Geochronology II:

U-Th-Pb
Chapter 3

U-Th-Pb
Three radionuclides decaying to 3 isotopes of Pb at different rates:
232Th decays to 208Pb with a half-life of 14 Ga.
238U decays to 206Pb with a half-life of 4.47 Ga.
235U decays to 207Pb with a half-life of 707 Ma.
The remaining isotope, 204Pb, is used for normalization.

Decay is a combination of alphas and betas:

238U: 8 alphas
235U: 7 alphas
232Th: 6 alphas
Energy emitted accounts for ~80% of radioactive energy production in the
Earth (most of the rest is

40

K)

The particular geochronological power comes from from two isotopes

of U, with identical chemical behavior (or nearly so), decaying to 2 Pb
isotopes at very different rates.

Chemistry of U, Th, and Pb

U and Th are actinide series REE. Both U and Th generally have a valence of +4,
but under oxidizing conditions, such as at the surface of the Earth, U has a valence of
+6.
In six-fold coordination, U4+ has an ionic radius of 89 pm and 100 pm in 8 fold (100
pico meters = 1 ). U6+ has an ionic radius of 73 pm in 6-fold and 86 pm in 8-fold
coordination
Th4+ has an ionic radius of 94 pm in 6-fold and 1.05 in 8-fold coordination.

The combination of somewhat large size and high charge makes them highly
incompatible.
Th slightly more incompatible than U.

Both Th and U are refractory, insoluble, and immobile, with the exception of U 6+
(favored at the Earths surface), which forms the UO 42- oxyanion, which is quite
soluble and mobile.

Pb is moderately volatile, chalcophile, and somewhat siderophile. Essentially

always in the 2+ state in nature. Ionic radius of Pb 2+ is 119 pm in 6-fold
coordination and 129 pm in 8-fold coordination. As a result, Pb is moderately
incompatible. Although not particularly soluble, it does seem to form some
soluble complexes and is somewhat mobile.

235

U/

238

The U-Pb system is generally regarded as the most precise, most

valuable dating system.

The power is a consequence of two U isotopes decaying at different

rates to two Pb isotopes.

Since

U and 238U are isotopes of the same element, we can expect

the to behave identically almost.
The canonical 235U/238U ratio is 1/137.88, but recent work suggests a value
235

of 1/137.82. Furthermore, small ( a few per mil) variations in the ratio

have now been demonstrated (resulting from small difference in chemical
behavior).
Mostly, these variation of negligible, but as precision improves, it will
become necessary to actually measure this ratio in each sample precisely.

Bear in mind that the ratio of 1/137.82 is the ratio today: it changes
continually with time as

235

U decays more rapidly.

Typo in book

Pb-Pb Dating
We can write two decay equations:
Pb*

235

206

Pb*

238

Divide one by the other:

207

Pb*

206
Pb*

Assuming

235

U (e235t 1)

207

U (e238t 1)

U (e235t 1)
238
U (e238t 1)
235

U/238U is constant:
Pb*
(e235t 1)

206
Pb* 137.82(e238t 1)
207

Ratio of radiogenic

Pb to 206Pb is proportional to time

(dont need to know U/Pb ratio).
207

Pb-Pb Isochrons
Jargon: geochronologists long ago named the
U/204Pb ratio .

238

Our isochron equation for the

206

For the

U-206Pb system is:

238

Pb/ 204 Pb ( 206 Pb/ 204 Pb)0 (e238t 1)

U207Pb decay, it is:

235

207

Pb/ 204 Pb ( 207 Pb/ 204 Pb)0

Suppose we plot

207

(e235t 1)
137.82

Pb/204Pb vs.

206

on such a plot is:

207 Pb/ 204 Pb
(e235t 1)

206 Pb/ 204 Pb 137.82(e238t 1)

Pb/204Pb. The slope

Pb-Pb Isochrons
On a plot of

Pb/204Pb vs.
206Pb/204Pb, the slope is
proportional to time and
is therefore an isochron.
207

Parent/daughter ratio not

needed.

Intercept does not give

the initial ratios.

Our
equation:(e
207 Pb/ 204 Pb

206

Pb/

204

235t

1)

Pb 137.82(e238t 1)

cannot be solved directly for

t.

Need to use indirect

method.

Figure 3.1

Total Pb Isochrons
The U-Pb system achieves its greatest power when we use the
238

U-206Pb,

235

U-207Pb, and

207

Pb-206Pb methods in combination.

Ideally, all ages should agree. If so, they are said to be

concordant.

Often, age information can still be extracted from non-concordant

systems.

One approach to testing for concordancy is to plot

U/206Pb
against 207Pb/206Pb - the Tera-Wasserburg diagram. Purely
radiogenic Pb will have unique 207Pb*/206Pb* and 238U/206Pb*
ratios at any given time and hence define a concordia curve
on such a diagram:
Pb* (e235t 1)( 238U / 206 Pb*)

206
137.82
Pb*

238

207

Error in
book

Tera-Wasserburg Diagram
Concordant data
containing variable
amounts of
common (i.e.,
initial) Pb will define
a lines with slope:
d(207 Pb/ 206 Pb)

d( 238 U / 206 Pb)

U 235t
(e 1)
238
U

235

207
206

Pb
(e238t 1)
Pb
i

The intercept with

the concordia curve

gives the age. Yintercept gives
initial 207Pb/206Pb.

Figure 3.2

Th/U ratios
We can, of course calculate ages using the
conventional isochron method based on
232Th/208Pb ratios.

Also if the age is known (for example from a PbPb

isochron) we can calculate Th/U ratios for that set
of samples, assuming constant Th/U. On a plot of
208Pb/204Pb vs 206Pb/204Pb, the slope will be given by:
( 208 Pb/ 204 Pb)

( 206 Pb/ 204 Pb)

Th/ 204 Pb(e232t 1) (e232t 1)

t
238
U / 204 Pb(e238t 1)
(e 238 1)

232

where =232Th/238U; BSE 4

Zircon Dating

Figure 3.3

Zircon Dating
Zircon (ZrSiO4) readily accepts U4+ in the Zr4+ site but
excludes Pb. To a first approximation, there will be no initial
or common Pb making it a very attractive target for
dating.

Zircon is also hard (Mohs 7.5) and chemically resistant. It is

a common accessory mineral in many (felsic) igneous and
metamorphic rocks.

Also preserved in coarse grained sediments.

The approach used for zircon dating can apply to other Urich minerals such as titanite (aka sphene, CaTiSiO 5),
monazite ( ThPO4), alanite (Ce2Al3(SiO4)3OH) and apatite
(Ca10(PO4)6(OH)2).

Concordia Diagram
A concordia diagram is a plot of
Pb*/238U vs. 207Pb*/235U, i.e., the
ratios of the number of atoms of
radiogenic daughter produced to the
number of atoms of radioactive
parent. These207are
to
Pb*proportional
t

(e

1)
time, e.g.:
235
206

235

This is, in effect, a plot of the

206

Pb age against the

235

207

U
Pb age.
238

Because only a small correction for

common Pb is necessary for most
zircons, these ratios are readily
calculated, and this diagram is very
often used in zircon dating.

If the ages agree, i.e., if they are

concordant, they will plot on the
concordia curve.
If not, we may still be able to extract
age information.

Thinking about the

concordia diagram
The best way to think
about how the
concordia diagram
works is to imagine
the points staying
fixed and the diagram,
and curve, growing as
time passes.

Imagine a zircon
formed at 3.0 Ga. At
the time of formation,
it would plot at the
origin and the diagram
would look like this.

Three billion years later

If the system
remained closed for
the subsequent 3
billion years, our
zircon would plot at
the 3 Ga point on the
concordia diagram.

Suppose, however, it
is now heated to the
point where Pb can
diffuse out and is lost
completely. How
would our point it
move on the diagram?

Concordia Curves

Now imagine zircons formed at 4 Ga partially losing Pb at

3 Ga. They move toward the origin at that time (left).

When we come along 3 Ga later, we find them on a chord

intersecting the concordia curve at 4 and 3 Ga.

U Loss
U gain essentially mimics
Pb loss in moving the
point toward the origin.

U loss is less common,

but age information can
still be retained.

Continuous Pb loss or Pb
gain (latter is particularly
unlikely) do not take
predictable paths and
destroy age information.

Issues and Solutions

What makes zircon a good clock, namely a high U
content, also tends to destroy that clock through
radiation damage.
Radiation damaged zircons are referred to a metamict.
Damaged zones are most likely to richest in U and to have
lost Pb and be discordant.

Various approaches have been taken effectively remove

metamict areas and analyze remaining zircon.
Abrasion: Krough (1982)
Annealing followed by step-wise dissolution (Mattinson,
2005).

Step-wise Dissolution

Figure 3.18

Step-wise dissolution

Figure 3.9

In situ Analysis

Figure 3.10

U-Th Decay Series Dating

Figure 3.11

Radioactive Equilibrium
The abundance of a nuclide that is both radioactive
and radiogenic (e.g., 230Th) will vary with time
according to:
dN
D

dt

P NP D ND

The r.h.s. is just the difference between its rate of

production and its rate of decay.

This nuclide is radioactive equilibrium if its abundance

does not change with time, i.e., dND/dt = 0.

What does this say about its ratio relative to its

parent?

Thought Experiment
Imagine a hopper with a
spring-loaded door into
which marbles fall.

As weight builds up, the door

opens and marbles fall out. If
weight builds more, the door
opens more until marbles fall
out faster.

Once marbles are falling out

of the hopper as fast as they
are falling in, the position of
the door remains stationary
and the rate of fall of
marbles becomes constant.

Non-equilibrium
When a system is perturbed and forced out of radioactive
equilibrium, it will return to the equilibrium state at a rate that
depends of the initial abundance of parent and daughter and
the two decay constants (assuming a closed system).

This predictable rate of return to equilibrium is the basis of

decay series dating.

No time information (other than a minimum elapsed time since

disturbance) can be obtained from a system at equilibrium.
Equilibrium also erases all previous history.

To know how the daughter nuclide abundance varies with time,

we integrate the previous equation:
ND

P
NPo e Pt e Dt NDoe Dt
D P

Activity
Because traditionally decay-series isotopes were
measured by measuring their radioactive decay
(counters of sorts), and many still are, we use activity
(decays per unit time) in place of abundance. Also
turns out this makes the math easier in many cases.

Activity is simply dN/dt and is related to molar

abundance by the decay constant:
dN
N
dt

We express activity by writing the nuclide or ratio in

parentheses, e.g., (234U) or (230Th/238U).

234

238

U dating

234U is the first reasonably long-lived (t1/2 = 245 ka) daughter

of

238

U.

At equilibrium (234U)=(238U) but in seawater (234U/238U) =

1.145. Why?
Deviations of the (234U/238U) ratio from the equilibrium value (1)
are usually expressed in per mil units and denoted as 234U.
Thus a (234U/238U) value of 1.145 would be expressed as 234U =
145.

Imagine a coral growing from seawater incorporating U into

calcite. We can divide the 234U into part that is supported by
238U and part that is not: the excess above the 238U activity:

(234U) = (234U)s + (234U)u = (238U) + (234U)u

234

The unsupported

238

234

U dating

U will decay according to:

( 234U )u ( 234U)uo e 234t

(234U)uo is just the initial excess of (234U) over (238U)

and because the activity of 238U does not change
measurably on the timescales of interest ( 238U) =
(238U)0 and:
(234 U)uo (234 U)o (238 U )

and
Total (234U) is:

( 234U )U ( 234U )o ( 238U) e 234t

( 234U) ( 238U ) ( 234U )o ( 238U) e 234t

234

Dividing through by the

238
238

U activity, we have:

U
( 238 ) 1
U

234

U dating

o
t
U
1
e 234
238
U

234

Thus we can date a sample provided we know the initial

U/238U
ratio. The application of 234U/238U has been largely restricted to
corals where we can assume the initial ratio is that of seawater.
It is not generally useful for freshwater carbonates because of
uncertainty in the initial activity ratio.
234

Mollusk shells and pelagic biogenic carbonate (e.g., foraminiferal

ooze) often take up U after initial deposition of the carbonate and
death of the organism, thus violating our closed system
assumption. The technique is typically useful up to 4 half-lives.

Th-

230
The

U Dating

238

U-238U technique does not have high temperature applications,

because at high temperature, radiation damage, which is the reason 234U is
removed in weathering more easily than 238U, anneals quite rapidly.
234

Disequilibrium between

Th (the daughter of 234U) and its U parents

provides useful geochronological information in both high- and lowtemperature systems.

We start with:

230

( 230Th)u ( 230Th)uo e 230t

Divide by normalizing isotope: 232Th

Th u

230

Th 230t
e
232
Th
u
230

In this case again we need to know the initial

230

Th/232Th ratio

Th Dating an Mn Nodule

230

Consider an Mn nodule
growing outward and excess
230Th decaying inward. We
assume the nodule grows at a
constant rate such that
z = st.

Substituting z/s for t and a full

expression for (230Th)u, we
have:

Th

232Th

230

Th 230 z/s
e

232
Th

230

U
1 e 230 z/s

Th

238
232

z/s = 0 at surface, so we know

initial ratio and can solve for
z/s the growth rate.

Figure 3.12

Th/U Carbonate Dating

Carbonates concentrate U and exclude Th. This
leads to (230Th/238U) ratios much smaller than 1
(the equilibrium value); indeed, (230Th/238U) in
many modern carbonates approaches 0.

The problem is complicated by the

disequilibrium that will generally exist between
234
U and 238U. The relevant equation is:
Th
1 e 230t
230
234 238
234U
( U / U) 230 234

230

( 230 234 )t

1 ( 234U / 238U) 1 e

Hence we must also measure (234U/238U).

Oldest European Cave

Paintings
Pike et al. (2012)
used th-U dating to
determine the age
of flowstone
coatings over cave
painting in northern
Europe.

The oldest, a simple

red disk, has a
minimum age of
40,800 years.

Modern humans or
Neanderthals?

A W G Pike et al. Science 2012;336:1409-1413

Published by AAAS

Fig. 5 The Panel de las Manos, El Castillo Cave, showing the location of samples O-82
overlaying a negative hand stencil, and O-83 overlaying a large red stippled disk.

A W G Pike et al. Science 2012;336:1409-1413

Published by AAAS

ThU Dating at High

Temperature
Now consider the case of high-temperature systems (e.g., magmas)
where it is reasonable to assume 234U and 238U are in equilibrium
(allowing us to treat 230Th as the direct daughter). Th and U have
different partition coefficients, so melting disturbs the radioactive
equilibrium.

We can derive the following:

Th

232Th

230

Th 230t
e

232
Th

230

U
1 e 230t

232

Th
238

The tricks to this derivation are to make the approximations 230 238 =
230 and e-238t = 1; i.e., assume 238 0; this is the mathematical
equivalent of assuming the activity of 238U does not change with time.

The first term on the right describes the decay of unsupported

while the second term describes the growth of supported

Th.

230

Th

230

Th-U Dating

Th

232Th

230

Th 230t
e

232
Th

230

U
1 e 230t

232

Th
238

Note that this equation has the

form of a straight line in
(230Th/232Th) (238U/232Th)
space, where the first term is
the intercept and (1 e230t) is
the slope; both are function of
t.
The intercept changes with
time.

The plot rotates

Tharound
U the
230

point where

232

Th

238

232

Th

After infinite time, the slope =

1 because lim

Th

232
t
Th

230

U
Th

238

232

Figure 3.13

Th-U Example
This is 230Th238U mineral isochron obtained on a
peralkaline lava from the African Rift Valley in Kenya
(Black et al., 1997) yields an age of 36,2002600 yrs.

However, the eruption of age of this lava is constrained

by stratigraphy and 14C dating to be less than 9,000 yrs.
Other lavas in the area yielded similarly precise, but
anomalously old, U-Th ages.

Black et al. interpreted the ages to reflect the time of

crystallization of the phenocrysts, which in this case
predates eruption by >25,000 yrs.

In most cases concentrations of Th and U are so low that

accurate measurement of Th-U disequilibria on mineral
separates remains challenging.

Some minerals, however, such as zircon, allanite

((Ca,Ce,Y,La)2(Al,Fe+3)3 (SiO4)3(OH)), pyrochlore
(Na,Ca)2Nb2O6(OH,F)), and baddeleyite (ZrO2), concentrate
uranium and/or thorium sufficiently that they can be
analyzed with the ion microprobe and U-Th ages
determined not only on individual grains, but also of
individual zones of mineral grains.

Using ion probe

Th238U dating of allanite from in the

pyroclastic products of the 73 Ka Toba eruption, Vazquez
and Reid (2004) found that allanite cores had crystallized
between 100 and 225 thousand years ago, whereas most
rims had ages identical with error of the eruption age.
230

Figure 3.14

Pa

231

U Dating

235

231Pa (t1/2 = 32.67 ka) is the granddaughter of

U, but the
intervening nuclide 231Th has a sufficiently short half-life that
we can always ignore it. Pa is typically in the 5+ valance
state with ionic radii of 92 pm and partitions readily into the
same igneous and metamorphic accessory minerals that
concentrate U and Th, but like Th, it is excluded from
carbonates. like Th and unlike U, Pa is generally immobile.

Precisely the same principles apply to

238

231

Pa-235U dating as to

U decay series dating:

Pa
231
235t
231t

235 U

231 235

231

And since we can assume 235 to be 0:

Pa
1 e 231t
235 U

231

Pa 231t
e
235
U

231

Pa 231t
e
235
U

231

235

Pa

231
Advantages

Disadvantages

No long-lived
intermediate comparable
to 234U

Abundance of 235U is less

than 1% of that of

238

231

U is less
than a sixth of that of
238
U and thus it decays
more rapidly

231

Pa decays more rapidly

230
Th.

Short half-life means it is

the half-life of

U Dating

235

231

235

Pa is less abundant
than 230Th
Pa decay constant is
less well known than that
of 234U or 230Th
useful over a smaller
range of time.

The value of the 231Pa-235U system, like that of the 207Pb235U one
comes in using it in combination with the 238U decay

Pa*/ U Th*/
Concordia

231

235

230

234

Figure 3.15

Concordia diagram analogous to the Pb one, defining concordant

ages.

Black line is for (234U/238U)=1; red line is for (234U/238U)=1.145

Ages of corals from Araki Island shown (Chiu et al., 2006)

Ra dating

226
226Ra t1/2 = 1600 yrs
No long-lived daughter isotope to

ratio it to, so ratioed to Ba instead.

226Ra/Ba230Th/Ba diagram is an
isochron diagram analogous to the
230Th/232Th238U/232Th one

Figure shows

Ra-230Th whole rock

isochron on obtained on a
sequence of trachytic lavas from
Longonot volcano in the Gregory
Rift of Kenya. The samples
represent roughly 5 km 3 of lava.
The degree of igneous fractionation
increases upward in the sequence,
resulting in a range of whole rock
compositions and Th/Ba ratios. The
age of 4274 years dates this
igneous differentiation and falls
within the range of eruption ages,
5650 to 3280 years BP, inferred
from 14C dating.
226

Figure 3.16

Pb Dating

210
t1/2 22 yrs.

6th great granddaughter of

Ra. One of these

intermediates is 222Rn (t1/2 3 days), a noble gas,
which can diffuse into the atmosphere before
decaying.

Essentially, this means

226

Pb can take a different

path into the hydrologic system than stable Pb
isotopes.
210

That leads to excess, or unsupported,

young sediments.

(210 Pb)u (210 Pb)u0 e t

210

Pb in

Compaction
For detrital sedimentation (as opposed to an Mn
nodule), we need to consider compaction.

To correct for compaction, we replace depth, z,

with a function called mass-depth, which has
units of depth per unit area. We define mi as
the mass of sediment in the core in depth
interval zi and the mass-depth, mi, at depth zi
as the total mass (per cross-sectional area of the
core) above zi.

Models of

210

Pb accumulation

Constant Flux or Constant Rate of Supply Model

Rate of fallout of

Pb constant and independent of

sedimentation rate.
210

Constant Initial Concentration or Constant Activity Model.

210Pb enters sediment absorbed onto particles accumulation
proportional to sedimentation rate; initial concentration is
constant.

Constant Sedimentation Model

Sedimentation rate is constant, but

Pb accumulation rate is not

210

Constant FluxConstant Sedimentation Model

Both sedimentation and Pb accumulation rate is constant.
Simplest assumption.

Pb Dating

210

1 (210 Pb)0
ti ln 210
( Pb)i

Constant Activity Model

1 A(0)
ti ln
A(i )

Constant Flux Model

A(i) is the integrated

Pb activity in the sediment

column beneath mass-depth i, A(0) is total
integrated activity in column.
210

Constant FluxConstant Sedimentation Model

(210 Pb) (210 Pb)0 e 210 m/r

and

ln( 210 Pb)i ln( 210 Pb)o

210
mi
r

Example of constant flux-constant

sedimentation model
Tehuantepec River
(Mexico) estuary

Changes at 1930
and 1972 due to
land use and
demographic
changes and
channelization

0.66 g/cm2-yr
0.22 g/cm2-yr

0.1 g/cm2-yr

Figure 3.17

210
210Pb decays (via

210

Po Dating

Bi) to

210

Po; t1/2

= 138 days

Po is volatile and is lost

quantitatively (or nearly so) from
lava upon eruption. It then returns
to equilibrium activity through:

210
210

Po

Pb

210 t

1 e

Po

210
210

Po

Pb

210 t

Po

Assume (210Po/210Pb)0 is zero (this

gives maximum age).

Rubin used this to date EPR lavas

collected after evidence of
eruption. Concluded eruption
occurred only months before
eruption

Figure 3.18