meteoric water and formation
 
waters of overlying strata would not reach temperatures close to 
those required to place illite crystallinity on the anchizone
-
epizone boundary. Although it is 
not 
definitive
 
that authigenic illite formed at temperatures of 200
-
300° C, these results indic
ate that 
significant heating was required to account for such well crystallized illite. 
 
 
 
123
 
 
Transmission Electron Microscopy
 
 
Many transmission electron microscope studies of clay minerals orient their samples with 
the basal planes 
perpendicular to the sample grid (parallel to the electron beam) in order to view 
the silicate layering. Doing so allows one to image layer spacings, layer sequences, and 
variation
s
 
in polytype. Sample preparation in this study consisted of depositing a dr
op of 
suspended powdered clay onto the sample grid. This method results in orientation of individual 
flakes with their basal (001) planes parallel to the sample grid. As a result, lattice fringe imaging 
w
as
 
not possible. Rather, crystal habit
, composition,
 
and stacking order
 
were the focus of 
investigation.
 
 
 
 
Illite crystallites have two commonly reported habits: lath shaped crystallites and 
euhedral pseudohexagonal crystallites
 
(Grubb and Peacor, 1991)
. The dominant habit in 
mudrocks is 
understood
 
to be c
ontrolled by polytypic variation associated with formation 
conditions, although illite habit is not yet fully understood. TEM studies of illite in sandstones 
are less common. Nadeau et al. (1985) and Güven (2001) present images of fibrous cements 
from sand
stones obtained with an identical sample preparation method as used in this study. All 
samples from five fibrous illite cemented sandstone units display lath morphology. Sizes of laths 
vary between units, but it appears that the lath crystal habit is domin
ant for fibrous cements. 
Güven presented evidence of authigenic illite laths nucleating on kaolinite cores
; however, 
no 
such phenomenon was observed within Michigan/Marshall samples
 
in TEM, although it was 
seen in SEM (Fig. 2
5
). 
 
 
Selected area electron di
ffraction (SAED) d
ata from illite cements prepared using 
identical
 
sample preparation procedure as this study were presented by Freed and Peacor (1992). 
Freed and Peacor (1992) demonstrated the effect of illite stacking disorder on SAED patterns. 
124
 
 
Coherent 
ordered stacking of illite cement layers results 
in 
SAED patterns with sharp, discrete 
spots as expected from a single crystal pattern, whereas disordered layer stacking results in 
ringed patterns that resemble powder SAED patterns. Illite cement SAED data
 
presented in this 
study display patterns with discrete spots and slight traces of rings (Fig. 6
7
, 6
8
)
, indicating 
dominantly ordered stacking with minor amounts of disorder. Ordered layer stacking is not fully 
understood, but is thought to be associated with higher temperature formation (
Grubb, 1991; 
Bauluz et al., 2002).
 
 
Illite Cement Morphology
 
Calcula
tions completed by Wilkinson and Haszeldine (2002) conclude that the growth of 
authigenic illite fibers as seen in sandstones is a process controlled by nucleation kinetics rather 
than growth kinetics. Based on observation of 
laths and 
fibers there appears
 
to be a maximum 
size to which an illite fundamental particle can grow, and once that size is reached growth of the 
fiber can only proceed if nucleation of another fundamental particle occurs
 
(Wilkinson and 
Haszeldine, 2002)
. This situation requires geolog
ical conditions that provide a high degree of 
pore fluid supersaturation with respect to illite. Thermodynamic stability is not sufficient to form 
illite fibers
;
 
the conditions must allow for continual, 
rapid
 
fundamental particle nucleation. 
According to W
ilkinson and Haszeldine 
(2002) 
these conditions must be more rigorous than 
slight supersaturation. They suggest that high temperatures and abnormally high pore fluid 
supersaturations are required for fibrous cement growth. These conditions are more readily
 
met 
by the deep brine K
+
 
delivery model. 
 
Geochemical Modeling
 
 
The geochemical scenarios resulting from the two proposed models of 
K
+
 
delivery can be 
simplified and modeled using model
ing software
. The goal of mixing modeling was to use 
125
 
 
compositional inputs from Michigan Basin pore fluid data and plot results on illite
-
inclusive 
K
2
O
-
Al
2
O
3
-
Si
O
2
-
H
2
O
 
mineral stability diagrams to evaluate 
whether
 
geologic scenarios 
representing the proposed K
+
 
deliver
y hypotheses would 
allow for thermodynamic stability of 
illite. 
 
 
The first step of modeling these scenarios was selection of a
 
solution
 
to represent 
Marshall pore fluid in sections of the basin where illitization has occurred. Illite was not present 
in the shallow sample well (CH101) or the intermediate depth sample well (RASA 5). Illite was 
only identified in the two deep wells which are both located 
in the middle of the basin. 
Unfortunately, USGS Open
-
File Report 90
-
368 did not collect groundwater samples at this depth 
in the Marshall. This is because the USGS report was part of the Michigan Basin Regional 
Aquifer
-
System Analysis program (RASA) which 
focused on sampling aquifers within the basin. 
The deepest central part of the Marshall/Michigan 
is
 
not considered part of the aquifer, and 
therefore w
as
 
not targeted for sampling
 
during the RASA program
. To 
compensate for
 
this
,
 
the 
well selected to repres
ent
 
solute chemistry in
 
the illite
-
bearing section of the Marshall/Michigan 
(Lake County Well #3) was the deepest and among the most saline of all Marshall wells sampled 
in the report. 
 
 
Using PHREEQC 
the composition of 
Lake County Well #3 was brought to 
u
ndersaturation with respect to illite (K
-
mica). 
A similar
 
calculation
 
was completed to bring 
meteoric water to equilibrium with respect to K
-
feldspar in order to represent enrichment in K
+
 
from K
-
feldspar dissolution. The deep brine compositions used were 
selected based on location 
and data availability and were unmodified.
 
 
According to Lanson (2002)
,
 
kaolinite
-
illitization reaction 3 
and illite neoformation 
reaction 4 both require an 
influx of fluid with a high K
+
/H
+
 
ratio
 
to initiate
. When plotting the 
126
 
 
r
esults from mixing models the data should be interpreted in the context of movement
 
of the 
solution composition
 
within/out of the illite stability 
field
 
and movement along the Y
-
axis which 
represents the K
+
/H
+
 
ratio. Regardless of movement along the X
-
axis, the 
formation of
 
illite will 
not be possible unless mixing causes an increase in the K
+
/H
+
 
ratio.
 
 
The meteoric water mixing model shows that increasing the amount of meteoric water 
entering the system (
desp
ite
 
the minor K
+
 
enrichment) dilutes the Marshall brines
 
with respect to 
the K
+
/H
+
 
ratio
. The 
compositional 
trend of the data decreases along the Y
-
axis, indicating that 
this model fails to provide 
the thermodynamic 
conditions necessary
 
for i
llite formatio
n
. 
 
 
The Devonian brine mixing model yielded similar results as the meteoric model. The 
slope of the mixing line is much nearer to zero, indicating that the Devonian brine is not diluting 
the Marshall brine, but is still failing to provide an increase in t
he K
+
/H
+
 
ratio. 
F
ailure to move 
datapoints up the Y
-
axis 
into the illite stability field 
indicates that this model also fails to provide 
a large enough K
+
/H
+
 
ratio increase to initiate 
an illite
-
forming 
reaction. 
 
 
Failure of the Devonian brine mixing mode
l necessitated that another brine mixing model 
be 
considered
. Higher salinity Silurian brines were selected as an input to address the issue of 
insufficient K
+
/H
+
 
ratio with mixing. The Silurian brine mixing 
trend
 
has a very 
steep
 
slope 
through 50% mixing. After 50% mixing the data begins to plot outside of the illite stability field 
and in the muscovite stability field. This is not an issue however, because movement of the 
mixing line up the Y
-
axis while remaining within the illit
e stability field indicates that mixing 
with Silurian brines could provide the K
+
/H
+
 
ratio increase required to initiate illitization reaction 
3. Once initiated this reaction will consume K
+
 
and release H
+
, the combination of which would 
begin to move 
comp
ositions
 
down the Y
-
axis, toward where present Marshall groundwater data 
plot. 
 
127
 
 
If mixing of pre
-
illite Marshall brines with Silurian brines were to occur, the resulting
 
solution
 
would 
plot increasingly upward
 
along the Y
-
axis. When a K
+
/H
+
 
ratio large eno
ugh to 
initiate reaction
 
3 or reaction 4
 
is reached the
(se)
 
reaction
(s)
 
would 
initiate and proceed
 
to d
rop 
the K
+
/H
+
 
ratio
 
along the Y
-
axis 
and 
into the range of modern data. What is not accounted for in 
this scenario is the slightly greater silica concent
ration observed in modern data than such a 
mixing
-
reaction path
 
would provide. To reconcile this, consider illite
-
forming reaction 2. This 
reaction was initially determined infeasible because it is 
understood
 
to occur
 
only
 
at high 
temperatures
 
(Lanson
 
et al.
, 2002)
. Illite crystallinity data indicate that high temperature 
formation of illite was likely. It is also possible that
,
 
contrary to the paragenetic sequence
,
 
minor 
amounts of K
-
feldspar dissolution 
could have occurred later than expected
. The po
ssibility of 
this late K
-
feldspar dissolution should not be rule
d
 
out, as skeletal K
-
feldspar grains are observed 
in open pore space (Fig. 1
8
) and may be contributing to the aqueous silica activity
. The 
coexistence of K
-
feldspar, kaolinite, and hot saline 
fluid could potentially initiate reaction
 
2 
in 
addition to reactions 
3
 
and 4
. Reaction 2 yields aqueous silica as a product phase, which could 
account for the X
-
axis increase our modeled data would require to return to modern groundwater 
chemistry. It is w
orth noting again that the slopes of experimentally determined kaolinite
-
illite 
stability boundaries
 
are oriented
 
in a top
-
left to bottom
-
right direction (
F
ig. 
7
8
), 
similar to
 
the 
trend of modern data and t
he expected outcome of
 
reaction 2 
occurring concur
rently with 
reactions 3 and 4. 
 
 
Re
-
evaluation of Potassium
-
Delivery Hypotheses
 
The 
potassium
-
enriched meteoric water hypothesis is 
refuted
 
by previous work a
nd 
results presented here. The enriched meteoric water hypothesis relies on the ability of 
these 
waters to percolate down through overlying strata
 
in order to deliver potassium as K
+
 
aqueous 
128
 
 
ions. Ging
 
et al.
 
(1996) report that the deepest portion of the Michigan/Marshall is hydrologically 
isolated from meteoric water recharge based on trends of
 

18
O values for much of 
the Marshall 
a
quifer are interpreted as a mixture of meteoric water and formational brine
 
(Ging 
et al., 1996)
;
 
however
,
 
values from the deeper central part of the basin become isotopically 
heavier
 
(Ging et al., 1996)
, indicating that meteoric water did not make it to this 
interval
. 
Authigenic illite cements were only located in the deepest part of the Michigan/Marshall
;
 
therefore
,
 
enriched meteoric water recharge could not provide the K
+
 
ions. 
 
The location of illite cementation in the center of the basin likely has little to do with the 
overburden pressure 
or maximum burial depth 
that this 
interval
 
was subjected to. The Marshall 
Sandstone has a maximum 
present 
depth of ~1500 ft. 
(~400 

 
450 m; T
able 1)
, which according 
to Lanson et al. (2002) is not sufficient burial depth to destabilize the K
-
feldspar
-
kaolinite 
mineral assemblage
 
(e.g., 3 

 
5 km
, 10,000 

 
16,000 ft.
 
in the North Sea)
. 
Past burial depths 
of 
Paleozoic strata in the Michigan Basin 
may have been greater at some point in the past
. 
However
,
 
in the absence of almost any post
-
Carboniferous and pre
-
Pleistocene strata throughout 
the Michigan Basin
 
(except ~100 m of Jurassic redbeds)
, burial history curves cannot be 
constructed
. Instead, 
ma
ximum burial depths of 
Paleozoic
 
strata 
in the Michigan Basin 
must be 
inferred from thermal maturity indicators, which in turn must be interpreted invoking 
assumptions about the nature and thermal conductivity of the missing strata
 
(
see references in the 
I
ntroduction, and review by 
Velbel, 2009). P
revious estimates
 
of maximum burial depth by 
such 
approach
es
 
have ranged from 
300

2000 m
 
(1000

6540 ft) of post
-
Pennsylvanian sediment
, 
which would have had to have been completely eroded by the Jurassic
 
(
see review by 
Velbel, 
2009)
. 
The timing of heating along the geothermal gradient at maximum burial (between Late 
Pennsylvanian and Jurassic; Velbel, 2009) is consistent with the previously acquire
d
 
K
-
Ar dates 
129
 
 
for the smallest Marshall Sandstone fine fracti
on of 294 and 275 Ma (Early Permian) (Cox, 
2015). However, e
ven with the maximum previously proposed thickness of now
-
missing former 
overburden (2 km)
, 
the maximum burial depth of the Marshall 
S
andstone would have been ~2.5 
km (8,000 ft.)
, shallower than t
he depth interval of illite authigenesis in sandstones (Lanson et 
al., 2002)
. The available
 
evidence 
does not favor the hypothesis 
 
that a burial model could 
account for temperatures required to initiate illitization in the 1500 ft
. present
 
depth range of 
the 
Marshall/Michigan. 
 
T
he spatial 
distribution
 
of illite cementation within the 
Marshall Sandstone
 
could be a 
function of proximity to
 
shallow 
structures formed by reactivation of faults in
 
the Mid
-
Continent 
Rift, as 
previously 
proposed by Cox (2015). 
Some structures in Michigan Basin Paleozoic strata 
are inferred to have formed by post
-
depositional and post
-
lithification reactivation of faults in the 
MCR beneath the Paleozoic strata
 
(Ma et al., 2009).
 
Such fracture and fault sys
tems may have 
served as conduits by which hydrothermal fluids could have been introduced into overlying 
strata
 
(Luczaj et al., 2006; Velbel, 2009)
. Cox 
(2015) 
observed authigenic illite in all four sample 
wells 
near
 
the Albion
-
Scipio and Lucas faults. The 
two samples wells from this study in which 
authigenic illite was observed were both 
located 
near
 
the MCR in Michigan (
F
ig. 
1
0
). Invocation 
of deep hot brine movement along MCR associated faults/fractures does an acceptable job of 
explaining authigenic illi
te spatial distribution within 
the basin. This correlation is supported by 
evidence of hydrothermal dolomitization
 
occurring along faults 
interpreted
 
to be associated with 
the MCR
 
(Luczaj et al., 2006). 
 
 
The precise origin of the suspected 
hydrothermal fluid is difficult to determine. A large 
K
+
/H
+
 
ratio increase is required 
to initiate illite
-
forming reactions
. Such an increase is possible 
from mixing with Michigan Basin Silurian brines, although not possible from 
mixing with 
 
130
 
 
other brines.
 
Enrichment of formation fluids in K
+
 
along upwardly mobile hydrothermal brine 
flow paths may be possible considering the presence of sylvite in the Silurian 
 
age Salina Group (Nurmi, 1977). 
Specific to the Michigan Basin, i
nteraction of sylvite with 
hydro
thermal fluid may allow less saline brines to reach K
+
/H
+
 
ratios large enough to initiate 
illitization. 
 
 
 
 
131
 
 
Recommendations for Future Work
 
 
XRD analysis remains a useful tool to investigate clay minerals and could be relied on as 
the primary 
analytical technique for further study. Us
e
 
of the Kübler Index for crystallinity is still 
recommended, but a more thorough XRD investigation of authigenic illite 
would
 
prove useful
 
to 
determine 
whether 
crystallinity values vary throughout the basin
. Incre
asing the spatial 
distribution and the number of samples 
may
 
provide a
 
better understanding of authigenic illite 
occurrence in the Michigan Basin
, provided that proper selection of anchizone boundaries is 
considered and results are interpreted with caution
. Identification of authigenic illite in sample 
wells 
more widely distributed
 
throughout the basin, and XRD analysis of a larger number of 
those samples
,
 
may reveal trends in illite crystallinity that could be used to investigate potential 
heterogeneity of
 
illitizing hydrothermal flow events.
 
Perhaps illite crystallinity 
varies 
considerably 
between illite cement
-
bearing wells.
 
Further i
nvestigation of illite polytypism
 
and 
stacking order
, although not fully understood, 
would
 
provide supportive evidence and reveal 
information about the nature of these hydrothermal events. Polytypism may be investigated by 
randomly oriented clay XRD, as well as TEM analysis of samples with bas
al
 
planes parallel to 
the electron beam. 
 
TEM analys
is of samples
 
extracted
 
from thin sections would allow imaging of lattice 
spacing in detrital grains experiencing alteration. This would resolve uncertainties regarding the 
degree of alteration 
in
 
detrital biotite
 
and muscovite. 
These alteration reactions 
would help to 
explain diagenetic changes in pore fluid composition, which can be incorporated into a more 
complete evaluation of 
Michigan Stray and Marshall Sandstone
 
diagenesis.  
 
Stable isotopic analysis of illite cements would yield an isotopic signatur
e that would 
provide strong support regarding the origin of formation fluid. This analysis would be difficult
,
 
132
 
 
however, because complete separation of illite from clay sized quartz would be required for 
accurate results, and such a complete separation is d
ifficult and time consuming. Any degree of 
quartz contamination would 
reduce the usefulness
 
of
 
the
 
results.
 
Mass
-
balance calculations using quantitative mineralogy 
and Marshall/Michigan pore 
fluid inputs 
can be applied to determine 
whether
 
the abundance of
 
K
-
feldspar and detrital mica in 
the Michigan/Marshall are sufficient to act as an internal source of K
+
 
for the observed amount 
of authigenic clay material
.
 
This calculation would definitively prove 
whether
 
an external K
+
 
source is required for illitizati
on.
 
 
Due to 
the 
thermodynamic data
set used in this study
,
 
K
2
O
-
Al
2
O
3
-
Si
O
2
-
H
2
O
 
mineral 
stability
-
field
 
boundaries and mixing model
s
 
were 
generated for
 
25°C. Illite crystallinity data 
indicate that formation temperatures of authigenic illite were much greater than 25°C. 
Rerunning 
mixing models at more appropriate temperatures (200
-
300°C) and plotting data on a stability 
diagram constructed using high tem
perature thermodynamic data
, and explicitly taking into 
account the effects of phengitic solid
-
solution effects in illite thermodynamics
 
would allow for a 
more realistic treatment of illite
 
thermodynamic
 
stability at the temperature 
at 
which it formed
.
 
 
In
 
addition to
 
illite cementation, intriguing questions regarding other clay minerals in the 
M
arshall/Michigan
 
exist. Kaolin material, identified in the Marshall as dickite based on 
morphology, was observed at depths much shallower than expected. The kaolini
te
-
to
-
dickite 
reaction is kinetically limited by temperature (Beaufort et al., 1998). We know that hydrothermal 
activity 
occurred
 
within these 
strata
, but dickite was not observed co
-
occurring with illite (our 
hydrothermal indicator) in this study. Investi
gation of Marshall dickite 
by
 
XRD and TEM may 
provide further insight into controls on hydrothermal activity
,
 
as well as contribute to the kaolin 
community. 
 
133
 
 
Authigenic chlorite w
as
 
observed as a
 
pervasive
 
diagenetic component in shallow 
portions of the Mi
chigan/Marshall but absent in such abundance in deeper portions. Focused 
study of chlorite cements (including chlorite crystallinity XRD work) would be useful to 
determine
 
whether the
 
distribution of chlorite cement is a function of depth, spatial position
, 
hydrology, aqueous geochemistry, mineralogy, or some combination of factors. Constraining the 
factors affecting 
authigenic 
chlorite distribution is a major component of reservoir quality 
evaluation, so perhaps the Michigan/Marshall could serve as a uniqu
e study area and an analog 
for similar systems. 
 
 
 
 
 
 
 
 
 
 
 
 
 
134
 
 
Conclusions
 
Late diagenesis in the Michigan/Marshall Sandstones consisted of pore
-
filling kaolinite 
cement formation, followed by subsequent illitization of that kaolinite
 
and neoformation of 
fibrous illite cements
. Detrital phyllosilicates, while abundant in 
some intervals
 
of these strata, 
are likely not providing the sole 
K
+
 
ion source for illitization
; this interpretation is
 
based 
mainly 
on the lack of co
-
occurrence of t
hese grains with illite cements
. Mixing of pre
-
illite Marshall 
pore fluid with Silurian brine allows for illite thermodynamic stability and provides the K
+
/H
+
 
ratio increase that
 
initiates
 
illite
-
forming reactions
, whereas mixing with enriched meteoric wat
er 
dilutes Marshall pore fluid with respect to this ratio. Illite crystallinity as determined by the 
Kübler Index indicates that authigenic illite formed in a high
-
temperature
 
thermal regime. These 
lines of evidence converge to support the previously propo
sed hypothesis that deep, hot brines in 
the Michigan Basin mov
ed 
vertically across formations 
along
 
reactivated faults and fractures to 
provide the 
potassium required
 
for the 
formation of illite
. 
Faults and fracture systems above the 
Mid
-
Continent Rift pro
bably s
erved as conduits for
 
hydrothermal flow
,
 
based on spatial 
distribution of illite cements within the basin.
 
These hot brines r
e
acted with kaolinite cements to 
form illite 
pseudomorphic after kaolinite 
by
 
transformation (reaction 3) and 
directly precipitated 
fibrous illite (
neoformation
;
 
reaction 4)
.
 
It is possible that high temperatures associated with this 
hydrothermal activity resulted in destabilization of the K
-
feldspar
-
kaolinite mineral assemblage 
an
d initiated an illite
-
forming reaction
 
(reaction 2)
 
in which K
-
feldspar and kaolinite are both 
consumed. 
 
 
 
135
 
 
 
 
 
 
 
 
 
 
 
 
 
 
APPENDI
CES
 
 
 
136
 
 
APPENDIX
 
A
: Point
-
count data
 
 
Table 10: 
Point
-
count data
: 
D
ata from point counting 200 points for each thin section. 
MQz 

 
monocrystalline quartz, PQz 

 
polycrystalline quartz, CQz 

 
chert, Kfs 

 
K
-
feldspar, 
-
Pl 

 
plagioclase, SRf 

 
sedimentary rock fragment, MRf 

 
metamorphic rock fragment, Carb 

 
unspecified car
bonate, Ms 

 
muscovite grain, Bt 

 
biotite grain , Py 

 
pyrite, Chl 

 
chlorite, Kln 

 
kaolinite, Pseudo 

 
pseudomatrix, Dol 

 
dolomite, Clay 

 
clay sized material (deemed not to be 
pseudomatrix). 
 
 
 
 
137
 
 
APPENDIX
 
A 
-
 
CONT.
 
 
Table 10 (cont.)
 
 
138
 
 
APPENDIX 
A 
-
 
CONT.
 
 
Table 10 (cont.)
 
 
139
 
 
APPENDIX
 
A
 
-
 
CONT.
 
 
Table 10 (cont.)
 
 
140
 
 
APPENDIX
 
B
: Thin section descriptions
 
 
 
Thin section descriptions
 
W
-
1002 Well:
 
1315 feet:
 
Sample consists of moderately well
-
sorted, very fine to fine
-
grained greywacke. QFR is 
95:3.33:1.66. Almost all porosity is occluded by detrital pseudomatrix. Pseudomatrix comprises 
36% of the entire sample, leaving 0.5% of porosity remaining. Authigenic pyrite comprises 3.5% 
of the sample.
 
 
 
1328 feet:
 
Sample consists of well
-
sorted, fine
-
grained sublitharenite. QFR is 81:2:17. Rock fragments 
consist of mudstones and metasedimentary fragments. The sample contains 11% porosity, and 
11% authigenic carbonate cement. Much of the carbonate cement is partially dissolved, resulting 
in
 
microporosity within the cemented areas. 
 
 
1344 feet:
 
Sample consists of well
-
sorted, very fine to fine
-
grained litharenite. QFR is 61.6:1.7:36.6. Rock 
fragments consist primarily of metasedimentary fragments. This sample contains 8.5% porosity. 
Muscovite
 
grains comprise 3% of this sample. 
 
 
1356 feet:
 
Sample consists of well
-
sorted, very fine to fine
-
grained litharenite. QFR is 66.3:1.2:32.5. Rock 
fragments consist primarily of metasedimentary fragments. This sample contains 7% porosity. 
Muscovite grains 
comprise 3.5% of this sample, and authigenic carbonate cements comprise 
2.5% of the sample. 
 
 
1363 feet:
 
Sample consists of well
-
sorted, very fine to fine
-
grained litharenite. QFR is 66.8:2.3:30.9. Rock 
fragments consist of metasedimentary fragments. This 
sample contains 8% porosity. Muscovite 
grains comprise 2% of this sample. Authigenic carbonate comprises 1%, and authigenic pyrite 
1% of the sample. 
 
 
141
 
 
APPENDIX
 
B
 
-
 
CONT.
 
1382 feet:
 

-
sized do
lomite comprises 
81.5% of the sample. Authigenic pyrite comprises 3.5% of the sample. This sample has 15% 
porosity.
 
 
1401 feet:
 
Sample consists of well
-
sorted, fine
-
grained sublitharenite. QFR is 78.2:9.7:12.1. The feldspar is 
entirely K
-
feldspar, and the 
rock fragments are entirely metasedimentary fragments. This sample 
contains 8.5% porosity. Authigenic pyrite is pore filling and framboidal and comprises 7% of the 
sample. Muscovite grains comprise 2% of the sample. 
 
 
1406 feet:
 
Sample consists of well
-
sor
ted, fine
-
grained sublitharenite. QFR is 84:1.2:14.8. The feldspar is 
entirely K
-
feldspar, much of which has been partially dissolved. The rock fragments are entirely 
metasedimentary lithic fragments. This sample contains 10.5% porosity. Authigenic pyrite 
is 
pore
-
filling and framboidal and comprises 7% of the total sample. 
 
 
1411 feet:
 
Sample consists of well
-
sorted, very fine to fine
-
grained greywacke. QFR is 94.4:2.8:2.8. No 
porosity is remaining in the sample as a result of deformed pseudomatrix comprisi
ng 24% of the 
sample. Framboidal authigenic pyrite comprises 4% of the sample, and muscovite grains 
comprise 2%. 
 
 
1420 feet:
 
Sample consists of moderately sorted, very fine to fine
-
grained litharenite. QFR is 70.2:2.4:27.4. 
All of the feldspar is K
-
feldsp
ar, and the rock fragments consist of sedimentary and 
metasedimentary fragments. This sample contains 5.5% porosity. Authigenic pyrite framboids 
comprise 8% of the sample. Muscovite grains comprise 1% of the sample. 
 
 
 
 
 
142
 
 
APPENDIX
 
B 
-
 
CONT.
 
1435 feet:
 
Sample consists of very well
-
sorted, fine to medium
-
grained sublitharenite. QFR is 
76.1:2.4:21.5. The feldspar is K
-
feldspar, the rock fragments are metasedimentary fragments. 
This sample contains 10% porosity. Authigenic pyrite, occurring as framboids and
 
pore
-
filling 
cement, comprises 7.5% of the sample. 
 
 
1444 feet:
 
Sample consists of well
-
sorted, fine
-
grained feldspathic litharenite. QFR is 70.3:10.3:19.4. The 
feldspar is dominantly K
-
feldspar. Much of the K
-
feldspar grains display overgrowths, many of 
which have been preferentially dissolved relative to the cores. The rock fragments are 
metasedimentary fragments. This sample contains 10.5% porosity. Framboids of authigenic 
pyrite comprise 5.5% of the total sample. 
 
 
1448 feet:
 
This sample consists of mo
derately well
-
sorted, fine
-
grained litharenite. QFR is 72.3:4.8:22.9. 
The feldspar is K
-
feldspar (some of which has been partially dissolved), and the rock fragments 
are metasedimentary. This sample contains 9% porosity. Authigenic pyrite framboids compris
e 
2.5% of the total sample. Muscovite grains comprise 3.5% of the total sample. 
 
 
W
-
962 Well:
 
1328 feet:
 
Sample consists of well
-
sorted, very fine to fine
-
grained litharenite. QFR is 64.1:0:35.9. The 
rock fragments are sedimentary mudstone and metasediment
ary fragments. This sample contains 
only 4% porosity. Muscovite grains comprise 6% of the total sample. 
 
 
1338 feet:
 
Sample consists of well
-
sorted, fine
-
grained litharenite. QFR is 67.5:3.7:28.8. The feldspar is 
plagioclase and K
-
feldspar, and the rock fr
agments are dominantly metasedimentary fragments, 
with some mudstone fragments. This sample contains 10% porosity. Authigenic pyrite cements 
(pore
-
filling and framboidal) comprise 5% of the sample, and muscovite grains comprise 3.5% 
of the total sample. 
 
 
 
143
 
 
APPENDIX 
B 
-
 
CONT.
 
1343 feet:
 
Sample consists of very well
-
sorted, medium to fine
-
grained sublitharenite. QFR is 
77.3:1.5:21.2. The rock fragments dominantly metasedimentary lithic fragments. This sample 
contains 14% porosity. Significant amounts of authigenic framboidal pyrite (9%) and authigenic 
carbonate (9%) o
cclude pore space.  
 
 
1373 feet:
 
Sample consists of well
-
sorted, very fine to fine
-
grained litharenite. QFR is 64.4:1.5:34.1. The 
rock fragments are metasedimentary fragments. This sample contains 10.5% porosity. This 
sample contains anomalously large amou
nts of pyrite cement. Framboidal and pore
-
filling 
authigenic pyrite comprises 20% of the total sample. 
 
 
1382 feet:
 

of this dolomite is silt sized, however some crystals are up to 0.55 mm in length. This sample 
contains 23% porosity. Authigenic framboidal pyrite comprises 11% o
f the total sample. 
 
 
1400 feet:
 

dolomite is made up entirely of silt sized crystals. This sample has almost no porosity remaining 
(0.5%). Framboidal authigenic p
yrite comprises 9.5% of the total sample. 
 
 
1420 feet:
 
This sample is a claystone. It consists of 60% clay sized grains with very fine
-
grained quartz 
(28%) and carbonate (9%) grains interspersed. Authigenic framboidal pyrite comprises 3% of the 
total sampl
e. 
 
 
1425 feet:
 
This sample is a well
-
sorted, fine
-
grained sublitharenite. QFR is 80.3:3.6:16.1. The feldspar is K
-
feldspar which has been partially dissolved, and the rock fragments are metasedimentary and 
mudstone fragments. This sample contains 10% poro
sity. Authigenic pyrite framboids comprise 
5% of the total sample. 
 
144
 
 
APPENDIX
 
B
 
-
 
CONT.
 
1430 feet:
 
Sample is a very well
-
sorted, fine to medium
-
grained sublitharenite. QFR is 80.5:4.9:14.6. The 
feldspar is K
-
feldspar, many of which contain overgrowths that 
appear to be preferentially 
dissolved. The rock fragments are metasedimentary fragments. This sample contains 10.5% 
porosity. Authigenic carbonate comprises 4% of the sample.
 
 
1448 feet:
 
Sample is a well
-
sorted, fine
-
grained sublitharenite. QFR is 79:7.4:1
3.6. The feldspar is entirely 
K
-
feldspar, and the rock fragments are entirely metasedimentary fragments. This sample 
contains 12.5% porosity. Authigenic pyrite (framboidal and pore
-
filling) comprises 5% of the 
total sample. 
 
 
1453 feet:
 
Sample is well
-
sort
ed, fine
-
grained sublitharenite. QFR is 77.5:4.7:17.8. The feldspar is K
-
feldspar and plagioclase, the rock fragments are entirely metasedimentary lithic fragments. This 
sample contains 9% porosity. Detrital mica grains comprise 4% of the total sample (mus
covite 
3%, biotite 1%). 
 
 
RASA 5 Well:
 
503 feet:
 
Sample is a very well
-
sorted, fine to medium
-
grained sublitharenite. QFR is 85:2.5:12.5. The 
rock fragments are metasedimentary and mudstone fragments. This sample contains 16.5% 
porosity. Kaolinite cements 
comprise 3.5% of the total sample. Very minor amounts of 
muscovite were observed, but not counted. 
 
 
514 feet:
 
Sample is a well
-
sorted, fine
-
grained litharenite. QFR is 72.4:4.1:23.5. The rock fragments are 
mudstone and metasedimentary fragments. This samp
le contains 10% porosity. Authigenic pyrite 
comprises 3% or the sample, and detrital muscovite 2% of the total sample.
 
 
 
 
145
 
 
APPENDIX
 
B 
-
 
CONT.
 
519 feet:
 
Sample consists of well
-
sorted, very fine to medium
-
grained sublitharenite. QFR is 
79.5:1.9:18.6. Rock fragments consist of metasedimentary and mudstone fragments. This sample 
contains 13% porosity. Authigenic components include pyrite, chlorite, and kaolinite. Muscovite 
comprise
s 2% of the total sample. 
 
 
531 feet:
 
Sample is well
-
sorted, fine
-
grained sublitharenite. QFR is 81.2:2.6:16.2. The feldspar is K
-
feldspar and plagioclase, and the rock fragments are metasedimentary and mudstone fragments. 
This sample contains 11.5% porosi
ty. Pore
-
filling kaolinite cements comprise 6% of the total 
sample. These cements contain microporosity within. 
 
 
588 feet:
 
This sample consists of moderately well
-
sorted, very fine to fine
-
grained micaceous 
sublitharenite. QFR is 76.2:1.6:22.1. The rock f
ragments are entirely metasedimentary 
fragments. This sample contains 33% detrital mica grains. 11.5% of the total sample is 
comprised of muscovite, and 21.5% of the total sample is biotite. The mica grains are all oriented 
in the same direction. This samp
le only contains 5% porosity. 
 
 
592 feet:
 
Sample consists of moderately well
-
sorted, very fine to fine
-
grained micaceous litharenite. QFR 
is 72.9:3.1:24. The rock fragments are entirely metasedimentary fragments. This sample contains 
25.5% detrital mica gr
ains. Muscovite comprises 11% of the total sample, and biotite comprises 
14.5% of the total sample. The mica grains are all oriented in the same direction. This sample 
contains only 5% porosity. 
 
 
598 feet:
 
Sample consists of well
-
sorted, very fine to fine
-
grained litharenite. QFR is 65.3:4.8:29.9. The 
rock fragments are metasedimentary and mudstone fragments. This sample contains 7.5% 
porosity. Chlorite (some weathered to a brown color) comprises 10% of the total sample. There 
are multiple types of authige
nic cement in this sample, including pyrite, carbonate, chlorite, and 
kaolinite. 
 
 
146
 
 
APPENDIX 
B 
-
 
CONT.
 
604 feet:
 
Sample consists of moderately
-
well sorted, very fine to fine
-
grained micaceous litharenite. QFR 
is 61.7:4.3:34. The rock fragments are metasedim
entary and mudstone fragments. Detrital mica 
grains comprise 39.5% of the total sample. Biotite (26.5%) and muscovite (13%) are 
preferentially oriented. 
P
yrite
 
cement
 
comprises 6% of the total sample
, and porosity only 3%.
 
 
607 feet:
 
Sample 
consists of moderately
-
well sorted, very fine to fine
-
grained micaceous litharenite. QFR 
is 68.1:3.4:28.5. The rock fragments are entirely metasedimentary fragments. Detrital mica 
grains comprise 30% of the total sample. Biotite (19%) and muscovite (11%) a
re preferentially 
oriented. This sample contains only 4% porosity. 
 
 
609 feet:
 
This sample consists primarily of detrital mica grains and authigenic pyrite with a preferred 
orientation. Biotite (29%) and muscovite (14.5%) comprise the majority of minerals 
in this 
sample. These mica grains are oriented 
with a fabric parallel to bedding
. Authigenic pyrite 
comprises 25.5% of the total sample. QFR make up only 24% of the total sample. QFR is 
75:8.3:16.7. Authigenic siderite micro
-
nodules comprise 1.5% of the sa
mple. This sample 
contains 5.5% porosity. 
 
 
611 feet:
 
This sample consists of 61.5% detrital mica grains. Biotite (38%) and muscovite (23.5%) grains 
are all preferentially oriented. Authigenic pyrite comprises 13% of the total sample. Much of the 
authigenic pyrite is also preferentially oriented along the sam
e plane as the mica grains. QFR 
combine for 24% of the total sample. QFR is 70.8:0:29.2. The rock fragments are mudstone and 
metasedimentary fragments. This sample contains only 1.5% porosity. 
 
 
612 feet:
 
This sample consists of moderately
 
well
-
sorted, ver
y fine to fine
-
grained sublitharenite. QFR 
comprises only 39% of the total sample. QFR is 83.3:0:16.7. Chlorite comprises 20% of the total 
sample. This chlorite is brownish green in color and occurs as discrete grains, many of which 
resemble the typical sh
ape of mica grains. Authigenic carbonate comprises 23% of the total 
sample. The carbonate is pore
-
filling calcite cement. Authigenic pyrite comprises 11.5% of the 
total sample. This pyrite is both pore
-
filling, and framboidal. This sample contains 5.5% 
por
osity. 
 
147
 
 
APPENDIX
 
B 
-
 
CONT.
 
CH101 Well:
 
20 feet:
 
Sample consists of very well
-
sorted, fine to medium
-
grained feldspathic litharenite. QFR is 
74.1:11.1:14.8. Feldspar grains are K
-
feldspar and plagioclase, and rock fragments are entirely 
metasedimentary frag
ments. Kaolinite cements with microporosity comprise 5% of the total 
sample. This sample contains 13% porosity. 
 
 
33 feet:
 
This sample consists of moderately well
-
sorted, very fine to fine
-
grained pyrite cemented 
sublitharenite. QFR is 77.2:3.5:19.3. The r
ock fragments are entirely metasedimentary 
fragments. This sample is cemented by authigenic pyrite. This pyrite cement comprises 41.5% of 
the total sample. As a result, this sample contains only 0.5% porosity. 
 
 
58 feet:
 
This sample consists of well
-
sorted
, very fine to fine
-
grained litharenite. QFR is 69.2:2.6:28.2. 
Rock fragments consist entirely of metasedimentary fragments. This sample contains 8% 
porosity. Authigenic components include pyrite, kaolinite, and chlorite. 
 
 
70 feet:
 
This sample consists of
 
wel
l
-
sorted, very fine to fine
-
grained litharenite. QFR is 69.7:2.7:27.6. 
Rock fragments consist entirely of metasedimentary fragments. This sample contains 19% 
porosity. Many of the grains are not in contact with one another. Detrital muscovite comprises
 
3% of the total sample. Authigenic components include pyrite, chlorite, and kaolinite. 
 
 
85 feet:
 
Sample consists of moderately well
-
sorted, very fine
-
grained litharenite. QFR is 68:2.7:29.3. The 
rock fragments consist entirely of metasedimentary fragment
s. Authigenic carbonate comprises 
7.5% of the total sample. This carbonate consists of rhombs that nucleate on detrital grains. 
Authigenic chlorite comprises 6% of the total sample. This chlorite is green in color and coats 
detrital grains. Other authigeni
c components include pyrite and kaolinite. This sample contains 
6% porosity. 
 
 
 
148
 
 
APPENDIX
 
B 
-
 
CONT.
 
100 feet:
 
Sample consists of well
-
sorted, very fine to fine
-
grained litharenite. QFR is 67.7:3.7:28.6. The 
feldspar is entirely K
-
feldspar, and the rock frag
ments are metasedimentary and mudstone 
fragments. Grain
-
coating authigenic chlorite comprises 5% of the total sample. Other authigenic 
components include carbonate, pyrite, and kaolinite. This sample contains 9.5% porosity. 
 
 
114 feet:
 
Sample 
consists of moderately well
-
sorted, very fine to fine
-
grained sublitharenite. QFR is 
76.5:2.9:20.6. The rock fragments are metasedimentary and mudstone fragments. Authigenic 
carbonate occurs as both pore
-
filling cement and rhombs nucleating on grains and c
omprises 
13.5% of the total sample. Authigenic chlorite occurs as grain
-
coatings and comprises 4% of the 
total sample. This sample contains 11.5% porosity. 
 
 
126 feet:
 
Sample consists of moderately well
-
sorted, very fine
-
grained litharenite. QFR is 69.2:2.
8:28. The 
feldspar is K
-
feldspar and plagioclase, and the rock fragments are metasedimentary and 
mudstone fragments. Authigenic carbonate rhombs comprise 12.5% of the total sample. 
Authigenic grain
-
coating chlorite comprises 5% of the total sample. Other a
uthigenic 
components include pyrite and kaolinite. This sample contains 5.5% porosity. 
 
 
133 feet:
 
Sample consists of well
-
sorted, very fine to fine
-
grained litharenite. QFR is 69.7:3.6:26.7. The 
feldspar is K
-
feldspar and plagioclase, and the rock fragmen
ts are metasedimentary and 
mudstone fragments. Authigenic grain
-
coating chlorite comprises 7% of the total sample. Other 
authigenic components include carbonate, kaolinite, and pyrite. This sample contains 7% 
porosity. 
 
 
140 feet:
 
Sample consists of well
-
s
orted, very fine to fine
-
grained litharenite. QFR is 70.8:4.5:24.7. The 
feldspar is K
-
feldspar, and the rock fragments are metasedimentary and mudstone fragments. 
Authigenic pyrite comprises 7.5% of the total sample. This pyrite is pore
-
filling and framboi
dal. 
Muscovite grains comprise 3% of the total sample. This sample contains 11.5% porosity. 
 
 
149
 
 
APPENDIX
 
C
: Marshall aquifer groundwater data
 
 
Table 11: 
Marshall aquifer groundwater data
: 
Selected aqueous geochemical data from the 
Marshall aquifer 
(compiled from USGS Open
-
File Report 90
-
368)
.
 
 
 
 
 
150
 
 
APPENDIX
 
C 
-
 
CONT.
 
 
Table 11 (cont.)
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
151
 
 
APPENDIX C 

 
CONT.
 
 
Table 11 (cont.)
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
152
 
 
APPENDIX
 
D
: Aqueous geochemical modeling well information
 
 
 
Table 12: 
Aqueous 
geochemical modeling well information
: 
Additional information regarding 
wells involved in geochemical modeling. 
 
 
 
 
 
153
 
 
APPENDIX E
: Pore fluid analyses used in aqueous geochemical modeling
 
 
Table 13: 
Pore fluid analyses used in aqueous 
geochemical modeling
:
 
All analyses used in 
modeling work. Compositions denoted with a * are initial inputs into mixing models. 
 
 
 
 
 
154
 
 
APPENDIX 
F
: XRD diffractograms
 
 
 
 
 
Figure 80: 
W
-
1002 XRD 
diffractograms
: 
Diffractogram
s
 
from
 
ungylocated (top) and 
glycolated
 
(bottom)
 
sample
s.
 
 
 
Sample d
epth: 1328 ft.
 
155
 
 
APPENDIX
 
F
 
-
 
CONT.
 
 
Figure 80 (cont.)
 
 
 
 
 
 
Sample d
epth: 13
44
 
ft.
 
156
 
 
APPENDIX
 
F
 
-
 
CONT.
 
 
Figure 80 (cont.)
 
 
 
 
 
Sample d
epth: 13
63
 
ft.
 
157
 
 
APPENDIX
 
F
 
-
 
CONT.
 
 
Figure 80 (cont.)
 
 
 
 
 
 
 
 
 
Sample d
epth: 1
406
 
ft.
 
158
 
 
APPENDIX
 
F 
-
 
CONT.
 
 
Figure 80 (cont.)
 
 
 
 
 
 
 
 
 
Sample d
epth: 1
420
 
ft.
 
159
 
 
APPENDIX
 
F
 
-
 
CONT.
 
 
Figure 80 (cont.)
 
 
 
 
 
 
 
 
 
Sample d
epth: 1
435
 
ft.
 
160
 
 
APPENDIX
 
F 
-
 
CONT.
 
 
Figure 80 (cont.)
 
 
 
 
 
 
 
 
Sample d
epth: 1
444
 
ft.
 
161
 
 
APPENDIX
 
F 
-
 
CONT.
 
 
 
 
 
 
Figure 81: 
W
-
962 XRD diffractograms
: Diffractograms from unglycolated (top) and glycolated 
(bottom) sample
s.
 
 
 
 
Sample d
epth: 13
38
 
ft.
 
162
 
 
APPENDIX 
F 
-
 
CONT.
 
 
Figure 81 (cont.)
 
 
 
 
 
 
 
Sample d
epth: 13
43
 
ft.
 
163
 
 
APPENDIX
 
F
 
-
 
CONT.
 
 
Figure 81 (cont.)
 
 
 
 
 
 
 
 
 
Sample d
epth: 1
425
 
ft.
 
164
 
 
APPENDIX 
F 
-
 
CONT.
 
 
Figure 81 (cont.)
 
 
 
 
 
 
 
 
 
Sample d
epth: 1
430
 
ft.
 
165
 
 
APPENDIX
 
F 
-
 
CONT.
 
 
Figure 81 (cont.)
 
 
 
 
 
 
 
 
Sample d
epth: 1
448
 
ft.
 
166
 
 
APPENDIX
 
G
: X
-
ray diffraction illite Ir 
number peak measurements
 
 
Table 14: 
X
-
ray diffraction illite Ir number peak measurements
: 
XRD Ir number peak 
measurements: Raw measurements of peak intensities used to calculate the Ir number.
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
167
 
 
APPENDIX H
: Transmission electron 
microscopy EDS data
 
 
Figure 82: 
W
-
1002 transmission electron microscopy 
EDS
 
data
: TEM
-
EDS data for illite lath at 
1406 ft. (depth). 
 
 
 
 
 
 
 
 
 
 
 
168
 
 
APPENDIX H 

 
CONT.
 
 
Figure 82 (cont.)
 
 
 
 
 
 
 
 
 
 
 
 
169
 
 
APPENDIX H 

 
CONT. 
 
 
Figure 82 (cont.)
 
 
 
 
 
 
 
 
 
 
 
 
170
 
 
APPENDIX H 

 
CONT. 
 
 
Figure 82 (cont.)
 
 
 
 
 
 
 
 
 
 
 
 
171
 
 
APPENDIX H 

 
CONT. 
 
 
 
Figure 83: 
W
-
962 transmission 
electron microscopy 
EDS data
: TEM
-
EDS data for illite lath at 
1425 ft. (depth). 
 
 
 
 
 
172
 
 
 
 
 
 
 
 
 
 
 
 
 
 
REFERENCES
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
173
 
 
REFERENCES
 
 
Adducci
, J.G., 2015, Geological 
c
ontrols on 
s
tratigraphy and 
s
edimentation of the 
 
Mississippian 
 
Marshall Formation, Michigan Basin, U.S.A.. Unpublished M.S. thesis.
 
Western
 
 
Michigan University, Kalamazoo, Michigan. 
 
 
Ahn, J.H., and Peacor
, D.R., 1987, Kaolinitization of biotite; TEM data and implications for an 
 
alteration mechanism: 
American Mineralogist
, 
v.
 
72
, p. 353
-
356. 
 
 

s
tability of endmember illite: I. Solution 
 
equilibration experiments a
t 100
-

 
G
eoc
hi
mica et Cosmochimica
,
 
v.
 
61
, no. 20, p. 4455
-
4460.
 
 
Aja, S.U., Rosenberg, P.E., Kittrick, J.A., 1991, Illite equilibria in solutions: I. Phase 
 
relationships in the system K
2
O
-
Al
2
O
3
-
SiO
2
-
H
2
O b
etween 25 and 250°C:
 
Geoc
hi
mica 
 
et Cosmochimica
, 
v.
 
55
, no. 5, p. 1353
-
1364.
 
 
Aldega, L., Botti, F., and Corrado, S., 2007, Clay mineral assemblages and vitrinite reflectance 
 
in the Laga Basin (Central Apennines, Italy): what do they record?: 
Clays and C
lay 
 
Minerals
, 
v.
 
55
, no. 5, p. 504
-
518.
 
 
Bailey, S.W., Structures of Layer Silicates: 
Quantitative X
-
ray Mineral Analysis of Clays
, 
 
Chapter 1, p. 1
-
125. Great Britain. Mineralogical Society.
 
 
Banfield, J.F., and Eggleton
, R.A., 1988, Transmission electron microscope study of biotite 
 
weathering: 
Clays and Clay Minerals
, 
v.
 
36
, no. 1, p. 47
-
60.
 
 
Battaglia, S., Leoni, L., and Sartori, F., 2004, The Kübler index in late diagenetic to low
-
grade 
 
metamorphic pelites: A critica
l comparison of data from 10 Å and 5 Å peaks: 
Clays and 
 
Clay Minerals
,
 
v. 52
, no. 1, p. 85
-
105.
 
 
174
 
 
Bauluz, B.,
 
Peacor, D.R., and Ylagan, R.F
.
,
 
2002
,
 
Transmission 
e
lectron 
m
icroscopy 
s
tudy of 
 
s
mectite 
i
llitization 
d
uring 
h
ydrothermal 
a
lteration 
o
f 
a
 
r
hyolitic 
h
yaloclastite
 
f
rom 
 
Ponza, Italy
:
 
Clays and Clay Minerals
,
 
v. 
50
, 
n
o. 2
, p. 
157
-
173.
 
 
Beautfort, D., Cassagnabere, A., Petit, S., and Lanson, B., 1998, Kaolinite
-
to
-
dickite reaction in 
 
sandstone reservoirs: 
Clay Minerals
, v
. 33
, no. 2, p. 297
-
316
.
 
 
Berger, G., Lacharpagen, J., Velde, B., Beaufort, D., and Lanson, B., 1997, Kinetic constraints 
 
on illitization reactions and effects of organic diagenesis in sandstone/shale sequences: 
 
Applied Geochemistry
,
 
v.
 
12
, no. 1, p. 23
-
25.
 
 
Berner, R.A., 1980, 
Early 
D
iagenesis: 
A
 
T
heoretical 
A
pproach
: Princeton University Press, 
 
Princeton, New Jersey.
 
 
Bjorkum, P.A., and Gjelsvik
, N., 1988, An isochemical model for formation of authigenic 
 
kaolinite, K
-
feldspar and illite in sediments: 
Journal of Sedimentary Research
, v. 
58
, no. 
 
3, p. 506
-
511.
 
 
Blatt, H., 1979, Diagenetic 
p
rocesses in 
s
andstones, in Scholle, P.A., and Schluger, P
.R., eds., 
 
Aspects of Diagenesis
, SEPM Special Publication 26, p. 141
-
157.
 
 
Boles, J.R., and Franks, S.G., 1979, Clay diagenesis in Wilcox sandstones of Southwest Texas; 
 
I
mplications of smectite diagenesis on sandstone cementation:
 
Journal of Sedimentary 
 
Research
, v. 
49
, no. 1, p. 55
-
70.
 
 
Boothroyd, J., 2012, 
Carboniferous 
p
rovenance 
t
rends from 
c
lastic 
s
trata of the Michigan Basin. 
Unpublished 
M

 
 
Carstens, H., 1986,
 
Displacive growth of authigenic pyrite: 
Journal of Sedimentary Research
: v. 
56
, no. 2, p. 252
-
257
 
 
Carroll, Dr., 1962, Rainwater as a chemical agent in geologic processes; a review: USGS Water 
 
Supply Paper 1535 Chapter G.
 
175
 
 
Castro, M.C., Ma, L., and Hall, 
C.M., 2009, A primordial, solar He
-
Ne signature in crustal 
 
fluids of a stable continental region: 
Earth and Planetary Science Letters
, v. 
279
, no. 3
-
4, 
 
p. 174
-
184
 
 
Cercone, K.R.
,
 
1984
,
 
Thermal 
h
istory of 
the 
Michigan Basin
: 
A
merican 
Association of 
Petroleum Geologist
 
Bulletin
, v. 
68
, 
no.
 
2
, p. 
130
-
136.
 
 
Cercone, K.R.
,
 
and Pollack, H.N.
,
 
1991
,
 
Thermal 
m
aturity of the Michigan Basin, in 
Catacosinos, P.A. and Daniels, P.A., eds.,
 
Early Sedimentary Evolution of the Michigan 
Basin:
 
Geologi
cal Society of America
,
 
Special Publication
 
256
, p. 1
-
11.
 
 
Cox, K.
,
 
2015
,
 
Anomalous 
t
hermal 
i
ndicators from 
a
uthigenic 
m
inerals in Upper Paleozoic 
s
trata 
of the Michigan Basin.
 
Unpublished 
M

Kalamazoo, Michigan. 
 
 
Craw, D., Coombs, D.S., and Kawachi, Y., 1982, Interlayered biotite
-
kaolinite and other altered 
 
biotites, and their relevance to the biotite isograd in eastern Otago, New Zealand: 
 
Mineralogical Magazine
, v. 
45
, no. 337, p. 79
-
85. 
 
 
Crowley, K.D.
,
 
1991
,
 
Thermal history of Michigan Basin and southern Canadian Shield from 
apatite fission track analysis
:
 
Journal of Geophysical Research
,
 
v.
 
96
, p.
 
697

711.
 
 
Dannemiller, G.T., and Baltusis, M.A. Jr., 1989, Physical and 
c
hemical 
d
ata for 
g
round 
w
ater 
 
in
 
the Michigan Basin, 1986
-
1989. U.S. Geological Survey Open
-
File Report 90
-
368.
 
 
Dong, H., Peacor, D.R., and Murphy, S.F., 1998, TEM 
s
tudy of progressive alteration of igneous 
 
biotite to kaolinite throughout a weathered soil profile:
 
Geoch
i
mica et 
Cosmochimica
, v. 
 
62
, no. 11, p. 1881
-
1887.
 
 
Doublier, M.P., Roache, T., and Potel, S., 2010, Short
-
wavelength infrared spectroscopy: A new 
 
petrological tool in low
-
grade to very low
-
grade pelites: 
Geology
, v. 
38
, no. 11, p. 1031
-
 
1034.
 
 
176
 
 
Eberl, D.D., 1984
, Clay mineral formation and transformation in rocks and soils: 
Transactions of 
 
the Royal Society of London,
 
v. 
311
, no. 1517.
 
 
Ehrenberg, S.N., 1990, Relationship between diagenesis and reservoir quality in sandstones of 
 
the Garn formation, Haltenbanken, mid
-
Norwegian Continental shelf: 
A
merican 
 
Association of Petroleum Geologists
 
Bulletin
, v.
 
74
, no. 10, p. 1538
-
1558. 
 
 
Ehrenbe
rg, S.N., 1991, Kaolinized, potassium
-
leached zones at the contacts of the Garn 
 
Formation, Haltenbanken, mid
-
Norwegian continental shelf: 
Marine and Petroleum 
 
Geology
, v. 
8
, no. 3, p. 250
-
269.
 
 
Eslinger, E., and Pevear, D., 1988,
 
Clay Minerals for Petrol
eum Geologists and Engineers
: 
 
Society for Sedimentary Geology
, s
hort 
c
ourse 
no. 22. 
 
 
Freed, R.L., and Peacor, D.R., 1992, Diagenesis and the 
f
ormation of 
a
uthigenic
 
i
llite
-
rich I/S 
 
c
rystals in Gulf Coast 
s
hales: TEM 
s
tudy of 
cl
ay 
s
eparates:
 
Journal of Sedimentary 
 
Research
, v. 
62
, no. 2, p. 220
-
234.
 
 
Folk, R.
L., 1974, 
Petrology of Sedimentary Rocks
: Austin, Texas, Hemphill Press, second 
 
edition, p. 182
 
 
Garrels, R.M., and Mackenzie, F.T., 1971, 
Evolution of Sedimentary Rocks
. Norton, New Yo
rk.
 
 
Girard, J.P., and Barnes, D.A., 1995, Illitization and paleothermal regimes in the Middle 
 
Ordovician St. Peter Sandstone, central Michigan basin: K
-
Ar, oxygen isotope, and fluid 
 
inclusion data: 
A
merican Association of Petroleum Geologists
 
Bulletin
, v. 
79
, no. 1, 49
-
 
69.
 
 
Ging, P.B., Long, D.T., and Lee, R.W., 1996, Selected geochemical characteristics of ground 
 
water from the Marshall aquifer in the central Lower Peninsula of Michigan: USGS 
 
Water
-
Resources Investigations Report 94
-
4220
.
 
 
177
 
 
Grubb, S.M.
, Peacor, D.R., and Jiang, W.,
 
1991
, 
Transmission 
e
lectron 
m
icroscope 
 
o
bservations
 
of 
i
llite 
p
olytypism
: 
Clays and Clay Minerals
,
 
v
. 
39
, 
n
o. 5, 
p. 
540
-
550.
 
 
Gualtieri, A.F., Ferrari, S., Leoni, M., Grathoff, G., Hugo, R., Shatnawi, M., Paglia,
 
G., and 
 
Billinge, S., 2008, Structural characterization of the clay mineral illite
-
1M: 
Journal of 
 
Applied Crystallography
, v. 
41
, p. 402
-
415.
 
 
Güven
, N., 2001, Mica structure and fibrous growth of illite: 
Clays and Clay Minerals
, v. 
49
, no. 
 
3, p. 189
-
196. 
 
 
Harrell, J.A., Hatfield, C.B., and Gunn, G.R., 1991, Mississippian 
s
ystem of the Michigan Basin 

 
s
tratigraphy, sedimentology, and economic geology, 
in
 
Catacosinos, P.A., and Daniels, 
P.A., Jr., eds., 
Early 
S
edimentary 
E
volution of the Michigan Basin
: 
Geological Society of 
America, 
Special Publication
 
256, p. 203
-
219.
 
 
Hinze, W.J., Kellogg, R.L., and 
O'Hara, N.W.
,
 
1975
,
 
Geophysical studies of basement geology 
 
of
 
Southern Peninsula of Michigan
:
 
A
merican Association of Petroleum Geologists
 
Bulletin
,
 
v
. 
59
, 
no
. 9
, p.
 
1562
-
1584.
 
 
Hurst, A., and Irwin, H., 1982, Geological modelling of clay diagenesis in s
andstones: 
Clay 
 
Minerals
, v. 
17
, p. 5
-
22.
 
 

 
scales used for definition: 
Journal of Metamorphic Geology
, v. 
8
, no. 1, p. 31
-
46.
 
 
Kübler
,
 
B.
, 1969,
 
Crystallinit
y of illite. Detection of metamorphism in some frontal part of the 
 
Alps, Referate der Vortrage auf der 47 Jahrstagung des 
D
eutsch. 
M
ineral. Ges.
,
 
p. 29

 
40.
 
 
K
ü
bler, B., and Jaboyedoff
,
 
M.
,
 
2000
,
 
Illite 
c
rystallinity
:
 
Comptes Rendus de 

 
 
 
Sciences
,
 
v.
 
331
,
 
no.
 
2
, p.
 
75
-
89.
 
 
178
 
 
Land, L.S., and Dutton, S.P., 1978, Cementation of a Pennsylvanian deltaic sandstone; isotope 
 
data: 
Journal of Sedimentary Research
, v. 
48
, no. 4, p. 1167
-
1176.
 
 
Land, L.S., and Milliken, K.L., Feldspar
 
diagenesis in the Frio Formation, Brazoria County, 
 
Texas Gulf Coast: 
Geology
, v. 
9
, no. 7, p. 314
-
318. 
 
 
Lanson, B., Beaufort, D., Berger, G., Bauer, A., Cassagnabere, A., and Meunier, A., 2002, 
 
Authigenic 
k
aolin and 
i
llitic 
m
inerals 
d
uring 
b
urial 
d
iagenesis of 
s
andstones: A 
r
eview
: 
 
Clay Minerals
, v. 
37
., p. 1
-
22.
 
 
Laughrey, C.D., Ruble, T.E., Lemmens, H., Kostelnik, J., Butcher, A.R., Walker, G., and 
 
Knowles, W., 2011, Black 
s
hale 
d
iagenesis: Insights from 
i
ntegrated 
h
igh
-
d
efinition 
 
a
nalys
es of 
p
ost
-
m
ature Marcellus Formation 
r
ocks, 
n
ortheastern Pennsylvania: 
 
A
merican Association of Petroleum Geologists
 
Annual Convention.
 
 

C
retaceous Gyeongsang 
 
Basin, Korea: implic
ations for basin evolution: 
Clays and Clay Minerals
, v. 
49
, no. 1, p. 
 
36
-
43.
 
 
Luczaj, J.A., Harrison, W.B. III
,
 
and
 
Williams, N.S.
, 
2006
,
 
Fractured hydrothermal dolomite 
reservoirs in the Devonian Dundee Formation of the central Michigan basin
:
 
A
merican 
Association of Petroleum Geologists
 
Bulletin,
 
v
. 
90
, p.
 
1787
-
1801.
 
 
Ma, L, Castro, M.C., Hall, C.M, and Walter, L.M., 2005, Cross
-
formation
al flow and salinity 
 
sources inferred from a combined study of helium concentrations, isotopic ratios, and 
 
major elements in the Marshall aquifer, southern Michigan: 
Geochemistry, Geophysics, 
 
Geosystems, 
v. 
6
, no. 10.
 
 
Ma, L., Castro, M.C., and Hall, C.
M, 2009, Atmospheric noble gas signatures in deep Michigan 
 
Basin brines as indicators of a past thermal event: 
Earth and Planetary Science Letters
, v. 
 
277
, no. 1
-
2, p. 137
-
147.
 
 
Macchi
, L., 1987, A review of sandstone illite cements and aspects of their significance to 
 
hydrocarbon exploration and development: 
Geological Journal
, v. 
2
, p. 333
-
345. 
 
179
 
 
Machel, H.G., and Lonnee, J., 2002, Hydrothermal dolomite 

 
A product of poor definition 
and 
 
imagination: 
Sedimentary Geology
, v. 
152
, no. 3
-
4, p. 163
-
171.
 
 
Mählmann, R.F., and Frey, M., 2012, Standardization, calibration and correlation of the Kübler
-
 
index and the vitrinite/bituminite
 
reflectance: an inter
-
laboratory and field related study: 
 
Swiss Journal of Geosciences
, v. 
105
, no. 2, p. 153
-
170. 
 
 
Mackinnon, I.D.R., 1990, Introduction to electron
-
beam techniques, in, 
Electron
-
Optical 
 
Methods in Clay Science
. The Clay Minerals Socie
ty workshop lecture v
.
 
2
. 
 
Boulder, Colorado. The Clay Minerals Society.
 
 
Merino
, E., 1975, Diagenesis in Tertiary sandstones from Kettleman North Dome, California. I
I. 
 
Interstitial solutions:
 
distribution of aqueous species at 100°C and chemical
 
relation
 
to 
 
the diagenetic mineralogy
. 
Geochimica et Cosmochimica Acta
, 
v. 
39
, 
p. 
1629
-
1645
.
 
 
Meunier A., and Velde B.
,
 
2004
, 
Illite
: Berlin, Springer, 286
.
 
 
Monnett
, V.B., 1948, Mississippian Marshall formation of Michigan: 
A
merican Association of 
Petroleum Geologists
 
Bulletin
, v. 
32
, no. 4, pg. 629
-
688.
 
 
Moore, D.M., and Reynolds, R.
,
 
1997
, X
-
Ray 
D
iffraction and the Identification and Analysis of 
 
Clay 
Minerals
. 
New York, New York. Oxford University Press. 
 
 
Morad, S., and Aldahan, A.A., 1986, Alteration of detrital Fe
-
Ti oxides in sedimentary rocks: 
 
Geological Society of American Bulletin
, v. 
97
, no. 5, p. 567
-
578.
 
 
Morad, S., and Aldahan, A.A., 1986, 
Diagenetic alteration of detrital biotite in Proterozoic 
 
sedimentary rocks from Sweden: 
Sedimentary Geology
, v. 
47
, no. 1, p. 95
-
107.
 
 
Morad, S., 
and 

 
in sandstones: 
Sedimentary
 
Geology
, v. 
51
, p. 147
-
153. 
 
 
180
 
 
Nadeau, P.H., Wilson, M.J., McHardy, W.J., and Tait, J.M., 1984, Interstratified 
c
lays as 
 
f
undamental 
p
articles: 
Science
, v. 
225
, p. 923
-
925.
 
 
 
Nadeau, P.H., Wilson, M.J., and McHardy, W.J., 1985, The conversion of smectite to illite 
 
during diagenesis: evidence from some illitic clays from bentonites and sandstones: 
 
Mineralogical Magazine
, v. 
49
, no. 352, p. 393
-
400.
 
 
Nesbitt, H.W., and Young, G.M., 1989, Formation and 
d
iagenesis of weathering 
profiles: 
The 
 
Journal of Geology
, v. 
97
, no. 2, p. 129
-
147.
 
 
Newbury, D.E., and Ritchie, N.W.M, 2010, Is scanning electron microscopy/energy dispersive 
 
X
-
ray spectrometry (SEM/E
DS) quantitative?: 
Scanning
, v. 
35
, no. 2, p. 141
-
168.
 
 
Noguera, C., Fritz, B., and Clement, A., 2011, Simulation of the nucleation and growth of clay 
 
minerals coupled with cation exchange: 
Geoc
hi
mica et Cosmochimica
 
, v. 
75
, no. 12, p. 
 
3402
-
3418.
 
 
Nunn, J.A., and Sleep, N.H., 1984, Thermal subsidence and generation of hydrocarbons in the 
 
Michigan Basin: 
A
merican Association of Petroleum Geologists
 
Bulletin
, v. 
68
, no. 3, p. 
 
296
-
315.
 
 
Nurmi, R.D., 1977, Sedimentology and depositional environments of basin
-
center evaporites, 
 
lower Salina Group (Upper Devonian), Michigan Basin: 
A
merican Association of 
 
Petroleum Geologists
 
Special Publication
, p. 23
-
52.
 
 
Osborne, M., Haszeldine, R.S., an
d Fallick, A.E., 1994, Variation in kaolinite morphology with 
growth temperature in isotopically mixed pore
-
fluids, Brent Group, UK North Sea: 
Clay 
Minerals
, v. 
29
, no. 4, p. 591
-
608.
 
 
Prouty, C.E., 1988
, 
Trenton exploration and wrenching tectonics in the 
Michigan Basin and 
environs, in Keith,
 
B.D., ed., 
The Trenton Group (Upper Ordovician Series) of Eastern 
North America
:
 
American Association of Petroleum Geologists
 
Studies in Geology
, v. 
29
,
 
p. 
207
-
236.
 
181
 
 
Raiswell
, R., and Berner, R.A., Pyrite formation in euxinic and semi
-
euxinic sediments: 
American Journal of Science
, v. 
285
, no. 10, p. 710
-
724. 
 
 
Rebertus, R.A., Weed, S.B., 
and 
Buol, S.W., Transformations of 
b
iotite to 
k
aolinite 
d
uring 
s
aprolite
-
s
oil 
w
eathering:
 
S
oil Science Society of America Journal
, v. 
50
, no. 3, p. 810
-
819
 
 
Robertson, I.D.M, and Eggleton, R.A., 1991, Weathering of granitic muscovite to kaolinite and 
 
halloysite and of plagioclase
-
derived kaolinite to halloysite: 
Clays and Clay Minerals
, v. 
 
3
9
, no. 2, p. 113
-
126.
 
 
Robie, R.A., Hemingway, B.S., and Fisher, J.R., 1979, Thermodynamic 
p
roperties of 
m
inerals 
 
and 
r
elated 
s
ubstances at 298.15 K and 1 Bar (10
5
 
Pascals) 
p
ressure and at 
h
igher 
 
t
emperatures: USGS Survey Bulletin 1452. 
 
 
Sass, B., Rosenberg, P.E., and Kittrick, J.A., 1987, The stability of illite/smectite during 
 
diagenesis: An experimental study: 
Geoch
i
mica et Cosmochimica
, v. 
51
, p. 2103
-
2115.
 
 
Segonzac
, G.D., 1970, Transformation of clay minerals during diagenesis and low
-
grade 
 
metamorphism: A review: 
Sedimentology
, v. 
15
, no. 3, p. 281
-
346.
 
 
Singh, B., and Gilkes, R.J., 1991, Weathering of a chromian muscovite to kaolinite: 
Clays and 
 
Clay Minerals
, v
. 
39
, no. 6, p. 571
-
579.
 
 
Smith, J.V., and Yoder Jr., H.S., 1954, Experimental and theoretical studies of the mica 
 
polymorphs: 
Mineralogical Magazine
, v. 
31
, p. 209
-
235. 
 
 

, J., 1980, Precise identification of illite/smectite interstratifications by X
-
ray powder 
 
diffraction: 
Clays and Clay Minerals
, v. 
28
, no. 6, p. 401
-
411.
 
 

-
ray powder diffraction identification of illitic materials: 
Clays and Clay 
 
Min
erals,
 
v. 
32
, no. 5, p. 337
-
349.
 
182
 
 

rodon, J., and Eberl, D.
, 
1984
,
 
Micas
: Illite
:
 
Mineralogical Society of America Reviews in 
 
Mineralogy
,
 
v
. 
13
, p.
 
495
-
544. 
 
 
Stearns, M.D., 1933, The petrology of the Marshall Formation of Michigan: 
Journal of 
 
Sedimentary Research
, v.
 
3
, no. 3, p. 99
-
112.
 
 
Taylor, T.R., Giles, M.R., Hathon, L.A., Diggs, T.N., Braunsdorf, N.R., Birbiglia, G.V., 
 
Kittridge, M.G., Macaulay, C.I., 
and Espejo, I.S., 2010, Sandstone diagenesis and 
 
reservoir quality prediction: Models, myths, and reality: 
A
merican Association of 
 
Petroleum Geologists
 
Bulletin
, v. 
94
, p. 1093
-
1132.
 
 
Teichmüller M., Teichmüller, R., and Weber, K., 1979, Inkohlung und 
i
l
lit
-
k
ristallinität 

 
 
v
ergleichende 
u
ntersuchungen im Mesozoikum und Palaozoikum von Westfalen: 
 
 
D
er Geologie von Rheinland und Westfalen
, v. 
27
, p. 201
-
276. 
 
 
Veblen, D.R., and Ferry, J.M., 1983, A TEM study of the biotite
-
chlorite reaction and 
 
comparison with petrologic observations: 
American Mineralogist
, v. 
68
, p. 1160
-
1168.
 
 
Velbel, M.A., 1983, A dissolution
-
reprecipitation mechanism for the pseudomorphous 
 
replacement of plagioclase feldspar by clay minerals during weathering: 
Sciences 
 
Geologiques
,
 
M
emoires
,
 
v. 
71
, p. 139
-
147. 
 
 
Velbel, M.A., 1985, Hydrogeochemical constraints on mass balances in forested watersheds of 
 
the Southern Appalachians: 
The Chemi
stry of Weathering
.
 
(ed. J.I. Drever)
, p. 231
-
247.
 
 
NATO ASI Series C, Mathematical and Physical Sciences. 
D. Reidel Publishing 
 
Company.
 
 
 

Michigan 
 
Geography and Geology
, 
Chapter 5, p. 60
-
68.
 
New York.
 
Pearson
 
Custom Publishing. 
 
 
Vugrinovich, R.
,
 
1988
,
 
Shale 
c
ompaction in the Michigan Basin: 
e
stimates of 
f
ormer 
d
epth of 
b
urial and 
i
mplications for 
p
aleothermal 
g
radients
:
 
Bulletin of Canadian Petroleum 
Geology
.
 
v.
 
36
, p. 
1
-
8.
 
 
183
 
 
Wagenwelt, K., 2015, 
Thermal 
h
istory of the 
M
ichigan Basin: 
r
esults from 
t
hermal 
m
aturation 
d
ata and 
g
eodynamic 
m
odelling.
 

University, Kalamazoo, Michiga
n.
 
 
Wang, H.F., Crowley, K.D., and Nadon
, G.C., 1994, Thermal history of the Michigan Basin 
from apatite fission
-
track analysis and vitrinite reflectance: 
A
merican Association of 
Petroleum Geologists
 
Special Volume
, Memoir 61, Chapter 12
, p. 167
-
177. 
 
 
Warr, L.N., and Cox, S.C., 2016, Correlatin

 
indices with metamorphic mineral zones of the South Island, New Zealand: 
Applied Clay 
 
Science,
 
v. 
134
, no. 3, p. 164
-
174. 
 
 
Weber, K., 1972, Notes on determination of illite crystallinity: 
Neues Jahr
buch für Mineralogie, 
 
Monatshefte,
 
p. 267
-
276.
 
 
Westjohn, D.B., and Weaver, T.L., 1998, 
Hydrogeologic framework of the Michigan Basin 
regional aquifer system
: USGS Professional Paper 1418. 
 
 
White, S.H., Shaw, H.F., and Huggett, J.M., 1984, The use of bac
k
-
scattered electron imaging 
for the petrographic study of sandstones and shales: 
Journal of Sedimentary Research
, v. 
54
, no. 2, p. 487
-
494.
 
 
Wilkinson, M., and Haszeldine, R.S., 2002, Fibrous illite in oilfield sandstones 

 
a nucleation 
 
kinetic theory of
 
growth: 
Terra Nova
, v. 
14
, no. 1, p. 56
-
60.
 
 
Wilson, M.D., and Pittman, E.D., 1977, Authigenic 
c
lays in 
s
andstones: 
r
ecognition and 
 
i
nfluence on 
r
eservoir 
p
roperties and 
p
aleoenvironmental 
a
nalysis:
 
Journal of 
 
Sedimentary Petrology, 
v. 
47
, p. 3
-
31.
 
 
Wilson, T.P., and Long, D.T., 1993, Geochemistry and isotope chemistry of Michigan Basin 
 
brines: Devonian formations: 
Applied Geochemistry
, v.
 
8
, no. 1, p. 81
-
100.
 
 
184
 
 
Wilson, T.P., and Long, D.T., 1993, Geochemistry a
nd isotope chemistry of CaNaCl brines in 
 
Silurian strata, Michigan Basin, USA: 
Applied Geochemistry
, v. 
8
, no. 5, p. 507
-
524.
 
 
Worden, R.H., and Burley, S.D., 2003, Sandstone diagenesis: the evolution of sand to stone. In: 
 
Burley, S.D., Worden, R.H. (Eds
.), 
Sandstone Diagenesis 

 
Recent and Ancien
t, v. 
4
, p. 
 
3
-
47. 
Blackwell, Berlin.
 
 
Worden, R.H., and Morad, S., 2003, Clay minerals in sandstones: controls on formation, 
 
distribution, and evolution
:
 
International Association of Sedimentology
, Special 
 
Pub
lication
, v. 
34
, p. 3
-
41. 
 
 
Yates, D.M., and Rosenberg, P.E., 1996, Formation and stability of endmember illite: I. Solution 
 
equilibration experiments at 100
-
250° C and Pv, soln: 
Geoch
i
mica et Cosmochimica
, v. 
 
60
, no. 11, p. 1873
-
1883.
 
 
Zacharias, K.F.
,
 
1992
,
 
Clay 
m
ineral 
d
iagenesis in the 
M
arshall Sandstone, Michigan Basin.
 
 

 
 
Zacharias, K.F., Sibley, D.F., Westjohn, D.B., and Weaver, T.L., 1993, Oxygen
-
isotop
e, X
-
ray 
 
diffraction and scanning electron microscope examinations of authigenic
-
layer
-
silicate 
 
minerals from Mississippian and Pennsylvanian sandstones in the Michigan Basin: USGS 
 
Open
-
File Report 93
-
103
.