GEOLOGICA CARPATHICA
, DECEMBER 2018, 69, 6, 558–572
doi: 10.1515/geoca-2018-0033
www.geologicacarpathica.com
In-situ U–Th–Pb geochronometry with submicron-scale
resolution: low-voltage electron-beam dating of complexly
zoned polygenetic uraninite microcrystals
MICHAEL WAITZINGER
and FRITZ FINGER
Department of Chemistry and Physics of Materials, University of Salzburg, Jakob-Haringer-Strasse 2a, A-5020 Salzburg, Austria;
Michael.Waitzinger@sbg.ac.at
(Manuscript received August 2, 2018; accepted in revised form November 5, 2018)
Abstract: Complexly zoned microcrystals of uraninite were encountered in orthogneiss from the central Tauern Window
in Austria (K1 gneiss, Felbertal scheelite mine) and analysed in-situ for U, Th and Pb with state-of-the-art FE-SEM/EDX
techniques. A three times finer spatial resolution was achieved using an acceleration voltage of 8 kV, compared to
the classic 15–20 kV set-up of U–Th–total Pb electron microprobe dating. The lower voltage allows a spheroid of material
with a diameter of only 0.3 µm to be selectively analysed. Careful tests on three uraninite reference materials show that
the low-voltage method yields sufficient precision and accuracy for U–Th–total Pb uraninite dating, with errors on
individual spot ages in the order of 10–30 Ma. By means of this innovative analysis technique, small-scale age zoning
patterns could be resolved and dated in the uraninite microcrystals from the orthogneiss. Based on microstructures
observed in backscattered electron images we interpret that an older uraninite generation in the rock, with a late Permian
formation age (~260 Ma), was recycled two times through a coupled dissolution–reprecipitation process at ~210 Ma and
at ~30 Ma. The younger dissolution–reprecipitation phase at ~30 Ma coincides with the Alpine regional metamorphism
(lower amphibolite facies). The two older ages (~210 Ma and ~260 Ma) have been previously recognized in rocks from
the Tauern Window by uraninite dating, but it is the first time here that both are recorded in the same rock and even
the same uraninite grain. The present study shows that recrystallized accessory uraninite can provide a sensitive geological
“hard disk” where several discrete thermal events of an area are stored. In addition, our work attests that the mineral
uraninite has an unexpected geochronological robustness, even on the microcrystal scale.
Keywords: uraninite, high-resolution in-situ U–Th–Pb dating, K1 gneiss, Tauern Window.
Introduction
U–Th–Pb geochronometry is undoubtedly one of the most
important analytical tools that are used in the Earth Sciences.
Since its beginnings (Holmes 1911), the field has continuously
improved with regard to better precision, more sophisticated
analytical techniques, lesser sample material, and in-situ
dating with high spatial resolution (e.g., Krogh 1982; Kober
1986; Williams 1998; Davis et al. 2003; Foster et al. 2004;
Mattinson 2005; Amelin & Davis 2006; Frei & Gerdes 2009;
Schoene et al. 2010; Schaltegger et al. 2015). Apart from
zircon and monazite, which are the dominant targets for
U–Th–Pb dating at present, minerals like titanite, rutile, xeno-
time, columbite, uraninite, thorite, and several others have
been used.
The common method for U–Th–Pb age dating is based
on isotopic measurements with mass spectrometers (TIMS,
SIMS, SHRIMP, Laser-ICP-MS). An alternative is the deter-
mination of total Pb contents in Th- and U-rich minerals using
electron beam methods (Parslow et al. 1985; Bowles 1990;
Suzuki et al. 1991; Montel et al. 1996, 2017; Williams &
Jercinovic 2002). The latter method has the advantage of a high
spatial analytical resolution (typically 1–2 µm in the tradi-
tional setups), thus enabling the targeting of small crystals as
well as the detailed study of age-zoned crystals. Furthermore,
chemical data is collected simultaneously with the age infor-
mation, which is helpful for many petrogenetic issues. Electron
beam analyses can be performed on minerals in thin section,
and, thus, the age dates can be directly related to mineral tex-
tures. Disadvantages of U–Th–total-Pb electron-beam dating
include its poorer precision and accuracy in the measured
ages, and poorer control on determining the possible effects of
Pb loss or common Pb presence. Electron beam dating is cur-
rently mainly applied to monazite, although successful studies
with other U–Th-bearing minerals have been made as well, for
instance with zircon (Geisler & Schleicher 2000), xenotime
(Bernhard et al. 1998), zirconolite (Tropper et al. 2007),
thoria
nite, thorite, huttonite and uraninite (Förster 1999;
Santosh et al. 2003; Cocherie & Legendre 2007; Naemura et
al. 2009; Yokoyama et al. 2010; Cross et al. 2011; Votyakov et
al. 2013; Allaz et al. 2015).
This work is focused on two examples of complexly zoned
microcrystals of uraninite that were encountered in an ortho-
gneiss from the Tauern Window in the Eastern Alps. In a pre-
vious study (Finger et al. 2017), a significant intra-crystal
variation of U/Pb ratios was observed in these zoned grains,
pointing to a polygenetic origin. Notably, other uraninite micro-
crystals in this gneiss showed a homogeneous intra-crystal
559
IN-SITU U–Th–Pb GEOCHRONOMETRY WITH SUBMICRON-SCALE RESOLUTION
GEOLOGICA CARPATHICA
, 2018, 69, 6, 558–572
U/Pb distribution with an either Permian or Paleogene total
Pb age. The origin and geological significance of the two
age-zoned grains thus constituted an interesting problem for
further research.
The investigation of small-scale compositional heteroge-
neities in micron-sized uraninite grains requires a particularly
high spatial analytical resolution. Using FE-SEM/EDX tech-
niques and a low-voltage (8 kV) electron beam, crystal volu-
mes with an approximately 0.3 µm diameter can be selectively
analysed. This low-voltage method is applied here for the first
time to the U–Th–Pb dating of minerals. Therefore, a signifi-
cant part of this paper will be devoted to methodology.
Methodology
General remarks
Using the standard electron microprobe setup with a 15–20 kV
acceleration voltage, electron beam dating excites a spheroid-
shaped volume with 1–3 µm diameter in the targeted minerals
(Jercinovic et al. 2012). The exact size and geometry of
the exci tation volume depends on mineral density and compo-
sition. This 1–3 µm resolution is too coarse for the study of
micron-sized crystals. A significantly better spatial resolution
of electron beam dating can be achieved by lowering the acce-
leration voltage. For instance, at an acceleration voltage of
8 kV, and a beam width of 100 nm, the excitation volume in
uraninite should theoretically be only ~ 0.3 µm in diameter,
compared to ~ 0.8 µm at 15 kV (Fig. 1). Low-voltage (i.e.,
8 kV) electron beam measurements on uraninite thus permit
U–Th–total Pb dating at a three times finer spatial resolution,
but the intensity of the generated X-ray signal that is needed
for chemical analysis is much weaker in such conditions.
Fortunately, special large-area EDX detectors have become
available over the past ten years that are more sensitive and
with much better counting statistics than previous generation
EDX detectors. In addition, new field emission (FE) cathodes
are a great advantage for high-resolution analysis, as they can
produce a small coherent beam with a high current density.
Analytical setup
In this study, uraninite analyses were carried out on polished
and carbon-coated rock thin-sections using a Zeiss ULTRA-
PLUS scanning electron microscope (SEM) equipped with
a field emission (FE) cathode. Uniform electron beam condi-
tions were used for quantitative analysis involving an acce le-
rating voltage of 8 kV, 2 nA beam current and 100 nm
beam width. EDX analysis was performed with the large-area
(50 mm
2
) silicon drift detector X-MAX 50 from Oxford
Instruments. A full spectrum range from 0 to 10 keV was
recorded with 2048 channels and a resolution of 5 eV per chan-
nel. Concentrations of U, Th and Pb were derived from count
rates obtained on the Mα lines using Oxford INCAEnergy
software. Background, matrix and fluorescence effects, as well
as line overlaps, are automatically computed and corrected by
this software. Quantification of concentrations is based on
intensity values taken from the Oxford INCAEnergy internal
standard data bank (Table 1) and a 180 s calibration measure-
ment on a Si wafer before every analytical session.
For quantitative uraninite measurements, the following
element list was used: Al
2
O
3
, SiO
2
, P
2
O
5
, SO
3
, CaO, TiO
2
,
FeO, Y
2
O
3
, La
2
O
3
, Ce
2
O
3
, PbO, ThO
2
, and UO
2
. Some of
these elements are uncommon in uraninite but were included
to identify contamination from inclusions or boundary
minerals. Oxygen was generally calculated using the stoi-
chiometry based on the cations and not from the measured
oxygen signal, as the latter is difficult to calibrate accurately
and is over sen sitive to surface contamination and oxidization.
However, the oxygen (and carbon) peak intensities were
routinely checked for anomalies in order to recognize any
local altera tion involving oxidization, hydration or carbo-
nation. Fur thermore, every spectrum was checked for the pre-
sence of any additional peak not included in the quantitative
analysis.
The challenge for U–Th–total Pb dating is to analyse these
three elements with such precision that permits sufficiently
precise age dates to be calculated. When analysing uraninite
under the aforementioned beam conditions, the Oxford
X-MAX 50 acquires a total count rate of ~2*10
6
counts within
Fig. 1. Spatial resolution of electron beam dating of uraninite illustrated by Monte Carlo simulations of electron scattering at 8 kV (A) and
15 kV (B). Assumed beam width is 100 nm. Red: backscattered electrons. Used software: CASINO, Drouin et al. (2007).
560
WAITZINGER and FINGER
GEOLOGICA CARPATHICA
, 2018, 69, 6, 558–572
c. 3 minutes (120 s live time). This results in an analytical
precision of approximately ± 0.4 wt. % (1σ) for U, while Pb
can be analysed at a detection limit of ~ 0.2 wt. %. Thorium and
other minor elements, like Si, Ca, Ce, and Y, that are occasio-
nally reported in uraninite (Allaz et al. 2015; Alexandre et al.
2016), can be analysed at detection limits of 0.1–1.3 wt. %
(Table 1). We are aware that these detection limits are far
beyond the sensitivity of WDX microprobe analysis, but
believe that this creates no essential problem for uraninite
dating. For instance, a 90 Ma old uraninite crystal (see Table 2)
contains ~1 wt. % radiogenic Pb, which is well within the mea-
surable range. Using an 8 kV/2 nA beam and a counting time
of 3 minutes, the analytical uncertainty for Pb in uraninite is
typically ~ 0.1 wt. %, which corresponds to an error in the order
of 20 Ma (1 σ) for a single spot age. Even when the counting
time per spot is reduced to one minute (30 s live time), reaso-
nable age errors in the order of 20–30 Ma can still be obtained.
Increasing the counting time per analysis spot to five minutes
or more improves detection limits and precision and could be
favourable for very young, Pb-poor uraninite. However, we
must be aware that this increases the risk for beam drift impe-
ding the desired spatial resolution, as well as the risk for beam
damage effects.
A sufficient analytical precision is a prerequisite for U–Th–
total Pb dating. However, the successful application of any
ana lytical method is also dependent on the accuracy of
the element analysis, in particular the determination of Pb.
As shown by Pyle et al. (2002) and Jercinovic et al. (2005,
2012) for electron-microprobe-based U–Th–total Pb monazite
dating, there are many possible pitfalls that can superimpose
a systematic analytical error on the Pb determination; the most
important of these refer to background and interference cor-
rection problems.
For uraninite dating, these analytical perils are less severe
because Pb contents are relatively high in uraninite; for exam-
ple, they are commonly an order of magnitude higher than in
monazite and generally reach a major element concentration.
Nevertheless, we have carried out several of the recommended
tests (Jercinovic et al. 2012) in order to assess the reliability of
our data (see next sections).
Testing the calibration with synthetic uraninite and other
mineral standards
A synthetic UO
2
crystal with 99.9 % purity was measured to
control and validate the so-called “standardless” calibration
provided by the Oxford INCAEnergy software. The obtained
UO
2
concentrations were between 98 and 102 wt. % (mean
99.6 ± 0.8 wt. %, n = 8), confirming that UO
2
concentrations can
be sufficiently accurately determined at 8 kV conditions with
the internal INCAEnergy calibration. For a synthetic ThO
2
crystal, we obtained a value of 101.10 ± 0.8 wt. % ThO
2
,
which is a little too high, but the effects of this ~1 wt. %
standar di zation inaccuracy on the age dating results are negli-
gible. Measurements on various Pb minerals also reproduced
the recom mended values within 3 % deviation. Even for
dif ficult-to-analyse Pb minerals such as galena (line overlap
of S Kα and Pb Mα) acceptable values were obtained (e.g.,
87.3 wt. % Pb vs. expected 86.6 wt. % for stoichiometric
galena).
Repeated measurements on the nominally Pb-free UO
2
showed that
the computation of the Pb by INCAEnergy
resulted in a small but systematic blank value of − 0.16 ±
0.06 wt. % PbO. This blank value was externally corrected for
all uraninite analyses, weighted by the individually measured
UO
2
. The vali
dity of this procedure was independently
assessed by measu rements on uraninite reference materials
(see below). The un certainty of the blank value determination
(± 0.06 wt. %) was incorporated in the total Pb error and
the age error.
Appreciable inaccuracies in the automatic INCAEnergy
background and interference correction for uraninite were also
encountered in the case of Th. A blank test on pure UO
2
gave
a slightly negative ThO
2
content of − 0.83 ± 0.37 wt. % (see
Table 1), probably because the peak interference of U Mα on
Th Mα was slightly overcorrected. Consequently, an external
Th correction was applied for every uraninite analysis,
weighted according to the individual UO
2
content. This empi-
rical Th correction is certainly not ideal, but the problem
with accurately determining the Th will generally have only
a minor influence on the final uraninite age calculations. For
instance, 1 wt. % more or less Th alters the calculated age by
less than 1 Ma. Nevertheless, it could be argued that the impre-
cise Th data may perhaps have some effect on the peak
interference correction for Pb, as there is a small Th line
close to the Pb Mα position. Indeed, Th and Y interference
effects create a well-known analytical problem for electron-
beam monazite dating (Scherrer et al. 2000). However, for
uraninite dating, this will, in general, be negligible, owing to
Element
Line
Calibration
standard
Detection
limit 120s*
Typical
error 120s*
Blank
correction
Al
K_SERIES
Al
2
O
3
0.06
0.02
Si
K_SERIES
SiO
2
0.15
0.05
P
K_SERIES
GaP
0.09
0.03
S
K_SERIES
FeS
2
0.06
0.02
Ca
K_SERIES Wollastonite
0.18
0.06
Ti
K_SERIES
Ti
0.36
0.12
Fe
L_SERIES
Fe
0.99
0.33
Y
L_SERIES
Y
0.36
0.12
Zr
L_SERIES
Zr
0.33
0.11
La
L_SERIES
LaB
6
1.29
0.43
Ce
L_SERIES
CeO
2
1.26
0.42
Pb
M_SERIES
PbF
2
0.21
0.12
− 0.16 ± 0.06
Th
M_SERIES
ThO
2
0.75
0.25
− 0.83 ± 0.37
U
M_SERIES
U
0.40
Table 1: SEM/EDX set-up used in this study for low-voltage (8 kV)
uraninite analysis. In columns 4–5, detection limits and typical errors
(1 σ) are listed for selected elements, at 3 min counting time (120 s
live time). The given detection limits and errors have been calculated
by the INCAEnergy software based on measurements on synthetic
uraninite and the Mitterberg uraninite standard, respectively. Blank
values refer to pure UO
2
.
561
IN-SITU U–Th–Pb GEOCHRONOMETRY WITH SUBMICRON-SCALE RESOLUTION
GEOLOGICA CARPATHICA
, 2018, 69, 6, 558–572
the commonly high radiogenic Pb concentrations. The excep-
tion would perhaps be when analysing young, Pb-poor ura-
ninite or very Th- and Y-rich uraninite.
Measurements on uraninite reference materials
Three in-house uraninite reference materials were measured
for control and to serve as sensitive secondary Pb standards for
comparison with the obtained U–Th–total Pb ages. The first
reference material is magmatic uraninite from the Kirchberg
granite, Erzgebirge, Germany, previously dated at 326 ± 4 Ma
(Förster 1999) and 330 ± 5 Ma (Kempe 2003) by EPMA.
The second is hydrothermal uraninite from Mitterberg,
Austria, dated at 90 ± 5 Ma by a concordant U
238
–Pb
206
TIMS age (Paar & Köppel 1978). The third reference material
is magmatic uraninite from the Königshain granite, Erzgebirge,
dated by EPMA at 328.6 ± 1.9 Ma (Förster et al. 2012).
Our low-voltage measurements reproduce the published
geochronological dates in all three cases surprisingly well
(Fig. 2), showing that the Pb analyses are sufficiently accurate.
Individual spot ages for the Mitterberg uraninite ranged
between 85 ± 11 Ma and 102 ± 11 Ma, with a calculated
mean age of 90.7 ± 6.2 Ma (95 % confidence level, n =12).
For the Kirchberg uraninite, individual ages ranged between
320 ± 12 and 335 ± 12 Ma, with a mean of 327.3 ± 6.8 Ma (95 %
confidence level, n =12). For the Königshain uraninite, we
obtained an age range of 320–331 Ma and a mean age of
325.9 ± 8.7 Ma (95 % confidence level, n = 8). The low analy-
tical totals of the Königshain uraninite (Table 2) result from
additional REE amounts, which can be qualitatively recog-
nized in the EDX spectra (Fig 3), but were
unquantifiable within the used analytical
setup. According to electron microprobe
data (Förster 1999; Förster et al. 2012),
the Königshain uraninite has a total REE
content of 4–8 wt. %.
The reproduction of the published age
for the Königshain reference material
shows that our analytical setup also pro-
vides sufficiently accurate Pb values for
Y-, REE-, and Th-rich uraninite, implying
that the Pb background and interference
corrections of the INCAEnergy software
are good. The latter is independently sup-
ported by blank tests made for Pb-free syn-
thethic ThO
2
and YPO
4
, which yield a zero
concentration result for Pb. In addition,
Pb blank testing was done on Pb-free
quartz and pyrite, and the measured Pb
values were also close to zero.
Finally, a test for beam damage and
charging effects was made using uraninite
from Mitterberg. Repeated 30 s measure-
ments on the same standard point yielded
constant U count rates over a five-minute
observation time, with a standard devia-
tion not greater than the counting statistics
of the instrument. The specimen current
also remained constant within these five
minutes of exposure time, suggesting that
beam damage and charging effects play no
or only a minor role at the given beam
energy dose.
Testing the spatial resolution of low-
voltage electron beam dating
The Mitterberg uraninite carries several
small inclusions and veins of gold (Fig. 4).
We recorded detailed chemical profiles
across the uraninite-gold grain boundaries
Sample
SiO
2
CaO
Y
2
O
3
PbO
ThO
2
UO
2
Total
Age
Error
Kirchberg
0.22
<0.11
<0.38
4.32
1.39
93.74
99.67
335
12
Kirchberg
<0.15
<0.11
<0.38
4.29
1.26
94.61
100.17
330
12
Kirchberg
0.21
<0.11
<0.38
4.21
1.21
94.43
100.05
325
12
Kirchberg
<0.15
<0.11
<0.38
4.28
1.42
94.32
100.01
330
12
Kirchberg
0.26
<0.11
<0.35
4.27
1.10
94.19
99.81
330
12
Kirchberg
0.33
<0.11
<0.38
4.11
1.02
92.20
97.66
325
12
Kirchberg
0.40
<0.11
<0.38
4.06
1.01
92.46
97.92
320
12
Kirchberg
0.42
<0.11
<0.35
4.24
1.05
94.54
100.25
327
12
Kirchberg
0.32
<0.11
<0.35
4.23
1.10
93.96
99.60
328
12
Kirchberg
0.37
<0.11
<0.35
4.25
0.95
93.95
99.51
329
12
Kirchberg
0.35
<0.11
<0.35
4.14
1.11
94.05
99.66
321
12
Kirchberg
0.31
<0.11
<0.35
4.16
1.00
92.69
98.17
327
12
Average age (95 % c. l.)
327.3
6.8
Mitterberg
0.41
0.41
0.77
1.29 <0.48
94.06
97.22
102
11
Mitterberg
0.37
0.47
0.81
1.12 <0.48
94.82
97.87
89
11
Mitterberg
0.34
0.53
0.83
1.16 <0.54
97.89
101.06
89
11
Mitterberg
0.32
0.58
1.49
1.08 <0.48
95.11
98.86
85
11
Mitterberg
0.36
0.44
1.42
1.16 <0.53
96.34
100.04
90
11
Mitterberg
0.33
0.48
1.31
1.09 <0.48
95.98
99.48
85
11
Mitterberg
0.20
0.67
1.11
1.12 <0.50
96.30
99.99
87
11
Mitterberg
0.26
0.65
0.89
1.18 <0.50
94.11
97.59
93
11
Mitterberg
0.25
0.66
0.79
1.24 <0.48
96.37
99.82
96
11
Mitterberg
0.30
0.44
0.85
1.17 <0.50
95.88
98.93
91
11
Mitterberg
0.27
0.63
0.64
1.18 <0.48
97.25
100.24
91
11
Mitterberg
0.31
0.49
0.61
1.16 <0.48
97.22
100.06
89
11
Average age (95 % c. l.)
90.7
6.2
Königshain
0.57
<0.14
4.23
3.56
7.09
76.90
92.34
329
13
Königshain
0.54
<0.14
4.31
3.56
6.58
76.51
91.49
331
13
Königshain
0.64
<0.14
4.66
3.54
6.90
77.25
92.99
326
13
Königshain
0.52
<0.14
4.36
3.48
6.57
75.99
90.91
326
13
Königshain
0.69
<0.12
3.48
3.58
7.49
79.47
94.72
320
12
Königshain
0.67
<0.12
3.77
3.61
6.81
79.49
94.35
324
12
Königshain
0.56
<0.12
3.20
3.63
7.61
80.08
95.08
322
12
Königshain
0.73
<0.12
3.55
3.60
7.60
77.42
92.90
330
13
Average age (95 % c. l.)
325.9
8.7
Table 2: Chemical data (wt. %) for the in-house uraninite reference materials Kirchberg,
Mitterberg and Königshain. Single point ages and errors (1 σ) were calculated based on
the measured U, Th and Pb concentrations, using the algorithm of Săbău (2012). Counting
time per analysis is c. 5 minutes (180 s live time). The age errors always include the analy-
tical uncertainties in the Pb and U determination as well as the error for the performed Pb
blank correction. Mean ages (95 % confidence level) were calculated using the program
ISOPLOT of Ludwig (2003).
562
WAITZINGER and FINGER
GEOLOGICA CARPATHICA
, 2018, 69, 6, 558–572
at 8, 10, and 15 kV, using a step width of 0.1 µm, to empiri-
cally determine the lateral analytical resolution and the size of
the excitation volume, respectively. Results are shown in
Figure 5. At 15 kV, we note a ~1 µm wide transition zone,
where the electron beam has excited both minerals, simulta-
neously producing signals for Au and U. The width of this
transition zone is reduced from ~1 µm to ~ 0.6 µm at 10 kV
and to ~ 0.3 µm at 8 kV, demonstrating that the spatial resolu-
tion of U–Th–Pb uraninite dating is about three times better at
8 kV than at 15 kV.
Recommended analysis strategy for uraninite dating
The suitability of monazite for electron beam U–Th–Pb
dating has been highlighted many times (no significant Pb
loss, little common Pb), but the potential of using uraninite as
a geochronometer is still insufficiently investigated. There is
agreement that the amount of common Pb in uraninite is
generally strongly subordinate compared to the vast amounts
of radiogenically produced Pb (e.g., Kotzer & Kyser 1993;
Fayek et al. 2002; Chipley et al. 2007). Thus, the common Pb
effect on a total Pb uraninite age date will be negligibly low in
most cases.
However, problems with Pb loss can be serious.
Investigations made on hydrothermally overprinted U depo-
sits show that primary uraninite can lose Pb and become
variably “rejuvenated” by interaction with fluids (Fayek et al.
2002; Alexandre & Kyser 2005; Decrée et al. 2011). More-
over, U gain can also occur when uraninite undergoes hydro-
thermal alteration (Kempe 2003). The key point here is to find
Fig. 2. Geochronological dates obtained for three uraninite reference materials using a low voltage electron beam (8 kV). Age data are sorted
from oldest to youngest left to right. Recording time is c. 5 minutes (180 s live time) per analysis point. Mean ages, errors and statistical
parameters were calculated with the program ISOPLOT (Ludwig 2003).
Fig. 3. Examples of EDX spectra (8 kV, 2 nA) for almost pure uraninite with ~1 wt. % PbO (Mitterberg), and Th, Y and REE bearing uraninite
with ~ 3.5 % PbO (Königshain).
563
IN-SITU U–Th–Pb GEOCHRONOMETRY WITH SUBMICRON-SCALE RESOLUTION
GEOLOGICA CARPATHICA
, 2018, 69, 6, 558–572
and develop an effective analytical strategy that enables us to
determine whether the Th–U–Pb system of a given crystal (or
crystal domain) is chemically disturbed or not (thermodyna-
mically a closed or open system). Measuring the U–Th–Pb
concentrations within single crystals at a high spatial reso-
lution, for example, along a profile, could provide such
a strategy. A constancy of ages would strongly suggest that
the U–Th–Pb system is undisturbed, whereas a disturbed sys-
tem will exhibit irregular U–Th–Pb distributions and inconsis-
tent ages (Kempe 2003). If several adjacent spot ages along
a profile are iden tical within the analytical error, they can be
statistically ave raged to produce a domain age. If all spot ana-
lyses in a grain are identical within error, they can be averaged
to give a single grain age.
The case study: complexly zoned uraninite
microcrystals in orthogneiss
Geological background and sample petrography
The Tauern Window exposes the deepest tectonic units
of the Eastern Alps (Fig. 6). It contains crustal rocks of
the European plate that were overridden by various nappes
derived from the southern Adriatic plate during the Alpine
orogeny (see Schmid et al. 2004, 2013 for review). As a con-
sequence of this tectonic burial, the Tauern Window experien-
ced upper greenschist to middle amphibolite facies regional
metamorphism in the late Paleogene, at c. 30 Ma. This phase
of regional metamorphism has been well constrained by
various geochronological studies, including K–Ar and Ar/Ar
mica and hornblende dating, Sm–Nd garnet dating (data com-
pilation in Pestal et al. 2009) and U–Pb allanite dating (Cliff et
al. 2015).
Approximately half of the Tauern Window complex com-
prises the so-called Penninic Units (Fig. 6) of Mesozoic
sedimentary and volcanic rocks, including obducted Jurassic–
Cretaceous ophiolites (Frasl & Frank 1966). An older, pre-
Mesozoic basement (Subpenninic Units in Fig. 6) of Early
Paleozoic, island arc-type crust was variably metamorphosed
during the Variscan period (Habach Complex and Altkristallin
in Fig. 6), and intruded by large volumes of Variscan grani-
toids, now the so-called Central Gneisses. The pre-Alpine
geo chronology of the Tauern Window is based mainly on
Carboniferous to Permian zircon dates from the Variscan
grani toids (Eichhorn et al. 2000; Vesela et al. 2012) and
Cambro–Ordovician to Devonian zircon dates from the older
arc-type crust (Eichhorn et al. 1995; Kebede et al. 2005).
Carboniferous (Variscan) metamorphism is recorded through
garnet and monazite ages (Von Quadt 1992; Finger et al.
2016).
Microcrystals of accessory uraninite have recently been
found in several of the Central Gneisses of the Tauern Window.
They have different formation ages (Finger et al. 2017).
The youngest generation of uraninite microcrystals formed
during the Alpine orogeny in the Paleogene at ~ 30 Ma.
A Permian (~ 265 Ma) generation of uraninite microcrystals
has been identified in the Central Gneiss types with a lower
Carboniferous intrusion age (K1 gneiss and Felbertauern
Augengneiss). Triassic (~ 215 Ma) uraninite microcrystals
were found in the Central Gneisses with Permian intrusion
ages (Granatspitz and Reichenspitz gneiss). The pre-regional
metamorphic uraninite microcrystals are interpreted as dating
Fig. 4. Intergrowths of uraninite (grey) and gold (bright) in
the Mitterberg sample (backscattered electron image).
Fig. 5. Profiles across a uraninite-gold boundary (uraninite left, gold right) with a stepwidth of 0.1 µm, showing the spatial resolution of
electron beam excitation at 15, 10 and 8 kV acceleration voltage (see text for further explanation). The gold occurs in solid solution with
~ 10 % silver.
564
WAITZINGER and FINGER
GEOLOGICA CARPATHICA
, 2018, 69, 6, 558–572
discrete low-T events related to increased heat and fluid
activity during the Permo–Triassic thinning of the European
plate (Finger et al. 2017).
The K1 Gneiss, the host rock of the two complexly zoned
uraninite crystals of this study, is a small orthogneiss body
in the realm of the prominent scheelite mine Felbertal
(47°13’27.1” N, 12°29’18.2” E; Fig. 6). It is a fine grained,
SiO
2
rich, metagranitic rock (Kozlik & Raith 2016) with
high concentrations of W (up to 3000 ppm), Nb (up to
100 ppm), U (up to 50 ppm) and has been regarded as
the source of the local tungsten mineralization. The quartz–
feldspar–mica fabric of the gneiss is entirely metamorphic
(Alpidic). The magmatic protolith age of the rock has been
constrained by zircon dating at 339.6 ± 1.2 Ma (Kozlik et al.
2016).
Microstructures, EDX analyses and U–Th–total Pb ages
On average, 10–20 uraninite microcrystals are observed in
a single thin section from the K1-gneiss. They are included in
feldspars, epidote, titanite or zircon, only a few are interstitial.
As mentioned earlier, most uraninite grains in the rock are
unzoned, with a homogeneous intra-crystal U/Pb distribution
that corresponds either to a Paleogene (~ 30 Ma) or a Permian
(~ 265 Ma) total Pb age (Finger et al. 2017). Compositionally
heterogeneous grains with internally variable U/Pb ratios and
total Pb ages, as described in the following, are thus rather
the exception than the rule.
Grain 1
This uraninite grain, although only some 6 µm wide,
presents a striking zonation in the BSE image with a darker
core and a brighter rim (Fig. 7). The boundary between the core
and the rim zone is sharp, but irregularly and multiply embayed
indicating replacement of older uraninite core substance by
younger rim substance via coupled dissolution–reprecipitation
(Putnis 2002; Harlov et al. 2011). The grain is enclosed in
titanite. Other uraninite micrograins occurring nearby (Fig. 7a)
are unzoned and entirely Permian in age. A small elongated
fluorite crystal, recognizable by an extremely bright CL signal,
adheres to grain 1 (Fig. 7c).
In order to obtain some three dimensional information,
the grain surface was reground in a second step of investiga-
tion with a layer of ~ 0.3 µm being removed. The bright rim
zone became wider (Fig. 7c), implying that larger volumes of
the bright substance reside at the bottom of the grain. A nar-
row, ~ 0.2 µm wide dark chip from the older uraninite conti-
nues into the bright rim zone, documenting that coupled
dissolution–reprecipitation can produce very fine-scale intra-
crystal heterogeneities. Such fine intergrowths are difficult to
resolve and to analyse even with FE-SEM methods and natu-
rally there is always a danger to obtain mix analyses and to
calculate mix ages here.
A chemical traverse with ~ 0.3 µm resolution (Figs. 7b, 8a;
Table 3) shows that the bright rim of grain 1 has a much lower
Pb content than the core (< 1 vs. ~ 3 wt. %), a lower Y
2
O
3
con-
tent (0–2 vs. 3–4 wt %), and a higher UO
2
content (~ 90 vs.
~ 80–85 wt. %). Notably, there is no significant change in
the Th content between core and rim, but the profile reveals
a Th-enriched inner core within the core zone with
~10–13 wt. % ThO
2
. This Th rich inner core is not very visible
in the BSE image but it could be approximatively delineated
by x-ray mapping. Four shorter chemical profiles were addi-
tionally recorded in different domains of grain 1 after it was
reground (Figs. 7c, 8b–e, Table 3). These profiles confirm
the crystallo-chemical heterogeneities within grain 1 as
described before.
Fig. 6. Tectonic overview map of the Eastern Alps after Schmid et al. (2013) and geological map of the Tauern Window after Pestal et al.
(2009).
565
IN-SITU U–Th–Pb GEOCHRONOMETRY WITH SUBMICRON-SCALE RESOLUTION
GEOLOGICA CARPATHICA
, 2018, 69, 6, 558–572
Total Pb ages (Fig. 8, Table 3) imply that the bright uraninite
rim formed during the regional metamorphic overprint of
the gneiss at ~ 30 Ma, although the three single point ages
measured in the rim sector of profile a (analyses a1–a3 in
Table 3) do not give fully consistent ages. This is probably
due to the presence of fine undigested remains of the precursor
uraninite, which became apparent in the BSE image after
regrinding (Fig. 7b and c). Profile c seems to contain lesser
impurities and indicates an age of 26 ± 11 Ma for the uraninite
rim.
Fig. 7. BSE images for uraninite grain 1, before (B) and after (C) regrinding (see text for further information). White arrows indicate the mea-
sured chemical profiles.
Fig. 8. Chemical profiles and total Pb ages (with 1 σ errors) recorded in grain 1. Positions of profiles are shown in Fig. 7. Given domain ages
(bold) are weighted average ages (95 % confidence level).
566
WAITZINGER and FINGER
GEOLOGICA CARPATHICA
, 2018, 69, 6, 558–572
The total Pb ages measured in the dark
core zone of grain 1 pivot around a mean
of ~ 262 Ma. This is consistent with
the Permian ages of most other uraninite
grains in the rock (Finger et al. 2017).
The total Pb age of the inner high-Th
core falls in the same range (261 ± 12 Ma).
The core thus represents a part of the Per-
mian growth history of the grain.
Grain 2
This relatively large subhedral urani-
nite grain with a size of approximately
20×40 µm is completely enclosed in epi-
dote. Along its periphery several small
fluorite crystals with bright CL signals
are identified (Fig. 9a). High resolution
BSE imaging (Fig. 9 b and c) reveals
a complex patchy zonation involving dark
grey domains, medium grey domains and
light grey domains. The light grey
domains are typically located along
the grain margin, while interior parts of
the grain are middle-grey. The contacts
between the middle grey central domains
and the light-grey marginal domains
appear rather sharp, but the BSE contrast
is not very strong. Embayment textures
suggest that the light-grey uraninite has
replaced the middle grey uraninite via
coupled dissolution–reprecipitation (Putnis
2002; Harlov et al. 2011). A thin dark
grey crystal domain with elongated shape
is seen in the left half of grain 2 (Fig. 9).
It is sharply bordered and almost com-
pletely embedded in the middle grey sub-
stance. In the upper part of the grain in
Fig 9, two puzzling small bright lenses
are observed (labelled 1–2 in Fig. 9).
Several short chemical profiles were
recorded in different parts of grain 2
in order to determine domain ages
(Table 4). Profiles b and g are positioned
in the dark grey domain. Both show
a relatively high Th content (~ 11–12 wt. % ThO
2
), while U
is relatively low (~ 80 wt. % UO
2
). The higher Th/U ratios
explain the darker BSE signal. The two measured domain ages
are 257 ± 7 Ma and 254 ± 8 Ma, respectively, indicating
that the dark grey zone has a Permian formation age, like
most other uraninite crystals in the K1 gneiss (Finger et al.
2017).
The middle-grey zone is represented by profiles c, h and i
(Figs. 9, 10). It has lower ThO
2
(~ 7 wt. %) and higher UO
2
(84–87 wt. %) content. The measured domain ages are
213 ± 8, 214 ± 7 and 200 ± 7 Ma. The middle grey part of
the crystal thus formed during the Triassic being clearly
younger than the remnant dark grey domain.
The light grey marginal zone of grain 2 (profiles d and f in
Figs. 9, 10) differs from the middle grey zone mainly by a
lower Y content (2–2.5 vs. 2.5–3 wt. % Y
2
O
3
). The measured
total-Pb ages are 199 ± 9 and 208 ± 7 Ma (Fig. 10). These age
dates do not differ from those measured in the middle grey
zone within the given analytical errors. Nevertheless, textures
do clearly suggest that the light grey uraninite is at least
slightly younger, as it has seemingly replaced the middle grey
uraninite via coupled dissolution–reprecipitation.
Grain 1
profile/point SiO
2
CaO
Y
2
O
3
PbO
ThO
2
UO
2
Total
Age
[Ma]
Error
(1 σ)
Live time
[s]
a1
0.81
<0.27
<0.63
0.71
6.35
88.69
96.56
59
27
60
a2
0.61
<0.27
1.41
0.04
7.50
91.83
101.40
3
25
60
a3
0.73
<0.27
2.00
0.90
6.52
86.87
97.01
76
27
60
a4
0.75
<0.27
3.80
2.71
7.39
81.74
96.38
238
29
60
a5
0.52
0.46
3.83
2.99
6.99
79.99
94.78
268
30
60
a6
0.52
<0.27
4.07
2.76
6.98
80.97
95.30
245
30
60
a7
0.64
<0.27
4.30
2.60
7.16
80.95
95.65
231
30
60
a8
0.67
<0.27
4.24
3.06
6.90
79.70
94.57
275
30
60
a9
0.54
<0.27
3.89
2.96
7.41
77.44
92.23
273
31
60
a10
0.66
0.48
3.87
3.20
7.85
79.06
95.13
288
30
60
a11
0.80
0.64
3.58
3.11
8.15
79.06
95.35
280
30
60
a12
0.62
<0.27
3.21
2.87
8.43
80.18
95.31
255
30
60
a13
0.61
0.46
3.05
3.11
9.44
80.76
97.42
273
30
60
a14
0.70
<0.27
3.28
2.74
10.85
79.38
96.95
244
30
60
a15
0.51
<0.27
3.22
2.94
11.31
80.17
98.15
259
30
60
a16
0.63
0.46
3.52
2.82
11.42
78.12
96.98
254
30
60
a17
0.59
<0.27
3.50
3.05
11.76
77.75
96.65
275
30
60
a18
0.66
<0.27
3.62
2.91
10.35
78.81
96.35
261
30
60
a19
0.69
<0.27
4.43
3.45
7.82
80.29
96.68
306
30
60
a20
0.68
<0.27
4.20
3.08
7.70
81.91
97.57
269
29
60
a21
0.68
0.46
4.57
3.25
6.84
82.31
98.11
283
29
60
a22
0.70
0.55
5.14
2.92
6.36
81.27
96.94
258
30
60
Mean
A4-22
261
15 (2σ)
b1
0.61
<0.18
4.49
2.84
12.79
74.60
95.33
266
22
120
b2
0.58
0.31
4.74
2.76
12.80
75.81
97.00
255
23
120
b3
0.60
0.48
4.51
2.97
12.66
74.60
95.82
278
23
120
Mean
266
26 (2σ)
c1
0.46
0.46
0.66
0.30
6.91
88.71
97.50
25
18
120
c2
0.49
0.33
0.90
0.15
6.88
87.87
96.61
12
19
120
c3
0.52
0.36
0.67
0.48
7.13
88.15
97.31
40
19
120
Mean
26
21 (2σ)
d1
0.51
0.43
4.32
2.80
7.01
80.77
95.85
249
21
120
d2
0.59
0.47
3.87
3.08
8.95
80.09
97.05
273
22
120
d3
0.55
0.49
3.68
3.07
9.70
79.64
97.12
273
22
120
d4
0.60
0.39
3.29
2.93
10.45
77.80
95.46
266
22
120
d5
0.68
0.35
3.58
2.71
10.05
78.81
96.18
244
22
120
Mean
261
19 (2σ)
e1
0.59
0.49
4.05
3.09
6.86
81.61
96.69
271
21
120
e2
0.61
0.29
4.23
2.83
7.37
81.07
96.39
250
21
120
e3
0.62
0.41
3.75
2.78
10.18
79.19
96.94
249
22
120
e4
0.59
0.33
3.12
2.72
10.28
79.08
96.13
244
22
120
Mean
254
21 (2σ)
Table 3: Chemical (wt. %) and age data for uraninite grain 1 from the Felbertal K1-gneiss
(profiles a – e; see Figs. 7, 8). Single point age errors are 1 σ. Note that a shorter counting time
(60 s detector live time) was used for profile a.
567
IN-SITU U–Th–Pb GEOCHRONOMETRY WITH SUBMICRON-SCALE RESOLUTION
GEOLOGICA CARPATHICA
, 2018, 69, 6, 558–572
In contrast to grain 1 we find no Paleogene rim zone in
grain 2. However, one of the small bright lenses (profile a in
Fig. 10) is very poor in Pb (PbO < 0.5 wt. %) and thus likely
represents a young regional metamorphic crystal domain. It is
possible that the “lens” is, in reality, the protrusion of a larger
Paleogene rim domain that is hidden underneath the crystal.
Profile a shows that there was no substantial exchange of Pb
between the lens and the surrounding older uraninite by solid
state diffusion.
The second bright lens (profile e) is, surprisingly, of dif-
ferent origin. Here, the PbO contents are in the order of
2.7 wt. % corresponding to a domain age of 215 ± 9 Ma. This
lens thus formed during the Triassic, like most other parts of
grain 2. Lens 2 has an unusually high UO
2
content of close to
92 wt. %, while ThO
2
(~3 wt. %), and Y
2
O
3
contents (<1 wt. %)
are significantly lower compared to the other parts of grain 2.
The distinctive chemical composition of lens 2 and its sharp
outlines would imply that practically no solid state diffusion
has taken place when the whole crystal underwent reheating to
~ 500 °C in the Paleogene, in connection with Alpine regional
metamorphism.
Discussion and conclusions
Uraninite recrystallization by coupled dissolution–
reprecipitation
The two complexly zoned uraninite crystals of this study
exhibit polygenetic, distinctly embayed zonation patterns,
which are indicative for recrystallization by coupled disso-
lution–reprecipitation (Putnis 2002; Harlov et al. 2011).
We have the suspicion that F-bearing fluids have caused
the dissolution of the uraninite, because both recrystallized
uraninite grains are bordered by small fluorite crystals.
The important role of F for the solubility of U in melts and
fluids has been pointed out many times (Keppler & Whyllie
1990; Peiffert et al. 1996).
We would argue that, in our case, the process of uraninite
recrystallization by coupled dissolution–reprecipitation invol-
ved only small local fluid volumes, because in the presence of
larger volumes of U-undersaturated fluids, the dissolved U
would have been most likely transported away and not repre-
cipitated at the spot.
The fact that dissolution–reprecipitation driven recrystalli-
zation of uraninite obviously can happen a couple of times
during the geological evolution of a rock, is extremely fortu-
nate for the field of geochronology. Multiply recrystallized
uraninite can be viewed as a sensitive “hard disk” where
several geological events of an area are stored. Most impor-
tantly, this “uraninite hard disc” may include information of
discrete low-T events that are commonly not recorded by other
geochronometers (Finger et al. 2017).
Geochronological robustness of uraninite
In-situ chemical dating of uraninite by means of electron
beam excitation and X-ray spectroscopy is not new (Parslow
et al. 1985; Bowles 1990). It was successfully used before
chemical U–Th–Pb dating of monazite started, but was never
widely applied. The method remained restricted to studies of
uraninite and pitchblende in U deposits (Alexandre & Kyser
2005; Cross et al. 2011; Shahin 2014; Ge et al. 2014) and to
accessory magmatic uraninite in granitic rocks (Förster 1999;
Fig. 9. CL and BSE images of grain 2. The CL image (A) reveals tiny fluorite crystals along the uraninite margin. The position of the measured
profiles and a tentative delineation of different dissolution–reprecipitation zones in grain 2 are given in B and C.
568
WAITZINGER and FINGER
GEOLOGICA CARPATHICA
, 2018, 69, 6, 558–572
Kempe 2003; Hurtado et al. 2007; Cocherie
& Legende 2007; Votyakov et al. 2013).
A more widespread application would
result from the study of uraninite micro-
crystals in metamorphic rocks. The increa-
sing use of electron microscopy in
petro graphic studies shows that uraninite
microcrystals occur in many metamorphic
rocks. The potential of uraninite dating in
metamorphic terranes will greatly depend
on the geochronological robustness of
the U–Th–Pb system in the mineral. We
present here unequivocal evidence that
even very small uraninite crystals can sur-
vive an amphibolite facies overprint, pre-
serving domains with intact U–Th–Pb
ratios. It must be taken into account,
though, that different age domains in poly-
genetic uraninite can be so small and inti-
mately intergrown with each other that
they are very difficult to resolve and to
analyse. Dating results can be problematic
in such cases. The risk of Pb loss through
solid state diffusion is low for unaltered
uraninite at low to medium temperatures.
However, fine scale uraninite alteration to
U silicate (coffinite) and other secondary
U minerals may cause Pb loss (Kempe
2003). It is therefore imperative that ura-
ninite age data is always collected in com-
bination with detailed backscatter electron
imagery and high-resolution composi-
tional profiles across single grains.
U–Th–Pb geochronometry with sub-
micron- scale resolution and the potential
of the SEM/EDX method
It is obvious that the geochronological
investigation of micron-sized zoning pat-
terns in uraninite microcrystals requires
an analytical resolution much finer than
what is presently possible with Laser-
ICP-MS methods (typically 15–30 µm
spots), and even with the latest and best
SHRIMPs (around 5–10 µm spot resolu-
tion). Among all the currently available
isotopic analytical methods, only the SIMS
technique possesses the required spatial
Grain 2
profile/point SiO
2
CaO
Y
2
O
3
PbO
ThO
2
UO
2
Total
Age
[Ma]
Error
(1 σ)
Live time
[s]
a1
0.31
0.41
2.53
2.71
10.03
81.70
97.69
236
29 60
a2
0.49
0.47
2.66
2.72
8.47
83.41
98.22
234
30 60
a3
0.46
0.40
3.20
2.64
10.72
79.98
97.40
234
31 60
a4
0.40
<0.27
2.71
0.62
11.43
81.20
96.35
55
28 60
a5
0.47
<0.27
2.58
0.38
10.12
84.60
98.15
32
27 60
a6
0.51
<0.27
2.38
0.24
9.49
83.95
96.57
21
28 60
a7
0.43
<0.27
2.29
0.36
10.09
84.59
97.76
31
27 60
a8
0.38
<0.27
2.27
0.30
10.26
85.42
98.63
26
27 60
a9
0.43
<0.27
2.20
0.45
9.94
85.71
98.74
38
27 60
a10
0.34
<0.27
2.18
0.33
10.76
84.06
97.67
29
27 60
a11
0.55
<0.27
2.75
0.15
9.40
84.90
97.74
12
27 60
a12
0.34
<0.27
2.38
0.28
9.52
84.94
97.46
24
27 60
a13
0.53
<0.27
2.61
0.61
9.68
85.16
98.58
52
27 60
a14
0.45
<0.27
3.19
1.17
11.43
80.88
97.12
104
29 60
a15
0.48
0.48
3.25
2.57
11.17
79.39
97.33
229
31 60
a16
0.48
0.49
3.32
2.73
11.39
78.01
96.42
247
32 60
a17
0.42
0.50
3.36
3.01
11.60
78.50
97.40
269
32 60
b1
0.48
0.53
3.07
2.94
10.95
80.72
98.69
257
21
120
b2
0.47
0.51
2.97
2.95
11.36
80.26
98.51
259
22
120
b3
0.60
0.51
3.00
3.00
11.30
79.53
97.94
266
22
120
b4
0.76
0.53
3.33
2.92
11.31
78.13
96.98
263
23
120
b5
0.84
0.38
3.29
2.94
11.57
78.63
97.65
263
22
120
b6
0.86
0.32
3.44
2.74
11.48
78.64
97.48
246
22
120
b7
0.67
0.49
3.12
2.89
11.48
79.33
97.98
257
22
120
b8
0.77
0.53
3.38
2.82
11.56
78.04
97.11
255
23
120
b9
0.75
0.63
3.07
2.75
11.55
79.54
98.29
244
22
120
b10
0.75
0.57
3.43
2.89
11.76
78.16
97.57
260
23
120
Mean
257 14 (2σ)
c1
0.52
<0.18
3.18
2.72
8.40
84.39
99.21
231
21
120
c2
0.73
<0.18
3.20
2.44
8.27
82.91
97.56
211
21
120
c3
0.75
<0.18
3.37
2.36
7.87
83.41
97.76
204
21
120
c4
0.73
<0.18
2.37
2.65
6.86
85.76
98.36
222
20
120
c5
0.71
<0.18
2.32
2.50
6.50
86.33
98.37
210
20
120
c6
0.70
<0.18
2.29
2.41
6.37
86.62
98.39
201
20
120
Mean
213 16 (2σ)
d1
0.43
<0.18
2.32
2.30
8.47
85.05
98.57
194
19
120
d2
0.48
0.24
2.07
2.39
8.62
83.70
97.51
205
20
120
d3
0.47
<0.18
2.21
2.50
8.19
83.31
96.70
216
20
120
d4
0.55
<0.18
2.44
2.19
8.31
82.90
96.39
190
20
120
d5
0.44
0.25
2.32
2.22
8.14
84.44
97.80
189
20
120
Mean
199 18 (2σ)
e1
0.34
<0.18
<0.39
2.63
3.10
92.62
98.69
208
18
120
e2
0.36
<0.18
0.54
2.71
3.49
92.57
99.66
214
18
120
e3
0.41
0.34
0.59
2.81
2.85
91.78
98.77
224
19
120
Mean
215 21 (2σ)
f1
0.44
<0.18
2.32
2.34
7.22
85.46
97.79
198
19
120
f2
0.43
<0.18
2.18
2.61
7.47
84.75
97.44
221
20
120
f3
0.45
<0.18
2.46
2.72
7.53
85.27
98.42
229
20
120
f4
0.39
0.26
2.14
2.39
7.44
85.09
97.71
202
20
120
f5
0.42
<0.18
2.26
2.52
7.17
84.56
96.93
215
20
120
f6
0.41
0.30
2.13
2.29
7.74
85.18
98.05
194
20
120
f7
0.45
0.25
2.39
2.34
7.19
84.81
97.43
199
20
120
f8
0.49
<0.18
2.46
2.52
7.19
84.71
97.37
214
20
120
f9
0.40
<0.18
2.16
2.35
7.62
85.57
98.10
198
20
120
Mean
208 13 (2σ)
g1
0.47
0.52
3.20
2.93
11.76
79.85
98.73
258
21
120
g2
0.44
0.64
3.16
2.72
11.73
78.29
96.99
245
22
120
g3
0.56
0.50
3.25
2.67
11.62
77.42
96.01
243
22
120
g4
0.42
0.49
2.99
2.91
12.01
78.11
96.93
261
22
120
g5
0.51
0.56
3.22
3.12
11.72
77.56
96.69
282
22
120
g6
0.55
0.52
3.07
2.70
11.79
77.99
96.61
244
22
120
g7
0.53
0.55
3.11
2.70
11.95
78.65
97.49
242
23
120
g8
0.56
0.35
3.19
2.81
11.71
78.16
96.79
253
22
120
Mean
254 16 (2σ)
h1
0.44
0.55
2.62
2.56
7.12
84.07
97.37
219
20
120
h2
0.38
0.63
2.59
2.70
7.32
84.20
97.80
230
21
120
h3
0.46
0.43
2.77
2.56
7.04
85.67
98.92
216
20
120
h4
0.47
0.54
2.64
2.75
7.12
84.41
97.92
235
21
120
h5
0.46
0.66
2.55
2.44
7.47
84.63
98.21
208
21
120
h6
0.50
0.52
2.78
2.38
7.00
84.38
97.56
204
21
120
h7
0.50
0.44
2.57
2.44
7.11
84.59
97.65
208
21
120
h8
0.54
0.79
2.81
2.30
7.22
85.26
98.91
194
21
120
h9
0.44
0.55
2.62
2.56
7.12
84.07
97.37
219
20
120
Mean
214 14 (2σ)
i1
0.50
0.49
2.73
2.40
6.92
85.38
98.41
203
20
120
i2
0.41
0.57
2.67
2.24
7.38
85.83
99.11
188
20
120
i3
0.51
0.42
2.83
2.53
6.79
84.39
97.48
217
21
120
i4
0.50
0.47
2.70
2.59
7.22
85.75
99.24
218
20
120
i5
0.49
0.39
2.76
2.37
6.97
87.02
99.99
196
20
120
i6
0.53
0.39
2.79
2.26
6.91
85.45
98.34
191
21
120
i7
0.54
0.66
2.79
2.26
6.87
85.58
98.70
191
20
120
Mean
215 15 (2σ)
Table 4: Chemical (wt. %) and age data for
uraninite grain 2 from the Felbertal K1 gneiss
(profiles a – i; see Figs. 9 and 10). Single point
age errors are 1 σ. Note that a shorter counting
time (60 s detector live time) was used for
profile a.
569
IN-SITU U–Th–Pb GEOCHRONOMETRY WITH SUBMICRON-SCALE RESOLUTION
GEOLOGICA CARPATHICA
, 2018, 69, 6, 558–572
resolution to analyse micron-sized crystal domains. The Cameca
nano-SIMS, for instance, allegedly can achieve a spatial reso-
lution of 50 nm (Kilburn & Wacey 2014). However, practical
applications of Th–U–Pb dating with nano-SIMS measure-
ments have not yet been made with spot sizes < 2–3 µm (Fayek
et al. 2002; Stern et al. 2005; Koike et al. 2014) as it is metho-
dically difficult, even with the latest nano-SIMS devices.
Thus, there appears to be currently no real alternative to elec-
tron beam dating on such small scales.
So far, the ~1 µm spot size obtained by electron micro-
probe-based dating in the classical 15 kV set-up was the lower
limit for quantitative Th–U–Pb geochronometry. In this study,
we have applied for the first time finer-scale total Pb dating on
the submicron level by means of low-voltage SEM/EDX tech-
niques. We have used so-called standardless EDX analysis for
uraninite dating. This method relies on factory-based intensity
ratios (counts per wt. %) that were initially measured for all
elements at 20 kV on appropriate element standards (see
Table 1). Therefore, the term “standardless” may be somewhat
misleading. These intensity values must be “updated” before
every analytical session by means of a monitor standard.
We did this by measuring Si count rates and background
signals on a silicon wafer.
We know that EDX analysis, especially when run in stan-
dardless mode, is viewed by many workers in the field with
scepticism with regard to whether it can provide sufficient
precision and accuracy for U–Th–total Pb dating. The main
problems are that the factory-based intensity values need to be
extrapolated to other beam conditions (in our case, 8 kV) by
using fundamental parameters, and that intensity ratios
between elements may change when the instrument ages.
However, based on systematic measurements on a number of
reference materials, we propose here that standardless SEM/
EDX analyses with the Oxford INCAEnergy system can pro-
vide geologically meaningful uraninite ages with a reasonable
precision and accuracy. The correctness of the primary stan-
dardization can easily be tested by means of control measure-
ments on synthetic UO
2
and ThO
2
and other element standards.
If these standards give systematically to high or too low
values, this could be reasonably considered with a later
external correction and recalibration. A certain analytical
problem could result from the fact that Th Mα and U Mα lines
overlap in the EDX spectrum (Fig. 2). This involves the risk of
an inaccurate U and Th determination in Th rich uraninite.
However, as mentioned earlier, slight errors in the U and Th
contents will have little effect on the age dates.
Fig. 10. Chemical profiles and total Pb ages recorded in grain 2 (positions of profiles are shown in Fig. 9). Shown error bars are 1 σ. Given
domain ages (bold) are weighted average ages (95 % confidence level).
570
WAITZINGER and FINGER
GEOLOGICA CARPATHICA
, 2018, 69, 6, 558–572
By far the most important factor for successful U–Th–
total Pb uraninite dating is certainly the accuracy of the Pb
determination. Fortunately, the latter can be sensitively con-
trolled with uraninite reference materials, but a variety of such
“age standards” with different compositions is needed to mini-
mize the risk of systematic errors caused by inaccurate peak
interference or the background correction. A minor systematic
error in the age, on the order of a few million years, may still
remain unrecognized though, due to the limited precision of
EDX analysis. Either way, the EDX data quality appears to be
definitely sufficient to study essential features of uraninite
recrystallization and alteration. The special construction of
scanning electron microscopes, directed towards a particularly
high beam stability, is advantageous for high spatial resolution
measurements, as is the ability of EDX detectors to analyse all
elements simultaneously.
A severe limitation of our analytical setup and of the EDX
method, in general, concerns the dating of very young ura-
ninite (< 20 Ma), where radiogenic Pb is close to or even
remains below the detection limit. WDX analysis would cer-
tainly have more potential here. Moreover, WDX analysis
could provide superior precision for U–Th–total Pb uraninite
dating. It seems possible that, with a good WDX setup,
uraninite spot analyses can be made with errors of less than
1 million years. This would theoretically permit high-preci-
sion dating of uraninite. However, due to the danger of small-
scale intra-crystal age heterogeneities, such application would
only make sense in combination with a high spatial analytical
resolution and a very high beam stability, respectively.
Whether electron microprobes can achieve equally good beam
stability behaviour as the state-of-the-art FE-SEMs still needs
to be tested.
Acknowledgements: Noreen Vielreicher is thanked for
stylistic and linguistic improvements of the text. Hans-Jürgen
Förster, Axel Renno and Olaf Tietz donated hand specimens of
uraninite bearing granite types from the Erzgebirge, which
served as valuable uraninite “age standards”. Reinhard
Wagner and Werner Paar provided uraninite crystals from
the Mitterberg ore deposit. Funding in the frame of the program
“Forschungspartnerschaften Mineralrohstoffe” of the Austrian
Geological Survey (GBA) and by the Austrian Science Fund
(P 31901-N29) is gratefully acknowledged. The paper bene-
fited from competent reviews by Daniel Harlov, Bernhard
Schulz and Armin Zeh.
References
Allaz J., Raschke M.B., Persson Ph.M. & Stern Ch.R. 2015: Age,
petrochemistry, and origin of a REE-rich mineralization in the
Longs Peak–St. Vrain batholith, near Jamestown, Colorado
(U.S.A.). Am. Mineral. 100, 2123–2140.
Alexandre P. & Kyser T.K. 2005: Effects of cationic substitutions and
alteration in uraninite, and implications for the dating of uranium
deposits. Can. Mineral. 43, 3, 1005–1017.
Alexandre P., Peterson R. & Joy B. 2016: Sector Zoning In Uraninite.
Can. Mineral. 53, 1–11.
Amelin Y. & Davis W.J. 2006: Isotopic analysis of lead in sub-nano-
gram quantities by TIMS using a
202
Pb–
205
Pb spike. J. Anal. At.
Spectrom. 21, 10, 1053–1061.
Bernhard F., Finger F., Schitter F., Berka R. & Schuster R. 1998:
Electron microprobe ages of monazite and xenotime from the
Austroalpine basement units of the Fischbacher Alpen, Styria,
Austria. Mitt. Österreich. Mineral. Gesellschaft 143, 246–248.
Bowles J.F.W. 1990: Age dating of individual grains of uraninite in
rocks from electron microprobe analyses. Chem. Geol. 83, 1–2,
47–53.
Chipley D., Polito P.A. & Kyser T.K. 2007: Measurement of U–Pb
ages of uraninite and davidite by laser ablation-HR-ICP-MS.
Am. Mineral. 92, 11–12, 1925–1935.
Cliff R.A., Oberli F., Meier M., Droop G.T.R. & Kelly M. 2015:
Syn-metamorphic folding in the Tauern Window, Austria dated
by Th–Pb ages from individual allanite porphyroblasts. J. Meta-
morph. Geol. 33, 4, 427–435.
Cocherie A. & Legendre O. 2007: Potential minerals for determining
U–Th–Pb chemical age using electron microprobe. Lithos 93,
3–4, 288–309.
Cross A., Jaireth S., Rapp R. & Armstrong R. 2011: Reconnais-
sance-style EPMA chemical U–Th–Pb dating of uraninite.
Australian J. Earth Sci. 58, 6, 675–683.
Davis D.W., Krogh T.E. & Williams I.S. 2003: Historical Develop-
ment of Zircon Geochronology. Rev. Mineral. Geochem. 53, 1,
145–181.
Decrée S., Deloule E., Putter T., Dewaele S., Mees F., Yans J. & Mari-
gnac C. 2011: SIMS U–Pb dating of uranium mineralization in
the Katanga Copperbelt: Constraints for the geodynamic con-
text. Ore Geol. Rev. 40, 1, 81–89.
Drouin D., Couture A.R., Joly D., Tastet X., Aimez V. & Gauvin R.
2007: “CASINO V2.42” — A fast and easy-to-use modeling tool
for scanning electron microscopy and microanalysis users. Scan-
ning 29, 92–101.
Eichhorn R., Loth G., Höll R., Finger F., Schermaier A. & Kennedy
A. 2000: Multistage Variscan magmatism in the central Tauern
Window (Austria) unveiled by U/Pb SHRIMP zircon data.
Contrib. Mineral. Petr
ol. 139, 4, 418–435.
Eichhorn R., Schärer U. & Höll R. 1995: Age and evolution of
scheelite-hosting rocks in the Felbertal deposit (Eastern Alps):
U–Pb geochronology of zircon and titanite. Contrib. Mineral.
Petrol. 119, 4, 377–386.
Fayek M., Kyser T.K. & Riciputi L.R. 2002: U and Pb Isotpe analysis
of uranium minerals by ion microprobe and the geochronology
of the McArthur River and Sue Zone uranium deposits,
Saskatchewan, Canada. Can. Mineral. 40, 6, 1553–1570.
Finger F., Krenn E., Schulz B., Harlov D. & Schiller D. 2016: “Satel-
lite monazites” in polymetamorphic basement rocks of the Alps:
Their origin and petrological significance. Am. Mineral. 101, 5,
1094.
Finger F., Waitzinger M., Förster H.J., Kozlik M. & Raith J. 2017:
Identification of discrete hydrothermal events in polymetamor-
phic basement rocks using high spatial resolution U–Th–Pb
chemical dating of uraninite microcrystals. Geology 45,
991–994.
Förster H.J. 1999: The Chemical Composition of Uraninite in
Variscan Granites of the Erzgebirge, Germany. Mineral. Mag.
63, 2, 239–252.
Förster H.J., Rhede D., Stein H.J., Romer R.L. & Tischendorf G.
2012: Paired uraninite and molybdenite dating of the Königshain
granite: implications for the onset of late-Variscan magmatism
in the Lausitz Block. Int. J. Earth Sci. (Geol. Rundsch.) 101,
57–67.
Frasl G. & Frank W. 1966: Einführung in die Geologie und Petrogra-
phie des Penninikums im Tauernfenster. Der Aufschluß, Spec.
Issue 15, Heidelberg, 30–58.
571
IN-SITU U–Th–Pb GEOCHRONOMETRY WITH SUBMICRON-SCALE RESOLUTION
GEOLOGICA CARPATHICA
, 2018, 69, 6, 558–572
Frei D. & Gerdes A. 2009: Precise and accurate in situ U–Pb dating of
zircon with high sample throughput by automated LA-SF-ICP-
MS. Chem. Geol. 261, 3–4, 261–270.
Ge X., Qin M. & Fan G. 2014: Study and Application of Electron
Probe Micro-analysis Dating for Uraninite. Acta Geologica Sini-
ca — English Edition 88, s2, 1345–1346.
Geisler T. & Schleicher H. 2000: Improved U–Th–total Pb dating of
zircons by electron microprobe using a simple new background
modeling procedure and Ca as a chemical criterion of fluid-
induced U–Th–Pb discordance in zircon. Chem. Geol. 163, 1–4,
269–285.
Harlov D.E., Wirth R.
&
Hetherington C.J. 2011
:
Fluid-mediated
partial alteration of monazite: the role of coupled dissolution–
reprecipitation in element redistribution and mass transfer.
Con-
trib. Mineral. Petrol.
162, 329–348.
Holmes A. 1911: The Association of Lead with Uranium in Rock-
Minerals, and Its Application to the Measurement of Geological
Time. Proceedings of the Royal Society of London A: Mathemat-
ical, Physical and Engineering Sciences 85, 578, 248–256.
Hurtado J.M., Chatterjee N., Ramezani J., Hodges K.V. & Bowring
S.A. 2007: Electron Microprobe Chemical Dating of Uraninite
as a Reconnaissance Tool for Leucogranite Geochronology. Na-
ture Precedings, hdl:10101/npre.2007.655.1.
Jercinovic M.J. & Williams M.L. 2005: Analytical perils (and prog-
ress) in electron microprobe trace element analysis applied to
geochronology: background acquisition, interferences, and
beam irradiation effects. Am. Mineral. 90, 526–546.
Jercinovic M.J., Williams M.L., Allaz J. & Donovan J. 2012: Trace
Analysis in EPMA. IOP Conf. Series: Materials Science and En-
gineering 32, 1–22.
Kebede T., Klotzli U., Kosler J. & Torbjorn S. 2005: Understanding
the pre-Variscan and Variscan basement components of the cen-
tral Tauern Window, Eastern Alps (Austria): constraints from
single zircon U–Pb geochronology. Int. J. Earth Sci. 94,
336–353.
Keppler H. & Whyllie P.J. 1990: Role of fluids in transport and frac-
tionation of uranium and thorium in magmatic processes. Nature
348, 531–533.
Kempe U. 2003: Precise electron microprobe age determination in
altered uraninite: consequences on the intrusion age and the
metallogenic significance of the Kirchberg granite (Erzgebirge,
Germany). Contrib. Mineral. Petrol. 145, 1, 107–118.
Kober B. 1986: Whole-grain evaporation for
207
Pb/
206
Pb-age-investi-
gations on single zircons using a double-filament thermal ion
source. Contrib. Mineral. Petrol. 93, 4, 482–490.
Kilburn M.R. & Wacey D. 2014: Nanoscale secondary ion mass spec-
trometry (NanoSIMS) as an analytical tool in the geosciences.
In: Grice K (Ed.): Principles and Practice of Analytical Tech-
niques in Geo sciences. The Royal Society of Chemistry, 1–34.
Koike M., Ota Y., Sano Y., Takahata N. & Sugiura N. 2014: High-
spatial resolution U–Pb dating of phosphate minerals in Martian
meteorite Allan Hills 84001. Geochem. J. 48, 5, 423–431.
Kotzer T.G. & Kyser T.K. 1993: O, U, and Pb isotopic and chemical
variations in uraninite; implications for determining the tempo-
ral and fluid history of ancient terrains. Am. Mineral. 78, 11–12,
1262–1274.
Kozlik M. & Raith J. 2016: Variscan metagranitoids in the central
Tauern Window (Eastern Alps, Austria) and their role in the for-
mation of the Felbertal scheelite deposit. Lithos 278, 303–320.
Kozlik M., Raith J.G. & Gerdes A. 2016: U–Pb, Lu–Hf and trace
element characteristics of zircon from the Felbertal scheelite
deposit (Austria): New constraints on timing and source of W
mineralization. Chem. Geol. 421, 112–126.
Krogh T.E. 1982: Improved accuracy of U-Pb zircon ages by the
creation of more concordant systems using an air abrasion tech-
nique. Geochim. Cosmochim. Acta 46, 4, 637–649.
Ludwig K.R. 2003: User’s manual for Isoplot 3.00: a geochrono-
logical toolkit for Microsoft Excel. Kenneth R. Ludwig, 1–74.
Mattinson J.M. 2005: Zircon U–Pb chemical abrasion (“CA-TIMS”)
method: Combined annealing and multi-step partial dissolution
analysis for improved precision and accuracy of zircon ages.
Chem. Geol. 220, 1–2, 47–66.
Montel J.-M., Foret S., Veschambre M., Nicollet C. & Provost A.
1996: Electron microprobe dating of monazite. Chem. Geol.
131, 1, 37–53.
Montel J.M., Kato T., Enami M., Cocherie A., Finger F., Williams M.
& Jercinovic M. 2017: Electron probe dating of monazite — the
story. Chem. Geol. 484, 4–15.
Naemura K., Hirajima T. & Svojtka M. 2009: The Pressure–Tempera-
ture Path and the Origin of Phlogopite in Spinel–Garnet Perido-
tites from the Blanský Les Massif of the Moldanubian Zone,
Czech Republic. J. Petrol. 50, 10, 1795–1827.
Paar W. & Koeppel V. 1978: The ‘pitchblende-nodule-assemblage’ of
Mitterberg (Salzburg, Austria). Neues Jahrb. Mineral. Abh. 131,
3, 254–271.
Parslow G.R., Bandstaetter F., Kurat G. & Thomas D.J. 1985:
Chemical ages and mobility of U and Th in anatectites of
the Cree Lake zone, Saskatchewan. Can. Mineral. 23, 4,
543–551.
Pestal G., Hejl E., Braunstingl R. & Schuster R. 2009: Erläuterungen
Geologische Karte von Salzburg 1:200,000. Geologische
Bundes anstalt, Wien, 1–162.
Peiffert C., Nguyen Trung C. & Cuney M. 1996: Uranium in granitic
magmas. Part II: experimental determination of uranium solubil-
ity and fluid-melt partition coefficients in the UO
2
-haplogran-
ite-H
2
O-halides system at 720-770°C, 200 MPa. Geochim.
Cosmochim. Acta 60, 1515–1529.
Putnis A. 2002: Mineral replacement reactions: from macroscopic ob-
servations to microscopic mechanisms. Mineral. Mag. 66,
689–708.
Pyle J.M., Spear F.S. & Wark D.A. 2002: Electron Microprobe Ana-
lysis of REE in Apatite, Monazite and Xenotime: Protocols and
Pitfalls. Rev. Mineral. Geochem. 48, 1, 337–362.
Săbău G. 2012: Chemical U–Th–Pb geochronology: a precise explic-
it approximation of the age equation and associated errors.
Geochronometria 39, 167 –179.
Santosh M., Yokoyama K., Biju-Sekhar S. & Rogers J.J.W. 2003:
Multiple Tectonothermal Events in the Granulite Blocks of
Southern India Revealed from EPMA Dating: Implications on
the History of Supercontinents. Gondwana Res. 6, 1, 29–63.
Schaltegger U., Schmitt A.K. & Horstwood M.S.A. 2015: U–Th–Pb
zircon geochronology by ID-TIMS, SIMS, and laser ablation
ICP-MS: Recipes, interpretations, and opportunities. Chem.
Geol. 402, 89–110.
Scherrer N., Engi M., Gnos E., Jakob V. & Liechti A. 2000: Monazite
analysis; from sample preparation to microprobe age dating and
REE quantification. Schweiz. Mineralog. Petrograph. Mitt. 80,
1, 93–105.
Schmid S.M., Fügenschuh B., Kissling E. & Schuster R. 2004:
Tectonic map and overall architecture of the Alpine orogen.
Eclogae Geologicae Helvetiae 97, 1, 93–117.
Schmid S.M., Scharf A., Handy M.R. & Rosenberg C.L. 2013:
The Tauern Window (Eastern Alps, Austria): a new tectonic
map, with cross-sections and a tectonometamorphic synthesis.
Swiss J. Geosci. 106, 1, 1–32.
Schoene B., Guex J., Bartolini A., Schaltegger U. & Blackburn T.J.
2010: Correlating the end-Triassic mass extinction and flood
basalt volcanism at the 100 ka level. Geology 38, 5, 387–390.
Shahin H.A. 2014: Geochemical Characteristics and Chemical Elec-
tron Microprobe U–Pb–Th Dating of Pitchblende Mineralization
from Gabal Gattar Younger Granite, North Eastern Desert,
Egypt. Open Journal of Geology 4, 24–32.
572
WAITZINGER and FINGER
GEOLOGICA CARPATHICA
, 2018, 69, 6, 558–572
Stern R.A., Fletcher I.R., Rasmussen B., McNaughton N.J. &
Griffin B.J. 2005: Ion microprobe (NanoSIMS 50) Pb-isotope
geochronology at <5 μm scale. Int. J. Mass Spectrom. 244, 2–3,
125–134.
Suzuki K., Adachi M. & Tanaka T. 1991: Middle precambrian
provenance of Jurassic sandstone in the Mino Terrane,
central Japan: Th–U-total Pb evidence from an electron
microprobe monazite study. Sediment. Geol. 75, 1–2,
141–147.
Tropper P., Harlov D., Krenn E., Finger F., Rhede D. & Bernhard F.
2007: Zr-bearing minerals as indicators for the polymetamorphic
evolution of the eastern, lower Austroalpine nappes (Stubenberg
Granite contact aureole, Styria, Eastern Alps, Austria). Lithos 95,
1–2, 72–86.
Vesela P., Soellner F., Finger F. & Gerdes A. 2012: Magmato–sedi-
mentary Carboniferous to Jurassic evolution of the western
Tauern window, Eastern Alps (constraints from U–Pb zircon
dating and geochemistry) (vol. 100, pg. 993, 2011). Int. J. Earth
Sci. 101, 2, 603.
von Quadt A. 1992: U–Pb zircon and Sm–Nd geochronology of mafic
and ultramafic rocks from the central part of the Tauern Window
(eastern Alps). Contrib. Mineral. Petrol. 110, 1, 57–67.
Votyakov S.L., Khiller V.V., Shchapova Y.V. & Erokhin Y.V. 2013:
Composition and chemical microprobe dating of U–Th-Bearing
minerals. Part 2. Uraninite, thorite, thorianite, coffinite, and
monazite from the Urals and Siberia. Geol. Ore Deposits 55, 7,
515–524.
Williams I.S. 1998: U–Th–Pb geochronology by ion microprobe. In:
McKibben M.A., Shanks W.C., Ridley W.I. (Eds.): Applications
of Microanalytical Techniques to Understanding Mineralizing
Processes. Reviews in Economic Geology, Littleton, 1–35.
Williams M.L. & Jercinovic M.J. 2002: Microprobe monazite geo-
chronology: putting absolute time into microstructural analysis.
J. Struct. Geol. 24, 6–7, 1013–1028.
Yokoyama K., Shigeoka M., Goto A., Terada K., Hidaka H. &
Tsutsumi Y. 2010: U–Th-total Pb ages of Uraninite and Thorite
from Granititc Rocks in the Japanese Islands. Bull. Natl. Mus.
Nat. Sci. 36, 7–18.